Dear all,
a brief question on posthoc tests: I have a design with two conditions (A, active - B, control). I assess between group differences (patient controls) with a two-sample t-test & the 1st level contrast A > B. I find differences between the groups & now want to determine whether patients specifically differ from controls in condition A - or whether there is actually a difference in the control condition. To this end I extract parameter estimates separately fro both conditions A & B - from a sphere centered at the between group differences. The plot shows that the group difference actually is larger in A than in condition B. How can I make a statistical inference about the differences to show that the difference was specifically in A. Just use 2-sample t-tests (patients vs controls) in a statistic software on the extracted parameter estimates & report the significance? Is there a more elegant way to determine the condition-specific effects?
Thanks in advance & best regards
Ben
|