[R-meta] Publication Bias
nzelle@k@yem @ending from ndm@ox@@c@uk
Wed Dec 19 15:51:51 CET 2018
Thank you very much.
From: Viechtbauer, Wolfgang (SP) <wolfgang.viechtbauer using maastrichtuniversity.nl>
Sent: 19 December 2018 14:30
To: Nzelle Kayem <nzelle.kayem using ndm.ox.ac.uk>; r-sig-meta-analysis using r-project.org
Subject: RE: [R-meta] Publication Bias
For this outcome, the 'Peters test' would in essence be:
dat <- escalc(measure="PFT", xi=xi, ni=ni, data=dat.debruin2009) dat
regtest(dat$yi, dat$vi, model="lm", predictor="ninv")
So, using the inverse sample size as predictor and using 'weighted regression with multiplicative dispersion' as the model.
From: R-sig-meta-analysis [mailto:r-sig-meta-analysis-bounces using r-project.org] On Behalf Of Nzelle Kayem
Sent: Wednesday, 19 December, 2018 13:32
To: r-sig-meta-analysis using r-project.org
Subject: [R-meta] Publication Bias
I found a blog on publication bias and unfortunately my questions were not answered so I thought I will write to you.
I am conducting a review on Ebola in pregnancy and I have completed the proportional meta-analysis.
I got into trouble though when I tried to do the publication bias tests. I used the double arcsine transformation and I have plotted the funnels using the sample size but I don't know the code to use for the peters test. I have used as a guide so far Wang et al 2018 and Viechtbauer W 2010.
Below is an example of the plot
More information about the R-sig-meta-analysis