Back to PLS HelpBetween groups PLS and unequal group sizes
Posted on 04/06/09 09:08:17
Number of posts: 6
Dear PLS experts,
I have a question regarding how robust the PLS procedure is in situations with unequal group sizes when running the analysis on two groups, either using SVD or non-rotated PLS.
It seems to me that since the statistical assessment is done with permutation and bootstrapping tests that do not require the data to meet parametric assumptions, the estimates of significance and reliability should be unaffected by differences in group size. But is there something I'm missing there? (I have one group with 21 and another with 23 participants.)
Thank you very much, any advice appreciated!
Zara
I have a question regarding how robust the PLS procedure is in situations with unequal group sizes when running the analysis on two groups, either using SVD or non-rotated PLS.
It seems to me that since the statistical assessment is done with permutation and bootstrapping tests that do not require the data to meet parametric assumptions, the estimates of significance and reliability should be unaffected by differences in group size. But is there something I'm missing there? (I have one group with 21 and another with 23 participants.)
Thank you very much, any advice appreciated!
Zara
Replies:
Untitled Post
I'm Online
Posted on 04/06/09 10:31:23
Number of posts: 291
PLS procedure is able to handle unequal group sizes when running the analysis on two groups. i.e. you can have one group with 21 and another with 23 participants. However, please make sure that the conditions for all the groups (including the number of conditions and the conditions' names) must be exactly the same for comparison.
Untitled Post
Posted on 04/06/09 10:33:48
Number of posts: 394
The permutation test is good guard against bias from differences in sample size since it creates new groups based on the N's you specify for each group.
Login to reply to this topic.
