<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<p>Am 10.11.17 um 11:17 schrieb Viechtbauer Wolfgang (SP):<br>
</p>
<blockquote type="cite"
cite="mid:88bef0e02cc64ef5ad7d436fd64cebce@UM-MAIL3216.unimaas.nl">
<pre wrap="">[...]
Note that "AS-Thompson" refers to using the 'rma' model.</pre>
</blockquote>
<br>
Wolfgang, I fear this is not the case.<br>
<br>
The "AS-Thompson" test refers to using (i) the arcsine difference as
effect measure which is unimportant for the following discussion and
(ii) method 3a in Thompson & Sharp (1999) which implements an
additive between-study variance component. This method is
implemented in metabias() of R package <b>meta</b> (argument <i>method
= "mm"</i>).<br>
<br>
I had a look at results of regtest() from <b>metafor</b> and
metabias() from <b>meta</b> using two (small) examples which are
part of the examples on the help page of metaprop(). The results are
summarized in the attached text file and show that p-values from
regtest() with argument <i>model = "rma"</i> (default) and
metabias() with argument <i>method = "mm"</i> do not agree. On the
other side, results from regtest() with argument <i>model = "lm"</i>
and metabias() with argument <i>method = "linreg"</i> (default) are
identical. Actually, in the second example, we see a similar pattern
for regtest() as observed by Laura: non-significant results for <i>model
= "lm"</i> and (highly) significant results for <i>model = "rma"</i>.
Clearly, it is not possible to deduce any general patterns from two
examples.<br>
<br>
I only had a quick glance at the R code of regtest(), however, I
assume that argument <i>model = "rma"</i> uses a multiplicative
overdispersion factor (see equation (2) in Thompson & Sharp,
1999).<br>
<br>
<br>
Main reason to implement an additive variance component in
metabias() is the following statement by Thompson and Higgins
(2002):<br>
<br>
"There is little to motivate the use of a multiplicative variance
adjustment factor in meta-regression, since the within-study
variances are known, although this is what is achieved by the
conventional use of weighted regression programs in most statistical
software. An additive component for the residual variance is more
reasonable in both meta-regression [9] [...]".<br>
<br>
See also Harbord et al. (2006) - including Matthias Egger as
co-author:<br>
<br>
"The alternative ‘weighted’ version of the test also suggested by
Egger et al. [7], denoted by ‘EW’ in Reference [14], is seldom used
and lacks a theoretical justification [24]."<br>
<br>
Furthermore, the test by Thompson and Sharp (1999), method 3a, is
the only test considering between-study heterogeneity mentioned in
Sterne et al. (2011), albeit in the setting of a binary outcome with
two groups.<br>
<br>
Best wishes,<br>
Guido<br>
<br>
References:<br>
<br>
Harbord RM, Egger M, Sterne JA, 2006, A modified test for
small-study effects in meta-analyses of controlled trials with
binary endpoints, Statistics in Medicine, 25(20), pp. 3443-57<br>
<br>
Sterne JAC et al., 2011, Recommendations for examining and
interpreting funnel plot asymmetry in meta-analyses of randomised
controlled trials, BMJ (Clinical research ed.), 343, p. d4002<br>
<br>
Thompson SG, Higgins JP, 2002, How should meta-regression analyses
be undertaken and interpreted? Statistics in Medicine, 21(11), pp.
1559-73<br>
<br>
Thompson SG, Sharp SJ, 1999, Explaining heterogeneity in
meta-analysis: A comparison of methods. Statistics in Medicine,
18(20): pp. 2693-708<br>
</body>
</html>