Dear colleagues, I am a PhD candidate. I am going to offer a conceptual model on teacher professional development. this model has one mediatory factor. I want to first test the mediation through Barron and Kenny or Bootstrap and then use SEM. but many say doing mediation analysis is an extra procedure. my questions are:
1. Is doing mediation analysis an extra thing to do? in addition to testing the model fit, does SEM really show the mediation and the degree of the effect of exogenous on endogeneous like regression coefficient?
2. in my PhD dissertation, should I write just one general question ( does x mediate the relationships between exohenous factors and endogenous factors? or I can write separate distinct questions for each path in the model (e.g. does x predict y?)
I reallllllllllllllllllllllly appreaciate your insightful answers