Message-ID: <6fd6936a54f042899614618981bc7d26@UM-MAIL3214.unimaas.nl>
Date: 2019-03-10T10:04:51Z
From: Wolfgang Viechtbauer
Subject: [R-meta] r-sig-meta-analysis: metafor with
In-Reply-To: <VI1PR0601MB255799E320F61F12D80D7D86DA710@VI1PR0601MB2557.eurprd06.prod.outlook.com>
Dear Guillem,
In essence, you have different operationalizations of the same concept. This happens all the time in meta-analyses. One can be strict and only combine studies using the same operationalization or, as Michael suggested, use the type of operationalization as a potential moderator.
I think chapter 2 ("Hypotheses And Problems In Research Synthesis" by Cooper) in the Handbook Of Research Synthesis And Meta-analysis (2nd Ed.) discusses this issue.
Best,
Wolfgang
-----Original Message-----
From: R-sig-meta-analysis [mailto:r-sig-meta-analysis-bounces at r-project.org] On Behalf Of Guillem Chust
Sent: Monday, 04 March, 2019 10:06
To: Michael Dewey; r-sig-meta-analysis at r-project.org
Subject: Re: [R-meta] r-sig-meta-analysis: metafor with
Thanks Michael,
Do you know a paper reviewing this issue or a paper conducting a meta-analysis in such conditions?
Sincerely,
Guillem
-----Mensaje original-----
De: Michael Dewey <lists at dewey.myzen.co.uk>
Enviado el: viernes, 1 de marzo de 2019 13:01
Para: Guillem Chust <gchust at azti.es>; r-sig-meta-analysis at r-project.org
Asunto: Re: [R-meta] r-sig-meta-analysis: metafor with
Dear Guillem
Opinions will doubtless differ over this one but my view, for what it is worth, is that if you consider there is an underlying latent variable of anthropogenic pressure and the proxies are all (imperfect) measures of that latent variable then I would proceed. I would include correlation_with as a moderator though whether it reaches some arbitrary level of statistical significance or not.
Michael
PS before you post again can you register so we do not have to manually authorise your post please.
On 01/03/2019 11:22, Guillem Chust wrote:
> Dear Dr Viechtbauer,
>
> I would like to apply metafor for conducting a meta-analysis in the following case, but I'm not sure wether this is correct or not.
>
> We have compiled studies on the correlation r between an index of water quality and anthropogenic pressures (see some examples below).
>
> The water quality index is always the same, but the pressure has been measured using different proxies: e.g ammonium, Cadmium, oxygen,...
>
> Can in this case apply a metanalysis? Or the metanalysis should be applied only with studies conducted with the same two variable pairs?
>
> Study N r correlation_with
> Mulik et al. (2017) 12 0.69 ammonium
> Carvalho et al. (2006) 24 0.78 As
> Santiba ez et al (2018) 27 0.86 Ba
> Borja et al. (2006) 120 0.45 Cd
> Borja et al. (2006) 120 0.47 Cr
> Carvalho et al. (2006) 24 0.83 Cr
> Spagnolo et al. (2014) 12 0.605 Cr
> Borja et al. (2006) 120 0.29 Cu
>
> Sincerely,
>
> Guillem Chust, PhD
> AZTI - Marine Research Division
> Txatxarramendi ugartea z/g; 48395 Sukarrieta (SPAIN)
> Tel: (+34) 667174396
> www.azti.es