Hi all,
I have a question. My study design is as follows:
Memory Task
Old words (emo, neu) + New words (not presented before)
Subject's performance measured: hits/misses/false alarms/correct rejections
I would like to compare hit-neu words with hit-emo words, neu words with emo words, hits vs correct rejections, & old vs new.
I've started by specifying 6 conditions hit-neu, miss-neu, hit-emo, miss-emo, cr and fa. Note that some subjects have 0 misses.
I've also created contrasts:
#1 hitneu > hitemo
#2 hitemo > hitneu
Q1: How can I create contrasts for the comparisons hits vs cr, old vs new and neu vs emo? The problem is that when I try to create a contrast for emo vs neu by using the vector [-1 -1 1 1 0 0), it says invalid contrast. What could be the issue here?
Q2: Some subjects have 0 misses (hit prob = 1). Since there are no onsets specified for the 'miss' condition, doesn't that pose a problem in the GLM?
Q3: I want to compare hitneu vs hitemo (emotion is a within-subject variable). Under 2nd level analysis, should I use the one-sample t-test option or the paired-samples t-test option?
Thanks!
Karen
|