Generalised Confidence Intervals for Meta-Regression
Thomas Friedrich, Guido Knapp
Faklutät Statistik, Technische Universität Dortmund, Dortmund, NRW, Germany; Faklutät Statistik, Technische Universität Dortmund, Dortmund, NRW, Germany

The explanation of heterogeneity that occurs when combining results from different studies is an important issue in meta analysis.

Besides including a heterogeneity parameter in the analysis, it is also important to understand the possible causes of heterogeneity. One possibility is to incorporate study-specific covariates in the analysis that account for such between-trial variability. This yields the random effects meta regression model.

We examine commonly used tests on the regression coefficients and propose a new method for constructing confidence intervals for the regression coefficients based on principles from generalised inference. The proposed method will be compared by simulation studies with respect to coverage probability and average length.

Keywords: meta analysis; confidence intervals; regression; generalised inference

Biography: After studying mathematics at the Freie Universität Berlin, Thomas Friedrich moved to the USA to study statistics at the University of Kentucky in Lexington. He then moved back to Germany to join the department of statistics at Dortmund University. His research interest centre around methods for meta-analysis and applications of generalised methods in this context.