Back to PLS HelpNon-rotated PLS question
Posted on 11/29/09 23:51:19
Number of posts: 16
Hi,
I've just started using PLS and have a simple question regarding non-rotated PLS.
My data has three conditions and I want to run a non-rotated analysis that compares conditions A and B to condition C.
I am unsure whether the contrasts should be specified as:
1 1 -1
or
0.5 0.5 -1
Any help would be greatly appreciated.
Thanks!
I've just started using PLS and have a simple question regarding non-rotated PLS.
My data has three conditions and I want to run a non-rotated analysis that compares conditions A and B to condition C.
I am unsure whether the contrasts should be specified as:
1 1 -1
or
0.5 0.5 -1
Any help would be greatly appreciated.
Thanks!
Replies:
Untitled Post
I'm Online
Posted on 11/30/09 10:09:08
Number of posts: 291
I suggest you to use -1 0 1, so that you can distinguish all 3 conditions.
Untitled Post
Posted on 11/30/09 14:50:17
Number of posts: 16
quote:
I suggest you to use -1 0 1, so that you can distinguish all 3 conditions.
For this contrast I am interested in the commonalities between A and B, and how they both differ from C, and am not sure which of the above contrast weightings in my first post is correct. I suggest you to use -1 0 1, so that you can distinguish all 3 conditions.
Thanks.
Untitled Post
Posted on 12/01/09 04:36:04
Number of posts: 60
Robert!
The non-rotated PLS has the advantage of allowing a direct assessment of hypothesized experimental effect.The only thing that you should be aware of is to have orthogonal contrasts (the inner product should be zero) if you are planning to specify more than one contrast otherwise you would have a hard time to interpret the contrasts in an independent way! If you have only one, I would go for .5 .5 -1.
/Alireza
The non-rotated PLS has the advantage of allowing a direct assessment of hypothesized experimental effect.The only thing that you should be aware of is to have orthogonal contrasts (the inner product should be zero) if you are planning to specify more than one contrast otherwise you would have a hard time to interpret the contrasts in an independent way! If you have only one, I would go for .5 .5 -1.
/Alireza
Login to reply to this topic.
