I am currently writing my thesis and I ran into some problems. My experiment has a 2x3 design (2 ethnicities x 3 types of advertisements). My main measures are “Felt targetedness” (how much they feel targeted by the ad) and “Attitude toward the ad” (how much they like the ad). The respondents (from 2 different ethnicities) were asked to view and rate one of the three advertisements (between design).
I want to test if felt targetedness mediates the variables ethnicity and attitude toward the ad by using regression. Do I need to do this on a total level or do I need to test this per advertisement type (so 3 tests in total)?
Also, after an ANOVA I used a LSD post hoc test but I have also seen articles using a Turkey post hoc. I tried to look up the differences but I do not really understand them. Which one is more common to use and why?