<html>
<head>
<meta name="viewport" content="width=device-width">
<meta http-equiv="Content-Type" content="text/vnd.ui.insecure+html;charset=utf-8">
</head>
<body style="overflow-wrap:break-word; word-break: break-word;"><div class="mail_android_message" style="line-height: 1; padding: 0.5em">Dear Gerta,<br/><br/>thank you very much for your answer. My confusion is cleared.<br/><br/>Regards <br/><br/>Tobias<br/><br/>--<br/>Diese Nachricht wurde von meinem Android Mobiltelefon mit GMX Mail gesendet.</div><div class="mail_android_quote" style="line-height: 1; padding: 0.3em"><html><body>Am 18.08.21, 15:51 schrieb "Dr. Gerta Rücker" <ruecker@imbi.uni-freiburg.de>:</body></html><blockquote class="gmail_quote" style="margin: 0.8ex 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<p>Dear Tobias,</p>
<p>Here I comment only to this particular point:<br> </p>
<div class="moz-cite-prefix">
Am 18.08.2021 um 14:12 schrieb Tobias Saueressig:
<br>
</div>
<blockquote>
<div style="font-family: Verdana;font-size: 12.0px;">
[...]
<div>
</div>
<div>
What puzzels me is the difference in the hetereogeneity estimates (Tau², I²..etc.). If I perform the analysis with the non-transformed means I get different estimates then the analysis with log-transformation.
</div>
</div>
</blockquote> This is correct.
<br>
<blockquote>
<div style="font-family: Verdana;font-size: 12.0px;">
<div>
If I use the metamean function in meta (for comparison) it shows the same heterogeneity estimates wheter I log-transform or not.
</div>
</div>
</blockquote>
<p>This I don't understand. Heterogeneity measures should in general strongly depend on whether the data are log-transformed or not. See attached (fictional and quite extreme) example.</p>
<p>Best,</p>
<p>Gerta<br> </p>
<blockquote>
<div style="font-family: Verdana;font-size: 12.0px;">
<div>
</div>
<div>
Regards,
</div>
<div>
</div>
<div>
Tobias
</div>
<div>
</div>
<div>
</div>
<div>
<div>
<div style="margin:10px 5px 5px 10px; padding:
10px 0 10px 10px; border-left:2px solid #C3D9E5;
word-wrap: break-word; -webkit-nbsp-mode: space;
-webkit-line-break: after-white-space;">
<div style="margin:0 0 10px 0;">
<b>Gesendet:</b> Mittwoch, 18. August 2021 um 13:59 Uhr
<br> <b>Von:</b> "Viechtbauer, Wolfgang (SP)" <a class="moz-txt-link-rfc2396E" href="mailto:wolfgang.viechtbauer@maastrichtuniversity.nl"><wolfgang.viechtbauer@maastrichtuniversity.nl></a>
<br> <b>An:</b> "Tobias Saueressig" <a class="moz-txt-link-rfc2396E" href="mailto:t.saueressig@gmx.de"><t.saueressig@gmx.de></a>, "Meta list" <a class="moz-txt-link-rfc2396E" href="mailto:r-sig-meta-analysis@r-project.org"><r-sig-meta-analysis@r-project.org></a>
<br> <b>Betreff:</b> RE: [R-meta] Question on combining means with robumeta
</div>
<div>
Dear Tobias,
<br>
<br> I don't understand what you mean by "Since in some cases I had negative values". Negative values where?
<br>
<br> This aside, 'var.eff.size' is for the *variances* but SEM is the SE of the means. So it should be:
<br>
<br> res <- robu(formula = Mean ~ 1, data = dat,
<br> studynum = Author, var.eff.size =SEM^2,
<br> rho = .8, small = TRUE, modelweights="CORR")
<br> res
<br>
<br> although the results are basically the same.
<br>
<br> Second, LogMean_SEM doesn't seem to be computed correctly in your example and let's not call it SEM since it's a variance:
<br>
<br> dat$LogMean_Var <- with(dat, SD^2/(N*Mean^2))
<br>
<br> Now:
<br>
<br> res1 <- robu(formula = LogMean ~ 1, data = dat,
<br> studynum = Author, var.eff.size = LogMean_Var,
<br> rho = .8, small = TRUE, modelweights="CORR")
<br> res1
<br>
<br> #backtransformed results from log scale
<br> round(exp(c(res1$reg_table$b.r[[1]],res1$reg_table$CI.L[[1]],res1$reg_table$CI.U[[1]])),2)
<br>
<br> gives very similar results as 'res' earlier.
<br>
<br> Best,
<br> Wolfgang
<br>
<br> >-----Original Message-----
<br> >From: R-sig-meta-analysis [<a class="moz-txt-link-freetext" href="mailto:r-sig-meta-analysis-bounces@r-project.org">mailto:r-sig-meta-analysis-bounces@r-project.org</a>] On
<br> >Behalf Of Tobias Saueressig
<br> >Sent: Tuesday, 17 August, 2021 17:48
<br> >To: Meta list
<br> >Subject: [R-meta] Question on combining means with robumeta
<br> >
<br> >Dear All,
<br> >
<br> >I have a question concerning the combination/pooling of means. Since most of my
<br> >studies are correlated I deceided to use robust variance estimation method to do
<br> >this. So metamean function in meta was out because of this. So I used robumeta.
<br> >Since in some cases I had negative values I deceided to use log transformed means
<br> >for pooling. My problem is that the estimates look reasonable now with robumeta
<br> >but the heterogeneity parameters(e.g. Tau, I²) are all estimated with zero. This
<br> >is strange. Does anybody know why that is?
<br> >
<br> >Regards,
<br> >Tobias Saueressig
<br> >
<br> >Reproducible example:
<br> >
<br> >#Example
<br> > Author <-c("Carboch et al. (2019)","Stare et al. (2015)","Mackie (2013)","Hornery
<br> >et al. (2007)",
<br> >"Takahashi et al. ( 2006)","Morante et al. (2005)","O'Donoghue, P. and Ingram, B.
<br> >(2001)",
<br> >"O'Donoghue, P. and Ingram, B. (2001)")
<br> >
<br> > Mean<-c(5.93,4.40,6.42,6.70,5.10,6.40,5.70,4.90)
<br> > SD<-c(0.67,1.73,1.27,2.20,5.19,1.40,1.00,1.40)
<br> > N<-c(7,5,9,17,16,12,24,35)
<br> > SEM<-c(0.25,0.77,0.42,0.53,1.30,0.40,0.20,0.24)
<br> > LogMean<-c(1.78,1.48,1.86,1.90,1.63,1.86,1.74,1.59)
<br> > LogMean_SEM<-c(1.50,3.40,2.72,3.57,6.62,2.59,1.16,1.16)
<br> >
<br> > dat<-data.frame(Author,Mean,SD,N,SEM,LogMean,LogMean_SEM)
<br> >
<br> > #Formulas used
<br> > #SEM<-SD/sqrt(N)
<br> > #LogMean<-log(Mean)
<br> > #LogMean_SEM<-SD^2/(N*Mean^2) => taken from metamean function in meta
<br> >
<br> > #normalscale
<br> > res <- robu(formula = Mean ~ 1, data = dat,
<br> > studynum = Author, var.eff.size =SEM,
<br> > rho = .8, small = TRUE, modelweights="CORR")
<br> > res
<br> >
<br> > #logscale
<br> > res1 <- robu(formula = LogMean ~ 1, data = dat,
<br> > studynum = Author, var.eff.size = LogMean_SEM,
<br> > rho = .8, small = TRUE, modelweights="CORR")
<br> > res1
<br> >
<br> > #backtransformed results from log scale
<br> > round(exp(c(res1$reg_table$b.r[[1]],res1$reg_table$CI.L[[1]],res1$reg_table$CI.U[
<br> >[1]])),2)
</div>
</div>
</div>
</div>
</div>
<br>
<pre class="moz-quote-pre">_______________________________________________
R-sig-meta-analysis mailing list
<a class="moz-txt-link-abbreviated" href="mailto:R-sig-meta-analysis@r-project.org">R-sig-meta-analysis@r-project.org</a>
<a class="moz-txt-link-freetext" href="https://stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis">https://stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis</a>
</pre>
</blockquote>
</blockquote></div></body>
</html>