ABSTRACT
Some reviews have claimed that dierences in eects and eect sizes are at worst only minimal.1,2 Others have concluded that these dierences can be as large as, or even larger than, the size of the eects that were to be detected.3 In a ‘review of reviews’, this is therefore hardly surprising: ‘Results of non-randomized studies sometimes, but not always, dier from results of randomized studies of the same intervention’.4
