[Back]

1 article(s) will be saved. To save, please use your browser's save option. Be sure to save as a plain text file (.txt) or an HTML file.

Record: 11
1816080091416919970601
Title: Rape in the criminal justice system.
Subject(s): RAPE -- Law & legislation
Source: Journal of Criminal Law & Criminology, Summer97, Vol. 87 Issue 4, p1194, 191p
Author(s): Bryden, David P.; Lengnick, Sonja
Abstract: Discusses heterosexual rapes while proposing reform of the original rape law from rape scholars. Reference to author Hans Zeisel book `The Limits of Law Enforcement, which observed that pervasive patterns of felony attrition plague all criminal justice systems; Attrition in rape case processing; Information on a Justice Department study of urban jurisdictions which found that rape and aggravated assault had comparable attrition patterns.
AN: 181608
ISSN: 0091-4169
Full Text Word Count: 97103
Database: Academic Search Premier

RAPE IN THE CRIMINAL JUSTICE SYSTEM

TABLE OF CONTENTS
  I. Introduction ....................................... 1195
 II. From Report to Verdict ............................. 1201
     A. Attrition Statistics and the Equivalent Crimes
       Fallacy .......................................... 1208
       1. Attrition in Rape Case Processing ............. 1210
       2. Rape v. Other Major Crimes .................... 1212
       3. Focusing on Acquaintance Rape ................. 1214
     B. Reporting ....................................... 1218
     C. The Police Founding Decision .................... 1230
     D. The Decision to Prosecute ....................... 1246
     E. The Jury ........................................ 1254
       1. Kalven and Zeisel's Classic Jury Study ........ 1255
       2. Other Research on Victims and Situational
          Variables ..................................... 1263
          a. The Defendant's Characteristics ............ 1274
          b. Jurors' Characteristics .................... 1278
          c.  The Evidence .............................. 1282
       3. Effects of Rape Law Reforms on Conviction
          Rates ......................................... 1283
III. Leniency in Acquaintance Rape Cases: Three Pervasive
     Issues ............................................. 1294
     A. Introduction .................................... 1294
     B. False Rape Reports .............................. 1295
        1. The Argument from Rationality ................ 1298
        2. The Argument From Underreporting ............. 1303
        3. The Argument from Unfounding Rates ........... 1303
        4. Other Empirical Research ..................... 1308
     C. The Burden of Proof ............................. 1315
     D. Victim Behavior as Ambiprobative Evidence ....... 1328
        1. Contributory Negligence ...................... 1333
        2. Alcohol and Drugs ............................ 1347
        3. Promiscuity and Prostitution ................. 1351
        4. Hitchhiking .................................. 1365
        5. Delinquency and Delinquent Peers ............. 1366
        6. Mental Problems, Runaways, and Truants ....... 1367
        7. Prior Intimacy with the Defendant ............ 1369
        8. Summary ...................................... 1373
 IV. Conclusion ......................................... 1377

I. INTRODUCTION

Modern rape scholarship has been informed by a number of empirical premises concerning the operation of the criminal justice system in rape cases. The most fundamental of these premises is that the justice system discriminates, at every stage, against rape victims.[1] The details of this charge can be briefly summarized. To begin, the case attrition rate in rape cases is shockingly high, and very few rapists are convicted of the crime.[2] Victims often do not report the rape, largely because they fear overbearing, hostile police,[3] and--should a trial ensue-vicious attacks on their character.[4] Although false reports of rape are no more common than of other crimes,[5] justice system officials are highly skeptical of women who claim to have been raped by acquaintances.[6] If the rape victim's conduct prior to the crime violated traditional sex-role norms, police commonly disbelieve her report or blame her for the rape.[7] Thus, officials deny justice to women who have engaged in nonmarital sex,[8] or other "improper" activities such as heavy drinking or hitchhiking.[9] None of these sex-role-norm violations is relevant to whether the woman was raped,[10] but the norms are enforced because they serve to keep women in their place[11] and because the men who control the justice system are irrationally obsessed with the danger of false rape accusations.[12]

Afraid that losing cases will look bad on their records, prosecutors are excessively reluctant to prosecute acquaintance rapists.[13] When they do prosecute, the system puts the victim rather than the defendant on trial.[14] Juries, motivated by the same biases as other participants in the system.[15] often blame the victim and acquit the rapist.[16]

Most rape scholars believe that, in large measure, these travesties of justice have been due to rules of law, fashioned by male judges over the centuries, that promote victim blaming.[17] Among the foremost such rules were the requirements that the victim resist her attacker[18] and that her testimony be corroborated by other evidence.[19] The "cautionary instruction,"[20] warning the jury about the danger of false rape accusations, has been another impediment to justice in rape cases[21] According to some scholars, the very name of the crime "rape" has perpetuated subtly harmful myths such as the idea that the perpetrator's motivation is sexual rather than violent.[22] Worst of all, the rule allowing testimony and cross-examination about the woman's sexual habits[23] distracted the jury's attention from the defendant's behavior, and often led to unjust acquittals.[24]

Most scholars have contended that, although attitudinal changes are also necessary, law reforms can play a major role in eliminating discrimination against victims.[25] By changing the name of the crime to "sexual assault" or some similar term,[26] legislatures can help to make the public aware, that the crime is motivated by a desire to dominate women, not sexual hunger, thus reducing public sympathy for the perpetrator.[27] By abolishing rules of rape law that embody sexist assumptions, courts and legislatures can encourage victims to report the crime, combat anti-victim stereotypes that pervade officialdom, and empower prosecutors with the legal tools to secure convictions.[28] Not every rape scholar has explicitly subscribed to every count of this indictment, but its main propositions have been the conventional wisdom.

Since the 1970s, most states have responded to this critique by enacting several reforms of their rape laws.[29] As one commentator notes, "[t]hese changes represent a shift away from laws containing misogynist assumptions and reflecting societal skepticism about the seriousness of rape and the veracity of women's accusations."[30] For example, the corroboration requirement is gone;[31] many states have eliminated the cautionary instruction;[32] and rape shield laws, with varying degrees of success, have restricted inquiries into the victim's sexual history.[33] In many states, rape has been redefined as "sexual assault" or some similar term, in an effort to emphasize that rape is a crime of violence and not due to an uncontrollable sexual passion.[34]

These reforms had multiple purposes. In part, they were designed to alleviate victims' ordeals during rape investigations and trials.[35] Another purpose was symbolic and educational: to abolish rules that were thought to embody sexist assumptions.[36] The reformers' main goals, however, seem to have been instrumental: to encourage reporting of rapes, and to facilitate prosecution of the perpetrators.[37]

There is growing evidence that, while the performance of the justice system in rape cases may have improved,[38] the legal reforms have generally had little or no effect on the outcomes of rape cases, or the proportions of rapists who are prosecuted and convicted.[39] What explains this failure? Some scholars believe that the original rape law reforms were too modest. These scholars have proposed new reforms, a process that continues to this day. Among these "second wave" reforms have been proposals to base rape liability on negligence[40] or even strict liability;[41] to redefine the crime as non-consensual sex, eliminating "force" from the definition;[42] and to require that the woman's consent be affirmatively given in advance of the sexual act.[43] So far, most of these proposals have not been widely adopted.

The purpose of this article is not to offer new ideas for improving rape law. We believe reforms need to be grounded in accurate perceptions of social reality, including a realistic understanding of the reasons why criminals escape justice and an awareness of the limitations of legal doctrines as instruments for changing case outcomes. Accordingly, our aim is to lay a foundation for discussion of reforms by reexamining the two central tenets of the rape law reform movement: (1) that discrimination against rape victims pervades the criminal justice system, and is the main reason why few rapists are punished; and (2) that the outcomes of forcible rape cases can be substantially changed by abolishing rules of law that are thought to foster such discrimination. In our judgment, a thorough analysis of these assumptions is a prerequisite to intelligent evaluation of new proposals for rape law reform.

Like most rape scholars, we will confine our discussion to forcible, heterosexual rapes.[44] We will not consider sentencing practices, a complex topic that would lead us far afield.

Does the social-scientific evidence vindicate or refute claims of bias against victims? In Part I we discuss this question in specific institutional settings, beginning with reporting of rape and continuing through evaluation of rape complaints by police, prosecutors, and juries. If anti-victim biases are indeed prevalent, have law reforms mitigated the problem? When objectionable laws are repealed or modified, do more victims report rapes, do prosecutors file more charges, and are juries more willing to convict? If not, why not? In answering these questions, we will analyze data from the latest social scientific research.

In Part II, we consider the problem of bias from a different perspective, focusing on three pervasive questions that recur in every institutional context:

1. How common are false reports of rape? Most academic discussions of this subject are cursory and tendentious. The prevalent assumption is that only a negligible proportion of rape reports are false. On that assumption, skepticism towards putative rape victims is hardly ever justifiable. But does the assumption rest on a solid empirical foundation? Here again we analyze social-scientific findings that have not yet been discussed by legal scholars.

2. To what extent is the difficulty of securing convictions in acquaintance rape cases due to the prosecution's burden of proof?. In other words, how often is leniency towards accused rapists attributable to genuinely reasonable doubts about the defendant's guilt, rather than unfair biases of police, prosecutors, and jurors? This difficult issue, though almost totally ignored in the rape literature, is central to an appraisal of the functioning of the criminal justice system in rape cases--not only the fact-finder's decision, but also the decisions of police and prosecutors, who often decline to pursue cases that they regard as unprovable.[45]

3. In an acquaintance rape case, does evidence that discredits the alleged victim's character have genuine probative value in determining whether the sexual encounter was consensual? Many scholars have discussed the relevance of evidence about the complainant's sexual history, particularly in connection with rape shield laws.[46] We will broaden the inquiry to include other types of evidence commonly used to discredit rape victims' characters.

Observers agree, for example, that juries tend to acquit men accused of raping women who violate sex-role norms by engaging in casual sex.[47] Is this because jurors are unfairly biased against such women, or because in such cases the defendant's claim that the woman consented is often plausible enough to raise a truly reasonable doubt?

Similarly, rape scholars report that, if the defendant and his accuser had previously been lovers, juries are extremely reluctant to convict him.[48] Is this because jurors harbor prejudices against nonmarital sex, or is it because in this type of case the defendant's version of events--seduction rather than rape--is more likely to be plausible? Our discussion of these issues will introduce the concept of "ambiprobative evidence," which, we believe, sheds new light on all of the major types of evidence used to discredit rape victims, from promiscuity to mental instability to hitchhiking or drunkenness.

Assuming that evidence of the woman's "misconduct" is irrelevant to the defendant's guilt, we conclude by considering whether such information can be kept from the jurY. Can law reforms force witnesses and juries to focus on the man's conduct rather than the woman's character?

We are not so naive as to suppose that all of these questions can be answered unequivocally, definitively, and to every reader's satisfaction. Our relatively modest ambition is to promote a more critical and nuanced appraisal of the empirical foundations of rape law reforms.

II. FROM REPORT TO VERDICT

The social-scientific and anecdotal evidence concerning "victim blaming" in rape cases distinguishes sharply between acquaintance rapes and stranger rapes. This is not a legal distinction: typically, modern statutes define forcible rape as nonconsensual sexual penetration obtained by physical force, or by threat of bodily harm, or when the victim is incapable of giving consent.[49] In this definition, and in most of the other rules affecting rape cases, no line is drawn between rapes by strangers and rapes by acquaintances, or between rapes with and without a weapon.[50] But rape scholars agree, with impressive unanimity,[51] that although forcible rape is nominally a single crime, for practical purposes it is two crimes. This is because the actors in the criminal justice system--police, prosecutors, juries, judges and even victims themselves--tend to be sympathetic towards some types of rape victims but skeptical towards others.[52] The traditional image of a rapist is a knife-wielding stranger[53] If the rapist conforms to this image, by being a stranger, or by inflicting some serious additional physical injury on the woman,[54] the public generally abhors the crime and sympathizes with the woman.[55] But most rapes are perpetrated by acquaintances of the victim: lovers, dates, coworkers, neighbors, relatives, and so on.[56] The rapist usually does not physically injure his victim.[57] In these typical cases, attention usually focuses on the woman's character[58] If her pre-rape behavior violated traditional norms of female prudence or morality,[59] many people blame her instead of the rapist.

A complete stranger cannot plausibly claim that his victim consented. Consequently, a stranger rape defendant almost always concedes that his accuser was raped; his usual defense is that she honestly misidentified him as the rapist.[60] He has no acute need, therefore, to portray her as vindictive, unbalanced, immoral, or dishonest. Because he denies that he had intercourse with the victim, he has no occasion to persuade the jury that she "led him on" or "was asking for it" or was promiscuous or failed to resist his advances. In a stranger rape, the possibility that the parties misunderstood each other's signals does not arise. As a result, the woman's character and all the controversial issues of appropriate sex roles and behavior in dating situations ordinarily are not issues. Toward the victim of a stranger rape, the public usually feels compassion, with a correspondingly severe attitude toward the rapist.[61]

In acquaintance rape cases, by definition, the parties knew each other beforehand, at least slightly. Consequently, the defense usually is consent rather than misidentification.[62] With a consent defense, the woman's character is inevitably a critical issue, because if their sexual encounter was consensual, then her story of rape must have been a fabrication or at least an exaggeration. In acquaintance rape cases the defendant often seeks to portray the woman as immoral, thus "putting the victim on trial."[63] Even if the evidence suggests an honest misunderstanding between the parties, the question of who was more to blame for the misunderstanding will then arise, and impressions of the woman's character probably will affect the jury's judgment on this issue.

Most proposals to reform rape law, such as rape shield laws, have been designed to improve the prosecution's chances in acquaintance rape cases, and to lessen the victim's ordeal by restricting inquiries into her relationships with other men.[64] From the prosecution's point of view, acquaintance rape cases are most difficult to win if the woman either had engaged in consensual sex with the defendant at some time before the alleged rape,[65] or behaved in ways that violated traditional norms of female propriety.[66] Common examples include being a prostitute,[67] or promiscuous,[68] or a truant or a runaway,[69] frequenting singles bars,[70] hitchhiking,[71] using drugs,[72] drinking heavily,[73] or wearing sexy attire.[74]

What is the origin of this prejudice against "bad victims"[75] who are raped by acquaintances? One theory is that a male ideology pervades the criminal justice system.[76] According to this view, the respectable males who dominate the system--police, prosecutors, jurors and judges--are sincerely horrified by stranger rapes, and rapes involving knives or guns.[77] They also abhor rapes that occur in "inappropriate" relationships such as incest.[78] In such cases, the rape scenario does not resemble ordinary, socially acceptable sexual relations.

In contrast, respectable men empathize with men who are accused of having raped a woman on a date.[79] The respectable man tends to identify with the male in a dating situation, and he can easily imagine being falsely accused of rape, or irresistibly provoked by a "loose woman." Accordingly, he wishes to ensure that "normal" aggressive male sexuality is protected by the law.[80]

Some authors contend that men, irrationally fearing false rape accusations, are deeply suspicious of women who claim to have been raped by an acquaintance.[81] Men (and women who have adopted the male ideology) often express this suspicion by blaming the female victim of the rape rather than the male perpetrator.[82] For example, some used to claim that it is impossible to rape a woman against her will.[83] According to another rape myth, all women want to be raped;[84] women who decline men's sexual advances are merely being coy: consciously or not, they secretly crave sex.[85] An equally venerable, and probably much more widespread, idea is that rapes are caused by a sudden, uncontrollable explosion of male desire, ignited by the behavior of a provocatively sexy woman.[86]

Some authors note that the effect of all this victim blaming is to preserve male sexual access to women, unencumbered by fear of being punished for rape.[87]

As modern rape scholarship proliferated, it became, in one scholar's words, "increasingly clear that rape victims were systematically subjected to institutionalized sexism, which began with their treatment by the police, continued through the legal system, much influenced by notions of victim precipitation, and ended with the acquittal of many de facto rapists."[88]

We now evaluate the validity of this meta-theory as an explanation of case attrition in rape cases, beginning with attrition statistics and then turning to the individual stages of the attrition process: the victim's report; police "founding" of the report; filing of charges by the prosecutor; and finally, the trial.

A. ATTRITION STATISTICS AND THE EQUIVALENT CRIMES FALLACY

The average citizen may suppose that when a serious crime occurs the victim reports it to the police, who verify the report, try to identify and arrest the perpetrator, and turn the case over to a prosecutor. The prosecutor files charges and usually proves the accused's guilt at a trial. The judge then sentences the defendant to an appropriately lengthy imprisonment, taking due account of the severity of the crime.

In practice, the system does not work that way? The justice system has been likened to a giant sieve, filtering out cases at every stage of the process,[90] Many victims do not report crimes, especially those committed by acquaintances.[91] Even if the victim reports the crime, the perpetrator usually is not convicted.[92] Sometimes the police cannot discover what they regard as sufficient corroborative evidence of his guilt,[93] Or the victim may decide, sooner or later, not to cooperate with the police, and the case may be dismissed for that reason.[94]

Even if the police believe the victim's report, they often fail to arrest the perpetrator--for instance, a stranger rapist--because they cannot identify him.[95] If they do arrest someone, the prosecutor's office may decide, for any of several reasons, not to file charges.[96] The crime may be minor and prosecutors may be preoccupied with cases they perceive as more important. A noncriminal solution, such as restitution of stolen property, may be worked out. In many cases the prosecutor decides that the evidence of guilt is not strong enough to persuade a jury beyond a reasonable doubt. To save the time and expense of a trial, and to ensure a conviction, the prosecutor often agrees to reduced charges (for instance, indecent assault instead of rape) or a light sentence in return for a guilty plea. Frequently, however, the case is so weak that the prosecutor, lacking the requisite bargaining power for a deal with the accused offender, simply does not file charges.

The relatively few cases that proceed to trial often end with an acquittal[97] or a hung jury. Even if the defendant is convicted, he may receive a suspended sentence. If he goes to prison, the chances are that the time served, once deductions are made for parole and good behavior, will strike the average citizen as much too short.[98]

In The Limits of Law Enforcement, Hans Zeisel observed that pervasive patterns of felony attrition plague all criminal justice systems.[99] In a single year in New York City, for example, of every 1000 "index felonies," only 54% were reported to police. Of these, only sixty-five led to an arrest, thirty-six resulted in a conviction, seventeen resulted in a custody sentence, and three criminals were incarcerated for over a year.[100] Although precise figures vary, a similar pattern of sharp attrition holds across different jurisdictions, countries and eras: 46 to 60% of felony arrests will result in a pretrial dismissal at some level of the criminal justice system.[101]

During the past two decades, many rape scholars have lamented the exceedingly high rate of attrition in rape cases, especially acquaintance cases. According to one authority, "the likelihood of a rape complaint actually ending in conviction is generally estimated at 25%."[102] Although the phenomenon of case attrition transcends crimes and even national boundaries,[103] many scholars believe that the problem is more acute in rape cases[104] and conclude that this is chiefly due to systemic discrimination against acquaintance rape complainants,[105] Analysis of this hypothesis has proceeded at several levels of sophistication.

I. Attrition in Rape Case Processing

Many commentators simply cite rape attrition statistics to buttress their contention that the justice system, at every stage, is hostile to claims of rape.[106] The attrition statistics for rape are indeed striking. Recent crime-victim survey data suggest that each year an estimated 500,000 women are victims of some form of rape or sexual assault.[107]

Yet in 1994, only 102,096 rapes were reported to authorities, and ultimately there were only an estimated 36,610 arrests for forcible rape.[108] Drawing on data from several jurisdictions, the Senate Judiciary Committee reported that 98% of rape victims "never see their attacker caught, tried and imprisoned."[109] Studies of individual jurisdictions have also found patterns of sharp case attrition for rape.[110]

Since the 1970s, such figures have led advocates and scholars to conclude that the criminal justice system discriminates against rape victims. The reluctance of victims to report and prosecute rape, they argue, can be attributed to "institutionalized sexism,"[111] which pervades a system hostile to rape complainants. Two scholars, discussing why increased reporting has not led to more rape convictions, conclude that while "[t] he reasons for this tremendous lack of convictions are complex, it is clear that a major factor is related to the judgmental policies of the police, prosecuting attorneys and the juries."[112] The Senate Judiciary Committee warned that low rates of reporting and processing of rape "provide dramatic testimony of the power of our stereotypes of crime--how these stereotypes distort our understanding of violence against women and deprive individuals of the equal protection of our laws."[113]

While these conclusions may be accurate, they are not self-evident from the attrition statistics. For the sake of illustration, suppose that of every 100 rapes, thirty are not reported for personal reasons that have nothing to do with anticipated official bias. In another fifteen cases the perpetrator was a stranger who cannot be identified. Assume that two false rape reports have been made,[114] and that in twenty cases the complainant, though honest, ultimately decided not to press charges,[115] Finally, suppose that in twenty-eight cases the rapist could not be proved guilty beyond a reasonable doubt to the satisfaction of an impartial factfinder.[116] In that event, if only five of every 100 rapes lead to criminal convictions, the justice system may be working perfectly, screening out false reports plus a much larger number of cases in which either the victim does not cooperate, or the perpetrator's identity is unknown, or his guilt cannot be proven.

But, of course, attrition statistics also are amenable to less sanguine interpretations. For example, rape victims might have decided to remain silent about the crime because they feared official mistreatment. Also, the system might have screened out some of the truthful complaints rather than the false ones. Detectives might have persuaded truthful complainants that it would not be in their interest to undergo the rigors of a trial.[117] Police or prosecutors might have brushed aside truthful accusations by women of whom they disapproved, or against ex-husbands or boyfriends,[118] while perhaps mistakenly believing some honest misidentifications in stranger rape cases.[119] And, of course, cases in which officials or juries decided that the crime did not occur, or was not proven beyond a reasonable doubt, may have been real rapes, amply proven. By themselves, rape attrition statistics do not provide a basis for excluding any of these hypotheses.

2. Rape v. Other Major Crimes

Some scholars have tried to improve the statistical analysis by comparing the attrition pattern for rape with that for comparable felonies. While these studies confirm that few rapes result in arrest or conviction, they also indicate that the same is true of most major felonies. For example, Galvin and Polk found that, in California, rape was treated comparably to other felonies by police, prosecutors, and judges.[120] They concluded that, "[w]hile probably somewhat more rape cases experience attrition in the justice system than is true for homicide, rape seems to show a level of loss comparable to robbery at most points; and certainly lower levels of loss than assault or burglary. These findings hold across time and jurisdictions."[121]

Myers and LaFree reached a similar conclusion in their study of felony prosecutions in Indiana.[122] Controlling for case differences such as evidentiary strength and crime seriousness, they found few differences in the processing of sexual assaults or in officials' reactions to rape complainants.[123] The differences that did exist were explained, in their view, not by victim characteristics but by the availability of evidence of legal guilt.[124]

A Justice Department study of thirty urban jurisdictions found that rape and aggravated assault had comparable attrition patterns.[125] Like Myers and LaFree, the study concluded that prosecutors generally decide whether to prosecute a rape case by the same criterion that they employ in nonsexual assault cases: the likelihood of obtaining a conviction.[126] Similarly, when Steffensmeier compared the disposition of rape cases and other felony offenses across a wide range of jurisdictions, he found that processing and conviction rates for rape were similar to those of other major felonies.[127]

Another similarity between case processing in rape and other crimes is the role of the victim-complainant in determining the outcome. Several studies have pointed to victims' unwillingness to cooperate as a major cause of attrition in rape cases.[128] This seems to mean that the attrition is voluntary, and thus not attributable to systemic bias; but perhaps officials sometimes discourage rape victims from pursuing meritorious complaints. Be that as it may, the phenomenon, like other sources of attrition, is not unique to rape. Analyzing national attrition data, Steffensmeier concluded that victim non-cooperation was the single most important reason for dismissals of all felonies except homicide,[129] a result consistent with other studies.[130] The percentage of dismissals due to victim withdrawal was 51% for rape, compared to 47% for robbery and 64% for nonsexual assault.[131]

3. Focusing on Acquaintance Rape

Some authors argue that attrition statistics, which do not distinguish between stranger and acquaintance rape, mask official bias against victims of acquaintance rape. When separate tabulations are made for different types of rape, the picture changes, often dramatically.[132] A study of Travis County, Texas, found that the probability of indictment was much higher in stranger rape cases (58%) than in acquaintance rape cases (29%).[133] A study of felony filings in Hawaii reached similar conclusions.[134] Of cases dismissed, 50% involved acquaintances, 38% persons who had just met, and only 12% strangers.[135] Of those cases in which felonies were filed, only 28% involved acquaintances, 20% persons who had just met, and 49% strangers.[136]

The single most important reason why most rapists are not punished is the failure of victims to report the crime to the police, or their later refusal to cooperate as a prosecution witness.[137] Reported and prosecuted rape cases are disproportionately stranger rapes, cases in which the rapist inflicted additional injuries, and other "strong cases."[138] Therefore, one might argue that the attrition rate for reported rapes should be lower than for other felonies; an equal attrition rate does not show equal treatment by justice system officials.

Lizotte analyzed National Crime Victim Survey (NCVS)[139] data on rape and assault and concluded that "factors which make a strong case for prosecution are more powerful predictors of reporting rape than reporting assault."[140] That is, rape victims tend to report to police only cases where there is strong evidence supporting conviction, while a wider variety of nonsexual assault cases are reported. Prosecuted rape cases are stronger, relative to all rapes (reported or not) than are prosecuted assaults, relative to all assaults,[141] Lizotte concludes from this that since only the most serious rapes are reported, studies showing comparable attrition rates for rape and other crimes do not rebut the theory that the system discriminates against rape complainants.[142]

Lizotte's data do not answer a key question: whether rape victims' tendency to report only strong cases is because they are discouraged by some potentially reformable aspect of the system, or because for reasons of their own, which may or may not include unfair public attitudes toward rape victims, they do not wish to seek legal redress.[143] Nor does his study reveal whether a "strong" acquaintance rape case involves as much powerful evidence of guilt, on average, as a "strong" nonsexual assault case. For these reasons, his findings are ultimately ambiguous.

The difference in case attrition between crimes committed by strangers and crimes committed by acquaintances is not unique to rape. In a comprehensive study of decision making in the criminal justice system, Gottfredson and Gottfredson conclude that three factors play a critical role in all phrases of felony case processing: the seriousness of the offense; the prior criminal conduct of the defendant; and the existence of a prior relationship between the victim and the defendant.[144] As the authors explain, the distinction between stranger and nonstranger offenders heavily influences every major criminal justice decision: Nearly every decision-maker in the process seeks alternatives [to prosecution] for criminal acts between relatives, friends and acquaintances. The most grave dispositions are reserved continuously for events between strangers. Victims report non-stranger events less frequently, police arrest less frequently, prosecutors charge less frequently, and so on through the system.[145]

This disparity in the legal processing of stranger and acquaintance offenders has been confirmed by numerous studies.[146] These studies suggest that acquaintance offenses are often perceived by decision-makers as inappropriate for criminal sanctions.[147] Albonetti found, for example, that prosecutors resist pursuing acquaintance cases for two reasons: first, a prior relationship between a victim and offender increases ambiguity about whether or not a crime has occurred; second, prosecutors perceive an increased risk that the victim will not cooperate in prosecution of acquaintance offenses.[148] This perception appears to be accurate. Hans Zeisel, analyzing data from a Vera Institute study, found that the factor most often correlated with a victim's withdrawal from prosecution was a prior relationship between the victim and offender.[149] Conceivably, this reluctance to press charges is partly due to systemic leniency toward violence by acquaintances, but in any event the phenomenon is not unique to rape.

Victim survey figures suggest that any general bias against processing acquaintance crimes falls disproportionately on female victims. In a 1987 comparison of violence by strangers and acquaintances, the Bureau of Justice statistics found that 70% of violent crimes committed by strangers, crimes which are pursued more vigorously by the system, were perpetrated against male victims.[150] National victim survey data indicate that, in 1993, 64% of all violent crimes against male victims were perpetrated by strangers; however, where the victim was female, only 38% were by strangers.[151] In that year, according to the victim survey, only 21.2% of rapes were committed by total strangers, the lowest percentage for any violent crime, compared with 80.1% of robberies and 57.7% of aggravated assaults.[152]

At this juncture, one might wish for a study comparing the attrition rates of acquaintance rape with the corresponding rates for other felonies committed by acquaintances of the victim. In a study making such a direct comparison, the results were mixed. Bachman and Paternoster analyzed data from several national surveys, in an attempt to measure the impact of rape law reforms on all phases of case processing from (1979 to 1991.[153] As expected, they found, in the pre-rape law-reform years, an "acquaintance discount" in processing outcomes for all crimes;[154] this discount was greater for rape.[155] More perplexing, however, is their finding that in the post-reform period this acquaintance discount almost disappeared for crimes other than rape, but at 18% was virtually unchanged for rape.[156]

Bachman and Paternoster's results are useful, but even the most sophisticated analysis of available data cannot determine whether differences in attrition are due to systemic bias. All such efforts rest on what we call the Equivalent Crimes Fallacy: the dubious assumption that, in the absence of official bias, crimes of comparable severity will have similar attrition rates. There is no reason to assume that the attrition rates of various crimes, even of various crimes by acquaintances, ought to be equal. Crimes differ in the rates at which they are reported, the typical amounts of corroborative evidence, and so on. Some of these differences can be quantified only in a crude fashion. No one has even tried, for example, to quantify and compare the frequency with which rape complainants and nonsexual assault complainants change their stories during police interrogation.

In short, attrition statistics are no substitute for the difficult and inevitably largely subjective task of trying to determine, with respect to each stage of case processing, whether official misbehavior is a serious problem. Here again, a caveat is needed. More often than we have space to describe, research results are based on questionable inferences from methodologically flawed investigations. Further, rape research, even if methodologically sound, often covers only a single jurisdiction (commonly a city) in a Single year. Findings from Detroit do not necessarily correspond to realities in, say, rural Texas. Not only that: much of the research is over ten-years-old. In a field like rape, where public attitudes and governmental practices are evolving, it is often impossible to tell whether an empirical study still reflects even the narrow slice of reality that it covers. We will try to minimize this problem by seeking corroboration of key findings, or hedging our conclusions when appropriate.

B. REPORTING

Is there really a "rape epidemic"?[157] If by "epidemic" one means an escalating incidence of rape?[158] the surprising answer is that the experts are not entirely certain. The two sources of national crime statistics, both managed by the Department of Justice, show markedly different trends, for rape as for some other crimes.[159] One source, the Federal Bureau of Investigation's annual Uniform Crime Report (UCR), tabulates crimes reported[160] to police departments throughout the country. According to the UCR, the number of rapes per 100,000 inhabitants rose from 24.5 in 1973 to 39.2 in 1994.[161]

A different trend emerged from the National Crime Victimization Survey (NCVS), which is based on interviews of all persons aged 12 and over in a stratified sample of households, asking whether they had been victims of various crimes.[162] According to the NCVS, the incidence of rape declined gradually during the same period in which the UCR showed an increase.[163]

By tabulating only reported crimes, the UCR understates the incidence of most crimes including rape.[164] In addition, any long-term change in the proportion of victims who report a crime will appear in the UCR as a (spurious) upward trend in the crime rate, The most common scholarly explanation of the discrepancy between the UCR and the NCVS is that rape victims have become more willing to report the crime.[165] This hypothesis is plausible: the rise in reported rapes coincided with the rise of modern feminism, including a great deal of consciousness-raising about rape, the development of rape crisis centers to support victims, and improvements in the police response to rape.[166]

Yet, the NCVS figures are also suspect. Mainly because it focuses on households with telephones, the NCVS systematically ignores several demographic groups that are highly vulnerable to rape and other crimes, such as transients, the homeless, and those without telephones.[167] However, these methodological flaws were constant during the period in question and therefore presumably did not create a spurious trend. Admittedly, NCVS interviewers may have become increasingly reluctant to venture into the most dangerous inner-city neighborhoods;[168] if so, this may have contributed to the downward trend in the NCVS rape rate. Everything considered, however, it seems likely that the NCVS's trend is more accurate than the UCR's.

Be that as it may, scholars agree that rape is still seriously under-reported. In apparently a large majority of acquaintance rapes, the victim either does not report the crime or, having reported it, subsequently decides not to press charges,[169] Estimates of the percentages of rapes that are reported vary, ranging from a low of well under 10% to a high of 60%.[170]

Scholars have advanced several reasons for victims' failure to report.[171] Some rape victims are too upset, or too embarrassed at the prospect of answering a stranger's intimate questions about the incident,[172] or so ashamed that they do not want anyone, even their friends, to know about it.[173] Sometimes the victim fears retaliation by her assailant.[174] In other cases, she wishes to conceal some aspect of her own behavior drug use, for example--immediately ,prior to the rape.[175]

It seems probable that unreported rapes are, disproportionately, acquaintance rapes,[176] In some acquaintance rape cases, the victim does not report the rape because she blames herself or does not regard the crime as a "real rape."[177] In other acquaintance cases, the rapist is a relative. For example, a rape by the victim's stepfather might not be reported because the victim's mother tells her not to notify the police.[178] Even if the rapist is not a relative, the victim's decision about whether to report the crime is often affected by the responses of her family or boyfriend.[179] She also may be swayed by the perceived expectations of her social group concerning the appropriate behavior of victims,[180]

Sometimes the victim does not report because she wishes to maintain her relationship with the offender,[181] One study found that 39% of rape victims date their attacker after the rape.[182] This is much less surprising than it sounds, because many rapes are perpetrated by lovers.[183]

Smith and Nelson found that the likelihood the victim would report varied directly with the amount of danger to her during the rape, the social distance between her and the offender, and the degree to which she expected support from relatives and friends for reporting the rape.[184] Nonreporting was most likely if the victim and the rapist were acquaintances, of about the same age and social class, and the rapist did not use a weapon or commit another crime (for instance, robbery) at the same time.[185] These findings, concluded the authors, illustrate "the widespread internalization among the [victim] sample of cultural stereotypes which tend to legitimate rape...."[186]

Although the ultimate culprit may be cultural stereotypes, the massive nonreporting by victims might be regarded as proof that they, not the justice system, bear the main responsibility for the fact that most rapists are not even charged with a crime. Many scholars believe, however, that the reluctance of rape victims to prosecute rapists is largely due to societal and official skepticism toward victims of acquaintance rape: (1) police and prosecutors who allegedly have deterred reporting by failing to take accusations of acquaintance rape seriously[187] and (2) a justice system that "puts the victim on trial" and thereby creates an ordeal that intimidates many women who might otherwise seek redress through a criminal prosecution.[188]

To the extent that this indictment refers to bad laws rather than bad attitudes, its validity can be tested by changing the laws. During the past twenty years or so, virtually every state legislature has enacted some mixture of rape law reforms.[189] One goal of these reforms, particularly shield laws, was to reduce the rape victim's ordeal in the justice system and thereby to increase reporting rates.[190] As we have seen, most scholars believe that the proportion of victims who report rapes has risen in recent decades. For example, one study concludes that the national proportion of rape victims who reported the crime rose by 10% between 1980 and 1990, a modest increase but one greater than for nonsexual assault and robbery (12% decrease).[191] If so, did legal reforms contribute to this trend?

In trying to answer this question, several possibilities should be kept in mind. First, the local publicity accompanying passage of reforms might lead to a perhaps temporary increase in reporting, due simply to victims' enhanced sensitivity to the importance of rape, or their impression of public support. Second, local publicity might increase reporting by making victims aware, for example, that their state has a rape shield law. This awareness might or might not be transient. Third, local reforms might improve the performance of the justice system, a result that, if sufficiently publicized, might encourage more reporting. Fourth, attitudinal changes due to national publicity about rape, or rape reforms, or feminism in general, might improve the reporting rate, for example by persuading women that female passivity is inimical to sexual equality.

In a careful study of the impact of reforms, Professors Cassia Spohn and Julie Homey analyzed data from six cities: Detroit; Chicago; Philadelphia; Washington, D.C.; Atlanta; and Houston.[192] The first three cities were in states with "strong" rape law reforms, while the last three represented states with "weak" reforms.[193] Michigan, often described as a model reform state,[194] has by far the strongest reforms, redefining rape, eliminating corroboration and resistance requirements, and including a highly restrictive rape shield law.[195] Illinois' reforms were similar, but without repeal of the corroboration requirement.[196] Pennsylvania enacted a strong rape shield law and eliminated the cautionary instruction, and the prompt complaint, corroboration, and resistance requirements, but retained a more traditional definition of rape.[197]

The "weak reform" cities (Houston, Atlanta, and Washington, D.C.) were all in jurisdictions with relatively weak shield laws, and in these jurisdictions the resistance requirement was not abolished,[198] Georgia and Washington, D.C. retained the traditional definitions of rape.[199] Texas redefined rape and modified the corroboration requirement, but these changes were much more limited than the corresponding changes in Michigan and Illinois.[200]

The authors chose jurisdictions that had enacted their reforms at different times, so that changes correlated with national publicity about rape could be distinguished from changes correlated with the legal reforms.[201] Analyzing monthly data from 1970 through 1984, they found that legal reforms had no positive effect on rape reporting in Chicago, Philadelphia, or Atlanta.[202] In the District of Columbia there was a decrease in reported rapes, no doubt coincidental, after elimination of the corroboration requirement, and the authors found no evidence that a new rape shield provision affected reporting rates.[203] Spohn and Homey found in Houston, the jurisdiction with the weakest reforms, and in Detroit, with the strongest reforms, an increase in reporting rates that appeared to be linked in some way to the reforms.[204]

In Detroit, reports increased fairly steadily before 1975, when Michigan enacted its comprehensive reforms. Reporting rates climbed even higher after the reforms, continued upward for a short period, and then stabilized at a fairly steady rate.[205]

A potentially important difference between Michigan and some other jurisdictions is that the Michigan reform package received a great deal of publicity.[206] Perhaps Michigan women apprehended a sea change in societal, or at least official, attitudes toward rape. In jurisdictions like the District of Columbia, where reforms were achieved by judicial decisions, and at different times, rape victims probably were unaware that the law had changed.[207] Spohn and Homey conclude: [T]he increase in reports probably was not caused by the substantive content of the rape law reforms. In both Detroit and Houston the increases in reporting appeared at the time the law reforms went into effect; this suggests that the increases resulted from publicity surrounding the reforms rather than from gradually acquired knowledge of improved treatment of victims under the new laws.[208]

Another Michigan study, statewide in scope, found that, after enactment of reforms in 1974, reports of forcible rape increased at a faster rate than reports of any other serious offense.[209] According to a survey of Michigan judges, prosecutors, defense attorneys and rape crisis counselors, increased reporting of "marginal" cases occurred after the Michigan law reforms.[210] Post-reform Michigan rape victims were reportedly less traumatized by their experience with the criminal justice system and "more willing to follow through on prosecution."[211] Nevertheless, the authors conclude that this phenomenon was independent of the reforms and may have been due to feminist consciousness-raising during the same period.[212]

If this conclusion is correct, perhaps the changes in Detroit's reporting rates had parochial causes. Spohn and Horney mention two possibilities. In Detroit, their interviews "revealed strong support for the reforms among criminal justice officials... [who]--in responding to our hypothetical cases--took the strongest positions on excluding sexual history evidence."[213] While granting that these attitudes may have been a cause rather than a result of Michigan's strong reforms, Spohn and Homey believe that the latter is more likely.[214]

In Detroit, the Rape Counseling Center is run through the police department, an unusual arrangement that "gives the Counseling Center earlier and greater access to victims than occurs in many cities and thus potentially greater influence in encouraging reporting and pressing for prosecution."[215] This relatively unheralded innovation may be one of the reasons for the increase in rape reporting in Detroit.

In sum, the evidence, although mixed, suggests that rape reporting rates are generally unresponsive to changes in a particular jurisdiction's rape law, even when those changes signal a desire to reduce victim blaming in rape trials.[216] This finding may surprise some readers who are familiar with the literature about rape law, which has a pronounced reformist orientation. Reformist authors naturally tend to highlight objectionable practices that they wish to see abolished, for instance, cross-examination of the alleged victim about her sexual habits.[217] Typically, such authors claim or at least strongly imply that objectionable rules have major instrumental effects. By the same token, they fail to emphasize--sometimes even fail to mention--those aspects of the legal process that contribute to the same harmful result but should not or cannot be reformed.[218] Yet the emotional rigors of a rape trial are due largely to intractable realities: the law's delay;[219] the unpleasantness of having to relive a horrifying and extremely intimate experience; fear, even apart from sexual history questions, of cross-examination, or of media publicity,[220] and embarrassment about the rape itself or about surrounding circumstances that inevitably will be revealed during the trial. All this is coupled with a desire, especially in some acquaintance rape cases, to let bygones be bygones,[221] and to get on with one's life.[222] The possibilities that the police will be unsympathetic, that the defense will be allowed to ask about previous love affairs, and that the rapist will escape justice anger any decent observer, but may not loom as large in many victims' calculations as more mundane, personal considerations,[223] and in any event, as we shall see, most reforms do not appear to have affected conviction rates.[224]

It remains possible, nevertheless, that law reforms have contributed to the national attitudinal changes that appear to have triggered increased reporting of rape. It seems likely that most rape victims are aware of national trends, perhaps including rape shield laws, but relatively hazy about local law. After all, most of the recent publicity concerning rape has emphasized national perspectives. Thus, law reforms in one state may indirectly encourage reporting in another state by providing fuel for the great bonfire of national publicity carrying the message that something is being done about rape. To the extent that this occurs, efforts to discover a correlation between passage of reforms in any particular state, and subsequent behavioral changes in that state, miss the point. Spohn and Homey do not deny this possibility, and it is consistent with their data.

C. THE POLICE FOUNDING DECISION

As with all crimes, the police decide whether a reported rape actually occurred, and attempt to determine who committed it. If they want the case to go forward, they "found" the complaint and transmit the file to the prosecutor's office[225] A few authors have depicted police rape investigators as sexist, tactless oafs.[226] While some policemen clearly deserve those labels,[227] a more balanced picture emerges from a leading study of case processing. Holmstrom and Burgess studied the cases of 146 reported rape victims admitted during a one-year period to the emergency wards of a large Boston hospital, plus a smaller number of victims from other sources.[228] By following these women through the entire legal process, Holmstrom and Burgess sought to determine "the part played in victimization by those groups and institutions that are supposed to protect the victim."[229] Concerning police behavior, Holmstrom and Burgess stressed that " [t]he nature of the material that is discussed--the details of an extremely frightening and humiliating sexual experience--mean that the encounter [with the police] may be unsettling for many victims even under the best of circumstances."[230] The victim is female,[231] the officer usually is male.[232] The victim knows that she really was raped; the officer often does not. The police must investigate, a task that cannot easily be combined with offering the emotional support that the victim needs. The detective presumably wishes to avoid an injustice to a wrongly accused individual. In addition, for reasons of professional pride, he does his best to avoid looking naive by falling for a story that turns out to be false.[233] Experienced investigators also know that many rape complainants ultimately decline to press charges, sometimes to the dismay of a detective who has worked hard to build a case.[234] Regardless of the nature of the crime, police tend to believe that people in general are untrustworthy, and that therefore the wisest course is to trust no one.[235]

As Richard Posner has observed, police generally come from a culture that has traditionally disapproved of women who engage in casual sex:

[N]onmarital intercourse[236] carries a stigma in some social strata--including those from which most police officers and many judges are drawn. These enforcement officials are sympathetic to "respectable" women raped by members of the criminal class but often not to sexually active women raped in compromising situations--say, while drunk on a date while scantily clad--or to women raped by their husbands. In such settings enforcement officials may believe that the woman has hiked up the rape charge for a strategic advantage, such as jockeying for favorable terms of divorce or trying to explain away a pregnancy.[237]

Despite these potential sources of friction, Holmstrom and Burgess found that most victims spoke favorably about the police who had questioned them:[238] only 10.3% expressed a negative reaction.[239] These negative evaluations were due to a variety of complaints, including police behavior perceived as overbearing, harsh, or moralistic.[240]

It is sometimes difficult to distinguish between appropriately aggressive investigation and tactless hostility. For example, one victim complained about "endless" questions,[241] which may have been asked in order to see whether the victim would change her story.

As an example of what they consider an "overbearing" approach, Holmstrom and Burgess describe a case in which investigators asked the complainant whether she had ever tried committing suicide, or had run away from home, taken drugs, or been hospitalized, (and why).[242] After this interrogation, the police concluded by telling her that rape carries a life sentence, implying that she ought to reconsider her report.[243] Perhaps that last observation was inappropriate, though such remarks serve to test the woman's resolve to press charges, a justification for which we feel some sympathy, given the large number of rape complainants who ultimately decline to cooperate with the police.[244] In any event, we have no objection to questions about attempted suicides and the like. Any competent prosecutor will want police to adduce any facts that may discredit the victim in the eyes of jurors. The persuasive objection is not to asking such questions but to doing so tactlessly, or with the attitude that affirmative answers should automatically lead to termination of the investigation.

Holmstrom and Burgess found a "striking absence" of the most absurd rape myths among the police they studied: that all women want to be raped, or enjoy it, or cannot be raped if they don't want to be.[245] Once they decide that a man is a rapist, the police often feel extremely punitive toward him, preferring prison to rehabilitation or psychotherapy.[246]

At the same time, police are afraid of being too gullible and appearing incompetent when the woman they believed changes her story[247] or decides not to press charges.[248] Police also know that the rape victim faces an ordeal in the legal system,[249] often culminating in an acquittal or a light sentence for the rapist. Understandably, some police harbor a sense of futility about rape cases.[250]

The national unfounding rate for rape is about 8%.[251] This means that the police believe over nine-tenths of the rape reports that they record.[252] On the other hand, the proportion of acquaintance rape complaints that are unfounded is higher than 8%.[253] Moreover, the unfounding rate for rape is roughly four times higher than for other major crimes.[254] Several scholars have tried to determine whether this is due to police biases against certain types of victims. For example, some scholars have tried to discover whether the police are mainly concerned, in processing rape complaints, with legitimate evidentiary questions. Two articles addressing this question depict the police in a basically favorable light. Wayne Kerstetter studied the sexual assault cases of the Chicago Police Department for 1979 and 1981.[255] Using a lengthy questionnaire, he classified factors that officers considered relevant to case processing,[256] under different headings, such as "administrative" (officer making a routine discretionary decision), "instrumental" (necessary to facilitate processing, but not within sole discretion of officer), matters of "evidence," "offense elements," or "aggravating" variables.[257] Kerstetter then analyzed the apparent effects of these variables on case processing using discriminant analysis.[258]

Discussing stranger rapes, Kerstetter notes that, "[w]hen a complainant alleges that she was attacked by an unknown assailant, the investigator faces three questions: (1) Did a sexual assault actually occur? (2) Who was the assailant? and (3) Where may he be found?"[259] The factors which correlated most with police founding decisions in stranger rape cases, "[W]ill the complainant prosecute?" and "[I]s the accused in custody?" relate directly to these questions?[260] Negative answers raised investigative costs and hampered prosecution.[261] As Kerstetter notes of the detectives in his study, "since they have more cases than they can handle, they have little incentive to pursue a case" if the complainant does not want to prosecute; her cooperation is crucial to a successful investigation.[262]

In acquaintance rape cases, the primary question for investigators is whether the putative victim consented.[263] Here too, Kerstetter concluded that the factors that are highly associated with the police founding decision relate to this central evidentiary question. Police were more likely to found acquaintance rape complaints if the suspect was in custody (low opportunity costs of investigating his state of mind) or if the victim had injury to her sex organs (corroborative evidence supporting lack of consent).[264] Conversely, the existence of nonsexual discrediting information, such as alcohol or drug use (lower victim credibility), often appeared to result in police unfounding complaints.[265]

Kerstetter concluded that instrumental and evidentiary variables structure police decision-making in both stranger and acquaintance rape cases.[266] Though these structures differ, both relate to proving the necessary elements of the crime. Kerstetter's study suggests that whether police found a rape complaint depends upon whether they believe that the government can prove these elements.

Gary LaFree reached similar conclusions after analyzing data concerning all 905 forcible sex offenses reported to police in a single city during 1970, 1973 and 1975.[267] LaFree concluded that, "The two best predictors of arrest were legal variables: the victim's ability to identify a suspect and her willingness to prosecute."[268] These findings dovetail with the Kerstetter study in that police are most likely to act when the elements necessary for a successful prosecution are present.

Yet, LaFree's data could just as easily be construed against the police. Although he found that victim misconduct was rarely alleged, he also found that such allegations, and delay in reporting, strongly correlated with a negative decision.[299] As LaFree acknowledges, "none of the 31 cases which alleged [victim] misconduct resulted in arrest."[270] Despite this startling finding, LaFree goes on to downplay the importance of extralegal variables:

The literature on official reactions to sexual assault emphasizes the victim's attributes and the interpersonal context of the crime for explaining variation in official decisions. For police in this city, at least, this emphasis appears to be greatly overstated. Victim's race, the location of the incident, victim resistance, victim injury, and witnesses had no effect on police decisions .... For all three outcomes, [arrests, charging, felony screening] legally-relevant variables were paramount.[271]

The proposition that "legally relevant variables" are "paramount" is not quite as informative as it may appear to be. One of LaFree's two major "legal variables," the victim's ability to identify a suspect, is nearly always present in acquaintance rapes,[272] and therefore is unhelpful in analyzing police decisions in those cases. The other major "legal" variable, the victim's willingness to prosecute, may be influenced by the detective's attitude toward the case.[273] This in turn is likely to be affected by matters, such as the woman's violation of sex role norms, whose relevance to whether she was raped is at best questionable.[274] The upshot is that LaFree's conclusions rest on evidence that does not differ greatly from the evidence adduced by scholars who are more critical of the police.

Even more plainly, Kerstetter's conclusion that police appraise acquaintance rape cases in light of legitimate evidentiary concerns is open to question. In acquaintance rape cases, the fact that a suspect is in custody seems to be, pace Kerstetter, a consequence rather than a cause of the police decision to found an accusation of rape. (If the police disbelieve the woman, they presumably will not arrest the man.) Phrases like "legitimate evidentiary concerns" obscure such critical issues as whether police give too much weight to, say, lack of physical injury to the woman, typical in acquaintance rapes,[275] in deciding whether to found the complaint.

Kerstetter also fails to distinguish between evidence that is certainly relevant to whether a rape occurred for instance, the consistency of the woman's story--and evidence of more debatable relevance, such as her drinking. There is a world of difference between these two types of "evidentiary concerns," even though both may affect jurors' attitudes.

Susan Meyers Chandler and Martha Torney are more critical of the police. They stress the role of extralegal variables in founding decisions.[276] After tracking sexual assault victims treated at a large urban hospital during the years 1976-78, they concluded that: "[t]he behavior and lifestyle of the victim prior to the assault, as well as at the time of the assault, are often considered by the processing agents to be as important as the actions of the assailant during the crime."[277] Chandler and Torney maintain that many of the extralegal concerns relevant to police processing of rape complaints reflect a "male bias toward acceptable female behavior."[278] They may be right, but their own analysis shows that this proposition is sometimes arguable. The variables that they characterize as "extralegal" are the same types of evidence that Kerstetter calls "legitimate evidentiary concerns." As Chandler and Torney concede, the "extralegal" variable "sobriety" might relate to the complainant's credibility.[279] Similarly, the "past sexual relations between the complainant and the alleged assailant" might be thought of as legitimate evidence concerning whether the sexual encounter was consensual or would be deemed so by a jury.[280]

Hubert Feild studied the attitudes of patrol police, rapists, crisis counselors, and citizens of a community using questionnaires designed to elicit their feelings about rape and to assess their general knowledge of the crime.[281] He found that the police officers were more similar to the rapists than to the counselors in their views of rape.[282] Police and rapists tended to agree on the basic motivation for rape, the lack of attractiveness of a rape victim after rape, and the mental abnormality of rapists.[283] Feild concluded that "the results do suggest a need for systematically evaluating the relations between police officers' attitudes toward rape and their behavior toward a rape victim."[284]

Likewise, Shirley Feldman-Summers and Gayle Palmer found that the police officers who participated in their study began with the expectation that three of every five rape complainants are either untruthful or mistaken.[285] They reasoned that, "It should not be surprising that such a belief is communicated to the victim in one way or another, even by police officers who try to treat a rape victim with courtesy and respect."[286]

If accurate, these studies are cause for concern. However, Feldman-Summers and Palmer only surveyed approximately fifteen ofricers.[287] The Feild study involved 254 officers, but its conclusion is debatable.[288] Underlying Feild's belief that police ought to think like rape counselors rather than rapists is the assumption that counselors are wiser about rape than rapists. This assumption is obviously valid on some issues--for example, whether some women deserve to be raped. But on some other issues, such as why men rape, conceivably rapists' opinions are more accurate than counselors'. Indeed, some feminists have quoted rapists' descriptions of their own motivations, as evidence that pornography causes rape.[289] Of course, these feminists may be wrong about the effects of pornography, but their contention cannot be dismissed simply by observing that rapists have the same opinion.

A number of authors have compiled lists of reasons given by police for unfounding rape complaints,[290] while other scholars have sought to determine, independently of the detectives' stated reasons for unfounding, which aspects of a rape case correlated, in particular cities, with decisions to unfound.[291] Combining several lists of both types, we compiled a master list of factors that have been found by at least one scholar to be the expressed reason for, or to be associated with, unfounding decisions in at least one city:

1. The woman (or girl) is unmarried but sexually experienced, or (worse) promiscuous?[292] or (even worse) a prostitute.[293]

2. She was drinking or taking drugs or is a drug addict.[294]

3. Her story contains inconsistencies[295] or she keeps changing it,[296] or it does not "check out."[297]

4. Police are unable to corroborate her story[298] for example, by injuries to her sex organs.[299]

5. She exposed herself to risk of rape--or in the vernacular, "she asked for it"--by accompanying the man voluntarily to the site of the alleged rape,[300] or inviting him into her home[301] or engaging in risky behavior such as hitchhiking.[302]

6. She has, apart from the rape, severe mental or emotional problems, perhaps evidenced by a history of psychiatric treatment or an attempted suicide, or running away from home.[303]

7. She does not appear to be upset by the alleged rape.[304]

8. She is unattractive.[305]

9. She knew the alleged rapist,[306] or, much worse, had been sexually intimate with him, voluntarily, on other occasions.[307]

10. There is no large discrepancy between her age and his.[308]

11. She is at least 12 years old.[309]

12. The alleged offender does not have a previous criminal record.[310]

13. She herself (rather than a relative or acquaintance) contacted the police.[311]

14. She alleges completed penile-vaginal intercourse rather than a lesser offense such as fondling or an attempt.[312]

15. She has demonstrable scratches.[313]

16. She is on welfare.[314]

17. She alleges that more than one offender was involved.[315]

18. She has had prior trouble with the police.[316]

19. She is an African-American.[3l7]

20. No woman (other than the complainant) was present during the interrogation.[318]

21. The offense was not promptly reported.[319]

22. The putative victim failed a polygraph test.[320]

23. She is uncooperative.[321]

24. She is inarticulate.[322]

Most of the above findings, though suggestive of improper police biases, are inconclusive even in the jurisdictions studied. As with studies of attrition rates, the most common problem is an excessively quantitative methodology: the scholar tabulates information from the often-sketchy police records, and then discovers, for instance, a correlation between the alleged victim's age and the founding decision. But, of course, most of the flesh and blood of a case is not preserved in a pithy police report, even before it has been reduced by a social scientist to a collection of measurable data. As a result, one cannot determine whether any particular correlate of unfounding is indicative of police bias or, rather, of the fact that the group in question files rape reports that are less credible, on average, than reports by other women. Subjectivity, so carefully excluded at the front door of the research design, enters triumphantly through the back door, when the reader intuits whether a particular correlation--say, between the complainant's social class and the unfounding decision--was due to reasonable police appraisals of all the circumstances bearing on vicrims' credibility or, on the contrary, the ignorance and prejudice of detectives.

Few of the items on our list are clearly indicative of either good detective work, on the one hand, or indefensible bias on the other.[323] Even an appropriate criterion such as the absence of corroboration can be misused. In acquaintance rape cases, corroboration, in the sense of evidence strongly indicating that rape rather than consensual intercourse occurred, is often lacking. In our judgment, it would be improper to unfound an otherwise credible acquaintance rape complaint merely because of the absence of this sort of corroboration. Yet, the rigorously quantitative social-scientific studies do not reveal how often this occurs. We learn that lack of corroboration is a factor, but we do not learn whether it is employed judiciously.

Conversely, even the most suspicious-sounding correlate of unfounding decisions may have an innocent explanation. No one knows which demographic groups, if any, are disproportionately likely to lie to the police about rape, or at least to tell stories that sound less credible than most rape reports. We may have our suspicions, but we do not know, for example, whether women on welfare are in fact more likely to lie about rape. Lacking such information, we cannot tell which of the correlations found by social scientists are due to unfair police discrimination.

Although some studies of founding decisions are not as useful as one might wish, several conclusions can be drawn. First, founding decisions in rape cases are highly subjective. This is evident, for example, from a study indicating that in one city the presence of another woman during interrogation of the complainant increased the likelihood that police would found the complaint.[324] Simulated jury studies reveal that appraisals of acquaintance rape scenarios are affected by one's gender and by whether one holds sexually egalitarian beliefs.[325]

A priori, then, male-dominated detective squads are likely to be at least somewhat too skeptical towards accusations of acquaintance rape. This conclusion does not require us to assume that police are uniquely biased; only that they are not uniquely free of bias. For if anything is clear about human nature, it is that groups tend, usually unconsciously, to interpret the world from parochial, self-interested points of view. While we cannot say exactly how much founding decisions are distorted by male bias, it seems highly likely that, except perhaps in the increasingly common sexually-integrated detective units,[326] some degree of distortion does occur.

Second, most observers agree that founding decisions in acquaintance rape cases are strongly affected by the purported victim's contributory negligence, and by her perceived immorality.[327] Unlike some of the other unfounding correlates, these recur in virtually every study of rape case processing, and police often consciously employ them. If police bias does indeed distort founding decisions, this appears to be the type of case in which most of the distortion occurs.

To defend the conscious use of contributory negligence/morality criteria, one must argue either that the trait in question, promiscuity for example, is relevant on the consent issue, or that such women are known to be disproportionately inclined to fabricate rape charges, or that the police should unfound these cases, regardless of their merits, because juries are unlikely to believe such women. We discuss the first two justifications in Part II of the article.[328]

The third justification, difficulty of obtaining a conviction, presumes that the police should be in the business of deciding, not only which crimes actually occurred, but which can be proved beyond a reasonable doubt, to the satisfaction of a possibly biased jury. In theory, such decisions are to be made by prosecutors.[329]

We will not dogmatically assert that departures from the theoretical division of functions between police and prosecutors are never justifiable on practical grounds. Once the police determine that the prosecutor is unlikely to file charges, for instance, because the woman's contributory fault makes a conviction too hard to obtain, further investigation may be a waste of time.

Reserving judgment on whether this attitude makes sense for some other crimes, in our opinion it is not an appropriate approach to rape complaints. Rape is one of the gravest and most devastating offenses. It is, even more than most crimes, an offense in which the appearance of official fairness is as important as the reality, if only because of the long history of inadequate societal responses to acquaintance rape. The rape victim's humiliation and sense of injustice are likely to be compounded if she is told at the outset that nothing will be done because she brought it on herself, say by hitchhiking, or because she is a moral outcast whom the law does not protect. Except when the report plainly appears to be false, the police should be as supportive as possible, consistent with the requirements of a thorough investigation. If it later becomes necessary to tell the woman that no formal charges will be filed, that should be done, with appropriate explanations, by the prosecutor. The prosecutor's claim to be motivated solely by problems of proof will usually be more credible than a comparable assertion by a detective who lacks legal training and who is probably more likely than a prosecutor, on average, to object strongly to the complainant's behavior on legally irrelevant moral grounds.[330]

A substantial minority of rape complainants ultimately decline to press charges.[331] Such withdrawals are most common in cases where, as in many acquaintance rapes, there is a close victim-offender relationship.[332] Holmstrom and Burgess found that only 41% of the Boston rape complainants they studied "clearly and unequivocally" wished to press charges.[333] Many others (31%) were ambivalent,[334] while a small minority (15%) definitely did not want their cases to proceed.[335] The authors concluded that, "[i]t is the rare victim who is unequivocally determined to press charges and who remains so throughout the entire court process."[336]

Complainants gave several reasons for refusing to proceed:

To avoid ordeal of court: 29 (24%)
Afraid of assailant taking revenge: 26 (21%)
To avoid sending a person to jail: 17 (14%)
What's the use--he'll get away with it anyway: 9 (7%)
Feel sorry for the guy: 9 (7%)
Just want to forget the whole thing: 9 (7%)
Scared of identifying wrong guy: 5 (4%)
Would look bad on my record: 3 (2%)
Other: 16 (13%)[337]

Judging by these figures, conventional accounts of victims' attitudes overemphasize feelings of hopelessness engendered by systemic bias against women. In this study, victims were not primarily motivated by a feeling that "he'll get away with it anyway," an apprehension that only 7% of them expressed, half as many as wanted "to avoid sending a person to jail" and exactly the same proportion as "felt sorry" for the rapist. To be sure, the single most common reason for failure to press charges was to avoid the ordeal of the courtroom, where the victim anticipated that she would be blamed for the crime and embarrassing questions would be raised about her sexual habitsin for example, an abortion, use of birth control, and that she "slept around."[338] Still, only about one out of four gave this reason.[339]

It remains possible, however, that police who are unduly negative affect the victims' decisions. A study in Denver found that 38% of rape accusers withdrew their charges soon after filing them. Among the reasons given was police resistance to prosecuting the case.[340]

Kerstetter investigated which variables were correlated with the victim's willingness to prosecute.[341] He concluded that the perpetrator's use of a weapon and the presence of a witness to the rape were important factors in both acquaintance and stranger rape cases.[342] To explain this finding, Kerstetter surmised that detectives influence the woman's decision to prosecute. If no weapon was used, and no witnesses exist, the police may feel that prosecution would be futile. In order to decrease the number of unsolved crimes on his record, the detective may attempt to dissuade the woman from pressing charges.[343]

In Michigan, a study of the impact of comprehensive rape law reforms found that the rate of unfoundings remained stable (less than 10%) after passage of the reforms.[344] Between 1973 and 1977, arrests for forcible rape rose by 61%, compared with 16% increases for both murder and aggravated assault and with robbery, which showed a decline of 8%.[345] Although the increased number of arrests was a predictable result of increased reporting coupled with a stable unfounding rate, the authors concluded that the rising arrest rate was also influenced by Michigan's strong rape law reforms.[346]

D. THE DECISION TO PROSECUTE

Once the police clear[347] a rape complaint, the prosecutor has, as with other crimes, virtually unlimited discretion in deciding whether to prosecute, whether the charge should be rape or some lesser offense, and whether to plea bargain.[348]

In some cases, the prosecutor dismisses the case because the putative victim has belatedly decided not to cooperate.[349] More often, prosecutors focus on proof problems. Frohmann studied prosecutorial decisions to reject rape cases in two West Coast communities in 1989 and 1990.[350] She concluded, as have others, that prosecutors screen out the "unwinnable" cases in order to improve their conviction rates, because this impresses their superiors and, if they are politically ambitious, the electorate.[351] As a rule, acquaintance rape cases are more difficult to win than stranger rape cases.[352] Thus, acquaintance rapes, especially those with "bad victims," are less likely than stranger rapes to result in prosecution of the offender.[353]

Holmstrom and Burgess found several similarities between the attirades of police and of prosecutors. Like the police, prosecutors were more likely to proceed with a case if the accused had a prior criminal record,[354] and if the complainant's account was consistent and unchanged.[355] District attorneys felt that their competence was on the line when a case was tried and did not want to look foolish. The prosecutors were wary, accordingly, of witnesses who did not tell the whole truth during preparation of the case, or who belatedly changed their stories.[356]

Prosecutors looked for medical evidence of force and penetration,[357] and took account of the woman's appearance and demeanor-for example, does she smile when she's upse?[358] Like all participants in the criminal justice system, the prosecutors gave weight to the woman's character, for example, her alcoholism, and her relationship with the accused.[359] Prosecutors were disappointed to learn, for example, that she had invited the accused into her apartment.[360]

Surprisingly, the prosecutors' most common judgmental comments concerned the women's intelligence. Holmstrom and Burgess speculated that perhaps prosecutors sought witnesses who could verbalize well.[361] Given the critical importance of precise testimony about the parties' words and actions, this explanation is plausible, An intelligent witness may be better able to respond to a vigorous cross examination. More broadly, intelligence may be a proxy for class. A solidly middle-class complainant would not only be more articulate, on average; she might also be more credible to many jurors than a lower-class woman, especially in a consent-defense case. After all, prosecutors want a witness whom jurors will regard as sexually restrained and honest, characteristics that jurors may associate with middie-class status. As an experienced sex-crimes prosecutor puts it:

Good Victims have jobs (like stockbroker or accountant) or impeccable status (like a policeman's wife); are well-educated and articulate, and are, above all, presentable to a jury: attractiverebut not too attractive, demure--but not pushovers. They should be upset--but in good taste--not so upset that they become hysterical.[362]

In rape cases, critics of prosecutors have focused mainly on their excessive use of "winnability" as a criterion for filing charges.[363] The basic point is well-taken, but some of this criticism strikes us as injudicious. Here, for example, is a passage from a Senate Judiciary Committee Report:

No one expects prosecutors to bring unwinnable cases--cases that could not meet the "beyond a reasonable doubt" standard. But prosecutors must also understand that there is more at stake in the prosecution of rape cases than an individual "win" or "loss." In rape cases, the reputation of the system is at stake in every prosecution. When survivors hear that prosecutors will not bring cases if they happen to recognize the perpetrator--the kid down the street, the UPS delivery man, the father of a friend--that has consequences. Why report the crime if you know it will not be prosecuted? Why pursue a case if you know you will have to fight barriers of institutionalized disbelief?. Why believe in our system of justice at all?[364]

This passage seems to conrate two quite different problems: satisfying courts and satisfying juries. "Unwinnable cases" are not usually ones that could not meet the "beyond a reasonable doubt" standard, as defined by courts. Even in the weakest acquaintance rape cases, the victim's testimony usually establishes a primafacie case, so the danger is not so much a directed verdict as an acquittal by the jury. For all its rhetoric, the quoted passage does not say what a prosecutor should do when a jury acquittal appears highly probable.

The authors of the Senate Judiciary Committee Report appear to believe that prosecutors' attitudes strongly affect the reporting rate for rape. This belief is commonly held by rape scholars,[365] and it is not demonstrably false, although we doubt that many victims know even approximately how often prosecutors decline to prosecute rape cases.[366] The Report's admonition against treating won-lost records as the sole gauge of prosecutorial success is fair enough. But the Report fails to consider the possibility that, if the prosecutor's pessimism about getting a conviction is well-founded, the rape victim will be disillusioned, sometimes even more sharply, by an acquittal after an emotionally-wrenching trial.[367] We do not say that this letdown is necessarily as severe as disappointment due to never having had a day in court, only that, if a jury will not convict, the choice is not between a bad alternative and a good one but between two bad alternatives.

The Senate Report also does not discuss the possibility that, in order to prosecute a losing rape case, the prosecutor will have to neglect other, more winnable types of cases, perhaps leaving other victims and potential victims disillusioned or unprotected[368] To make these abstractions more concrete, consider a conversation between Professor Susan Estrich and a prosecutor about a rape case:

The victim came to ... [the assistant district attorney's] office ... dressed in a pair of tight blue jeans. Very tight. With a see-through blouse on top. Very revealing. That's how she was dressed. It was, he tells me, really something. Something else. Did it matter? Are you kidding!

The man involved was her ex-boyfriend. And lover: well, ex-lover. They ran into each other on the street. He asked her to come up and see Splash on his new VCR. She did. It was not the Disney version--of Splash, that is. It was porno. They sat in the living room watching. Like they used to. He said, let's go in the bedroom where we'll be more comfortable. He moved the VCR. They watched from the bed. Like they used to. He began rubbing her foot. Like he used to. Then he kissed her. She said no, she didn't want this, and got up to leave. He pulled her back on the bed and forced himself on her. He did not beat her. She had no bruises. Afterward, she ran out. The first thing she did was flag a police car. That, the prosecutor tells us, was the first smart thing she did.

The prosecutor pointed out to her that she was not hurt, that she had no bruises, that she did not fight. She pointed out to the prosecutor that her ex-boyfriend was a weight lifter. He told her it would be nearly impossible to get a conviction. She could accept that, she said: even if he didn't get convicted, at least he should be forced to go through the time and the expense of defending himself. That clinched it, said the D.A. She was just trying to use the system to harass her ex-boyfriend. He had no criminal record. He was not a "bad guy." No charges were filed.[369]

In discussing this case with Professor Estrich, the prosecutor emphasized the "likely response of juries," and he "leaned heavily on the 'neutrality' of his decision."[370] It was, he assured Estrich, the sort of decision he makes every day, in all sorts of criminal cases, as he decides whether to file charges. The critical evidentiary factors, in his view, were lack of resistance, the parties' prior relationship, and the absence of corroboration. As with other crimes, he told Estrich, he was simply calculating the odds of success in a trial.[371]

Estrich rejects this explanation, not as factually inaccurate, but as insensitive to the uniqueness of rape cases. "Because of the nature of the crime, rape is less likely to be supported by corroboration than... other crimes."[372] Rape is also unique in that, "Because of the sex and socialization of the victim, it may require less force and generate less resistance."[373] Finally, "[t]o take into account prior relationship in rape in the same way as in other crimes communicates the message that women victims, particularly of ... [nonaggravated acquaintance] rapes, are to blame for their victimization--precisely the sort of judgment that leads them to remain silent."[374] Since "rape is different from assault or robbery or burglary it should be treated differently; to treat them the same is not truly 'neutral.'"[375]

Estrich has a point. If a particular type of crime is both serious and characteristically hard to prove, there is much to be said for prosecuting more long-shot cases, especially if the crime harms a class of people, in this case women, who have understandable suspicions of the justice system. It should be recognized, however, that prosecutors have often faced long odds in acquaintance rape cases. For several reasons, conviction-rate statistics are not always available, and are not a wholly reliable guide to understanding the problem,[376] but a few examples should suffice to instill at least a modicum of empathy with prosecutors.

According to a nationwide study, prosecutors in the 1950s had only a 7% conviction rate in nonaggravated acquaintance rape cases.[377] In Travis County, Texas, between 1970 and 1976, 91% of the stranger rape defendants were convicted, compared to only 25% of the acquaintance rape defendants.[378] In Philadelphia, researchers found that the conviction rate varied from zero in rape cases where there was some ',victim precipitation" and no physical roughness or weapon to about 90% where those factors were absent.[379] A study of rape cases processed in the District of Columbia from 1971 to 1976 found that only 9% of defendants who were cohabiting, or were exspouses or ex-boyfriends of rape victims, were convicted.[380] Of defendants who were friends of the victim, 10% were convicted; of acquaintances, 19%; and of strangers, 27%.[381] These figures show a fairly linear relationship between the degree of intimacy between the victim and defendant and the likelihood of acquittal. The figures also show, contrary to the inference a reader of Estrich might make, that some prosecutors are already employing the sort of flexible standard that Estrich proposes: they have taken much greater risks of losing in acquaintance cases than in stranger cases. To be sure, few if any are willing to prosecute many cases in which an acquittal seems highly probable.[382]

The question is one of degree. In the case Estrich describes, the prior relationshipof the parties, the lack of corroborative evidence, and the woman's rather suggestive behavior made a conviction unlikely.[383] The defendant easily could have invented a motive for a false accusation: for example, by testifying that after they had consensual sex, the woman pleaded for a resumption of their old relationship, which he declined, after which she bolted from the room in a fury.

Let us assume that at least some of these weak cases should be tried. The question then becomes how many, how weak, and for which crimes? Judging by her unequivocal condemnation of the prosecutor in the case she discusses, Professor Estrich seems to believe that nearly all acquaintance rape cases should be prosecuted, regardless of the odds, provided of course that the prosecutor believes the woman and has a primafacie case.

That position would certainly make sense if losing cases had no costs. In fact, of course, a futile prosecution entails several potential costs. The first, arguably, is unfairness to the defendant. Estrich's prosecutor evidently regarded a predictably futile prosecution as unethical "harassment" of the defendant. On that issue, we agree with Estrich that prosecution is ethically permissible, if the prosecutor considers the defendant guilty in fact, and believes that the admissible evidence would support a conviction if one were obtained.[384] Under such circumstances, the accused has no legal right to a directed verdict by the judge,[385] and we see no reason why he should have an ethical right to what amounts to a directed verdict by the prosecutor. This is particularly true when, as in some acquaintance rape cases, the predicted acquittal would be based, at least in part, on legally improper or factually iraplausible considerations.[386] To be sure, the jury is entitled (in the sense that nothing will be done about it) to acquit in the teeth of the law and the facts,[387] But that does not mean that the prosecutor is obligated to engage in anticipatory nullification.

In defense of Estrich's apparently unqualified desire to prosecute difficult acquaintance rape cases, one can cite the psychological importance to rape victims of tangible manifestations of official concern. But, Estrich does not fully explore the costs of such manifestations. She informs us that the victim in this case was willing to endure the emotional roller-coaster of a trim and acquittal. However, she fails to discuss the victims of other crimes, perhaps neglected because of this probably futile prosecution. Some of those victims' stories might have been equally poignant; some of their cases might have been more winnable. This is not to say that we necessarily disagree with Estrich's conclusion that the case should have been prosecuted; only that we want more facts about the prosecutor's options.

Undoubtedly, some prosecutors have been too concerned about their "won-lost" statistics. This attitude obviously has a greater impact on crimes like acquaintance rape that are characteristically difficult to prove. But, the potential benefits and costs of unsuccessful prosecutions are difficult to calculate accurately even in a single case, and vary from case-to-case and office-to-office. For such a problem, no categorical answer is defensible.

The evidence is mostly negative concerning the impact of legal reforms on prosecutors' willingness to file charges in rape cases. Once again, Michigan is the only state in which reforms seem to have affected official behavior. In Detroit, Spohn and Homey found "an increase in the percentage of reported rapists indicted even as the number of reported cases increased."388 While conceding that this may reflect decisions by the police rather than by prosecutors, they think the latter is more likely, since "most of the legal changes involved evidentiary rules affecting the likelihood of obtaining convictions at trial."[389] Anecdotal evidence supported this conclusion.[390]

In Houston, however, the percentage of reported rapists indicted decreased as the number of reports rose, perhaps because of limited prosecutorial resources.[391] In Chicago, Philadelphia, Washington, D.C., and Atlanta, law reforms had no apparent impact on indictment rates.[392]

Similarly, Loh concluded that reforms in the state of Washington had not affected prosecutors' standards for determining "convictability."[393] Studying California data, Polk found no significant post-reform change in the police clearance rate, but a slight increase in the percentage of arrests for rape that resulted in the filing of a felony complaint.[394]

E. THE JURY

Criticisms of police, prosecutors, and especially judges are more prominent in the rape literature than criticisms of jurors.[395] Of course, legal scholars can be expected to focus on appellate decisions, but we suspect that there are other explanations as well. Few Americans are entirely free of the idea that "the people" are exceedingly wise, a tribute that is less often paid to police or lawyers. Unlike juries, appellate judges write opinions explaining their decisions. A bad opinion offers a much better target than an unjust but unexplained acquittal by a jury that can only be second-guessed by someone who attended the trial, and then only by writing a prolix discussion of the evidence. In rape law, as in other fields, we all seek judicial opinions that confirm our values.

If reformers wish to improve the chances of conviction, however, the main limiting factor is not skeptical police, cautious prosecutors, or sexist judges, but biased jurors. The empirical evidence suggests that the kinds of cases that police tend to unfound, and in which prosecutors are reluctant to file charges, and in which appellate courts occasionally reverse a conviction, are the kinds in which most juries are unlikely to convict.[396] This is not to say that the attitudes of police, prosecutors, and judges are always defensible, or always harmless, but only that they generally reflect public attitudes that are also common, apparently more common,[397] among jurors. For this reason, coupled with the prosecution's burden of proof at trial, the biases of police and prosecutors toward acquaintance rape victims probably do not substantially reduce the numbers of men who are convicted of rape. It is critical, therefore, to understand jurors'--which is to say, the public's--attitudes towards rape.

1. Kalven and Zeisel's Classic Jury Study

Kalven and Zeisel's The American Jury[398] is the most comprehensive study of jury verdicts in American criminal cases. As part of the Chicago Jury Project, the authors analyzed a nationwide sample of criminal trials. For each trial, they asked the trial judge to indicate the jury's verdict and what the judge's own verdict would have been, if the case had been tried without a jury. By comparing these hypothetical verdicts (and the judges' comments) with the juries' actual verdicts, Kalven and Zeisel were able to determine the types of cases in which judges tend to disagree with juries.

Of the 3576 trials studied in this project, 106 involved a charge of forcible rape.[399] Analyzing the 64 cases of "aggravated rape" (stranger rapes, gang rapes, and rapes in which the victim suffered an additional injury), Kalven and Zeisel found that the jury was more lenient than the judge in only 9.4% of the cases.[400] In the 42 other rape cases,[401] acquaintance rapes with no aggravating factor, the jury returned a verdict of guilty of rape only three times (7%); the trial judges who heard these cases would have convicted seven times as often, in about half of the cases.[402] As Professor Berger has remarked, this is "an amazing statistic in light of demonstrated judicial skepticism toward complainants of sexual assault."[403] Especially, we would add, since the study took place in the 1950s, when judges probably were much more likely to blame rape victims than they are today.

The jury acquitted when the judge would have convicted in 36.9% of all the nonaggravated acquaintance rape cases.[404] This computation treated a conviction on a lesser charge as a conviction. If such convictions are treated as acquittals, the judge-jury disparity widens further, with the jury being more lenient than the judge in almost 50% of the nonaggravated acquaintance rape cases.[405]

Bearing in mind that the 42 nonaggravated acquaintance rape complaints that went to trial survived scrutiny by victims themselves, who tend to report only the relatively strong cases,[406] and by police and prosecutors, two groups with a well-documented aversion to weak cases,[407] the 7% conviction rate represents a near-total nullification of the crime of rape in cases where the parties knew each other and no aggravating factor was present.

For each of 42 crime categories, Kalven and Zeisel calculated the "net jury leniency," arrived at by subtracting the percentage of cases in which the judge was more lenient than the jury from the percentage in which the jury was more lenient than the judge.[408] The result was a positive figure for all 42 categories, reflecting the juries' consistently greater leniency.[409] But, there were marked differences from one crime to another, ranging from game4aws cases, where the net jury leniency was highest (+43)[410] to "miscellaneous public disorders," where the net jury leniency was only +4.[411]

The net jury leniency for forcible rape was +18,[412] identical to simple assault,[413] and lower than murder (+29),[414] manslaughter (+29),[415] aggravated assault (+27)[416] and negligent homicide (+26),[417] as well as a number of less serious crimes.[418] Like most rape statistics, this net jury leniency includes both stranger rape and acquaintance rape, leading to a result that is as misleading as the average temperature in a Minnesota year. Using Kalven and Zeisel's data and terminology, we divided forcible rape into aggravated rape (stranger rape, plus aggravated acquaintance rapes) and simple rape (nonaggravated acquaintance rape). The net jury leniency for simple rape (+43)[419] tied for the highest of any crime category, and was much higher than any crime of comparable gravity.[420] Aggravated rape, by contrast, plummeted to the bottom of the list (+2).

The comments of the trial judges about acquittals in simple rape cases fell into a consistent pattern. Time after time the judge explained the jury's verdict by pointing to the alleged victim's behavior prior to the rape.

Consider, for example, one judge's explanation of a case wherein a young defendant was charged with raping a seventeen-year-old girl and the jury acquitted: A group of young people on a beer drinking party. The jury probably figured the girl asked for what she got.[421]

Or consider another judge's reference to the alleged victim's behavior through his narration of the facts: Complaining witness and defendant were formerly married and had two children. During the past year they had been going together with a view toward reconciliation and remarriage. The defendant had apparently spent much time and many evenings at the complaining witness' home. She denied any prior intercourse during the period since the divorce, but he claimed it continued after the divorce. The jury was of the opinion that if it was in a course of conduct which she had accepted, she was in no position to complain of her leading him on.[422]

Kalven and Zeisel also described a series of other cases in which "the point is not quite so explicit," but the judge nevertheless suggested something like an "assumption of risk" on her part:

The complaining witness alleged after several beers she entered car with defendant and three other men and was driven to cemetery where act took place.[423]

Woman involved went to public dance and was picked up by defendant. Then went to night club and permitted defendant to take her home over unfrequented road.., woman involved twice married and divorced, age 33.[424]

Prosecutrix and defendant strangers to each other; met each other at dance hall. He undertook to take her home.., rape occurred in lonely wooded area, she drinking but not drunk. He much more under influence.[425]

Kalven and Zeisel concluded that juries in simple rape cases in effect rewrite the law of rape by importing the tort concept of contributory fault or assumption of risk, acquitting the defendant of rape when they perceive that the alleged victim's conduct helped to precipitate the rape.[426]

Concerning this thesis, several comments are in order. In the first place, the concepts of contributory negligence and assumption of risk, though useful as shorthand terms, should not be taken too literally. The jurors' inclinations are clear, but their precise reasons are much less so, and may be subconscious. One ingenious theory is that people need to believe that the world is controllable.[427] When they learn that someone has suffered a minor loss, they can easily attribute her misfortune to chance. As the loss becomes more severe, chance becomes a more unsettling explanation, because it means that "it could happen to me." To alleviate this fear, people tend to attribute the victim's fate to her own misbehavior. We will call this the it won't happen to me hypothesis. Whatever its validity in other contexts, this hypothesis does not account for Kalven and Zeisel's finding that jurors are much more lenient in nonaggravated acquaintance ("simple") rape cases than in other serious crimes including aggravated (usually stranger) rape cases.[428]

Another possibility is that jurors are consciously swayed by evidence of the woman's contributory fault. Ignoring thejudge's instructions, the jurors decide that victim negligence should be a defense to a charge of rape. This might be called the conscious nullification/negligent victim hypothesis. This hypothesis does not account for some typical instances of jury leniency in acquaintance rape cases, notably cases in which the parties had been lovers prior to the alleged rape[429]--no one would suggest that it is negligent to date your lover. Contributory negligence also is an inapt term for cases in which the woman's character violated traditional moral standards but her conduct immediately prior to the alleged rape was reasonably careful: for instance, a sexually-unselective woman who was raped on a date during which her behavior was prudent.[430] This suggests that "contributory negligence" may not be the whole story.

One supposes, moreover, that contributory negligence also occurs in some, perhaps many, cases of stranger rape, for instance when the victim leaves her apartment unlocked after retiring for the night, or walks home through a dangerous area late at night when she should have taken a bus or hailed a taxi. Yet Kalven and Zeisel offer no evidence that juries tend to acquit in such cases.[431] One searches in vain for evidence that juries are lenient toward other types of criminals--thieves, for example--who prey on negligent victims. Why then do juries single out negligent victims of acquaintance rape?

In some cases the jury may feel that the woman teased the man beyond endurance, so that his rape, while legally and perhaps morally wrong, was understandable enough to deserve leniency--the provoked rape hypothesis.

Perhaps in some types of cases, including but not necessarily limited to those in which the woman behaved imprudently, jurors regard her as a disreputable character who might well lie about being raped--the untrustworthy accuser hypothesis. It is not altogether clear why this theory would not apply to stranger rapes, but one possible motivation for a false report, to obtain revenge against the accused man, would not apply if the man was a stranger.

Perhaps juries interpret the woman's contributory negligence, or her violation of sex-role norms, as enough evidence of consent to raise a reasonable doubt in a case where the corroborative evidence is weak: the evidence of consent hypothesis. Because consent is hardly ever an issue in a stranger rape case, this hypothesis would explain the absence of comparable leniency in stranger rape cases. It may also explain why juries so often acquit in cases where the parties had been lovers prior to the alleged rape: here too, the jury may regard prior intimacies as enough evidence of consent to raise a reasonable doubt.

The evidence of consent hypothesis is consistent with Kalven and Zeisel's "liberation" theory: "The jury does not often consciously... yield to sentiment in the teeth of the law. Rather it yields to sentiment in the apparent process of resolving doubts as to evidence."[432] In other words, where they regard the facts as unclear, jurors, while perhaps influenced by subconscious prejudice, interpret evidence of negligence, or perhaps other "misconduct," as evidence of consent. Thus, judge-jury disagreement is due to a combination of a genuine "evidentiary difficulty" and jurors' prejudices.[433]

In other cases, the jury may suspect that the sexual encounter was due to a misunderstandingrathe ambiguous situation hypothesis.

The lenient jurors, perhaps unconsciously, may simply be punishing women who violate traditional sex-role norms, at least if the violation increased the danger of being raped. Although such violations can occur in connection with a stranger rape, for example, if the victim walked home from a bar late at night, they may be more common in acquaintance rape cases. This might be called the sex role norm hypothesis. As we will see, this hypothesis needs some refining, because apparently only certain types of violations of sex-role norms affect the likelihood of an acquittal.

In legal theory, juries in criminal cases decide whether the defendant's guilt has been proved beyond a reasonable doubt, not how important it is to incarcerate him. But jurors may be swayed, in marginal cases, by perceptions of dangerousness.[434] They may perceive acquaintance rapists as less dangerous than stranger rapists, because a woman has greater control over her voluntary social companions than over a stranger who grabs her in a parking lot. Similarly, those who rape prudent women may seem to some jurors more dangerous than those who rape women who "asked for it." If the woman was careful, but was raped anyway, then the rapist is potentially dangerous to all women, including respectable women like many jurors and relatives of jurors. If, on the other hand, the woman behaved carelessly, for example by hitchhiking, going to the home of a stranger, or getting drunk, some jurors may perceive the man as less threatening to the safety of other, more prudent women. In a close case, this may affect the verdict. Although victim negligence occurs in stranger rapes as well as acquaintance rapes, stranger rapists are generally thought to be psychopaths who are dangerous to all women. We label this the less dangerous rapist hypothesis.

A final possibility is that the jury regards the complainant (a prostitute, for example, or the defendant's girlfriend) as the kind of woman who would not be seriously hurt by forced sex, either in general or with this manrecall this the relatively harmless rape hypothesis. Again, this attitude may be most often decisive in cases that are close on the merits.

These hypotheses are not mutually exclusive. They overlap in many cases, and interact with the burden of proof; usually several of them are plausible explanations, though of course not necessarily justifications, of an acquittal. The correct hypothesis may vary from one kind of victim "misconduct" to another. For example, Kalven and Zeisel found evidence of jury leniency in other types of drinking cases: where both parties had been drinking prior to a brawl between two men, or a domestic quarrel between husband and wife, juries tended to acquit the defendant even if his guilt was clear. Similar leniency occurred in drinking cases involving robbery, auto theft, and even in a drinking and driving scenario where the victim was a passenger eventually killed in an automobile accident.[435] Since the victims in many of these cases were men, it seems that hostility toward rape victims in drinking cases is not due, or at least not solely due, to jurors' sex-role norms. Suspicion of "loose women," on the other hand, may be due to traditional attitudes about female sexuality.

Without purporting to resolve these difficult motivational issues, or to pass judgment on the rape complainant's behavior, we will use terms like "nontraditional victims" and "sex-role-norm violations" as convenient shorthand descriptions of all rape cases in which the putative victim's behavior is likely to be criticized.

Most of the motivational hypotheses that we have described are inconsistent with the law. Contributory negligence, provocation, departure from sex-role norms, the "less dangerous rapist," and the "relatively harmless rape" theory are all invalid defenses to rape.[436] The "ambiguous situation" defense is valid, in most jurisdictions, only if the defendant made a reasonable mistake as to consent.[437] While this may have occurred in some of Kalven and Zeisel's cases, it seems unlikely that the male trial judges of the 1950s would have wanted to convict men where the evidence raised a reasonable doubt on this score. Similarly, if the "disreputable accuser" hypothesis reflected genuinely reasonable doubts about the woman's credibility, one would expect the trial judges to share those doubts.

Likewise, if the misconduct evidence were truly suggestive of consent, one would expect the trial judges of that era to think so too. In addition, we believe that there are strong theoretical objections to this justification, which we discuss in Part II of the article.[438]

Many rape scholars imply that leniency in acquaintance rape cases has no parallels in other criminal cases, at least where the victims are male. It is probably true that victim blaming defenses are more often successful in acquaintance rape cases than in any other type of criminal case. But victim blaming by defense counsel also is common in some other crimes. We have already mentioned drunken victims. In addition, there is homicide: one judge, has described "putting the deceased on trial" as "the oldest, most common and most successful tactic in homicide cases."[439]

Ultimately, whether acquaintance rape is unique is less important than whether serious injustices often occur. Although Kalven and Zeisel's study is badly dated[440] and has several methodological weaknesses,[441] it remains the best available evidence, of national scope, about the performance of juries in rape cases. Most subsequent studies and anecdotal evidence have been at least broadly consistent with what we regard as Kalven and Zeisel's central conclusion about rape cases: too often, jurors acquit acquaintance rapists.[442] We believe, therefore, that The American Jury's analysis of rape cases was fundamentally sound.

Whether it is equally sound today is more doubtful. Kalven and Zeisel performed their research during the 1950s, prior to the modern feminist movement. At least in some states, most participants in the criminal justice system believe that convictions for acquaintance rape are now easier to obtain than they used to be.[443] In federal court rape trials, the conviction rate has risen steadily since the early 1960s.[444] Linda Fairstein, a New York sex-crimes prosecutor of long experience, affirms that acquaintance rape juries have become substantially more sympathetic to the prosecution during recent years, but adds that this tends to be less so in rural areas and in jurisdictions that still lack specialized sex-crimes units in police departments and prosecutors' offices.[445]

2. Other Research on Victims and Situational Variables

For reasons of convenience and economy, most rape scholars have eschewed studies of actual rape trials, choosing instead to use college students (or, rarely, members of the general public) as the subjects of research into popular, and therefore presumably jurors', attitudes concerning rape. Typically, the scholar asks questions designed to elicit opinions about, for example, whether a rape occurred and if so whether the victim's behavior contributed to the crime.

The results of such research should be accepted with caution.[446] Obviously, student volunteers are not a representative sample of the adult population, or even, perhaps, of the student population. One's answers to a questionnaire about a hypothetical rape do not necessarily correspond to one's behavior in a jury room, where the evidence is more voluminous, decisions are collective,[447] real people are affected by the verdict, and the judge instructs the jury to apply the correct legal standards.[448] Another difference is that peremptory challenges may eliminate some of the more extreme opinions from real-world juries.

Although most studies have echoed Kalven and Zeisel's finding that victims of acquaintance rape are judged more harshly, the results have not been entirely consistent. Some studies have found that when the woman knows her attacker she is held more responsible for the rape; some have found that she is held less responsible; and a couple that it makes no difference.[449] One possible explanation for these divergent results is that some scholars have asked whether the woman was "at fault" or deserves "blame," while others asked how "responsible" she was, a less pejorative (though still imprecise) word.[450] Badlywritten scenarios may also have skewed results.[451] For example, in a famous study finding that women are blamed more when they do not know the rapist,[452] neither of the scenarios used involved a dating situation; in both scenarios the assault occurred as the victim walked alone through a wooded park at 11:30 p.m. In one scenario the rapist knew the victim; in the other he did not. Subjects may have regarded the victim's behavior as having played a causal role in the rape by a stranger, but not in the rape by an acquaintance, who they probably figured would have attacked her sooner or later in any event.[453] This is an arguable assumption, perhaps, but not one that sheds light on the behavior of real jurors, who display no tendency to treat stranger rapists more leniently than acquaintance rapists.

In a real case, the woman's causal contribution is not evaluated in isolation; it is intertwined with the jurors' evaluations of the defendant's guilt: to the degree that the woman is thought to be responsible, the defendant is not. In this critical respect the scenarios in the study were extremely artificial: the men were clearly guilty of rape but the students were asked to assess the degree to which the woman's behavior had caused the rape. In the scenarios used in this study, neither rapist contended that the woman had consented, and the students were not asked whether the men should be convicted and punished. Thus, there is no reason to believe that in a real case student jurors' causal attributions would have led to an acquittal of the acquainted rapist.

More predictably, other scholars have found that if the rapist knew the victim, subjects perceive her as less truthful,[454] as more likely to have desired or enjoyed the rape,[455] and as more responsible for it.[456] Klemmack and Klemmack studied the attitudes of women obtained from a random sampling of dwelling units listed in a city directory. They asked these women to indicate whether a rape had occurred in each of several scenarios, all of which met the legal definition of rape. The authors found that if "any relationship is known to exist between the victim and the accused, no matter how casual, the proportion of those who consider the event rape drops to less than 50 percent."[457] Another study found that subjects, especially women, felt more similar to the victims in the stranger rape scenarios than to the victims in the date rape scenarios.[458]

Studies of actual trials, though rare, generally confirm Kalven and Zeisel's finding that rape defendants who were acquainted with the victim are less likely to be convicted. LaFree, Reskin, and Visher interviewed 331 jurors who heard cases of forcible sexual assault in Marion County (Indianapolis).[459] They asked the jurors the reasons for their verdicts in thirty-eight forcible rape cases, and found that jurors were less likely to believe in the defendant's guilt if the putative rape victim had been acquainted with the defendant, however briefly, prior to the alleged assault.[460] The authors had supposed that, where the defense was misidentification instead of consent, evidence that the parties had known each other would enhance rather than diminish the credibility of the woman's testimony. But they found that "as was true for the consent/no sex cases, any prior acquaintance between the victim and the defendant had the opposite effect, net of the other variables."[461] Other studies have found that subjects recommend a less severe sentence for a convicted rapist if there had been a prior dating relationship between the parties.[462]

Some scholars have stressed another factor: the victim's departure from traditional sex-role norms.[463] Since traditional norms of female propriety coincide in varying degrees with the dictates of sexual prudence, this factor is closely related to Kalven and Zeisel's "contributory fault" theory.[464] For example, in one study participants evaluated two rape scenarios.[465] In one scenario, the victim was assaulted on her way home from a university library at dusk. This scenario was designed as one in which the woman had not violated any sex-role norm. In the other scenario, the victim was a young woman employed at a service station who, in the middle of the night, voluntarily provided a man with a ride to his car "which had run out of gas," instead of asking her male co-worker to attend to this task. The authors of the study found that subjects blamed the "norm violating" victim more for the rape, and the rapist slightly less, than those reading the "no norm violation" scenario. On the other hand, there was no difference in how serious the subjects rated the crime, and no significant difference in the behavioral intentions they attributed to either the victim or the rapist.[466] It is not at all clear, therefore, that the backwardness of some subjects' opinions about appropriate female behavior would have induced them to vote for acquittal on these facts.

Researchers have found that subjects rate a woman's willingness to have sex, and the justifiability of the rape, higher if she initiates the date, or visits the man's apartment instead of going to the movies with him.[467] One might surmise that these findings reflect traditional sexrole norms. But the same study found that subjects rated the woman's willingness to have sex, and the justifiability of rape, higher if the man paid for the date than if the woman did.[468] This presumably was due to some subjects' belief that a man who pays for a date is entitled to sex in return. Yet the woman who lets the man pay all the expenses is following the traditional sex-role norm.

If jurors were simply punishing women who violate sex-role norms, they presumably would do so in stranger rape cases as well as acquaintance rapes. Most of the evidence suggests otherwise. Studying jurors' attitudes in actual rape trials, LaFree, Reskin and Visher found that, where the defendant claimed either that no sex had occurred or that the woman had consented (typically acquaintance cases), the woman's moral character was related to verdicts. If she drank, used drugs or had reportedly engaged in sex outside marriage, jurors were less likely to believe that the defendant was guilty.[469]

In contrast, when the defendant claimed that he had been misidentified as the rapist (stranger cases), the victim's moral character was unimportant in the jurors' determination of guilt.[470] The authors concluded that, "contrary to feminist claims, our measures of victims' sex-role behavior had little effect on jurors' judgments in the identification-diminished responsibility [mostly stranger rape] cases."[471]

In another study, the authors asked 598 adult subjects to read one of several rape vignettes, which differed on the victim-defendant 0relationship, the amount of force used, and the victim's reputation.[472] The victim was raped by either a date or a stranger, who forced her to have sex either by covering her mouth with his hand and forcing her onto a couch, or by holding a knife to her throat. The victim was described as either being easy, or a virgin who did not drink alcohol. Subjects indicated their willingness to convict the assailant, whether the event was rape, the reasons for their judgments, and their opinions about the man and the woman.[473]

Using multiple regression, the authors sought to find the best model to predict the two dependent variables, definition of the situation as rape and willingness to convict. The best predictor of subjects' likelihood of defining the situation as rape was their opinion of the woman in the vignette: the more positive they were about her, the more likely they were to favor conviction.474 The second best predictor was whether the subjects thought the victim had precipitated the event.[475]

Research by Pugh supports the proposition that the victim's contributory fault affects verdicts.[476] The undergraduate subjects in Pugh's experiment read a case summary and transcripts of testimony in a rape trial. The victim had met a man in a bar and he had gone to her apartment thereafter. She said he came in to use the bathroom and then raped her; he claimed that the sex was consensual. Three versions of the transcripts, differing on the woman's moral character, were distributed to subjects. In one version the woman acknowledged that she had had intercourse before with men she picked up in bars; in another she said that she had never done this; and in the third the topic was not mentioned. Pugh asked his subjects to assess her fault and the defendant's guilt.

With a loglinear analysis, the model that best predicted verdicts contained subject gender, victim contributory fault rating, and victim moral character. Of the mock jurors who rated the victim high in fault, 58% found the defendant guilty; of those rating her low in fault, 86% found him guilty.[477]

Studies of youths as well as adults have found considerable support for the idea that if a female "leads on" a male, raping her is at least somewhat justifiable.[478] Here again the explanatory power of the concept of a "violation of a sex-role norm" depends on the nature of the violation. True, traditional morality disapproves of sexual teasing. But traditional sex-role norms also dictate that girls should not aspire to be doctors or lawyers. Yet, there is no evidence that rapes of women who have careers outside the home are more likely to lead to lenient treatment of the rapist than rapes of homemakers. It seems, therefore, that the precise nature of the "sex-role norm violation" is critical. In all likelihood, people object to women who "lead men on" simply because they regard sexual teasing as unfair, if not unbearably provocative.

Many scholars have tried to determine which attributes of the rape victim affect subjects' sympathy for her. The most spectacular finding concerned the victim's respectability. Common sense, reinforced by anecdotal evidence, tells us that, all else being equal, people will blame respectable rape victims less than, say, prostitutes, strippers, runaways, welfare mothers and others who lack middle class credentials. Yet pioneering research with student subjects seemed to prove the opposite: the more respectable the victim, the more she was blamed for having been raped.[479] To explain this amazing result, scholars offered the "just world theory." People need to believe that the world is just. A rape of a disreputable woman seems just. But a rape of a respectable woman appears to be unjust, To reconcile this apparent injustice with the premise of a just world, people need to blame the victim. So they attribute fault to the respectable victim, more so than to the disreputable victim, whose rape needs no such explanation.[480]

Other social scientists, having failed to replicate this finding, discovered methodological flaws in the original "just world" research design, and ended the silliness by concluding that society does not discriminate against respectable women.[481] On the contrary, social science has added its imprimatur to conventional courthouse wisdom, by confirming that subjects blame victims of acquaintance rape more if their sexual history violates traditional norms of female restraint,[482] or if they are intoxicated[483] or provocatively dressed.[484]

In one study, some subjects read a scenario in which the victim's prior sexual experience was not mentioned, while in another scenario she had limited sexual experience, and in a third she was promiscuous.[485] When her sexual history was not mentioned, the subjects' recommended sentence for the rapist was 8.5 years; when she had limited experience, he received 7.3 years; and when she had many previous casual sexual encounters, the average sentence was only 4.5 years.[486] The subjects perceived the rape as most serious when the victim's past was not mentioned, less serious if she had limited experience, and least serious when she was promiscuous.[487]

Thus, most subsequent research has supported Kalven and Zeisel's finding that acquaintance cases, especially those with nontraditional victims, are more difficult for the prosecution than stranger cases. However, two exceptions should be noted. LaFree, Reskin and Visher found that although in general the putative victim's characteristics were important only in acquaintance cases,[488] evidence that she drank or used drugs was associated with leniency toward the defendant even in stranger cases.[489] This finding is noteworthy because it accords with Kalven and Zeisel's report that drinking leads to victim blaming in various sorts of cases.

In a much more radical departure from Kalven and Zeisel, Spohn and Horney recently found that in Detroit the characteristics and outcomes of stranger rape and nonaggravated acquaintance rape cases "are surprisingly similar, and that there is little evidence of an interaction between type of case and victim characteristics."[490] Apparently, attitudes in Detroit have evolved to the point where Kalven and Zeisel's dichotomies between stranger rape and nonaggravated acquaintance rape, and between cases in which the victim was contributorily negligent and those in which she was not, have become invalid. The weight of the evidence suggests, however, that these findings will not be replicated in most other jurisdictions.[491]

One of the more interesting lines of research involves the victim's appearance. Some studies have found that portraying the rape victim as physically attractive increases perceptions of the defendant's guilt, perhaps because rape of an attractive woman seems more plausible, and leads to increases in the sentence suggested by subjects, perhaps because harm of a valued person is thought to warrant a more severe penalty.[492] But other studies have found no effect of victim attractiveness.[493]

Several scholars have studied the effects of racial prejudice in rape cases. Such prejudice has been most heavily publicized in cases where a black man is accused of raping a white woman.[494] Given the terrible history of lynching, and of racial discrimination in trials of black men for raping white women,[495] the resulting publicity has been entirely appropriate. But, regardless of the race of the defendant, when the putative rape victim is black, the jury is more likely to acquit than when she is white.[496] Since most rapes are intraracial, black rape defendants are more likely to be acquitted than white rape defendants--not, of course, because they are black but because their victims are usually black.

a. The Defendant's Characteristics

Considerable research has been performed on the effects of various attributes of the defendant on judgments in hypothetical (and occasionally real) rape cases. The attributes that have been studied the most are physical attractiveness, race, and socio-economic status.[497]

Research results consistently indicate that the rape defendant's socio-economic status (including his unemployment) affects perceptions of his guilt. One study found that some subjects are more likely to blame the rape victim if her assailant is well-dressed than if he is badly-dressed.[498] With otherwise identical scenarios, mock jurors were more certain of a janitor's guilt of rape than of a scientist's,499 and they imposed longer sentences on "socially-unattractive" defendants than on socially-attractive ones.[500]

Studying actual rape trials, LaFree, Reskin, and Visher created a "loser" scale. Variables that loaded highly on this scale included being unmarried, childless, unemployed, lacking a sexual partner, and presenting a negative appearance to jurors.[501] The authors found that the losers (and those with criminal records that were mentioned in court) were significantly more likely than other defendants to be thought guilty by jurors.[502] Subsequent research found that jurors are less likely to find the defendant guilty if he is employed.[503] While such findings raise legitimate suspicions of prejudice, perhaps actual guilt is statistically associated with the same "loser" qualities.

Reskin and Visher found that the defendant's race, when substituted for his employment status, showed a significant effect. However, when they included his race in the same equation as his employment status, race did not show a significant effect.[504]

Studies in other contexts suggest that subjects are often more sympathetic to simulated defendants who are good-looking, a tendency that has been attributed to a "beauty is good" stereotype.[505] A notable exception is when the crime could have bee0 facilitated by attractiveness--for example, a swindle.[506] Although in some cases the defendant's good looks might be seen as having facilitated a rape, it seems more likely, given the traditional assumption that rape is due to sexual frustration,[507] that jurors would conclude that a handsome man must be innocent, because he doesn't need to resort to rape.

In apparent confirmation of this hypothesis (or of the "beauty is good" hypothesis), some studies indicate that physical attractiveness is indeed an advantage to rape defendants.[508] Under simulated conditions, this has been found to be true even if the scenario plainly reveals that the man raped the woman.[509] Female subjects, though not males, blamed the victim more if the rapist was attractive than if he was unattractive.[510] Subjects perceived the likelihood of future antisocial behavior by the rapist as especially high if he was both unattractive and acquainted with the victim, perhaps because "this description fits the script of an unappetizing character picking out a victim and waiting for an opportunity, whereas situational explanations may be more plausible in the other conditions."[511]

Some studies have found that mock jurors' verdicts are unaffected by the defendant's looks.[512] But, a study of actual rape cases found that jurors were less likely to regard attractive defendants as guilty.[513]

The jury, in some cases, becomes aware of the defendant's prior criminal record.[514] The impact of such evidence in rape trials has rarely been studied, but common sense tells us that it may tip the balance in close cases, and a few scholarly investigations have supported this hypothesis, at least in consent-defense cases.[515]

b. Jurors' Characteristics

Scholars have devised several scales to measure subjects' attitudes toward women and rape: Rape Myth Acceptance;[516] Sex Role Stereotyping;[517] Adversarial Sexual Belief;[518] Acceptance of Interpersonal Violence.[519] Attitudes Toward Rape.[520] Attitudes Toward Women,[521] and Rape Empathy.[522] Although the results have not been uniform, a number of scholars have found, unsurprisingly, that one's general attitudes on abstract questions about relations between the sexes strongly affect one's opinions about concrete rape scenarios.[523] For example, Hubert Feild found that the attitudinal factors measured by the Attitude Toward Rape scale were more predictive of mock jurors' decisions than were the jurors' demographic characteristics.[524] Subjects imposing shorter sentences in a precipitory stranger rape case believed it was the woman's responsibility to prevent rape, that rapists should not be punished severely, that rapists are normal men, that a raped woman is no longer desirable, and that women should not resist rapists. Combining all attitudinal items, the correlation between attitudes and sentence was .51 in the nonprecipitory rape case and .45 in the precipitory case--a strong relationship. All of the mock jurors' demographic characteristics combined correlated with their sentences of the rapist only .17 in the nonprecipitory case and .18 in the precipitory case--a weak relationship, albeit statistically significant.

Most research supports the commonsensical supposition that females are more likely than males to believe an accusation of rape.[525] One assumes that both sexes tend to empathize with their own kind, at least in an ambiguous dating situation. No citation is necessary for the proposition that men, even if they are basically peaceful, tend to abhor violence less than women do. Research has revealed another gender difference: men are more inclined than women to discern sexual motivations in the behavior of a woman,[526] a tendency that presumably leads to a greater sympathy for some acquaintance rapists. Nevertheless, the evidence on gender bias is somewhat mixed,[527] and the subjects' values, traditional or feminist, appear to be more decisive than gender as such.[528]

In one study, subjects read about a date where either the man had bought the concert tickets, or the man and woman split the cost, or they happened to meet at the concert and the woman accepted a ride home with the man.[529] All three scenarios ended with forced sex in the dormitory parking lot. Subjects were then given a Sexual Experience Survey asking females about prior sexual victimization and males about prior sexual aggression, as well as the Rape Myth Acceptance Scale, measuring subjects' agreement with common skeptical attitudes toward rape victims. The authors found that males were less likely than females to label the forced sex as rape; males who had themselves been sexual aggressors were the least likely to call the scenario a rape.[530] Both males and females who were high in Rape Myth Acceptance were also less likely to view the encounter as rape.[531] On the other hand, females (but not males) were less likely to label the forced sex as rape if the man had paid for the tickets rather than meeting the woman at the concert.[532]

The critical role of traditional versus feminist values is understandable, at least in acquaintance rape cases, where the defense usually blames the putative victim. Traditionally, women have been regarded as sexual gatekeepers. The woman's responsibility, in this traditional view, is to avoid situations in which the man may misinterpret her desires, or lose control of his own. Accordingly, she should not behave in a manner that men may mistakenly interpret as indicative of sexual availability, and she should verbally make her position clear. By contrast, feminists, of either sex, are more likely than traditionalists to regard sexual equality as trumping prudential concerns, and accordingly are less willing to endorse standards of behavior that assign a major role in avoiding rape to the woman. Feminists are therefore less censorious toward female promiscuity, drinking, and the like, which often are part of the evidence in a consent-defense case.[533] At the same time, feminists are less likely than traditionalists to accept the idea that, when the woman's conduct is highly "provocative," it may ignite uncontrollable passion in the man.[534]

Age is a related factor. Women who came of age prior to the sexual revolution of the 1960s are presumably more inclined, on average, to subscribe to relatively traditional sexual norms[535] One might surmise that older women are more shocked by the behavior of "bad victims" and as a result more inclined to acquit in such cases. One study did find that older people of both sexes are more likely to believe myths about rape.[536] But on the whole the literature on this topic is too skimpy to support any firm conclusions.[537]

The relationship between feminist values and attitudes toward rape cases suggests that, if the public's attitude toward sex roles becomes increasingly egalitarian,[538] juries will be more willing to convict acquaintance rape defendants.

c. The Evidence

Most of the literature about public attitudes towards rape cases is devoted to identifying extralegal factors that affect jurors. Improper biases make a better story than rational evaluation of the evidence. What is more, an expose of prejudice is potentially more valuable to society. We have no quarrel with this, but obviously, the evidence in the case may also play a powerful role. Even the most blatantly prejudiced people are generally not entirely oblivious to reality. According to Kalven and Zeisel's "liberation hypothesis,"[539] jurors' biases should exert a more powerful influence in cases in which the prosecution's case is strewn with seeds of doubts that will germinate if watered by prejudice. A more recent study of rape trials confirms the liberation hypothesis, finding, for example, that when the prosecution's case is relatively weak, juries are more likely to convict if the accused rapist was unemployed.[540] But if the prosecution's evidence is strong, the defendant's employment status has been found to have no relationship to verdicts.[541] Of course, this does not justify any unfair discrimination against the unemployed,[542] but it does confine its scope.

Similarly, scholars have found that racial factors are most influential in cases where the evidence is relatively weak.[543]

On the other hand, one study of rape trials found that the measurable quality of the prosecution's evidence, though paramount in cases (typically stranger cases) where the defense was misidentification or diminished responsibility, did not affect jurors' judgments in cases where the defense was consent or that no sex had occurred (typically acquaintance cases).[544]

3. Effects of Rape Law Reforms on Conviction Rates

One of the main goals of rape law reforms has been to facilitate prosecution of acquaintance rape cases?[545] In analyzing whether the reforms achieved this goal, scholars have studied trends in conviction rates, measured either as the proportion of rape indictments that led to a conviction or the proportion of rape reports that led to a conviction. The former statistic, though it superficially seems to measure juries' willingness to convict, may only reflect prosecutors' willingness to file charges in weak cases. For example, if the conviction rates for rape and nonsexual assault are both 75%, the most likely explanation is not that convictions are equally easy to obtain for the two crimes, but that prosecutors are, on average, willing to take about one chance in four of losing, and do a good job of calculating the odds. This of course tells us nothing about the relative difficulty of winning rape and assault cases.

An upward trend in the conviction rate would be more informative. Again, however, one conceivable explanation is that overburdened prosecutors are taking fewer tough cases.

Most of the available empirical evidence suggests that rape law reforms are not significantly related to the percentages of indictments that lead to convictions. In the most sophisticated study of this subject, Spohn and Homey found no evidence that the conviction rate (as thus defined) had changed after state reforms in five of the six cities they studied.[546]

The sole exception, Detroit, was factually ambiguous. Although the authors found no increased proportion of convictions (relative to indictments) in Detroit, the indictment rate had increased in Detroit. Therefore, "the fact that the [conviction] rate did not decline is important," because, "if the increase in indictments represented more borderline cases entering the system, a decline in convictions might have followed without the evidentiary changes that were part of the reform."[547] Because the number of convictions increased, Spohn and Homey concluded that "defendants in these borderline cases are being convicted."[548] Whether this was due to the reforms is unclear. Michigan judges interviewed by the authors felt that juries have become more willing to convict accused rapists, not because of the legal reforms but because of the attitudinal impact of the women's movement.[549]

In a follow-up study, Spohn and Homey studied a random sample of sexual assault cases bound over for trial in Detroit Recorder's Court from 1970 through 1084, a span covering both pre-reform and post-reform periods.[550] Unlike prior research on the impact of rape law reforms, this study focused on the relationship of victims' characteristics to case outcomes. The authors expected that more cases involving nontraditional victims would be bound over for trial in the post-reform period than in the pre-reform period.[551] Their findings partially confirmed this hypothesis. On the one hand, there was no post-reform increase in the proportion of cases in which the victim had a prior sexual relationship with the offender, or in which she did not scream, physically resist, or report the crime promptly.[552] On the other hand, a larger proportion of the post-reform cases involved questions about the victim's moral character[553] and evidence of risktaking behavior.[554] The percentage of cases involving acquaintances also increased significantly.[555]

Spohn and Homey concluded that the larger proportion of nontraditional-victim cases in the post-reform period was probably due to a combination of two developments: (1) a greater willingness of such victims to report the crime; and (2) a more positive attitude toward these cases on the part of police and prosecutors.[556] Whether these attitudinal changes were due to the reforms, or instead to the changed public attitudes that led to the reforms, remains unclear.

The authors found, however, that the impact of most victim characteristics on the ultimate outcomes of cases did not decline in the post-reform period.[557] Thus, this study's findings were consistent with the earlier study by the same authors of the effects of reforms in Detroit.[558]

The statewide Michigan study relates that Michigan criminal lawyers believe that the state's rape shield law made rape easier to prove and, therefore, rape cases more winnable.[559] Supporting this anecdotal evidence, the authors found a post-reform increase in the number of rape convictions, from an average of eight forcible rape convictions per month between 1972 and 1974 to an average of 21 per month from 1976 to 1978.[560] They concluded that this development was attributable to the rape law reforms[561] Spohn and Horney argue persuasively, however, that it was probably due to an increase in the number of cases getting into the system, rather than a greater likelihood of conviction once charges had been filed.[562]

The more important question is whether reforms have reduced the post-report attrition rate, or in other words the conviction rate as a percentage of rape reports. Spohn and Homey found that in four of the cities they studied (Chicago, Detroit, Atlanta, and Washington, D.C.) the conviction rates (as a percentage of rape reports)were virtually identical, varying only by two percentage points.[563] This is a remarkable finding, because Illinois and Michigan had "strong" reforms, while Georgia and the District of Columbia were "weak" reform jurisdictions.[564] The city with the lowest proportion of reported rapes leading to convictions (9%) was Houston, which contrasted sharply with Philadelphia, where 20% ofreported rapes led to a conviction. Spohn and Homey found no evidence that either city's conviction rate had been affected by reforms.

Summarizing their findings, Spohn and Homey concluded that, "[t]he reforms did not produce an increase in the likelihood of conviction, and they produced an increase in reports and in the likelihood of indictment in only one of the six jurisdictions."[565]

Gilchrist and Homey used time-series analysis to evaluate the moderately reformed rape statutes passed by Nebraska in 1975. They found no evidence that reforms were associated with any increase in the conviction rate.[566] Loh reached a similar conclusion after studying data from Washington state, but he found that an increased proportion of convictions were for rape rather than for another charge growing out of the same encounter.[567]

It is not difficult to find plausible explanations for the apparently low impact of rape law reforms on conviction rates and case attrition.[568] Not only were reforms modest in some of the states studied; even in states with the most impressive reforms prosecutors continue to face weighty practical problems in many acquaintance rape cases. Although some reformers hoped to shift the balance of power in acquaintance rape trials,[569] forcing defendants to go on the defensive for a change, the burden of proof remains on the prosecution. As one Michigan prosecutor put it: "[y]ou can,t legislate a credible rape victim."[570] Evidence of the woman's sexual habits is still admissible in some cases, either because the shield law contains exceptions, or because judges make a discretionary decision to admit sexual history evidence when they deem it relevant. Some rape shield laws have too many loopholes,[571] but even after the loopholes are closed some sorts of evidence about the complainant's character and habits will inevitably be available to the jury. It would be unjust to prohibit all references to her sexual history; such evidence is sometimes necessary in order "to provide context" or for some other legitimate purpose such as suggesting a motive for a false accusation.[572] Even when sexual history evidence as such has been excluded, the accuser's contributory negligence and other nontraditional victim characteristics usually will be revealed in the testimony about events leading up to the rape. One cannot prosecute a man who allegedly raped a hitchhiker, for example, without revealing how it was that the victim came to be seated in a stranger's car. If the parties went to the woman's apartment to smoke marijuana and then the rape occurred, the illicR purpose of the visit cannot honestly be suppressed during the woman's testimony, and even if it could the jury might then infer that the true purpose must have been sexual. Not infrequently, the jury can draw speculative inferences about the alleged vicitm's sexual habits from the legitimate evidence about events prior to the rape, without having to hear testimony specifically devoted to that subject. All this is not to say that rape shield laws are wholly ineffective; only that the woman's lifestyle will often be fairly evident to the jury even in a jurisdiction with a reasonably tough shield law.

A few of the reformers' expectations were obviously unrealistic. To take the most extreme example, some reformers hoped that by changing the name of the crime from rape to "sexual assault" (or some similar term) they could "divert attention from questions of the victim's consent, for assault is, by definition, something to which the victim does not consent."[573] Those who entertained this hope evidently failed to consider that forcible rape is also "by definition, something to which the victim does not consent."

Another misguided reason for changing the name of the crime was to draw attention "to the seriousness and the violent nature of rape."[574] The idea that an "assault" sounds more serious, or more violent, than a "rape" is almost too fantastic to discuss, Rape has always been regarded as a more heinous crime than assault. The problem has been that many rapes are not regarded as, in Estrich's words, "real rape." If there is anything to be gained by changing the name of the crime, it can only be by making it seem less serious and therefore more like the average juror's view of many acquaintance rapes.[575] Kalven and Zeisel concluded that acquaintance rape jurors, despite their biases, often were willing to convict rape defendants of lesser crimes such as indecent assault.[576] Perhaps by calling rape "Sexual Assault," or "Criminal Sexual Conduct," and dividing it into degrees with variable penalties, reformers can make jurors more willing to convict defendants in cases where jurors tend to perceive the man as culpable, but not culpable enough to be punished as a "rapist."[577] But, since jurors usually do not know the maximum penalty for, say, "Criminal Sexual Conduct: Second Degree," this hope seems dubious. By contrast, it will be obvious to jurors that a lesser included offense like indecent assault carries a lesser penalty and stigma than rape, and so they may regard it as an appropriate compromise verdict in cases in which they believe that forcible sex occurred under extenuating circumstances.

It is easy to understand the tendency of ordinary citizens to assume that legal changes usually lead to corresponding changes in social reality. One might have expected greater sophistication from legal scholars. On the whole, however, legal scholars have been as credulous as other reformers about the likely effects of reforms. Of course, lawyers and law professors cannot be expected to reject a worldview that glorifies their own role. Even the top law schools do very little to familiarize students with the literature about the practical consequences of legal change. Law students and professors argue constantly about whether Rule X is better than Rule Y. But they relatively rarely discuss whether the choice between two rules of law will often affect the outcomes of cases; still less often do they engage in informed discussions of the effects of legal rules on social reality outside the courtroom. If taught at all, empirical studies of such matters are customarily relegated to small, optional courses like Law and Social Science. In regular courses, class discussions often include wildly implausible speculation about the potential impacts of, for instance, changes in criminal law rules, even in circumstances where few would know about the rule, or where other, more powerful motivations are obviously present.

Whether legally trained or not, reformers usually do not distinguish between bad rules of law that are often outcome-determinative and those that are not.[578] There is some evidence, for example, that the cautionary instruction in rape cases increases the likelihood of an acquittal.[579] Other rules of law, though perhaps equally objectionable, are less likely to be outcome-determinative. The corroboration requirement, for example, has appeared in every reformist critique of rape law, but in most states it may have had little or no effect on case outcomes.[580] Even prior to the reforms of the 1970s and 1980s, many states had no corroboration requirement.[581] Those that did require some corroboration usually did not insist that every element of the offense be corroborated.[582] Since prosecutors and juries want some corroboration even if it is not legally required,[583] abolition of the rule did not necessarily make convictions easier to obtain.[584]

Similarly, reformers had supposed that the resistance requirement was a major obstacle to convictions in acquaintance rape cases, a supposition encouraged by many modern commentaries on rape law.[585] Estrich, on the other hand, shrewdly observed that, since "force" and "resistance" are two sides of the same coin, it does no good to abolish the resistance requirement while continuing to define rape as non-consensual intercourse obtained by force.[586] Without resistance, there will be no "force." But not even Estrich discussed the possibility that, even if neither resistance nor force were a required element of the crime, police, prosecutors, and juries would still need to decide whether the woman consented, and in doing so would still attach great weight, in acquaintance rape cases, to the degree of force and resistance.[587]

Reformers also hoped that, by dividing the crime into degrees, they would encourage plea bargaining and discourage jury nullification.[588] But Spohn and Homey found little or no evidence that reforms had this effect.[589] Conceivably, however, the increase in the number of cases in the system in Michigan, leading to a larger number (though not proportion) of convictions was due in part to definitional changes in Michigan law.[590]

In Detroit, according to Spohn and Homey, prosecutors were reluctant to take advantage of lesser charges that had been created by Michigan's reform, because they feared that juries would become confused if they were given definitions for four different degrees of "Criminal Sexual Conduct."[591]

The reader should not infer that those who drafted rape reform legislation were unusually naive. Some feminists were pessimistic about the potential instrumental effects of reforms.[592] The optimism of the majority was characteristic, not just of rape law reformers, but of citizen-reformers in general, including: many lawyers. In fields as diverse as the Miranda rule and environmental rights, both sides usually talk as if legal doctrines are major determinants of what happens in everyday life, when in fact they often do not even determine what happens in court.[593] As often as not, the fierce doctrinal battle is largely symbolic, and the predictions (on both sides) of major practical effects are hyperbolic.[594]

This is not to say that symbolic victories are meaningless; as Holmes said "we live by symbols." A reform whose short-term impact is purely symbolic may affect attitudes that eventually change society in ways more subtle than we understand or can trace. On a cumulative, national basis, publicity about rape law reforms may already have helped to increase reporting of rape. Then too, a reform like a shield law may reduce some victims' ordeals even if it has little effect on case outcomes. Yet, it is also true that reformers, fixated on abhorrent legal rules, frequently expect more short-term instrumental benefits from law reform then mere laws can deliver. Like their allies in citizens' groups, reformist law professors often attribute excessive importance to what the law says, frequently ignoring more decisive but less tractable realities such as the burden of proof.[595] Indeed, several prominent rape scholars criticize "the law" when the context makes it clear that they mean to criticize, not rules of rape law, but how those rules are administered.[596]

Law review articles, like Hollywood movies, usually have happy endings. In traditional doctrinal scholarship a happy ending is almost guaranteed by the standard conception of the enterprise. Typically, the ultimate issue is, "What should be done?" The author begins by describing a problem--a social evil, a questionable rule of law, or a division of authority among courts. The next step is to consider the history of the problem, its ramifications and illustrations, and several possible solutions. Above all, one must be constructive; the article closes with specific proposals for legal action. One shows, for example, that a proposed rule would be more just, or that it seems likely to have valuable utilitarian consequences.

Ideally, a scholar tries to record the truth, including all the uncertainties and grounds for pessimism about law reforms. But, there is a triumphalist bias in reformers' legal scholarship. For a scholar trying to publish an article, doubts about the practical importance of the choice among possible legal rules are self-defeating: they make the article seem pointless. In order to show the importance of their proposals, many authors tend to exaggerate the effects of disfavored legal rules, unconsciously subordinating if not entirely suppressing any evidence that the objectionable rules are not outcome-determinative and that reforms will be inefficacious. The messiness and uncertainties of the real world, and our imperfect understanding of it, must be tidied up, just as they are in a movie or a lawyer's brief. This leads to the happy ending: change the law and we will improve the world.

The triumphalist bias is most pronounced in controversial publiclaw fields like rape, where readers crave solutions to newly-recognized problems. Many student editors and professors entered the profession at least partly because they were inspired by the prospect of achieving social justice through law. From their point of view, an article that acknowledges a problem but offers no legal solution, or, even worse, suggests that legal solutions are likely to fail, is at best a disappointment, undermining the profession's noblest ideals, which enshrine the goal, not so much of understanding the world realistically, as of changing it beneficially.[597]

In discussions of rape law reforms, the problem of unrealistic expectations is further exacerbated by one of the most basic rules of criminal procedure: the state cannot appeal from an acquittal. As a result, appellate decisions generally occur only in cases in which the jury convicted the defendant. This greatly affects the impression created by the cases, because it means that the court never looks more pro-prosecution than the jury. Judging by Kalven and Zeisel's findings, juries are much more inclined to acquit in nonaggravated acquaintance rape cases than trial judges would be.[598] If the appellate courts were entitled to reverse rape acquittals, and of course especially if they could do so merely because they themselves would have handed down a different verdict, their collective performance in rape cases probably would look vastly more "pro-victim" than it does, especially when compared with the relatively "pro-defendant" views ofjuties. But, since appellate courts only hear rape cases at the behest of convicted defendants, reversals create the impression that judges are more skeptical than juries toward victims. This in turn creates the impression that, since judicial decisions are the problem, new rules of law are the solution. Sometimes, of course, that is indeed the case. But sometimes the other realities dwarf the doctrines to which legal scholars devote most of their attention.

III. LENIENCY IN ACQUAINTANCE RAPE CASES: THREE PERVASIVE ISSUES

A. INTRODUCTION

The charge of massive, systemic bias against acquaintance rape victims rests at bottom on three assumptions:[599]

1. That only a minuscule proportion of acquaintance rape reports are false.[600]

2. That the guilt of most acquaintance rapists can be proven beyond a genuinely reasonable doubt, to the satisfaction of unbiased factfinders.[601]

3. That evidence about the contributory fault and "immoral" behavior of the acquaintance rape complainant, and particularly about her sexual habits, generally has no significant probative value in determining whether she consented.[602]

All of these propositions are critically important; none of them is self-evidently true; and yet none of them, except the probative value of sexual history evidence, has been analyzed at length in rape scholarship. In this portion of the article we subject the three assumptions to critical scrutiny.

B. FALSE RAPE REPORTS

In contemporary rape scholarship, the topic of false rape reports is normally discussed, if at all, only in passing.[603] Such discussions usually assume that an accusation of rape is simply an assertion of historical fact.[604] This assumption is accurate enough in stranger rape cases, where the defense is typically misidentification. With that defense, either the defendant raped the complainant or he did not; one usually cannot imagine a plausible intermediate scenario. But in acquaintance rape cases, even if the historical facts were known, and the abstract definition of rape in the jurisdiction were clear, some cases would arise in which knowledgeable observers would disagree about whether a rape had occurred. Some might say that the woman's words and acts signified consent, while others disagreed. Or some might say that the man made a reasonable mistake as to consent, while others believed that he knew what he was doing or that his mistake was unreasonable. Similar disagreements might arise about whether the alleged rapist used "force" and whether the woman was too drunk to consent.

While these ambiguities suggest that all purported tabulations of "false rape reports," whether high or low, should be taken with a grain of salt, the inquiry is not meaningless. For one thing, we suspect that most "ambiguous situation" rapes and near rapes are screened out by victims, police, and prosecutors. At least in cases that lead to formal charges, we suspect that usually the defendant is simply lying. Be that as it may, no one denies that there are at least a few deliberately false rape reports. These are our present topic.

Prior to the rise of modern feminism in the 1970s, male judges and scholars often indicated that false allegations of acquaintance rape were common.[605] Rape, Sir Matthew Hale instructed an 18th century jury, "is an accusation easily to be made and once made, hard to be proved, and harder to be defended by the party to be accused, tho never so innocent."[606]

In the early 20th century, several students of the mind, including Freud, contributed to a theory about a distinctively female type of mendacity. Women, so the theory went, are by nature masochistic. Subconsciously desiring to be raped, they fantasize about it, and some of these women develop a false but unshakable belief that their fantasies were actual rapes.[607]

Evidently influenced by this sort of thinking, some prestigious medical figures concluded that false rape accusations were widespread.[608 A number of legal scholars endorsed this position,[609] andsome recommended that rape complainants be routinely subjected to psychiatric examinations.[610] Wigmore, the most prominent of these scholars, explained:

The unchaste (let us call it) mentality finds ... expression in the narration of imaginary sex incidents of which the narrator is the heroine or the victim. On the surface the narration is straightforward and convincing. The real victim, however ... is the innocent man; for the respect and sympathy naturally felt by any tribunal for a wronged female helps to give easy credit to such a plausible tale.[611]

Without necessarily subscribing to the details of any psychological theory, many Anglo-American judges felt that false accusations were a serious danger in acquaintance rape cases. Accordingly, they fashioned legal rules, most obviously the resistance requirement, the corroboration requirement, the fresh complaint rule, and the cautionary instruction, designed to protect innocent defendants from convictions based on false rape accusations.[612]

The rape-fantasy theory has disappeared from scholarly discourse, supplanted by modern scholars' radically different analysis. Nevertheless, some men still believe that false rape reports are common.[613] Nearly all modern rape scholars consider such fears to be unwarranted,[614] and routinely dismiss Hale and Wigmore as discredited sexists who propagated myths about rape.[615] The conventional wisdom now is that the proportion of false reports is negligible, perhaps as low as 2%, a figure said to be comparable to that for most other major crimes.[616]

This new orthodoxy may well be correct, but, like the contrary opinions of Hale and Wigmore, it derives more from intuition than from common experience or scientific evidence. Although many have tried, no one has succeeded, either deductively or empirically, in demonstrating that the proportion of false rape reports is either low or high. Consider these typical arguments:

1. The Argument From Rationality

Most people generally avoid extremely irrational, serf-destructive conduct. From this premise, some scholars have reasoned to the conclusion that false reports of rape must be rare.[617] It is, to say the least, foolish to file a false rape report. Any advantage the woman might derive from a false accusation seems to be greatly outweighed by the potential adverse consequences of this type of deceit: an unpleasant, intrusive investigation;[618] a trial at which the defendant and his attorney are likely to assail the alleged victim's character;[619] possible exposure as a liar, perhaps followed by prosecution for perjury, or for falsely reporting a crime.[620] Given all these hazards, a false rape complaint is as reckless as it is unethical. Lacking any strong evidence that rape fantasies are as common as Wigmore thought,[621] it seems reasonable to suppose that false rape reports are indeed rare.

This argument is not entirely convincing. To begin, a false accusation of rape is not always as pathological or stupid as it may sound. In some cases a false rape story provides a convenient (if shortsighted) explanation of an embarrassing teenage affair or pregnancy, or an adulterous liaison.[622] According to one study, where the putative victim is a minor--as many, and perhaps most, forcible rape victims are[623]--the rape report is almost always made by someone else, usually a member of her family.[624] Even if the victim is an adult, in most cases someone else--family, friends, or a stranger--"made the decision to contact the police, acted as intermediary at the request of the victim, or persuaded the victim herself to call."[625] Of course, this does not mean that the putative victim probably is lying. But it does undercut the idea that a false rape report is necessarily a wildly irrational act, committed by a woman of mature years. In many cases, the alleged victim's only decision was to tell someone, say, her mother, that she had been raped. Lying to your mother about why you did not come home until 3:00 a.m. does not sound nearly as deviant, nor as obviously dangerous, as walking into a police station with a premeditated plan to make a false crime report.

Even if the woman's motivation is vindictive, it may be rational in a sense. For example, a prostitute whose customer fails to pay for sex may retaliate by falsely accusing him of forcible rape.[626] This is "irrational" only if she should acquiesce in her victimization, or has a better method of revenge. No doubt the most prudent course is to forget the incident, but there is no reason to assume that prostitutes are as risk-averse as most lawyers or homemakers.

Many rape complainants ultimately decide not to cooperate with the police and prosecutors. While there are other, highly plausible explanations,[627] perhaps some of these women are motivated by a dawning realization that a false report of rape is a dangerous business. If so, their "irrationality" was only momentary.

Assuming arguendo that a false rape report is invariably extremely irrational, and therefore not the sort of thing that most women would do, it does not follow that nearly all reports of rape must be true. In assessing the likelihood of false reports, the relevant population is not women in general but rather women who file reports of rape. Although it is true, as feminists have repeatedly stressed, that "any woman can be raped,"[628] it is not true that the women who claim to have been raped are a representative sample of the female population,[629] As a class, rape complainants are, disproportionately, extremely young girls or women,[630] who statistically tend to have distinctive behavioral traits. For example, rape complainants seem to be, on average, more sexually adventuresome than most women.[631] They are, according to some studies, much more likely to have a record of juvenile delinquency.[632] This is not to say that rape complainants are necessarily less honest on average than women in general; only that one cannot rule out that possibility. Of course, even if a statistical difference between putative rape victims and women in general were known to exist, it would not follow that most rape complainants are lying, nor that most of them are disreputable in any way, nor even that young, promiscuous, hitchhiking, substance-abusing runaways usually lie about rape. Our point is simply that there is a logical jump between the premise that the great majority of women would not lie about rape and the conclusion that therefore it is implausible that, say, 15% of those who file rape complaints are lying.[633]

In any event, the proportion of false rape reports cannot be determined simply by deciding whether such reports are "rare." The question is, rare compared to what? In the present context, the relevant comparison is not to all rapes that occur, most of which are unreported, but to true rape reports to the police. Yet, the same logic that might lead us to suppose that false reports are rare might also lead to the conclusion that true rape reports, at least of stranger rape, are rare. After all, in most circumstances rape is an irrational act:[634] for a momentary pleasure, the rapist creates a risk that he will be interrogated, perhaps arrested and perhaps tried for a heinous crime. Even if acquittal were certain, the crime, judged by the standards of a reasonable man, would usually be imprudent as well as immoral. This is most clearly true of stranger rape by a prior offender, where there is a substantial risk of a lengthy imprisonment.[635] Yet, we would not say that therefore most men accused of a second or third stranger rape must be innocent.

However ordinary the personalities of men accused of acquaintance rape may seem to be, they too differ, as a class, from a random sample of the male population.[636] Thus, accused rapists may tend to be less prudent and more violent than the normarive reasonable man; possibly a substantial minority of their accusers are also less honest and prudent than a "reasonable woman," although of course they may nevertheless be telling the truth about their rapes.

Although we are unpersuaded by the Argument from Rationality as applied to the original report, that argument becomes increasingly plausible the longer the putative victim persists in her accusation despite repeated opportunities to recant, and in the face of an impending trial.[637]

Discussions of false rape accusations may seem pedantic or even heartless to women such as counselors in rape crisis centers, who have been raped or who have comforted rape victims. We do not doubt the value of practical experience in judging the veracity of rape complaints. But practical experience has not produced a consensus about the danger of false rape reports. Some policemen have concluded, after many years of investigating rape charges, that false rape reports are very common.[638] For all their biases, the police deal with every formal rape complaint, and so their database, for an analysis of formal rape complaints, is superior to those of rape counselors and prosecutors. Rape counselors talk to some women who never report the crime to the police, and do not talk to some who have reported it but who choose not to seek help from rape counselors. Depending on their motivations, women who have filed false rape reports may hardly ever talk to rape counselors. Prosecutors deal only with cases that have already been founded by the police, and where the victim has not changed her mind about pressing charges.[639] It seems certain, therefore, that police hear a substantially higher proportion of false rape reports than do prosecutors. Although some police are unduly suspicious of rape complaints,[640] some prosecutors and counselors may also be biased by their roles and gender.

2. The Argument From Underreporting

Many rape victims do not report the crime.[641] Although the precise incidence of rape is uncertain, scholars agree that rape is much more common than police statistics, based on reported rapes, reveal.[642] Susan Brownmiller cites this underreporting as one of her reasons for rejecting concerns about false rape accusations.[643] But Brownmiller's argument mixes apples and oranges. To be sure, if one were trying to estimate the number of rapes that occurred in a given year, the frequency with which victims fail to report rapes would be an extremely important datum. For that purpose, one supposes that unreported rapes would greatly outnumber falsely reported rapes. But, when one is trying to determine, not how many rapes occur, but what proportion of rape reports are false, the number and proportion of unreported rapes are of little or no value. Unreported rapes and falsely reported rapes may both be common;[644] they presumably involve different types of women and girls--some perhaps too fearful, and some not fearful enough.[645]

3. The Argument from Unfounding Rates

The ratio of true to false rape reports cannot be determined by deductive logic. Can it be calculated empirically? Despite the obvious difficulties, several scholars have attempted to estimate how often rape reports are false. The conclusions of these studies vary radically, with the estimated percentages of false reports ranging from 2%[646] to 41%.[647]

The simplest way to count "false" reports is by relying on police evaluations. Since the F.B.I. compiles police unfounding rates, broken down by crime, this solution has the apparent advantage of providing a definitive, national answer. Relying on unfounding rates, Professor Alan Dershowitz argues that false rape reports are more common than false reports of other crimes:

According to FBI crime statistics, 8.4 percent of all reported rapes turn out to be "unfounded." That percentage translates into more than eight thousand false rape reports each year. This number is dramatically higher than the number of false reports of other serious crimes. The comparable figures for assault, for instance, are 1.6 percent; burglary, 3.8 percent; larceny, 1.2 percent; motor-vehicle theft, 4.2 percent; murder, 2.3 percent; and robbery, 3.5 percent.[648]

The accuracy of Professor Dershowitz's Argument from Unfounding Rates depends on the extent to which police departments make accurate determinations of whether reported rapes actually occurred and report to the FBI as "unfounded" only those reports that they believe were false. In both respects, Professor Dershowitz's faith in the FBI figures is unwarranted. For several years, the UCR figure for unfounded rape reports has been about 8%,[649] but it formerly was 15%.[650] Does Professor Dershowitz believe that women have become more truthful?

No one knows how often the police mistakenly conclude that the alleged victim is lying. This much is clear: evaluating the truthfulness of a rape complainant is often a difficult and highly subjective enterprise. Perceptions of the complainant's truthfulness seem to be influenced by the investigator's gender,[651] experience,[652] and attitude toward the role of women in society.[653] According to many studies, the police tend to disbelieve rape reports by women who suffer from mental problems[654] or whose conduct was contributorily negligent or violative of traditional moral standards or who are black or poor.[655] Conceivably, some or even all of these groups are indeed disproportionately likely to file false rape reports. However, since they are all groups against which it is plausible to suppose that some detectives are biased, the elevated unfounding rates are at least suspicious. To the extent that improper biases lead to unfounding of truthful rape reports, police statistics are inaccurate, perhaps grossly so.

Even if we were to assume that police investigators generally make accurate appraisals of the woman's veracity, unfounding rates often do not reflect those appraisals. Police departments employ the term "unfounded" to cover not only reports that they believe to be false, but also cases that they choose not to pursue for various other reasons,[656] such as the lateness of the victim's report to the police, lack of corroborative evidence, lack of cooperation by the victim or some other witness, reporting in the wrong jurisdiction, the victim's drunkenness, her drug usage, or the police perception that she is a prostitute.[657] "[S]ome police agencies report all of their unfounded rape cases to be due to false allegation, while other agencies report none of their unfounded declarations to be based on false allegation."[658]

It is unlikely that most police investigators distinguish sharply between deciding that a rape complainant is not credible, meaning that she seems to be lying about having been raped; deciding that she is not credible, meaning only that she would not be a good witness, for example because she is promiscuous or "a mental case" or was "asking for it"; and deciding that "it's her word against his" and therefore no jury will convict.[659] In theory, the decision about whether a case is worth prosecuting is made by the prosecutor's office,[660] but in practice the police often unfound rape complaints by disreputable women, even when there is no additional reason for doubting their stories.[661] In some jurisdictions, prosecutors advise police about unfounding decisions,[662] a practice that seems likely to blur even further any distinction between what is true and what is thought to be provable beyond a reasonable doubt to the satisfaction of a skeptical jury.

The UCR's rape unfounding rate has declined over the years.[663] Moreover, in any given year there are sometimes enormous differences in unfounding rates from one city to another.[664] The most plausible explanation of inter-city variations is that different departments employ different founding criteria, and sometimes even different definitions of "rape," which they periodically revise. Two recent newspaper stories illustrate this phenomenon. One article discusses a rather high number of unfounded rape reports in the Washington, D.C. area.[665] The article notes that in the Northern Virginia suburbs, the city of Fairfax had the highest rate of unfounded rape reports, apparently because Fairfax, unlike other nearby jurisdictions, includes cases in which the victim decides not to cooperate in prosecution.[666] Since the alleged victim's decision not to press charges may be motivated by other considerations,[667] it is not a reliable measure of the veracity of her original report.

The second article discusses an Oakland police department decision to reopen 90% of the cases that they unfounded in 1989 and 1990.[668] The department stated that among their reasons for the unfoundings were victims who were uncooperative, difficult to locate, engaged in prostitution, or known to their assailants, and lack of time to conduct proper investigations.[669]

The lesson, in brief, is that unfounding criteria vary from jurisdiction to jurisdiction, and the UCR figures are not a reliable tabulation of false rape reports. This is not to say that the UCR's unfounded rate necessarily overstates the percentage of false rape reports. It may, but this cannot be proved. Some police departments do not even record the complaints that they regard as least credible,[670] creating the possibility that some false reports are not tabulated at all. The point, in other words, is that unfounding rates are untrustworthy evidence of the proportion of false reports, not that they are necessarily too high.

4. Other Empirical Research

Undaunted by the obvious difficulty of the task, several scholars have tried to calculate the proportions of false rape reports in various samples. Arthur Frederick Schiff studied 100 cases of alleged rape seen in the Dade County (Florida) Medical Examiner's Office over a fourteen-month period. He concluded that "seven percent were definitely not rape and fifteen percent were questionable."[671]

In a more comprehensive 1977 study,[672] Geis, Wright, and Geis obtained the opinions of 128 of the approximately 500 members of the Association of Police Surgeons of Great Britain. "In your estimation," asked the authors, "about what percentage of these claims [of rape that you examined during the past year] were false (that is, no crime had taken place)?"[673] Of 1361 rape reports seen by these physicians, 426 (31.4%) were judged to be false.[674] For the five female surgeons, however, the proportion of perceived false complaints fell to 43 out of 185 cases (23.2%).[675] Agreeing almost exactly with these women, the dozen surgeons who had examined the most complainants (655 in all) thought that 155 (23.5%) had made false charges of rape.[676] But the three surgeons who had seen the very largest number of cases (cumulatively 300) thought that only 12 (4%) were based on false complaints.[677] As the authors concluded, "the wide discrepancies in estimates suggest that in all probability something other than (or in addition to) the realities of the situation contributed to the answers."[678]

Further evidence of the subjectivity of "false-report" figures is provided by a study done by the New York City Sex Crimes Analysis Squad.[679] This investigation covered all allegations made to the Unit over a two-year period. They found that the rate of false allegations for rape and sexual offenses was around 2%, which was comparable to the rate for other crimes.

Susan Brownmiller cites the New York experience as evidence that "false" rape reports are due mostly to male bias: "When New York City created a special Rape Analysis Squad commanded by policewomen, the female police officers found that only 2 percent of all rape complaints were false--about the same false-report rate that is usual for other kinds of felonies."[680]

At least for academic purposes, it might be more revealing to give rape reports credibility ratings rather than conjectural "true" or "false" designations. In the only example we found of this methodology, investigators asked social workers to interview 577 women who had reported rapes to the Philadelphia General Hospital. (Evidently they did not interview the alleged rapists.) Twenty-four of these women admitted to the social workers that their rape reports had been false.[681] The social workers rated the credibility of all 577 cases, on a scale of one to four, with one and two defined as "partially incredible" and three and four as "credible." (The lower scores signified inconsistencies in the women's stories, but not necessarily that they had not been raped.) On this basis, the social workers gave high credibility ratings (three or four) to all but 12.7%.[682]

In some jurisdictions, even the police seem to believe nearly all rape reports. Carolyn Hursch and James Selkin studied Denver police records for 1973. They designated rape reports as false "only if (1) the victim finally admitted that she had originally lied about being attacked, or (2) the police concluded from other evidence that she had lied."[683] Employing these criteria, Hursch and Selkin found that only 2 to 3% of rape reports were false.[684]

Professor Eugene Kanin reached a drastically different conclusion after examining the 109 consecutive rape reports during a nineyear period in a "small metropolitan community."[685] He labeled a complaint false only if the complainant later admitted to the police that no rape occurred.[686] In the city Kanin studied, departmental policy forbade police officers to use their discretion in deciding whether to officially acknowledge a questionable rape complaint.[687] The recantations tallied by Kanin "did not follow prolonged periods of investigation and interrogation"[688] and those who chose to recant were subsequently "informed that [they would] be charged with filing a false complaint, punishable by a substantial fine and a jail sentence."[689]

Kanin's finding was perhaps the most striking and counter-intuitive in the history of rape scholarship: of 109 forcible, completed rapes, 45 (41%) were false by his definition.[690] He found no evidence that false rape charges were made by delusional women or for the purpose of extortion.[691] Instead, "false rape allegations appear to serve three major functions for the complainants: providing an alibi, seeking revenge, and obtaining sympathy and attention."[692]

The false complainants were all white, and largely of lower socioeconomic background (only three had any post-secondary education). Their mean age was twenty-two. From the sketchy information available, Kanin could discern no difference between the characteristics of this group and rape complainants who had not recanted.[693]

Of the forty-five charges that were recanted, 56% "served the complainants' need to provide a plausible explanation for some suddenly foreseen, unfortunate consequence of a consensual encounter, usually sexual, with a male acquaintance."[694] Kanin provides several examples of these "alibi" cases:

[(1)] A divorced female, 25 years of age, whose parents have custody of her 4-year-old child. She lost custody at the time of her divorce when she was declared an unfit mother. She was out with a male friend and got into a fight. He blackened her eye and cut her lip. She claimed she was raped and beaten by him so that she could explain her injuries. She did not want to admit she was in a drunken brawl, as this admission would have jeopardized her upcoming custody hearing.

[(2)] A 16-year-old complainant, her girlfriend, and two male cornpanions were having a drinking party at her home. She openly invited one of the males, a casual friend, to have sex with her. Later in the evening, two other male acquaintances dropped in and, in the presence of all, her sex partner "bragged" that he had just had sex with her. She quickly ran out to another girlfriend's house and told her she had been raped. Soon, her mother was called and the police were notified. Two days later, when confronted with the contradictory stories of her companions, she admitted that she had not been raped. Her charge of rape was primarily motivated by an urgent desire to defuse what surely would be public information among her friends at school the next day, her promiscuity.

[(3)] A 37-year-old woman reported having been raped "by some nigger." She gave conflicting reports Of the incident on two occasions and, when confronted with these, she admitted that the entire story was a fabrication. She feared her boyfriend had given her "some sexual disease," and she wanted to be sent to the hospital to "get checked out." She wanted a respectable reason, i.e., as an innocent victim of rape, to explain the acquisition of her infection.[695]

Relying on the women's own verbalizations during recantation, as recorded in police files, Kanin calculated that 27% of the false rape reports were motivated by a desire for revenge against a male who had rejected the complainant, either initially or after an affair.[696] For example, "a 16-year-old reported she was raped, and her boyfriend was charged. She later admitted that she was 'mad at him' because he was seeing another girl, and she 'wanted to get him into trouble.'"[697]

According to Kanin, about 18% of the false charges clearly served the function of getting attention or sympathy for the complainant.[698] For example, "[a]n unmarried female, age 17, had been having violent quarrels with her mother who was critical of her laziness and style of life. She reported that she was raped so that her mother would, 'get off my back and give me a little sympathy.'"[699]

Kanin later examined police records of all forcible rape complaints during the previous three years at "two large Midwestern state universities." At both universities "the taking of the complaint and the follow-up investigation was the exclusive responsibility of a ranking female officer. Neither [university] employed the polygraph and neither declared the complaint false without a recantation of the charge."[700] Of sixty-four forcible rape complaints in the two universities, exactly half had been recanted.[701]

Regrettably, Kanin's tabulations did not distinguish between acquaintance rapes and stranger rapes.[702] Is there a more basic flaw in his research design? It seems unlikely that he happened to study three exceedingly atypical jurisdictions, and he assures the reader that the city he studied was not distinctive in any obvious way such as being a resort town.[703] Why, then, have other scholars failed to uncover similar recantation rates in other jurisdictions? Professor Kanin's explanation is that the chief of police in the city he studied insisted that every rape report, however implausible, be thoroughly investigated, a practice that was feasible because of the low crime rate in that city.[704] In other cities, busy police sometimes do not even record the most obviously false rape reports.[705]

One possibility is that the women who retracted their accusations in the jurisdictions Kanin studied are largely the same women who are described in other studies as simply unwilling to press charges. Other scholars have found that a sizeable proportion of those who report rapes are highly ambivalent about whether to press charges;[706] many of them ultimately decide not to do so.[707] Perhaps many of these women and girls had filed false reports; their subsequent unwillingness to cooperate with the prosecution may have been motivated by fear of exposure as a perjurer, or by a feeling at some point that the false accusation had achieved its purpose without a trial. This hypothesis does not explain why, in the jurisdictions that Kanin studied, the women who had filed false reports chose to recant rather than offering less embarrassing explanations for their decisions not to press charges.

Perhaps the recantations rather than the original accusations were false. Although Kanin was not present during the interrogations, he effectively rebuts the possibility that recantations were signed by women trying to escape overbearing police.[708] However, he does not consider the possibility that the police said or implied that without a recantation the case would proceed to trial irrespective of the woman's wishes to the contrary.[709] That sort of police policy might well generate a spate of false recantations, motivated by a desire to withdraw the complaint for some reason other than its falsity. At the same time, such a policy would presumably eliminate the phenomenon of investigations terminated because of the victim's failure to cooperate.

The only conclusion that safely can be drawn from these studies is that false rape complaints may be much more common than most recent rape scholars have supposed. If Kanin's findings are replicated by other researchers,[710] the skepticism of police and prosecutors toward some rape complaints may appear to be based on a wise appraisal of the facts rather than an improper bias against certain types of victims. But, it would not follow' that innocent men are being convicted of rape, nor that many acquittals in rape trials are justifiable. All that Kanin found was that many false reports of rape are made and later retracted. As Kanin acknowledges, this claim does not entail the quite different claim that many false reports are made, never retracted, and believed by police, prosecutors, and juries. We do not know, for example, what proportion of the rape complaints that were retracted in Kanin's study would have been founded by the police if they had not been retracted. In some of Kanin's false-charge cases, the woman claimed that she had been raped by a stranger, whom she never identified.[711] Obviously, such cases would not have led to a prosecution. In the remaining cases, Kanin reports: "These women were not inclined to put up a steadfast defense of their victimization, let alone pursue it into the courtroom. Recantation overwhelmingly came early and relatively easily."[712]

One supposes'that the purposes of false rape accusations can usually be achieved without the trouble and risk of accusing an acquaintance and testifying in court. For example, if the putative victim is a girl who craves attention and sympathy from her parents, she presumably need only persuade them that a stranger raped her. Similarly, if she needs an excuse for venereal disease or pregnancy, an imaginary attack by an unidentified stranger suffices. Even if her motive is revenge, she may feel satisfied after the male has been embarrassed, grilled by the police, and perhaps forced to retain' counsel.

Considering the high pre-trial case attrition rate,[713] the skepticism of some police toward acquaintance rape accusations,[714] the large numbers of women who everitually decide not to cooperate with the prosecution,[715] the reluctance of prosecutors to take on weak cases,[716] and the historic reluctance of jurors to convict men charged with acquaintance rape,[717] mistaken convictions may well be less likely in acquaintance rape cases than in, for example, stranger rapes and other cases based on eyewitness identifications.[718]

Nevertheless, it is important to find out whether Kanin's findings can be replicated. A false accusation of rape is an injustice, even if later withdrawn. While there are many reasons, including Kanin's study, for concluding that few false report cases will be pursued to trial and conviction, the pressures on a rape complainant are not all in the direction of abandoning the case; having made the accusation, she sometimes comes under pressure from police and others to assist in the prosecution.[719] If the rate of false accusations is found to be high, this will inevitably color one's judgments about uncorroborated accusations. Indeed, if the rate of false accusations is even half as high as Kanin found, the literature about official discrimination against rape complaints will need some major revisions.

C. THE BURDEN OF PROOF

Rape scholars commonly ignore the burden of proof, treating bias against victims as the major cause of attrition in rape cases.[720] Beyond question, bias is a serious problem, but to focus exclusively on bias is misleading, because many acquaintance rapes present difficult proof problems that help to explain the high attrition rate.

Consider, for example, a fairly realistic hypothetical case. Sally and Ted were college students on a date. After watching a movie, they went to his room for a drink and conversation. Sally says that when she resisted his advances, Ted raped her. He claims that he seduced her with professions of love but then told her that he did not want to be "tied down" to one woman, "and so we'd better not date anymore." Sally became angry and got revenge, Ted's lawyer suggests, by telling the police that he had raped her. She had a scratch on her left forearm but no other physical injury.[721] She filed her rape report promptly, or, if you prefer, she did not.[722] Both parties sound sincere and tell stories that are internally consistent and consistent with the physical evidence.

On this evidence, a jury may choose to believe Sally rather than Ted, or vice-versa.[723] In that sense, the prosecution has carried its burden of proof. But to appraise the jury's behavior, one needs to ask, not merely what the jury was legally entitled to do, but what it should have done. (After all, the jury is also legally entitled to acquit.) In our hypothetical case, with proof problems that are presumably typical of many acquaintance rapes, what should the jury do?

Certainly one can make a powerful argument for acquittal. In the first place, there should be no criticism of "victim blaming" until the legal standard of proof has been met. To rebuke those who "blame the victim" is to assume the very point at issue in a consentdefense case: that a rape rather than a seduction occurred. We do not doubt the defendant's factual guilt in the great majority of criminal cases, including rape. But of course, the presumption of innocence requires the jury to focus, not on whether Ted is probably guilty in fact, but on whether the prosecution has proved his guilt beyond a reasonable doubt. To the extent that the difficulty of convicting acquaintance rapists is due to genuinely reasonable doubts about defendants' guilt, case attrition is not evidence of a malfunctioning criminal justice system. Quite the contrary, under such circumstances the rapist's escape from justice, while deplorable in a sense, is evidence that the system is functioning exactly as it should.

Despite the centrality of proof problems, the most influential modern book about rape law, Estrich's Real Rape, alludes to the burden of proof only in passing, and never intimates that it might be responsible for much of the difficulty of punishing acquaintance rape.[724] Given Estrich's reformist agenda, we do not criticize her for confining her discussion to rules that she thought needed to be reformed. But, the inevitable effect of focusing on questionable rules of law is to imply that the main obstacles to rape convictions are egregious judicial doctrines and the cultural attitudes that the doctrines express. The reader is led to understand that the problem is solvable, if not by changing rules of law then by changing the attitudes that lawinterpreters bring to their task. Nothing so intractable and sacrosanct as the burden of proof stands in the way.

Estrich's failure to deal with the burden of proof is typical of modern rape scholars. For example, some authors complain that, as one book puts it, "[t]he victim is placed in the precarious position of having to prove that her reaction [to the rapist's advances] was sufficient to establish nonconsent."[725] That sounds unfair. Why should the victim have to prove anything? The answer, of course, is that the law presumes that she is not a victim, unless and until the jury finds the defendant guilty. Because Sally is the prosecution's leading witness, and the prosecution has a heavy burden of proof, "she" quite properly "has to prove" beyond a reasonable doubt that the defendant raped her. So let us rephrase the scholars' complaint: "In rape, the prosecution must prove the essential elements of the crime, including the alleged victim's nonconsent, beyond a reasonable doubt, relying chiefly, especially in the absence of strong corroborative evidence, on credible testimony by the alleged victim." As thus rephrased, the proposition is unremarkable and connotes no unfairness. Yet the essential practical difficulty remains: the burden of proving nonconsent is on the government and therefore, in a loose sense, on the government's leading witness.

Since suspicion of acquaintance rape complaints is observable even outside the criminal process, and sometimes is carried to outrageous extremes, it seems clear that improper biases do hamper some rape prosecutions. Our point is that the two elements, legitimate burden of proof concerns and excessive suspicion, are often hard to disentangle. Indeed, both elements tend to peak in the same sorts of cases: acquaintance rapes with little or no corroborative evidence, and a woman whose behavior was at least slightly suggestive of consent. Arguably, police and even prosecutors should pay less attention than they often do to a jury's likely verdict, and thus to the burden of proof. But one's evaluation of these officials is likely to be more charitable if many or even most of the cases that they decline to pursue are ones in which a jury neither would nor should return a verdict of guilty.

It is obviously impossible to determine precisely what proportion of the men who commit rape could be proven guilty beyond a genuinely reasonable doubt. That the prosecutor's burden is a serious obstacle, however, seems clear. Even without an instruction on the burden of proof, most jurors would surely vote to acquit in cases in which they felt some doubt about the defendant's guilt. Until relatively recently, rape was a capital offense in some jurisdictions,[726] and it still carries an enormous stigma and a relatively severe maximum penalty.[727] The reluctance of jurors to convict a man of such a serious crime on the basis of the largely uncorroborated testimony of a single witness is at least somewhat understandable, and does not necessarily signify discrimination against women.

To be sure, even in a simple dating situation, there may be a great deal of additional evidence. But as the televised William KennedySmith case illustrates, the reasonableness of doubts may still be debatable. The alleged victim met Smith in a bar in Palm Beach, Florida.[728] She later drove him to the house where he was staying with several members of the Kennedy family.[729] They went for a late-night walk along the beach.[730] She testified that he suddenly disrobed, went for a swim, and as she tried to depart he tackled her and raped her.[731] He testified that the sexual encounter was consensual, and continued until, by a slip of the tongue, he called her "Cathy," which was not her name.[732]

Smith's accuser claimed she was raped on the lawn in front of the Kennedy house, and that she yelled while being attacked.[733] Stephen Barry, a Manhattan prosecutor and friend of the Kennedy family, was staying in a bedroom of that house with his wife on the night of the alleged rape.[734] He testified that from the bedroom he had often overheard conversations of people sitting on a bench near the area where the alleged rape occurred.[735] Yet on the night in question he heard no screams nor sounds of any struggle.[736] Senator Edward Kennedy echoed this testimony. He testified to hearing no screams, even though his bedroom windows were open.[737]

Jean Smith, the defendant's mother, was also in the house on the night of the alleged rape.[738] She testified to hearing nothing.[739] The defense added to the impact of this testimony by calling a Miami meteorologist to the stand.[740] He said that on the night of the rape, there was little wind.[741] His testimony suggested that if the woman had been raped where she said she was, someone would have heard her.[742]

Perhaps the most controversial of the defense witnesses was Raphael Good, a professor at the University of Miami.[743] Smith's accuser alleged that Smith's penis was only partially erect during the rape, and that she struggled to prevent him from penetrating her.[744] Good testified that under such circumstances, penetration would be highly unlikely.[745] Yet Smith's semen was found inside his accuser.[746] This suggested that penetration did occur, and that perhaps it was consensual, as Smith claimed.

As further evidence that the encounter was consensual, the defense introduced the woman's undamaged clothing. After the trial, one juror pointed to this evidence as a reason for the acquittal.[747] To explain why the clothing was not ripped, the prosecution called a forensics expert in the Palm Beach County Sheriff's Department.[748] She testified that clothing is damaged in only about 10% of rape cases.[749]

Because Smith claimed that he and the woman were intimate on the beach, much of the trial concerned the evidence of sand found in the woman's underwear. The head of Michigan State University's forensic science programs testified that the inorganic material "could not have come from the lawn, but could have come from the beach."[750] An FBI authority on sand and soil countered by noting that sand "can go wherever wind, water and people move it, including the oceanfront lawn."[751]

Because the woman claimed that Smith attacked her on the lawn, grass was also a major issue. A botanist testified that "tiny traces of a plant like material did turn up on the woman's clothing."[752] However, the defense's botanist testified that fragments of grass could have easily blown onto the beach.[753] Also for the defense, the director of the Connecticut State Police crime laboratory testified that had Smith actually thrown the woman to the lawn, trace material would have remained on her outer clothing, not just inside her panties."[754]

Smith's attorney emphasized a number of alleged gaps and inconsistencies in the putative victim's account of the rape. For example, she told police that she thought they had kissed after Smith emerged nude from his ocean swim; in court, she testified that the kiss occurred before he went swimming.[755] She was unable to recall whether Smith had any clothes on after he went for his swim, when he chased her.[756]

The prosecution had one important piece of uncontroverted physical evidence. The alleged victim's panties were covered with Smith's semen, though Smith claimed she removed them before the encounter.[757]

When one considers the evidence at trial plus inadmissible evidence of three prior sexual attacks by Smith,[758] it seems probable that he did rape the complainant. Even without the evidence of prior sexual assaults by Smith, we believe that his guilt was established by a preponderance of the evidence. But we are not at all sure that the jury should have convicted him under the reasonable doubt standard. At best, the prosecution's case was exceedingly close to the reasonable doubt borderline. Yet the major law review article on the case, by Professor Estrich, does not even discuss the burden of proof. To Estrich, the verdict reveals "how far we have yet to go," "how difficult it still is for a woman to be believed," and that the public "presumes" that victims of date rape are lying.[759]

One traditional way to deal with swearing contests in rape cases was the much-criticized corroboration requirement. Similar requirements have been imposed in a few other circumstances. For example, in some states a conviction for obtaining property by false pretenses cannot rest on the uncorroborated testimony of a single witness.[760] Many states do not allow criminal convictions on the uncorroborated testimony of alleged accomplices.[761]

In general, however, a crime victim's testimony need not be corroborated,[762] and most of the handful of exceptions seem to reflect distrust of certain types of allegations--for example, hearsay statements by alleged victims of child abuse[763]--rather than a general principle of proof. Unless false rape accusations are extraordinarily common, abolition of the corroboration requirement was justifiable.

To break the evidentiary stalemate in a sweating contest, one naturally focuses on the relative credibility of the parties. The accused has a motive to lie. Does the accuser? If the defendant offers no plausible theory as to why he has been falsely accused, then the prosecution may well prevail. That is what happens when, for example, a store clerk identifies the defendant as the one who robbed him. (In fact, misidentification is often a plausible defense, but juries are notoriously inclined to overrate eyewitness testimony.)[764] Unfortunately, the acquaintance rapist often can concoct a fairly plausible story of a seduction followed by a false accusation. By its nature, the crime occurs in private. Kalven and Zeisel found that eyewitnesses other than the victim testified in only 4% of rape trials, much less frequently than in cases of homicide (44%) or burglary (20%).[765] (For acquaintance rape, the figure might have been even lower than 4%.)

Some studies indicate that the level of violence and injury in acquaintance rapes tends to be greater than in stranger rapes.766 However, other studies find violence more common in stranger rapes.767 In any event, most rape victims are not physically injured,768 and minor injuries such as bruises and scratches usually can be explained away as due to drunken stumbling, rough (but consensual) sex, sports, or accidents.[769]

In the kinds of acquaintance rape cases in which convictions are most difficult to obtain, the consensual sex scenario offered by the defense usually is not inherently implausible. In such cases the parties are of a similar age, and the woman is obviously sexually active--sometimes even a prostitute. Voluntary sex, unlike, say, large gifts of money to strangers on the street, is a common activity. Proof that intercourse occurred does not help to rebut a consent defense. Since ex4overs frequently become bitter enemies, and people often wish to conceal nonmarital, consensual sex, the defendant often can easily invent a grievance or other motive for his usually very youthful accuser to lie. In the absence of powerful corroborative evidence, it is his word against hers, and to a respectable citizen sometimes both parties look unsavory.

To make matters worse, rape complainants sometimes give inconsistent testimony or lie about some of the circumstances of the date, for example in order to hide embarrassing conduct (drugs, "provocative" behavior, infidelity, etc.), seriously undermining their credibility on the consent issue.[770]

Most lawyers seem to believe that juries can resolve swearing contests by considering the demeanor of the defendant and his accuser. A mass of social-scientific evidence suggests that this is a myth: people generally cannot determine whether someone is lying by observing his or her demeanor.[771]

Thus, the evidence in an acquaintance rape case will often be inconclusive. If both the defendant and his accuser offer plausible accounts of what transpired, one might say that, while legally free to decide either way, the jury ought to conclude that the prosecution has not met its heavy burden of proof.

Rape is not the only crime that is difficult to prove beyond a reasonable doubt without corroborative evidence. In several other types of cases that are at least arguably analogous to typical acquaintance rapes, most civil libertarians would surely favor acquittal. For example, what if Sally had accused Ted of soliciting the crime of theft? According to her story, they went to her room for a drink after a movie. During conversation about teachers, they discovered that they both knew Professor Dixby. Sally was Dixby's research assistant; Ted was enrolled in one of Dixby's courses. After a few drinks, Ted offered to pay her $500 for a copy of the forthcoming exam in the course, which he suggested that she could find in Dixby's office some evening. She refused, and promptly reported the incident to the police.

Ted told a different story. In his version, Sally had made sexual advances to him, which he had rudely turned aside, making some tactless comments about her homeliness. Incensed, she had ordered him to leave, and the next day he found himself charged with soliciting theft.

As in the Sally-Ted rape hypothetical, a jury would be legally entitled to believe Sally and disbelieve Ted. But we think that in the absence of corroborative evidence most juries should and would acquit.

Consider, for another example, what would probably take place if, instead of prosecuting Ted for rape, the government prosecuted Sally for filing a false report. In that event, the evidentiary shoe would be on the other foot and the jury probably would acquit her. In that kind of case, the man would be the accuser; usually his character would come under vigorous attack by the defense, and in most cases it would be fairly easy to raise a reasonable doubt about his credibility.[772]

To obtain firsthand accounts of similar proof problems, we interviewed several veteran criminal lawyers in the Twin Cities (Minnesota). Each was asked to, "Name some crimes where it is difficult to obtain a conviction because the only strong evidence is the victim's testimony, thus reducing the trial to a swearing contest." The lawyers were then asked to recall the usual outcome of such cases in their own experiences.

All the lawyers named nonsexual assault as a crime where swearing contests often occur. One defense attorney observed that such cases, "generally involve a race to the courthouse where whoever wins is the 'victim.'"[773] He noted that in his experience the prosecutor's burden of proving the elements of nonsexual assault beyond a reasonable doubt creates a strong advantage for defendants.[774] The same attorney explains that, "without statements indicating premeditation, the constitutional burden of proof makes it nearly impossible for the prosecutor to prevail."[775]

A prosecutor agreed:

We have to prove guilt beyond a reasonable doubt, and this tips the scales in favor of the defendant. As prosecutors, we try to counteract this by turning the trial into a credibility game, and asking the jury to simply find for who they believe. If we didn't, it would be hard to get a conviction in assault cases."[776]

A public defender concurs that defendants can often raise a reasonable doubt in a nonsexual assault case, by attacking the character of the alleged victim.[777]

Such comments are impressive in light of a study finding that prosecutors enjoy certain advantages in nonsexual assault cases, as compared to rape cases: witnesses are more often available; victims tend to be older; and the assailants are more often strangers whom the victims had resisted and who had inflicted physical injuries.[778]

Recalling several cases in which he defended women charged with prostitution, another attorney notes that, in his experience, the female defendant usually prevails despite the testimony of a male undercover officer, in the absence of physical evidence to support his claim that the woman offered sex for money.[779] If bias against women who have violated traditional morality were overriding all other factors, the contrary should be true: juries should tend to rule against women charged with prostitution, most of whom are obviously prostitutes, or at least "loose women," even if not guilty on the occasion in question. That jurors often acquit such women suggests that at least in some contexts the burden of proof is more decisive than the woman's respectability.

An experienced prosecutor claims that even in child sexual abuse cases, where the victim is most sympathetic, convictions are difficult to obtain.[780] "Without physical evidence, a past history of abuse [by the accused], or some other type of corroborative evidence, often the alleged perpetrator won't even be charged because we can't realistically prove the case. The child's claim that he was fondled simply isn't enough."[781]

If the presumption of innocence protects suspects who are charged with sexual abuse against children, where the child's past sexual history is not an issue, it may also protect suspects charged with rape, even when a rape shield law makes the complainant's past sexual history inadmissible. Certainly, it presents an alternative explanation to claims of systemic bias against rape victims. To be sure, the alleged victims of forcible rape are usually adults or at least teenagers, less suggestible and possessing much better memories than small children. Yet in some cases they may also seem more likely to have a motive to lie, even if in fact they are telling the truth.

In assessing the legitimate role of the burden of proof, much depends on how one phrases the question. If the question is why acquaintance rapes are not punished more often, the burden of proof is an important part of the answer. Even Estrich stresses the difficulty of corroborating the testimony of an alleged victim of acquaintance rape. In Estrich's view, this difficulty justified abandoning the corroboration requirement.[782] But it is also a reason why, even without a corroboration requirement, not all rapists can be punished--a reality that legal scholars seem loath to acknowledge.

If we focus instead on the cases that survive screening by police and prosecutors, the role of the burden of proof is presumably less great. As Estrich notes, pure swearing contests rarely get to court.[783] The weakest rape cases are either unreported, unfounded, or dismissed by prosecutors who insist on corroborative evidence, even though it is not legally required, in order to improve the chances of obtaining a conviction.[784]

With powerful reasons to fear the investigation and the trial, and many opportunities to change their minds, few women are likely to persist in a false accusation of rape all the way through the criminal process. In this important respect, swearing contests in rape cases are not precisely analogous to, say, a dispute between a policeman and a prostitute or between two men with conflicting stories about who started a fight. The lying woman has more to fear (and perhaps less to gain?), especially in the kind of case where her character is most vulnerable to attack. For these reasons, the prosecutor's burden of proof, though significant at both stages, is more cogent as an explanation of case attrition prior to trial than as a justification of acquittals at trial.

How often is the corroborative evidence introduced in a simple rape trial strong enough to remove all truly reasonable doubts about the defendant's guilt? The best way to obtain an informed, disinterested answer to this question is by comparing jury verdicts with the conclusions of the trial judges who heard the same evidence. Kalven and Zeisel's study of judge-jury disagreement found that juries were more lenient than trial judges in almost 50% of nonaggravated acquaintance rape cases.[785] Of course, trial judges are fallible. Perhaps the judges were mistaken and the juries were right to acquit in these cases. But we believe that that possibility is remote. Although most observers believe that judges are on average less lenient than juries,[786] the trial judges of that period were almost all men. Both common sense and social scientific research indicate that men tend to be relatively suspicious toward acquaintance rape accusations.[787] Indeed, the rape literature is peppered with anecdotes about male judges who were crudely biased in favor of acquaintance rape defendants.[788] Kalven and Zeisel's study was conducted prior to the anti-rape campaign of the women's movement, in an era when men were less afraid to blame rape victims.[789] One supposes, therefore, that the judicial bias in these cases, if any, would tend to favor the male defendant, perhaps strongly so. Indeed, the trial judges themselves favored acquittal in about half of the nonaggravated acquaintance rape cases in the study.[790] Equally important, trial judges can be expected to take account of the prosecution's burden of proof. They are not likely to disagree with a verdict merely because they intuit that the defendant is guilty in fact.

Despite some methodological shortcomings,[791] Kalven and Zeisel's study remains the best available evidence of national scope on the propriety of acquittals in acquaintance rape cases. It is supplemented, moreover, by more recent anecdotal evidence to the same effect.[792] Our conclusion is that the prosecution's heavy burden of proof has played an important role in the justice system's treatment of acquaintance rape cases, but so have public biases against certain classes of alleged rape victims. To neglect either source of case attrition is to distort reality.

D. VICTIM BEHAVIOR AS AMBIPROBATIVE EVIDENCE

By nearly all accounts, the prosecutor's proof problems are most acute when the putative acquaintance rape victim's behavior violates traditional standards of female morality or prudence. Indeed, one major study of rape cases found that, in cases where the victim's truthfulness is a primary issue, jurors are influenced more by her character than by the corroborative evidence introduced by the prosecution.[793]

To the extent that negative evidence about the alleged victim's behavior has genuine probative value on the consent issue, the leniency of acquaintance rape juries may be justifiable, especially in the absence of strong evidence corroborating the woman's account. If, however, the evidence used to discredit rape complainants lacks genuine probative value, the jurors' leniency in such cases is much harder to defend, given that contributory negligence and "provocation" are not, as such, defenses to rape.

One might suppose that rape shield laws should put to rest controversy about such matters, but the terms of those laws vary enormously and judges sometimes ignore them and admit evidence that they consider relevant.[794] Besides, shield laws are limited to sexual history evidence and do not cover other matters such as hitchhiking, truancy, and drug or alcohol use.[795] As well, a great deal of evidence about the putative victim's behavior is admissible in order to provide context by explaining the circumstances leading up to the rape.[796]

We believe that the kinds of evidence that are typically used to discredit the rape complainant's character are generally devoid of demonstrable evidentiary value on the issue of whether she consented to sex with the defendant. We reach this conclusion by a very different route than most commentators, and we disagree with a number of arguments that are commonly invoked to demonstrate that the putative victim's violation of sex-role norms has no bearing on whether she consented. Although our theory has obvious implications for evidence law, we will not try to deal with all of the evidentiary issues raised by the theory. Our purpose is to evaluate, in an exploratory fashion, a possible explanation of jurors' leniency in victim "misconduct" cases, not to rewrite the rape shield laws.

Responding to popular prejudices against "bad victims," rape scholars have often adopted a dismissive or argumentative stance about the subject. For example, many scholars have endorsed an untenable, generic objection to such evidence. The objection, voiced repeatedly in rape literature, is that "victim blaming" shifts our attention away from the man's conduct, leading us to focus instead on the woman's conduct. Not only is this diversion helpful to the defense, it is based, say many authors, on a popular myth about rape. To describe and rebut this so-called myth, scholars often argue as follows:[797]

1. Evidence about the rape complainant's character, and her conduct immediately prior to the rape, is usually calculated to show that she "led him on" or that she was "asking for it."

2. The implicit major premise of such evidence is that rape is the result of an uncontrollable explosion of sexual desire, precipitated by the woman's provocative conduct.

3. In fact, however, rape is a crime of violence, not of sex.[798] Men rape out of a desire for power over women, not because they are sexually needy.

4. Evidence about the woman's supposedly provocative conduct is therefore irrelevant in a rape trial.

The idea that rape is a crime of violence, not of sex, has become a shibboleth in rape literature.[799] On one level, the shibboleth is obviously true: by definition, forcible rape is a crime of force or threatened force and thus of violence. Although the most interesting cases today often involve minimal force,[800] and many scholars now seek an expanded definition of rape,[801] it remains true that adult rape, as defined by most people (and courts) is almost always a crime of violence.[802] But this truism sheds no light on the rapist's motivation; just as robbers usually employ violence, not for its own sake but to obtain money, perhaps most rapists employ violence simply in order to obtain sex.[803]

Authors sometimes describe rape as a crime of violence in order to stress that acquaintance rape is a serious matter, not a mere sexual peccadillo.[804] Again, this truism sheds no light on the more problematic argument that the alleged rape victim's sexual history, and her conduct prior to the sexual encounter, are irrelevant to whether she was raped. In this latter context, the idea that rape is a crime of violence refers to the rapist's motivation. If he rapes because he is sexstarved, then some will regard him as a sympathetic character led on by a provocative woman. If, on the other hand, he rapes because he is violent, or wants to control and dominate women, then the woman's behavior looks irrelevant, the rape no longer even half-resembles innocent lovemaking, and people will feel less empathy for the man.

Despite these rhetorical advantages, speculation about the motivations of rapists[805] sheds no light on the probative value of evidence about the woman's conduct. First, the crime-of-violence theory does not even purport to address the defendant's nominal defense. The defendant in an acquaintance rape case contends that the complainant consented, not that she provoked a rape. At least formally, the issue is whether a rape occurred, not why it occurred. In strictly legal terms, what needs to be rebutted is the proposition that her behavior prior to the rape suggests that she consented, not the proposition that her conduct caused and justified his rape. True, the jury may use the evidence for the latter, improper purpose, but that danger does not necessarily justify depriving the defendant of the opportunity to offer the evidence for a legitimate purpose.[806]

In any event, to appraise the likely prejudicial effect of evidence it is necessary, first, to consider the extent to which it has genuine probative value.[807] One cannot avoid this task simply by characterizing rape as a crime of violence. Suppose for the sake of argument that rape is indeed basically a "crime of violence, not of sex."[808] It does not follow that the victim's sexual allure cannot have been one of the causes of the crime. Research has shown that even homicide, the most violent of crimes, is often provoked, though of course not necessarily excused, by the victim's behavior, sometimes even by the victim's sexual behavior.[809]

The "crime of violence" maxim sets up a false dichotomy between sexual and other motivations. Rape is not simply "violence"; it is sexual violence. Even if the rapist's underlying motivation is to control, humiliate, dominate or hurt the victim, we need to account for the fact that he does so by a sexual assault rather than an ordinary assault. Not only that, unless he is sexually aroused, he cannot consummate the assault. Castration, whatever its defects from a humanitarian perspective, prevents recidivism.[810] For all these reasons, the acquaintance rapist's motivation, even in the relatively rare cases where it plainly goes beyond mere sexual lust, must be at least partly sexual.[811]

We need not decide how often the rapist's motivation is exclusively or primarily rather than only secondarily sexual. For a rapist who desires to control, hurt, or degrade a woman in a sexual manner may be aroused by the same sorts of sexual stimuli that arouse men who have no desire to harm women. Like other men, most (though not all) rapists prefer women who are young. Even if the rapist wants to hurt or control women, he may prefer to do so with "sexy women," rather than all women, or maybe he intuits that the easiest marks for his violent impulses are imprudent or sexually liberal women.

One must also consider the possibility, rejected by many feminists, that different types of rapists have different motivations. Vague umbrella concepts like "control over women" obscure important differences among rapists. Although stranger rapists sometimes inflict gratuitous, sadistic injuries,[812] such terror for the sake of terror is much rarer in acquaintance rapes,[813] One need not accept the near equation between sex and rape that has been advanced by a few authors,[814] to recognize that the motivations of a drunken fraternity boy who rapes his companion may be at least approximately similar to the motivations of a drunken fraternity boy who seduces his companion, especially in cases that are close to the rape-seduction borderline.

Check and Malamuth found that male subjects were sexually aroused to the same extent by written descriptions of consensual sex and acquaintance rape, while a description of a stranger rape elicited significantly less arousal. In addition they found that male subjects' self-reported likelihood of committing acquaintance rape correlated with their degree of sexual arousal from reading the description of such a rape, a result that was not found in the stranger rape condition.[815] These findings suggest that the motivation for acquaintance rapes is closer to ordinary sexual desire than is the motivation for stranger rapes.

According to some scholars, rapists often misinterpret their vicrims' verbal resistance as mere coyness; the rapist thinks that "'no' doesn't mean no."[816] If so, the motivation of these rapists seems to be sexual. For if their motivation were solely to hurt or dominate the woman, presumably the rapists would be more aroused by a "no" that they perceived as sincere than by one that they interpreted as mere coyness.

Two other standard responses to evidence of the woman's violation of sex-role norms are also inadequate. Some scholars point out that the mere fact that a woman engages in careless or provocative behavior does not give a man the right to rape her.[817] This is of course true, but the question we are considering is not whether he had a right to rape her but whether her behavior is some evidence that he did not rape her.

Some scholars point out that the woman's consent to sex on previous occasions does not "prove" consent on the occasion in question.[818] That argument confuses relevance with conclusiveness. Other scholars contend that the probative value of, say, evidence of the woman's promiscuity,[819] is only slight.[820] In most cases, this is a fair characterization, but it remains possible that, if the prosecution lacks strong corroborative evidence, "slight" evidence of consent sufrices to raise a reasonable doubt. It is in the weak cases, where corroborative evidence of the defendant's guilt is lacking, that juries are most strongly influenced by evidence about the putative victim's character.[821]

Inescapably, then, scholars must consider whether the putative victim's violation of sex-role norms casts any light on the consent issue. We now offer a new approach to that question as we examine the most common types of "victim blaming."

1. Contributory Negligence

Kalven and Zeisel argue that the explanation forjudge-jury disagreement in simple rape cases is that jurors do not limit themselves to the legal issues of penetration and consent. Rather, juries in effect nullify the law of rape by importing the tort concept of contributory negligence or assumption of risk into the criminal case, acquitting the defendant when they believe that the victim behaved carelessly.[822]

Kalven and Zeisel discovered a similar phenomenon in homicide prosecutions growing out of drag racing accidents in automobiles, or Russian roulette. When one participant is killed, and his confederate is prosecuted for manslaughter, juries tend to acquit even if the defendant's legal guilt has been plainly established.[823] The juries believe, evidently, that the victim assumed the risk. This belief, although not sanctioned by the law of homicide, seems to be factually accurate. In drag racing, as in Russian roulette, the one who dies usually has "lost" in accordance with the rules of the game as understood by the parties. For independent reasons of public policy, the law of homicide seeks to override theparties' rules. Rightly or wrongly, the juries disagree, and nullify the law by enforcing the parties' deal.

Is a similar analysis applicable to rape? Even if one believes that a rape complainant "led him on," that sort of negligence is not closely analogous to drag racing. The victim's negligence in an acquaintance rape case is more analogous to a homeowner who negligently fails to lock her front door or a banker who strolls through Central Park at midnight carrying a wallet full of $100 bills. No one suggests that a burglar who takes advantage of an unlocked door, or a mugger who attacks an imprudent banker, ought to be acquitted, and we have found no evidence, anecdotal or otherwise, that juries often acquit thieves whose victims were negligent.[824]

On the other hand, the analogy to theft is often carried too far by rape scholars intent on demonstrating bias against rape victims. Here, for example, is an imaginary cross-examination, written by a scholar trying to show the unfairness of focusing on the woman's behavior prior to the rape:

"Mr. Smith, you were held up at gun point on the corner of First and Main?"

"Yes."

"Did you struggle with the robber?"

"No."

"Why not?"

"He was armed..."

"Did you scream? Cry out?"

"No. I was afraid."

"I see. Have you ever been held up before?"

"No."

"Have you ever given money away?"

"Yes, Of course."

"And you did so willingly?"

"What are you getting at?"

"Well, let's put it like this, Mr. Smith. You've given money away in the past. In fact, you have quite a reputation for philanthropy. How can we be sure that you weren't contriving to have your money taken from you by force?"

"Listen, if I wanted..."

"Never mind. What time did this holdup take place, Mr. Smith?"

"About 11:00 p.m."

"You were out on the street at 11:00 p.m.? Doing what?"

"Just walking."

"Just walking? You know that it's dangerous being out on the street that late at night. Weren't you aware that you could have been held up?"

"I hadn't thought about it."

"What were you wearing at the time, Mr. Smith?"

"Let's see ... a suit. Yes, a suit." "An expensive suit?"

"Well--yes. I'm a successful lawyer, you know."

"In other words, Mr, Smith, you were walking around the streets late at night in a suit that practically advertised the fact that you might be a good target for some easy money, isn't that so? I mean, if we didn't know better, Mr. Smith, we might even think that you were asking for this to happen, mightn't we?"[825]

The suggested analogy between rape and robbery is less apt than the author implies. In a consent-defense rape case the defendant, at least formally, claims that consensual sex occurred, not that he committed a justifiable rape. If a robbery defendant were to offer a similar consent defense, for instance by claiming that he was a panhandler and the "victim" freely gave him $300, then it would surely be appropriate to inquire into the putative victim's habits, particularly any propensity to make large gifts of money to panhandlers or other strangers, perhaps even seeking them out for that purpose late at night. In such a case, the hypothetical cross-examination would not seem ludicrously inappropriate. Of course, in real robbery cases no such defense is offered, because the circumstances almost always make it extremely implausible; hardly anyone gives large sums of money to strangers on the street. In this respect, robbery is analogous to stranger rape. But in stranger rape cases, as in robbery cases, people rarely blame the victim.

Does a woman's contributory negligence sometimes raise a reasonable doubt about whether she consented to sex? Consider again the imprudent banker who walks through Central Park late at night. One difference between the banker and a woman who is careless about the danger of being raped is that, although some psychologists might surmise that the banker has a subconscious desire to be mugged, the risky stroll is not an ambiguous signal. No one is going to rob the banker under the illusion that it was a consensual transaction and the banker's protestations to the contrary were merely coyness.

In an acquaintance rape case, on the other hand, the woman's adventuresome conduct may indeed, in some circumstances, create a more or less reasonable inference that she will consent to sex. But we are considering cases in which the woman's signals are at least somewhat ambiguous. For example, a couple meet in a bar and she accepts his invitation to have one last drink in his apartment at 1:00 a.m. At this juncture, let us assume for the sake of argument that it would be reasonable for the man to predict that the woman will later consent to sex. If his prediction turns out to be mistaken, this is not a "reasonable mistake" in the legal sense. It is one thing to predict a favorable response to a sexual request and quite another to dispense with the legal necessity of obtaining that response before having intercourse. In order to have a legally valid defense, the defendant must be reasonably mistaken, not about whether the woman will consent, but about whether she is consenting at the time of the rape.

In some cases, the woman does indeed deserve censure for imprudence or for having created sexual expectations, not altogether unreasonable, that she does not intend to fulfill. Blaming the victim is not always wholly unjust. But, as with other crimes, victim blaming should be divorced from perpetrator-absolving: to say that a rape victim's conduct was foolish and irresponsible is not to say that the perpetrator should be acquitted. Many criminals prey chiefly on careless victims.

In a dating situation, the man can resolve any uncertainty about his companion's desires by asking, just as he would if he were a buyer of jewelry who felt sure, perhaps on reasonable grounds, that the owner was ready to sell it. If instead he rapes her, it makes no sense to say that she tacitly agreed that rape would be an acceptable outcome. It may be foolish for a woman to hitchhike or to drink several beers and then visit the apartment of a man she met only an hour earlier in a bar, but without other indicia of consent there is no good reason to treat this negligence as tantamount, either factually or legally, to consent to sexual intercourse. Granted, the woman's conduct may have kindled a flame of desire, and in some cases she may even have been wrong to do so, but in our sex-saturated culture a man who does not control his sexual desires when "unfairly" aroused is both highly abnormal and highly dangerous.

Of course, the defendant in an acquaintance rape case does not concede that he forced the woman to have sex, or in other words that he treated her perhaps imprudent behavior as consent to sex. In most cases, he testifies that she explicitly consented or at least that she did not resist his advances. The real question, therefore, is what weight the jury should give to her imprudence in determining who is telling the truth. As such, contributory negligence is not a defense, but does it have some tendency to prove that the defendant's story is true? If so, does this circumstantial evidence of consent help to explain the high acquittal rate in such cases?

Evidence is considered relevant for purposes of the rules of evidence only if it is both probative and material. It is material if it relates to an issue in the case, such as whether the rape complainant consented. To be probative, the evidence must tend "to establish the proposition for which it is offered or--to be precise--if the proposition is more likely to be true given the evidence than it would be without the evidence."[826]

Is evidence of the complainant's sexual habits usually relevant to whether she consented to sex with the accused? At common law, the woman's general reputation for chastity and specific instances of prior sexual activity were admissible evidence that she had consented to sex with the defendant.[827] Modern scholarly and judicial opinions about this issue fall into three groups, corresponding to the three standard characterzations of evidence whose relevance is challenged. Some say that the sexual history evidence does indeed tend to show consent;[828] some say that it does not;[829 and some say that, while the evidence tends to show consent, its probative value is very slight and greatly outweighed by the danger of prejudicing the jury.[830]

All three of these positions assume, in accordance with standard evidentiary analysis, that the probative value of relevant evidence cuts in favor of one party's position and against the opposing position taken by the other party. We believe, to the contrary, that evidence about the rape complainant's sexual history usually "cuts both ways" on the issue of consent and that common experience usually affords no basis for saying which of the two opposing inferences is stronger. We will call such evidence "ambiprobative."

Consider an example from another field of criminal law. Suppose that Jones has been charged with murdering Smith. His defense is self-defense; he killed Smith while fending off a murderous attack by Smith. In this case, is evidence of a previous affair between the defendant and Smith's wife relevant? We have no doubt that a court would admit such evidence,[831] but notice that it bolsters both the prosecution's charge of murder and the defendant's plea of self-defense. The affair supplies a motive for Jones to kill Smith, but also a motive for Smith to have tried to kill Jones.

In conjunction with other evidence, ambiprobative evidence may be more helpful to the prosecution than to the defense, or vice-versa, but it is not intrinsically more probative of either side's version of the facts. For example, if a woman is charged with having murdered her husband, and she pleads self-defense, evidence that he had battered her on other occasions would tend to prove both that she acted in selfdefense and that she had a motive to murder him even if he was not attacking her at the time. Since men are, as a rule, much more violent than women, her claim of self-defense sounds more plausible than the prosecution's charge of murder. Of course, there may be other evidence that suggests murder:for example, evidence that he was asleep on a couch when she killed him. In such a case, the main gap in the prosecution's evidence might be lack of an apparent motive for murder, and the prior batterings might fill that gap. The defense might need the evidence less, because it possessed independent evidence that the husband planned to attack the wife. But unless the evidence that the husband was sleeping is conclusive, his prior attacks would also tend to corroborate her version of the killing.

Another example of ambiprobative evidence is when a personal injury plaintiff seeks to introduce evidence that the defendant had liability insurance. The plaintiff may argue that the insurance is relevant on the negligence issue because insured people have less incentive to avoid liability and therefore presumably tend to be less careful.[832] In contexts where the defendant himself would not be endangered by his own negligence, this theory is at least moderately plausible. The defendant may respond by suggesting a countervailing inference: the prudent type of person who buys insurance is less likely to be negligent.

If the only problem in such a case was the ambiprobative quality of the evidence, courts might let the jury choose between the two competing inferences.[833] The main problem, however, is not that the evidence is ambiprobative, but that the jury is likely to use it for the forbidden purpose of finding in favor of the plaintiff because the. insurance company will bear the cost. Given this third possibility, courts wisely exclude such evidence.[834]

We regard the insurance cases as analogous to many "bad victim" rape cases. As in the insurance cases, the evidence is ambiprobative, the jury lacks sufficient basis in common experience for choosing between the inferences, and the danger that the jury will misuse the evidence is great.

Most evidence treatises ignore the problem of ambiprobative evidence.[835] We suspect that there are several reasons for this. First, the phenomenon is relatively rare. Second, the opposing party, instead of objecting, may cite the evidence as probative of its own position. Third, in some cases one of the two conflicting inferences may seem stronger than the other. In such situations the net probative value cuts in favor of one party and the evidence is relevant by the ordinary definition because it makes one side's position more plausible than it would otherwise have been.[836] In other cases, including rape, the ambiprobative nature of the evidence is not obvious or well-known and so the opposing lawyer usually either does not object to the evidence, or objects on some other ground. In acquaintance rape cases, for example, objections to sexual history evidence are usually on the ground that it lacks probative force (in any direction) and is too prejudicial.[837] Finally, even if the ambiprobative quality of the evidence is apparent, we can safely assume that many lawyers and judges believe that the jury should decide which of the two conflicting inferences is stronger.

Two treatise writers do address the problem. Wigmore discusses evidence that could support either of two opposite conclusions; he indicates that many courts admit such evidence only if the inference for which it is offered is more probable than the opposing inference.[838] Jones also mentions the problem of what he calls "equivocal" or "ambignous" evidence.[839] He believes that such evidence should be admitted, allowing the jury to assess its value.[840]

The caselaw on this issue is sparse and divided.841 We have not found a rape case in which the issue was raised. For present purposes it suffices to note that in a typical rape-victim "misconduct" case the jury has no basis in common experience for choosing between the conflicting inferences. For example, consider again the negligent rape victim. If the issue before the jury were simply whether two people had consensual sex on a particular occasion, the woman's antecedent suggestive behavior would indeed be relevant, though of course not conclusive, evidence that they did. Imagine an observer who is trying to guess whether a man had consensual sex with a woman who entered his apartment at 1:00 a.m., shortly after having met him in a bar. Plainly, their behavior is not conclusive proof that the parties engaged in consensual sex. But all else being equal, the woman in this scenario is much more likely to have consented than a woman whose conduct was more circumspect. On these facts, with no allegation of rape, an observer probably would have at least a reasonable doubt about whether consensual sex occurred.

Rape scholars often deny, or at least seem to deny, this obvious truth,[842] no doubt because they suppose that such a concession undermines any subsequent accusation of rape, either by falsely implying that the woman must have consented, or by leading to the conclusion that even if she didn't exactly consent, she "got what she was asking for." These scholars seem to assume that any evidence that has a tendency to show consent must have a corresponding tendency to show that no rape occurred. That assumption, though superficially axiomatic, ignores the possibility that certain kinds of evidence are suggestive both of consent and of rape.

In most acquaintance rape cases, the jury must make a comparative evaluation of two conflicting scenarios. According to one scenario the woman consented and later falsely accused the man of rape; according to another she was raped. Anything that increases the likelihood that one scenario is true, but equally increases the likelihood that the other is true, has no bearing on which of the two scenarios is more credible. In our hypothetical case, the woman's behavior increases the plausibility of the consent scenario but also increases the plausibility of her accusation of rape. After all, most observers would regard her conduct as imprudent, precisely because it increased the danger of rape, unless she wanted to have sex.

Admittedly, consensual sex is much more common than rape.[843] When the couple in our hypothetical enter the man's apartment, an observer, knowing nothing about the parties, would probably consider the likelihood of consensual sex much greater than the likelihood of rape. But this does not mean that if the woman later reports that she was raped, her story is less credible than the man's defense of seduction. Assuming that most accusations of rape are true, an accusation alters the balance of probabilities, probably very greatly, although we cannot say by precisely how much. A denim of that accusation, on the other hand, is what we would expect of a rapist, no less than of an innocent man, and so it scarcely affects the probabilities.

The woman's contributory negligence, though not increasing the likelihood that she consented, may have played a causal role in precipitating the crime. Research has demonstrated that the victim's behavior is one of the causes of many crimes.[844] A "victim precipitated" crime is one that might not have occurred but for the victim's conduct: for example, leaving his window open while he was absent on vacation, or insulting a gang leader.[845] Female victims are not prominent in most of the literature about victim precipitation. Indeed, victim precipitated homicides, as compared to other homicides, are disproportionately ones in which the victim was male.[846]

A judgment that the crime was victim precipitated does not logically entail absolving the perpetrator. Most traditional types of voluntary manslaughter are by definition victim precipitated:[847] for example, when the defendant catches his spouse in adultery and kills her or her lover.[848] Since the crime would otherwise be murder, one can say that victim precipitation reduces the gravity of the offense, but it remains one of the most serious crimes. A thief who goes through an open window is guilty of burglary under modern statutes, notwithstanding the victim's precipitative role.[849] Likewise, in rape cases, the woman's contributory negligence is not a defense unless the man makes a reasonable mistake,[850] not as to whether she is likely to consent, but as to whether she does consent.[851] To speak of a "victim precipitated rape,' is, therefore, not oxymoronic.

Yet much of the literature about victim precipitation has strong pejorative overtones. Especially in the context of rape, the very term "victim precipitated" has an innuendo that the victim was the main culprit and that the perpetrator's actions were a normal response to her provocation. Further, the concept of victim precipitation is often applied to cases in which the victim (usually a male) was careless at best and a bully or a cheat at worst. One type of victim precipitated homicide, for example, is where the victim starts a fist fight and is then killed by the defendant.[852] Definitions of victim precipitation usually exclude involuntary and status risk factors such as age, race, gender, physical condition and wealth,[853] further contributing to the vaguely pejorative innuendo of the term.

As these examples suggest, the concept of victim precipitation involves several distinct questions. The first question is non-pejorative: Did any aspect of the victim's character or conduct increase the risk that she would be raped? Another possible question is whether she should be criticized for this. We would not criticize her, for example, for being young, though that increased the risk of being raped, nor for dating men, though that too increased the risk. But a relative or friend might justly criticize her for, say, going camping with a man she had just met, and with whom she had no desire for sex. Finally, there is the legal question: Did her conduct partly or wholly excuse the defendant's behavior?

In rape cases, commentators of all ideologies usually confiate all these questions, treating the victim as responsible for the crime in every sense or in none. The oft-quoted list of "Rape Myths" exemplifies this problem. Stevi Jackson, for example, asserts that the popular myth that males have an "uncontrollable" sexual response that, once set in motion, leaves the man "totally at the mercy of his desires" is merely a "technique of neutralization available to the rapist: denial of responsibility.[854] If a man attributes this to himself, perceives himself as a helpless slave to his desire, then he will be less inclined to curb himself in the face of a woman's refusal and more inclined to resort to force to attain his ends."[855]

Jackson's purpose, as she makes plain, is to insist that the rapist be held accountable for his crime. We applaud that goal, but why should one suppose that any idea that can be invoked by a rapist in an effort to escape accountability must be false? That the truth may be misused does not make it a myth. The notion of "uncontrollable" sexual desires may indeed be false--as with most issues of free will no one really knows--but the notion that the conduct of a woman sometimes increases the risk that she will be raped is surely true. Our point is simply that certain types of behavior by potential victims increase the risk of victimization, for rape as for every crime. Whether a potential victim should therefore change her behavior is a different question, and whether to exonerate a man who rapes her is yet another question.

Another so-called "myth" is that, "[w] hen women go around braless or wearing short skirts and tight tops, they are just asking for trouble."[856] This is called a myth, presumably, because some men falsely believe that provocatively-dressed women invariably want sex or even want to be raped. But those are not the only reasonable interpretations of "asking for trouble." If the phrase means "creating a potentially misleading impression of a desire for casual consensual sex, which may lead to unwanted and perhaps even violent sexual advances by men," the proposition doesn't sound mythical at all. Common male attitudes, as many scholars have stressed, create precisely those dangers.

Perhaps the best example of the myth-that-may-not-be-a-myth genre is a proposition used by a scholar testing the attitudes of police and other citizens toward rape. The proposition was that, "A woman should be responsible for preventing her own rape."[857] Although the meaning of "responsible" in this proposition is obviously ambiguous, an affirmative answer was interpreted as evidence of a "pro-rape" attitude.[858] This interpretation is correct only if one reads some word or phrase like "solely" or "rather than the rapist" into the proposition. Otherwise, the "myth" is just as true as the proposition that, "a homeowner should be responsible for preventing burglaries of her house." As with a homeowner, some precautions are obviously called for, some are obviously excessive, and some are debatable.

Feminists are not the only ones who confound the various issues raised by victim precipitation. Even greater confusion appears in Menachem Amir's pathfinding study of Patterns in Forcible Rape.[859] Amir, the leading student of victim precipitated rape, defines the term twice. First, he says it means "that in a particular situation the behavior of the victim is interpreted by the offender either as a direct invitation for sexual relations or as a sign that she will be available for sexual contact if he will persist in demanding it."[860] He excludes "situations where no interaction was established between the offender and the victim, and when the offense was a sudden event which befell the victim."[861] Thus, stranger rapes are excluded even if the woman was negligent, and acquaintance rapes are included, apparently even if she was not negligent, whenever the man interprets her behavior as suggestive: "Whether it is really so is not as important as the offender's interpretation of her actions within the then current situation. Because even if erroneous, it leads to action."[862] So if a woman invites a man with whom she has a platonic relationship to dinner in her apartment, and he interprets the invitation as a sexual overture, any subsequent rape was "victim precipitated"!

In another passage, Amir tries again to define victim precipitated rapes:

The term "victim precipitation" describes those rape situations in which the victim actually, or so it was deemed, agreed to sexual relations but retracted before the actual act or did not react strongly enough [in whose opinion?] when the suggestion was made by the offender(s). The term applies also to cases in risky situations marred with sexuality, especially when she uses what could be interpreted as indecency in language and gestures, or constitutes what could be taken as an invitation to sexual relations.[863]

Amir's second definition, like his first, appears to exclude most of all stranger rapes even if the victim was negligent, while covering virtually any acquaintance rape preceded by drinking, a visit to the man's apartment, or other even very mildly suggestive behavior depending on how the vague language about "risky situations marred with sexuality" is interpreted. Noting that the law of rape "does not recognize precipitation, provocation, and seduction,"[864] Amir suggests that, "[i]f any causal connection exists between precipitation and rape, we must assess it with the aim of educating the lawto recognize it too."[865] The implication seems to be that victim precipitated rapes should not be punished, or perhaps should be punished less severely than other rapes. Since Amir seems to define victim precipitation broadly enough to cover cases in which the woman was not even negligent, his suggestion that victim precipitation should be a defense is patently unjustifiable.

The National Commission on the Causes and Prevention of Violence published crime-specific definitions of victim precipitation. For forcible rape, the definition was, "[w]hen the victim agreed to sexual relations but retracted before the actual act or when she clearly invited sexual relations through langnage, gestures, etc."[866] Whereas Amir, with his vague and apparently broad definition of victim precipitation, concluded that 19% of all the rape cases he examined had been victim precipitated,[867] a Commission task force concluded that only 4.4% of rapes are victim precipitated, a lower figure than for homicide (22%), nonsexual assault (14.4%), armed robbery (10.7%) or unarmed robbery (6.1%).[868]

Susan Brownmiller discusses three situations in which some might say that a rape was victim precipitated. First, a homemaker accepts a strange man's offer to rake her lawn for a fee, and when he later asks for a glass of water, she lets him into her living room. Second, a female hitchhiker. If a rape ensues, Brownmiller "would consider the housewife and the hitchhiker insufficiently wary, but in no way would I consider their actions provocative or even mildly precipitant."[869] Finally, Brownmiller argues that a woman should not be deemed guilty of precipitant behavior if she "engages in sex play but stops short of intercourse."[870] The reader is left to wonder whether Brownmiller would ever describe a rape as "victim precipitated." Evidently, she regards use of the term as tantamount to excusing the rape.

Recognizing the pejorative innuendo of the concept, many scholars avoid--indeed, frown upon--discussions of the rape victim's role in precipitating the crime.[871] Although this attitude is eminently understandable, some rapes do involve negligent victims.[872] As with a homeowner who fails to lock a door, this sense of the term does not logically entail absolving the perpetrator: unless the man made a reasonable mistake about consent, the woman's negligence has no bearing on his legal guilt.[873] Nevertheless, it seems best to eschew negligence and other normative terminology, particularly in a field like rape, where evolving conceptions of gender equality complicate the task of determining what precautions a "reasonable woman" would take. To emphasize this point, we will avoid loaded terms like "victim precipitated." Instead, we will refer to "risk factors" for rape or the woman's "vulnerability to rape." These terms refer to any attribute of the victim or her conduct, whether voluntary and blameworthy or not, that increased the likelihood that she would be raped: her age, her income, her drinking, her sexual habits, her companions, her hitchhiking, and so on. Thus, our definition is broader but less pejorative than the definitions others have employed.

2. Alcohol and Drugs

"Candy is dandy, but liquor is quicker," wrote Ogden Nash.[874] Folk wisdom holds that alcohol reduces sexual inhibitions and so increases the likelihood of consent. Of course, sex with an unconscious victim is rape, but if the woman was conscious, but intoxicated, numerous authorities hold that she may have been capable of validly consenting to sex.[875] Even if she does not in fact consent, a woman who has been drinking may respond in ways that give the impression of consent and so mislead her parmer into an arguably reasonable mistake. Perhaps for these reasons, rape juries are less likely to convict if the evidence reveals alcohol (or drug) use by the woman.[876] Indeed, Kalven and Zeisel found that drinking by the woman was the factor that most often induced jurors to acquit in rape cases.[877] As one would expect, some prosecutors are leery of such cases.[878]

A few scholars have tried to brush this problem aside by observing that, "the fact that some victims of such assaults had taken drugs or had one drink too many ... does not mitigate the seriousness of the act [of rape]."[879] This perfectly reasonable assertion assumes that a rape occurred and the only question is whether to excuse ("mitigate") it. At some, perhaps subconscious level, that may indeed be the question. However, it is not the legal issue: rather, in most acquaintance rape cases, the defendant claims that no rape occurred. What if the defendant's lawyer argues that the woman's drinking, prior to the sexual encounter, tends to support the defendant's testimony that she consented? Let us assume that she was not so drunk as to be legally incapable of consent. Her drinking is not dispositive evidence of consent, but is it not relevant? It would be relevant evidence of subsequent negligence in, say, driving a car, because people who drink are more likely, all else being equal, to drive carelessly.[880] Aren't they also more likely to consent to sexual propositions? What is the difference?

One important difference is that in the case of rape, we deal not only with an obvious causal connection, but also with a contested social construct: valid consent. At least in the context of heavy drinking, consent is less a condition of the drinker's mind than a reaction in the observer's mind. The point at which pre-coital libations invalidate what would otherwise be consensual conduct cannot be resolved by any merely empirical observation.

We suggest another perspective on the problem. In the automotive context, the probative force of evidence of drinking cuts in only one direction, while in the context of rape it is ambiprobative. Drinking prior to driving is not conclusive proof that one committed any subsequent negligent act such as failure to signal a turn, but it does increase the likelihood of negligence. Drinking may also increase the likelihood of genuine or at least apparent consent to a sexual overture. But in the sexual context evidence of prior drinking is ambiprobative. In addition to increasing the likelihood of consent, it also increases the likelihood of rape.[881] In what appears to have been a conservative estimate, Amir found that in Philadelphia at least 31% of those who reported a forcible rape had been drinking prior to the rape.[882] He listed drinking as one of several "situational" factors that make rape more likely to occur.[883]

Why should this be so? In the first place, the woman's drinking may be associated with rape because it is associated with the man's drinking, which is perhaps the real catalyst. Just as one would expect, studies show that the man's drinking greatly increases the likelihood that he will try to rape his companion.[884] Alcohol reduces men's inhibitions against violence, including sexual violence.[885] The woman's drinking may also play a causal role: alcohol may make her less able to ward off the rapist.[886] In addition, some people regard forcible sex as more acceptable when the victim is drunk.[887] Amir speculated that by drinking in public places with strangers, a woman "defines herself as 'prey' to her drinking companions."[888] There must also be some women who drink precisely in order to reduce their own inhibitions. (Whether this is as common as some men suppose is a different question.) In such cases, any subsequent consent is genuine by the 'usual standards of voluntary behavior.

It seems that a drinking woman is more likely to apparently or actually consent, more likely to be raped, and more likely to enter a legal twilight zone where the proper characterization of what occurred is debatable. Since we do not know which of these alternative inferences is generally strongest, and since the subject is full of tricky normative and factual questions, it seems best not to dogmatize.[889] Granted, by conventional standards, there are probably a lot more instances of inebriated consensual sex[890] than of inebriated forcible rape, but this does not mean that accusations of the latter are likely in fact to be the former.

Kalven and Zeisel found that victims' use of alcohol negatively affected jurors' perceptions of them in cases involving various offenses, with both female and male victims.[891] One might surmise, therefore, that jurors' skepticism rests, not so much on sex-role stereotypes, as on doubts about the accuracy of the recollections of one who had been drinking prior to the encounter.[892] On the other hand, since the man usually has also been drinking, one would suppose that such skepticism would undermine his credibility as much as hers. Yet, social-scientific research indicates that when both parties drink many observers are more inclined to blame the woman for the rape than when neither party engages in drinking.[893] In a trial context, this phenomenon might be due to the burden of proof,[894] but in the context of atfitudinal research by opinion polls based on rape scenarios, it seems to reflect notions about the appropriate roles of the victim in preventing certain crimes--notions that are not, however, limited to rape cases nor to cases in which the victim is female.

3. Promiscuity and Prostitution

Promiscuity is perhaps the most effective charge that a rape defendant can level against his accuser.[895] Prior to the advent of rape shield laws, courts generally regarded the woman's "character for chastity," her propensity to engage in nonmarital intercourse,[896] as relevant to whether she had consented.[897] Despite shield laws, juries often become aware of the alleged victim's sexual habits.[898] Does such evidence serve any legitimate purpose? Wigmore thought so: "The character of the woman as to chastity is of considerable probafive value in judging the likelihood of consent."[899] Even today, it is not hard to find men, including legal scholars, who privately agree with Wigmore. But, modern rape scholars usually argue that sexual history evidence is irrelevant and prejudicial. They maintain that the complainant's sexual history has little or no tendency to show consent, because nonmarital sex has become common,[900] and consent to sex on one occasion, with one man, is not evidence of consent on another occasion or with another man.[901] Estrich, for example, would exclude evidence of the woman's past sexual encounters with other men because "most women, like most men, have had past sexual encounter s" and so "unless the pattern is indeed peculiarly close, they don't prove a thing."[902]

Such arguments are not an adequate justification for describing evidence of promiscuity as irrelevant. To be sure, research confirms the popular impression that a "sexual revolution" beginning in the 1960s and 1970s has led to a much higher incidence of nonmarital sex than in previous eras.[903] But to say that sexual mores have evolved is only to say that the baseline of normalcy should be changed,[904] not that evidence of a major departure from the new baseline is irrelevant. If, by the standards of our time,[905] the alleged victim was sexually unrestrained, why is this not relevant (albeit inconclusive) evidence that she consented?

An issue that some regard as analogous arises when the prosecution seeks to introduce evidence that an accused rapist has committed other sexual assaults. Urged on by feminists, Congress has amended the Federal Rules of Evidence to enlarge the circumstances in which evidence of the accused's alleged prior sexual assaults is admissible.[906] Yet some of the same arguments that are commonly used to justify excluding evidence of the woman's sexual history could also be invoked by a defendant to just'fly exclusion of evidence about his prior crimes. He could say, for example, that "just because a man raped one woman, on one occasion, doesn't mean that he did it on another occasion, with another woman."

The beginning of wisdom on this subject is to recognize that, in Paul Meehl's words, "behavior science research ... shows that, by and large, the best way to predict anybody's behavior is his behavior in the past."[907] Suppose that the defendant in a nonsexual assault or murder case claims that he acted in self-defense. He is permitted to introduce character evidence to show that his victim had a reputation for attacking other men.[908] It would be ridiculous to exclude such evidence on the ground that '[just because he attacked people in the past doesn't prove he did it this time." Again, that argument confuses relevance with conclusiveness. Admittedly, the probative value of the past assaults is probably much greater than the probative value of a putarive rape victim's promiscuity, but we are discussing the admissibility of the evidence, not its relative strength.

It is true, of course, that consent on one occasion does not "prove" consent at another time or with a different man, even if the woman often engages in casual "one-night stands." Like all character traits, promiscuity is a relative term. Just as a "liar" usually tells the truth, and a "violent" person usually behaves peaceably, a "promiscuous" woman usually does not desire sex with the nearest man. Nor, for that matter, does a rapist attack most of the women he encounters. Even lovers spend most of their leisure time doing something else, and from time to time they decline each other's advances. None of this shows, however, that sexual history evidence is irrelevant, as opposed to inconclusive.

Many people think the evidence is relevant. As an English legislator put it, "a woman with a past.., is less likely to be the victim of rape than a maiden aunt, an unpromiscuous virgin, or a respectable married woman."[909] Evidently, jurors agree. According to a study of English rape trials,

There is a staggering difference in the conviction rates of those defendants whose victims were virgins, or of whose sexual past the jury knew nothing, and those accused of raping women known to have had prior experience. This highlights the perceived importance of chastity in the (genuine) victim of rape: virginity all but guarantees a conviction. In this sample [of 81 randomly-selected trials], only one out of a total of seventeen defendants was acquitted of the rape of a virgin or a woman whose sexual past was not referred to during the trial. That is a conviction rate of 94 percent.

At the other end of the spectrum, the conviction rate among those accused of raping a woman whose sexual reputation was markedly discredited during the trial was 48 percent. This includes women who had in the past suffered from sexually transmitted diseases, those who had a reputation in the local community for being sexually available, those who had been involved in sexual intercourse with a number of persons within a short period of time and those who were alleged to be prostitutes. It is also interesting.., that nearly all those defendants who were acquitted on the judge's direction . . . were accused of raping women alleged to have such sexual histories.[910]

Abundant anecdotal evidence suggests that American juries, like their English counterparts, are powerfully influenced by the complainant's sexual habits, and that, despite rape shield laws, defense counsel often manage to get such evidence before the jury.[911]

If one were simply trying to calculate the likelihood that a given woman would consent to a sexual proposition, her sexual selectivity (or lack thereof) would be useful, albeit far from conclusive, evidence. It is no answer to say that "even a prostitute is entitled to decline sex;" the issue is not her undoubted right to do so, but whether she exercised that right on a particular occasion. The case for admitting evidence of the alleged victim's sexual history is, therefore, stronger than many scholars have acknowledged, provided that the evidence reveals, not merely an ordinary amount of sexual experience, but an unusually high propensity, judged by contemporary standards, to engage in casual sex.

We see two major justifications for excluding the evidence. The first is the well-known argument that the probative value of the evidence is outweighed by the danger that it will prejudice the jury. As there is copious evidence of jury prejudice against sexually adventurous women, we believe that this argument is usually valid, subject to exceptions that need not be discussed here.

In judging the relevance of sexual history evidence, the ultimate issue is not the woman's propensity to consent, viewed in isolation, but rather the relative plausibility of her account of rape and his account of consent. Evidence of the woman's promiscuity is ambiprobative. It enhances the plausibility of the defendant's tale of seduction, but it also, much less obviously, enhances the plausibility of her tale of rape. That is the critical difference between evidence that the putatire victim of a nonsexual assault is a bully who beats people up, and evidence that the putative victim of a rape is promiscuous. In most cases, the violent tendencies of a bully-"victim" tend to corroborate the defendant's claim of self-defense, and not the government's claim that the defendant himself was the aggressor.[912] Evidence that the putative victim of a rape is promiscuous, by contrast, tends to corroborate both parties' versions of the encounter--and therefore neither's.

First, the woman's failure to accuse any of her prior lovers of rape tends to show, albeit weakly, that she does not make false rape accusations: "If the victim has had twenty instances of prior sexual conduct with rock stars ... without claiming rape, in the absence of other evidence of motivation the most reasonable inference is that she claimed rape this time because she was raped."[913]

More important, sexual permissiveness is associated with rape victimization. The best predictor of whether one will be a rape victim, according to research by Koss and Dinero, is a composite model based on a history of child sexual abuse, liberal sexual attitudes, higher-thanaverage alcohol use, and above-average sexual activity.[914] In other words, the women who are most likely to consent are also the women who are most likely to be raped when they do not consent.

Similarly, Essock-Vitale and McGuire found a relationship between rape victimization and the victims' voluntary sexual practices. Women who reported being raped or molested had almost twice as many consensual sexual partners (14.2) as those who had not been raped or molested (7.2).[915]

Philip Belcastro investigated the sexual behavior patterns of 437 female college students916 in order to determine whether the 40 who had been raped had different sexual habits from the 397 who had not. Belcastro found that the raped and never-raped groups were similar on a number of demographic variables,[917] but the raped students reported significantly more (1) heterosexual partners; (2) heterosexual partners whom they had exclusively dated for six months or longer; (3) heterosexual cohabitation partners; (4) self-masturbation; (5) use of contraceptive techniques; and (6) coitus on the first date with a man.[918] The raped group was also significantly younger at their first pregnancy than the nonvictim group.[919] The most interesting finding was that the rape victims had more homosexual experiences than the nonvictims. Unfortunately Belcastro was unable to determine whether these experiences were pre-rape or post-rape.[920]

Belcastro's findings do not necessarily mean that, on any given date, a woman with more sexual experience is more likely to be raped. Conceivably, the more sexually active women were more vulnerable to rape at least partly because they had more dates. But whatever the causal chain, Belcastro's research suggests that there is no reason to be extraordinarily suspicious of a rape report by a woman who has had an unusually large number of sexual partners.

Amir collected data for all cases of forcible rape listed by the Philadelphia police in 1958 and 1960. He found that, of 640 women who reported rapes to the police, 124 (19%) had an arrest record, and 37.9% of these arrests were for sexual misconduct including "promiscuity" and prostitution.[921] If juvenile misconduct is counted, 56.4% of the rape complainants with a criminal record had engaged in sexual misconduct.[922] About 20% of all reported victims had a "bad" reputation.[923] Perhaps some of these "bad" women lied about being raped, but Amir surmised that rapists had intuited that the criminal justice system would not punish them for raping such women.[924] Confirming Amir's explanation, a study of gang rape found that the rapists usually select victims known to have a "bad reputation."[925]

Several studies of women who have been raped more than once have found that multiple victimization is associated with multiple consensual sexual partnerships.[926]

A recent, comprehensive, national survey of sexual practices found that nearly 22% of American women have been "forced" to have sex with a man.[927] As a group, the women who had been forced to engage in sex "were more likely to have had more than ten sex partners in their lifetime than women who had never been forced."[928] They were also more adventuresome in their sexual practices.[929]

The demographic groups that are disproportionately likely to engage in consensual sex are also disproportionately likely to be raped. For example, consensual sex is statistically more common among young women than among middle-aged women, but so is rape.[930] Similarly, poor, single, inner-city residents may be on average more sexually active than the general female population.[931] Perhaps, at least partly for this reason, police may be more skeptical of rape accusations by such women.[932] But according to some studies poor, single women are also statistically more likely to be raped than other women,[933] so one could just as easily regard their rape accusations as more credible.[934]

Prostitutes are the most extreme example of the ambiprobative quality of evidence that the woman was promiscuous. Judges often admit evidence that the rape complainant was a prostitute, on the theory that it is relevant on the issue of consent.[935] Juries are sometimes reluctant to convict men charged with raping prostitutes.[936] Is this defensible?

In analyzing the cases involving prostitutes, one should distinguish among several ways in which a rape defendant or an official may seek to discredit the prostitute's testimony. Some have said that rape of a prostitute is a contradiction in terms.[937] Perhaps they mean that a prostitute, by definition, is willing to have sex with anybody and therefore cannot be raped; strictly speaking, the "rapist" is guilty, at most, of failing to pay (theft) which is a lesser crime. According to this theory, even if the defendant overpowered the prostitute he did not commit rape. This argument is an obvious nonsequitur: merchants are usually willing to sell to anyone, yet theft of a merchant's wares is a crime. Similarly, "theft" of a prostitute's services, if forcible, can be punished for that reason.

A slightly more sophisticated theory is that, since prostitutes are inured to indiscriminate sex with strangers, rape does not cause them the anguish that gives the crime its peculiar horror and therefore calls for a penalty that is among the most severe in our law.[938] The short answer to this argument is that the law is to the contrary: a man who obtains sex by force or threat of force is guilty of rape even if his victim is a prostitute.[939] This is surely just. Although prostitutes are inured to sex with strangers, they are not necessarily less traumatized by the forcible aspect of the assault, and the violation of their autonomy. Furthermore, analogous "harmless rape" arguments have been rejected by courts in cases where the parties were lovers.

A different analysis is required if the defendant claims that he did not employ force; rather, he cheated the prostitute by failing to pay her fee. She concocted her accusation of rape as a retaliatory measure. Although the criminal nature of the transaction does not preclude a theft charge, this sort of sex by deception is generally held not to be rape.[940] Some police believe that it is a common explanation of rape reports by prostitutes.[941] In any event, this defense, in combination with lack of corroboration of the prostitute's story, may sometimes produce an evidentiary stalemate leading to a genuinely reasonable doubt about the defendant's guilt of forcible rape.

Regardless of its bearing on the issue of consent, evidence that the complainant was a prostitute may be admissible in order to "provide a context" explaining the circumstances of the complainant's encounter with the accused and the ground for his claim that she is lying about forcible rape.[942]

The most difficult question is whether the mere fact that the complainant is or used to be a prostitute is relevant on the issue of whether she consented to intercourse with the defendant. By occupation, prostitutes have an abnormally high propensity to consent to sex and therefore acquittals in cases involving prostitutes have a superficially rational explanation. As one court put it,

Who is more likely to consent to the approaches of a man, the unsullied virgin and the revered, loved, and virtuous mother of a family, or the lewd and loose prostitute, whose arms are open to the embraces of every coarse brute who has enough money to pay for the privilege.[943]

Although it is easy to dismiss such nineteenth-century rhetoric as absurd and sexist, the court's logic is impeccable as far as it goes. For a price, prostitutes obviously are more likely to consent to casual, nonmarital sex than virgins and homemakers. But just as an urban grocer is more vulnerable to robbery than, say, a professor, a streetwalker is more vulnerable to rape, because she frequents dangerous places and interacts with sexually aggressive males, some of whom probably believe that prostitutes are fair game for rape.[944] Consequently, prostitutes are more likely than the average woman to be forcibly raped.[945]

It is therefore fallacious to assume, merely because a woman is a prostitute, that her accusation of rape is necessarily even slightly less likely to be true than a similar accusation by another woman. To be sure, prostitutes, though more likely to be raped, may also be much more likely than other women to file false rape reports. On average, prostitutes' rape allegations may be less truthful than those of other women. But, that intuition is not self-evidently valid, and it may even be the reverse of the truth.

Mimi Silbert studied 200 juvenile and adult street prostitutes in the San Francisco Bay area.[946] Even though the study excluded rapes by the prostitutes' customers,[947] Silbert found that almost threefourths (73%) of the subjects had been raped, a total of 193 rapes.[948] Of these rapes, 71% occurred after the women became prostitutes.[949] Not only were prostitutes more likely to be raped than the average woman,[950] a large majority of these rapes (84%) were by strangers.[951] In most cases the prostitute-victim suffered some physical injury,[952] and the emotional impact of the rapes seemed to Silbert to be even greater for prostitutes than for other women.[953] Only 19% reported their rapes to the police.[954]

Summarizing her findings, Silbert noted that "the rapes in the study ..., especially those occurring after the subjects entered prostitution, involved more rapists who were strangers, more use of force, and more serious injury to the victim than rapes of women who were not prostitutes (as described in published research on rape)."[955]

Having excluded rapes by customers, Silbert concluded that the extraordinarily high rape rate revealed in this study was not directly due to the women's work as prostitutes. "Instead, the rapes were associated with the victims' vulnerability as women living in high-crime areas of the city and working during high-crime hours."[956] Given these findings, it makes no sense to treat prostitutes' rape complaints as paten fly less credible than those of other women.

In other words, on average, rape victims tend to be less sexually restrained than non-victims. Sexually adventurous women are, it seems, both more likely to consent and more likely to be raped. That is why evidence of the alleged victim's nontraditional sexual habits, standing alone, does not necessarily justify skepticism toward her testimony.

One must distinguish, however, between questions of probative value and questions of admissibility, and between the usual case and the exceptional one. We have been discussing the probative value of evidence of promiscuity on the consent issue in the general run of cases. There are situations in which, notwithstanding our generalizations, evidence of promiscuity should be admissible. In some exceptional cases, evidence about the victim's sexual history offers substantially more support to the defendant's version of events than to the complainant's.[957] Even in the more typical case, where the opposing inferences seem equally plausible, a few precedents from other legal contexts uphold admission of arguably analogous ambiprobative evidence, trusting the jury to decide which of the conflicting inferences is stronger.[958] Further, the evidence may be offered for some purpose other than to show that the complainant has a propensity to consent. For example, in some cases the complainant's sexual history is relevant evidence of a motive to falsely accuse the defendant of rape.[959] The evidence may also be necessary in order to "provide context" by explaining how the parties met and what they were doing immediately prior to the rape.[960]

For example, if the defendant offers to prove that the alleged victim used to be a prostitute, that evidence usually should be excluded as irrelevant and prejudicial. But what if he claims that their sexual encounter was consensual and that he then refused to pay the agreed sum (for instance, a heroin dealer reneged on a promised "fix"), which led to her false rape report?[961] It is difficult to see how that sort of testimony can be excluded without violating the defendant's due process right to make his defense.[962] In this hypothetical case, he cannot even tell his version of their encounter without testifying that she was a prostitute. The evidence, therefore, should be admitted to "provide context" and to supply a motive for a false accusation, even though, in our judgment, the fact that she was a prostitute, without more, does not enhance the credibility of his claim that no rape occurred.

4. Hitchhiking

Several authorities report that juries tend to disbelieve rape accusations made by female hitchhikers.[963] In support of this result, one might speculate that hitchhikers are less risk-averse and more willing to violate traditional norms of female behavior. Therefore, the argument might go, they probably are statistically more likely to engage in casual sex. Even if this conjecture is valid, it does not follow that a rape accusation by a hitchhiker is less likely to be true than an accusation by another woman. Hitchhiking is a risk factor for rape.[964] Therefore, even if one assumes that hitchhikers are on average more likely than other women to engage in casual sex, the evidence is ambiprobative and provides no basis for treating rape complaints by hitchhikers as less credible.

5. Delinquency and Delinquent Peers

Some scholars report that police are more likely to unfound a rape complaint if the woman or girl has a police record.[965] Beyond that, evaluations of rape victims' characters are also inevitably affected by the kinds of people with whom they associate. In the context of rape, the temptation is to assume that a woman or girl who associates with delinquent peers is a less credible rape complainant. Perhaps such women do tend to be less honest and less sexually selective, but it is equally plausible to suppose that they are more likely to be raped than women who have no police record and who keep better company.

One research project supports this supposition. By comparing victims of sexual assaults with a control group of nonvictims, Professor Suzanne Ageton sought to determine the best predictors of vulnerability to sexual assault.[966] The single best predictor, by far, was exposure to delinquent peers.[967] Less predictive, but also a factor, was the victim's own delinquent behavior.[968] Professor Ageton concluded:

These findings suggest that the cumulative effect of associating with delinquent peers and engaging in a fair amount of delinquency may be to raise substantially the risk of sexual assault. We speculate that there are two primary reasons why this may be so. First, the settings and circumstances in which delinquency occurs are likely to be conducive to many forms of deviance. For example, even though a female victim may have intended only to steal some drugs and get high with her friends, the situation could easily evolve into one that ends with a forced sexual experience. Second, a female involved in delinquent behavior may project a generally deviant image, which carries with it expectations about sexual behavior. Consequently, she may not be successful in restricting her deviant behavior to nonsexual acts.[969]

If this study is accurate,[970] the very sort of young woman whose rape report is often viewed with suspicion because of doubts about her character and that of her associates is, in fact, more likely to be raped than a woman of impeccable character and associations. Once again, evidence that seems to discredit the complainant proves on closer inspection to be ambiprobative.

6. Mental Problems, Runaways, and Truants

One of the major evidentiary issues in modern rape trials concerns evidence of the putative victim's mental problems and her record, if any, of psychiatric care.[971] By portraying her as "troubled," defendants seek to discredit her testimony and render their own fabricated-charge scenario more plausible.[972] Most courts authorize release of the putative victim's psychiatric records under "extreme circumstances," leaving the evidentiary door at least slightly ajar.[973]

Although post-rape psychological problems are obviously distinguishable, common sense seems to dictate admission of at least some sorts of evidence about the complainant's pre-rape mental instability. In the first place, some women who are mentally retarded or have psychiatric problems are easier to seduce than the average woman.[974] At least in extreme cases, this is relevant to the consent defense. In addition, Wigmore believed that mentally unstable women often make false rape accusations.[975] Even if Wigmore exaggerated this phenomenon, it is hard to deny that mental instability sometimes affects veracity. Given these arguments, it comes as no surprise that police often reject the rape claims of women with a psychiatric history.[976]

On the other hand, evidence of mental problems may well be ambiprobative: some unstable women are more vulnerable to rape than the average woman,[977] and no one knows whether that vulnerability is less significant than any enhanced propensity to consent or to lie.[978] Since the answer may vary from one syndrome to another, it seems best to deal with these issues on a case-by-case basis.

The criminal justice system tends to reject accusations of rape by runaways and truants.[979] Perhaps such youths are indeed more likely than other females to fabricate a rape charge, but it is at least equally plausible to suppose that they are also more likely to be raped.[980] Female juvenile delinquents presumably tend, more often than other girls, to associate with violent boys and men and are therefore at greater risk of becoming victims of violence.

7. Prior Intimacy with the Defendant

Evidence that the defendant and the complainant previously were lovers is extremely effective in persuading police not to investigate a rape complaint[981] and jurors not to believe it.[982] Judges also have been known to take a dim view of such cases.[983] Although previous consensual sex is obviously not conclusive evidence of consent on the occasion in question, nearly all commentators regard it as relevant, including thinkers as diverse as Susan Brownmiller, Herbert Wechsler, Susan Estrich, and Menachim Amir.[984] At least superficially, this sort of evidence seems superior to evidence of intercourse with other men. There are, after all, two grounds on which one may decline a sexual overture: "You're the wrong person" or "this is the wrong time." Evidence of sex with the defendant appears to eliminate the former possibility, something that evidence of sex with other men, or even of promiscuity, never does. Of course, the second possible ground for a refusal remains open, but once we know that the complainant previously had been receptive to sex with the defendant, we 0seem to have advanced the inquiry considerably. One might contend, therefore, that the consent defense is most plausible in cases where the defendant was the husband or lover of the alleged victim. This is why rape shield laws, either explicitly or by judicial gloss, contain an exception for testimony about prior sexual relations between the complainant and the defendant.[985]

What legislators and rape scholars have overlooked is the possibility that evidence of prior intimacies is ambiprobative. True, the fact that the defendant and the complainant were lovers is much more probative of consent on the occasion in question than evidence of intimacies with other men. A prior relationship may also supply a motive for a false accusation. But apparently men are more likely to rape women with whom they previously have been intimate than women with whom no such relationship existed. Koss found that 31% of acknowledged rape victims, and 76% of those who in fact were raped but did not regard it as rape, had been involved romantically with the offender.[986] According to a Bureau of Justice Statistics survey of rape victims, 28% of rapists were current or former husbands or boyfriends of the victim.[987] If such figures are even approximately correct, then it would not be implausible (though it might be mistaken) to speculate that the false proportion of rape reports directed at former lovers is lower than the false proportion of rape reports directed at other men.[988]

A thought experiment may serve to illuminate the point. Imagine that a woman has confided to you that last night her boyfriend raped her. Then, as an alternative hypothetical, imagine that the same woman has instead confided that one of her colleagues at work, with whom she had never been sexually intimate, raped her. Assume that both men, questioned by the police, denied the accusation and claimed that the intercourse was consensual. All else being equal, is the accusation against the boyfriend less credible than the accusation against the colleague?

Even if there have been no prior intimacies, a long-term dating relationship increases the popular acceptability of forced sex. In one survey of high-school students, 43% of the males (and 32% of the females) said that it is acceptable for a man to force a woman to have sex if they have dated for a long time.[989] According to a study of college students' attitudes, if a woman has been sexually intimate with a man on ten previous occasions, a subsequent incident of forcible sex is less likely to be deemed rape than if the parties had no previous intercourse.[990]

In appraising this phenomenon, one should bear in mind that in some of the cases involving prior intimacy, the parties had separated prior to the alleged rape.[991] No doubt prior intimacy increases the likelihood that the woman would consent, except perhaps in some cases of unusually acrimonious separation, but it probably also increases the likelihood that the man would feel entitled to force himself upon the woman. In addition, the quarrel that led to the separation, perhaps coupled with sexual frustration, may have inflamed the man's anger, further increasing the risk of violence toward the object of his hostility.

Prior intimacy may give the man a proprietary feeling. At common law, this feeling was sanctioned, for married men, by the rule that married women were not entitled to decline their husbands' requests for sex.[992] Even in jurisdictions that have abolished this rule, the underlying attitude presumably persists in some husbands and low ers.[993] The man may feel that it isn't really rape if the two have previously been intimate. Further, a prior relationship may give him a greater sense of security. He may know (or think he knows) that "she would never go to the police." Of course, this attitude is especially likely if he has raped her before. He may also intuit that, because of their prior relationship, "no one will believe I raped her."

In rare circumstances, a prior sexual relationship might provide a basis for a reasonable-mistake defense. For example, suppose that two lovers have previously had forcible sex, but the woman acted afterwards as if she enjoyed the experience. In such an unusual relationship, the man might reasonably come to believe that the woman's verbal or even physical resistance does not signify nonconsent. The parties have developed, so to speak, a private language. He might have a valid mistake defense if, contrary to her previous practice, she suddenly decided to treat one of these encounters as rape. Apart from such exceptional cases, we see no reason to assume, though it may be so, that those who know each other well make more mistakes than those who do not.

One might justify admission of evidence of a long-term prior relationship on the ground that such evidence is less likely to be prejudicial than evidence of promiscuity, especially in an age when monogamous, nonmarital sex has become very common. But the issue of prejudice arises, theoretically, only after a preliminary judgment that the evidence is relevant.[994]

This is not to say that the evidence of a prior relationship should be excluded. Even more often than in cases in which the woman is a prostitute, it may be essential to mention the prior relationship in order to provide a context for testimony about such matters as why they were together on the evening of the alleged crime. To take an extreme case, it would be absurd to exclude evidence that they were husband and wife, forcing the jurors to speculate about why they occupied the same house. A prior relationship may also furnish a possible motive for a false accusation. But acquittal of the defendant cannot be justified on the theory that a prior relationship increases the likelihood that the sexual encounter was consensual. We do not know whether that is true.

8. Summary

The types of evidence that are commonly used to discredit rape complainants do indeed often tend to show at least a slightly aboveaverage propensity to consent to sex, either in general or with the defendant. But they also tend to show an above-average vulnerability to rape, either in general or by the defendant.

As an empirical matter, one of these conflicting inferences may be much stronger than the other. For example, it is possible that sexually unselective women are more likely than the average woman to lie about rape.[995] Judge James Horton, in his memorandum granting a new trial in the infamous Scottsboro Case,[996] treated this as common knowledge: "History, sacred and profane, and the common experience of mankind teach that women of the character shown in this case are prone to make false accusations both of rape and of insult upon the slightest provocation or without even provocation, for ulterior purposes."[997]

Modern scholars invariably regard this sort of folk wisdom as paten fly sexist and mythical. They may be right, but in the present state of our knowledge it would be more accurate to say that (1) the proportion of chronic liars among promiscuous women may be much higher than among women in general; but (2) this is not demonstrated by "history" or "common experience;" and (3) even assuming arguendo that such women are on average less honest, we have no way of knowing whether this outweighs, statistically, the contrary inference that one can draw from their enhanced vulnerability to rape.

By the same token, a rape accusation by a man's girlfriend or wife in fact may be more (or less) likely to be false than an accusation by another woman; but nothing in common experience or the currently available academic research provides a firm basis for such a conclusion.

Most lawyers probably would argue that ambiprobative evidence should be admitted, allowing the jury to decide which inference is superior. One problem with that position is that it seems to invite unwarranted speculation on the jury's part. Another problem, particularly acute in acquaintance rape cases, is that ambiprobative "misconduct" evidence seems to have a severely prejudicial effect. In other words, juries treat the evidence as if it were much more helpful to the defense than to the prosecution' Of course, the trial judge could instruct the jury that the evidence is ambiprobative, but in some other contexts analogous instructions have not been highly effective.[998]

For these reasons, we favor a general rule excluding all ambiprobative "misconduct" testimony if offered as evidence on the issue of consent. However, even if this radical reform were adopted, it probably would not affect the outcome of many cases. Most types of ambiprobative evidence in rape cases can be offered for some purpose other than as evidence of consent. For example, evidence that the woman had been drinking would be relevant to the reliability of her memory. For this purpose, the evidence would not be ambiprobative. More often than not, ambiprobative evidence is essential to provide context by explaining how the parties met (they were lovers, or she was a prostitute or a hitchhiker) or why they behaved as they did (the reason they went to his house was to smoke marijuana, etc.).

Perhaps the public antipathy toward nontraditional rape victims can be justified by combining some of the arguments that appear to be specious or inadequate when examined individually. Consider, for example, a case in which the complainant is a hitchhiking, promiscuous, substance-abusing, teenage runaway. While it is true that such a woman is statistically more vulnerable to rape than most women, it may not be true that this enhanced vulnerability totally offsets her presumptively enhanced propensity to consent, coupled with the possibility that such a woman is more likely than the average woman to be a chronic liar.[999] Taken together, these two grounds for viewing her uncorroborated rape complaint with some suspicion may outweigh the rape-vulnerability reason for presuming that her accusation is extraordinarily credible.

The question is empirical rather than logical. We must therefore rely, so the argument might go, on the judgments of police and juries who have immersed themselves in the facts of individual cases. The biases of such decisionmakers, serious though they sometimes are, may be outweighed by the advantages of having investigated concrete cases. Only such decisionmakers are in a position to perceive nuances--for example, of evasion and inconsistency--that for all we know may correlate with one or more of the "bad victim" traits. If these decisionmakers tend to disbelieve the "bad" woman's rape complaint, perhaps that recurrent disbelief reflects a cumulative wisdom, distilled from many thousands of individual cases and evaluators, that no cloistered theorist or social-scientific experiment can match.

We think this argument has some merit. This is a field in which first-hand experience in investigating rape complaints would be invaluable, but there is no consensus among those who have immersed themselves in the facts of rape cases.[1000]

The social-scientific literature canvassed earlier in this article demonstrates that many people, especially men with traditional values, blame the rape victim for legally irrelevant reasons,[1001] Therefore, it seems certain that, as feminists contend, some detectives and jurors are improperly predisposed to reject accusations of acquaintance rape. While the opposite bias undoubtedly exists, it appears to be much less common.

Jurors who are lenient toward men accused of raping a nontraditional victim may not be expressing a judgment about the comparative plausibility of the prosecution's rape scenario and the defendant's seduction-fabricated charge scenario. Their motivation instead may be a combination of legally appropriate concerns about whether guilt has been proved beyond a reasonable doubt and legally inappropriate (but common) feelings that not all rapes are equally harmful, and not all rapists are equally dangerous. Consider, for example, a case in which the complainant and the defendant used to be lovers. She claims that he raped her; he claims that the sex was voluntary and that later, after he told her that he didn't love her any more, she fabricated a rape charge. Assume that neither story is patently untrue. Why is it that jurors are especially likely to acquit in such cases?

It is hard to believe that modern juries strongly disapprove of monogamous, nonmarital sex. Absent other discrediting evidence, we doubt that jurors regard the woman in this type of case as morally repugnant. But, if they are uncertain about who is telling the truth, decision theory tells us that they will try to take account of the potential impact of an erroneous verdict. That is, they will seek to minimize their expected regret from having reached an erroneous decision,[1002] If they erroneously acquit the defendant, they will have failed to punish a man who raped a woman with whom he had previously been intimate. Although many authors stress that this sort of rape can have a devastating psychological impact,[1003] it seems probable that most jurors are not terribly disturbed by the prospect of having failed to incarcerate a single rapist of this sort.[1004] For retributive purposes and deterrence of the defendant ("specific deterrence"),[1005] they may regard the ordeal of being publicly accused and tried as sufficient. Without necessarily condoning the man's conduct, or even feeling that it should not be a crime, jurors may be comfortable taking the risk of an erroneous acquittal in this type of case if they are unsure what occurred. This acquittal has a price, but it is not like the risk of acquitting an armed serial rapist who has terrorized a city.[1006] If, on the other hand, they make the opposite mistake, convicting an innocent man falsely charged with raping his girlfriend, they will have "destroyed an innocent man's life." Decision theory predicts that on these assumptions jurors who aren't quite certain what happened will vote to acquit,[1007] and that is indeed what they usually do.[1008] The same sorts of feelings may explain acquittals in some other nontraditional victim cases, particularly those involving prostitutes.

IV. CONCLUSION

The major premises of the first wave of rape law reforms were that the criminal justice system discriminates against victims of nonaggravated acquaintance rape, and that law reforms can substantially alleviate the problem. To provide a foundation for consideration of more recent reform proposals, we reevaluated those premises from several perspectives. We began by considering evidence of anti-victim bias. The attrition rate in acquaintance rape cases is extremely high, but acquaintance cases have high attrition rates for all crimes, and, given the uniqueness of each crime's proof problems, there is no reason to expect equal attrition rates in, for example, sexual and nonsexual assaults by acquaintances.

The main source of case attrition in acquaintance rape cases is the victim's reluctance to pursue legal redress. A substantial, if not enormous, majority of acquaintance rape victims either do not report the crime or, having reported it, later decline to press charges. Scholars often attribute this phenomenon, in large part, to victims' fears of a hostile or overly skeptical criminal justice system, and especially the practice of "putting the victim on trial" by exposing her sexual history. So far, empirical studies indicate that this hypothesis exaggerates the role of legal rules in deterring reporting of rape. The weight of the evidence suggests that (1) the incidence of rape has declined in recent decades; (2) the proportion of victims who report the crime has risen during the same period; and (3) this rising propensity to report has been due, mainly if not entirely, to attitudinal changes associated with the women's movement rather than law reforms, such as shield laws, in individual jurisdictions. Although we cannot be certain, our guess is that rape reporting is inhibited chiefly by nonlegal factors such as embarrassment about having been raped, or about the surrounding circumstances, self-blaming, fear of the investigatory and adversarial processes (quite apart from the specific rules of law criticized by reformers), the victim's desire to resume a normal life, her peers' and family's attitudes, and often even her desire to preserve her relationship with the rapist. For an increasing percentage of victims, these disincentives are outweighed by a commitment to see justice done. Although this commitment generally does not appear to be enhanced by legal changes in the jurisdiction where the crime occurred, law reforms throughout the country may have contributed, cumulatively, to a growing sense that the victim is not to blame, will be treated fairly, and may prevail. In this indirect way, reforms in one state may encourage reporting in other states. If so, reforms may be somewhat more effective than they appear to be when studied in individual states. We simply do not know.

Some police detectives are suspicious of reports of acquaintance rape, especially when they disapprove of the woman's character or conduct prior to the rape. In our judgment, this bias cannot be justified by any presently available evidence that rape complaints by normviolating women tend to be less credible. Nor can it be justified by the difficulty of obtaining convictions in such cases: that sort of decision should be made by prosecutors rather than by police, although to the extent that police accurately predict prosecutorial decisions we cannot confidently say that this procedural point matters much to victims. In any event, there is some evidence that most women who report rape have a favorable opinion of their treatment by the police.

Several studies have found correlations between the police response to a rape complaint and the complainant's race, her welfare status, her age, her lifestyle, and other presumably irrelevant factors. These studies, however, typically cover only one police department during a short period and are often ten or twenty years old. More fundamentally, the studies of police unfounding decisions do not (and obviously cannot) reveal which rape complaints were untruthful. Typically, the studies are rigorously quantitative and do not even reveal what proportion of rape complaints contain inconsistencies or other indicia of untruthfulness. Lacking such information, it is impossible to say whether a correlation between some attribute of the putative victim and the likelihood that her complaint will be unfounded is due to fabricated complaints, truthful but suspicious sounding complaints, unjustifiable police bias, or some combination of the three. Nor do the studies provide evidence from which a reader can infer how often police unfound a rape complaint that might have led to a conviction. Since prosecutors often decline to file charges, and juries often acquit, even in the relatively strong cases in which the police regard the complaint as credible, it seems probable that police attitudes, however mistaken they may be in some or even many cases, are rarely outcome determinative.

This is not to say that excessive police skepticism is harmless. Unfair treatment of a crime victim is still unfair, even if it does not affect the likelihood that the offender will be punished. Recognizing this, many police departments have created sexually-integrated units to deal with rape and other sex-crime investigations. Some scholars believe that this administrative change has contributed to victims' willingness to report rapes.

Prosecutors naturally seek high conviction rates, and consequently are reluctant to prosecute cases in which the odds are long against a conviction. Given the difficulties of corroborating accusations of acquaintance rape, this prosecutoria! attitude has a more severe impact on such cases than it does in most other areas of the criminal law. For this reason, we concur with scholars who urge prosecutors to take more chances in acquaintance rape cases than they customarily do in other types of cases. Comparative conviction-rate data suggest that in some cases prosecutors are already doing so. In any event, the potential benefits of a futile prosecution should be balanced against potential psychic costs to the rape victim, and missed opportunities to pursue other, possibly more winnable cases.

There is a great deal of anecdotal and social-scientific evidence of public (and jury) bias against norm-violating victims of acquaintance rape. This bias is most prevalent among people, especially but not exclusively males, with traditional sex-role values. Given the variety of such evidence, the methodological weaknesses of individual studies and sources do not seem fatal.

This bias appears to have declined in recent decades, with the result that juries are at least somewhat more sympathetic to the prosecution in acquaintance rape cases. Although we lack definitive, national studies, the role of law reforms in achieving this result appears to have been peripheral: as with reporting rates, rape conviction rates have not been highly responsive to legal reforms. This should not be surprising. We favor rape shield laws, for example, but it is a major mistake to suppose that, with a properly-drafted shield law, "victim blaming" will or should disappear from the courtroom in acquaintance rape cases. Given our adversarial system, the defendant in a consent-defense case will inevitably try to discredit the alleged victim.[1009] How else can he account for the fact that, as he contends, she has falsely accused him of rape? Even if he claims that the "rape" was due to a misunderstanding (rather than a fabricated accusation), he will try to persuade the jury that his mistake was reasonable, or in other words that the woman was partly if not entirely to blame for the misunderstanding.

Thus, the law cannot simply "shift the focus of the trial from the woman's conduct to the man's," as so many reformers have urged,[1010] To a considerable degree, evidence about her conduct will be admitted and weighed as evidence about his state of mind and conduct. In consent-defense rape cases, as in civil contracts cases, the two are inseparable. The jury, after all, has to decide who is lying or, if the evidence suggests an honest mistake as to consent, whether the mistake was reasonable. There is no way to engage in this inquiry without examining both parties' alleged behavior.

The problem is not that juries focus on the woman's conduct but that they sometimes draw unwarranted inferences of consent from her violation of sex-role norms. It is difficult to see how this can be prevented by rules of law. To be sure, a well-drafted shield law will exclude certain kinds of evidence about the woman's sexual history with other men. Such laws are desirable even if they do not change many case outcomes. But defendants still have opportunities, many of them legitimate or inevitable, to discredit the rape complainant's character. Even without formal sexual history evidence, the jury will learn, for example, that the woman was hitchhiking, or that the parties had been drinking, or that the woman had behaved in a "lewd" manner shortly before the rape, or that she had been the defendant's girlfriend, or even, sometimes, that she was a prostitute.

Even without a formal corroboration requirement, juries still want some corroboration. And even without an instruction that resistance by the victim is an essential element of the crime, juries will still consider the degree of resistance in determining whether the woman consented or at least misled the defendant.

For all these reasons, conviction rates cannot easily be changed by waving a legal wand. This is not a counsel of despair. Even when law reforms are ineffective, attitudinal changes, perhaps indirectly helped along by the reforms, may achieve the desired result. Besides, conviction rates are a poor measure of the success of a criminal justice system. Few readers will need to be reminded that fairness to defendants is also important. No less important, and traditionally underemphasized in most of the criminal law literature, is fairness to victims. Whether the rapist is punished may be less important to his victim than whether she perceives that officials have responded empathetically and fairly to her report of a terrible crime. In this respect, there have been some improvements in the justice system,[1011] and no doubt further progress can be made.

In Part II of the article we evaluated three general defenses of leniency in acquaintance rape cases, beginning with the formerly popular idea that many rape reports are false. Although most scholars seem to believe that this issue has been settled by common sense, or by research showing that nearly all reported rapes actually occurred, the question is still open. There is no consensus among the handful of scholars who have tried to evaluate the veracity of a sample of rape complaints. Nor is there any way to demonstrate, deductively, that the proportion of false reports is minuscule.

While it is impossible to know exactly what proportion of rape complaints are false, some scholars have tabulated the percentages of complaints that were subsequently retracted. These studies covered only a couple of jurisdictions and yielded radically different results. Pending the results of further research, we can only say that the proportion of rape reports that are false may be higher than modern scholars usually suppose.

At first blush, this possibility seems to threaten the whole edifice of modern rape scholarship. If false reports are common, the oft-criticized public and official skepticism will indeed appear in a much more favorable light. But false reports do not necessarily lead to mistaken convictions. There is no reason to suppose that many women and girls make false reports, do not retract them, and manage to deceive a detective, a prosecutor, and a jury. On the contrary, the most impressive "false report" study was based on quickly-obtained recantations during police investigations and therefore provides no evidence that false reports survive the screening process. Although any such claim is necessarily speculative, the anecdotal and social-scientific evidence both suggest that the proportion of erroneous convictions in acquaintance rape cases may be lower than in many other types of criminal cases, especially those in which the government relies on an eyewitness.

We considered the possibility that acquittals in acquaintance rape cases are attributable to the prosecution's heavy burden of proof, and the difficulty of winning a "swearing contest." We found that the prosecution is at a disadvantage in pure swearing contests, not only in acquaintance rape cases but in some other crimes as well. The burden of proof is a legitimate reason for case attrition in those acquaintance rape cases where both parties' versions of the encounter are plausible and corroboration of the alleged victim's account is lacking. Although uncorroborated accusations probably are not a high proportion of the rape cases in which charges are filed, they presumably are common among rapes that are not reported or that do not survive pre-trial screening by police and prosecutors. In such cases the problem is not the oft-criticized corroboration requirement, but rather the difficulty of eliminating reasonable doubts without corroboration.

We are not suggesting that it is often legally impossible to secure a conviction for an acquaintance rape. Ordinarily, the victim's testimony is sufficient to create a jury issue. Our point, rather, is that juries, even if they have no improper biases, are likely to acquit if the prosecution lacks strong corroborative evidence and the defendant suggests a plausible motive for a fabricated rape charge. In such cases, much depends on whether the jurors find the defendant's "false accusation" scenario plausible. This in turn depends partly on jurors' intuitions about the rate of false rape charges.

According to Kalven and Zeisel's data, trial judges in the 1950s frequently disagreed with acquittals in nonaggravated acquaintance rape cases. This finding is broadly consistent with most of the more recent anecdotal and social-scientific evidence about acquaintance rape cases, particularly those with "nontraditional" victims. It seems clear, therefore, that the burden of proof does not fully explain the historic leniency of juries in these cases. Some evidence suggests that excessive leniency has substantially declined in recent years, but the reports are fragmentary, and regional variations probably persist.

Approaching the problem from a different perspective, we inquired whether rape complaints are less credible when the evidence reveals prior intimacy with the accused, contributory negligence, prostitution, promiscuity, substance abuse, hitchhiking, or some other violation of traditional norms of female conduct. To analyze this subject, we introduced the concept of ambiprobative evidence. Applying that concept to rape-victim-norm-violation evidence, we concluded that in most if not all such cases the woman's behavior is indeed at least slightly suggestive of consensual intercourse. None of the conventional grounds for rejecting this inference is wholly persuasive. But the same evidence is, we submit, suggestive of vulnerability to rape. Perhaps one of these inferences is empirically stronger than the other, but we know of no basis for that surmise in the academic literature or common experience. This being so, jurors' tendency to acquit the defendant in these types of cases is not defensible by reference to the known probative value of the evidence.

Once more, however, the potential effect of law reform is limited. Although we believe that courts should usually exclude ambiprobative evidence when offered to show consent, such evidence must often be admitted to provide context by explaining the circumstances of the parties' encounter. The evidence may also be admissible for other legitimate purposes such as to show a motive for a false accusation. Once the evidence is admitted, it is difficult to prevent juries from misusing it.

We have speculated that acquittals in nontraditional victim cases are sometimes due to mixed motives. Uncertain who is telling the truth, jurors may prefer to run the risk of having acquitted the rapist of a prostitute, or of a former lover, rather than running the risk of ruining an innocent man's life. While we might balance the interests differently, in close cases this is an understandable choice. It is difficult to see how new rules of law would alter these perhaps unconscious calculations; such reasoning is already legally improper, because the law makes no such distinctions among victims.

We conclude that although official bias has played an important role, most rape-case attrition appears to be due to a combination of the victim's unwillingness to seek legal redress, the prosecution's burden of proof in criminal cases, and jurors' attitudes. The first and third factors are affected by evolving public attitudes toward rape, which in turn may be enhanced by national publicity given to law reforms. But in most individual jurisdictions changes in legal rules do not seem to generate noticeable changes in case outcomes in that jurisdiction.

In sum, there is good news and bad news. The bad news is that rape law reforms appear to play a much more secondary instrumental role than legal scholars like to believe. The good news is that, for whatever reason, progress has occurred: the rape rate appears to have declined since the 1970s; victims seem to be more willing to report to the police; police and prosecutors are more likely to employ specialized sex-crimes units with expertise in the emotional as well as the legal aspects of the crime; and there is anecdotal evidence that juries have become more sympathetic to acquaintance rape victims.

From the victims' perspective, these developments are as welcome as they would be if they were mainly due to law reforms. From the legal reformer's perspective, realism about the limits of law reform is essential if false hopes are to be avoided. In our next article, we will revisit this theme as we examine the latest proposals for changing rape law.

We are very grateful to Vivian Berger, Linda Fairstein, Mary Louise Fellows, Richard Frase, Patricia Frazier, Julie Horney, Mary Koss, David Lykken, Roger C. Park, Barbara Reskin, Marian Saksena, Stephen Schulhofer, Emily Sherwin, Michael Tonry, and Frank Zimring for suggestions or encouragement based on an earlier version of this article. Of course, they are not responsible for any remaining errors, and they do not necessarily endorse all of our conclusions. Diligent research assistance or citechecking was provided by Robert Ballieu, Staci Bartsch, Jason Hannan, Kristen Ludgate, David Marcus, Erick Ottoson, Robin Preble, David Shamla, Rebecca Simpson, Andrew Wronski, and Susie Zamzow. We are also indebted to Mary Jane Luedtke for typing services above and beyond the call of duty.

1 See, e.g., SUSAN BROWNMILLER, AGAINST OUR WILL: MEN, WOMEN AND RAPE (1975); SUSAN ESTRICH, REAL RAPE (1987). Brownmiller and Estrich, in particular, have had a major influence on subsequent discussions about rape and rape law. See also Vivian Berger, Man's Trial, Woman's Tribulation: Rape Cases in the Courtroom, 77 COLUM. L. REv. 1 (1977); Gerald D. Robin, Forcible Rape: Institutionalized Sexism in the Criminal Justice System, 23 CRIME & DELINQ. 136 (1977).

2 See infra notes 102-13 and accompanying text (discussing high attrition rates in rape case processing).

3 See infra notes 187-88, 226-50 and accompanying text.

4 See, e.g., LYNDA LYTLE HOLMSTROM & ANN WOLBERT BURGESS, THE VICTIM OF RAPE: INSTITUTIONAL REACTIONS 56 (1978). Despite the passage of rape shield laws, evidence of a complainant's sexual habits is still admissible in some cases, either because the law was badly-drafted, or because the evidence is relevant for a legitimate purpose such as suggesting a motive for a false accusation or providing a context for the defendant's consent defense. See generally Harriett R. Galvin, Shielding Rape Victims in the State and Federal Courts: A Proposal for the Second Decade, 70 MINN. L. REV. 763, 812-905 (1986).

5 See infra note 616 and accompanying text.

6 See, e.g., ESTRICH, supra note 1, at 27-56. See generally infra notes 132-38, 154-56 and accompanying text.

7 See, e.g., Alan C. Acock & Nancy K. Ireland, Attribution of Blame in Rape Cases: The Impact of Norm Violation, Gender and Sex Role Attitude, 19 SEx ROLES 179, 187 (1983); Judith S. Bridges, Perceptions of Date and Stranger Rape: A Difference in Sex Role Expectations and RapeSupportive Beliefs, 24 SEX ROLES 291, 304-05 (1991).

8 See, e.g., RICHARD A. POSNER, SEX AND REASON 388 (1992).

9 See, e.g., HOLMSTROM & BURGESS, supra note 4, at 39-43 (unfounding of rape reports correlated with victim's drinking or hitchhiking). See generally infra notes 863-83, 952-53 and accompanying text.

l0 See, e.g., BROWNMILLER, supra note 1, at 353-55; Berger, supra note 1, at 55-56. See a/so infra note 594.

11 See, e.g., BROWNMILLER, supra note 1, at 11-30; ESTRICH, Supra note 1, at 56, 23-25, 62-69.

12 See, e.g., ESTRICH, supra note 1, at 56, 42-56. See generally infra notes 603-719 and accompanying text (discussing false rape reports).

13 See, e.g., LINDA A. FAIRSTEIN, SEXUAL VIOLENCE: OUR WAR AGAINST RAPE 151-52 (1993); ALICE VACHSS, SEX CRIMES 140-43 (1993); Lisa Frohmann, Discrediting Victims'Allegations of Sexual Assault: Prosecutorial Accounts of Case Rejections, 38 SOC. PROCS. 213, 217-24 (1991).

14 See, e.g., LEE MADIGAN & NANCY C. GAMBLE, THE SECOND RAPE (1991);JOYCE. WILLIAMS & KAREN A. HOLMES, THE SECOND ASSAULT (1981); Berger, supra note 1.

15 See infra notes 516-28 and accompanying text.

16 See infra text accompanying notes 400-07 (describing juror leniency in aggravated versus non-aggravated sexual assault cases).

17 See, e.g., BROWNMILLER, supra note 1, at 19-24. For an overview of the resistance requirement at common law, see ESTRICH, supra note 1, at 29-41.

18 In recent decades, the resistance requirement has been eroded by reforms. Where they existed, requirements that the victim resist continuously and to the utmost have been universally repealed; most states have gone further, eliminating all requirements of physical resistance. See Susan Estrich, Rape, 95 YALE L.J. 1087, 1123-24 (1986) [hereinafter Estrich, Rape]. However, most states continue to regard force as an essential element of rape or sexual assault. See generally Cynthia Anne Wickton, Note, Focusing on the Offender's Forceful Conduct: A Proposal for the Redefinition of Rape Laws, 56 GEo. WASH. L. REv. 399 (1988). As Estrich points out, force and resistance are two sides of the same coin, since force is necessary only if the woman resists. ESTRICH, supra note 1, at 58-71.

19 The Model Penal Code's version of the corroboration requirement states: "No person shall be convicted on any felony under . . . [the sexual offenses] article upon the uncorroborated testimony of the alleged victim." MODEL PENAL CODE (Section) 213.6(5) (1985). Almost all states, with the exception of Nebraska, have now eliminated the requirement that a rape complainant's testimony be corroborated in cases of forcible rape. SANFORD H. KADISH &: STEPHENJ. SCHULHOFER, CRIMINAL LAW AND ITS PROCESSES 402 (5th ed. 1989). See generally ESTRICH, supra note l, at 43-47 (overview of the corroboration requirement).

20 Although the wording varies from one state to another, Morris notes that:

most cautionary instructions contain three common elements...: (1) rape is a charge that is easily made by the victim, (2) rape is a charge that is difficult for the defendant to disprove, and (3) the testimony of the victim requires more careful scrutiny by the jury than the testimony of the other wimesses in the trial.

A. Thomas Morris, Note, The Empirical, Historical and Legal Case Against the Cautionary Instruction: A Call for Legislative Reform, 1988 DUKE LJ. 154, 154-55. See infra note 606 and accompanying text.

21 See, e.g., Cheryl J. Oros & Donald Elman, Impact of Judges' Instructions Upon Jurors' Decisions: The "Cautionary Charge" in Rape Trials, 10 REPRESENTATIVE RES. Soc. PSYCH. 28, 34 (1979).

22 See, e.g., BROWNMILLER, supra note 1, at 14-15; ESTRICH, supra note 1, at 8283.

23 See, e.g., Berger, supra note 1, at 12-100; Galvin, supra note 4, at 778-802.

24 See, e.g., Galvin, supra note 4, at 767.

25 See, e.g., LINDA BROOKOVER BOURQUE, DEFINING RAPE 110 (1989) (the "overriding" objective of reforms was to shift the burden of proof from the victim to the defendant); see alsO JEANNE C. MARSH ET AL., RAPE AND THE LIMITS OF LAW REFORM 21-23 (1982); Patricia Seades & Ronald J. Berger, The Current Status of Rape Reform Legislation: An Examination of State Statutes, 10 WOMEN'S RTS. L. REP. 25, 26 (1987).

26 See, e.g., Martin D. Schwartz & Todd R. Clear, Toward a New Law on Rape, 26 CRIME & DELINQ. 129, 135-36 (1980); Searles & Berger, supra note 25, at 26.

27 See, e.g., Stevi Jackson, The Social Context of Rape: Sexual Scripts and Motivation, in RAPE SOCIETY. READINGS ON THE PROBLEM OF SEXUAL ASSAULT 16, 19-20 (Patricia Seades & Ronald J. Berger eds., 1995) [hereinafter RAPE & SOCIETY].

28 See, e.g., Susan Estrich, Palm Beach Stories, 11 LAW & PHm. 5 (1992) [hereinafter Estrich, Palm Beach Stories]; Robin, supra note 1, at 150-53; MARSH ST AL., Supra note 25, at 2123; Helene Sasko & Deborah Sesek, Note, Rape Reform Legislation: Is it the Solution ?, 24 CLEV. ST. L. REV. 463 (1975).

29 See generally Searles & Berger, supra note 25, at 25. 30 Id. The same authors list the goals of the rape law reformers: 1) increasing the reporting of rape and enhancing prosecution and conviction in rape cases; 2) improving the treatment of rape victims in the criminal justice system; 3) achieving comparability between the legal treatment of rape and other violent crimes; 4) prohibiting a wider range of coercive sexual conduct; and 5) expanding the range of persons protected by law. Id. 31 Almost every state has eliminated the requirement that a rape complainant's testimony be corroborated in cases of forcible rape. KADISH & SCHULHOFER, Supra note 19, at 402. Some retain a corroboration requirement for statutory rape. Id.

32 See RonaldJ. Berger et al., The Dimensions of Rape Reform Legislation, 22 L. & Soc'Y REv. 329, 332 (1988). As of 1988, over half of the states still allowed a cautionary instruction in forcible rape trials. Morris, supra note 20, at 156.33 See generally Galvin, supra note 4, at 765-66.

34 Seades & Berger, supra note 25, at 25-26.

35 See infra notes 187-90 and accompanying text.

36 See infra text accompanying notes 75-86.

37 See infra note 545.

38 See infra notes 443-45 and accompanying text.

39 See infra text accompanying notes 546-94.

40 See, e.g., ESTRICH, supra note 1, at 94. This is already the law in a number of states.

41 Lynne N. Henderson, Review-Essay: What Makes Rape a Crime?, 3 BERKELEY WOMEn'S LJ. 193, 211-19 (1987-88).

42 See, e.g., ESTRICH, supra note 1, at 84-86 (elimination of resistance requirement is ineffective unless force requirement is also eliminated); see generally Wickton, supra note 18.

43 See, e.g., StephenJ. Schulhofer, Taking Sexual Autonomy Seriously: Rape Law and Beyond, 11 LAW & PHIL. 35, 71-77 (1992) (stating that: "nothing less than a crystallized attitude of positive willingness should ever count as consent"); Beverly Balos & Mary Louise Fellows, Guilty of the Crime of Trust: NonstrangerRape, 75 MINN. L. REv. 599, 607 (1991) (observing that: "[s] ilence or other passive behavior by the victim would not constitute sufficient evidence of consent between nonstrangers").

44 Forcible rape differs from statutory rape in that the prosecutor must ordinarily prove that the victim did not consent, while in statutory rape the underage victim's youth nullifies her consent. S WHARTON's CRIMINAL LAw (Sections) 280, 285 (Charles E. Torcia ed., 15th ed. 1995) [hereinafter WHARTON'S]. In certain narrowly defined and relatively rare types of cases an adult is presumed incapable of consent because of her mental defect or unconsciousness or the extortionate or deceptive conduct of the rapist. See generally Ernst Wilfred Puttkammer, Consent in Rape, 19 ILL. L. Rev. 410 (1925). For a brief outline of the legal issues involved in statutory rape, see WHARTON'S, supra, (Section) 285.

For an introductory bibliography on homosexual rape, see Panel Discussion: Men, Women and Rape, 63 FORDHAM L. REV. 125, 127 nn.9, 10 (1994).

45 See infra notes 225-80, 347-94 and accompanying text.

46 See, e.g., Berger, supra note 1, at 12-100; Galvin, supra note 4, at 778-801.

47 See, e.g., Galvin, supra note 4, at 767. In this article, "promiscuous" means much less sexually restrained than most women of the same age and marital status. Unfortunately, we know of no alternative term that is either more precise or less pejorative.

48 See, e.g., SUZSANNA ADLER, RAPE ON TRIAL 89-93 (1987).

49 See WHARTON'S, supra note 44, (Section) 276; Leigh Bienen, Rape III: National Developments in Rape Reform Legislation, 6 WOMEN'S RTS. L. REP. 170, 174 (1981). Forcible rape is often divided into degrees, in order to permit more serious punishment in cases in which the offense is aggravated by physical injuries or multiple offenders. See WHARTON'S, supra note 44, (Section) 284.

50 See generally WHARTON'S, supra note 44, (Section) 276.

51 But see Julie Horney& Cassia Spohn, The Influence of Blame and Believability Factors on the Processing of Simple Versus Aggravated Rape Cases, 34 CRIMINOLOGY 135 (1996) (outcomes of non-aggravated and aggravated rape cases do not differ in Detroit).

52 See, e.g., ESTRICH supra note 1 at 17-20. Some scholars use the term "aggravated rape" to cover rapes (1) by strangers, or (2) with multiple assailants, or (3) resulting in extrinsic injuries. All other rapes are called "simple rape." See, e.g., ESTRICH, supra note 1, at 19-20; HARRY KALVEN, JR. & HANS ZEISEL, THE AMERICAN JURY 252 (1966). While we have no serious objection to this terminology, our research indicates that for purposes of legal sociology, gang rapes may be functionally equivalent to ordinary acquaintance rapes rather than to the other types of aggravated rape. Although it obviously makes sense to treat gang rapes as "aggravated" for sentencing purposes, in most rape scholarship the main point of the distinction between aggravated and simple rape is that the justice system is thought to work reasonably well in aggravated rape cases, while functioning poorly in simple rape cases. See United States v. X-citement Video, Inc., 115 S. Ct. 464, 475 (1994); see, e.g., Estrich, supra note 1, at 3-4, 60; FAIRSTEIN, supra note 13, at 151-52.

For this purpose, the correct classification of gang rapes is uncertain. Susan Estrich believes that the criminal justice system is more sympathetic to the prosecution in gang rape cases than in ordinary acquaintance rape cases. ESTRICH, supra note 1, at 10, 99. But see Gary D. LaFree, Official Reactions to Social Problems: Police Decisions in Sexual Assault Cases, 28 Soc. PROBS. 582 (1981) [hereinafter LaFree, Official Reactions]. In a major study of a police department's processing of sexual assault cases, LaFree found that:

assaults involving more than one offender were less likely to be filed as felonies. Interviews indicated that detectives were suspicious of sexual assaults involving more than one offender--particularly when these cases also involved more than one victim, when victims and offenders were acquainted prior to the incident, or when victims and of* fenders were young. Several detectives referred to cases with these characteristics as 'party rapes,' and suggested that such incidents deserved less serious attention than other complaints.

Id. Accord THoMAs W. MCCAHILL ET AL., THE AFTERMATH OF RAPE 112 (1979) (stating that Philadelphia police believed "only about two-thirds of gang rape allegations"). These findings suggest that gang rapes by acquaintances are not treated by the police as "aggravated" rapes.

53 ESTRICH, supra note 1, at 8.

54 Minor injuries such as a scratch or a bruise can often be explained as selfinflicted or due to other causes. See infra note 313.

55 See, e.g., KALVEN AND ZEISEL, supra note 52, at 252-55. Kalven and Zeisel found that juries were four times as likely to convict in aggravated rape cases (where the accused was a stranger to the victim, or extrinsic injuries to victim resulted, or multiple assailants were involved) as they were in simple rapes. Id. See also ESTRICH, supra note 1 at 34, 13-14; Judith S. Bridges, Perceptions of Date and Stranger Rape: A Difference in Sex Role Expectations and Rape-Supportive Beliefs, 24 SEx ROLES 291 (1991) (subjects more likely to incorporate sexrole expectations and rape-supportive beliefs into perceptions of acquaintance rape than stranger rape; less likely to attribute causes of rape to victim's failure to control situation, to a "misunderstanding" between parties, or to victim's desire for intercourse if parties were not previously acquainted).

56 The National Victim Center's survey data revealed that only 22% of rape victims had been raped by someone they had never seen before, or did not know well; 9% by husbands or ex-husbands; 11% by fathers or step-fathers; 10% by boyfriends or ex-boyfriends; 16% by other relatives; 29% by other non-relatives such as friends and neighbors; and 3% were unsure or refused to answer. NATIONAL VICTIM CENTER, CRIME VICTIMS RESEARCH AND TREATMENT CENTER, RAPE IN AMERICA 4 (1992) [hereinafter RAPE It4 AMERICA]. According to the most methodologically sophisticated national survey of sexual practices, only 4% of women who have been "forced to do something sexual that they did not want to do" were victimized by a stranger. EDWARD O. LAUMANN ET AL., THE SOCIAL ORGANIZATION OF SEXUALITY. SEXUAL PRACTICES IN THE UNITED STATES 338 (1994). In 9% of the incidents of "forced" sex, the perpetrator was the woman's spouse; in another 46% he was someone with whom she was in love; in 22% he was someone she knew well; and in another 19% he was an acquaintance. Id. Although the authors caution that their concept of forced sex is broader than the legal definition of rape, id. at 335, it seems clear that most rapes are perpetrated by men who know the victim. See, e.g., FAIRSTEIN, supra note 13, at 129 (stating that "[M]ore than 50 percent of reported rapes were--and continue to be--assaults by men who were known to their victims"); Mary P. Koss, The Hidden Rape Victim: Personality, Attitudinal and Situational Characteristics, 9 PSYCHOL. WOMEN Q. 193, 206 (1985) (reporting that most rapes involved acquaintances or romantic intimates). According to one estimate, only 15% of rapes are perpetrated by strangers. JULIE A. ALLISON & LAWRENCE S. WRIGHTSMAN, RAVE: THE MISUNDERSTOOD CRIME 51 (1993). Surveys to determine the proportion of rapes committed by acquaintances are likely to err on the low side, unless the survey methodology takes account of what Koss calls "hidden rape"--i.e., the fact that many women whose answers to survey questions reveal that they have in fact been raped do not label the experience as such. Koss, supra, at 206. Victimization surveys which "require that a woman conceptualize and report her experience as rape" are therefore suspect, Id. In Koss's study of university women, 43% of the women whose answers revealed that they had been raped denied that they had been raped, and most of these women "were assaulted by an acquaintance or romantic intimate." Id. Of course, researchers must also take care not to define "rape" too broadly. See Neil Gilbert, The Phantom Epidemic of Sexual Assault, 103 PuB. INTEREST 54 (1991).

Some rapists are on the borderline between acquaintances and strangers: for example, a neighbor or co-worker whom the victim recognized but had not met, or a driver who rapes a hitchhiker after preliminary conversation. In addition, some types of acquaintances are functionally equivalent to strangers because they cannot plausibly claim that the victim consented or because the nature of the relationship is such that the public blames the man even if the relationship was consensual. The latter type of case is often statutory rape or incest, but if the man employed force he can be charged with the more serious crime of forcible rape.

57 See infra notes 768439 and accompanying text.

58 See ESTRICH, supra note 1, at 4-6.

59 Many of the traditional norms of female morality can be characterized as at least partly prudential; even today sexual activity generally entails greater risks for females than for males.

60 See infra notes 470-71 and accompanying text.

61 See, e.g., ESTRICH, supra note 1, at 11-14.

62 Other defenses--for example, diminished responsibility--are, of course, possible. See generally WHARTON'S, supra note 44, (Section) 286; JOSHUA DRESSLER, UNDERSTANDING CRIMINAL LAW (Sections) 16.01-16.03, 24.01-24.07, 26.01-26.03 (1987).

63 See, e.g., Karen S. Calhoun & Ruth M. Townsley, Attributions of Responsibility for Acquaintance Rape, in ACQUAInT^NeE RAPE 57, 65 (Andrea Parrot & Laurie Bechhofer eds., 1991) [hereinafter ACQUAINTANCE RAPE]; FAIRSTEIN, supra note 13, at 122.

64 See, e.g., ESTRICH, supra note 1, at 7, 92-104; Balos & Fellows, supra note 43, at 601-02 (arguing that rape law should incorporate the common law confidential relationship doctrine).

65 According to one scholar, a prior sexual relationship between the alleged victim and the defendant is a "real trump card for the defense," which makes it "practically impossible to convince the jury that the incident in question was anything other than one in a long series of consensual acts." ADLER, supra note 48, at 53. After studying a random sample of English rape trials, Adler concluded that "[w]hen the victim agrees that there had been sexual involvement [with the defendant] in the past, the defendant is almost invariably acquitted." Id. at 91. Adler notes that the prior relationship not only discredits the claim of rape in thejury's eyes, but also is suggestive of motives for a fabricated rape charge. Id. at 92-93.

Some police are also swayed by evidence of a prior relationship between the parties. Holmstrom and Burgess quote a Boston policeman: "[T]he officer said loudly, with derision and contempt in his voice, 'Her boyfriend did it. He rapes her every Monday, Wednesday, and Friday--when she wants it.'" HOLMSTROM & BURGESS, supra note 4, at 40. In a study of rape cases processed in the District of Columbia from 1971 to 1976, Williams found that only 9% of defendants who were cohabiting, or were ex-spouses or ex-boyfriends of putative rape victims, were convicted. Kristen Williams, Few Convictions in Rape Cases: Empirical Evidence Concerning Some Alternative Explanations, 9 J. CRIM. JusT. 29, 36 (1981). On the other hand, only 10% of the defendants who were friends of the victim were convicted and 27% of strangers. Id.

In an experimental study, L'Armand and Pepitone included three relationship conditions in their rape stories: dating with prior consensual sex; dating with no prior sex; and strangers. K. L'Armand & A. Pepitone, Judgements of Rape: A Study of Victim-Rapist Relationship and Victim SexualHistory, 8 PERSONALITY & SOC. PSYCH. BULL. 134, 135 (1982). Subjects estimated victim blame for the rape on a 0-100 scale, giving the victim with prior consensual sex with the defendant an average blame score of 39.3, the victim who had dated the defendant without sex 32.5, and the victim who did not know her attacker 30.1. Id. The sentence imposed varied significantly, with the stranger rapist receiving 13.5 years, the dating without prior sex rapist 8.5 years, and the rapist who had a previous consensual relationship with the victim 5.1 years. Id.

66 An experienced prosecutor, writing about trial preparation for rape cases, discusses the "unpopular victim." William Heiman, Prosecuting Rape Cases: Trial Preparation and Trial Tactic Issues, in PRACTICAL ASPECTS OF RAPE INVESTIGATION 343 (Robert R. Hazelwood & Ann Wolbert Burgess eds., 1987). According to Heiman, the unpopular victim is one,

who, by virtue of her background or lifestyle in general or because of the particular activity in which she was engaged just prior to the rape, can be expected to elicit biased or negative feelings from the average juror. Examples of this include prostitutes, hitchhikers, drug and alcohol abusers, runaways and truants.

Id.

Also, a study conducted by a prominent rape scholar concluded that, "jurors who held conservative notions regarding appropriate behavior for women tended to absolve a defendant of guilt if the victim allegedly violated conservative notions of 'proper' female behavior by drinking or using drugs or by being sexually active outside marriage." Gary D. LaFree et al., Jurors' Responses to Victims' Behavior and Legal Issues in Sexual Assault Trials, 32 Soc. PRoas. 389, 401 (1985).

67 One study compared victim responsibility ratings and sentences for the rapist of a prostitute, divorced woman, married woman, single virgin and single non-virgin. Shirley Feldman-Summers & Karen Lindner, Perceptions of Victims and Defendants in Criminal Assault Cases, 3 CRIM. JusT. & BEHAV. 135 (1976). Results on sentence length were insignificant but victim responsibility ratings did differ in the expected direction. The prostitute was assigned significantly more responsibility for the sexual assault than the married woman and the two single women. Id. at 141. The prostitute was also assigned more responsibility than the divorced woman, although not significantly so. Id. See also LaFree et al., supra note 66, at 401.

68 In another study, mock jurors were given three different rape scenarios to read and evaluate. In one scenario, no mention was made of the victim's prior sexual history; in another scenario she was described as having limited sexual experience; and in a third she was promiscuous. Significant differences were found over both stranger and dating relationships for recommended sentence, seriousness of the crime, and damage to the victim. When the victim's sexual history was not mentioned, the defendant was sentenced to 8.5 years. When she was described as having limited experience, he received 7.3 years. But when the victim was described as having had many previous casual sexual relationships, the rapist received only 4.5 years. L'Armand & Pepitone, supra note 65, at 135.69 Heiman, supra note 66, at 343. 70 GARY LAFREE, RAPE AND CRIMINAL JUSTICE: THE SOCIAL CONSTRUCTION OF SEXUAL ASSAULT 51 (1989) [hereinafter LAFREE, RAPE & CRIMINAL JUSTICE]. 71 See LaFree, Official Reactions, supra note 52, at 586; HOLMSTROM & BURGESS, supra note 4, at 43.

72 See LaFree et al., supra note 66, at 392 (if victim used drugs jurors were less likely to believe in defendant's guilt than if she had not); ADLER, SUpra note 48, at 102-03.

73 Not surprisingly, studies also show that the victim's alcohol consumption affects mock jurors' judgments about her and their recommended sentences for the rapist in a date rape situation. E.g., Deborah Richardson & Jennifer L. Campbell, Alcohol and Rape: The Effect of Alcohol on Attributions of Blame for Rape, 8 PERSONALSaY & Soc. PSYCHOL. BULL. 468 (1982). If the victim of a rape was drunk, she was assigned more responsibility than if sober. Id. at 472.

74 POSNER, supra note 8. Holmstrom and Burgess explain that sexy attire is perceived as an indicator of sexual behavior. By wearing sexy clothing the victim was thought to be "asking for it." HOLMSTROM & BURGESS, supra note 4, at 134-35, 178-79.

75 Cf VACHSS, supra note 13, at 90.

76 See, e.g., ESTRICH, supra note 1, at 16, 60.

77 Id. at 34, 10.

78 See, e.g., id. at 32-38.

79 Se.e ADLER, supra note 48, at 129-30.

80 ESTRICH, supra note 1, at 56.

81 Id. at 102. Estrich states that distrust of women claiming to have been raped by an acquaintance has been incorporated into the law through evidentiary rules such as requiring corroboration and admitting evidence of the victim's norm-violating character. Id. at 2756.

82 Susan Brownmiller traces rape victim-blaming to ancient times. BROWNMILLER, supra note 1, at 19.20, 22.

83 Id. at 311-12.

84 Id. at $11-13.

85 Martha R. Burt, Rape Myths and Acquaintance Rape, in ACQUAINTANCE RAPE, supra note 63, at 3031.

86 Id. at 32.

87 ESTRICH, supra note 1, at 2325, 43, 62-69.

88 ADLER, supra note 48, at 17.

89 See generally CALEB FOOTE & ROBERTJ. LEVY, CRIMINAL LAW: CASES AND MATERIAILS 301-27 (1981).

90 See, e.g., id. at 306.

91 Results of the National Crime Survey show that between 1973 and 1990, the percentage of all criminal victimizations reported to the police ranged from 32% to 38%. MARTIN S. GREENBERG &: R. BARRY RUBACK, AFTER THE CRIME: VICTIM DECISION MAKING 7-8 (1992). The reporting rates vary by type of crime: 75% of motor vehicle thefts were reported in 1990, compared to 54% of rapes and 29% of personal thefts. Id. See a/so Jim Galvin & Kenneth Polk, Attrition in Case Processing: Is Rape Unique?, 20J. REs. CreME DELINQ. 126, 138 (1983) (finding that the percentage of all victimizations in the U.S. reported to the police in 1980 ranged from 42% for rape to 57% for robbery). See infra notes 169-88 and accompanying text (discussing the reporting of sexual assaults).

92 In 1980, the percentage of reported offenses resulting in arrest ranged from 14% for burglary to 72% for homicide. Galvin & Polk, supra note 91, at 138. Attrition continues as cases progress through the system. Thus, in California, Galvin and Polk found that only 17% of those arrested for rape eventually received an institutional sentence, compared with 3% for assault and 38% for homicide. Id. at 136. See also Martha A. Myers & Gary D. LaFree, Sexual Assault and Its Prosecution: A Comparison With Other Crimes, 73 J. CRIM. L. & CRIMINOLOGY 1282, 1288 (1982) (45.6% of sexual assaults proceeding to trial result in acquittal versus 31.8% for other violent crimes). Analyzing data from several previous studies, Frazier and Haney concluded that 12% of reported rapes result in conviction. Patricia A. Frazier & Beth Haney, Sexual Assault Cases in the Legal System: Police, Prosecutor and Victim Perspectives, 20 LAw & HUM. BEHAV. 607, 617 (1996).

93 The police may "unfound" a complaint for any of several reasons. See generally infra notes 225-346 and accompanying text (discussing the police founding decision in rape cases).

94 See HANS ZEISEL, THE LIMITS OF LAW ENFORCEMENT 26 (1982).

95 One compilation of arrest rate statistics indicates that, depending on the jurisdiction, 40% to 52% of all founded rape complaints led to an arrest. Kim Lane Scheppele, The ReVision of Rape Law, 54 U. Cm. L. REV. 1095, 1097 (1987) (book review). Likewise, Frazier and Haney found that suspects in a midwestern jurisdiction were identified in only 48% of rape cases reported to the police in 1991; only 47% of those identified were subsequently referred to the prosecutor. Frazier & Haney, supra note 92, at 617.

96 Galvin and Polk found that California prosecutors filed charges against 52% of suspects arrested for rape, compared to 66% of homicide suspects and 4% of those arrested for burglary. Galvin & Polk, supra note 92, at 141. A more recent study found that onequarter of rape cases referred to prosecutors in a midwestern jurisdiction were not charged. Frazier & Haney, supra note 92, at 617-18. See generally WAYNE R. LAFAVE & AUSTIN W. SCOTT, JR., HANDBOOK ON CRIMINAL LAW (Sections) 4, 18 (1986) (describing the charging decision).

97 Acquittals have historically been common for many major crimes. See KALVEN & ZEISEL, supra note 52, at 42 (jury acquittal rates for major crimes). For a discussion of jury leniency in rape cases, see infra text accompanying notes 398426. Although we lack empirical studies, some anecdotal evidence indicates that juries in acquaintance rape cases are now less prone to acquit than in previous cases. See infra text accompanying notes 44345.

98 For example, in Minnesota, the executed sentence for aggravated robbery committed by a first-time offender would be 48 months in prison. MINN. 8TAT. ANN. (Sections) 244.101, 244 App. IV (West Supp. 1995). Provided the prisoner commits no disciplinary offenses, he is released after serving 32 months--two-thirds of the executed sentence. MINN. STAT. ANN. (Section) 244.101 (West Supp. 1995). The remaining 16 months are served on supervised release. Id. Nationwide estimates find that on average prisoners only serve 33% of their sentences. PATroCK A. LANGAN & RICHARD SOLAm, U.S. DEP'T OF JUSTICE NATIONAL JubrCIAL REPORTING PROGRAM, 1990 tbl., at 1.5 (1993).

99 ZEISEL, supra note 94, at 17-25.

100 Id. at 18.

101 Id. at 26.

102 Joan McGregor, Introduction to Symposium on Philosophical Issues in Rape Law, 11 LAW & PHIL. 1, 2 (1992).

103 The Vera Institute estimated that only 15% of those arrested for felonies in New York were convicted of a felony. Indeed, only 2.3% of the cases went to trial and in the vast majority of those, the conviction was for a misdemeanor or a lesser felony. The case was almost always either dismissed (43%) or resulted in a guilty plea (55%). Of the guilty pleas, only 5% were for the felony charged; another 20% were for lesser felonies. About 75% of all guilty pleas for felony charges reduced the offense to a misdemeanor. VERA INSTITUTE OF JUSTICE, FELONY ARRESTS: THEIR PROSECUTION AND DISPOSITION IN NEW YORK CITY'S COURTS 6 (1977) [hereinafter VEgA INSTITUTE STUDY]. Only murder failed to conform to this pattern of attrition. Id. at 52.

104 See generally HOLMSTROM & BURGESS, supra note 4, at 152-54.

105 See, e.g., ESTRICH, supra note 1, at 15-26; Deborah Gartzke Goolsby, Comment, Using Mediation in Cases of Simple Rape, 47 WASH. & LEE L. REv. 1183, 1184 (1990) (stating that the criminal justice system accepts the defendant's claim that the victim consented and now either regrets the encounter or wishes to retaliate against the defendant); Lynne Henderson, Getting to Know: Honoring Women in Law and in Fact, 2 TEX. J. WOMEN & L. 41, 41 (1993) (claiming that victim-blaming in the criminal justice system makes a successful prosecution in non-aggravated acquaintance rape cases improbable).

106 See, e.g., supra note 105.

107 BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, VIOLENCE AGAINST WOMEN: ESTIMATES FROM THE REDESIGNED SURVEY 1 (1995) [hereinafter VIOLENCE AGAINST WOMEN].

108 FEDERAL BUREAU OF INVESTIGATION, U.S. DEP'T OF JUSTICE, CRIME IN THE UNITED STATES, 1994 UNIFORM CRIME REPORTS 376 (1995) [hereinafter 1994 UCR].

109 STAFF OF SENATE COMM. ON THE JUDICIARY, 103D CONG., lST SENN., THE RESPONSE TO RAPE: DETOURS ON THE ROaD TO EQUAL JUSTmE iii (Comm. Print 1993) [hereinafter THE RESPONSE TO RAPE].

110 For a recent review of studies of rape case processing, see Ruth Triplett & Susan L. Miller, Case Processing in the Harris County, Texas Criminal Justice System: A Comparison Across Crime Types, 22 J. CRIM. JusT. 13, 13-15 (1994).

111 Robin, supra note 1, at 142.

112 FELDMAN-SUMMERS 8C LINDER, supra note 67, at 135-36.

l13 THE RESPONSE TO RAPE, supra note 109, at 13.

114 For a discussion of false rape reports, see infra notes 603-16 and accompanying text.

115 See infra notes 331-37 and accompanying text.

116 For a discussion of the difficult proof problems in rape cases, see infra text accompanying notes 720-71.

117 See infra notes 337-40 and accompanying text.

118 See infra notes 307, 359-60 and accompanying text.

119 See ELIZABETH F. LOFTUS, EYEWITNESS TESTIMONY 142-44 (1979) (stating that unconscious transference can cause a witness to identify a suspect because the witness saw the suspect, or a photo of the suspect, in a context other than the crime); Stephanie J. Platz & Harmon M. Hosch, Cross-Racial/Ethnic Eyewitness Identification: A Field Study, 18 J. APPLIED Soc. PSYCHOL. 972, 981-83 (1988) (discussing the inaccuracy of cross-racial identifications); Elizabeth F. Loftus et al., Some Facts About "Weapon Focus, "11 LAW & HUM. BEHAV. 55, 61-62 (1987) (stating that "weapon focus" often interferes with identification capacity). See also Fredric D. Woocher, Note, Did Your Eyes Deceive You?: Expert Psychological Testimony on the Unreliability of Eyewitness Identification, 29 STAN. L. REv. 969, 978-86 (1977) (describing variables that sometimes lead to misidentification).

120 Galvin & Polk, supra note 91, at 146. The researchers found that in 1980 in California, 43% of reported rapes were cleared by arrests, compared to 63% of homicides, 56% of assaults, 23% of robberies, and 13% of burglaries. Id. at 141. Charges were filed in 52% of rape arrests, compared to 66% of homicides, 54% of robberies, 43% of burglaries, and 25% of assaults. Id. Similar patterns were revealed in conviction and sentencing, leading the authors to conclude that "[r]ape has no unique pattern of attrition." Id. at 152.

121 Id. at 126.

122 Myers & LaFree, supra note 92, at 1288.

123 Id. at 1300.

124 Id. at 1301.

125 BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, THE PROSECUTION OF FELONY ARRESTS, 1988 15-23 (1992) [hereinafter PFA]. For example, in 1988 in San Diego, 42% of rape cases and 41% of assault cases were dismissed; guilty verdicts were obtained in 55% of rape cases and 50% of assault cases. Id. at 21. In the District of Columbia, 61% of rape cases and 71% of assault cases were dismissed, while guilty verdicts were obtained in 29% of rape and 25% of assault cases. Id. at 23.

126 Id.

127 Darrell J. Steffensmeier, The Uniqueness of Rape?: Disposition and Sentencing Outcomes of Rape in Comparison to Other Major Felonies, 72 Soc. SCI. REs. 192, 193 (1988). For example, once a case was filed as a felony, rape was found to have a 73% conviction rate, compared to 76% for homicide, 67% for assault and 86% for burglary. Id.

128 For example, of 109 reported rape cases in Boston, 26 were aborted by the victim's request or her failure to press charges or her disappearance. This was the second largest cause of pre-trial attrition, after the victim's inability to identify her assailant. HOLMSTROM BURGESS, supra note 4, at 150-51. See also C. HURSCH, THE TROUBLE WITH RAPE 110-14 (1977) (finding 38% of rape victims in Denver withdrew their charges after reporting the crime but prior to prosecution).

129 Steffensmeier, supra note 127, at 193. The author reasons that the phenomenon does not apply to homicide "because living witnesses are unlikely to be available." Id.

130 See, e.g., ZEISEL, SUpra note 94, at 26, 111 (concluding that the withdrawal of victims from prosecution is the primary cause of felony attrition); MICHAEL R. GOTTFREDSON & DON M. GOTTFREDSON, DeCISION-MAKING IN CRIMINAL JUSTICE: TOWARD THE RATIONAL EXERCISE OF DISCRETION 154 (1980) (finding that witness problems were the most significant reason given by prosecutors for dismissing charges for violent crimes).

131 Steffensmeier, supra note 127, at 193.

132 See infra notes 251-53, 351-53 and accompanying text.

133 Robert A. Weninger, Factors Affecting the Prosecution of Rape: A Case Study of Travis County, Texas, 64 VA. L. Rev. 357, 377-78 (1978).

134 Susan Meyers Chandler & Martha Torney, The Decisions and Processing of Rape Victims Throughout the Criminal Justice System, 4 CAL SOCIOLOGIST 155 (1981).

135 Id.

136 Id.

137 See Steffensmeier, supra note 127, at 195.

138 See Alan J. Lizotte, The Uniqueness of Rape: Reporting Assaultire Violence to the Police, 31 CreMe & DELINQ. 169 (1985) (analysis of NCVS results finds that thecloser the relationship between a rape victim and offender, the less likely the victim is to report the rape).

139 For an explanation of the NCVS, see infra text accompanying note 162.

140 Lizotte, supra note 138, at 181.

141 Id. at 185.

142 Id. at 185-86.

143 See infra text accompanying notes 171-86.

144 GOTtgranDSON & GOTTFREDSON, supra note 130, at 330-33.

145 Id. at 331.

146 For a thorough review of studies comparing case processing for stranger and nonstranger offenses, see Lenore M.J. Simon, Legal Treatment of the Victim-Offender Relationship in Crimes of Violence, 11 J. INTERPERSONAL VIOLENCE 94 (1996).

147 Id. at 94-95. See also VERA INSTITUTE STUDY, supra note 103, at 15 ("Expression of anger results in the commission of technical felonies, yet defense attorneys, judges, and prosecutors recognize that in many cases conviction and prison sentences are inappropriate responses").

148 Celesta A. Albonetti, Prosecutorial Discretion: The Effects of Uncertainty, 21 L. & SOC'Y REv. 291,300 (1987). Albonetti argues that prosecutors prefer to avoid uncertainty, which translates into a bias against prosecuting non-stranger offenses.

149 ZEISEL, supra note 94, at 114-15.

150 ANITA D. TIMROTS, BUREAU OF JUSTICE STATISTICS, U.S. DEP'T. OF JUSTICE, VIOLENT CRIME BY STAGERS AND NONSTRANGERS 3 (1987).

151 BUREAU OF JUSTICE STATISTICS, U.S. DEP'T. OF JUSTICE, CRIMINAL VICTIMIZATION IN THE UNITED STATES 1993, 35 (1994) [hereinafter 1993 NCVS].

152 VIOLENCE AGAINST WOMEN, SUpra note 107, at 33.

153 Ronet Bachman & Raymond Paternoster, A Contemporary Look at the Effects of Rape Law Reform: How Far Have We Really Come? 84J. CraM. L. & CRIMINOLOGY 554 (1993). The authors used data from the National Crime Victimization Survey (NCVS), the Uniform Crime Reports (UCR), National Prisoners Statistics Program (NPSP) and the National Corrections Reporting Program (NCRP). Id. at 560451.

154 Id. at 572 (comparing NCVS data on victims' descriptions of their relationship with their perpetrator with imprisoned offenders' descriptions of their relationship with their victim); id. at 562-63.

l55 From 1979-86, there was a 20% difference between the expected and actual incarceration rates for acquaintance rapists, compared to a 7% difference for robbery and a 14% difference for assault. Id. at 572.

156 The authors estimated that in 1991, 58% of those incarcerated for rape should have been acquaintance rapists, but only 40% actually were. Id. In the same period, there was only a 1% difference between expected and actual incarceration rates of non-stranger perpetrators of robbery and assault. Id. at 573.

157 Many authors, ignoring the limitations of the UCR, cite it as authority for the proposition that the incidence of rape (or crime in general) has risen alarmingly. E.g., LARRY BARON & MURRAY A. STRAUS, FOUR THEORIES OF RAPE IN AMERICAN SOCIETY 3 (1989) (stating that "official statistics" show 440% rise in rape incidence over 28 years). Of course, even if the incidence of rape is declining, it should be punished appropriately, so the question has little if any bearing on law reform issues.

158 Controversy also surrounds statistics on the prevalence of rape, defined as the proportion of women who have been raped during their lives. Compare Gilbert, supra note 56, at 54-55, with Mary P. Koss, Defending Date Rape, 7J. INTERPERSONAL VIOLENCE 122 (1992). See generally Mary P. Koss, Detecting the Scope of Rape: A Review of Prevalence Research Methods, 8 J. INTERPESONAL VIOLENCE 198 (1993).

159 See generally Gary F. Jenson & Mary Altani Karpos, Managing Rape: Exploratory Research on the Behavior of Rape Statistics, 31 CRIMINOLOGY 363, 366 fig. 1 (1993).

160 Reports that the police disbelieve (and therefore "unfound") are tabulated separately. 1994 UCR, supra note 108, at 376.

161 Id. at 23. The number of reported rapes per 100,000 inhabitants has recently declined somewhat from 42.8 in 1992 to 39.2 in 1994. FEDERAL BUREAU OF INVESTIGATION, U.S. DEP'T. OF JUSTICE, CRIME IN THE UNITED STATES, 1992 UNIFORM CRIME REPORTS 58 tbl. 1 (1993) [hereinafter 1992 UCR]; 1994 UCR, supra note 108, at 23.

162 BUREAU OF JUSTICE STATISTICS, U.S. DEP'T. OF JUSTICE, CRIMINAL VICTIMIZATION IN THE UNITED STATES, 1991, iii (1992) [hereinafter 1991 NCVS].

163 Id. at 6 tbl.3.

164 No scholar contends that the number of falsely reported "rapes" exceeds the number of unreported rapes. But, deliberately false reports are presumably more common in the UCR than in the NCVS, since a false response to a survey serves little or no purpose.

165 In Michigan, where reported rapes rose sharply after rape law reforms in the mid-1970s, criminal justice officials attributed the rise to "enlightened attitudes that encourage victims to come forward," rather than to an increase in the incidence of rape. MARSH ET At.., supra note 25, at 42-43. Cf Jensen & Karpos, supra note 159, at 368. However, demographic changes sometimes make the NCVS figures somewhat misleading. As the proportion of elderly in the population increases, the violent crime rate can be expected to decline, because violent criminals and their victims are both disproportionately young. Consequently, the incidence of a violent crime may decline during a period in which the likelihood that any given youth will commit a violent crime has risen substantially. DAVID T. LYKKEN, THE ANTISOCIAL PERSONALITIES 4 (1995).

166 See generally Jensen & Karpos, supra note 159, at 371-83. Jensen and Karpos believe that the UCR's upward trend was mainly due to improvements in police management of rape cases such as increased use of female and civilian investigators. Id. Other scholars have noted that, ironically, improvements in law enforcement sometimes lead to increased reporting of crime and thus to spurious "paper increases" in the incidence of crime. RICHARD S. FRASE & FRANKLIN E. ZIMRING, THE CRIMINAL JUSTICE SYSTEM: MATERIALS ON THE ADMINISTRATION AND REFORM OF THE CRIMINAL LAW 51 (1980). Irrespective of improvements in police techniques and personnel, victims may feel more assertive and less isolated-hence more willing to report--because of the feminist anti-rape campaign. Cf Ronet Bachman, Predicting the Reporting of Rape Victimizations: Have Rape Reforms Made a Difference? 20 CRIM. JUST. & BEHAV. 254, 265 (1993) (victims of acquaintance rape found to be as willing as victims of stranger rape to report the crime, probably because of legal reforms and "media campaigns").

167 Jensen & Karpos, supra note 159, at 368. For criticisms of the NCVS, see generally Mary P. Koss, The Measurement of Rape Victimization in Crime Surveys, 23 GRIM. JusT. & BEHAr. 55 (1996); Mary P. Koss, The Underdetection of Rape: Methodological Choices Influence Incidence Estimates, 48J. Soc. ISSUES 61 (1992); Mary P. Koss et al., Criminal Victimization Among Primary Care Medical Patients: Prevalence, Incidence, and Physician Usage, 9 BEHAV. ScI.& L. 85 (1991). Recent reforms of the NCVS methodology, prompted by criticism of the survey's ability to accurately gauge the incidence of certain crimes, have increased the number of rape reports garnered by the survey. BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, CRIMINAL VICTIMIZATION IN THE UNITED STATES, 1993, at 2 (1994) [hereinafter 1993 NCVS]. The new NCVS methodology allows for more direct questioning about sexual crimes and covers coerced and threatened sexual activity in addition to rape and attempted rape. Id. at 150. The new methodology has "elicited information on about 3 to 4 times as many sexual crime victimizations" as appeared prior to 1993. I& at 2. Despite this sudden increase, there is no reason to expect that the rate of reported rape measured by the NCVS will not continue to fall in future years.

168 LYKKEN, supra note 165, at 3.

169 See infra text accompanying notes 331-37. A study in Denver found that 38% of rape victims withdrew their charges soon after reporting the crime. HURSCH, supra note 128, at 110-14.

170 McCAroLL ET AL., Supra note 52, at 82 ("An unreported to reported ratio of 10:1 is often quoted, but ratios as high as 100:1 are not unheard of."); Scheppele, supra note 95, at 1096 ("Only slightly over half of all rapes and rape attempts by strangers are reported to the police.., and those figures only count the discrepancies between rapes that women report to the police and those that women report to survey interviewers. Rapes reported to neither simply aren't included in the figures."). Estimates of the reporting rate for acquaintance rapes vary from 5% to nearly 50%. Id. See generally CASSIA SPOHN & JULIE HORNE'/, RAPE LAW REFORM: A GRASSROOTS REVOLUTION AND ITS IMPACT 18 (1992). The NGVS states that 50-60% of rape victims report the crime to the police. BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, CRIMINAL VICTIMIZATION IN THE UNITED STATES, 1992, at 7 tbl.5 (1993). Other studies have found lower reporting rates. See DIANA H. RUSSELL, SEXUAL EXPLOITATION: RAPE, CHILD SEXUAL ABUSE, AND WORKPLACE HARASSMENT 31 (1984) (9.5%o); RAPE IN AMERICA, supra note 58, at 3 (16%); Dean G. Kilpatrick e t al., Mental Health Correlates of Criminal Victimization, 53 J. CONSULTIma & CLINICAL PSYCHOL. 866 (1985), cited in Koss, Detecting the Scope of Rape, supra note 160, at 202 (29%); Crystal S. Mills & BarbaraJ. Granoff, Date and Acquaintance Rape Among a Sample of College Students, 37 Soc. WORK 504, 506 (1992) (finding that of 20 student rape victims, none told the police and only 15% told someone).

According to federal victimization surveys, 64.6% of rape victims (and 24.1% of attempted rape victims) report the crime. The former figure is higher than the corresponding figures for aggravated assault (with an injury) (58.1%) or simple assault completed with an injury (51.6%), and for Crimes of violence in general (47.9%). BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, A NATIONAL CRIME SURVEY REPORT 80 (1988). The same source indicates that 38% of all crimes were reported to the police and that the reporting 0rate increased for crimes in general, from 32% in 1973 to 38% in 1990; for crimes of violence in general from 46% to 48%; and for completed rape, from 49% to 54%. Id. These figures, while perhaps accurately depicting trends, should be interpreted cautiously in light of the methodological weaknesses of the NCVS. See supra notes 169-71 and accompanying text.

In a survey of randomly-selected university women, Koss found that, of the 38% who reported sexual victimizations that met the legal definition of rape or attempted rape, "only 4% of these women had reported their sexual assault to the police." Koss, The Hidden Rape Victim, supra note 56, at 206. In Koss's study, unlike many victimization surveys, the questions were phrased in such a way as to detect women who had in fact been raped but who did not label the experience as "rape." Id. See also Mary P. Koss, Hidden Rape: Sexual Aggression and Victimization in a National Sample of Students in Higher Education, in 2 RAPE & SEXUAL ASSAULT 3 (A.W. Burgess ed., 1988) (8%); Mary P. Koss et al., The Scope of Rape: Incidence and Prevalence of Sexual Aggrression and Victimization in a National Sample of Higher Education Students, 55J. CONSULTING & CLINICAL PSYCHOL. 162 (1987) [hereinafter Koss et al., The Scope of Rape] (finding prevalence rates for rape victimization 10 to 15 times higher than estimates from the Bureau of Justice Statistics and that virtually none of these victims had been involved in the justice system). But see Gilbert, supra note 56.

171 One victim survey found that the most important reason why 33% of rape victims did not report the crime to the police was that they regarded it as a private or personal matter or took care of it themselves informally. An additional 20% indicated that the most important reason for non-reporfng was because they believed the police would not help, for reasons including disinterest, ineffectiveness, or bias. Another 6% of victims felt that the police could not do anything, either because of the victim's inability to identify the rapist, or the police's inability to find him, or a lack of proof. Finally, 13% did not report because they feared reprisal by the victim or someone else. Bachman, supra note 166, at 263.

172 See HOLMSTROM & BURGESS, supra note 4, at 58.

173 "Victims of rape frequently hide their assault, even from significant others. For example, 42% of women college student rape victims indicated on a self-report survey that they had never told anyone at all about the incident." Koss et al., The Scope of Rape, supra note 170, at 162.

The Victim Center's data revealed that rape victims were at least somewhat or extremely concerned about: (1) her family knowing (71%); (2) people thinking it was her fault or that she was responsible (69%); (3) people outside her family knowing (68%); (4) her name being made public by the media (50%). Id. Only 17% had a medical exam following the assault, and of these 40% did not do so within 24 hours. RAPE IN AMERICA, supra note 56, at 4.

Even in cases of stranger rape some victims are reluctant to report the crime because of embarrassment. Linda Fairstein recounts the case of an 82 year-old woman who was embarrassed that she had been raped by her 30 year-old window washer. FARSTEIN, supra note 13, at 59. Another possible factor is the victim's pessimism about the likelihood of a conviction: "In no other category of crime does the victim approach the criminal justice system with lower expectations of a successful resolution than in the area of sex offenses." Id. at 67. See als0 VACHSS, supra note 13, at 62 (victim felt guilty about pressing charges against her father). For a thoughtful discussion of the role of shame in non-reporting of rape, see Berger, supra note 1, at 23.

174 Interviews at Beth Israel's Rape Crisis Center indicated that 7 of 12 victims feared retaliation. D.C. SILVERMAN & S.M. Bowre, FROM OUT OF THE SHADOWS: DISCLOSING THE PATTERNS AND CIRCUMSTANCES OF RAPE VICTIMIZATION (1986), cited in Gail Stetekee & Anne H. Austin, Rape Victims and the Justice System: Utilization and Impact, 63 Soc. SERF. REv. 285, 301 n.23 (1989). Another study found that 75% Of minority group rape victims listed fear of retaliation as their reason for not reporting the crime. MichaelJ. Hindelang & BruceJ. Davis, Forcible Rape in the United States: A Statistical Profile, in FORCIBLE RAPE: THE CRIME, THE VICTIM AND THE OFFENDER 87, 97 (D. Chappell et al. eds., 1977) [hereinafter FORCIBLE RAPE]. In one study, over one-quarter of rape victims stated a "fear of reprisal" as inhibiting their decision to report the rape to the police. Joanne Belknap, The Sexual Victimization of Unmarried Women by Nonrelative Acquaintances, in VIOLENCE IN DATING RELATIONSHIPS 205, 214 (Maureen A. Pirog-Good & Jan E. Stets eds., 1989). The greatest inhibitor of reporting rapes was, however, the view that the rape was a "private or personal matter." Id.

175 Some victims try to solve this dilemma by lying to the police, a practice that creates problems for prosecutors. See FAIRSTEIN, supra note 13, at 205-16.

176 See, e.g., RUSSELL, supra note 170, at 97 tbl.3.7; Carol A. Skelton & Barry R. Burkhart, Sexual Assault: Determinants of Victim Disclosure, 7 CRIM. JusT. & BEhAr. 229, 229-36 (1980); Linda S. Williams, The Classic Rape: When Do Victims Report?, 31 Soc. PRODS. 459, 459-67 (1984); see also Barry Burkhart, Acquaintance Rape Statistics and Prevention, Paper Presented at the Rape Prevention on College Campuses Conference (Dec. 1983) (stating that fewer than 1% of acquaintance rape victims report the crime to the police). For a contrary finding, see Belknap, supra note 174, at 212. Accord Bachman, supra note 166, at 265 (reporting not significantly influenced by victim/offender relationship; result attributed to recent feminist publicity). For a review of studies of demographic and other variables that affect the likelihood that a woman will report a rape, see Steketee & Austin, supra note 174. Since many acquaintance rapes are by the victim's lover, and victims often do not label these as rapes, and often choose to continue the relationship, see infra notes 17779, we think it likely that acquaintance rapes continue to be underreported relative to stranger rapes, although the gap may have narrowed somewhat as a result of anti-rape publicity. Cf. R. Barry Ruback, Comment on Bachman (1993): The Victim-Offender Relationship Does Affect Victims' Decisions to Report Sexual Assaults, 20 CRIM. JusT. & BEHAV. 271 (1993).

177 Calhoun & Townsley, supra note 63, at 66. In one study, less than 20% of a sample of adult women who read a description of forced sex on a date labeled it as rape. Susan H. Klemmack & David L. Klemmack, The Social Definition of Rape; in SEXUAL ASSAULT 135, 142 (Marcia J. Walker & Stanley L. Brodsky eds., 1976). See generally Koss, The Hidden Rape Victim, supra note 56. In her study of the sexual experiences of 2016 randomly-selected university women, Koss found that, of the 12.7% whose answers to preliminary questions indicated that they had in fact been raped, only 57% labeled the encounter as rape in response to a final question whether they had been raped. Id. at 197. "IN]one of the unacknowledged rape victims reported their experience to the police, a rape crisis center, or a hospital emergency room," and "more than half of them did not reveal their assault to anyone." Id. The victim's failure to perceive her experience as rape may be due to her romantic involvement with the perpetrator. Id. In the acknowledged rape victim group, only "8% of the victims reported their experience to the police, while 13% went to a rape crisis center or hospital emergency room." Id.

Another study found that the probability that the victim would report the rape varied directly with the amount of danger to the victim during the rape, social distance between the offender and the victim, and the perceived probability of "compensation from significant others" for reporting. Linda C. Smith & L.D. Nelson, Predictors of Rape 'Victimization Reportage, cited in HOLMSTROM & BURGESS, supra note 4, at 34, 61. Non-reporting was apt to occur among victims whose offenders were acquaintances of approximately the same age and social class, and who did not use a weapon or commit another crime at the same time. Id. See generally BOURQUE, supra note 25.

178 JOHN M. MACDONALD. B, RAPE: CONTROVERSIAL ISSUES 5 (1995). Kanin studied 71 white, undergraduate students who admitted that they had raped a woman on a date. Only six were certain the rape had been reported to the police, and in every case the victim refused to prosecute. Seven other cases were not included because the victim-offender relationship continued for at least four months after the rape. E.J. Kanin, Date Rape, 9 VICTIMOLOGY 95 (1984). Other studies have found that remarkably high proportions of the rape victims who knew their assailant continued their previous relationship with him after the rape. Mills & Granoff, supra note 170, at 506 (45%); R. Lance Shotland, A Theory of the Causes of Courtship Rape, 48 J. Soc. IssuEs 127, 128 (1992) (39% of rape victims date their attacker after the incident).

179 Stetekee & Austin, supra note 174, at 291.

180 Id.

181 See generally BOURQUE, supra note 25, at 45 (victims with a prior sexual relationship with the offender are the least likely to report a rape); Mills & Granoff, supra note 170, at 506; Shotland, supra note 178, at 128.

182 Shotland, supra note 178, at 128.

183 See, e,g., Mary P. Koss, The Hidden Rape Victim: Personality, Attitudinal, and Situational Characteristics, 9 PSYCHOL. WOMEN Q. 193 (1985). Koss collected data from a sample of 195 women who responded to a "Sexual Experiences Survey." Id. at 195. Those subjects who reported having ever been forced to submit to acts of intercourse against their will were split up according to whether or not they viewed themselves as "rape victims." Id. at 196. Koss reported that 31% of the acknowledged rape victims and 76% of the unacknowledged victims were romantically involved with their attackers prior to the rape. Id. at 197.

184 They excluded victims under 18 and those who did not personally report the rape. Smith & Nelson, supra note 177, at 34.

185 Id. Other investigators have also found that victims are more likely to report to the police if they experienced severe physical force or were injured during the crime. See, e.g., Bachman, supra note 166, at 266.

186 Id.

187 SPOHN & HORNEY, supra note 170, at 18; Morris, supra note 20, at 158; Deborah P. Kelly, Delivering Legal Services to Victims: An Evaluation and Prescription, 9 JusT. SYS. J. 62, 64 n.3 (1984); ESTRICH, supra note 1, at 14-15 (failure of criminal justice system to treat rape complaints as legitimate is a major reason for non-reporting by victims). Cf. HOLMSTROM & BURGESS, supra note 4, at 39 (low reporting rate blamed partly on "horror stories" and television portrayals of police mistreatment of victims); MARSH ET AL., supra note 25, at 1 (victims often discouraged from reporting and prosecuting cases because of "humiliating and degrading treatment by hospital staff, police officers, prosecutors, defense attorneys, and judges").

188 THE RESPONSE TO RAPE, supra note 109, at 38 ("It is the fear of what a jury will think that drives [rape] survivors not to report, police to refuse to arrest in 'futile' cases, and prosecutors to dismiss prosecutions as 'unwinnable.' "). Berger, supra note 1, at 24 ("The law itself, however, is widely regarded as a prime deterrent to would-be complainants in cases of rape.").

189 Leigh Bienen, Rape Reform Legislation in the United States, 8 VICTIMOLOGY 139, 139 (1983) [hereinafter Bienen, Rape Reform]. Between 1975 and 1980,

[t]raditionally conservative state legislatures enacted changes in state criminal law which radically redefined the offense of rape and introduced new state statutes which either limited the admissibility of evidence concerning the prior sexual conduct of the victim, or required ajudicial screening before such evidence could be placed before a jury. Often rape reform legislation redefined and recharacterized all sex offenses, including rape, sodomy, incest, statutory rape, assault with intent to commit rape and the sexual molestation and abuse of children. Mandatory minimum penalties were introduced in some states; other states incorporated an across-the-board reduction of penalties while simultaneously introducing graded, or "staircased," offenses to replace the single category of rape. The acts constituting the most serious sex offense were recharacterized, and in more than half of the states new offenses based upon sexual contact were defined. These offenses had no analog in prior laws against rape. Changes in the legal definition of the crime of rape, with their accompanying statutes directed at limiting the admissibility of evidence regarding the prior sexual history of [the] victim at trial, were enacted in response to vigorous, nationally coordinated lobbying by feminists. Id.

190 See SPOHN & HORNEY, supra note 170, at 18; Bienen, supra note 49, at 179.

191 Bachman & Paternoster, supra note 153, at 565436. These authors relied on the NCVS, which tabulates not only victimization data, but also the proportions of victims who reported the crime to the police, ld. at 565. "Looking at the original data points, there was a 28% increase in rape victims who reported to the police from 1980 to 1990," but when these data are smoothed the increase is only 10%. ld. at 565-66.

192 SPOHN & HORNEY, supra note 170, at 35-36. Cf Bienen, Rape Reform, supra note 189, at 14749 (finding "minimal" practical effects from New Jersey's repeal of the spousal exemption and decriminalization of incest).

193 SPOHN & HORNEY supra note 170, at 46.

194 E.g., MARSH ET AL., supra note 25, at 2. The Michigan reforms took effect in April, 1975. Id.

195 SPOHN & HORNEY, supra note 170, at 46.

196 Id.

197 Id.

198 Id.

199 ld.

200 Id.

20l Id. at 81.

202 Id. at 92-98.

203 Id. at 96-97.

204 Id. at 101. Since reports of other violent crimes did not show a similar upward trend, Spohn and Horney decided that the rising volume of rape reports probably indicated a rising propensity to report the crime rather than a rising incidence of rape. Id. at 86-87.

205 Id. at 86.

206 Id. at 1014)2.

207 Id. at 96.

208 Id. at 101-02.

209 MARSH ET AL., supra note 25, at 26. A similar phenomenon occurred in Canada and California. Julian V. Roberts & Robert J. Gebotys, Reforming Rape Laws: Effects of Legislative Change in Canada, 16 LAw & HUM. BEHAV. 555, 561 (1992); James L. Lebeau, Statute Revision and the Reporting of Rape, 72 SOC. SCI. REs. 201 (1988). Over the five-year period prior to Canada's 1983 reforms, the overall reporting rate of all three levels of sexual assault increased 22%, while reports of nonsexual assault increased by 20%. Roberts & Gebotys, supra, at 561. There was a significant increase in reporting when the legislation was introduced, and a 110% increase--more than double the increase of reports of nonsexual assault-in the five years after enactment of reforms. Id. at 561-63. The authors believe that the coincidence of the 1983 legislation with a relatively sharp increase in reporting rules out an explanation based on evolving attitudes unrelated to the reforms; they also reject several other explanations, concluding that "it seems likely that the increase is due to a change in victims' attitudes, brought about by the reform legislation and the publicity surrounding its passage." Id. at 568.

This conclusion may not be relevant in the United States, where nearly all rape law reform occurs at the state level, yet most media publicity is national and consequently unrelated to passage of specific statutes. For this reason, one would expect the correlation between reporting increases and enactment of particular reforms to be weaker in the United States.

210 MARSH ET AL., Supra note 25, at 41-43. "Marginal" means, of course, a case in which a conviction is less likely.

211 Id. at ix.

212 Id. at ix, 25, 36. Other scholars have identified several factors, other than the law reforms, that affect reporting of rape: (1) preconceived ideas about how the criminal justice system will respond to the report; (2) situational characteristics surrounding the rape; (3) trauma suffered by the victim; and (4) advice offered by family, friends, lovers, and others. They conclude that "if conditions in any of these stages favor non-report, then the victim probably will not report the crime to the police." Richard L. Dukes & Christine L. Mattley, Predicting Rape Victim Reportage, 62 Soc.& SOC. RES. 63, 68 (1977).

213 SPOON & HoRley, supra note 170, at 172.

214 Id.

215 Id.

216 See id. at 18-20.

217 See generally Berger et al., supra note 32.

218 See, e.g., Carol Bohmer, Acquaintance Rape and the Law, in ACQUAINTANCE RAPE, supra note 63, at 317, 317-26; Henderson, supra note 105, at 87.

219 One survey of rape victims found that 50% said that they would be "a lot more likely" to report if the law prohibited the media from obtaining and disclosing their names and addresses, with an additional 16% saying that they would be "somewhat more likely" to report with these safeguards. RAvE IN AMERICA, supra note 56, at 6.

220 See HOLMSTROM & BURGESS, supra note 4, at 58.

221 See id. Which is not to say that the likelihood of a conviction, as reported to the victim, has no effect. See Steketee & Austin, supra note 174, at 293.

222 See, e.g., VACHSS, supra note 13, at 203 (discussing a victim of a serial rapist who had married and had a child before the trial and was reluctant to prosecute because she wanted to put the rape behind her).

223 See Steketee & Austin, supra note 174, at 293-94. Perhaps victims in some cities believe that police will do more to solve their crimes. Rapid City, South Dakota, a prairie community of 83,000 people, had the highest rape rate in the nation. The police chief attributed this phenomenon to a "small town" effect: women in Rapid City, he claimed, tend to believe that the police will do more to solve their crimes and so they are more likely to report the rape, a judgment in which the manager of a shelter for battered women concurred. Joe Hallinan, No Safety in Smaller Towns, Cities Have No Patent on Murder and Rape, CHI. SUN-TIMES, Feb. 20, 1994, at 37. That might explain why every one of the ten cities with the highest murder rates were in the South, while all but one of those with the highest rates of reported rapes were in the North--half of them in Michigan, the state with the most ambitious rape law reforms. Id. Nearly all murders are reported, but rape is notoriously under-reported. As a consequence the reported murder rate is likely to be highly accurate, while the "rape rate" may reveal more about victims' attitudes than about the frequency of the crime.

224 See infra notes 545-67 and accompanying text.

225 A case is "cleared" when a suspect is arrested. The clearance rate (over 50%) is much lower than the founding rate (92%). See 1994 UCR, supra note 108, at 24. See generally KENNETH CULP DAVIS, DISCRETIONARY JUSTICE: A PRELIMINARY INQUIRY 3-26 (1980). In some cases, where the rape is accompanied by another crime such as a robbery, and the evidence of the rape is judged to be too difficult to prove, the police choose to classify the case under the label of the other crime. Duncan ChappeU & Susan Singer, Rape in New York City: A Study of Material in the Police Files and Its Meaning, in FORCIBLE RAPE, supra note 174, at 245, 256.

226 ESTRICH, supra note 1, at 1-2, !5-16; BROWNMILLER, supra note 1, at 364-68.

227 Examples are legion. For instance, one policeman told a rape victim that, "[u]sually if women are out at this time of night, it was their fault." VERNON R. WIEHE & ANN L. RICHARDS, INTIMATE BETRAYAL: UNDERSTANDING AND RESPONDING TO THE TRAUM OF ACQUAINTANCE RAPE 32 (1995). Another declared that, "[it's not right] for a girl to go into a stranger's apartment, drink beer, and then be upset when the guy makes advances." HOLMSTROM & BURGESS, SUpra note 4, at 39.

228 HOLMSTROM & BURGESS, supra note 4, at 9-11. Holmstrom and Burgess followed these victims from the time they were admitted to the hospital until the final outcome of the legal process. Id. at 1-2.

229 Id. at 4.

230 Id. at 35.

23l Like nearly all rape scholars, we choose to focus only on women as victims of rape, and we excluded statutory rape, which raises distinctive problems. For information on male victims of rape, see generally Cindy Struckman-Johnson, Male Victims of Acquaintance Rape, in ACQUAINTANCE RAPE, supra note 63, at 192.

232 While historically most interviews have been conducted by male officers, some police departments have put female officers in charge of interviewing rape complainants. See, e.g., BROWNMILLER, supra note 1, at 387 (New York City); LAFREE, RAPE & CRIMINAL JUSTICE, supra note 70, at 68 (Indianapolis).

233 HOLMSTROM & BURGESS, supra note 4, at 46.

234 Id. at 47-48. As one officer put it, "We go to so much trouble to work up a case and then take it to court and the girl doesn't show and the judge gets mad and throws it out and it makes us look foolish." Id. at 47. In Sweden, according to one scholar, 25% of rape complainants withdrew the complaint. The closer the relationship between the parties, and the lower the level of physical violence, the more frequent the withdrawal. For example, one Swedish study found that of those who gave a reason for withdrawing the complaint, 21% said that the investigation was mentally burdensome, 16% felt that they shared responsibility for the alleged rape, and 12% felt that the authorities did not believe them. Annika Snare, Sexual Violence Against Women: A Scandinavian Perspective, 9 VICTIMOLOGY 195, 206 (1984).

235 HOLSTROM & BRUGESS, supra note 4, at 47.

236 It would be more accurate, today, to say "casual nonmarital sex by females."

237 POSNER, supra note 8, at 388.

238 HOLMSTROM & BURGESS, supra note 4, at 51. Anotherstudy found that 68% of private agencies that provide counsel to rape victims had an "excellent" or "good" evaluation of police performance in rape cases, based on their own experiences and what they had heard from victims. RAPE IN AMERICA, supra note 56, at 11. The proportion giving that evaluation of prosecutors was 59%, of judges 46%, and of probation departments 40%. Id. A somewhat higher proportion of the agencies said that they had an excellent or good "working relationship" with the police (86%), prosecutors (79%),judges (68%), probation officers (61%), prisons (27%), and parole boards (18%). Id.

239 HOLMSTROM & BURGESS, Supra note 4, at 51.

240 Id. at 35-37.

241 Id. at 54-55.

242 Id. at 36-37.

243 Id.

244 See infra text accompanying notes 331-87.

245 HOLMSTROM 8C BURGESS, supra note 4, at 44.

246 Id. at 45.

247 id. at 46.

248 Id. at 47.

249 Id.

250 See, e.g., HOLMSTROM & BURGESS, supra note 4, at 44-50.

251 Similarly, in the United Kingdom the official unfounding rate is about 8-9%. S. Grace et al., Rape: From Recording to Conviction, in Kimberly A. Lonsway & Louise F. Fitzgerald, Rape Myths: In Review, 18 PSYCHOL. WOMEN Q. 133, 135 (1994).

252 Since police unfound all rape reports that they believe to be false, plus some for other reasons, the unfounding rate overstates the proportion of recorded reports that police believe to be false. But in some jurisdictions police do not record some rape complaints that they regard as palpably false. See infra notes 646-48, 694-95 and accompanying text.

253 See ChappeU & Singer, supra note 225, at 245-71.

254 See sources cited infra notes 648-49, 663-64.

255 Wayne A. Kerstetter, Gateway to Justice: Police and Prosecutorial Response to Sexual Assaults Against Women, 81 J. CRIM. L. & CRIMINOLOGY 267, 276 (1990).

256 Id. at 269-84. Kerstetter identified five major theories about the police response to rape: formal; criminological; conflict (as to race and class); conflict (as to gender); and instrumentalist. Id. at 271-76. He selected variables calculated to test the theories' validity.

257 Id. at 282.

258 Id. at 278. "Discriminant analysis is a statistical procedure for distinguishing between two or more groups of cases by using a collection of variables that measure characteristics on which the groups are expected to differ." Id.

259 Id. at 288.

260 Id. Although the complainant can be required to testify, if she does not wish to do so, her attitude is likely to gravely undermine the prosecution's case, in which she is the leading witness. LaFree, Official Reactions, supra note 52, at 588. The fact that the accused is in custody, however, while obviously significant in stranger rape cases, because it means that he has been identified, is presumably a consequence rather than a cause of founding in an acquaintance rape case.

261 Kerstetter, supra note 255, at 289. It may appear callous to analyze rape processing in economic terms, but with limited resources police inevitably ration their efforts. Kerstetter's analysis suggests that in some jurisdictions part of this process involves unfounding complaints where the chances of a successful prosecution are slight.

262 Id. at 290.

263 Id. at 297.

264 Id.

265 Id.

266 Id. at 302-06.

267 LaFree, Official Reactions, supra note 52, at 583, 585.

268 Id. at 588.

269 Id.

270 Id.

271 Id. at 592.

272 The only exceptions that come to mind are relatively rare, for example: a victim raped while in a drunken stupor at a fraternity party where the perpetrator might have been any of several men without alibis.

273 See, e.g., Kerstetter, supra note 255, at 292-93; LaFree, Official Reactions, supra note 52, at 585.

274 See, e.g., LaFree, Official Reactions, supra note 52, at 588-89.

275 See generally 11 CRIMES oF VIOLENCE, A STAFF REPORT TO THE NATIONAL COMMISSION ON THE CAUSES AND PREVENTION OF VIOLENCE, D. Mulvihill & M. Tumin 234-37 (1969); Wallace D. Loh, The Impact of Common Law and Rape Reform Statutes on Prosecution: An Empirical Study, 55 WasH. L. REV. 543, 590 (1980).

276 Chandler & Torney, supra note 134, at 155.

277 Id.

278 Id.

279 Id.

280 See supra note 65 and accompanying text.

28l Hubert S. Feild, Attitudes Toward Rape: A Comparative Analysis of Police, Rapists, Crisis Counselors, and Citizens, 36J. PERSONALITY & SOC. PSYCHOL. 156, 157 (1978).

282 Id. at 169.

283 Id.

284 Id. at 176.

285 Shirley Feldman-Summers & Gayle C. Palmer, Rape as Viewed by Judges, Prosecutors, and Police Officers, 7 CRIM. JusT. & BEHAV. 19, 36 (1980).

286 Id.

287 Id. at 25. Of the 62 officers to whom questionnaires were sent only 24.2% responded. Id. at 24-25.

288 Feild, supra note 281, at 176.

289 CATHARINE A. MACKINNON, ONLY Wows 18-19 (1993).

290 See, e.g., HOLMSTROM & BURGESS, supra note 4, at 39-44; J.C. LeDoux & R.R. Hazelwood, Police Attitudes and Beliefs Toward Rape, 13J. POLICE SCI. & ADMIN. 211, 219 (1985); LaFree, Official Reactions, supra note 52, at 583-88.

291 See Kerstetter, supra note 255, at 286-301; MCCAroLL ET AL., supra note 52, at 116-21.

292 HOLMSTROM & BURGESS, supra note 4, at 40, 43. Although this variable does not explicitly appear in the list published by MCCAHILL ET AL., supra note 52, at 116, several of the items on their list have sexual overtones--for example, "prior trouble with the police" sometimes means prostitution or juvenile promiscuity. MENACHEM AMIR, PAl-rEmuS I~ FORCIBLE RAPE 271-72 (1971). Similarly, "victim-precipitation" cases sometimes have overtones of promiscuity.

293 HOLMSTROM & BURGESS, SUpra note 4, at 41.

294 Id. at 43. Cf MCCAHiLL ET AL., Supra note 52, at 116 (no findings concerning alcohol or drugs, but "victim-precipitated" rapes more likely to be unfounded). Cf Kerstetter, supra note 255, at 297 (nonsexual discrediting information such as alcohol or drug use often led to unfounding). Kerstetter concluded, however, that police and prosecutors who take account of a woman's drinking are motivated by evidentiary considerations rather than their sex-role norms. Id., at 305.

295 HOLMSTROM & BURGESS, Supra note 4, at 42 (police "most impressed" by "a consistent and unchanging story"). Cf MCGILL ET AL., supra note 52, at 116 (32.9% unfounded if social worker evaluated her story as partially incredible; otherwise 13.7%).

296 HOLMSTROM & BURGESS, Supra note 4, at 40-42. Cf MCCAHILL ET AL., Supra note 52, at 116.

Even if the victim has lied in her report, she may have been raped. It is not unusual for a victim to be embarrassed over the fact that she accepted a ride from a stranger, so she may say that a car drew up beside her, the driver kidnapped her and raped her. She was raped but not kidnapped.

MACDONALD, supra note 178, at 145.

297 HOLMSTROM & BURGESS, Supra note 4, at 46.

298 Id. at 42. Although this does not appear on the list in MCCAHILL ET AL., supra note 52, at 116, common sense tells us that a rape report is more likely to be believed when it is corroborated by other evidence. This is a factor that may also be reflected in social workers' judgments as to whether the report is "partially incredible." See id.

299 Kerstetter found that in acquaintance rape (where of course, consent looms as an issue) this was one of the major determinants of founding decisions. Kerstetter, supra note 255, at 297.

300 HOLMSTROM & BURGESS, supra note 4, at 3940, 42. Cf MCCAHILL ET AL., SUpra note 52, at 116 (23% of cases that a social worker later evaluated as victim-precipitated were unfounded, versus 12.3% of other cases).

301 HOLMSTROM & BURGESS, supra note 4, at 3940, 42. Cf. McCAroLL ET AL., supra note 52, at 116 (victim-precipitation associated with unfounding).

302 HOLMSTROM AND BURGESS, supra note 4, at 4243.

303 Id. at 43; United States v. X-citement Video, Inc., 115 S. Ct 464, 475 (1994); MCCAHILL ET AL., supra note 52, at 116 (23.9% of cases in which victim reported to a social worker that she had seen a psychiatrist prior to the alleged rape were unfounded versus 11.8% with no such report). Cf. HOLMSTROM & BURGESS, supra note 4, at 74-78 (hospital saws negative reactions to "mental cases"); id. at 40 (negative po1ice reaction to a "mental case" who kept changing her story and had a history of psychiatric problems). Concerning runaways, see id. at 42; McCAHILL ET AL., supra note 52, at 116 (proportion unfounded for runaways was 24.1 versus 13.7 for others).

304 HOLMSTROM & BURGESS, supra note 4, at 43.

305 MCCAHILL ET AL., SUpra note 52, at 116 (53.3% of rape reports by obese women unfounded, as opposed to 14.2% of reports by others); HOLMSTROM & BURGESS, supra note 4, at 67 (police more sympathetic to "pretty and articulate" victims). Perhaps police find attractive women more likeable, or more poignant victims; or they may find rape of an attractive woman more plausible. See generally notes 48485 and accompanying text (for a discussion of the effects of victim attractiveness).

306 HOLMSTROM & BURGESS, SUpra note 4, at 4043. Another study found that 249'0 of the rape complaints in nonstranger cases were considered without merit, compared with under 5% in stranger cases. Chappell & Singer, supra note 225, at 245-71.

307 HOLMSTROM & BURGESS, SUpra note 4, at 4043; LeDoux & Hazelwood, supra note 290, at 219. Cf. MARSH ET AL., supra note 25, at 98 (victim's sexual relations with accused rapist rated more important than drinking or use of force).

308 HOLMSTROM & BURGESS, supra note 4, at 43. This item is not on the list in McCAHILL ET AL., SUpra note 52, at 116, but the victim's age (over 12) was found to correlate with unfounding, and presumably there is some overlap between this factor and the "different ages" factor, id.

309 MCCAHILL ET AL., supra note 52, at 116 (16.9% of reports by females in this age range unfounded, versus 6.6% of reports by younger females); HOLMSTROM & BURGESS, Supra note 4, at 67 (police more sympathetic if victim is extremely young).

310 HOLMSTROM & BURGESS, supra note 4, at 43 ("Nothing makes ... [the police] more enthusiastic about a case than to find out the assailant has other charges against him or a prison record."). For an analysis of the admissibility of prior offenses in sex-crime cases,

see generally David P. Bryden & Roger C. Park, "Other Crimes" Evidence in Sex Offense Cases, 78 MINN. L. REV. 529 (1994).

311 MCGILL ET AL., supra note 52, at 116 (21.8% unfounded versus 13% when someone else reported the rape). This is interesting, because one possible source of false reports is when a teenage girl lies to a parent in order to conceal a voluntary sexual encounter, leading the parent to report the incident to the police. See infra notes 622-25 and accompanying text. One wonders whether political reasons lead police to regard reports by parents as more credible, or as more worthy of respectful treatment.

312 MCCAHILL ET AL., supra note 52, at 116 (16.5% unfounded versus 7% where another crime occurred). Perhaps the police figured that anyone who is going to lie about a sexual encounter will probably allege a completed rape rather than a mere attempt or some other lesser offense. Or perhaps, in cases that are doubtful on other grounds, the detective is more willing to found the report if the consequences for the man are less severe.

313 Id. (25.3% unfounded versus 13.9%). Presumably, these are cases in which the police conclude that the victim scratched herself. For an example of a false rape report by a woman with self-inflicted injuries, see MACDONALD, supra note 178, at 86. In such cases, "[t]he cuts usually are superficial, and are on the front of the body, the cheeks, chest, abdomen, and thighs .... IT] he wounds tend to be within reach, at unusual angles, and often conform to the range of motion of the person's arm or hands." Id. at 106. See generally C.P. McDOWELL & N.S. HIBLER, False Allegations, in PRACTICAL ASPECTS OF RAPE INVESTIGATION (R.R. Hazelwood & Ann Burgess eds., 1987).

314 MCGILL ET AL., supra note 52, at 116 (18.9% unfounded versus 11.9%). As such, race was not significantly associated with unfounding in this study. Id. See LaFree, Official Reactions, supra note 52, at 592. Cf. Mimi H. Silbert, Compounding Factors in the Rape of Street Prostitution, in Raw & SEXUAL ASSAULT II 75 (A.W. Burgess ed., 1988) (academic studies show people discriminate against women of low respectability). A study of putative sexual assault victims in Philadelphia revealed that of the 121 adolescents between the ages of 13 and 17, 42% were from families receiving some form of public assistance. WILLIAM KRASNER ET AL., VICTIMS OF Rays (1976).

315 MCCAHILL ET AL., supra note 52, at 116 (21% unfounded versus 13.9%).

316 Id. (22% unfounded versus 13.9%). In evaluating this correlation, one should bear in mind that many cases of "prior trouble" involve voluntary sexual offenses such as prostitution. See AMIR, supra note 292. For a critical evaluation of this factor, see infra text accompanying notes 853-57, 869-84. See also MAcDoNALD, supra note 178, at 145 ("A routine part of any [police] investigation, not just of rape, is to check whether the victim or the suspect has a criminal record.").

317 See LAFREE, RAPE AND CRIMINAL JUSTICE, supra note 70, at 84-89. Cf. Eugene Kanin, False Rape Allegations, 23 ARCHIVES SEXUAL BEHAV. 81, 84-85 (1994).

318 McCAHILL EST AL., supra note 52, at 115-16 (finding this was the most important of 13 variables associated with unfounding: with a woman present the unfounding rate dropped from 21.9% to 9.2%).

319 LaFree found a strong correlation between late reports and negative founding decisions. LaFree, Official Reactions, supra note 52, at 588. The National Women's Study found that, of the rapes that were reported to the police, 25% were reported over 24 hours after the rape. RAPE IN AMERICA, supra note 56, at 5.

320 "Most detectives ask relatively few victims of sexual assault to take the [polygraph] test. It is more likely to be used in an acquaintance rape case, where there is no bodily injury, no other physical evidence, no eyewitnesses, and both victim and witness [putative offender?] give apparently truthful accounts of the time spent together." MACDONALD, supra note 178, at 146-47. Although some states prohibit use of the polygraph in sexual assault cases, detectives in other states routinely administer the test. Id. Prosecutors sometimes offer accused rapists a deal: "if you will take a polygraph test and waive your right to have the result excluded from evidence, we promise not to prosecute you unless you fail the test." Interview with Professor David Lykken, Psychology Depart., University of Minn., expert witness on the reliability of polygraph tests, in Minneapolis, Minn. (Oct. 1, 1995). Professor Lykken believes that polygraph tests are unreliable. Id.

321 WIEHE & RICHARDS, supra note 227, at 37 ("Police generally have far more cases than they can handle and thus are reluctant to pursue the investigation of rape cases where the victim is uncooperative. Also, police experience suggests that a face-saving way of recanting a complaint is to decide not to pursue it.").

322 HOLMSTROM & BURGESS, supra note 4, at 67 (police more sympathetic to "pretty and articulate" victims).

323 Confining ourselves to criteria explicitly invoked by the police, and putting aside for later analysis all of the contributory fault/morality factors, we would select the woman's failure to act upset as perhaps the least defensible criterion of her truthfulness. This phenomenon is likely to be due to differences in personality and experience among victims, rather than to differences between truthful and untruthful reports. Interview with Professor David Lykken, Psychology Dept., University of Minn., in Minneapolis, Minn. (July 24, 1995).

324 MCCAHILL ET AL., supra note 52, at 115-16.

325 See sources cited infra notes 516-28.

326 The trend toward use of female investigators goes back over 20 years. See Berger, supra note 1, at 7.

327 See, e.g., HOLMSTROM & BURtESS, supra note 4, at 39-40, 42-43; Ledoux & Hazelwood, supra note 290, at 219; MACDONALD, supra note 178, at 145; MARCH ET AL., supra note 25, at 98; MCGILL ET AL., supra note 52, at 116.

328 See infra text accompanying notes 631-33, 895-934.

329 For a discussion of prosecutors' attitudes, see HOLMSTROM & BURGESS, supra note 4, at 121-56.

330 See supra text accompanying note 236.

331 See infra notes 128-30 and accompanying text.

332 ZEISEL, supra note 94, at 26.

333 HOLMSTROM & BURtESS, supra note 4, at 56.

334 Id.

335 Id.

336 Id. at 60.

337 Id. at 58 tbl.4. Cf. supra note 173 (reasons given for failure to report the crime).

338 HOLMSTROM & BURGESS, supra note 4, at 58.

339 Id. at 58 tbl.4.

340 Other reasons were fear of a trial and intimidation by friends of the rapist. HURCH, supra note 128, at 110-14.

341 Kerstetter, supra note 255, at 267.

342 Id. at 308.

343 Id. at 309.

344 MARSH ET AL., supra note 25, at 86. Similarly, Canadian reforms apparently did not affect the unfounding rate. Roberts & Gebotys, supra note 209, at 571. But since the volume of rape reports rose after reforms in both Michigan and Canada, perhaps a higher proportion of marginal cases entered the system. If so, a stable unfounding rate would be consistent with a greater willingness to found complaints.

345 MARSH ET AL., Supra note 25, at 28.

346 "The monthly pattern of arrests for forcible rape from 1972 through 1978... reveals a relationship between arrests and the law's enactment, and this is confirmed by the timeseries analysis." Id.

347 1992 UCR, supra note 161, at 210. A reported crime is said to be "cleared" when at least one person is arrested, charged with an offense, and turned over for prosecution. Id. at 202.

348 See generally U.S. PRESIDENT'S COMM'N ON LAW ENFORCESSENT & THE ADMIN. OF JUSTICE, THE CHALLENGE OF CRIME IN A FREE SOCIETY 329-33 (1967) [hereinafter THE CHALLENGE OF CRIME]; WAYNE R. LAFAVE &JEROLD H. ISRAEL, CRIMINAL PROCEDURE (Sections) 13.1-13.2 (Hornbook series, student ed. 1985).

349 MCCAHILL ET AL., SUpra note 52, at 170-71 (explaining that over 15% of all rape charges are dismissed because the victim is afraid to face her attacker and to discuss the rape before strangers).

350 Frohmann, supra note 13.

351 Id. at 215. For similar conclusions, see, e.g., FAIRSTEIN, supra note 13, at 151-52; VACHSS, supra note 13, at 140-43; MARSH ET AL., supra note 25, at 101.

352 See, e.g., FAIRSTEIN, supra note 13, at 133-36 (describing the obstacles to conviction in acquaintance rape cases).

353 See id. at 151-52; VACHSS, supra note 13, at 90-91, 231. 354 HOLMSTROM & BURGESS, supra note 4, at 138. 355 Id. at 142. 356 ld. at 145-46.

357 Id. at 144.

358 Id. at 143.

359 Id. at 144.

360 Id.

361 Id.

362 VACHSS, supra note 13, at 90.

363 See supra text accompanying notes 350-51. In general and within reason, scholars agree than winnability is a legitimate factor. E.g., THE RESPONSE TO RAPE, supra note 109, at 11; FRANK W. MILLER, PROSECUTION: THE DECISION TO CHARGE A SUSPECT WITH A CRIME 4-7, 11-19 (Frank J. Remington ed., 1969).

364 THE RESPONSE TO RAPE, supra note 109, at 11.

365 See, e.g., ESTRICH, supra note 1, at 15-26; HOLMSTROM & BURtESS, supra note 4, at 133-48; SPOHN & HORNEY, supra note 170, at 18-20.

366 Our intuition would perhaps be more debatable if the question were rephrased: What proportion of rape victims have heard about a rape case (or cases in general) in which a prosecutor refused to file charges? Never having heard such rumors ourselves, we doubt that they are as widespread as the Senate report implies. But perhaps in some highcrime-rate subcultures such rumors abound, especially if one cumulates rape with other crimes, or prosecutors' responses with those of other officials.

367 See, e.g., HOLMSTROM & BURGESS, Supra note 4, at 254-55.

Surprisingly, the effect of trial verdicts on long-term response [of the victim to the rape] has not been studied to any great extent. [One study].., found that conviction led to fewer symptoms 6 months later. But some researchers suggest that the verdict is often less distressing to the victim than her experiences during the actual trial proceedings. Many victim/witness advocates, however, express frustration in attempting to explain to a woman whose assailant was adjudged not guilty why this verdict has been handed down; convincing her that the jury did not necessarily disbelieve her testimony is not an easy task. It would be surprising if the verdict had little effect on the victim.

Steketee & Austin, supra note 174, at 299. Nevertheless, these authors believe that "involvement in the criminal justice system, particularly trial proceedings, potentially has a positive effect on the victim with regard to expression of anger, emotional processing of the event, a sense of empowerment, and reduction of feelings of victimization." Id. at 299-300. This effect is, however, mediated by various aspects of the trial including the outcome, ld. Another author states that in Michigan, "victims who had received counseling were often well prepared psychologically for the ordeal of a trial; and they made better witnesses." MARSH ET AL., supra note 25, at 13.

368 Most prosecutors in urban jurisdictions are overloaded with cases. See, e.g., FAIRSTEIN, supra note 13, at 91 (stating that prosecutors use plea bargaining as a way to manage their workload); THE CHALLENGE OF CRIME, supra note 348, at 323-24 (arguing that a substantial number of cases must be dismissed or negotiated). Prosecutors often "receive cases at the last moment and are faced with the task of preparing several cases simultaneously." HOLMSTROM & BUROESS, supra note 4, at 148. One harried district attorney, evaluating the merits of a rape case, said: "I don't want to waste a lot of time on this and have the jury come back in two minutes with a not guilty verdict." I& In addition, a number of prosecutors work only part-time and maintain a private practice. THE CHALLENGE OF CRIME, supra note 348, at 362-63.

369 ESTRICH, supra note 1, at 9.

370 Id. at 20.

371 Id.

372 Id.

373 Id.

374 Id. at 20-21.

375 Id. at 21.

376 Nearly all rape trials occur in state courts. While annual conviction rates for federal courts are available, broken down by specific crimes, no comparable data of national scope exists for state trials. Scholars have sometimes published such statistics, but the jurisdictions studied may not be representative, and the statistics must be interpreted carefully. Official statstics invariably lump together stranger rapes and acquaintance rapes; the difficulties of the latter type of prosecution are thus obscured. This problem is not wholly solved by compiling separate acquaintance rape figures: one wants subdivisions for each type of case--for example, cases in which the putative victim's sexual history was in evidence, cases in which she had a pre-rape sexual relationship with the perpetrator, and so on.

Even if all this information is available, numerous questions remain: Is the jurisdiction typical? Is the data sufficiently recent? (Scholars usually cannot say, for example, whether attitudinal changes have made a ten-or fifteen-year-old study obsolete). If the conviction rates are high, may this be due to the reluctance of prosecutors to try difficult cases, rather than the willingness of judges and juries to convict? Conversely, does a disappointingly low rate mask prosecutorial aggressiveness? Due account must also be taken of plea bargaining--did prosecutors inflate conviction rates by agreeing to excessively lenient deals?--and convictions on lesser charges that may or may not be due to excessive jury leniency or prosecutorial timidity.

One also should ask how the conviction rate compares with the rate for comparable crimes, bearing in mind that for some purposes the "comparable" crime may be, for example, nonsexual assaults in which the defense was serf-defense and adverse character evidence about the putative victim was introduced rather than nonsexual assaults in general. Finally, and most obviously, conviction rates reveal what happened, not what should have happened. A low conviction rate may be distressing and yet partly due to legitimate problems of proof.

Despite all these complexities, conviction rates are sometimes informative. If we learn, for example, that juries in New York convicted acquaintance rapists in 60% of the cases, but in none of the 50 cases in which the parties had been lovers, we can safely conclude that the latter type of case raises difficult problems for prosecutors.

377 KALVEN & ZEISEL, supra note 52, at 252-54.

378 Weninger, supra note 133, at 360-62.

379 McCAHILL ET AL., supra note 52, at 191.

380 Williams, supra note 65, at 36.

381 Id. The victim-defendant relationship was unknown in about 20% of the cases; in these cases the conviction rate was 41%. Id. Since the relationship was not mentioned in these case records, probably most were stranger rapes. Other estimates of acquittal rates vary somewhat, but are almost always lower for nonaggravated acquaintance rape than for stranger rape. See HOLMSTROM & BURGESS, supra note 4, at 247 (citing conviction rate of 11.3% for all stranger rape cases reported to police, and 2.5% for corresponding acquaintance rapes); Tamar Lewin, Tougher Laws Mean More Cases Are Called Rape, N.Y. TIMES, May 27, 1991, at A8 (citing conviction rate of 60% for all stranger rape cases going to trial and 40% for the corresponding acquaintance rape cases); see generally Gary D. LaFree, Variables Affecting Guilty Pleas and Convictions in Rape Cases: Toward a Social Theory of Rape Processing, 58 Soc. FORCES 833 (1980) (reporting a significant correlation between victim-defendant relationship and probability of conviction, with acquainted defendants more likely to be acquitted). But see Homey & Spohn, supra note 51 (finding no difference in convictions of stranger versus acquaintance rape in Detroit).

382 This can be inferred from the small numbers of prosecutions for acquaintance rape in the jurisdictions studied, and from anecdotal evidence.

383 See supra notes 65, 380-81 and accompanying text.

384 In deciding whether a conviction is sustainable, courts ask "whether upon the evidence, giving full play to the right of the jury to determine credibility, weigh the evidence and draw justifiable inferences of fact, a reasonable mind might fairly conclude guilt beyond a reasonable doubt." LAFAVE & ISRAEL, supra note 348, (Section) 23.6(a).

385 See id.

386 See infra text accompanying notes 446-96.

387 See LAFAVE & ISRAEL, supra note 348, (Section) 21.1 (g).

388 SPOHN & HORNEY, supra note 170, at 102.

389 Id. at 102-03.

390 Id.

391 Id. at 102.

392 Id. at 93-97.

393 Loh, supra note 275, at 613-I4. An analysis of the impact of Canadian reforms found that the rate at which rape cases led to the filing of charges rose after reforms, but since a comparable increase occurred for other crimes of violence, this was not due to the reforms. Roberts & Gebotys, supra note 209, at 569.

394 Kenneth Polk, Rape Reform and Criminal Justice Processing, 31 CRIME & DELINQ. 191, 194-98 (1985).

395 See, e.g., ESTRICH, supra note 1, at 15-25.

396 Compare supra text accompanying notes 291-322, 349-62 (describing the types of cases that are more often unfounded by the police and dismissed by prosecutors), with infra text accompanying notes 389-405 (describing the types of cases that result in acquittal by juties). Our point is further illustrated by data from a study of requests by police (which must be approved by prosecutors) for arrest warrants for criminal sexual conduct in Kalamazoo County, Michigan from 1975 to 1977. In 64% of the cases in which the police founded the complaint and requested a warrant, no conviction ensued, either because the prosecutor denied the request, or because an acquittal, dismissal or nolle ended the proceeding. Susan Caringella-MacDonald, Sexual Assault Prosecution: An Examination of Model Rape Legislation in Michigan, 4 WOMEN AND POL. 65, 71 tbl. 1 (1984). Thus, even in the cases in which the police believed the complainant, the accused usually was not convicted. In cases in which the police unfounded the complainant because of unjustifiable doubts about the complainant's veracity, or because they considered prosecution to be futile, the chances of obtaining a conviction were probably on average much weaker.

397 Although recent national studies do not exist, the conviction rate in nonaggravated acquaintance rape cases has historically been below 50%. See supra text accompanying notes 377-81. It follows that, in most of the simple rape cases in which police and prosecutors believed both that the complainant was telling the truth and that the case was worth pursuing, jurors did not find her version credible beyond a reasonable doubt.

398 KALVEN & ZEISEL, supra note 52.

399 fid. at 67.

400 Id. at 253.

401 Id.

402 Id. at 253-54 & tbl.73.

403 Berger, supra note 1, at 30. See generally Carol Bohmer, Judicial Attitudes Toward Rape Victims, 57 JUDICATURE 303 (1974).

404 KALVEN & ZEISEL, supra note 52, at 253-54.

405 Id. at 254.

406 See ALLISON & WRITGHTSON, supra note 56, at 52 (stating that "more violent rapes are more likely to be reported to the police" and that "the greater the amount of physical violence used by the rapist, the more likely the attacker was to succeed and the more likely the rapist was reported to the authorities"); Bachman, supra note 166, at 265-66 (concluding that victims are more likely to report rape if they experience severe physical force or are injured); Lizotte, supra note 138, at 181-85 (stating that "victims report rape to the police in response to factors that make the incident more serious and hence make prosecution easier").

407 For a discussion of factors affecting the decisions of police and prosecutors in rape cases, see supra text accompanying notes 225-397.

408 KALVEN & ZEISEL, supra note 52, at 28 n.28.

409 Id. at 69-75.

410 Id. at 74.

411 Id. at 71.

412 Id. at 70.

413 Id. at 69.

414 Id.

415 Id.

416 Id.

417 Id.

418 Game Laws (+43); Indecent Exposure (+41); Gambling (+35); Statutory Rape (+32); Fraud (+25); Drunken Driving (+25); Petit Larceny (+25); Receiving Stolen Goods (+24); Molestation of a Minor (+21); Burglary (+21); Other Liquor Offenses (+20). Id. at 69-75.

419 See id. at 253-54.

420 The nearest major crime is Statutory Rape (+32), followed by Murder (+29) and Manslaughter (+29). Id. at 69-75.

421 Id. at 250 422 Id.

423 Id.

424 Id.

425 Id.

426 Id. at 249-51.

427 Elaine Walster, Assignment of Responsibility for an Accident, 3 J. OF PERSONALITY & Soc. PSYCHOL. 73, 73-79 (1966).

428 This same objection applies to two related theories: (1) that hindsight leads jurors to overestimate the causal role of the victim's behavior; and (2) that people need to believe in a "just world" and consequently tend to assume that the victim must have deserved her fate. See Ronnie Janoff-Bulman et al., Cognitive Biases in Blaming the Victim, 21 J. EXPERIMENTAL SOC. PSYCHOL. 161 (1985)(hindsight theory); see also MELVIN J. LERNER, THE BELIEF IN A JUST WOMB: A FUNDAMENTAL DELUSION (1980); MelvinJ. Lerner & D. T. Miller, Just World Research and the Attribution Process: Looking Back and Ahead, 85 PSYCHOL. BULL. 1030 (1978); Melvin J. Lerner & Carolyn H. Simmons, Observers' Reaction to the Innocent Victim: Compassion or Rejection ?, 4 J. oF PERSONALITY & SOC. PSYCHOL. 203 (1966).

429 For a discussion of these cases, see infra text accompanying notes 981-94.

430 For a discussion of the relevance of evidence of promiscuity, see infra notes 895-934 and accompanying text.

431 Some studies have found that contributory fault of the putative victim affects verdicts in stranger rape scenarios as well as in date rape. Martha R. Burt & Rochelle S. Albin, Rape Myths, Rape Definitions and Probability of Conviction, 11J. APPLIED SOC. PSYCHOL. 212 (1981). But see LaFree, Variables Affecting Guilty Pleas, supra note 381, at 833 (rape victim's character related to verdicts in consent defense cases, but not in cases of mistaken identification).

432 KALVEN & ZEISEL, supra note 52, at 165.

433 Id. at 164435 ("The closeness of the evidence makes it possible for the jury to respond to sentiment by liberating it from the discipline of the evidence.").

434 One study found that in cases where the defendant either claims that he has been misidentified as the rapist, or argues for diminished responsibility due to insanity or drug use, testimony about his use of a weapon strongly inclined jurors to believe that he was guilty; jurors told the researchers that evidence of a weapon made them anxious to incarcerate the defendant. LaFree, Jurors'Responses, supra note 66, at 399.

435 KALVEN & ZEISEL, supra note 52, at 244-57.

436 See generally 3 WHARTON'S, supra note 44, (Section) 276.

437 See, e.g., People v. Mayberry, 542 P.2d 1337 (Cal. 1975).

438 See infra text accompanying notes 793-873.

439 Jahnke v. State, 682 P.2d 991, 1010 (Wyo. 1984) (Brown, J., concurring).

440 The trials on which Kalven and Zeisel's findings were based occurred between 1954 and 1958. KALVEN & ZEISEL, supra note 52, at 33 n.1.

441 Kalven and Zeisel asked 3500 judges to participate in the study, but only 555 did so. KALVEN & ZEISEL, supra note 52, at 35-36. The judges who chose to participate may have differed in some way from the larger group who were asked to participate, for instance by being more (or less) conviction-prone. In addition, some judges submitted over 50 trial reports, while others submitted only one. Id. at 39. Thus, the responses of these extremely conscientious judges were weighted heavily, perhaps skewing the results.

The judges' opinions may also have been affected by their hypothetical nature: Unlike the jurors' verdicts the judges' "verdicts" had no effect on the defendants' lives. Perhaps the opinions of judges participating in the study were affected by their knowledge that scholars were observing their performance.

Because Kalven and Zeisel could not observe jury deliberations, they had to rely on their own interpretations of the trial judges' comments about the cases. Since the judges usually did not explicitly adopt the contributory fault hypothesis, perhaps Kalven and Zeisel's preconceptions affected their interpretations. This problem could have been averted by adopting a coding scheme and asking independent evaluators (who were blind to the contributory fault hypothesis) to read the judges' comments and assign them to categories. Kalven and Zeisel could then have performed statistical significance tests on these categorical codings. For a sharp critique of the study's methodological faults, see Michael H. Walsh, The American Jury: A Reassessment, 79 YALE LJ. 142 (1969).

We would be more concerned about these methodological flaws if the evidence from other sources contradicted Kalven and Zeisel's findings, but on the whole other studies have reached conclusions that are consistent with Kalven and Zeisel's.

442 See, e.g., ADLER, supra note 48, at 119-21; FAIRSTEIN, supra note 13, at 133-36; VACHSS, supra note 13, at 90-92.

443 MARSH ET AL., SUSa note 25, at 44. One defense attorney expressed the shift in public attitudes graphically: "Over the last five or ten years, Detroit has changed. You just can't win by showing that the victim wasn't wearing underwear." Id. at 54.

444 James P. Levine, Using Jury Verdict Forecasts in Criminal Defense Strategy, 66 JUDICATURE 448, 455 (1983).

445 Telephone Interview with Linda Fairstein (July 19, 1996). Cf. FAIRSTEIN, supra note 13, at 134-36.

446 Some methodologies are high in external validity, meaning the results are generalizable to the "real world," while others are high in internal validity, meaning that the cause and effect relationship of the independent variables on the dependent variables is certain, even though the context of the experiment is artificial. Studies that are high in both internal and external validity are very rare in the context of jury decisionmaking. The most externally valid study would include real jurors, in real trials, actually deliberating. A scholar would gather data on many actual trials, coding for variables such as juror sex, defendant race, and situational factors revealed by the evidence such as the woman's drinking or the degree of force employed by the defendant. Statistical analyses would then be performed to determine whether any of the coded factors are related to the verdict, and which are most important. While high in external validity, this methodology is low in internal validity (or ability to prove cause and effect), because the scholar has no control over factors such as jury selection that vary among the trials.

Another type of study, fairly high in external validity, and also high in internal validity, employs mock jurors and videotaped trials differing along the dimensions of experimental interest. For example, mock jurors would view a tape of a reenactment of a rape trial in groups of six (or twelve). Half of the groups of jurors would watch a trial where the defendant and the putative victim had previously had a relationship and half would see a trial in which they were strangers. The jurors would then deliberate and reach verdicts. If the verdicts for the two types of cases differed over many groups of jurors, one could infer that the type of relationship between the victim and the defendant caused the differences in conviction rates. This type of experiment is also fairly high in realism and thus external validity, especially if the mock jurors are drawn from an actual jury pool as opposed to a college population. But such studies are rare because they are extremely time-consuming and expensive.

A third type of study, the most common in the literature, is high in internal validity but not very realistic. In this type of study, subjects read an abbreviated transcript of a rape case or a short rape scenario. Different subjects receive different versions of the rape, varying along the dimensions of experimental interest. Typically, the subjects are asked to make individual judgements (verdicts, attributions of fault, etc.) without collective deliberation. As a result of the experimenter's full control over the experimental variables and random assignment of subjects to each experimental condition, this type of study has high internal validity and cause and effect can be inferred from significance testing. But, such experiments obviously lack realism.

447 See generally KALVEN & ZEISEL, supra note 52, at 489; Garold Strasser et al., The Social Psychology of Jury Deliberations, in THE PSYCHOLOGY OF THE COURTROOM 221 (Norbert L. Kerr & Robert M. Bray eds., 1982); see a/so Jeffrey Kerwin & David R. Shaffer, Mock Jurors versus Mock Juries: The Role of Deliberations in Reactions to Inadmissible Testimony, 20 PERSONALITY & SOC. PSYCHOL. BULL. 153 (1994). Kerwin and Shaffer found that mock juries that deliberated were significantly less likely to be influenced by extralegal factors than were individual mock jurors, who did not deliberate. Id. at 159. They conclude that "allowing verdicts to unfold through the process of deliberation may be an important component of realistic mock juridic research." Id. at 161.

448 Although the consensus is that jurors ignore trial judges' instructions to disregard inadmissible evidence, experimental studies usually do not include group deliberations. Kerwin found that mock jurors who participate in deliberations are more likely to follow the (mock)judicial instructions that they receive. Kerwin & Shaffer, supra note 447, at 159.

449 See generally Mark A. Whatley, Victim Characteristics Influencing Attributions of Responsibility to Rape Victims: A Meta-Analysis, 1 AGGRESSION & VIOLENT BEHAV. 81, 83 (1996); Susan T. Bell et al., Understanding Attributions of Blame in Stranger Rape and Date Rape Situations, 24 J. APP. SOC. PSYCH. 1719, 1722 (1994).

45O Whatley, supra note 449, at 84.

451 Id. at 90-91.

452 Ronald E. Smith et al., Ro/e and Justice Considerations in the Attribution of Responsibility to a Rape Victim, 10 J. RES. PERSONALITY 246 (1976),

453 Id. at 348.

454 See, e.g., Joyce E. Williams, Secondary Victimization: Confronting Public Attitudes About Rape, 9 VICTIMOLOGY 66, 70-71 (1984); Cynthia E. Willis, The Effect of Sex Role Stereotype, Victim and Defendant Race, and Prior Relationship on Rape Culpability Attributions, 26 SEX ROLES 213, 218 (1992)(sexual egalitarians showed a stronger belief in victim truthfulness than traditionalists, but only if accuser and accused were acquainted).

455 See, e.g., Burt & Albin, supra note 431, at 218-24; Williams, supra note 454, at 70-71.

456 With few exceptions, studies comparing victim responsibility in stranger and acquaintance rape cases overwhelmingly support the hypothesis that victims of acquaintance rape are blamed more than those assaulted by strangers. This is true over a variety of measures of victim fault. See, e.g., Bell et al., supra note 449, at 1729; Judith S. Bridges & Christine A. McGrail, Attributions of Responsibility for Date and Stranger Rape, 21 SEX ROLES 273 (1989); James D. Johnson & Inger Russ, Effects of Salience of Consciousness-Raising Information on Perceptions of Acquaintance Versus Stranger Rape, 19 J. APPLIED SOC. PSYCHOL, 1182, 1187 (1989); L'Armand & Pepitone, supra note 65, at 134; Williams, supra note 454; James D. Johnson, The Effect of Rape Type and Information Admissibility on Perceptions of Rape Victims, 30 SEX ROLES 781,787 (1994); Cynthia E. Willis & Lawrence S. Wrightsman, Effects of Victim Gaze Behavior and Prior Relationship on Rape Culpability Attributions, 10 J. INTERPERSONAL VIOLENCE 367, 374 (1995).

Some studies contradict or only partially confirm the majority view that subjects treat stranger rapists differently from acquaintance rapists. See, e.g., HORNEY, & SPOHN, supra note 51; Willis, supra note 454, at 219-20. Willis found a difference in victim responsibility ratings for acquaintance and stranger rapes when the victim was black but not when she was white. She found a significant effect of prior relationship on sentence, however, with the date rapist receiving only 8.98 years and the stranger rapist receiving a much harsher sentence of 13.77 years. Id. at 222. In another study, subjects read a rape scenario in which the victim and defendant were described as having dated a few times or as strangers. Patricia Tetreault & Mark A. Barnett, Reactions to Stranger and Acquaintance Rape, 11 PSYCHOL. WOMEN Q. 353 (1987). All subjects then viewed an identical five-minute videotape of the victim talking to a therapist about her rape. They were then asked to rate the victim's responsibility for the rape on a 7-point scale. The difference in responsibility ratings did not approach significance, but since the average rating was rather low (about 1.65), id. at 355, perhaps viewing the victim in therapy made subjects more sympathetic to the date-rape victim than they would have been in a more realistic setting.

457 Klemmack & Klemmack, supra note 177, at 135, 144. See also R.D. Stacy et al., A Comparison of Attitudes Among College Students Toward Sexual Violence Committed By Strangers and By Acquaintances, 18 J. SEX EDUC. & THERAPY 257, 259451 (1992) (finding that both men and women consider stranger rape to be a more serious offense than acquaintance rape).

458 See, e.g., HOLMSTROM & BURGESS, supra note 4 at 279; Galvin, supra note 4, at 812-905.

459 Gary D. LaFree et al., Jurors' Responses to Victims' Behavior and Legal Issues in Sexual Asault Trials, 32 SOC. PROB. 389, 393 (1985).

460 Id. at 393, 397.

461 Id.

462 See, e.g., Suresh Kanekar et al., The Acquaintance Predicament of a Rape Victim, 21 J. AS'PLIED SOC. PSYCH. 1524, 1537 (1991); L'Armand & Pepitone, supra note 65, at 137 tbl. 1; Willis, supra note 454, at 222.

463 See, e.g., Acock & Ireland, supra note 7, at 187; Bridges, supra note 7, at 304-05.

464 Of course, the labels have opposite pejorative innuendoes: "Contributory fault" is explicitly derogatory, while "violation of a sex-role norm" has, at least in the modern academic culture, a mildly favorable innuendo. Even apart from their pejorative connotations, both terms are inexact. "Contributory fault" sounds like negligence, but why is victim "negligence" more important in acquaintance rapes than in stranger rapes or burglaries? Similarly, if "violation of sex-role norms" were the decisive factor, one would expect to find discrimination against female professionals, in all kinds of rape cases and presumably in other types of cases as well. It would be better, we believe, to speak of "violation of traditional rules of female sexual prudence."

465 Acock & Ireland, supra note 7, at 184.

466 Id. at 187.

467 See, e.g., Charlene Muehlenhard, Misinterpreted Dating Behaviors and the Risk of Date Rape, 6J. Soc.& CLINICAL PSYCH. 20, 24-25 (1988).

468 Id. at 25-26.

469 LaFree et al., supra note 66, at 397.

470 Id.

471 Id.

472 BURT & ALBIN, supra note 431. This is one of the few studies using adult subjects comparable to actual jurors. The average age of the sample was 42 years, with 12.8 years of schooling, ld. at 216. Subjects indicated their opinions about the two actors in the scenario on a semantic differential scale, which rates the actors on several dimensions such as bad-good, worthless-valuable, etc. Ratings for all dimensions are added to obtain a composite score of the subjects' evaluations of the actors. Id.

473 Id. at 218-19.

474 Id. at 224 tbl.3.

475 Id.

476 M.D. Pugh, Contributory Fault and Rape Convictions: Loglinear Models for Blaming the Victim, 46 SOC. PSYCH. Q. 233 (1983).

477 Id. at 236. Females were more likely (79%) than males (64%) to convict. For the female subjects, whether the victim had previously picked up men in bars had no effect on their verdicts. But only 37% of the males found the defendant guilty after reading that evidence, while 72% of the males found him guilty without the evidence. Id.

478 See, e.g., Bridges & McGrail, supra note 456, at 277-79; Megan Jenkins & Faye Dambrot, The Attribution of Date Rape: Observer's Attitudes and Sexual Experience and the Dating Situation, 17J. APPLIED PSYCH. 875, 882-84 (1987); Andrea Parrot & Laurie Bechhofer, What Is Acquaintance Rape, in ACQUAINTANCE RAPE, supra note 63, at 9, 22; Pugh, supra note 476, at 236. Of course, the idea that a "teaser" is fair game for rape is not confined to youths. See, e.g., R. Giarrusso et al., Adolescents' Cues & Signals: Sex and Assault (paper presented at the Western Psychol. Ass'n Meeting, San Diego, 1979), quoted in BOURQUE, supra note 25, at 66 (54% of males and 31% of females consider aggressive sex appropriate if the woman has "led him on"); E.J. Kanin, Date Rapists: Differential Sexual Socialization and Relative Deprivation, 14 ARCHIVES OF SEXUAL BEHAV. 219 tbl.2 (1985) (81% of rapists and 40% of controls believed aggressive sex is appropriate ff the woman is "a known 'teaser' ").

479 Cathaleene Jones & Elliott Aronson, Attribution of Fault to a Rape Victim as a Function of Respectability of the Victim, 26 J. PERSONALITY & SOC. PSYCH. 415 (1973).

480 Id. at 418-19.

481 See, e.g., Feldman-Summers & Lindner, supra note 67, at 59, 146 (responsibility assigned prostitute greater than other women; contrary result of Jones & Aronson explained as due to their excessively narrow range of respectability conditions); A. Kahn et al., Attribution of Fault to a Rape Victim as a Function of Respectability of the Victim, 8 REP. RES. ON SOC. PSYCHOL. 98 (1977). Cf. Charles H. Moore et al., The Effects of Physical Attractiveness and Social Desirability on Judgments Regarding a Sexual Harassment Case, 9 J. SOC. BEvy. & PERSONALITY 715, 725 (1994) (mock jurors reached more frequent guilty verdicts in a sexual harassment case when they perceived the plaintiff to be socially desirable). See generally Whatley, supra note 449, at 87, 90.

482 See, e.g., BOURQUE, supra note 25, at 66 (table showing circumstances in which poll respondents believed aggressive sex appropriate); Feldman-Summers & Lindner, supra note 67, at 141; LaFree et al., supra note 66, at 397; C. Neil MacRae &John W. Shephard, Sex Differences in the Perception of Rape Victims, 4 J. INTERPERSONAL VIOLENCE 278, 284 (1989); Pugh, supra note 476, at 236-38; Jody Miller & Martin Schwartz, Rape Myths and Violence Against Street Prostitutes, 16 DEVIANT BEHAV. 1, 9-17 (1995).

483 See Richardson & Campbell, supra note 73, at 472. On a seven-point scale, subjects assigned the drunk victim a responsibility score of 2.75, and a sober victim 2.34. If the defendant was drunk, by contrast, subjects assigned him less responsibility than if he was sober. Id. at 472. The victim's intoxication also affected female subjects' recommended sentences, which were shorter for one who had raped a drunk victim. Id. at 473. Male subjects' sentences were unaffected by this factor. Id. Subsequent studies have confirmed Richardson and Campbell's finding that victims are held more responsible if drunk, while drunk perpetrators are seen as less culpable. See, e.g., Georgina S. Hammock & Deborah R. Richardson, Blaming Drunk Victims: Is It Just World or Sex Role Violation?, 23 J. APPLIED SOC, PSYCHOL. 1574, 1583-85 (1993); Antonia Abbey & RichardJ. Harnisch, Perceptions of Sexual Intent: The Role of Gender, Alcohol Consumption and Rape Supportive Attitudes, 32 SEX ROLES 297., 307 (1995). Cf. William E. Snell & Lisa Godwin, Social Reactions to Depictions of Casual and Steady Acquaintance Rape: The Impact of AIDS Exposure and Stereotypic Beliefs About Women, 29 SEX ROLES 599, 607 (1993) (subjects believed excessive drinking by male and female equally contributed to the rape on their first date, but thought that for long-term dating partners intoxication contributed more to the rapist's behavior than to the victim's). But see GloriaJ. Fischer, Effects of Drinking by the Victim or Offender on Verdicts in a Simulated Trial of an Acquaintance Rape, 77 PSYCHOL REP. 579, 583-86 (1995) (finding college students did not attach greater or less responsibility for an assault if either the victim or perpetrator had been drinking, a result which may be unique to the campus context).

484 See, e.g., Deborah G. Schult & Lawrence J. Schneider, The Role of Sexual Provocativeness, Rape History, and Observer Gender in Perceptions of Blame in Sexual Assault, 6 J. INTERPERSONAL. VIOLENCE 94, 99 (1991) (subjects attributed more blame to victim of sexual assault as the victim's provocativeness intensified).

Some evidence suggests that the victim's provocative dress does not affect subjects' responses to a stranger rape scenario. Scroggs found that subjects assigned the rapist the same sentence regardless of whether the victim of a stranger rape was wearing a raincoat that hid her figure or a low-cut top that displayed her "rather large bosom." James R. Scroggs, Penalties for Rape as a Function of Victim Provocativeness, Damage, and Resistance, 6 J. AS'PLIED SOC. PSYCH. 360, 364 (1976). Another study found that a victim's provocative dress did not affect subjects' recommended sentences in stranger rape scenarios, but did affect subjects' attributions of responsibility for the rape. Suresh Kanekar & Maharukh B. Kolsawalia, Responsibility of a Rape Victim in Relation to her Respectability, Attractiveness, and Provocativeness, 112 J. SOC. PSYCH. 153 (1980). See generally Whatley, supra note 449, at 86-87, 90.

485 L'Armand & Pepitone, supra note 65, at 135.

486 Id. at 137 tbl.1.

487 Id. The pattern for the victim's psychological damage paralleled these findings. Id. Similar findings were reported in another study, of acquaintance rapes in which the victim was described either as a virgin or as promiscuous. Subjects rated the defendant as deserving more punishment for raping a virgin, believed the virgin was more psychologically damaged, felt that the incident was more serious, and blamed the victim less. MacRae & Shepherd, supra note 482, at 284 tbl.2.

The victim's sexual history may have less effect on reactions to stranger rapes, where her sexual history is plainly irrelevant to the man's usual defense (misidentification) than in acquaintance rapes, where the victim's experience is at least superficially relevant to the consent defense. In addition, people may regard rape by stranger as a horrible experience for any woman, while believing that rape by an acquaintance is less devastating if one is accustomed to casual sex. See Hubert S. Feild, Rape Trials and Jurors' Decisions: A Psycholegal Analysis of the Effects of Victim, Defendant and Case Characteristics, 3 LAW & HUM. BEHAr. 261, 271 (1979) (with stranger rape scenario, subjects assigned virtually identical sentences to rapist of virgin and sexually-experienced woman).

488 More precisely, cases in which the defense asserts either that no sex occurred or that the defendant was misidentified as the rapist--typically stranger cases. LaFree et al., supra note 66, at 397, 399.

489 Id. (But it only "approached statistical significance").

490 Homey & Spohn, supra note 51.

491 As the state with rape law reforms that are generally considered to be exemplary, Michigan may be legally unique; there is also evidence that officials there are unusually sensitive to feminist concerns about rape. See infra text accompanying notes 537-39. See generally MARSH ET AL., supra note 25.

492 See, e.g., Sheila R. Deitz et al., Attribution of Responsibility for Rape: The Influence of Observer Empathy, Victim Resistance, and Victim Attractiveness, 10 SEX ROLES 261, 274-75 (1984) (subjects who scored low in rape empathy rated unattractive victim as more likely to have encouraged the rape); Marsha B. Jacobson, Effects of Victim's and Defendant's Physical Attractiveness on Subjects' Judgments, 7 SEX ROLES 247, 252 (1981) (rapist of unattractive victim less likely to be seen as guilty); Neil Rector et al., The Effect of Prejudice and Judicial Ambiguity on Defendant Guilt Ratings, 133J. SOC. PSYCH. 651, 657 (1993); Billy Thornton, Effect of Rape Victim's Attractiveness in a Jury Simulation, 3 PERSONALITY & SOC. PSYCHOL. BULL. 666, 668 (1977) (attractiveness of victim had no effect on evaluations of her credibility and responsibility, but resulted in longer sentences).

493 See generally Ronald Mazzella & Alan Feingold, The Effects of Physical Attractiveness, Race, Socioeconomic Status and Gender of Defendants and Victim's on Judgments of Mock Jurors, 24 J. APPLIED SOC. PSYCHOL. 1315, 1325 (1994). The authors analyzed results from 80 mock juror studies. They found that victim attractiveness had no significant effects on the judgments of mock jurors. Accord Whatley, supra note 449, at 82, 87, 90.

Victim attractiveness has been used as an independent variable in several studies of mock jurors' decisionmaking in stranger rape cases. In these studies, in addition to the rape scenario and testimony, subjects received a photograph of the victim, showing either an attractive or an unattractive woman. No main effects for attractiveness were found in any of these studies. See Deitz et al., supra note 492; Feild, supra note 487, at 261; Eugenia P. Geldes et al., Perceptions of Rape Victims and Assailants: Effects of Physical Attractiveness, Acquaintance and Subject Gender, 19 SEX ROLES 141, 150 (1988); Marsha B. Jacobson, supra note 492, at 247; Clive Seligman et al., Rape and Physical Attractiveness: Assigning Responsibility to Victims, 45J. PERSONALITY 555 (1977); Nora K. Villemur &Janet Shibley Hyde, Effects of Sex of Defense Attorney, Sex of Juror, and Age and Attractiveness of the Victim on Mock Juror Decision Making in a Rape Case, 9 SEX ROLES 879 (1983). The experimenters always have a large number of independent raters judge the photographs for attractiveness, and often the subjects are also asked to do so.

Jacobson did report an effect for victim attractiveness, finding that it makes subjects more likely to regard the man as guilty, but this effect was significant only at the p < .1 level, which does not meet the usual .05 criterion for significance. Jacobson, supra, at 252. Some of the above studies did find interactions between attractiveness and other independent variables.

494 See, e.g., BROWNMILLER, supra note 1, at 210-55.

495 See generally id.

496 A district attorney told Holmstrom and Burgess that "you can't get a conviction [with a black victim] even if you have a good case." HOLMSTROM & BURGESS, Supra note 4, at 144. Another scholar found that when the victim was black, mock jurors treated black and white offenders equally. But, black offenders who assaulted white women received much more severe sentences than white offenders who did so. J.S. Feild, Juror Background Characteristics and Attitudes Toward Rape: Correlates of Jurors ' Decisions in Rape Trials, 2 LAW & HUMAN BEHAV. 73 (1978).

497 See generally Mazzella & Feingold, supra note 493. The authors found jurors were more lenient toward defendants who were attractive and/or had a high socioeconomic status. Id. at 1319, 1325, 1327. The effect of a defendant's race on jurors' decisions varied across crimes and punishment scenarios. Id. at 1325.

A few studies have dealt with miscellaneous attributes such as the defendant's criminal record (of which jurors sometimes become aware), and his age. See, e.g., R. Barber, Judge and Jury Attitudes to Rape, 7 AUST. & N.Z.J. CRIMINOLOGY 157 (1974).

498 A. Daniel Yarmey, Older and Younger Adults' Attributions of Responsibility Toward Rape Victims and Rapists, 17 CANADIAN. J. BEHAV. Sex. REv. 327 (1985). Young subjects and females perceived the victim as more responsible for the assault when she resisted a welldressed assailant than when she resisted a poorly-dressed assailant. Older subjects and males were unaffected by this factor. Id. at 331.

499 Sheila R. Deitz & Lynne E. Byrnes, Attribution of Responsibility for Sexual Assault: The Influence of Observer Empathy and Defendant Occupation and Attractiveness, 108 J. PSYCHOL. 17, 23 (1981). However, subjects' perceptions of the janitor and scientist defendants differed only when they were both physically attractive, perhaps because subjects found it difficult to believe that a man who was both physically and occupationally attractive would need to resort to force to obtain sex. Id. at 26. Describing the defendant as a janitor or a scientist had no effect on subjects' recommended sentences. Id. at 23. Accord Kahn et al., supra note 481, at 104-05 (mock jurors assigned an auto mechanic and a math teacher similar sentences in a stranger rape case); Kanekar et al., supra note 462 (subjects assigned the same sentence to defendants described as either a manager, a watchman, or a stenotypist regardless of whether victim knew defendant). Of course, jurors may be more inclined to find a high-status defendant innocent, even if they do not discriminate in (mock) sentences of those they consider guilty.

500 NJ. Barnett & H.S. Feild, Character of the Defendant and Length of Sentence in Rape and Burglary Crimes, 104J. SOC. PSYCHOL. 271 (1978); H.S. Feild & NJ. Barnett, Simulated Jury Trials: Students vs. "Real"People as Jurors, 104J. Soc. PsYcHOL. 287 (1978). Cf. Moore et al., supra note 481, at 721 (socially undesirable defendants significantly more likely to be judged guilty in a mock sexual harassment case). But see Kahn et al., supra note 481, at 98 (authors varied the "social attractiveness" of the defendant in a rape case and found no main effects or interactions involving defendant attractiveness on mock jurors' sentencing of the defendant.).

501 LaFree et al., supra note 66, at 36.

502 Id. at 397.

503 Barbara Reskin & Christy Visher, Research Note: The Impacts of Evidence and Extralegal Factors in Jurors'Decisions, 20 L. & Soc'Y. REv. 423, 431 (1986). But this was true only if the evidence of guilt was relatively weak; in "strong cases" the defendant's employment status had no effect. Id. at 434-35.

504 Id. at 431 n.9. Another study found that the defendant's race was not related to verdicts in 38 rape cases, regardless of whether the defense was consent or misidentification. LaFree et al., supra note 66, at 387. Another study, of 124 actual rape cases, found that the defendant's race did not affect the verdict, but that African-American defendants were significantly less likely to plead guilty than white defendants. LaFree, supra note 381, at 844-45. LaFree speculated that the insignificant effect of race on verdicts may have been due to multicollinearity among the variables, or in other words a high correlation between race and other variables, which may be more related to verdicts. Id. at 845.

Some experimental studies have found that the defendant's race does not significantly affect sentencing. Oros & Elman, supra note 21, at 34; Willis, supra note 454. In contrast, in an experimental study using community members as subjects, Feild found that the defendant's race was related to sentence in a case of stranger rape. Feild, supra note 487, at 271 (mock jurors assigned the Black defendant an average sentence of 17.67 years; the White was sentenced to only 11.47 years).

505 Geldes et al., supra note 493, at 14142 (discussing the "beauty is good" theory). For an interesting example of the advantage of good looks, see D. Landy & E. Aronson, The Influence of the Character of the Criminal and His Victim on the Decisions of Simulated Jurors, 5 J. EXPERIMENTAL SOC. PSYCHOL. 141 (1969) (mock jurors in negligent automobile homicide case sentenced the unattractive defendant to a longer prison term than the attractive one). It would be interesting to know whether (1) subjects would discriminate against an unattractive but well-dressed and evidently respectable defendant, as compared with, say, a handsome but proletarian-looking one, and (2) whether the persons who actually commit certain types of crimes are on average less physically attractive than law-abiding citizens. The latter hypothesis seems plausible, given the correlations between looks and affluence, and between certain types of crime and social class. Of course, a study of such matters would raise difficult methodological problems, because those who have been convicted of a certain crime are not necessarily representative of those who have committed it.

506 Geldes et al., supra note 492, at 14142; Jacobson, supra note 493, at 248-49. For reviews of the literature concerning crimes other than rape, see Harold Sigall & Nancy Ostrove, Beautiful but Dangerous: Effects of Offender Attractiveness and Nature of the Crime on Juridic Judgment, 31 J. PERSONALITY & SOC. PSYCHOL. 410 (1975); Mazzella & Feingold, supra note 493.

507 One study found that males are more likely to believe that sex is the motivation for rape, while females are more likely to believe power is the motivation. L.A. Syzmanski et al., Gender Role and Attitudes Toward Rape in Male and Female College Students, 29 Sex ROLES 37, 55 (1993). Although feminists generally reject the traditional explanation in favor of explanations that emphasize violent, nonsexual motivations, see infra text at notes 786-89, Richard Posner defends the traditional view. POSNER, supra note 8, at 106-07.

508 In addition to the studies mentioned in the text, see Jacobson, supra note 492, at 251-53 (stranger rape, misidentification defense: attractive defendant was less likely to be seen as guilty and received a shorter recommended prison term); Rector et al., supra note 492, at 657. But, one study indicates that the advantage of the good-looking defendant is reversed when the subjects have agreed to be impartial. R.M. Friend & M. Vinson, Leaning Over Backwards: Jurors' Responses to Defendants' Attractiveness, 24 J. COMM. 124 (1974). Another study found that although male subjects recommended a shorter prison term for the attractive defendant, female subjects gave him a longer term than the unattractive defendant. Jacobson, supra note 492, at 251-53. Attractiveness has also been found to decrease the sentence suggested by subjects for a hypothetical rapist. Marsha B. Jacobson & Paula M. Popovich, V{ctim Attractiveness and Perceptions of Responsibility in an Ambiguous Rape Case, 8 PSYCHOL. WOMEN Q. 100, !02-03 (1983). Deitz and Byrnes found that female subjects assigned a longer sentence to attractive defendants, while males assigned a longer sentence to unattractive ones. Deitz & Byrnes, supra note 499, at 23-24.

One study found that the effect of defendant attractiveness varies depending on the presence or absence of acquaintance and the observer's gender. Geldes et al., supra note , 493, at 150-51. The perceived likelihood of future anti-social behavior by the perpetrator increased for attractive perpetrators who were acquainted with the victim. Id. 5o9 Subjects, 32 male and 32 female students in introductory social science courses, participated in groups. Within each gender, subjects were randomly assigned to read one of eight versions of a "newspaper article" accompanied by pictures of the rapist and the victim. In each story, the defendant grabbed the woman while she was walking alone at night, raped her in a deserted stairwell, and was caught in the act by a group of students who heard the victim's screams. Stories differed in the presence or absence of prior acquaintance. Victim and defendant were described as average, middle-class students in each story; only their pictures were varied (attractive or unattractive victim, attractive or unattractive defendant). Female subjects considered the victim less responsible when the defendant was unattractive than when he was attractive. Geldes et al., supra note 493, at 144.

510 Id. at 146-47.

511 Id. at 151.

512 Deitz & Byrnes, supra note 499 (mock jurors in stranger rape case gave attractive and unattractive defendants equivalent sentences on average); Geldes et al., supra note 493, at 145 (sentences imposed by mock jurors unaffected by defendant's attractiveness in both acquaintance and stranger rape cases).

513 Reskin & Visher, supra note 503, at 431.

514 Although the general rule in state courts excludes such evidence, it is often admissible under any of several exceptions. See generally Bryden & Park, supra note 310, at 534430.

515 One study found that if the accuser's criminal record was mentioned in court, he was more likely to be found guilty in consent-defense (but not in misidentification-defense) cases. LaFree et al., supra note 66, at 397-400. LaFree's study of rape trials found a highly significant relationship between a defendant's criminal record of forcible sex offenses and guilt. LaFree, supra note 381, at 843, 844 tbl.3. Cf Bryden & Park, supra note 310, at 575-78 (evidence of the defendant's prior rapes has more probative value in acquaintance rape cases than in stranger rapes).

516 See generally Martha R. Burt, Cultural Myths and Supports for Rape, 38 J. PV. PESONALITY & SOC. PSYCHOL. 217 (1980). The Rape Myth Acceptance (RMA) scale is the best-known psychometrically-valid scale assessing beliefs about rape. To construct this scale, Burt generated numerous untrue statements about rape from several sources. She then asked a sample of 598 adults to indicate the extent of their agreement with these myths on sevenpoint Likert-type scales, and employed item analysis to select the 19 most internally-consistent items, resulting in a Cronbach's alpha of .875. Id. at 223. One obtains a composite score for each respondent by adding the scores of individual items, indicating the extent to which the respondent holds favorable or unfavorable attitudes about rape or rape victims.

517 Id. at 222.

518 Id.

519 Id.

520 See generally Feild, supra note 281.

521 See generally Gloria J. Fischer, Cognitive Predictors of Not-Guilty Verdicts in a Simulated Acquaintance Rape Trial, 68 PSYCHOL. REP. 1199 (1991).

522 See generally Sheila R. Deitz et al., Measurement of Empathy Toward Rape Victims and Rap/sts, 43 J. PERSONALITY & Soc. PSYCHOL. 372 (1982).

523 One study found that subjects high in RMA attributed more responsibility to the victim for a rape and less responsibility to the defendant. Barbara Krahe, Victim and Observer Characteristics as Determinants of Responsibility Attributions to Victims of Rape, 18J. APPLIED SOC. PSYCHOL. 50 (1988). Surprisingly, RMA was not related to subjects' willingness to convict in a sample of adults. Burt & Albin, supra note 431, at 222 tbl.2. But subjects higher in RMA were less likely to define a forced sexual encounter as rape. Id. In the final regression model, the authors found that the best predictors of willingness to convict were the subjects' willingness to define the situation as rape, their opinions of the man and the woman in the scenario, and their Acceptance of Interpersonal Violence. Id. at 224 tbl.3. Attitudinal variables (though not RMA) were better predictors of willingness to convict a rapist than any demographic subject variables. Id.

Mock jurors higher in Rape Empathy were found to be more certain of the defendant's guilt in a stranger rape case, assigned him a longer sentence, and believed he was more responsible for the rape than subjects lower in Rape Empathy. Dietz et al., supra note 522, at 380, Subjects high in Rape Empathy had more negative feelings about the defendant, more positive feelings about the victim, and felt that the victim was less responsible than those low in Rape Empathy. Id. Those high in Rape Empathy believed that the psychological impact on the victim was more devastating and that the rape was more serious than did those low in Rape Empathy. Id. Similar results were reported in two other studies. Deitz & Byrnes, supra note 499; Deitz et al., Attribution of Responsibility for Rape, supra note 492.

Muehlenhard measured his subjects' traditionality by the Attitudes Towards Women Scale, which assesses whether subjects tend to believe that women should occupy their traditional roles. Muehlenhard, supra note 467, at 30. Those with more traditional attitudes towards women thought that the man was significantly more justified in forcing the woman to have intercourse over all dating situations, especially if the woman had violated traditional norms by asking the man out, splitting the cost of the date, or going to the man's apartment. Id. Of the more traditional subjects, 29.7% said that rape is sometimes justifiable, compared to 14.7% of the more egalitarian subjects. Id. at 31. In addition, men thought that the woman was more willing to have sex than did female subjects, across all conditions. Id. at 26. Males (especially traditional ones) were more inclined (27% versus 17.5%) to believe that forcible sex is sometimes justifiable. Id. at 28.

524 Feild, Attitudes Toward Rape, supra note 281.

525 See, e.g., Bell et al., supra note 449, at 1726, 1729; Giarrusso et al., supra note 478, at 66 (males more likely than females to consider aggressive sex appropriate in each of nine different circumstances); Kelly McLendan et al., Male & Female Perceptions of Date Rape, 9J. SOC. BEHAV. & PERSONALITY 421,425 (1994); Schult & Schneider, supra note 484, at 99-100. See generally Christine Brems & Patricia Wagner, Blame of Victim and Perpetrator in Rape Versus Theft, 134J. SOC. PSYCHOL. 363 (1994).

Several studies have found that females favor longer sentences for rape than males: Arnie Cann et al., Attributing Responsibility to the Victim of Rape: Influence of Information RegardingPast SexualExperience, 32 HUM. RELATIONS 57, 62 (1979); Feldman-Summers & Lindner, supra note 67, at 141; Geldes et al., supra note 493, at 145; Kanekar et al., supra note 462; Smith et al., supra note 452, at 352. Cf. Richard L. Wiener et al., Empathy and Biased Assimilation of Testimonies in Cases of Alleged Rape, 13 LAW & HUM. BEHAV. 343, 347-54 (1989) (female "jurors" in mock trial said to be more prone to base their judgments on sentiments rather than the legal elements of the case). This was also true of community members (as opposed to student jurors), those who knew a rape victim, and those who did not know a man accused of rape. See also LONSWAY & FITZGERALD, supra note 251, at 142; Richard L. Wiener et al., The Social Psychology of Jury Nullification: Predicting When Jurors Disobey the Law, 21 J. APPLIED SOC. PSYCHOL. 1379 (1991) ("Numerous demographic and background variables have been examined in relation to Rape Myth Acceptance, the only consistent relationship being with the sex of the respondent.").

526 See generally BOURQUE, supra note 25, at 8.

527 See generally id. at 134-40; P. Pollard, Judgments About Victims and Attackers in Depicted Rapes, 31 BRIT. J. SOC. PSYCHOL. 307 (1992). Pollard reviewed research on gender differences in perceptions of acquaintance rape. He found that while there is considerable evidence for the existence of gender differences in perceptions, the studies are not unanimous. See, e.g., HUBERT S. FEILD 8C LEIGH B. BIENEN, JURORS AND RAPE 50-51 (1980). The authors found that men and women give nearly identical answers to questions about the appropriate legal standards in rape cases, for example whether the degree of the woman's resistance should be the major factor in determining whether a rape occurred, the significani:e of a delay in reporting the alleged rape, and the appropriate sentence length. Other questions, such as whether it would do some women good to be raped, and whether a woman's appearance provokes a rapist, elicited somewhat different replies from men and women. Id. In the same authors' study of jurors in actual rape trials, no sex differences were found in the likelihood of voting for conviction. Id. at 121.

Some studies have found that mock jurors do not differ by gender in the sentences they prefer in a stranger rape. Feild, supra note 496; Jones & Aronson, supra note 479; Scroggs, supra note 484. Another study found that males impose harsher sentences than females on rape defendants. Richardson & Campbell, supra note 73.

528 See, e.g., Wayne P. Anderson & Kimberly Cummings, Women's Acceptance of Rape Myths and Their Sexual Experiences, 34 J. C. STUDENT DEV. 53, 56 (1993) ("[w]omen who accepted traditional definitions of women's roles and rape myths were more likely to disbelieve another woman's report that she had been raped"); Feild, supra note 487 (mock jurors' recommended sentences affected by beliefs that women should be responsible for preventing their own rapes, that women provoke rape by their appearance or behavior, that women should do everything they can to resist a rape, that rapists are normal men, and that rapists should not be treated harshly); Muehlenhard, supra note 467; R. Lance Shotland & Lynne Goodstein, Just Because She Doesn't Want to Doesn't Mean It's Rape: An Experimentally Based Causal Model of the Perception of Rape in a Dating Situation, 46 Soc. PSYCHOL. Q. 220, 224-26 (1983) (sexually egalitarian subjects more likely to label the encounter as rape and less likely to blame victim); Snell & Godwin, supra note 483, at 606, 612 (undergraduate female subjects with nontraditional sex-role views saw rapists as more responsible than did traditionalists, especially if the rape occurred in a long-term dating relationship). Willis found that sexually egalitarian subjects showed a stronger belief in the acquaintance rape victim's truthfulness than did traditionalists; attributed less responsibility to her; were more certain she had not consented; were less likely to think she had promoted the incident; perceived her suffering as greater; and perceived the perpetrator as more likely to commit further, similar offenses. Willis, supra note 454, at 218-22. See also Abbey & Harnisch, supra note 483, at 305-06 (male subjects who were high in rape myth acceptance perceived greater sexual intent by a female target than males low in rape myth acceptance).

Gender and values may also interact. One study found that men who were high in rape myth acceptance perceived mundane dating behaviors more sexually (and as more indicative of consent) than low rape myth acceptance men or women high or low in rape myth acceptance. R.M. Kowalski, Inferring Sexual Interest from Behavioral Clues: Effects of Gender and Sexually Relevant Attitudes, 29 SEX ROLES 13, 33 (1993). See also Beverly A. Kopper, Gender, Gender Identity, Rape Myth Acceptance, and Time of Initial Resistance on the Perception of Acquaintance Rape Blame and Avoidability, 34 SEX ROLES 81, 90-91 (1996) (finding that the critical variable for predicting perceptions of rape is level of rape myth acceptance, which is generally lower for women than for men).

529 Jenkins & Dambrot, supra note 478.

530 Id. at 882.

531 Id. Thus, in this study only the female subjects seemed to be affected by the victim's counter-normative behavior.

532 Id.

533 See infra text accompanying notes 874980.

534 See infra note 797 and accompanying text.

535 See generally LAUMANN ET AL., supra note 56, at 194-220.

536 Burt, supra note 516, at 227 tbl.5.

537 One study that used members of the community (with an average age of 35) reported a weak age effect: the older the mock juror (male or female), the more lenient he or she tended to be. This study found no male-female difference. Feild, supra note 496, at 84. Another mock jury study investigated the effect of age on verdicts, using 257 adult subjects, aged 21 to 67. The results of a Chi-square analysis showed a significant effect for age, with those under 25 and those over 40 being the most likely to acquit the rape defendant. A. Philip Sealy, Another Look at Social Psychological Aspects of Juror Bias, 5 LAw & HUM. BEHAV. 187, 192 (1981). However, this age effect was found for only one of two trials used in the study. In the trial showing age and gender effects on verdicts, the defendant had almost no evidence against him. Id. at 193. In the other trial, with a defendant who had incriminated himself in court, no significant age or gender effects were found. Id. at 19293. Another study found older African-American and Mexican-Americans more supportive of rape victims than younger ones, but found no age differences for Whites. Willis, supra note 454, at 75-76.

538 As measured by the responses of high school students to questions about sex roles, endorsement of feminist gender beliefs has increased in recent decades. Jensen & Karpos, supra note 159, at 370.

539 KALVEN & ZEISEL, supra note 52, at 164-65. See also supra notes 432-33 and accompanying text.

540 See Reskin & Visher, supra note 503, at 430-35. Some experimental studies, however, have found that the strength of the evidence did not moderate the impact of stereotypes. See, e.g., Randall A. Gordon, The Effect of Strong Versus Weak Evidence on the Assessment of Race Stereotypic and Race Nonstereotypic Crimes, 23J. APPLIED SOC. PSYCHOL. 734, 747 (1993).

541 Id. at 431, 434-35.

542 Whether unemployed men are in fact more likely to be rapists than those with jobs, and if so whether juries are excessively affected by this factor, are questions that have not been definitively answered.

543 See Reskin & Visher, supra note 503. One study found that while African-Americans and Whites both discriminated in favor of same-race defendants in judging rape scenarios, the African-Americans (but not the Whites) did so even ff the evidence was strong. In other words, they tended to give the African-American defendant the benefit of the doubt even when the evidence against him was powerful. Denis Chimaeze E. Ugwuegbu, Racial and Evidential Factors in Juror Attribution of Legal Responsibility, 15 J. EXPERIMENTAL SOC. PSYCH. 133, 139-42 (1979). But cf. Brems & Wagner, supra note 525, at 373 (subjects attributed more fault to victims in ambiguous situations than in "clear-cut" rapes).

544 LaFree, et al., supra note 66, at 389.

545 MARSH ET AL., supra note 25, at 22-23; Searles & Berger, supra note 25, at 25.

546 SPOHN & HORNEY, supra note 170, at 92-100.

547 Id. at 103 (no increase), 104 (but absence of decline is favorable sign).

548 Id. The percentage of cases in which there was a conviction on the original charge rather than a lesser offense increased but not to a statistically significant degree.

549 Id. at 80.

550 Cassia Spohn &Julie Homey, Rape Law Reform and the Effect of Victim Characteristics on Case Processing, 9 J. QUANTATIVE CRIMINOLOGY 383 (1993). Because all of the cases in their sample were bound over for trial, Spohn and Horney did not examine decisions made by police and prosecutors prior to arraignment in Recorder's Court. Id. at 393 n.8.

551 Id. at 397-98.

552 Id. at 398.

553 P = 0.001. Id.

554 p = 0.01. Id.

555 = 0.001. Id.

556 Id. at 398-99.

557 Id. at 405. The only exceptions were that (1) in the pre-reform period, prosecutors seem to have been more willing to reduce charges as part of a plea bargain, where the victim did not report the crime promptly, id. at 400, 405, and (2) that the parties were strangers increased the likelihood of incarceration in the pre-reform period and had no effect in the post-reform period. Id. at 403, 405.

558 A different result was reached by Bachman and Paternoster, who sought to determine, for two time periods, whether there were differences between the situations in which rape occurred and the situations in which it led to incarceration. For this purpose, they compared NCVS data concerning acquaintance rapes with national prison statistics concerning the offenses committed by the convicted rapists. During the 1979-1986 period, the NCVS indicated that 41% of rapes had been committed by acquaintances, 14% by relatives, and 45% by strangers. Bachman & Paternoster, supra note 153, at 570 tbl. I. National prison statistics for 1986, by contrast,, revealed that only 29% of those incarcerated for rape had raped an acquaintance, 15% a relative, and 56% a stranger. Id. Similar data for 19871990 (NCVS) and 1991 (National Prisoner Statistics) showed a greater correspondence between characteristics of offenders and characterstics of prisoners, with 48% of the offenders having raped an acquaintance (and 16% a relative), while 40% of the prisoners had raped an acquaintance, and 17% a relative. Id. Concerning these findings, several caveats: First, the NCVS data, based on household surveys, may understate rapes by relatives. Second, Since the NCVS relied on self-classification by victims, id. at 561 n.30, it probably understated acquaintance rapes, especially by lovers. See Koss, supra note 170, at 4. Third, because state reforms have occurred at different times "it would be virtually impossible to establish a single time point by which to distinguish between pre- and postreform periods." Bachman & Paternoster, supra note 153, at 564. If the authors had chosen an earlier period for their "pre-reform" data, the change might have been much more dramatic. Fourth, although the authors surmise that the changes "may" have been due to reforms, id. at 571, they offer no evidence to that effect. Attitudinal changes on the part of acquaintance rape victims and justice system officials are an alternative hypothesis. Finally, the assumption that prisoner data ought to correspond to NCVS data is true only if the corroborative evidence is on average equally good for acquaintance and stranger rapes.

559 MARSH ET AL., supra note 25, at 44-45.

560 Id. at 29.

561 Id. at 43-45. This conclusion was supported both by interviews with criminal justice system participants and by time-series analyses. Id.

562 SPOHN & HORNEY, supra note 170, at 104.

563 Id. at 70-71.

564 Id. at 36-41.

565 Id. at 160.

566 K. Gilchrist & J. Homey, Assessing the Impact of Changes in the Nebraska Rape Statutes: Effect on Prosecution, cited in SPOHN & HORNEY, supra note 170, at 31.

567 Loh, supra note 275, at 591-93.

568 One scholar contends that reforms seldom achieve their instrumental goals and serve chiefly to provide visibility and legitimacy to certain attitudes and values. See JOEL F. HANDLER, SOCIAL MOVEMENTS AND THE LEGAL SYSTEM (1978).

569 A respected social scientist goes so far as to say that the "overriding" objective of reforms was to "move the burden of proof from the victim to the defendant." BOURQUE, supra note 25, at 110. We doubt that many reformers literally believed this, but some reformers do seem to have believed that shield laws would so sharply restrict attacks on the victim's character that the defense, abandoning its former aggressive tactics, would be forced to treat the victim respectfully, and to focus, with the prosecution and the jury, on the defendant's conduct rather than that of his victim. It would be hard to imagine a better scenario for conviction, which explains both its attractiveness and its unrealism.

570 MARSH ET AL., supra note 25, at 52.

571 See, e.g., HOLMSTrOM & BURGESS, supra note 4, at 279; Galvin, supra note 4, at 812-905.

572 See generally Galvin, supra note 4, at 807-12.

573 Seades & Berger, supra note 25, at 26.

574 Ronald J. Berger et al., The Impact of Rape Law Reform: An Aggregate Analysis of Police Reports and Arrests, 19 CRIM. JusT. REv. 1, 1 (1994).

575 Perhaps this is analogous to what Professor Vivian Berger meant when she suggested that, "[s] ex-neutral redefinitions of 'rape' (or criminal sexual conduct in general) enhance the protection of victims of homosexual attack while subtly freeing this branch of the law from outmoded notions harmful to women." Berger, supra note 1, at 12.

576 KALVEN & ZEISEL, supra note 52, at 250-54.

577 Kalven and Zeisel point out that in the simple rape cases where the jury was given the option of convicting the defendant of a lesser charge, they found the defendant guilty of the lesser charge in all but one. ld. Of course, instructing the jury on lesser crimes might have an effect on juries that simply reducing the statutory penalty for rape would not have, either because juries are influenced less by the potential sentence than by their notion of which rapes deserve that label and its stigma, or because they do not know the statutory maximum penalty, but do understand that it is higher than for lesser sex crimes.

578 In Michigan, several radical feminists disassociated themselves from the rape law reform effort, reportedly because they felt that the males who would administer the law would prevent it from being efficacious. MARSH ET AL., supra note 25, at 17. To the extent that the system's performance has improved, albeit not necessarily because of particular law reforms in individual jurisdictions, they too were unrealistic.

579 Oros & Elman, supra note 21, at 34.

580 SPOHN & HORNEY, supra note 170, at 24-25, 163-64. In New York, however, the exceptionally strict corroboration requirement appears to have been a major obstacle to convictions. See Irving Younger, The Requirement of Corroboration in Prosecutions for Sex Offenses in New York, 40 FORDHAM L. REv. 263, 268 (1972).

581 ESTRICH, supra note 1, at 43; Younger, supra note 580, at 264.

582 See, e.g., DonaldJ. Friedman, Note, The Rape Corroboration Requirement: Repeal Not Reform, 81 YALE L.J. 1365, 1367-70 (1972).

583 See SPOHN & HORNEY, supra note 170, at 161454. Commenting on repeal of the California corroboration requirement, a Los Angeles deputy prosecutor remarked: "Legal theory is not legal reality.., and in California just like anywhere else in this country, a woman who hopes to win a rape case better have plenty of corroboration." Chappell & Singer, supra note 225, at 266. See also Feild, supra note 487 (mock jurors whose version of the case included strong corroborative evidence--torn clothing and scratches--were more likely to recommend a long prison sentence).

584 In a jury simulation study, mock jurors were given a rape case containing the same evidence but various corroboration instructions. Juries that had received corroboration instructions tended to be less likely to convict the defendant of rape. Analysis of the recorded jury deliberations revealed that jurors who had received corroboration instructions were not more likely to scrutinize the putative victim's credibility or the corroborative evidence. Rather, the jurors who had received such instructions seemed to have a generally more positive opinion of the defendant (and a less positive opinion of the woman). VALERIE P. HANS & NEIL VIDMAR, JUDGING THE JURY 207 (1986).

585 See, e.g, BROWNMILLER, supra note 1, at 385;John Dwight Ingram, Date Rape: It's Time For "No" to Really Mean "No", 21 AM.J. CRIM. L. 3, 10-13 (1993); Lani Anne Remick, Comment, Read Her Lips: An Argument for a Verbal Consent Standard in Rape, 141 U. PA. L. REv. 1103, 1110-13 (1994).

586 ESTRIGH, supra note 1, at 58-66.

587 See SPOHN & HORNEY, supra note 170, at 161-64; Judith E. Krulewitz &Janet E. Nash, Effects of Rape Victim Resistance, Assault Outcome, and Sex of Observer on Attributions About Rape, 47 J. PERSONALITY 557, 562-64 (1979) (as victim resistance increased, subjects became more certain that rape had occurred); Shodand & Goodstein, supra note 528 at 224 (when man used low degree of force subjects judged woman's desire for sex as greater than when moderate force was used). In addition to interpreting the absence of force and resistance as evidence of consent, jurors may feel that in low force situations the woman, though not necessarily consenting, could have averted the rape and so is partly responsible for her own victimization. See Bell et al., supra note 449, at 1720; Kopper, supra note 528, at 89 (when initial resistance occurred early in the encounter, subjects attributed significantly less blame to the victim and situation and more blame to the perpetrator). To the extent that jurors adopt either of these positions, the degree of force and resistance will be influential irrespective of whether resistance or force is a legal element of the crime.

588 SPOHN & HORNEY, supra note 170, at 160-61.

589 Id. at 161.

590 Spohn and Horney leave this up in the air. Compare id. at 161 (no increase in conviction rate in Detroit despite definitional changes), with id. at 104 (more borderline cases in the system post-reform, hence absence of decline in conviction rate seems to mean convictions are easier to obtain).

591 Id. at 161.

592 BOURQUE, supra note 25, at 109.

593 See generally GERALD N. ROSENBERG, THE HOLLOW HOPE (1991).

594 In the battle over conservatives' efforts to overrule Roe v. Wade, for example, both sides seemed to believe that the clock could be rolled back to the era when legal abortions were extremely rare.

595 See, e.g., David P. Bryden, Environmental Rights in Theory and Practice, 62 MINN. L. REV. 163, 163-76 (1978). Although the accusation of legal unrealism is directed most often at liberals, conservatives are often vulnerable to the same charge. For example, consider the National Review's farewell to Justice Brennan:

In areas as diverse as criminal justice, civil rights, federalism, pornography, the role of religion in public life and education policy, we are governed by his decisions. And such decisions have had consequences: urban decay, racial polarization, soaring violent crime, the growth of centralized government, deteriorating public schools, and loss of community.

NATIONAL REVIEW, Sept. 1, 1997, at 19-20. For an excellent discussion of the unrealism of some conservative legal nostrums, see Charles Krauthammer, A Social Conservative Credo, 121 THE PUBLIC INTEREST 15 (1995).

596 See, e.g., Kimberle Crenshaw, Mapping the Margins: Intersectionality, Identity Politics, and Violence Against Women of Color, 43 SWAN. L. REV. 1241, 1275 (1991) ("rape laws serve to regulate the sexual conduct of women by withholding from rape victims the ability to invoke sexual assault law when they have engaged in nontraditional behavior"). As Crenshaw states: "Rape law.., serves not only to penalize actual examples of nontraditional behavior but also to diminish and devalue women who belong to groups in which nontraditional behavior is perceived as common." Id. at 1280. Although Estrich says that "It]he problem has never been so much the terms of the statutes as our understanding of them," ESTRICH, supra note I at 90, she seems to be talking about judicial interpretations rather than, for example, juries' verdicts. Be that as it may, her book is almost entirely devoted to criticizing judicial decisions and legal doctrines. This of course is a legitimate enterprise, but such criticisms of the judiciary do have the unfortunate side effect of implying that judicial decisions are the main obstacle on the road to justice in rape cases.

597 Most public law scholars are liberals; as a result, most examples of the can-do creed in public law scholarship involve liberal causes. See, e.g., David P. Bryden, Brandeis's Facts, 1 CONST. COMM. 281, 281-83 (1984) (recounting fatuous praise for Brandeis's briefs). But conservative scholars display the same symptoms. Rather than describing liberal reforms as ineffective, they usually prefer to warn that liberal panaceas will have terrible consequences. As for the conservative agenda, what liberal hope is more forlorn than the idea that the decline of religion can be arrested by allowing school prayers, or that abolition of Miranda warnings would substantially reduce crime?

598 Perhaps part of the difference between juries' verdicts and judges' "verdicts" was that the former but not the latter affected the defendant's fate. It is easier to be severe in principle than in practice. Of course, this would not account for the differences in net jury leniency among various crimes. See supra notes 408-20 and accompanying text.

599 A fourth assumption is that there are no significant differences in the appropriate legal response to rapes by lovers or acquaintances and rapes by strangers, nor between rapes of women who were figuratively "asking for it" and those who behaved with reasonable prudence. This assumption can best be evaluated in a separate article.

600 See, e.g., BROWNMILLER, supra note 1, at 387 (asserting that the percentage of false rape reports is the same as for other crimes); Patricia Frazier & Eugene Borgida, Juror Common Understanding and the Admissibility of Rape Trauma Syndrome Evidence in Court, 12 LAw & HuM. BEHAV. 101, 106-07 (1988) (assessing sparse data about false reports of rape and concluding that the rate of false reports is the same or lower than for other crimes); Goolsby, supra note 105, at 1189-91 (false reports of rape are estimated at about 2%, the same as for other crimes); Morrison Torrey, When Will We Be Believed?: Rape Myths and the Idea of a Fair Trial in Rape Prosecutions, 24 U.C. DAviS L. REV. 1013 (1991) (relying on estimated 2% rate of false reports to argue that the fresh complaint rule is out-of-date).

601 See, e.g., Goolsby, supra note 105, at 1188-92 (attributing high attrition rates for simple rape cases to male biases, prior relationship issues, force requirements, corroboration requirements, and fresh complaint requirements, but not to the burden of proof); Henderson, supra note 105, at 4143. See also supra notes 712-17 and accompanying text.

602 Privacy of Rape Victims: Hearings on H.R. 14666 Before the Subcomm. On Criminal Justice of the House Comm. On the Judiciary, 94th Cong. 36 (1976) (statement of Mary Ann Largen) (declaring that no "right-minded citizen" could believe that the rape complainant's prior sexual activities are relevant); BROWNMILLER, supra note 1, at 370-71 (discussing the admissibility into evidence of prior sexual conduct and criticizing the conclusion "[t]hat a reputation of 'loose moral character' probably has a basis in fact and that a girl with such a character is more likely than not to consent to intercourse in any given instance"); Sakthi Murphy, Comment, Rejecting Unreasonable Sexual Expectations: Limits on Using a Rape Victim 's Sexual History to Show the Defendant's Mistaken Belief in Consent, 79 CAL. L. REV. 541, 552 (1991) (declaring a consensus that a woman's sexual past has low probative value, if any value at all, in a rape trial).

603 See, e.g., ESTRICH, supra note 1, at 4346; Goolsby, supra note 105, at 1188-91; Torrey, supra note 600, at 1027-31.

604 See, e.g., ESTRICH, supra note 1, at 4346; Goolsby, supra note 105, at 1188-91; Torrey, supra note 600, at 1027-31.

605 See generally AMIR, supra note 292, at 254; ESTRICH, supra note 1, at 5-6.

606 1 M. HALE, THE HISTORY OF THE PLEAS OF THE CROWN 635 (1778). Hale's admonition is not literally false: it is indeed often difficult to prove that the defendant raped the putative victim and also presumably sometimes difficult for an innocent defendant, particularly in a nonaggravated acquaintance rape case, to establish his innocence to the satisfaction of an unbiased third party. But bias and the burden of proof usually solve the defendant's problem while compounding the acquaintance rape victim's problem. Hale's failure to mention this should be evaluated within his seventeenth-century English context, which included not only the death penalty for rape but lack of a right to counsel and the rule prohibiting defendants from testifying on their own behalf. See generally Berger, supra note 1, at 31.

607 See 1 H. DEUTSCH, THE PSYCHOLOGY OF WOMEN (1944); S. FREUD, NEW INTRODUCTORY LECTURES ON PSYCHOANALYSIS (1933); K. HORNEY, FEMININE PSYCHOLOGY (1933). For a critical evaluation of this literature, see S. BESSMER, THE LAWS OF RAPE (1984); S. EDWARDS, FEMALE SEXUALITY AND THE LAW (1981); S. Edwards, Sexuality, Sexual Offenses, and Conceptions of Victims in the Criminal Justice Process, 8 VICTIMOLOGY 113 (1983); E.J. Kanin, Female Rape Fantasies: A Victimization Study, 7 VICTIMOLOGY 114 (1982).

608 M. GuTTMACHER & H. WEIHOFEN, PSYCHIATRY AND THE LAW 360-63 (1952); K.A. Menninger, (Menninger Clinic of Psychiatry & Neurology, Topeka, Kansas) M.S. letter (Sept. 15, 1933), cited in 3A WIGMORE, EVIDENCE IN TRIALS AT COMMON LAW (Section) 924(a) (James H. Chadbourn rev., 1970).

609 3A WXGMORE, supra note 608, (Section) 924(a); R.A. Hibey, The Trial of a Rape Case: An Advocate's Analysis of Corroboration, Consent, and Character, 11 AM. CRIM. L. REv. 309 (1973); M. Juliver, Psychiatric Opinions as to Credibility of Witnesses: A Suggested Approach, 48 CAL. L. REV. 648 (1960).

610 R.V. Sherwin, Commentary, 7 MED. ASPECTS HUM. SEX. 193 (1973). See also GUTTMACHER & WEIHOFEN, supra note 608.

611 3A WIGMORE, supra note 608, (Section) 924(a). Citing the discoveries of "modern psychiatry," Wigmore warned that women's testimony is untrustworthy:

Their psychic complexes are multifarious, distorted par fly by inherent defects, partly by diseased derangements or abnormal instincts, partly by bad social environment, partly by temporary physiological or emotional conditions. One form taken by these complexes is that of contriving false charges of sexual offenses by men .... Judging merely from the reports of cases in the appellate courts, one must infer that many innocent men have gone to prison because of tales whose falsity could not be exposed.

Id. at 736.

612 See supra notes 18-20.

613 When Michigan's rape law reforms were being considered by the state legislature, some male legislators expressed concern that vindictive wives would use the new law to win property settlements in divorce proceedings. To satisfy these critics, and improve the bill's prospects, the reformers agreed to a spousal exemption. See MARSH ET AL., supra note 25, at 15. Perhaps the legislators were thinking not so much of false accusations as of true accusations used to extort favorable settlements.

614 For a brief but cogent response to the idea that many women have deep-seated masochistic needs that lead them, subconsciously, to seek situations in which there is a high risk of rape, see Berger, supra note 1, at 27-29.

615 See, e.g., FAIRSTEIN, Supra note 13, at 15-16.

616 See, e.g., ALLISON & WRIGHTSMAN, supra note 56, at 205 (2%); BROWNMILLER, supra note 1, at 386-87; SEDELLE KATZ & MARY ANN MAZUR, UNDERSTANDING THE RAPE VICTIM 209 (1979) (citing unpublished study by Carolyn Hursch & James Selkin); Harry J. O'Reilly, Crisis Intervention with Victims of Forcible Rape: A Police Perspective, in PERSPECTIVES ON RAPE AND SEXUAL ASSAULT 89, 96-97 (June Hopkins ed., 1984) [hereinafter PERSPECTIVES]; Torrey, supra note 600, at 1028 (asserting that rape complaints are as likely to be true as reports of any other crime).

617 For a particularly nice example of this fallacy, see Bryden & Park, supra note 310, at 577. The argument presented by Bryden and Park, however, did not depend on the assertion that few false rape reports are filed, but only on the far safer assumption that few false rape reports survive screening by police and prosecutors, as well as numerous opportunities for the woman to change her mind or conclude that her lie has achieved its purpose.

618 See HOLMSTROM & BURGESS, supra note 4, at 63-156 (describing the inevitably somewhat and sometimes highly unpleasant processes of the hospital examination and the police investigation).

619 See Galvin, supra note 4, at 764-66.

620 Perjury prosecutions are rare. Barbara A. Babcock, Taking the Stand, 35 WM. & MARY L. REv. 1, 9 (1993). But presumably few nonlawyers know this. The woman also runs the risk of being prosecuted for filing a false police report. See, e.g., MINN. STAT. ANN. (Section) 609.505 (West 1994).

621 Even the author of a study finding an extremely high rate of false rape complaints states that there was no evidence that false complaints occurred as a result of delusions. Kanin, supra note 317, at 85.

622 This may be what happened in People v. Olsen, 685 P.2d 52 (Cal. 1984).

623 See RAPE IN AMERICA, supra note 56. According to one authority, over half of all reported rape victims are under 20; three-quarters under 26. AMIR, supra note 292, at 51-52 & tbl.8; D. MULVIHILL & M. TUMIN, supra note 275, at 212 & tbl.3.

624 HOLMSTROM & BURGESS, supra note 4, at 33. Of 23 victims aged 16 and under, only two had contacted police on their own initiative. Id. at 31 tbl. I. Of 94 victims aged 17 or older, only 22 had done so. Id.

625 Id. at 31-33. Another study found that of 761 rapes reported to police, the victim was the one who contacted the police in only 23% of cases, followed by her mother or stepmother (22.1%), hospital authorities (12.2%), and only small percentages of a long list of other categories of reporters. MCCAHILL ET AL., supra note 52, at 83. See generally Richard O. Hawkins, Who Called the Cops: Decisions to Report Criminal Victimization, 7 LAw &Soc. REv. 427 (1973).

626 See infra note 940 and accompanying text. Nonpayment of the prostitute's fee is not deemed to vitiate her consent to the sexual act, and therefore the customer is not guilty of rape merely because of this misdeed. See generally WHARTON'S supra note 44, (Section) 283 (consent vitiated in certain instances of fraud). At common law, consent to intercourse is vitiated if the fraud is in the facturn. DRESSLER, supra note 62, (Section) 33.05(c).

627 See supra note 337 and accompanying text.

628 See, e.g., BROWNMILLER, Supra note 1, at 8.

629 For an unsupported assertion to the contrary, see Jackson, supra note 29, at 16 (observing that "[a]ll the evidence suggests that Mr. Average rapes Ms. Average"). At least with respect to incarcerated serial rapists, "Mr. Average" is a patently inaccurate characterization. For example, one study of 38 rapists incarcerated in 12 states found that 27 (71%) reported "chronic" stealing and shoplifting as children or adolescents; "many" had broken into nearby homes; 63% reported temper tantrums/hyperactivity; a like number acknowledged "alcohol abuse"; 55% had assaulted adults; 54% were chronic liars; and 62% reported "isolation/withdrawal." Janet I. Warren et al., Serial Rape: The Offender and His Rape Career, in RAPE AND SEXUAL ASSAULT III: A RESEARCH HANDBOOK 275, 289 (Ann Wolbert Burgess ed., 1991). Of course, much depends on how one defines "average." When a serial killer is arrested, it is not uncommon for his neighbors to describe him as apparently ordinary, but no reputable scholar has suggested that therefore serial killers are "Mr. Average." In saying this, we do not mean to suggest that date rapists are as psychologically deviant as serial killers, but only that the concept of "Mr. Average" should not be taken literally. See infra note 636.

630 The National Victim Center study found that most forcible rapes occurred in childhood or adolescence: a shocking 29% when the victim was less than 11 years old; another 32% between the ages of 11 and 17; 22% between 18 and 24; 7% between 25 and 29; and only 6% when the victim was older than 29. RAPE IN AMERICA, supra note 56, at 3. See also AMIR, supra note 292, at 61 (alleged victims in Philadelphia were generally "below the age of marriage"). Another study, limited to incarcerated serial rapists, found that the victims' average age for the rapist's first offense was 22.8 years, followed by 26.1 and 24.4 for the second and last offenses, respectively. Warren et al., supra note 629, at 294.

631 See infra text accompanying notes 914-20.

632 AMIR, supra note 292, at 116-17. Amir found that 18.5% of Philadelphia rape complainants had a record of juvenile delinquency.

633 See Kanin, supra note 317, at 89.

634 For a description of the "rewards" that rapists derive from their crime, see Diana Scully &Joseph Marolla, "Riding the Bull at Gilley's": Convicted Rapists Describe the Rewards of Rape, in RAPE & SOCIETY, supra note 27, at 58. In our view, these rewards do not gainsay the irrationality of the act, judged by ordinary standards of prudence. After all, many forms of imprudent behavior, for instance speeding, excessive drinking, and compulsive gambling also have large psychic rewards, as do various crimes. See, e.g., NICHOLAS PILEGGI, WISEGUYS 19 (1985). Violent criminals tend to be remarkably fearless compared to the ordinary man. See generally LYKKEN, supra note 165.

635 See, e.g., MINN. STAX. ANN. (Section) 244 (West Supp. 1995) (sentencing guidelines grid proriding for a presumptive 98-month sentence for a prior sexual offender convicted of criminal sexual conduct in the first degree).

636 Since convicted rapists are not a representative sample of all rapists (e.g., minorities, extremely violent rapists and stranger rapists are over-represented, see Judith Lewis Herman, Considering Sex Offenders: A Model of Addiction, in RAPE & SOCIETY, supra note 27, at 78), the characteristics of the average rapist cannot be reliably determined by studying convicts. But several studies of reported (or serf-acknowledged) rapists reveal ways in which they differ, statistically, from men in general. One study found that rapists are more hostile toward women than nonrapists. N.M. Malamuth et al., Characteristics of Aggressors Against Women: Testing a Model Using a National Sample of College Students, 59 J. CONSULTING CLINICAL PSYCH. 670 (1991). Boys who commit sexual assaults are likely to belong to a peer group that accepts all forms of interpersonal violence. SUZANNE AGETON, SEXUAL ASSAULT AMONG ADOLESCENTS 106-12 (1983). Young men who acknowledge a completed or attempted rape are more likely than other men to endorse "rape-supportive attitudes."

A Swedish study found that one-quarter of all accused rapists were unemployed, and that "close to 70%" had an arrest record, "as opposed to an expected 10%." Snare, supra note 234, at 198. Since rape, especially acquaintance rape, is grossly underreported, the characteristics of accused rapists are not necessarily representative of the characteristics of all offenders. It would not be surprising, therefore, if the proportion of unemployed men among actual rapists were much lower than among reported rapists. See Herman, supra note 630, at 79.

637 See supra text accompanying notes 331-37.

638 A California police manual described rape as one of the most often falsely reported crimes, and claimed that most second-day reported rapes never happened. BROWNMILLER, supra note 1, at 364. Holmstrom and Burgess found that some Boston police have reached similar conclusions. HOLMSTROM & BURGESS, Supra note 4, at 50. It seems likely, however, that some police confuse the question whether a rape occurred with the question whether the victim was contributorily negligent or provocative. For example, Adler quotes an English policewoman who told a rape victim that "she had been in the force for seven years, and not once had she come across a genuine rape case. We get a lot of people who claim to have been raped, girls who hitchhike and ask for trouble and then are surprised when they get it .... "ADLER, supra note 48, at 5. Police skepticism may be less common today than it was twenty or thirty years ago, prior to the modern feminist movement. See, e.g., Jensen & Karpos, supra note 159, at 373.

639 See supra text accompanying notes 331-37.

640 See supra text accompanying notes 236-37.

641 See supra text accompanying notes 169-70.

642 See supra text accompanying notes 157-70.

643 BROWNMILLER, SUpra note 1, at 387.

644 One survey indicates that the police surgeons of Great Britain believe both that false rape reports are common (31.4% of all reports) and that rape is vastly under-reported (10:1 ratio of unreported to reported rapes). Robley Geis et al., Police Officer or Doctor?: Police Surgeons'Attitudes and Opinions About Rape, in PERSPECTIVES, supra note 610, at 58-59.

645 Cf LYKKEN, SUpra note 165.

646 KATZ & MAZUR, supra note 616, at 209.

647 Kanin, supra note 317, at 84-85.

648 ALAN DERSHOWITZ, THE ABUSE Excuse 275 (1994).

649 FEDERAL BUREAU OF INVESTIGATION, U.S. DEP'T OF JUSTICE, CRIME IN THE UNITED STATES, 1995 UNIFORM CreME REPORTS 24 (1996); 1994 UCR, supra note 108, at 24; FEDERAL BUREAU OF INVESTIGATION, U.S. DEP'T OF JUSTICE, CRIME IN THE UNITED STATES, 1993 UNIFORM CreME REPORTS 24 (1994); 1992 UCR, supra note 161, at 24; FEDERAL BUREAU OF INVESTIGATION, U.S. DEP'T. OF JUSTICE, CreME IN THE UNITED STATES, 1991 UNIFORM CreME REPORTS 24 (1992).

650 FEDERAL BUREAU OF INVESTIGATION, U.S. DEP'T OF JUSTICE, CRIME IN THE UNITED STATES, 1974 UNIFORM CreME REPORTS 10 (1975) [hereinafter 1974 UCR].

651 See supra notes 525-28 and accompanying text.

652 See supra notes 234-35 and accompanying text.

653 See supra note 523 and accompanying text.

654 See, e.g., HOLMSTROM & BURGESS, supra note 4, at 43; MCCAHILL ET AL., supra note 52, at 116.

655 See, e.g., HOLMSTROM & BURGESS, supra note 4, at 39 (negative police reaction to rape complainant who went to stranger's apartment to drink beer); id. at 40 (negative reaction to teenage girls who had considerable freedom to be out on their own late at night and were considered promiscuous as well as drunk); id. at 41 (rapes of prostitutes not considered "real"); McCAHILL ET AL., supra note 52, at 112 (police often disbelieve young blacks' allegations of gang rape); id. at 116 (police more likely to unfound if complainant is a welfare recipient). A police officer can express sympathy for the victim's situation, listen to her claim seriously, not make any derogatory remarks, "but still not see it as a rape case worth pursuing through the courts." HOLMSTROM 8(: BURGESS, supra note 4, at 41. Similarly, Adler quotes an English policewoman who regards rapes of hitchhikers as not "genuine" because such women are "asking for it." ADLER, supra note 48, at 5. 656 The FBI's instructions for local police, though not always perfectly clear, generally seem to equate "unfounded" with "false." The UCR says that forcible rape complaints, like reports of other crimes, are unfounded if the police determine that they are "false or baseless." 1992 UCR, supra note 161, at 24. Complaints of crime that are determined through investigation to be unfounded or false are eliminated from an agency's count. Id. Thus, if the police believe a rape occurred, but the victim refuses to cooperate because of fear of embarrassment, the police should found her complaint. FEDERAL BUREAU OF INVESTIGATION, UNITED STATES DEP'T OF JUSTICE, UNIFORM CRIME REPORTING HANDBOOK 11 (1984). But when we asked an FBI official what the police should do if they believe that the report is true, but the woman would not be a credible witness, he said that they should unfound her report. Telephone Interview with George Dixon, Training Instructor, Uniform Crime Reports, Crim. Just. Info. Services Division 0an. 27, 1995).

657 Kanin, supra note 317, at 81-82; ESTRICH, supra note 1, at 15-16.

658 Kanin, supra note 317, at 89 ("[s]ome of these [police unfounding] policies are really nothing more than statistical and procedural legerdemain").

659 For a poignant example of such a case, see WIEHE & RICHARDS, supra note 227, at 32-33.

660 Estrich, Rape, supra note 18, at 1159 (stating that prosecutors carefully consider the rape victim's credibility when deciding whether to charge someone with rape); Kerstetter, supra note 255, at 282 (describing prosecutorial consideration of whether to file a felony charge, which is based, in part, on the victim's credibility as a witness).

661 Kerstetter concluded that founding decisions in Chicago were strongly influenced by evidentiary variables that related more to whether the crime could be proved than to whether it occurred--for example, the presence of a witness, the perpetrator's use of a weapon, and the complainant's willingness to prosecute. Kerstetter, supra note 255, at 267. Of course, some of these variables do offer additional assurance that the complainant is telling the truth, but their absence is not indicative of a false report.

662 In Philadelphia, for example, a police investigator may discuss the case with the prosecutor if he or she is unsure of what the charge should be. MCCAHILL ET AL., supra note 52, at 94.

663 In 1974, the FBI reported that 15% of rape complaints were unfounded. 1974 UCR, supra note 650, at 10. The following data give an idea of the changing unfounded rate in the 1980s: 1983: 10.1 1984: 9.1 1985: 8.6 1986: 9.0 1987: 9.0 1988: 8.9 1989: 8.4 1990: 8.6 1991: 7.6

Unfounded Offenses 1983-1991, In-House Table, FFDERAL BUREAU OF INVESTIGATION, UNITED STATES DEP'T OF JUSTICE (1992). In 1992 and 1993, the unfounded rate steadied out at an estimated 8%. See sources cited supra note 649. The discrepancy between the 7.6% figure shown at the table above for 1991, and the 8% figure at supra note 639 for 1991 is probably due to rounding by the department.

664 See MCGILL ET AL., supra note 52, at 108.

665 Stephen Buckley, Unfounded Reports of Rape Confound Area Police Investigators, WASH. POST, June 27, 1992, at B1.666 Id.

667 See supra notes 331-37 and accompanying text.

668 Candy J. Cooper, Oakland Admits 184 Rapes Ignored, S.F. EXAMINER, Feb. 1, 1991, at A1.

669 Id. Police in Berkeley, California, "take every woman's case so seriously that not one [in 1989] was found to be false." Candy J. Cooper, Berkeley Unit Takes All Cases as Legitimate, S.F. EXAMINER, Sept. 16, 1990, at A10. During the same year, 24.4% of Oakland's rape cases were classified as unfounded. CandyJ. Cooper, Nowhere to Turn for Rape Victims: High Proportion of Cases Turned Aside by Oakland Police, S.F. EXAMINER, Sept. 16, 1990, at A1.

670 McCAHILL ET AL., supra note 52, at 112-13, 115. In Philadelphia, the "true" rate of police unfounding (50%) was once estimated to be far in excess of the official rate (20%). Comment, Police Discretion and the Judgement that a Crime Has Been Committed--Rape in Philadelphia, 117 U. PA. L. REv. 277, 279 n.8 (1968).

671 Arthur Frederick Schiff, StatisticalFeatures of Rape, 14J. FORENSIC SCI. 102, 108 (1969).

672 The authors sent questionnaires to police surgeons attending the annual meeting of the Association of Police Surgeons of Great Britain, obtaining 54 responses; another 74 responded to an inquiry by mail. Geis et al., supra note 644, at 57.

673 Id. at 58.

674 Id.

675 Id.

676 Id. A similar conclusion was reached in a recent Scottish study which analyzed the outcome of 196 reported incidents of sexual assault, finding that the police had classified 22% of them as "no crime," a "very much higher" rate than for other crimes. ADLER, supra note 48, at 5-6.

677 Geis et al., supra note 644, at 58.

678 Id. For a summary of other, equally inconclusive British studies, see ADLER, supra note 48, at 25.

679 REPORT OF THE NEW YORK CITY RAPE ANALYSIS SQUAD, cited in POLLY PATTULLO, JUDGING WOMEN 18 (1983).

680 BROWNMILLER, supra note 1, at 366.

681 MCCAHILL ET AL., supra note 52, at 113.

682 Id.

683 KATZ & MAZUR, supra note 616, at 209.

684 See id.; Hugsell, supra note 128, at 84.

685 Kanin, supra note 317, at 81.

686 Id. at 84.

687 Id. at 83.

688 Id. at 85.

689 Id. Kanin discusses the possibility that the recantations were false and due to the complainants' desire to avoid aggressive police investigations. He thinks this unlikely:

First, with very few exceptions, these complainants were suspect at the complaint or within a day or two after charging. These recantations did not follow prolonged periods of investigation and interrogation that would constitute anything approximating a second assault. Second, not one of the detectives believed that an incident of false recantation had occurred. They argued, rather convincingly, that in those cases where a suspect was identified and interrogated, the facts of the recantation dovetailed with the suspect's own defense. Last, the policy of this police agency is to apply a statute regarding the false reporting of a felony. After the recant, the complainant is informed that she will be charged with filing a false complaint, punishable by a substantial fine and a jail sentence. In no case, has an effort been made on the part of the complainant to retract the recantation.

Id.

690 Id. at 84.

691 Id. at 88.

692 Id. at 85-87. Kanin stresses that his characterizations of the women's motives were not speculative; they were based on the women's own recorded reasons for having filed false charges. Id. at 85.

According to four experienced detectives in a sex crimes unit, the motives for false reports of rape, in order of frequency, are: (1) to provide an alibi or excuse; (2) revenge; (3) financial or other gain; and (4) attention and sympathy. MACDONALD, supra note 178, at 87. Examples of "alibi" cases are when a married woman has an affair and contracts a venereal disease from her lover, or when an unmarried girl or the wife of a prison inmate becomes pregnant because of an affair. Id. at 88.

693 Kanin, supra note 317, at 84.

694 Id. at 85.

695 Id. at 86.

696 Id.

697 Id. at 87.

698 Id.

699 Id.

700 Id. at 90.

701 Id. According to the women's recantations, 53% were motivated by the need for an alibi after consensual sex had led to unanticipated problems, 44% were for revenge, and only one was due to a need for attention or sympathy. Id.

702 The "revenge" cases must have involved acquaintances, but we suppose that in most of the other types of cases those who filed false reports found it more expedient to claim that they had been raped by a stranger. See supra note 692.

703 Id. at 83.

704 Id.

705 See Geis et al., supra note 644, at 58-59.

706 In one study, 38% of the putative victims refused to cooperate with the authorities at some stage. See Hursch, supra note 128, at 110-14.

707 See supra notes 331-37 and accompanying text.

708 See supra note 689.

709 Kanin, supra note 317, at 83.

[A]gency policy forbids police officers to use their discretion in deciding whether to officially acknowledge a rape complaint, regardless how suspect that complaint may be. Second, the declaration of a false allegation follows a highly institutionalized procedure. The investigation of all rape complaints always involves a serious offer to polygraph the complainants and the suspects. Additionally, for a declaration of false charge to be made, the complainant must admit that no rape had occurred. She is the sole agent who can say that the rape charge is false. The police department will not declare a rape charge as false when the complainant, for whatever reason, fails to pursue the charge or cooperate on the case, regardless how much doubt the police may have regarding the validity of the charge. In short, these cases are declared false only because the complainant admitted they are false.

Id.

In response to our questions, Kanin assured us that the police in the jurisdictions studied would dismiss a case whenever the woman, for whatever reason, asked them to do so. Telephone Interview with Eugene Kanin (Feb. 15, 1996).

710 Kanin suggests that the major difference between the city he studied and cities studied by other scholars may be that "his" police conducted unusually thorough investigations. Kanin, supra note 317, at 90. Cf. WIEHE & RICHARDS, supra note 227, at 37 (stating that "police experience suggests that a face-saving way of recanting a complaint is to decide not to pursue it").

711 In cases of false reports motivated by a desire for attention and sympathy "no one was identified as the rapist." Kanin, supra note 317, at 87. This was also true in at least some of the "alibi" cases. Id. at 85-86.

712 Id. at 88.

713 See supra text at notes 107-10.

714 See supra text accompanying notes 236-37.

715 See supra notes 331-37 and accompanying text.

716 See supra text accompanying notes 349-60.

717 See supra text accompanying notes 398-420.

718 See generally JEFFREY T. FREDERICK, THE PSYCHOLOGY OF THE AMERICAN JURY 259 (1987) (stating that general conclusion of researchers to be that "eyewitness testimony is relatively unreliable"). Eyewitness testimony increases conviction rates, but jurors appear to do a poor job of distinguishing between accurate and inaccurate eyewitnesses, relying largely on the serf-confidence of the witness, an attribute that has been found to be irrelevant to accuracy. Id. at 259-60.

719 See, e.g., HOLMSTROM &: BURGESS, supra note 4, at 49.

720 See, e.g., Henderson; supra, note 105, at 41 (describing acquittals in recent rape cases as evidence that men are not being held "responsible for raping women," with no discussion of burden of proof). In stranger rape cases the normal defense is misidentification and the prosecution's proof problems are mitigated by the well-documented tendency of jurors to overrate the probative value of eyewitness identifications, an advantage that does not apply when the parties are acquainted and the standard defense is consent rather than misidentification. See generally Bryden & Park, supra note 310, at 576-78.

721 See infra notes 766-69 and accompanying text. Psychological damage is more common among rape victims than physical injuries. Compared to women in general, rape victims are more likely to have Post-Traumatic Stress Disorder, or a major depression, or to have considered suicide. RAPE IN AMERICA, supra note 56, at 7.

722 The National Victim Center's survey of rape victims found that, of those who had reported the crime to the police, 25% did so more than 24 hours after the crime. RAPE IN AMERICA, supra note 56, at 5.

723 Legally, corroboration is no longer required for rape prosecutions; in practice, however, juries usually seek corroborative evidence when assessing a complainant's credibility. See supra notes 580-84 and accompanying text. See generally J. WIGMORE ON EVIDENCE (Section) 2034, at 343 (Chadburn rev. 1978) (testimony of a single witness is legally sufficient for guilty verdict).

724 ESTRICH, supra note 1, at 42-56.

725 RITA GUNN & CANDICE MINCH, SEXUAL ASSAULT: THE DILEMMA OF DISCLOSURE, THE QUESTION OF CONVICTION 29 (1988). "By virtue of the relationship of the perpetrator and victim, the burden of proof appears to shift toward the victim." WIEHE & RICHARDS, supra note 227, at 95. A Swedish scholar has expressed the same complaint: "IT]he rape victim is actually placed on trial with the burden of proving her innocence." Snare, supra note 234, at 206-07. See also HOLMSTROM & BURGESS, supra note 4, at 157 ("victim is as much on trial as the defendant").

726 In Coker v. Georgia, 433 U.S. 584 (1977), the Supreme Court declared that the death penalty is an unconstitutionally disproportionate sentence for a rape conviction. At the time of this decision, Georgia was only state to authorize the death penalty for rapes, but several years earlier 16 states imposed capital punishment in rape cases. Id at 59-96.

727 For example, under Minnesota's Sentencing Guidelines, first degree criminal sexual conduct is assigned a severity level of eight on a scale of 10. MINNESOTA SENTENCING GUIDELINES COMMISSION, MINNESOTA SENTENCING GUIDELINES AND COMMENTARY 41 (1993). The maximum sentencing range for first degree criminal sexual conduct under the guidelines is 153-183 months, surpassed only by the ranges for murder (up to 246 months for third degree or felony murder, up to 433 months for second degree, and mandatory life sentence for first degree). The maximum sentencing range for manslaughter does not surpass the range for first degree criminal sexual conduct. Id.

728 Timothy Clifford, Smith on Stand: "I Got Picked Up': Says Sex Consensual; Calls Accuser a 'Nut', NEWSDAY, Dec. 11, 1991, at 5.

729 Id.

730 Id.

731 Id.

732 Id.

733 Id.

734 Paul Richter, Heard No Noises, Smith Case Witness Says, L.A. TIMES, Dec. 9, 1991, at A20.

735 Id.

736 Id.

737 Larry Tye, Sen. Kennedy Denies Hearing Alleged Rape, BOSTON GLOBE, Dec. 7, 1991, at 10.

738 Shirley E. Perlman, Experts on Sex, Sand, Sky, NEWSDAY, Dec. 10, 1991, at 3.

739 Id.

740 Mary Jordan, Smith Expected to Testify Today; Queries on Other Women Unlikely, WASH. POST, Dec. 10, 1991, at A3.

741 Id.

742 Id.

743 Id.

744 Id.

745 Id.

746 Perlman, supra note 738.

747 Mary Jordan, Smith's Lawyer Credits Kennedy's 'Charisma': Prosecution's Use of Senator Called a Mistake, WasH. POST, Dec. 13, 1991,' at A3.

748 Tye, supra note 737.

749 Id.

750 Richter, supra note 734.

751 Tye, supra note 737.

752 Id.

753 Perlman, supra note 738.

754 Richter, supra note 734.

755 Mary Jordan, Smith Defense Hammers at His Accuser's Account, WASH. PosT, Dec. 6, 1991, at A1.

756 Ken Ringle, Trial Diary: The Women Behind the Mask, WASH. PosT, Dec. 6, 1991, at D[1].

757 Clifford, supra note 728.

758 Two of the women reported that Smith suddenly became aggressive, pinned them down, and pawed them, but that they were able to repulse him. A third woman reported that after she became intoxicated at a party that she and Smith attended, they returned to Smith's apartment where he made unwelcome sexual advances and, although she tried to fight him off, he forced her to have intercourse with him. Larry Tye et al., Alleged Assaults by Smith Described: Accounts By 3 Women Are Similar To Charges In Palm Beach Rape Case, BosTON GLOBE, July 24, 1991, at 8.

759 Estrich, Palm Beach Stories, supra note 28, at 10.12. Estrich never flatly States that in her judgment the jury should or should not have acquitted the defendant. At times, she implies that the evidence left reasonable grounds for doubt: "I'm not sure any of us knows for sure what really happened at the Kennedy compound .... "Id. at 9. "Even ff the jury was right to acquit, they were certainly wrong to do it because she could not remember exactly who said what and when." Id. at 30. But for the most part she stresses that the moral of the case is that the public has yet to decide that "date rapists deserve to be punished." Id. at 9. Absent evidence of a defendant's other rapes, prosecutors face "the overwhelming reluctance of jurors, or the public for that matter, to believe a woman when others do not come forward to support her." Id. at 26. The jury may have held Bowman to a "higher standard" of testimonial consistency than victims of other crimes. Id. at 29. She closes by asserting, once again, that "the question" is whether we are ready to punish date rapists rather than merely deploring the crime in the abstract. Id. at 32-33.

760 See, e.g., CAL. PENAL CODE (Section) 532 (West 1996); IDAHO CODE (Section)19-2116 (1995); NEV. REv. STAT. ANN. [section sign[ 175.

761 (Michie 1995). 761 See Krulewitch v. United States, 336 U.S. 440, 454 (1949) (Jackson, J., concurring).

762 The general rule is that the testimony of a single witness is a legally sufficient basis for a verdict of guilty. 7 J. WIGMORE ON EVIDENCE (Section) 2034, at 259 (3d ed. 1940).

763 See e.g., ALA. CoDE (Section) 15-25-34 (1995); CAL. EVID. CODE [section sign] 136 (West 1996); COLO. REv. STAT. ANN. (Section) 13-25-129 (West 1995); FL~. STAT. ANN. (Section) 90.803 (West 1996); MINN. STAT. ANN. (Section) 595.02 (West 1996).

764 See infra note 119.

765 KALVEN & ZEISEL, supra note 52, at 142 tbl.40. A study of rapes in the District of Columbia found that witnesses were present in 9% of the cases. KATZ & MAZUR, supra note 616, at 198-99.

766 Snare, supra note 234, at 197. Another study found that at least half of the victims of acquaintance rape reported receiving bruises, cuts, or black eyes, while 7% received internal injuries or were knocked unconscious, and 40% reported requiring medical attention. JOANNE BELKNAP, supra note 174, at 212. One reason this finding is counterintuitive is that the reader may unconsciously think of an "acquaintance" as a college boy on a date, forgetting that the rapist may instead be a violent lover. According to a Justice Department survey injuries are almost twice as likely to occur if the attacker was a husband or boyfriend rather than a stranger. Most Women Knew Their Attackers, NEWSDAY, Jan. 31, 1994, at 17. Stranger rapes might be less violent on average because the stranger's threat, for example with a knife, induces a terrified compliance, while the acquaintance relies more often on actual rather than threatened force.

767 Joan M. McDermott, Rape Victimization in 26 American Cities (1979), cited in Eleanor R. Hall & Patricia J. Flannery, Prevalence and Correlates of Sexual Assault Experiences in Adolescents, 9 VICtIMOLOGY 398, 404 (1985). McDermott found that nonstranger rapes were "less likely to involve a weapon," and "less likely to involve physical injury aside from the rape."

768 In a national survey, 70% of the rape victims reported no physical injuries, 24% reported minor injuries, and only 4% said that they had suffered serious physical injuries. RAPE IN AMERICA, supra note 56, at 4.

769 Ironically, women's increasing participation in traditionally "male" activities, such as sports, make such explanations often plausible.

770 See FAIRSTEIN, supra note 13, at 213-16. "In Michigan a prosecutor opined that the incident usually starts consensually and ends forcibly. In cases like these, he continued, the victim may be likely to falsify certain aspects of the crime or events which preceded the assault in order to strengthen her case." MARSH ET AL., supra note 25, at 94-95 (internal quotation marks omitted).

771 See Olin Guy Wellborn III, Demeanor, 76 COm~ELL L. Rev. 1075, 1104-05 (1991).

772 Two experienced criminal lawyers had never heard of such a prosecution and considered it nearly impossible to prove the woman's guilt beyond a reasonable doubt. Interview with Julius Nolan, Assistant Hennepin Gounty Attorney, in Minneapolis, Minn. (Oct. 3, 1995). Interview with Stephen Simon, Director of the Misdemeanor Prosecution and Defense Glinics at the University of Minnesota Law School, in Minneapolis, Minn. (Oct. 5, 1995).

773 Interview with Professor Stephen Simon, Director of the Misdemeanor Prosecution and Defense Clinics at the University of Minnesota Law School (Nov. 10, 1994).

774 Id.

775 Id.

776 Telephone Interview with E. George Widseth, Assistant Hennepin Gounty Attorney (Nov. 14, 1994).

777 Interview with John Stewart, Assistant Public Defender for Hennepin County, Minn. (Nov. 11, 1994).

778 BOURQUE, supra note 25, at 102-03.

779 Simon interview, supra note 773.

780 Widseth interview, supra note 776.

781 Id.

782 See ESTRICH, supra note 1, at 21-22.

783 Id. at 56.

784 "Both analogue and clinical studies indicate that greater force and injury were associated with higher probability of reporting to police." Steketee & Austin, supra note 174, at 288. "IF]actors that make a strong case for prosecution are more powerful predictors of reporting rape than of reporting assault." Lizotte, supra, note 138, at 185. Cf. Arnold S. Kahn et al., Rape Scripts and Rape Acknowledgment, 18 PSYCHOL. WOMEN Q. 53 (1994) (unacknowledged rape victims tend to have been raped in a less forceful manner than acknowledged victims). See also Kerstetter, supra note 255 (police try to dissuade "weak case" victims from pressing charges).

785 KALVEN & ZEISEL, supra note 52, at 253-54.

786 Conventional courthouse wisdom holds that juries in criminal cases tend to be more lenient than judges. Levine, supra note 444, at 457. Although Kalven and Zeisel's findings support this conclusion, another scholar suggests that in felony cases juries are more prone to convict than judges. Compare supra text accompanying notes 408-20, with Levine, supra note 444, at 464-66, and James P. Levine,Jury Toughness: The Impact of Conservatism on Criminal Court Verdicts, 29 CRIME & DELINQUENCY 71 (1983). Professor Levine contends that felony juries have become more conviction-prone as a result of increasingly harsh public attitudes toward crime. Id.

787 See supra text accompanying notes 525-26.

788 According to one study, where the evidence shows third-party sexual relations by the complainant, prior consensual sex with the accused, or a delayed rape report, many trial judges disbelieve the woman or conclude that she got what she deserved. Bohmer, supra note 403, at 304-07.

789 The view of the women's movement did not become orthodox in educated circles until roughly the 1970s and 1980s.

790 KALVEN & ZEISEL, supra note 52, at 253-54.

791 Deservedly famous though it is, the Kalven & Zeisel study is weakened by several methodological flaws. See supra note 441.

792 E.g., FAIRSTEIN, supra note 13, at 131-36.

793 The measures of evidence that were found to have little effect on jurors' pre-deliberation beliefs about the defendant's guilt included eyewitnesses, the number of prosecution witnesses and exhibits, the use of a weapon and injury to the victim. See LaFree, supra note 66, at 397-99.

794 Galvin, supra note 4, at 774.

795 Id. at 765456.

796 McCoRMICK, EVIDENCE, (Section) 193 (John W. Strong et al. eds., 4th ed. 1992).

797 E.g., Deanna Crawford, The Rape Shield Law: MakingIt Work, 24 N.H. BAR J. 109, 111 (1983). As Crawford explains: Experts agree that the myths about rape support the defense attorneys' practice of delving into the rape victim's life. Perhaps the most persistent myth about rape is that it is a sexually motivated crime, an act of passion, when, in fact, it is a crime of violence. The rapist is attempting to prove his strength and power, to control and exploit, to express his hostility toward women and to compensate for feelings of inadequacy. No one denies the complexity of the rapist's motivation or the variety of social and psychological causes for it. But the fact remains clear that although sex may be the result, it is not the cause of this crime. Every time rape is misconstrued as sexually motivated, it detracts [sic] attention from the offender to the victim. Every time this attention shifts from the offender to the victim, some of the blame is shifted as well. The result is to lessen the blame on the offender and to diminish the credibility of the victim. Id. See also ADLER, supra note 48, at 9.

798 "IT]he feminist perspective views rape as not primarily a sexual act but rather a crime of violence and humiliation." WIEHE & RICHARDS, supra note 227, at 80. This theory is sometimes combined with the idea that consensual heterosexual intercourse is also more political than sexual. See, e.g.,Jackson, supra note 27, at 16, 19 (observing that "[s]exual conquest becomes an acceptable way of validating masculinity, of demonstrating dominance of and superiority over women... If sexuality was [sic] not bound up with power and aggression, rape would not be possible"). Such statements are commonly coupled with insistence that rapists are simply ordinary men. See, e.g., id. at 16. Rape is thus perceived as similar to ordinary heterosexual acts, whose purpose is political and oppressive, not to "relieve an itch." A thorough evaluation of this ideology is beyond the scope of the present article.

799 See, e.g., BROWNMILLER, supra note 1, at 14-15; ESTRICH, supra note 1, at 82-83; GERMAINE GREER, THE FEMALE EUNUCH, 24748 (1970).

800 E.g., Rusk v. State, 406 A.2d 624 (Md. Ct. Spec. App. 1979) (en banc), rev'd, 424 A.2d 720 (Md. 1981); State ex rel. M.T.S., 609 A. 2d 1266 (NJ. 1992); Commonwealth v. Berkowitz, 609 A. 2d 1338 (Pa. Super. Ct. 1992), aff'd in part, 641 A.2d 1161 (Pa. 1994).

801 See, e.g., ESTRICH, supra note 1, at 102-03 (arguing that the definition of rape should be expanded to include "extortionate threats and misrepresentations of material fact").

802 There are a few narrowly-circumscribed exceptions to the requirement of force (or threat of force), such as where the victim is unconscious. See generally Puttkamer, supra note 44, at 417.

803 The topic is obviously too complex for adequate treatment here.

804 See supra notes 797-98 and accompanying text.

805 For a summary of studies of rapists' characteristics, see BOURQUE, supra note 25, at 59-76.

806 "Although relevant, evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice .... "FEn. R. EVID. 403. Rule 403 codified the common law power of the judge to exclude relevant evidence that "arouses the jury's hostility or sympathy for one side without regard to the probative value of the evidence." McCORMICK, supra note 796, (Section) 185, at 340.

807 Id.

808 Richard Posner argues to the contrary, that "rape appears to be primarily a substitute for consensual sexual intercourse rather than a manifestation of male hostility toward women or a method of establishing or maintaining male domination." POSNER, supra note 8, at 384.

809 MARVIN E. WOLFGANG, PATTERNS IN CRIMINAL HOMICIDE 206-07 (1958).

810 See generally N. Heim & C. J. Jursch, Castration for Sex Offenders: Treatment or Punishment?: A Review and Critique of Recent European Literature, 8 ARCHIVES SEX. BEHAV. 281, 300 (1979).

811 Although we have no statistical data on the proportion of acquaintance rapes that include some clearly gratuitous violence, there is evidence that the woman usually suffers no extrinsic physical injury. See Ian T. Bownes et al., Rape: A Comparison of Stranger and Acquaintance Assaults, 31 MEn., SCI.,& THE L. 102, 104-06 (1991) (52% of sampled victims reporting no injuries, 34% reporting minor abrasions/bruises, 14% reporting further injuries).

812 Although the typologies overlap to a large degree, Nicholas Groth first described stranger rapists' motivations as coming in three types: the power rapist; the anger rapist; and the sadistic rapist. See ALLISON & WRIGHTSMAN, supra note 56, at 54-55. The power rapist is motivated by a desire to overcome his own sense of inferiority and inadequacy; he achieves control and authority through rape. Id. at 56. The anger rapist uses sex as a weapon to vent the extreme rage built up inside him; he often humiliates the victim and uses much more force than is necessary to overcome the victim's resistance. Id. at 55. The sadistic rapist gratifies himself by brutally torturing his rape victim, and sometimes even culminates the ritual by murdering her; he is the least common, but the most violent type of rapist. Id. at 57-58. As thus stated, none of the three motivational theories explains why the rapist does not simply batter or kill his victim.

813 See id. at 65 (stating that acquaintance rapists tend to employ "more subtle types of coercion" and "verbal 'manipulation'" rather than threats, physical violence, or weapons). See also supra note 811.

814 E.g., ANDREA DWORKIN, PORNOGRAPHY: MEN POSSESSING WOMEN 18 (1981); Catharine A. MacKinnon, Feminism, Marxism, Method, and the State: Toward a Feminist Jurisprudence, 8 SIGNS 635, 650 (1983).

815 See J.V. Check & N.M. Malamuth, Sex-Role Stereotyping and Reactions to Depictions of Stranger Versus Acquaintance Rape, 45 J. PERSONALITY &Soc. PSYCHOL. 344 (1983).

816 See, e.g., WIEHE & RICHARDS, supra note 227, at 77, 96; Scheppele, supra note 95, at 1110-11. See generally BOURQUE, supra note 25, at 7.

817 See, e.g., HELEN BENEDICT, VIRGIN OR VAMP 16 (1993).

818 E.g., MARSH ET AL., supra note 25, at 2.

819 By "promiscuity" we mean an unusually high propensity to engage in casual, nonmarital sex, or in other words a lack of sexual selectivity that is extraordinary by contemporary standards. As thus defined, the term, applied to females, may be less clear at the margin than, say, "negligence," but some behavioral patterns certainly meet our definition. We regret its pejorative innuendo, but have not found aconcise synonym.

820 See, e.g., Berger, supra note 1, at 55-57.

821 Reskin & Visher, supra note 503, at 43536. Cf. KALVEN & ZEISEL, supra note 52, at 165.

822 KALVEN & ZEISEL, supra note 52, at 249-50. In a more recent research project, Hubert Feild asked a sample of 896 white adults to respond to a six-page summary of a rape case. Different respondents received different summaries; all were asked to recommend a prison sentence. Feild found that subjects who read that the woman voluntarily allowed a (stranger) rapist into her apartment gave the rapist a lighter sentence than subjects who were told that the rapist forced his way into the apartment. Feild, supra note 487. Perhaps the subjects regarded the rapist who forced his way in as more dangerous to the average woman than one who talked his way in, a debatable assumption.

823 KALVEN & ZEISEL, SUpra note 52, at 243-44.

824 Interview with Professor Stephen Simon, Director of the Misdemeanor Prosecution and Defense Clinics at the University of Minnesota Law School (Mar. 7, 1996); Telephone interview with John Stuart, Minnesota State Public Defender (Mar. 11, 1996).

825 The Legal Bias Against Rape Victims, 61 A.B.A.J. 464 (1975) (Connie K. Borkinhagen speaking to the House of Delegates). Cf. Estrich, supra note 1, at 59 (observing that "if a thief stripped his victim, flattened that victim on the floor, lay down on top, and took the victim's wallet or jewelry, few would pause before concluding forcible robbery"). In a similar vein, another scholarly work says that there is "no other crime in which consent figures in the law as an issue." NANCY GAGER & CATHLEEN SCHURR, SEXUAL ASSAULT: CONFRONTING RAPE IN AMERICA. 139-42 (1976). Of course, the statement is not literally true: consent is a defense to, for example, theft, kidnaping, and nonsexual assault (sometimes called boxing). To be sure, defendants rarely raise the consent defense in prosecutions for theft and kidnaping, but that is because in such crimes (unlike date rape) the circumstances usually make the consent defense patently untrue. In nonsexual assault cases, defendants commonly claim that the complainant started the fight, which is analogous to a consent defense.

826 KADISH 8c SCHULHOFER, supra note 19, at 24.

827 ROLLIN M. PERKINS, CRIMINAL LAW 158 (2d ed. 1969).

828 Today the issue is largely governed by rape shield laws. Apart from such laws, many scholars and jurists have considered such evidence relevant. See, e.g., 1 WIGMORE, EVIDENCE. (Section) 62 (3d ed. 1940) (general character for unchastity); 3A WIGMORE, supra note 723, (Section) 924 (specific acts of unchastity). "Fortunately the character of the woman as to chastity or unchastity is admissible in evidence because of its probative value in judging whether she did or did not consent to the act in question." PEavaNs, supra note 827, at 158. For illustrative judicial opinions, see Brown v. State, 280 So. 2d 177, 179 (Ala. Crim. App. 1973) and People v. Stephens, 310 N.E.2d 824, 830 (I11. App. Ct. 1974). See generally AMIR, supra note 292, at 22-23; Galvin, supra note 4, at 783 n.96 (referring to a former California jury instruction that stated "[a] women of unchaste character can be the victim of a forcible rape but it may be inferred that a woman who has previously consented to sexual intercourse would be more likely to consent again"). Cf Men, Women and Rape, 63 FORDHAM L. REv. 125, 142 (1994) (remarks of Professor Donald Dripps) (observing that "[s]o long as the law requires the jury to focus on the victim, rather than the defendant, the shield laws are pretty difficult to justify on the basis of relevancy, at least in some of their more interesting applications").

829 MARSH ET AL., supra note 25, at 22-23 (victim's sexual history with other men is irrelevant and prejudicial). See also Berger, supra note 1, at 15-22; Estrich, Palm Beach Stories, supra note 28, at 26 (past sexual encounters with others are irrelevant except perhaps if defendant knew of them); Abraham P. Ordover, Admissibility of Patterns of Similar Sexual Conduct: The Unlamented Death of Character for Chastity, 63 CORNELL L. REV. 90, 97-102 (1977). f

830 E.g., Leon Letwin, "Unchaste Character," Ideology and the California Rape Evidence Laws, 54 S. CAL. L. REV. 35, 57-58 (1980).

831 See Covington v. State, 342 So. 2d 1339, 1340 (Ala. Crim. App. 1977). In a murder prosecution, the defendant claimed self-defense, stating that the victim attacked him with a knife, and he shot the victim to defend himself, hL at 1341. The wife of the victim testified that she and the defendant had been having an affair for some time. Id. at 1340. No issue was raised as to whether the wife's testimony was admissible. Id.

832 The Advisory Committee's Note to Federal Rule of Evidence 411 states that the inference of fault from the fact that a person carries insurance is tenuous. Rules of Evidence for United States Courts and Magistrates, 56 F.R.D. 183, 230 (1973). 833 One wonders, however, whether a jury can know which inference is stronger.

834 See McCoRMICK, supra note 796, (Section) 201. The Federal Rules of Evidence exclude such evidence. FED. R. Eviu. 411.

835 Concerning the meaning of relevancy, see MCCORMICK, supra note 796, (Sections) 184-185; see also 1 CHRISTOPHER B. MUELLER & LAIRD C. KIRKPATRICK, MODERN EVIDENCE (Section) 83 (1995); 1 MICHAEL GRAHAM, HANDBOOK OF FEDERAL EVIDENCE (Sections) 401-404 (3d ed. 1991); IRVING YOUNGFR & MICHAEL GOLDSMITH, PRINCIPLES OF EVIDENCE 44-49 (1984); GRAHAM C. LILLY, AN INTRODUCTION TO THE LAW OF EVIDENCE 21-28 (1978); 1 EDWIN C. CONRAD, MODERN TRIAL EVIDENCE (Sections) 21-37 (1956).

836 See FED. R. EVID. 401; McCoRMICK, supra note 796, (Section) 185.

837 See, e.g., Galvin, supra note 4, at 791-801; Murphy, supra note 602, at 552-53.

838 1A WIGMORE, EVIDENCE (Section) 32 (Tillers rev. 1983).

839 2 JONES, EVIDENCE (Section) 11.8 (Clifford S. Fishman ed., 7th ed. 1992).

840 Id.

841 See, e.g., United States v. Anderson, 575 F. Supp. 31, 33 (S.D.N.Y. 1983) (tape-recorded reaction of defendant after being told that an associate had named him as perpetrator held admissible despite defendant's claim that it was merely an angry reaction to a false accusation); Standafer v. First Nat'l Bank, 52 N.W. 2d 718, 720 (Minn. 1952) (evidence should be excluded unless one inference outweighs the other); People v. Yazum, 196 N.E.2d 263, 264-65 (N.Y. 1963) (attempted escape from custody admissible to show guilt even though defendant was also being held for another crime).

842 In other words, rape scholars customarily and understandably focus on whether the woman's conduct justifies rape, or proves that she consented, rather than on whether it has some tendency to indicate that consent will occur or did occur.

843 Some feminists contend that in our society women are not truly free to decline heterosexual relations, and therefore heterosexual sex is inherently coercive. E.g., CATHERINE A. MACKINON, TOWARD A FEMINIST THEORY OF STATE 168439, 171-83 (1989); LORENNE M.G. CLARK & DEBRA J. LEWIS, RAPE: THE PRICE OF COERCIVE SEXUALITY' 125-32 (1977).

844 See AMIR, supra note 292, at 259-76 (discussing the victim's contribution to forcible rape); HANS YON HENTIG, THE CRIMINAL AND HIS VICTIM 383450 (1948) (discussing the "duet frame of crime"); WOLFGANG, Supra note 809, at 245-65 (discussing victim precipitated homicide).

845 SeeJamesJ. Gobert, Victim Precipitation, 77 COLUM. L. P, REV. 511,518 (1977).

846 WOLFGANG, supra note 809, at 255.

847 Wharton defines voluntary manslaughter as an "intentional killing in the heat of passion as a. result of severe provocation." WHARTON, supra note 44, (Section) 155. See generally id. (Sections) 155.66.

848 See id. (Section) 165; see also Maher v. People, 10 Mich. 212 (1862).

849 See, e.g., Gobert, supra note 845, at 518-21.

850 People v. Mayberry, 542 P.2d 1337, 1346 (Gal. 1975) (en banc) (holding that a reasonable mistake about consent is a defense to forcible rape); WHARTON, supra note 44, (Section) 47 (stating that the contributory negligence of the rape victim is no defense).

851 Mayberry, 542 P.2d at 1346.

852 See generally WOLFGANG, supra note 809, at 247-52.

853 See, e.g., Gobert, supra note 845, at 513-14 & n.ll (excluding from definition of victim precipitation factors over which the victim has no control but that increase the likelihood of victimization, such as poverty and age).

854 Jackson, supra note 27, at 19-20.

855 Id.

856 Burt, supra note 516, at 217-30.

857 Feild, supra note 281, at 161.

858 Id.

859 AMIR, supra note 292, at 259-76.

860 Id. at 260-61.

861 Id. at 261.

862 Id.

863 Id. at 266.

864 Id. at 265.

865 Id. at 266.

866 BROWNMILLER, supra note 1, at 354.

867 AMIR, supra note 292, at 266.

868 BROWNMILLER, supra note 1, at 355.

869 Id. at 354.

870 Id.

871 See, e.g., ESTRICH, Supra note 1, at 24-25 (noting that "[in rape] when we ask 'Who started it?' we imply that if the woman agreed to give the man a ride home, or to go to his office or apartment, she is to blame for her subsequent rape and should not complain").

872 See AMm, supra note 292, at 259-66 (reporting that of the 646 rapes studied, 122 fit the definition of victim- precipitated). Cf ALLISON & WRIGHTSMAN, supra note 56, at 244-54 (discussing strategies of rape prevention that address behaviors of the potential victims while still acknowledging that rape victims are not to blame for attacks against them). Even Susan Estrich has noted that the rape victim is sometimes negligent. ESTRICH, supra note 1, at 24.

873 WHARTON, supra note 44, (Section) 47 (stating that the contributory negligence of the rape victim is no defense). See also People v. Mayberry, 542 P.2d 1337 (Cal. 1975) (holding that a reasonable mistake about consent is a defense in a forcible rape prosecution).

874 JOHN BARTLETT, FAMILIAR QUOTATIONS 855 (15th ed. 1980).

875 See generally Karen M. Kramer, Rule by Myth: The Social and Legal Dynamics Governing Alcohol-Related Acquaintance Rapes, 47 STAN. L. REV. 115, 126 (1994). When a woman's ability to appraise her own sexual conduct is impaired by the voluntary use of intoxicants or drugs, any resulting intercourse cannot be charged as rape, unless "the victim is manifestly unable ... to make a reasonable judgment as to the nature or harmfulness charged to constitute the offense." WAYNE R. LAFAw & AUSTIN W. SCOTT, JR., SUBSTANTIVE CRIMINAL LAW (Section) 5.11 (1986). See, e.g., Cole}, v. State, 616 So. 2d 1017 (Fla. App. 1993) (sexual battery conviction reversed despite evidence of victim's inability to consent because of cocaine use, based on "an absence of evidence to show.., that it was apparent to defendant that the victim was unable to give a knowing and voluntary consent").

Most state codes have rape provisions that include scenarios where the victim was drunk to the point of being unconsciousness. See, e.g., MODEL PENAL CODE (Section) 213.1 (A)-(C) (1985) (defining one kind of rape as a male having sexual intercourse with "a female not his wife if the female is unconscious"). Note, however, that the degree of required incapacitation is extreme. MODEL PENAL CODE (Section) 213.1 (1985) commentary at 318-19. Defense counsel often suggest that the woman's drinking is evidence of consent, or of impaired memory. See Adler, supra note 48, at 103-04.

876 Drinking also affects police founding decisions. See supra note 294.

877 KALVEN & ZEISEL, supra note 52, at 254-57.

878 See FAIRSTEIN, supra note 13, at 151-52; HOLMSTROM & BURGESS, supra note 4, at 144; Frohmann, supra note 13, at 215.

879 ALLISON & WRIGHTSMAN, supra note 56.

880 See, e.g., J.D. LEE & BARRY A. LINDailL, MODERN TORT LAW: LIABILITY & LITIGATION [section sign] 3.25, at 87 (rev. ed. 1988) (asserting that driving a car under the influence is evidence of negligence when there is a causal connection with the injury).

881 See, e.g., Charlene L. Muehlenard & Melaney A. Linton, Date Rape and Sexual Aggression in Dating Situations: Incidence and Risk Factors, 34J. COUNSELING PSYCH. 186, 187 (1987); M.B. Myers et al., Coping Ability of Women Who Become Victims of Rape, 52 J. CONSULTING & CLIN. PSYCH. 73 (1984) (rape victims exhibit higher rates of pre-rape alcohol and drug abuse). On one college campus, female victims of unwanted sexual contact reported male use of alcohol for over 75% of all types of experiences, and reported their own use of alcohol in over half of all experiences. Sally K. Ward et al., Acquaintance Rape and the College Social Scene, 40 FAMILY RELATIONS 65, 68 (1991). According to another study, 73% of men and 55% of women involved in acquaintance rapes had been drinking or taking drugs at the time of assault. Koss, supra note 56, at 16. Of college men who acknowledged having raped a woman, 74% said that they were drinking or using drugs at the time of the assault, and 75% recalled that their victims were doing so. Id. at 35, 45. Another study found that, of 236 rapists 142 (60%) were using alcohol or drugs at the time of the rape, and another 6% may have been. Of the victims, 36% indicated that they had been using drugs or alcohol. WIEHE & RICHARDS, supra note 227, at 17. For a discussion of other studies reaching similar conclusions, see Shotland, supra note 178, at 139-40. See also Snare, supra note 234, at 199-200 (most Swedish men accused of rape had been drinking, as had 30% of the alleged victims, 10-15% of whom were "highly intoxicated"). Cf. VACHSS, supra note 13, at 91 (asserting that "[r]apists... look for whatever vulnerability might insulate them from capture and punishment."). In other words, sexual predators--like all predators--have an eye for vulnerability, and situations that are suggestive of consent increase vulnerability.

The National Victim Center reports higher rates of drug and alcohol consumption among rape victims, who are 5.3 times more likely to have used prescription drugs nonmedically (14.7% vs. 2.3% of general population); 3.4 times more likely to have used marijuana (52.2% vs. 15.5%); 6 times more likely to have used cocaine (15.5% vs. 2.6%); 10.1 times more likely to have used hard drugs other than cocaine (12.1% vs. 1.2%); and 6.4 times more likely to have used hard drugs or cocaine (19.2% vs. 3.0%). RAPE IN AMERICA, supra note 56, at 7. However, for most victims, the age at which the first rape occurred was younger than the age at which they first became intoxicated or began using marijuana or cocaine. "Only 21% first became intoxicated at an earlier age, and only 32% of those who used marijuana did so earlier than their first rape. The corresponding figure for cocaine use was 11%." Id. at 8. In some instances, the rape may have been a cause of the drug or alcohol use. In others, intoxication may have been a cause of the rape, and in still others, there may have been a common cause, such as delinquent peers or abusive parents.

882 AMIR, supra note 292, at 99.

883 Id. at 258.

884 "The degree of intoxication of the man is the single most important factor in determining whether acquaintance rape will occur." Parrott & Bechhofer, supra note 478, at 23. Cf MACDONALD, supra note 178, at 57 (50% of rapists drank beforehand, the majority heavily).

885 Muehlenard & Linton, supra note 881, at 187.

886 Id. See also AMIR, supra note 292, at 343.

887 ld. One study reported that if the woman is stoned or drunk, 29% of males and 18% of females consider aggressive sex to be appropriate. Giarrusso et al., supra note 478, at 66.

888 AMIR, supra note 292, at 343.

889 For an impressive analysis from a pro-prosecution point of view, see Kramer, supra note 875.

890 See supra note 875 (discussing the legal relevance of alcohol use to the element of consent).

891 See supra note 435.

892 See Kerstetter, supra note 255, at 274. Kerstetter does not consider a third alternative: that the jurors blame drunken victims, whether male or female, for helping to precipitate the crime, rather than supposing that they cannot recall such an important detail as whether their sexual resistance was forcibly overcome.

893 See supra note 483.

894 It is unclear how often the victim's drinking casts doubt on her claim that she was raped. Of course, much depends on how inebriated she was. The nature of the conflict in the evidence is also important. In the Kennedy-Smith case, for example, the woman claimed that the defendant had tackled her from behind as she was leaving the Kennedy property, and then raped her. He claimed that he had not tackled her and their encounter was consensual. When the difference between the parties' stories is so stark, it is unlikely to be due to alcoholimpaired perceptions or memories. On the other hand, one can imagine cases in which the parties' stories were virtually identical, except for a gesture, a word, or a facial expression. In such a case, the victim's intoxication might well create a reasonable doubt.

895 According to Estrich, "[t] he defendant lucky enough to find out, albeit later, that his victim was sexually experienced could and would try to hide behind that fact at trial, if she was willing even to proceed to trial." ESTRICH, supra note 1, at 53. Similarly, Fairstein notes that, "[a]s most people are aware, and as Susan Brownmiller's history describes in exquisite detail, the most common defense in sexual assault cases--and one that had worked for centuries--was an incorporation of details or innuendos about the sexual history of the victim." FAIRSTEIN, supra note 13, at 122. See also ADLER, supra note 48, at 94 (observing that "[i]nsofar as a sexually active woman will be seen as a 'bad' woman, and therefore an unlikely rape victim, it becomes important for the defense in the trial to portray her as such").

In a jury simulation study, subjects from the Minneapolis-St. Paul area saw a videotape of a mock rape trial, condensed into about two hours. The defense was consent. Some subjects received information about the putative victim's sexual relations with other men; others did not. Jurors who had received such information inferred that the woman had consented to sex with the accused. They evaluated the woman's conduct more carefully, more unfavorably, and attributed more responsibility to her for the encounter. Even in scenarios where the defendant's guilt was highly probable, jurors who had learned of the woman's sexual history were reluctant to convict. Eugene Borgida, Legal Reform of Rape Laws, in 2 APPLIED SOC. PSYCH. ANNUAL 211 (L. Bickman ed., 1982).

Judges have also been prejudiced against rape complainants with "loose" habits. Bohmer, supra note 403, at 306.

896 State v. Bird, 302 So. 2d 589, 592 (La. 1974).

897 See WiGMOrnS ON EVIDENCE (Section) 200 (3d ed. 1940). To show lack of chastity, most courts allowed reputation evidence, but not testimony about specific sexual acts. WIGMORE, supra note 723, at 682. An exception allowed such testimony with regard to the defendant. See, e.g., Bedgood v. State, 17 N.E. 621 (Ind. 1888). Most courts held that lack of chastity, though relevant to consent, could not be used to impeach overall credibility. This distinction seems devoid of practical importance in this context because, in most cases, proof that a woman consented is proof that she is lying or exaggerating. At least some of those courts which allowed evidence of promiscuity to impeach the woman's credibility did not apply this rule to men, an apparent case of sex discrimination. Berger, supra note 1, at 16. But the real problem is the tacit assumption that one type of violation of conventional ethical norms imports another. See it/. at 17; Bryden & Park, supra note 310, at 561-81. If this assumption is accurate, then the differential treatment of men and women was compelled by the ethical double standard. Since male promiscuity is not condemned by society, promiscuous men have not displayed a willingness to flout society's norms. Promiscuous women, by contrast, were flouting a societal norm, however unfair that norm may have been. This violation may not shed much light on willingness to violate an unrelated and more defensible norm, and its prejudicial effect on the jury is likely to be great in some cases.

898 Exceptions built into rape shield laws permit limited admission of evidence of the victim's sexual history. For example, Federal Rule of Evidence 412, the federal rape shield law, and state laws patterned thereafter, permit evidence of a victim's past sexual behavior in two situations. First, the defendant may offer evidence of the victim's sexual activity with someone other than the defendant to explain the presence of physical evidence of rape, such as semen, injury, or disease. FED. R. EVID. 412 (b) (2) (A) . Second, when the defense is consent, the defendant may offer evidence that the victim engaged in sexual behavior with him in the past. FED. R. Eyre. 412(b)(2)(B). Other exceptions have been created by state statutes and judicial decisions. See generally Galvin, supra note 4, at 883-93 (discussing in detail the federal rape shield law and similar state laws).

899 WIGMORE, supra note 723, (Section) 62. However, even Wigmore recognized that, "It]he fact that a woman may have been guilty of illicit intercourse with one man is too slight and uncertain an indicator to warrant the conclusion that she would probably be guilty with any other man who sought such favors of her." Id.(Section) 200 (quoting Rice v. State, 17 So. 286, 287 (Fla. 1895)).

900 See, e.g., Lisa Van Areburg & Suzanne Rechtin, Rape Evidence Reform in Missouri: A Remedy for the Adverse Impact of Evidentiary Rules on Rape Victims, 22 ST. Louis U. L.J. 367, 385 (1978); Berger, supra note 1, at 55-56; Evelyn Stroufe, Evidence--Admissibility of the Victim's Past Sexual Behavior Under Washington's Rape Evidence Law--Wash. Rev. Code [section sign] 9.79.150 (1976), 52 WASH. L. REv. 1011, 1031-32 (1976); Note, If She Consented Once, She Consented Again--A Legal Fallacy in Forcible Rape Cases, 10 VAL. U. L. REv. 127, 138-43 (1976).

901 See, e.g., Berger, supra note 1, at 55-56.

902 Estrich, Palm Beach Stories, supra note 28, at 26. Estrich grants that if these encounters were known to the defendant they may be relevant to his mens rea. Id. We believe that this is so only if the circumstances and testimony raise the possibility that the defendant made a reasonable mistake about whether the woman had consented, which is not the same as the far more common reasonable mistake about whether a woman will consent:

903 See generally LAUMANN ET AL., supra note 56, at 172-224.

The women [like the men] also exhibit quite different proportions of having many partners, with only 2.6 percent of the 1933-42 [birth] cohort, nearly 18 percent of the 1943-52 cohort, and over 22 percent of the 1953-62 cohort having had five partners [by the age of thirty]. Here again, there is quite compelling evidence of a period effect on numbers of sex partners that can indeed be considered a sexual revolution.

Id. at 200.

904 For example, a national survey found that of women born between 1963 and 1972, only 13.3% had five or more partners by the age of twenty. Id. at 198. Although this proportion was higher than in previous birth cohorts, it would still be accurate to describe such behavior as unusual. Similarly, "Over 90 percent of the women . . . in every [birth year] cohort report fidelity within their marriage, over its entirety." Id. at 214. Therefore, it is still true that a married woman who has had several affairs is unusual. Even among single women, some patterns of sexual behavior are still extraordinary.

905 Estrich points out that prior rapes are more powerful evidence of guilt than prior sex is of consent. Estrich, Palm Beach Stories, supra note 28, at 25. We agree, even if the promiscuity is extreme. However, we are discussing the relevance of the evidence, not its probafive weight, and Estrich's observation does not rebut the evidence's relevance. Estrich also asserts that the jury is more likely to overrate sexual history evidence than prior crimes evidence. Again, we are inclined to agree, though this is debatable. See generally Bryden & Park, supra note 310.

906 SeeFED. R. EWD. (1996) (referring to Rule 413 added Sept. 13, 1994, Pub. L. 103-322, Title XXXII, (Section) 320935(a), 108 Stat. 2135, effective July 9, 1995). Rule 413 now provides that for a criminal defendant accused of an offense of sexual assault, evidence of his commission of another sexual assault is admissible. Jurors may consider this evidence for its bearing on any matter to which it is relevant. Id. See generally Bryden & Park, supra note 310.

907 Paul E. Meehl, Law and the Fireside Inductions (with Postscript): Some Reflections of a Clinical Psychologist, 7 BEHAVIORAL SCI. & L. 521, 532 (1989).

908 The admission of character evidence is governed by Federal Rule of Evidence 404. Generally, evidence of prior misconduct by the accused is not admissible to prove he acted in conformity with his character, although it may be admitted for other purposes. FED. R. EVID. 404(b). See generally Bryden & Park, supra note 310.

Evidence about a victim's character may be admitted under certain circumstances. Rule 404(a) (2) provides:

(a) CHARACTER EVIDENCE GENERALLY. Evidence of a person's character or a trait of character is not admissible for the purpose of proving action in conformity therewith on a particular occasion, except:

(b) CHARACTER or VICTIM. Evidence of a pertinent trait of character of the victim of the crime offered by an accused, or by the prosecution to rebut same, or evidence of a character trait of peacefulness of the victim offered by the prosecution in a homicide case to rebut evidence that the victim was the first aggressor.

FED. R. EVID. 404(a)(2) (emphasis added).

As McCormick points out, in murder and assault cases, when the identity of the first aggressor is in doubt, the probative value of evidence of a victim's character normally justifies the risk that a jury could erroneously conclude a victim got what he deserved. McCoRMICK, supra note 796, (Section) 193. In cases of rape or sexual assault, prior misconduct evidence to prove a victim's character is admissible if it is probative of consent; however, since the 1970s, such evidence cannot be admitted merely to establish (or rebut) the victim's chastity, which is no longer considered pertinent. Id. See also FED. R. EVID. 412 (allowing the admission of evidence of specific instances of sexual behavior to prove consent).

909 ADLER, supra note 48, at 36. See also AMIR, supra note 292, at 22-23 (lack of chastity is strong evidence of consent).

910 ADLER, supra note 48, at 101.

911 See supra note 898 and accompanying text; see also FAIRSTEIN, supra note 13, at 139-54.

912 One can imagine a case in which such evidence would be ambiprobative: for example, if the prior assaults by the putative victim had been on the defendant. In such a case, we concede, one is tempted to admit the evidence despite its ambiprobative quality, on the theory that some other evidence will reveal the superiority of one of the two conflicting inferences. Like promiscuity evidence in a rape case, this evidence may prejudice the jury against the victim, but we doubt that the problem is as severe in this context. Moreover, we think that in the case of the bully, the inference of a retaliatory assault is probably less strong than the inference of yet another assault by the bully. Therefore, the evidence may be relevant rather than ambiprobative.

913 22 CHARLES A. WRIGHT & KENNETH W. GRAHAM, FEDERAL PRACTICE AND PROCEDURE (Section) 5287, at 594 n.51 (1980).

914 M.P. Koss & T.E. Dinero, Discriminant Analysis of Risk Factors for Sexual Victimization Among a National Sample of College Women, 57 J. CONSULTING & CLINICAL PSYCHOL. 242 (1989). Cf Koss, The Hidden Rape Victim, supra note 56, at 208 (survey of university women found that the acknowledged rape victims reported significantly more liberal sexual values, a greater number of sexual partners and an earlier age for first intercourse than nonvictimized women did). Conceivably, these findings merely reflect the fact that, all else being equal, those who date the most are at the greatest statistical risk of rape. Koss reports that "38% of the rape victims had never had sexual intercourse before their sexual assault." Id. Cf Muehlenard & Linton, supra note 881, at 192 (college women who had experienced sexual aggression had less traditional attitudes toward women than those who had not, while men who had been guilty of sexual aggression had more traditional attitudes than those who had not).

915 S.M. Essock-Vitale & M.T. McGuire, Women's Lives Viewed From an Evolutionary Perspective, 6 ETHOLOGY & Soc. 137 (1985).

916 Philip A. Belcastro, A Comparison of Latent Sexual Behavior Patterns Between Raped and Never Raped Females, 7 VICtIMOLOGY 224, 225 (1983). The sample was limited to non-virgins who had never been married. Victims of homosexual rape were excluded as were rapes by multiple assailants. Id.

917 Id.

918 Id. at 22526.

919 Id.

920 Id. at 226.

921 AMIR, supra note 292, at 116. A Swedish study found that 15% of reported rape victims had a criminal record (predominantly for property crimes), as compared to an expected 3%. Snare, supra note 234, at 200.

922 AMIR, supra note 292, at 116-17.

923 Id. at 117. Another study of Philadelphia sexual assault victims included 121 adolescents between the ages of 13 and 17. Over half of them had intercourse before the sexual assault, 30% had been truant, 23% had run away from home, and 19% had prior police trouble. KaisER ET AL., supra note 314.

924 AMIR supra note 292, at 257. Street gangs and motorcycle gangs often select a girl with a promiscuous reputation for a gang rape. MACDONALD, supra note 178, at 69.

925 Scully & Marolla, supra note 634, at 58, 69.

926 Catalina Mandoki & Barry R. Burkhart, Sexual Victimization: Is There a Vicious Cycle?, 4 VIOLENCE & VICTIMS 179 (1989), cited in Susan Sorenson et al., Repeated Sexual Victimization, 6 VIOLENCE & VICTIMS 299, 301 (1991).

927 EAUMANN ET AL., Supra note 56, at 335. The respondents in this survey were asked whether they had ever been "forced to do something sexual that they did not want to do." Id. at 338. The authors caution that "the reported experience involving force may not constitute rape in a legal sense or in the minds of the female respondents." Id. at 335. In ordinary usage, the word "forced" is often employed loosely to cover situations in which there is pressure but no physical violence or threat of violence. For example, some might say that a boyfriend who threatened to terminate the relationship unless his sexual request was granted "forced" his girlfriend to accede. Similarly, if a woman threatened to terminate a relationship unless the man married her, some might say she "forced him" to marry her.

928 Id. at 339.

929 They were more likely to have engaged in group sex, active and receptive oral sex, and anal sex. Id.

930 Rape victims and women who have had numerous sexual parmers both tend to be younger, more urban, and poorer than the general population. Kathryn Kost &Jacqueline D. Forrest, American Women's Sexual Behavior and Exposure to Risk of Sexually Transmitted Diseases, 24 FAM. PLAN. PERSP. 244, 24647 (1992) (stating that the women most likely to have many sexual partners are 20-34 years of age, with an income below the poverty level, and living in an urban area). According to another study, 16-19 year-olds are the group with the highest rape victimization rate, followed by 20-24 year-olds. Koss et al., The Scope of Rape, supra note 170, at 162. Amir found that rape victims in Philadelphia were generally "below the age of marriage." AMIR SUpra note 292, at 61. Rape victimization rates are "clearly related to family income," and higherin low-income families. Id. at 169. National Crime Survey data confirm this assertion. Belknap, supra note 174, at 209-10. Data from a national survey of reported sexual violence in Sweden led one scholar to characterize rape as "an urban phenomenon." Snare, supra note 234, at 196. Half of the Swedish victims fell in the 15-25 age group. Id. at 199. Amir concluded that in Philadelphia "the potentiality of the Negro female to become a victim of rape is almost twelve times greater than that of the white female." AMIR, supra note 292, at 44. "Women with incomes at or above poverty level are less likely to report having had more than one partner than are those with incomes below poverty level" but there is no difference between low-income and middleincome groups, and there is no difference by poverty status in the proportion of women who have had more than 10 partners." Kost & Forrest, supra, at 24546. "Women living in urban areas are more likely than rural residents to have had more than one partner and to have had more than 10." Id. The greatest proportion of women with multiple partners during the past three months were 20-29 years of age, and 86% lived in urban areas. Id. at 249.

931 Id.

932 See MCCAHILL ET AL., supra note 52, at 112 (police suspicion of gang rape complaints affects chiefly poor, black victims); Id. at 116 (police more likely to unfound case if victim is a welfare recipient). See generally Crenshaw, supra note 596, at 1251, 1267-69 (describing violence against victims who are women of color).

933 Cf supra note 930. According to national victimization data, rape occurs three times as often among unemployed women, proportionate to their number, as it does among women who are either employed or "housekeepers." CRIMINAL VICTIMIZATION SURVEY 31-32 (1981). A study of Philadelphia rape victims, including 121 adolescents between the ages of thirteen and seventeen, found that in 42% of the cases, the victim's family received some form of public assistance. KRASNER ET AL., supra note 314. See generally BOURQUE, supra note 25, at 40 (some studies indicate young, poor, single, non-Anglo women are at greatest risk of being raped). But see id. at 4042 (data challenging the idea that lower-class minority women are overrepresented among rape victims); Hall & Flannery, supra note 767, at 404 (found no relationship between race or class and Sexual assault victimization).

934 We have omitted from consideration the possibility that poor people are, as a statistical matter, less honest than affluent people, an inquiry that would take us far afield, particularly in view of the literature suggesting that character traits are situational, so that one may tend to be dishonest concerning some matters yet honest about others. For a discussion of this problem, see Bryden & Park, supra note 310, at 561435.

935 See, e.g., State v. Williams, 487 N.E.2d 560 (Ohio 1986) (evidence of victim's reputation as a prostitute held admissible to impeach her testimony that she had not consented); ADLER, supra note 48, at 69, 83 (even the "most enlightened" English judges admit such evidence). Cf Commonwealth v. Joyce, 415 N.E.2d 181, 187 (Mass. 1981) (evidence of complainant's prostitution inadmissible to show consent but admissible to show motive to lie). See a/so Galvin, supra note 4, at 828-902.

936 Although the law is to the contrary, prostitutes--in the opinion of many-- annot be raped. Silbert, supra note 316, at 75. Some prosecutors describe prostitute-rape cases as "difficult," but some jurors and officials take these cases seriously. BOURQU% supra note 25, at 5.

937 E.g., BOURQUE, supra note 25, at 4 (quoting judge to the effect that a prostitute by definition cannot be raped); Adler quotes an English judge who regarded rape of a prostitute as "like a contradiction in terms." ADLER, supra note 48, at 3.

938 "[T]o speak of sexual intercourse with a prostitute without her consent as an 'outrage to her person and feelings' is in the nature of mockery." PERKINS, supra note 827, at 158.

939 E.g., Haynes v. State, 498 S.W.2d 950 (Tex. Crim. App. 1973); 4 W. BLACKSTONE, COMMENTARIES 213 (1765).

940 See ROLLIN M. PERKINS & RONALD N. BOYCE, CRIMINAL LAw 205, 1080 (3d ed. 1982).

941 See BROWNMILLER, supra note 1, at 364435; HOLMSTROM & BURGESS, supra note 4, at 4950. This theory is not limited to policemen who reject all rape accusations by prostitutes. According to a newspaper account, the area in Los Angeles with the largest number of rape reports abuts corridors known for drug-dealing and prostitution. A policeman explains that, "[t]here are some people out there that don't like women, period, and use prostitutes as a real up-front, easy-to-identify example of their hatred." But the same policeman "said some rape reports filed by prostitutes turn out to be false claims against men who don't pay, including drug dealers who offer quick hits of heroin and crack cocaine in exchange for sex, then renege on their promises." Leslie Berger, Prostitutes are Prime Victims in Leading Rape Area, L.A. TIMES, May 8, 1994, at B1.

942 McCORMICK, supra note 796, (Section) 193, at 350.

943 State v. Camp, 3 Ga. 417, 422 (1847).

944 Kurt Weiss & Sandra S. Borges, Victimology and Rape: The Case of the Legitimate Victim, 8 IssuEs CRIMINOLOGY 71, 95 (1973).

945 A study in Winnipeg found that 78% of the city's prostitutes had experienced non-sexual assaults, while over 50% had been victims of both types of assault, since becoming a prostitute. GUNN & MINCH, supra note 725, at 24.

946 Silbert, supra note 314, at 221.

947 The 73% figure did not include attempted rapes, rapes by customers, or other sexual abuses. Id. at 87.

948 Id. at 79.

949 Id.

95O The number of women who are victims of sexual assault has become a controversial subject. See, e.g., Gilbert, supra note 158, at 57-65 (arguing that variances among findings are a result of confusing incidence rates and prevalence rates, or broad versus narrow definitions of sexual assault). The 1-in-4 figure often appearing in feminist literature includes attempted rapes as well as completed rapes. See, e.g., Andrea Parrot & Laurie Bechofer, Preface to ACQUAINTANCE RAPE, supra note 63, at ix. Prevalence studies have arrived at various conclusions. One study found that over 27% of the women interviewed in Cleveland, Ohio, had been victims of rape or attempted rape since the age of 14. Mary P. Koss et al., Criminal Victimization Among Primary Care Medical Patients: Prevalence, Incidence, and Physician Usage, 9 BEHAV. SCI. & L. 85, 90 (1991). Half of these women had been victims of rape and haft of attempted rape. Id. In a Los Angeles study, 25% of the Black women interviewed had experienced rape or attempted rape since the age of 18. Gail Elizabeth Wyatt, The Sociocultural Context of African American and White American Women's Rape, 48J. SOC. IssuEs 77, 82 (1992). Of white women, 20% had been raped or subjected to an attempted rape, a non-significant difference between the two racial groups. Id. In a larger Los Angeles study of 3000 adults, 16.7% of the women and 9.4% of the men had experienced completed or attempted sexual assault. Susan B. Sorenson & Judith M. Siegel, Gender, Ethnicity, and Sexual Assault: Findings from a Los Angeles Study, 48 J. SOC. IssuEs 93, 96 (1992). In Russell's study of 930 women in San Francisco, 24% of the women had experienced completed rape in their lifetimes. RUSSELL, supra note 170, at 35. Other studies have found lower prevalence rates of rape. Gordon and Riger found that 6% of volunteers surveyed by telephone and 11% of volunteers surveyed by interview had experienced rape. Stephanie Riger & Margaret T. Gordon, The Fear of Rape: A Study in Social Control, 37 J. SOC. ISSUES 71, 76, cited in Koss, Detecting the Scope of Rape, supra note 158. Another study found that 8.8% of women surveyed by telephone in Charleston, South Carolina had experienced forcible rape or attempted forcible rape. Kilpatrick et al., supra note 170, at 866.

951 Silbert, supra note 314, at 79.

952 Id. at 81.

953 Id. at 84.

954 Id. at 83.

955 Id. at 87.

956 Id.

957 If Jones, an artist, accuses Smith of having stolen a painting from him, Smith may contend that Jones gave him the painting. At least if the painting is valuable and the parties know each other only slightly, this explanation is likely to strike the fact-finder as implausible. To correct this assumption, Smith should be permitted to offer evidence that Jones has previously given away valuable paintings to mere acquaintances. The same logic applies, in some cases, to the rape complainant's sexual history. As Park explains:

Suppose that a rape prosecution arises from sexual intercourse that occurred on the pavement of a parking lot on a cold rainy night. The jury is likely to think that the complainant would not have consented because of the circumstances. Evidence that she had consensual intercourse with another man on the same pavement earlier in the evening, however, leads to inferences that help the defense more than the prosecution, even if she did not claim rape on the earlier occasion.

Roger C. Park, The Crime Bill of 1994 and the Law of Character Evidence: Congress Was Right About Consent Defense Cases, 22 FORDHAM URB. L.J. 271, 278 (1995).

A more realistic (and more difficult) case would be one in which the complainant was a respectable married woman--perhaps even a minister's wife--and the consent-defense rape defendant sought to introduce evidence of previous affairs in order to rebut the inference of nonconsent that might be drawn from her marital status and apparent moral stature.

958 See supra text accompanying notes 831-$4. Exclusion of evidence of the putative victim's promiscuity can be justified on grounds of extrinsic policy: to encourage reporting of rape and to preserve the woman's privacy. Bryden & Park, supra note 310, at 568. In addition, the jury is likely to overvalue such evidence. See supra note 895. Therefore, exclusion of the evidence can be justified even if one believes that ambiprobative evidence generally should be admissible.

959 If she was a prostitute, the defendant may claim that she falsely accused him after a dispute about her fee. If she and the defendant were once lovers, their breakup may give rise to a dispute that, according to the defendant, motivated a false accusation. C POSNER, supra note 8, at 388.

960 McCORMICK supra note 796, (Sections) 193, 350.

961 See supra notes 940-41.

962 Cf. State v. Colbath, 540 A. 2d 1212, 1216-17 (N.H. 1988). In Colbath, the defendant was convicted of aggravated felonious sexual assault after the judge instructed the jury that, under the state's rape shield law, evidence of the complainant's behavior with other men was irrelevant to the issues before them. The evidence in question indicated that the complainant directed sexually provocative attention toward several men in a bar, including the defendant. The defendant testified that he had fondled her breasts and bottom and she had "rubbed his crotch" before the two of them eventually left the tavern and went to the defendant's trailer, where sexual intercourse occurred: consensual by his account, forcible by hers. The two of them were joined unexpectedly by a young woman who lived with the defendant. Suspecting that the defendant was indulging in faithless behavior, she came home at an unusual hour. With her suspicion confirmed, she flew into a rage and assaulted the complainant. As soon as the complainant returned to town, she accused the defendant of rape. The New Hampshire Supreme Court, in an opinion by Judge (later Justice) Souter, held that evidence of the complainant's behavior immediately prior to the rape, apparently inviting sexual advances from other men, such as sitting in one man's lap, was relevant on the issue of consent. Judge Souter explained that "the jury could have taken evidence of the complainant's openly sexually provocative behavior toward a group of men as evidence of her probable attitude toward an individual within the group," and "the sexual activities of a complainant immediately prior to an alleged rape may well be subject to a defendant's constitutional right to present evidence." Id. at 1217.

963 See, e.g., HOLMSTROM & BURGESS, supra note 4, at 42-43; Scully & MaroUa, supra note 634, at 66.

964 A study of 55 rape victims found that 20% had been hitchhiking. QUEEN'S BENCH FOUNDATION, RAPE VICTIMIZATION STUDY. PRELIMINARY RESEARCH FINDINGS AND REGOMMENnATIONS 13 (1975), cited in BROWNMILLER, supra note 1, at 186 [hereinafter Rase VICTIMIZATION STUDY]. Two sociologists who examined police reports in Boston and Los Angeles found that gang rape was more common in the latter city. They attributed this difference to a transportation and mobility problem that encourages the practice of hitchhiking in Los Angeles. Id. at 186. According to data compiled by the Federal Bureau of Investigation, many rapes of college students occur while hitchhiking. Id. at 350. Brownmiller is surely correct in asserting that "[r]ape-minded men" would consider hitchhiking "tantamount to an open invitation." Id. at 354. One study of convicted rapists' attitudes found that a number of them regarded hitchhiking by females as "a signal of sexual availability." Scully & Marolla, supra note 634, at 66.

A common kind of gang rape is hitchhike-abduction rape. As Scully & Marolla note:

In these cases, the gang, cruising an area, 'looking for girls,' picked up a female hitchhiker for the purpose of having sex. Though the intent was rape, a number of men did not view it as such because they were convinced that women hitchhiked primarily to signal sexual availability and only secondarily as a form of transportation. In these cases, the unsuspecting victim was driven to a deserted area, raped, and in the majority of cases physically injured.

Id. at 69. Cf. Hall & Flannery, supra note 767, at 404 (hitchhiking found to be "significantly related" to sexual assault).

965 MGGAHILL ET AL., supra note 52, at 116. A prior felony conviction also may affect the complainant's credibility.

966 Suzanne S. Ageton, Vulnerability to Sexual Assault, in RAPE AND SEXUAL ASSAULT II, supra note 314, at 221.

967 Id. at 239. Cf. Hall & Flannery, supra note 767, at 402 (finding a slight but statistically significant relationship between the "sexual climate of the peer group" and sexual victimization).

968 Ageton, supra note 966, at 238.

969 Id. at 238-39. Another study found that while rape victimization was significantly related to membership in a sexually active peer group, the relationship was "very slight." Cf Hall & Flannery, supra note 767, at 403. By the same token, some studies indicate that religious females, with fewer sexually active female friends, are statistically less likely to be raped. Id. at 405.

970 Ageton's definition of rape was vague and her methodology has been criticized. BOURQUE, supra note 25, at 132-33.

971 See Estrich, Palm Beach Stories, supra note 28, at 14-20.

972 Id. at 15. This was done, in the press and in court filings, prior to the Kennedy-Smith trial. Id.

973 Id. at 16-17.

974 This may be because they are dependent on caregivers and thus reluctant to resist, because they are unusually passive, or because they have difficulty expressing their lack of consent. See generally Sherene Razack, From Consent to Responsibility, From Pity to Respect: Subtexts in Cases of Sexual Violence Involving Girls and Women With Developmental Disabilities, 19 LAW & SOC. INQUIRY 891 (discussing the difficulty of applying traditional consent standards to victims with developmental disabilities).

Of course, these qualities also may render mentally ill and mentally retarded women more vulnerable to rape. A study of sexual abuse trends analyzed results from several Canadian and American studies and found the rate of sexual abuse and assault for developmentally disabled women was four times greater than the national average. CHAInruNE Y. SENN, VULNERABLE: SEXUAL ABUSE & PEOPLE WITH AN INTELLECTUAL HANDICAP 4-8 (1989).

975 See supra note 611.

976 MCCAHILL ET AL., supra note 52, at 116.

977 One study reported that rape victims, when compared to controls, had higher rates of pre-rape suicidal thoughts and psychiatric hospitalization. Myers et al., supra note 879, at 76. A nationwide Swedish study concluded that 10% of rape complainants "seem to have had some serious mental problems prior to the rape incident." Snare, supra note 234, at 200. Of 55 American rape victims, 40% had psychiatric problems prior to the rape. RAPE VICTIMIZATION STUDY, supra note 948, at 19. Estrich perceptively adverted to the possibility that psychiatric evidence about the rape accuser cuts both ways because perhaps "rape victims may be especially likely to have led lives of trauma or pain; that rapists may sense, and prey, on the vulnerable and the needy .... "Estrich, Palm Beach Stories, supra note 28, at 17.

978 For an extreme example, see the dialogue reproduced in HOLMSTROM & BURGESS, supra note 4, at 75.

979 See sources cited supra note 316.

980 E.M. Ellis et al., An Examination of Differences Between Multiple-and Single-lncident Victims of Sexual Assault, 91J. ABNORMAL PSYCH. 221-24 (1982) (multiple-incident victims significantly more dysfunctional in six of eight areas of functioning, and significantly more likely to have attempted suicide). But see Sorenson et al., supra note 910, at 299-308.

981 MARSH ET AL., supra note 25, at 96.

982 See supra note 65 (reviewing the impact of prior intimacy with the defendant on a victim's credibility); note 380 and accompanying text (9% conviction rate).

983 See, e.g., Bohmer, supra note 403.

984 According to Brownmiller, although evidence about the complainant's sexual history with other men should be excluded, "[p]rior consensual intercourse between the complainant and the defendant does have some relevance, and such information probably should not be barred." BROWNMILLER, supra note 1, at 386. Similarly, Wechsler claimed that "it simply ignores reality to suggest that past practice with the accused is not relevant to the issue of consent on a given occasion, though it is equally a distortion to regard proof of such experience as dispositive." MODEL PENAL CODE (Section) 213.1 commentary at 350 (1980). See also AMIR, supra note 292, at 231.

Estrich offers a somewhat different rationale for admitting the evidence. In her view, "evidence of any relationship between these two people is almost invariably relevant to his state of mind, or mens rea, as to consent." Estrich, Palm Beach Stories, supra note 28, at 26. It is true, of course, that a previous relationship might include events that tended to validate--or for that matter, to rebut--a defendant's claim that he made a reasonable mistake as to the woman's consent. But this is not necessarily what the defendant will claim, or the evidence will plausibly suggest, actually occurred, in a rape case involving former lovers. For example, the woman may testify that after they had separated the man requested sex, she declined, and he then overpowered her. Unless he testifies that this sort of behavior was a recurrent and apparently mutually enjoyable feature of their relationship, we fail to see the relevance of that relationship to his "state of mind." In most cases, he will testify that she unambiguously assented. Cf. supra text accompanying note 369 (discussing such a case from Estrich's REAL RAPE).

985 The least controversial exception to rape shield laws allows testimony about prior intimacy between the parties. See, e.g., ADLER, supra note 48, at 86-87 (such inquiries permitted under English law); Galvin, supra note 4, at 807 (observing that "[e]ven the most ardent reformers acknowledged the high probative value of past sexual conduct.., when the defendant claims consent and establishes prior consensual sexual relations between himself and the complainant"). The theory behind this exception seems to be, not that such testimony is an inevitable preliminary to testimony about the encounter in question, but that it is relevant to whether the alleged victim consented. Adler, supra note 48, at 88. Adler argues that this sort of evidence, relies on the rather dubious assumption that once a woman agrees to intercourse with a man, the likelihood is that she will continue to consent at any later stage in their relationship or indeed when the relationship has ended. The rationale here is rather reminiscent of Hale's famous dictum regarding the permanent nature of consent given on marriage.

Id.

But, there is a huge difference between a rule that a married woman is not entitled to decline sex and a rule that merely allows the jury to learn about the marriage in determining whether she in fact declined. It is true, of course, that prior intimacy between the parties is far from being conclusive evidence that consent was given on the occasion in question. But Adler's argument does not explain why such evidence is irrelevant rather than merely inconclusive. If the issue were whether the woman had agreed to go fishing with the man on a particular occasion, or had agreed to sell him widgets, evidence that she had often done so in the past would be useful, even though she might have declined on the occasion in question. It is only by focusing on the ambiprobative aspect of the evidence that one can plausibly describe it as irrelevant. But it will probably be admissible anyway, "to provide context."

986 Koss, supra note 170.

987 Most Women Know Their Attackers, supra note 766, at 17. Another survey found that 9% of victims had been raped by their husband or former husband, and another 10% by boyfriends or ex-boyfriends. RAPE IN AMERICA, SUpra note 56, at 4.

988 It is probably impossible to fashion a research design that would resolve this issue. What scholars have studied, instead, is a related question: whether most rape victims are raped by strangers, casual acquaintances, dates, or steady boyfriends. Even on this relatively simple issue the findings are mixed. Some studies find that most rapes occur on casual dates. Belknap, supra note 174, at 205; Ward et al., supra note 881, at 65. Others find that more rapes occur in long-term relationships. C. Kirkpatrick & E. Kanin, Male Sex Aggression on a University Campus, 22 AM. SOC. REv. 52 (1957) (more advanced forms of sexual aggression found to have occurred in more durable relationships); Mary P. Koss et al., Stranger and Acquaintance Rape: Are There Differences in the Victim's Experience?, 12 PSYCHOL. WOMEN Q. 1 (1988); Muehlenhard & Linton, supra note 879, at 186; R. Lance Shodand & Lynn Goodstein, Sexual Precedence Reduces the Perceived Legitimae of Sexual Refusal: An Examination of Attributions Concerning Date Rape and Consensual Sex, 18 PERSONALITY & SOC. PSYCHOL. BULL. 756 (1992). The most methodologically sophisticated, national survey of sexual practices found that, of the women who had been "'forced to do something sexual that they did not want to do,'" 46% were victimized by someone with whom they were in love, and in another 9% of the incidents the perpetrator was their spouse. LAUMANN ET AL., Supra note 56, at 338. This finding is highly suggestive, but "forced," as understood by the respondents, does not necessarily mean raped in the legal sense. Cf. Shotland, supra note 178 (39% of victims dated the rapist after the rape).

989 Goodchilds et al., Adolescents and Their Perceptions of Sexual Interaction Outcomes, in RAPE AND SEXUAL ASSAULT II, supra note 314, at 245-70.

990 Shotland & Goodstein, supra note 988, at 756. Such attitudes are not confined to students. See Giarrusso, supra note 525, at 66 (43% of men and 32% of women consider aggressive sex appropriate if the parties have dated a long time).

991 See, e.g., KALVEN & ZEISEL, supra note 52, at 250.

992 See DRESSLER, supra note 62. The Model Penal Code retains the common law rule that a husband cannot rape his wife. MODEL PENAL CODE (Section) 213.1 (1985).

993 ESTRICH, supra note 1, at 107 n.3. Fifteen states had abolished the marital exemption for rape as of January, 1990. In 23 states it is punished less severely than non-marital rape, and in eight states, it is punishable only if the couple was living apart and in the process of legal separation or divorce. DIANA E.H. RUSSELL, RAPE IN MARINADE 375-82 (1990).

994 See, e.g., FED. R. EVID. 403 (relevant evidence may be excluded). See also McCORMICK supra note 796, (Section) 185.

995 On the other hand, the character of the average rape defendant is probably more suggestive of dishonesty than that of the average rape complainant, regardless of her sexual habits. Moreover, if he is guilty, he has an obvious, extremely powerful motive to lie about the rape, while in her case the suggested motive for a false accusation is usually weaker and more speculative. Given the burden of proof, however, any strong correlation between promiscuity and dishonesty about forced sex would tend to justify, if only sightly, acquittals in such cases.

The converse question also arises. What if the woman's character is exemplary, and she is a churchgoer? There is some evidence that churchgoers are less likely to be raped. BOURQUE, supra note 25, at 49. But this may be more than offset by their presumptively greater truthfulness, a possibility that raises multiple difficult issues, including whether characteristics such as honesty are general or situational, and, even if the suggested inference is valid, to what degree juries are likely to overrate its probative force. See generally Bryden & Park, supra note 310, at 561-65.

996 Powell v. Alabama, 287 U.S. 45 (1932); see also Norris v. Alabama, 294 U.S. 587 (1935); Patterson v. Alabama 294 U.S. 600 (1935).

997 A NEW DICTIONARY OF qUOTATIONS 1004 (H.L. Mencken ed., 1942).

998 See, e.g., S. Penrod & E. Borgida, Legal Rules and Lay Interference, REV. PERS. &Soc. PSYCH. 151 (1983); S. Sue et al., Effects of Inadmissible Evidence on the Decisions of Simulated Jurors: A MoralDilemma, 4J. APPLIED Soc. PSYCH. 345 (1973); S. Wolf & D.A. Montgomery, Effects of Inadmissible Evidence and Level of Judicial Admonishment to Disregard on the JuSts of Mock Jurors, 7J. APPLIED SOC. POCH. 205 (1977). For a recent review of the research on the effectiveness of jury instructions, see Kerwin & Shaffer, supra note 447, at 153-55.

999 The experts disagree about whether character traits such as honesty are situational rather than general. Bryden & Park, supra note 310, at 561-65.

1000 See supra text accompanying notes 603-719.

1001 See supra text accompanying notes 446-96.

1002 RICHARD O. LEMPERT & STEPHEN A. SALTZBURG, A MODERN APPROACH To EVIDENCE 162 (2d ed. 1982) (discussing prejudice in terms of regret matrix of jurors).

1003 Women who are raped by someone they know have to cope not only with the physical violation of their bodies, but also the betrayal of trust. JEAN O'GORMAN HUGHES & BERNICE R. SANDLER, "FRIENDS" RAPING FRIENDS: COULD IT HAPPEN TO You? 6 (1987) (explaining the adverse effects of date and acquaintance rape on the victims); MARY P. Koss & MARY R. HARry, THE RAPE VICTIM: CLINICAL AND COMMUNITY APPROACHES To TREATMENT 42 (1987) (discussing the difficult post-rape adjustment for acquaintance rape victims). In addition, acquaintance rape victims are more likely to be blamed for being raped, and therefore they may receive less emotional support from others, than women who were raped by strangers. Id. See also ROBIN WARSHAW, I NEVER CALLED IT RAPE 65 (1988) (citing a study by researchers at the Urban Institute that concluded acquaintance rape victims rate themselves less recovered than do stranger rape victims for up to three years following their rape experiences). On the other hand, some studies indicate that a substantial minority of rape victims continue to date the rapist, suggesting that acquaintance rapes are sometimes less devastating than modern scholars uniformly assert. See supra notes 181-82 and accompanying text.

1004 Cf MODEL PENAL CODE (Section) 213.1 (1985) commentary at 307 (stating that "when previous sexual liberties have been allowed and the persons involved are voluntary companions on the occasion of the offense, the gravity of the wrong is arguably less severe"). The MPC classifies aggravated rapes as first degree rape and simple rapes (parties acquainted, no physical injury) as second degree. It seems likely that many jurors agree that acquaintance rapes are less serious. If so, they presumably are more willing, all else being equal, to acquit in close acquaintance rape cases.

1005 Criminal law jurisprudence customarily distinguishes between general deterrence, deterring potential offenders in the community through the threat of punishment, and specific deterfence, deterring the defendant himself from committing the crime again by punishing him. See generally KADISH & SCHULHOFER, supra note 19, at 149.

1006 Jurors seem to be understandably eager to incarcerate rapists who use weapons. See, e.g., LaFree et al., supra note 66; LaFree, supra note 381.

1007 LEMPERT & SALTZBURG, supra note 1002, at 162.

1008 The psychological impact of a rape on a prostitute is perceived by many as being less acute than the impact on other victims. Silbert, supra note 314, at 76. The same authority believes that in reality the emotional impact of rape is compounded when the victim is a prostitute. Id. at 82-83.

1009 See, e.g., this cross examination:

Q: Isn't it a fact that at the point when your girdle came off, you assisted in the taking of that girdle off?.

A: Assisted? I wouldn't use the word assisted.

Q: Did you participate in taking that girdle off?....

Q: Did you in any way pull that girdle down?

A: No, I wouldn't say that, sir.

Q: Isn't it a fact that you helped those men take that girdle off your body?...

Q: Isn't it a fact, further, that you did not resist their taking off those underpants from your body?

A: That's not true. That's not true ....

Q: Is it not a fact that on the occasion of the third intercourse, you said to the man "come on, come on"?...

A: If I used the words "come on," it meant please leave me alone, come on, don't do this to me .... But I didn't say "come on" in the sense the other way.

United States v. Thorne, 406 F.2d 995 (D.D.C. 1969), quoted in Richard A. Hibhey, The Trial of a Rape Case, in RAPE VICTIMOLOGY 180-81 n.48 (Leroy G. Schultz ed., 1975). This is an example of the sort of brutal cross-examination that is permissible, and should be permissible, notwithstanding shield laws and other reforms. Although the questions all focus on the putative victim's behavior, they are all relevant both as to whether she consented and whether the defendants reasonably believed so. Our point is not that shield laws are useless, but that one should not expect a rape trial to focus on the defendant's behavior to the exclusion of the victim's behavior.

1010 CHARLES W. DEAN & MARY DEBRUYN-KOPs, THE CRIME AND THE CONSEQUENCES OF RAPE 86 (1982) (finding that rape trial victims are treated worse than defendants).

1011 See, e.g., FAroSTEIN, supra note 13, at 270-73 (remarking upon the urgent need to continue striving for improvement). See also Cassia Spohn & Julie Homey, "The Law's the Law but Fair is Fair:" Rape Shield Laws and Officials' Assessment of Sexual History Evidence, 29 CRIMINOLOGY 137 (1991);Wade Barber & Pat DeVine, Prosecuting Cases of Sexual Assault, 5 POPULAR Gov. 1, 1-5 (1985). The authors stress the positive effects of a number of reforms designed to make the victim's experience more bearable--for example, steps to protect the her from unnecessary interviews and court appearances and to advise her weekly about the progress of the case. Id. at 4-5.

~~~~~~~~

By DAVID P. BRYDEN[*] & SONJA LENGNICK[**]

[*] Gray, Plant, Mooty, Mooty & Bennett Professor of Law, University of Minnesota. All rights reserved, David Bryden

[**] Member of the Illinois bar.


Copyright of Journal of Criminal Law & Criminology is the property of Northwestern University School of Law and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Source: Journal of Criminal Law & Criminology, Summer97, Vol. 87 Issue 4, p1194, 191p.
Item Number: 181608

[Back]