The meta-analysis of multi-centre trials can be based on either fixed or ra
ndom effect models. This argues for the general use of random effect models
, and illustrates the value of non-parametric likelihood (NPML) analysis of
such trials. The same general approach unifies administrative 'league anal
yses in epidemiological and other studies. Several examples of the NPML ana
lysis are given, a 70-centre trial. Copyright (C) 1999 John Wiley & Sons, L
td.