Back to PLS HelpUntitled Post
Posted on 09/16/08 13:31:12
Number of posts: 29
j crofts
Hi,
I have been using seed pls (1 seed) and have been able to interpret the results. I am now doing behaviour PLS (more than 1 behaviour) and I am a little unsure how to interpret the results. For example, with 3 groups and 2 behaviours the matrix of group saliences is now 6 by 6, now say the first latent variable for the groups (right singular vector) is given by [0.51,-0.12,0.456,-0.876,-0.89,-0.654]. Do components 1, 3 and 5 correspond to behaviour 1, and 2, 4 and 6 behaviour 2? Can I conclude that the first LV distinguishes between group 3 and groups 1 and 2 in the case of the first behavioural condition? Also, that the second behavioural condition does not distinguish between groups?
Any help would be most appreciated.
Cheers,
Jonathan
I have been using seed pls (1 seed) and have been able to interpret the results. I am now doing behaviour PLS (more than 1 behaviour) and I am a little unsure how to interpret the results. For example, with 3 groups and 2 behaviours the matrix of group saliences is now 6 by 6, now say the first latent variable for the groups (right singular vector) is given by [0.51,-0.12,0.456,-0.876,-0.89,-0.654]. Do components 1, 3 and 5 correspond to behaviour 1, and 2, 4 and 6 behaviour 2? Can I conclude that the first LV distinguishes between group 3 and groups 1 and 2 in the case of the first behavioural condition? Also, that the second behavioural condition does not distinguish between groups?
Any help would be most appreciated.
Cheers,
Jonathan
Replies:
Untitled Post
Posted on 09/16/08 16:00:15
Number of posts: 394
Hi Jonathan,
Your interpretation of the singular vector is perfect. You may also want to check the variable "lvcorrs" or "orig_corr" which are the correlations of the two behaviors with the brain scores by group. In the bootstrap results there are also ulcorr and llcorr that are the upper and lower bounds of the confidence interval around the correlations. These are helpful in grounding your interpretation inferentially.
cheers
Randy
Your interpretation of the singular vector is perfect. You may also want to check the variable "lvcorrs" or "orig_corr" which are the correlations of the two behaviors with the brain scores by group. In the bootstrap results there are also ulcorr and llcorr that are the upper and lower bounds of the confidence interval around the correlations. These are helpful in grounding your interpretation inferentially.
cheers
Randy
Login to reply to this topic.
