Valid methods for meta-analyses with few studies or small numbers of cases - Part II

PIs:

Markus Pauly (TU Dortmund University)

Tim Friede (University Medical Center Göttingen)

Project Descritption

The project continues our successful collaborations from DFG project 413270747 and a VW project on Bayesian and Nonparametric Statistics and other research activities. Our joined aim is to develop valid inference methods (confidence and credible intervals, hypothesis tests) for meta-analyses that lead to reliable results even under difficult conditions such as

(i) heterogeneity between studies

(ii) unbalanced designs

(iii) small or moderate numbers of cases and

(iv) few studies.

To this end, we use and combine meta-analytical methods with approaches from Bayesian and non-parametric statistics as well as machine and statistical learning. Moreover, we collaborate with several renowned experts. In particular, we have the following

Mercator Fellows:

• Sofia Dias (University of Bristol)

• Christopher Schmid (Brown University)

• Wolfgang Viechtbauer (Maastricht University)