Comments on: Stats and Methods Urban Legend 3: Myths About Meta-Analysis

By: Richard N. Landers

Richard N. Landers — Thu, 12 May 2011 19:26:58 +0000

While that would be fantastic, I think the biggest roadblock is that the current academic atmosphere does not support such a thing. Releasing data from a meta-analysis is akin to sharing data – and even if that data has been sliced and diced, there’s always a voice in the back of your head that says “maybe there’s another publication in that data.” Given the current reward system in academic, I don’t see this happening any time soon (even if it would be great move for science).

By: Jeromy Anglim

Jeromy Anglim — Tue, 03 May 2011 16:14:04 +0000

I like the point about reproducible research. It would be good if more meta-analyses included the complete database of studies and accompanying effect sizes, sample sizes, and study properties in a form ready for re-analysis. It would be even better if the meta-analyses were produced using some form of reproducible statistical document like Sweave.

Meta-analyses could be seen as an ongoing collective endeavour, where alterations to methodology could be easily examined by future researchers.

By: Jeromy Anglim

Jeromy Anglim — Tue, 03 May 2011 16:12:02 +0000

I like the point about reproducible research. It would be good if more meta-analyses included the complete database of studies and accompanying effect sizes, sample sizes, and study properties in a form ready for re-analysis. It would be even better if the meta-analytic analyses were embodied in some reproducible statistical document.