Hi. I have a dataset that contains two categorical variable. One is something I varied in a scenario (number of witnesses; 0, 1, 2, or 3) and the other is a dichotomous factor given by my subject about how they would plead in court (guilty=1 or not guilty=2). I want to determine if the number of subjects giving each plea varies as a function of the number of witnesses present. I am guessing this would be something like a 4 X 2 chi-squared test, if such a thing even exists...is that right? If not, could you guide me to the correct test? My second question is: if the overall chi-square is significant (assuming that is the correct test), how do I do a post-hoc analysis to determine which levels differ from the others? I should also mention that I do not have equal N's in all of my conditions. Also, I will be doing the analysis on SPSS (PASW 17) if you can give any advice on how to accomplish the analysis with this software that would save me posting the same question in the SPSS group. I apologize if this question doesn't make any sense. It has been about 30 years since I last had a stat course Any help will be much appreciated.