Data:  K:\nd.edu\user22\yli\Public\593sp01\Data\gss_logit.dta

You are interested in the determinants on people's opinion on interracial marriage (whether it should be banned or not).  Specifically, you want to assess the effects of race, sex, and education on one's attitude on interracial marriage issue.  Use logistic regression technique and the GSS data to examine these hypotheses.  Try a few different models (you may add other independent variables that you think are important) and identify one that best satisfies your research curiosity within the limit of the data.

Interpret the results in the following way:

• Does your best model significantly explain the variation of individuals' response to the interracial marriage issue?  Is it better than the other models you tried?  Cite/calculate supporting statistics.
• What are the significant predictors of the opinion?  What impact do they have on individual's opinion? i.e., Are people with more education more likely or less likely to oppose interracial marriage?  Are whites more likely or less likely than non-whites to oppose interracial marriage?  What about men and women?  Are men as likely as women to be against interracial marriage?  Are there any interaction effects among race, sex, and education?
• What do the slope coefficients tell us?
• Give a few substantial examples of your prediction.  For example, what is the chance of a white woman with average years of education to be against interracial marriage?  What is the chance of a non-white man with a college degree (approximately 16 years of education) to be against interracial marriage?  Use both odds and probability in your discussion.  (You may present this in a table.)