Entering predictors sequentially for multinomial logit?

#1
Hello,

I have a set of data with a 3-level outcome variable and multiple predictor variables; I am testing a hypothesis about whether a certain continuous predictor has a bigger effect on the outcome for one group (a level of a categorical predictor) than another group.

I understand that, with 3 levels of the outcome variable, I should use multinomial logistic regression. Since I also want to test the significance of interactions between predictor variables, I assume I should input variables sequentially. However, in SPSS (which is what I'm using for analysis right now), it seems that multinomial logistic regression doesn't allow sequential entry of independent variables (unlike binomial logistic regression, which does).

Does anyone have suggestions about other software packages that would allow this, or other analyses which would be more appropriate?

Thank you very much,
Steve
 
Last edited:
#3
Thank you very much, I'll try that out!

I also looked in another stats book recently and it looks like I can do the sequential regression myself, even in SPSS, by just doing a baseline regression and then separately running the next step of the regression, and calculating the model chi-squared test by hand based on the baseline and new chi-squareds.