Seb Schmoller points to a really interesting site that summarizes a couple dozen popular educational interventions and assesses their effectiveness in promoting positive educational outcomes. The presentation is a bit unintuitive - the interventions are scored in terms of cost (that's clear enough), stars (representing the number of studies supporting the assessment), and effectiveness (confusingly represented in terms of months - so, for example, the 'average impact' of peer tutoring is '+6 additional months', whatever that means). While a site like this is useful, it needs to be read with caution - your "improved educational outcome" might be my "severe psychological trauma". That said, the sort of things one would think are useful - like feedback, meta-cognition and self-regulation, and yes, peer tutoring - are at the top of the list, while those things we would think are useless - like school uniforms, performance pay, physical environment and ability grouping - are at the bottom. It would have been nice to see what the Toolkit says about standardized tests, but it unfortunately ducks that issue completely.