How big an impact does a service need to have before it should be declared a success story? Peter Tymms reported on a study of a school-based peer tutoring program which identified an effect size of 0.2. He expressed disappointment with this: according to Cohen's accepted rule of thumb, it was small.

However, Robert Slavin, from the University of York's Institute for Effective Education, argued that this should be considered a large difference for a social or educational intervention and that it represented a "policy-ready finding".

Larger differences would only be achieved, he suggested, in tightly controlled settings. And there was value even in small effects if they were cumulative. Later, Slavin presented evidence showing that there was a negative