Hi, I apologize in advance if this question is obvious... I have a group with N=17, four conditions (all within-subject). I have two planned contrasts between conditions (1 -1 0 0 and 0 0 1 -1). So, I ran an ANOVA, which gave me [4,64] as dfs for the omnibus. For the contrasts, i get t tests with 64 degrees of freedom.
Now, had I run t tests directly on the pairs of conditions of interest (i.e., without running the ANOVA), the degrees of freedom would have been 16 (instead of 64) for each comparison. So, I'm confused, which one is "better"? What is the rationale for one vs the other? Obvioulsy, I get more activations when I run the contrasts within the ANOVA....
Thanks in advance...
-Mike