Sociology 3155: Cumulative exercise 3-- 50 points possible

Voting Behavior: the 2000 Election

This is one of the best survey studies of the 2000 election, in which Bill Clinton defeated incumbent president George W. Bush. Stratified sampling was used to select a nationally representative sample, and more than 200 questions were asked. This makes it a nice final excercise for our class. I hope you do well with it and that you enjoy it. Bring it to our regular classroom at 2 p.m. on Monday, May 11, or to my office (Cina 104C) before that time.

You will have received the data set by email, and in class I will provide a copy of the codebook. You may work singly, in pairs, or even in a three-some on this assignment. You may email the SPSS to me as an attachment, and turn in a detailed commentary on what you've done and the conclusions you reached in a separate, typed document. Or you may copy the SPSS tables to Microsoft Word, reduce the size of the tables, and turn in the whole document on Monday, May 11, or before.

The challenge in this exercise is to figure out on your own what kinds of statistical procedures are needed to describe voting patterns in the 2000 federal elections and to find the important causal relationships in this data. (In other words, I'm not going to tell you which crioss tabs to run, which regressions, or which control variables. I will give you just a few overall goals.

Goals of this exercise:
1. To establish the major voting patterns in the 2000 election in relation to the various demographic variables.

2. To find out which of the other variables are the best predictors first, of whether people vote or not; second of what candidate/party they support; and third, of whether they engage in other kinds of political activity (donating to candidates, for example). It will probably also be interesting to see whether the party they vote for in the congressional elections is the same as in the presidential campaign.

3. To add some control variables for the picture... for example, if concern about foreign affairs and defense is a good predictor of Republican voting, does this dynamic work the same for both women and men, or for different races.

4. Notice that there's a special section of interval/ratio variables, allowing both bivariate and multivariate analysis. Include some of these and their interpretation, if you are able.

5. You may want to specialize somewhat, concentrating on some kinds of items to the exclusion of others. I'm expecting 3-5 hours of work on this assignment, including class time.

Notice that the questions are divided into categories which might be helpful in your initial thinking about patterns and relationships. These categories are:

A. Voting behavior and related items
B. Political involvement items
C. Media involvement items
D. Candidate image items
E. Presidential approval and government performance items
F. Economic Conditions items
G. Ideology Items
H. Social welfare and domestic spending items
I. Social Issue Items
J. Civil rights and equality issue items
K. Environmental Issue Items
L. Foreign Affairs and defense Issue items
M. General Government Orientation Items
N. Demographic Items
O. Interval-level variables for further analysis