My research is related to vicarious embarrassment, which is an embarrassment because of the actions of other people (the protagonist). My hypothesis is that the friendship relationship with the protagonist who violates the norm will have a higher vicarious embarrassment than the vicarious embarrassment in the protagonist's condition as a stranger.

I ran a repeated measures analysis with a within-subject factor RELATION that has two levels: friendship X strangers. My dependent variable is vicarious embarrassment. The results of the analysis show that the relationship factor has a significant impact on VE. The difference between the conditions of friendship X strangers is also significant.

Besides the main effect analysis, I suspect that friendship collectivism functions as a covariate. I included it in the analysis as a covariate.

Problem:

When friendship collectivism was included in the repeated measures GLM analysis as a covariate the results were as follows.

in the table of tests of within subjects contrasts, the RELATION factor changed to be insignificant, and the interaction between RELATION X friendship collectivism was also not significant. I also don't know how to interpret the parameter estimates table. Can anyone help interpret these results? Thank you very much for your time and knowledge.