I want to compare three groups (three different ethnic groups) on one dependent variable that is categorical (3 categories). What test do I use?
I am using SPSS, so some advice on how to do it would also be greatly appreciated. Thank you!
2
- $\begingroup$ What is your dependent variable? I don't assume it's ethnicity membership... $\endgroup$Stephan Kolassa– Stephan Kolassa2017-05-08 08:29:45 +00:00Commented May 8, 2017 at 8:29
- $\begingroup$ The dependent variable is voting behavior with three possible categorical outcomes. Ethnicity membership is the variable used to create the groups from the sample. $\endgroup$Stefan– Stefan2017-05-08 08:33:09 +00:00Commented May 8, 2017 at 8:33
Add a comment |
1 Answer
$\begingroup$ $\endgroup$
If all you are interested in is the relationship between a single categorical IV and a single categorical DV, then a contingency table analysis is appropriate. You will find a lot of info online and in the SPSS help. This may be a good place to start.
If you have more IVs, and possibly also interactions, things get a little more complicated.
Your response is categorical. Specifically, it is multinomial, i.e., you have more than two possible outcomes. (With just two outcomes, you could use a logistic regression. With three outcomes, that is not a straightforward possibility.) The standard approach would be a multinomial logistic model, with your groups and other IVs as predictors, potentially modeling interactions. This earlier question might be helpful (and also contains SPSS info): Multinomial logistic regression vs one-vs-rest binary logistic regression
