There is no need to weight by number of trials. The GLM will do fine without anything like that, and adding weights may result in unexpected side effects (false positive or negatives).
Unbalanced designs are Ok, especially since it doesn't seem so bad (it not like 10 trials vs. 70).
Good luck,
Colin Hawco, PhD
Neuranalysis Consulting
Neuroimaging analysis and consultation
www.neuranalysis.com
[log in to unmask]
-----Original Message-----
From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] On Behalf Of Michael L.
Sent: August-25-17 6:12 PM
To: [log in to unmask]
Subject: [SPM] weigth by number of trials
Hi,
I'm posting again my question from a few days ago, because there were no replies and I'm still not sure how to continue with my analysis.
I have four conditions at first level: taskA-conditionA, taskA-conditionB, taskB-conditionA and taskB-conditionB. I was wondering if I should weight my contrasts by the number of events in each condition. For example, for testing the taskA-conditionA effect should I use [1 0 0 0] (no weighting) or should I weight this contrast by the number of events in taskA-conditionA?
If I have 11 trials in taskA-conditionA, 10 trials in taskA-conditionB, 14 trials in taskB-conditionA and 11 trials taskB-conditionB, how can I do the weighting properly?
Thank you.
Best regards,
Michael
|