• Can you do a repeated measures multinomial logistic regression using SPSS?


I need to do a regression on data at two points of time and I think this maybe the only way to go(?).

To elaborate: I work for a national health service supporting individuals with psychosis. I want to investigate whether any factors (age, gender etc) predict the vocational outcome of a client group (3 categories: partial vocation, full vocation or no vocation) at entry of service and at 18 months into the service. I want to consider whether there is a difference at 18 months compared to entry to the service in regards to vocational status.

  • 2
    $\begingroup$ I don't use SPSS, so I can't help with that, but I think you want to at least explore an ordinal logistic model. If the assumptions are met, you will have a much simpler model to deal with. $\endgroup$
    – Peter Flom
    Jan 27, 2011 at 13:18

1 Answer 1


One way is to build an SPSS PLUM or NOMREG model that checks for an interaction between each predictor and a binary predictor, “time.” In that scenario you'd use just a single column for all the values of your outcome variable. For 1/2 the data set, time would be marked 0, and for the other half it'd be marked 1. Essentially you’d be treating time as if it were like gender or any other binary predictor that potentially could interact with other predictors.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy