Dear all,
when analyzing my data, I tried two approaches:
first, I calculated first-level con-images of my two conditions to then t-
test those con-images against 0 on the second-level.
Secondly, I calculated first-level con-images of each of my two conditions
vs. 0 to then t-test those con images against eachother in a factorial 2nd-
level design.
Why aren't the two approaches equivalent (the first approache always
yields much higher t-values than the second)? Which one should I choose?
I hope you can help me. Kind regards,
Tim
|