ANOVA- interaction effects

#1
hey,
for my dissertation i am conducting a 2 (race) x2 (emotion) within subjects anova. I've basically conducted the Anova and found a significant interaction. I then decided that to look at this interaction in more detail i needed to conduct paired t-tests. I've conducted them but im not sure if i have done it correctly. Basically because it's a within subjects design my accuracy results consist of one condition of race and one condition of emotion together (for instance black faces exhibiting neutral expression/ white faces exhibiting angry expression). For the paired t-tests i need to look at race and emotion seperately. So for instance to look at race would i conduct a paired sample t-test on black neutral and black angry faces and then do another for white neutral and white angry faces? then for emotion would i differ the race within the paired test and then change the emotion between the two tests? im not sure if this makes much sense?
thanks!!!!!
 
#2
Actually, since you have only 2 levels of each factor, you don't need any post-hoc tests at all.

The interaction is telling you that the difference in means between neutral and angry faces is different for black faces and white faces. Just look at the means to see which difference is bigger.

Karen