I'm using spss for windows, and have a 3x2x2 (with the first two factors within subjects, and the third between) with a covariate. The full factorial model results results in the covariate being significant in only one interaction. I want to keep that interaction in the model, but drop all the other nonsignificant covariate interactions, and re-run the anaysis.

My question: first is this possible with spss? It sure doesn't seem possible with just using the windows, and looks like I'll need to get into the syntax.

If I am using the syntax, what command allows me to specify what covariate interactions are allowed into the model?

Any help would be much appreciated,

Reid