This is a synopsis of Lindblom's classic article (1959), "The Science of 'Muddling Through'". We are studying it to highlight a long-standing division in public administration approaches. The phrase, "muddling through", comes of course from the pride the British take in "muddling through somehow", seeing their bureaucratic decision process as a muddle but still getting the job done. But why the oxymoron of calling this muddle a "science"? This was Lindblom tweaking the whiskers of the dominant approach to public administration at the time: an approach that believed that scientific analysis could solve the political problems faced by public administrators.

The late 1950s, when the article was written, was the heyday of rationalism, where it looked like various forms of rational / technical analysis (game theory; optimization algorithms) would bring a scientific approach to what had hitherto been a distasteful political process. (This was the period when Robert MacNamara and his "whiz kids" were revolutionizing the administration of Ford through use of such analyses. He went on to attempt the same thing in the Pentagon, where he ran into the political realities of the various armed services.) This same feeling—that everything could be solved by a scientific approach—was found in social science generally with the domination of "behaviorism". (Which continues to this day [2003].)

Lindblom essentially says that a rational / technical approach is not possible, and he offers an alternative—or at least an explanation of why the despised political maneuvering of administrative decisionmaking may not be as bad as it appears.

Lindblom's perspective is seen in two other areas related to public administration: budgeting (incrementalism [Wildavsky]) and political structure (pluralism, e.g., Robert Dahl). In economics it is related to Nobel Prize-winner Herbert Simon's work on limited rationality.

The "Root" (a.k.a. "Rational-Comprehensive" or "Means-Ends Relationship" ) Method

• Specify all ends (distinct from means). [NOTE: "Ends" means "values to be pursued" or "goals", not "specific outcomes" or "ways the situation should end up". Thus "going to law school" is probably not an "end" to you; rather, it would be one specific means toward satisfying the end of "being able to get a job and survive".]
• Specify weights for all the ends. In other words, indicate how important the various goals are relative to each other. [Note the assumption that different parties will be able to agree on the relative importance of the various goals.]
• Examine all possible sets of means. [Note the assumption here that we can think of all possible sets of means—or even all reasonable sets of means.]
• Evaluate each set of means against ends, assigning a score to how well the given set of means achieves each end. [Note the assumption that we have the ability and resources to do these evaluations.]
• For each set of means, calculate its overall measure based on the weighted average of its scores on achieving the different ends.
• Choose the set of means with the highest weighted score.

The "Branch" (a.k.a. "Incremental") Method

• Ends and means are intimately intertwined, i.e., we often know our ends only from consideration of the means we are contemplating.
• Only a few means are considered ... [Assumes that managers have limited time and other resources (including information) to decide, so they can only do non-comprehensive analysis.]
• ... and only those which don't represent too much of a departure from the status quo. (Thus the name, "branch method", where each policy branches off to another.)
• Evaluation of the means is crude, in that many consequences are ignored. (This ignoring of consequences often occurs because a full analysis is not possible.)
• Choice among the means is determined by agreement among interested parties rather than by summary indicators arising from the analysis.
• Agreement is the only empirical indicator of virtue, because values are not usually clear-cut or even shared.

Conclusion:  Public administrators must concentrate on agreement on actions (actual policies), not on abstract arguments for adopting those policies.

Comparing Two Approaches to Administrative Management and Decisionmaking

Questions for Study and Review

1. Lindblom says a lot about the interrelationship of means and ends in decisions. (a) Take any reasonably complex decision in your own life (e.g., what to major in; who to marry; where to attend college or graduate school; what car to buy; what friends to invite over for pizza), and try to list and rank all the ends without consideration of the available means. (You can stop after you've listed at least fifteen ends.) (b) Discuss how successful you were in separating means and ends. [The example we used in class was helping a friend celebrate his 21st birthday, presumably by taking him out to a restaurant for dinner. Another example discussed briefly in class was that of the Director of Parks & Recreation considering what policy opportunities exist to meet complaints by high school students that there aren't any places in Duluth that have things for people their age to do.]
2. Analyze the following dialogue you might have with yourself one evening in terms of Lindblom's perspective:
• "What book should I read?"
• "Well, what are my goals?"
• "Amusement. Learning something about people. Increased professional knowledge."
• "What about that book lying beside my bed?"
• "Yeah!--and that reminds me that my bedside is too messy, so I guess I should add 'cleaning up messy spots in the house' to my list of goals. Now, where does that leave me?"
• [As youlook around, you notice that messing up the top of your TV there are four videotapes you ordered but haven't watched. "Maybe I should make that 'reading books and viewing videotapes'."
3. According to Lindblom, what problems prevent us from doing comprehensive analyses, and how do they prevent us?
4. "The science of muddling through" is an oxymoron (self-contradictory phrase). In what sense can Lindblom call his "branch" method a "science"?