[R-meta] Spatial-temporal meta-analysis

Rafael Rios biorafaelrm at gmail.com
Fri Nov 10 13:37:10 CET 2017


Dear Wolfgang,

Thank you for the reply. I am sorry to complicate something that should be
simple. I want to evaluate if the effect size vary throughout the time, for
example. Therefore, I thought to use a measure that estimate such temporal
variation in effect size, like CV. I did not consider to use temporal
categories as a moderator because I have different categories for each
species. For instance, I have three effect sizes for three breeding seasons
in a species, while I have twelve measures for another species. Even in
this situation, may I use temporal categories in the function rma.mv? If I
can do this, it will be amazing.

Best wishes,

Rafael.

Em 10 de nov de 2017 8:45 AM, "Viechtbauer Wolfgang (SP)" <
wolfgang.viechtbauer at maastrichtuniversity.nl> escreveu:

Dear Rafael,

It is still not really clear to me what want to do. You say that you want
to "use coefficient of variance (CV) and number of replicates in rma.mv
function of metafor package to built a random model controlling
phylogenetic signal". I have no idea what that means, so I cannot say if
this approach is appropriate or not.

When you say that you want to examine if effect sizes vary in space and
time, then why not include some measure of time and some measure of
location in the model as predictors/moderators?

Best,
Wolfgang

>-----Original Message-----
>From: Rafael Rios [mailto:biorafaelrm at gmail.com]
>Sent: Thursday, 09 November, 2017 14:51
>To: Viechtbauer Wolfgang (SP); r-sig-meta-analysis at r-project.org; Michael
>Dewey
>Subject: Spatial-temporal meta-analysis
>
>Dear Wolfgang and All,
>
>I am still working on an approach to evaluate variations in effect sizes
>in space and time. I have several replicates of effect sizes (Fisher's Z)
>in different populations and/or breeding seasons for each species. My
>hypothesis is that effect size varies in space and time. Then, I want to
>use a random model to evaluate if a estimative of variation in effect
>size is high when compared to a null model. I am wondering if I can use
>coefficient of variance (CV) and number of replicates in rma.mv function
>of metafor package to built a random model controlling phylogenetic
>signal. Do you agree with my approach or suggest to use another measure
>of variation in replicates of effect size or a different estimate of
>weight?
>
>Best wishes,
>
>Rafael R. Moura.
>scientia amabilis
>
>Doutorando da Pós-graduação em Ecologia e Conservação de Recursos
>Naturais
>Universidade Federal de Uberlândia, Uberlândia, MG, Brasil
>
>ORCID: http://orcid.org/0000-0002-7911-4734
>Currículo
>Lattes: http://buscatextual.cnpq.br/buscatextual/visualizacv.do?id=K42449
>08A8
>Research Gate: https://www.researchgate.net/profile/Rafael_Rios_Moura2

	[[alternative HTML version deleted]]



More information about the R-sig-meta-analysis mailing list