A US economist makes the case for putting statistical effect sizes and the rudiments of cost benefit analysis into the same equation, and proposes a new interim “law” of prevention: that an intervention should be adopted only if there is no other way to create the same scale of improvement more cheaply.