Iago, The PNG file is very hard to read. Could you upload a PDF file? Not sure what you mean by "second option feasible". Could you please elaborate. Best, Guido
[R-meta] A dataset with 0 heterogeneity
3 messages · Iago Giné Vázquez, Guido Schwarzer
Guido, I share the pdf file here: https://ann.nl.tab.digital/s/aNPBFedC6s8eXMR Regarding the "second option feasible", the second option is the second point you mentioned in your first mail: - all differences between individual study results can be explained by chance, i.e., random deviations due to (limited) sample sizes in individual studies. Could it explain the 0 heterogeneity in my example? Best, Iago
De: Dr. Guido Schwarzer <guido.schwarzer at uniklinik-freiburg.de>
Enviat el: diumenge, 2 de juliol de 2023 14:35 Per a: Iago Gin? V?zquez <iago.gine at sjd.es>; R Special Interest Group for Meta-Analysis <r-sig-meta-analysis at r-project.org> Tema: Re: [R-meta] A dataset with 0 heterogeneity Iago, The PNG file is very hard to read. Could you upload a PDF file? Not sure what you mean by "second option feasible". Could you please elaborate. Best, Guido
11 days later
Overall, 21 of 22 confidence intervals include the meta-analysis estimate. This is exactly what we would expect under homogeneity for 95% confidence intervals, i.e., 19 of 20 CIs include the meta-analysis estimate. Accordingly, the estimates of tau2 and I2 are 0 which means that all observed differences between point estimates can be explained by chance / random deviations. Best, Guido