How to conduct a repeated measures regression with ordinal outcome variable?


I am trying to predict an ordinal variable (success: 0, 1, 2, 3). I have 3 predictor variables I want to include in the analyses together with their interaction terms.

The first predictor (skill level: expert, novice) contains independent groups. The second (# of Info statements of two types; critical, non-critical) and third (# of goal statements of two types: prediction, decision) are repeated measures (on an interval scale)

The issues I am having problems with are (a) regression that has repeated measures, and (b) use of an ordinal variable as the dependent measure.

If I didn't want to predict success, in ANOVA terms this would simply be a 2 [skill level-BP] x 2 [Info type-WP] x 2 [goal type-WP] ANOVA. However, I really want to predict 'success', and given that I have repeated measures on some factors (that are also correlated) a mixed model seems to be the best way to go. Problem is, I am mostly familiar with SPSS and it doesn't seem to like the ordinal dependent variable in MIXED Linear.

Am I on the right track with linear mixed models? If the ordinal dependent variable cannot be handled within SPSS, can anyone suggest a package that can do this?