Back to PLS HelpPosted on 09/14/14 20:15:03
Number of posts: 26
Hi experts,
I have already run a PLS analysis on 3 groups with 3 conditions and it has worked brilliantly. I now want to analyse the trials that were correct and this results in uneven groups for the conditions. I tried running this and it returned an uneven matrix columns error message. Am I able to do this? Any assistance is greatly appreciated.
Regards
Andrew
Replies:
Posted on 09/15/14 07:03:29
Number of posts: 394
Hi experts,
I have already run a PLS analysis on 3 groups with 3 conditions and it has worked brilliantly. I now want to analyse the trials that were correct and this results in uneven groups for the conditions. I tried running this and it returned an uneven matrix columns error message. Am I able to do this? Any assistance is greatly appreciated.
Regards
Andrew
Hi Andrew,
I need a bit more detail on what exactly you are trying to do on the follow-up analysis. When you say uneven groups per condition, does that mean that for some groups one of the three conditions does not exist?
Posted on 09/15/14 07:25:39
Number of posts: 26
Hi,
Not necessarily, just whereas before all 3 conditions contained 4 video blocks, I now want to allocate the videos according to the subjects response and therefore the numbers in some conditions are anywhere from 1-8 and uneven between subjects. I do have a couple of subjects with 0 for one of the conditions but I had removed these.
Regards
Andrew
Posted on 09/15/14 07:32:04
Number of posts: 394
Hi,
Not necessarily, just whereas before all 3 conditions contained 4 video blocks, I now want to allocate the videos according to the subjects response and therefore the numbers in some conditions are anywhere from 1-8 and uneven between subjects. I do have a couple of subjects with 0 for one of the conditions but I had removed these.
Regards
Andrew
Ok, so then I am not sure how you can compare conditons when they are not balanced across the groups. What question are you asking of the data? Can I get a bit more detail on the design
Posted on 09/15/14 08:46:07
Number of posts: 26
I am looking at the network commonalities and differences across the 3 conditions and 3 groups. If one condition has 2 videos in for one participant and 4 for the next participant, can it not look to see which regions covary?
Posted on 09/15/14 08:54:41
Number of posts: 394
I am looking at the network commonalities and differences across the 3 conditions and 3 groups. If one condition has 2 videos in for one participant and 4 for the next participant, can it not look to see which regions covary?
yes, you should be able to so long as all subjects have one video. When you are creating the sessiondata files for this, are you merging across or within runs? Also, is this an event-related or block design?
Posted on 09/15/14 09:06:48
Number of posts: 26
block design and i believe it is within runs (but need to check tomorrow)
I'm Online
Posted on 09/15/14 09:35:28
Number of posts: 229
I am looking at the network commonalities and differences across the 3 conditions and 3 groups. If one condition has 2 videos in for one participant and 4 for the next participant, can it not look to see which regions covary?
the challenge will be with the interpretation - if you have widely varying response numbers across subjects or across group/condition, then you will also vary the SNR with respect to the signal - As Randy says, you can run the analysis with only one observation in a condition(s) for a given subject(s), but your confidence that the signal actually reflects the task demands will be diminished -
nancy
Login to reply to this topic.
