This is a special post existing to allow users to post comments and questions about other meta-analytic models, software, and evidence base problems that are not dealt with elsewhere on this website.
As time passes, if some of the topics become more popular, we will move the comments to a new post and write about it there.
Subscribe to get the latest posts sent to your email.
Hi Robert and Gian Luca,
not sure if this is the right spot to post (and happy to be referred elsewhere!) but I think I noted a typo in the Bayes MA book, specifically in the BUGS code for multivariate analysis in Section 15.2.3: The first loop uses “j” as the loop iteration index but then refers to “i” throughout the covariance matrix code in the loop.
Thank you for writing, Johannes. I looked at this (p.254) and I think you are right. Sorry everyone, the index should be j as it indicates the study. We used j for that throught the book but obviously slipped up here.
In Chapter 3, just before section 3.2.1, we made a mistake. Sorry everyone! We show formulas for a manual bare-bones common effect meta-analysis, and encourage you to try it out yourself on the green tea data. Then we showed the results that should be obtained… but we gave the random effects output by mistake. The stats should be:
theta-hat = -0.703
SE(theta-hat) = 0.033
Many thanks to Jak Laight from Cancer Research UK Clinical Trials Unit for bringing that to our attention.
Leave a Reply