No significant LV's!
jcrofts
Posted on 07/04/09 11:50:46
Number of posts: 29
j crofts
Hi,
If the permutation tests return only one significant latent variable - which can't be used to infer differences between classes right? Does this mean my data is garbage or that PLS (in this form anyway) is inappropriate?
Cheers,
Jonathan
Untitled Post
rmcintosh
Posted on 07/05/09 12:51:03
Number of posts: 394
Its hard to know how to interpret your outcome unless you can be a bit more specific on the pattern on your significant LV. In general, however, I wouldn't say that non-significant results mean either that your data are garbage or that PLS is inappropriate - it really depends on what question you are trying to answer. We have data sets where there were no significant changes in activity, yet quite robust brain-behavior and functional connectivity patterns were identified. Given the richness of imaging data, its not wise to focus on a single analysis to make the determination of whether you have anything in your data set.
Randy