From: The normality assumption on between-study random effects was questionable in a considerable number of Cochrane meta-analyses
Binary outcome
(N = 6162)
Non-binary outcome
(N = 4014)
Originala
ORb
RRb
RDb
REML failing to convergec
259
273
272
217
101
\({\widehat{\tau }}^{2}\)=0d
1669
1678
1719
1219
480
Eligible meta-analyses
4234
4211
4171
4726
3433