Hi,
I read the e-mail reply, 'Re: covary with emotion scale' from Dr.
Weissman about performing t-contrasts in the fix effect analysis and
using the results to correlate with outcome measures. I also appreciate
his message greatly as I am working on a similar problem.
I have a related question. I did a longitudinal study on multiple
subjects. But some of them had 1 run of 60 time-frames (scans), and
others had 2 runs of 70 frames each. Most subjects had the same no. of
frames from the studies at both time points. But, a couple of subjects
had 1 run of 60 frames on day 1 and 1 or 2 runs of 70 frames on day2.
When I perform the fix effect analysis, should I enter 1s (ones) and -1s
on all the runs (of all subjects, both time points)? Or, should I
calculate the ratio of the total number of frames between each subject
and each time point and enter those ratios to counter-balance the
statistical power of each subject at each time point?
Any advice or suggestion will be greatly appreciated.
Yi-Fen
@Wake Forest
|