Open dialogue peer review: a response to Tymms, Merrell & Coe

We welcome Peter Tymms, Christine Merrell and Robert Coe's paper as a timely contribution to an important issue. For precisely the reasons that they state, this is an area of current concern. We are writing to suggest that for complex interventions involving educational programmes an even more complicated sequence of investigations could be useful, taking as the model the medical approach as detailed in MRC (2000). We agree that the RCT is an essential tool to investigate the efficacy of programmes. There is no other way to know if, on the whole, a programme works across a variety of contexts and if some programmes should 'work' better than others. Pragmatic randomisation as described in the Fife study outlined by Tymms et al. should be appropriate although it is a pity that it appears no children are continuing with their current exposure to peer learning, which would allow for the possibility that this is just as good as the new interventions. Blind assessment of outcomes is of course essential.