How much do solutions to social problems differ in their effectiveness? A collection of all the studies we could find.

Abstract

The article analyzes previous research that has studied the differences in effectiveness of social interventions to understand how much interventions differ in their effectiveness. Different studies claim that the best interventions within a given problem area may achieve 10,000 times more good than the worst interventions or 100 times more than typical interventions. The article’s analysis of multiple datasets suggests that these claims generally hold and that interventions’ effectiveness tends to follow a heavy-tailed distribution, with the most effective interventions many times more impactful than the average. However, the analysis concludes that empirical estimates from previous research likely overstate the true differences, and accounting for various factors like measurement error and the potential for non-primary outcomes to be important suggests that the upper bound of the potential difference in effectiveness is likely only 10 times or less compared to the typical or average intervention within a particular problem area. – AI-generated abstract.