Back to PLS HelpPosted on 04/27/14 19:55:16
Number of posts: 6
We are running a task pls (using the command line) with 2 groups and 48 tests. We set up 2 datamats, one for each group - with 48 columns in each datamat (rows equal to number of subjects). We have one significant LV and would like to know the saliences (and high/low CIs) for each test (column) for each group. However, we can only get overall saliences - not specific to each group.
To get the saliences we are using: result.boot_result.u
For CI's we are using: CI=3*result.boot_result.u_se(:,1)/result.s(1) ; CI_low=(result.boot_result.u)-CI ; CI_high=(result.boot_result.u)+CI
Can anyone in PLS land give us some advice on this?
Thanks!
Replies:
Posted on 04/27/14 20:25:46
Number of posts: 394
Hi Carrie,
the saliences are best thought of as parameter estimates for an effect, like a regression coefficient. If I understand your design, the saliences would reflect how strong each test reflects a group difference. You can look at the group means on the test to understand the nature of the effect.
Randy
Posted on 04/27/14 21:16:32
Number of posts: 6
Thank you!!
Carrie
Login to reply to this topic.
