Methodological Insights for Impact Evaluations

This brief is part of a series uncovering lessons learned from four impact evaluations of promising reading interventions funded by USAID as part of the Latin America and the Caribbean Reads (LAC Reads) project. The evaluations were conducted by Mathematica.

Randomized controlled trials of education interventions present evaluators with many choices in the form of both challenges and opportunities. Impact evaluations of programs designed to improve early grade reading in Guatemala, Honduras, Nicaragua, and Peru generated methodological insights that could help researchers, implementers, and donors make decisions that improve evidence. These include decisions about baseline and interim data collection, the unit of randomization, and the contrasts that provide meaningful learning.