"The use of Canonical Correspondence Analysis depends if you have abiotic data
and if you want to test the
relationship among biotic and abiotic variables."
Yes, but the abiotic variables can simply be nominal ("dummy") variables
representing "treatments vs. controls" or even "before and after".
It is true that many people use RDA or CCA for continuous variables, but the
real strength of such techniques is their flexibility. The ability to factor
out nuisance variables (such as block effects) is another strength. As
previously mentioned, RDA with nominal variables is actually a special case of
MRPP.
I agree Nonmetric Multidimensional Scaling results in 'just a graph' - but its
graph is very limited in its interpretibility, as it does not maximize gradient
structure, it does not maximize the explained relationship between species and
the treatments (indeed, all species information is discarded), and it has no
clear ties to ecological theory. A big advantage of RDA and CCA is that crisp
(and flexible) hyopthesis tests are feasible, AND there are simultaneous
graphical representations of the observed differences.
--Mike Palmer
________________________________________
From: Ecological Society of America: grants, jobs, news
[[email protected]] On Behalf Of Gabriel Barros
[[email protected]]
Sent: Sunday, April 10, 2011 4:20 PM
To: [email protected]
Subject: Re: [ECOLOG-L] Community Analysis Hypothesis Tests
Hi Travis,
I believe that the use of ANOSIM or MANOVA (I do prefer PERMANOVA) depends
of your question.
The multivariate Analysis of variance will test the significance in the
difference among location (means). Thus, you could test if sites are
significant differents (sum or average of years) or test if there are
significant differences among years (sum or average of sites).
Using ANOSIM you could test the same things, but this analysis will test the
similarity (through a similarity matrix using a distance index) among
factors defined a priori (years or sites).
Using MDS will allow you to observe the possible configurations of the
groups tested in ANOSIM (different years or sites). Following the Tony
Underwood's opinion: "MDS is just a graph."
I think that sounds pretty enough. The use of Canonical Correspondence
Analysis depends if you have abiotic data and if you want to test the
relationship among biotic and abiotic variables.
I agree with James Novak about the importance of statistical tests.
Regards.
*Gabriel Barros Gonçalves de Souza*
Bachelor in Biological Sciences
Master in Ecology and Biomonitoring
CRBio: 59.707/05-D
Cel: (71) 8744-5246
CV Lattes: http://lattes.cnpq.br/4012374701934609
-------------------------------------------------------------------------------------------------------------------------------
Laboratório de Ecologia Bentônica (LEB)
Universidade Federal da Bahia (UFBA)
2011/4/2 Isabelle Wolf <[email protected]>
> Hi Travis, I concur with using PRIMER ver 6 and the two manuals that come
> with it for the proposed analysis. Further to this, I highly recommend
> using
> PERMANOVA+, an add-on to Primer ver 6, and its manual. It accommodates very
> complex designs and any number of factors and should be well suited for your
> purpose.
>
> Dr. Isabelle Wolf
> *********************************************************************
> Research and Analysis Officer
> Parks and Wildlife Group
> Department of Environment, Climate Change and Water (NSW)
> Level 2, Hurstville, NSW 2220
> [email protected]
> Ph: 02 9585 6672
> Fax: 02 9585 6601
>
>
> ----- Original Message ----- From: "Liz Pryde" <[email protected]>
>
> To: <[email protected]>
> Sent: Wednesday, March 30, 2011 8:01 PM
>
> Subject: Re: [ECOLOG-L] Community Analysis Hypothesis Tests
>
>
> Hi Travis,
>
> As it is a community comparison I think ANOSIM is still widely accepted to
> test for differences between treatments/time, particularly in the marine
> literature. Manuel is correct, it will not tell you the biological reason
> behind a difference and cannot really give a magnitude of difference per se
> (depending on the design of the study) but it does give a meaningful p
> value
> based on permutations of configurations of the presence-absence matrix. In
> this way its limitations are analogous with ANOVAs. As this is a
> comparative
> study I would think that a statistically significant p value would be
> meaningful.
>
> However, teasing out the question "why the difference" requires closer
> examination and comparisons between the community matrices and say,
> environmental variables. Examples of these types of further analyses can be
> found in the PRIMER manuals (ver 6) and in the extensive literature that
> can
> be found on the PRIMER-E website.
>
> Hope that helps,
> Liz
>
>
>
> On Tue, Mar 29, 2011 at 11:12 PM, Manuel Spínola <[email protected]
> >wrote:
>
> Hi Travis,
>>
>> I don't think that a p-value is going to tell you if there is a biological
>> meaningful difference between community.
>>
>> What will be the metrics that you are planning to use with ANOSIM ?
>>
>> Without seeing the data I can tell you that there is a difference between
>> communities, but the important question is how different they are, so you
>> can assess a practical or biological significance and a p-value is not
>> going
>> to tell you that.
>>
>> Best,
>>
>> Manuel
>>
>>
>> On 28/03/2011 07:14 a.m., T. Travis Brown wrote:
>>
>> Hello, I am trying to determine the best way to test for a difference in
>>> the overall mussel community found in a stream between 1980 and 2008. I
>>> have seven sites with presence/absence data. In addition to various
>>> descriptive statistics and graphs (nonmetric multidimensional scaling) I
>>> would like to use ANOSIM because it offers a P-value, and answers the
>>> question: "well, is there a difference or not?". I am not as up-to-date
>>> on
>>> this literature as I would like to be. Does anyone know if this is still
>>> an
>>> accepted test? Would some type of multi-response permutation procedure
>>> be
>>> better?
>>>
>>>
>>> T. Travis Brown
>>> [email protected]
>>>
>>>
>>>
>> --
>> *Manuel Spínola, Ph.D.*
>> Instituto Internacional en Conservación y Manejo de Vida Silvestre
>> Universidad Nacional
>> Apartado 1350-3000
>> Heredia
>> COSTA RICA
>> [email protected]
>> [email protected]
>> Teléfono: (506) 2277-3598
>> Fax: (506) 2237-7036
>> Personal website: Lobito de río <
>> https://sites.google.com/site/lobitoderio/>
>> Institutional website: ICOMVIS <http://www.icomvis.una.ac.cr/>
>>
>>
>
>
> --
> Liz Pryde
> PhD Candidate (off-campus)
> School of Earth and Environmental Sciences
> James Cook University
>
> Thornbury, Melbourne
> 0418551570
>
