Discrepancy between meta-analysis results of new and old version of R

I have a question about Meta package and its updates.

I ran meta-analysis previously for my data using the previous version of Meta and the results are slightly different with the new update of Meta package. The overall direction of results, p-value and heterogeneity are the same but the weighing of each study and overall results are changed.

Has anyone experiences such discrepancy of results before? Is there any reason why different versions of a same package use different algorithms for their analysis? this could change the overall meta-analysis results for those 'on edge' findings!

Cheers,
Sam

Ok, I think my smart colleague and I have found the issue, looks like the new update of Meta package used Restricted Maximum Likelihood (REML) instead of DerSemonian-Laird (DL) for between study heterogenity estimator as a diffult. So anyone still interetsed to use DL should be able to add method.tau="DL" to thier code and still get the same results as previously.

1 Like

This topic was automatically closed 42 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.