Comparing Groups in Ordinal Logistic Regression


I seem to have gotten a little confused about the possibility of comparing groups in ordinal logistic regression.

Here is my issue:
I gave people one of three framing conditions for the goal of a task, which varied in the wording of how the goal was presented. Then I let people perform task, that resulted in a performance score and asked for their satisfaction with that performance (which was measured on an ordinal scale, hence the ordinal logistic regression). What I am trying to test now is, whether the participants' satisfaction with performance varies between framing conditions over the course of performance scores, i.e. whether the differences between regression slopes is singnificant. (So I have the continuous IV performance and the ordinal DV satisfaction, plus the three framing conditions)

When I run the regressions independently, test for different models and select the best fitting models for each framing condition individually, I get results that suggest differences in the framing conditions. However, I would like to test whether the between-group effects are significant -- this is my basic question: Whether, and if so, how this is possible?

From what I've found this seems to be a very particular issue with not a lot of information on it out there. Infering from this document however, I can not really do that: (p.20) I might however, be just confused, as I am quite new to ordinal logisitc regression as such.

Your help is very much appreciated,

Last edited: