Examining how meta-analytic methods perform

Abstract

I read with interest the paper by Bramley et al.1 that concluded (regarding error estimation) that the IVhet method often outperforms the Henmi-copas (HC) method with fewer than 10 studies. They also concluded that the IVhet method is never the best choice of interval because for small numbers of studies Knapp–Hartung (KH) method is better and for larger numbers HC is better.