Turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- JMP User Community
- :
- Discussions
- :
- Discussions
- :
- If I do a LSM means planned contrast for an anova with Fit Model, are the result...

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Jul 27, 2016 10:58 PM
(2099 views)

If I do a LSM means planned contrast for an anova with Fit Model, are the resulting p values adjusted for the multiple corrections? If not, how do I adjust the p- value?

1 REPLY

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Jul 29, 2016 6:42 AM
(1961 views)
| Posted in reply to message from nsalmano_stanfo 07/28/2016 01:58 AM

They're not adjusted. If you devise orthogonal planned contrast, then you don't need to adjust your p-values because this would be equivalent to partitioning your data so that each contrast involves nonredundant pieces of information. This is why orthogonal planned contrasts can be so beneficial! Normally, with orthogonal contrasts, one wouldn't be interested in the omnibus F test, but it's interesting to note that summing the sum of squares of all your orthogonal contrasts will add up to the sum of squares between of the omnibus ANOVA. This makes sense because it confirms the notion that one is partitioning the sum of squares between into a series of nonredundant tests.

If your planned comparisons are not orthogonal, then you do have to adjust your p-values and there are many alternatives for doing so. It's hard to recommend a specific approach without knowing more about your design. Some approaches are more conservative than others (e.g., Bonferroni), but you can also look into Dunn's test or Scheffe's tests.

Laura C-S