Abstract
This article argues that the evidence-based policy movement needs to supplement its current emphasis on program evaluations with an approach that uses data at a much higher frequency to improve the administration and impact of government-funded social service programs.
Doing so offers the best chance of making significant progress in ameliorating challenging social problems. Liebman describes how an idealized government social service agency could use data and data analysis to improve its results, review the barriers that prevent agencies from operating this way, and outline how targeted resources and technical assistance can help overcome these barriers.
The article also includes strategies for moving beyond the effective administration of siloed service programs to the improvement of population-wide outcomes, especially among individuals and families who need multiple services.