I am doing a repeated measures ANCOVA.

My independent variable is visual clutter measured as a between-subjects 2 levels categorical variable (Higher visual clutter versus lower visual clutter). For either the Higher visual clutter or lower visual clutter conditions, the participants in those respective groups either see the SPORT TYPE soccer first or basketball first (Note that everything is randomized so there are no order effects).

The dependent variable is consumer satisfaction and I have measured this separately for basketball and soccer. so two continuous dependent variables satisfaction (SATS) soccer and SATS basketball.

In the covariates box, I added the moderator: the continuous variable escapism and I added the interaction effect of visual clutter and escapism gender and age

My supervisor who is now unavailable told me in her feedback that if the difference in the means of sport type is significant I should include it as a covariate in the ANCOVA model.

But either I completely misunderstood her feedback or I am correct that it is impossible to create a dummy variable for a within-subjects continuous variable (sport type satisfaction).

And even so, in the repeated measures ANCOVA, the within-subjects test shows that there is no significant difference in means between basketball and football. But a simple paired t-test shows that for the higher visual clutter condition, there is a significant difference in means for the satisfaction levels between basketball and soccer while in the lower visual clutter condition there is not.

Can I drop sports-type based on the results of the repeated measures ANCOVA, the within-subjects effect being insignificant and conduct a regular ANCOVA afterwards?

My thesis is due in five days and I would deeply appreciate the help.

Sebastian Brom