In her Comment (Nov 8, p 1610) on our second phase of evaluation of Sure Start local programmes in England, Penny Kane makes several points with which we could not agree more. Like her, we would have much preferred to see a randomised controlled trial done, since this would have afforded much stronger causal inferences than the quasi-experimental investigation we undertook. We also agree that the fact that we drew on data collected by two different research teams raises questions about the confidence that can be placed in conclusions drawn.
Belsky, J., Leyland, A., Barnes, J. & Melhuish, E. (2009). Sure Start in England. The Lancet, 373 (9661), 381-381.