Hello Rich, I may be able to provide some insight into this... Apologies in
advance if my response is too big. (First time user)
1:
I would think they are included because of familiarity and different
usage scenario's. (see 2). ANOSIM has some advantages under very
specific usage scenario's, which would make it hang around, and also,
when it comes to things like this, where there are multiple options for
multiple different usage scenarios, it makes sense to keep as many
packages "online": This would provide alternative choices in the event
of a package failing to operate, or user preference.
2: The main thing that ADONIS has over ANOSIM/MRPP is partitioning of variance,
allowing for interaction terms.
ANOSIM
isn't as nice in it's approach, simply clumping together "replicates"
and doing seperate permutations (As you know already). mrpp also does
this (Does not allow for interpretation of complex experimental design).
ANOSIM/MRPP
would be used for simple experiments, where you would believe there to
be no interactions, typically 1 factor designs (eg; Is there a
difference in species composition at two sites on an urban road). Also,
the similarity co-efficients used are different; As far as I am aware,
ANOSIM uses Bray-Curtis measures, whereas MRPP has got options.
This
further differentiates ANOSIM from MRPP: Essentially, MRPP can be used
for different data types (in combination with experimental designs)
wheras in my experience, ANOSIM is restricted to Bray-Curtis, eg: I'd
use MRPP for 1 factor presence/absence, and ANOSIM for 1 factor
abundance indices.
Having said all of that; from my personal
experience I'd recommend to just default to ADONIS: It allows for
multi-factorial designs with interactions, multivariate designs with
interactions, the amount of information it retains is better than ANOSIM
and MRPP (ANOSIM/MRPP are rank based, whereas ADONIS uses the actual
indice), and generally speaking, the results are easier to interpret.
Cheers
Date: Mon, 4 Nov 2013 08:17:13 -0800
From: ml-node+s471788n757848...@n2.nabble.com
To: alexander.angelo...@hotmail.com
Subject: Multivariate Analyses of Ecological Communities
I've read the docs for anosim, adonis, and mrpp and am currently
reading
the vegan tutorial. Toward the end of the anosim and mprr docs the author
writes that each has potential shortcomings and recommends the adonis
package be used instead. This leads to my two questions:
1) Why are anosin and mrpp packages still maintained and included? That
they are suggests question
2) Under what circumstances are these two packages appropriately used?
TIA,
Rich
--
Richard B. Shepard, Ph.D. | Have knowledge, will travel.
Applied Ecosystem Services, Inc. |
<http://www.appl-ecosys.com> Voice: 503-667-4517 Fax: 503-667-8863
_______________________________________________
R-sig-ecology mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
If you reply to this email, your message will be added to the
discussion below:
http://r-sig-ecology.471788.n2.nabble.com/Multivariate-Analyses-of-Ecological-Communities-tp7578489.html
To unsubscribe from r-sig-ecology, click here.
NAML
--
View this message in context:
http://r-sig-ecology.471788.n2.nabble.com/Multivariate-Analyses-of-Ecological-Communities-tp7578489p7578493.html
Sent from the r-sig-ecology mailing list archive at Nabble.com.
[[alternative HTML version deleted]]
_______________________________________________
R-sig-ecology mailing list
R-sig-ecology@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
