post hoc test for perManova

[Francisco Cabrera Rodriguez]

Hello,

I would like to compare plant life forms from samples of different ages. I  have used a perMANOVA, it was significant. But I do not how to use any post boc test to find where differences are!

I would appreciate any help.

Sincerely,

Fco.

[Bruce McCune]

Francisco,

 

At the bottom of the perMANOVA Setup dialog, check the box, "Make pairwise comparisons"

 

Bruce McCune

--
You received this message because you are subscribed to the Google Groups "PC-ORD" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pc-ord+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Francisco Cabrera Rodriguez]

Bruce,

Thanks a lot!

 My design is unbalanced. Is possible to make a perMANOVA with unbalances group sizes?  I want to compare 3 groups (unbalanced).

Sincerely,

Fco

[Bruce McCune]

With the version in PC-ORD (i.e. the pure nonparametric form of permanova, as presented in Anderson's original paper), you must have a balanced design. But Jeri Peck in her Step-by-Step book (see PCORD.COM) gives detailed instructions on how to use a resampling procedure to do this: making a stratified random sample of the largest possible size to achieve balance, then repeating this a large (1000X) number of times using PC-ORD's batch processing.

 

If you have a 1-way design, I would suggest MRPP as an alternative that is likely to give similar results to perMANOVA. PC-ORD does not require a balanced designed for a simple 1-way MRPP.

 

Bruce McCune

[Francisco Cabrera Rodriguez]

Thanks a lot!

Fco