Simple enough: policy makers want to know what works and systematic reviews are supposed to tell them. The reality is more complicated, however, as Mark Newman from the EPPI-Centre at the Institute for Education, London, demonstrated.

One of the problems is that there is no agreed formula for deciding how many studies, of what quality, are needed to claim that a particular type of intervention is effective. Several sets of guidelines exist but they differ in terms of how they grade studies, the quality threshold they require, and their approach to interpreting the results. Worse still, some systematic reviews do not use guidelines of any kind.

Consequently, different reviews of the same subject, using the same studies, may reach contradictory conclusions: “it works and should be adopted”; “it is promising but should be