Program evaluation

Program evaluation is the systematic assessment of a government, nonprofit, or grant-funded program to determine whether it is working, for whom, and at what cost. It sits between simple activity reporting, how many people were served, and full causal analysis of impact, and is a core discipline in public-sector analytics because most programs are funded on the assumption that they change outcomes, not just deliver services.

Evaluators typically start from a logic model that lays out the chain from inputs and activities to outputs and outcomes, then choose a design suited to the question. A process evaluation checks whether the program was implemented as intended; an impact evaluation estimates the program's causal effect, using a randomized controlled trial where feasible or a quasi-experimental method such as difference-in-differences when random assignment is not possible. Many evaluations rely on administrative data already collected by agencies rather than new surveys.

Program evaluation matters because funders, legislatures, and agencies use it to decide whether to continue, expand, or cut a program. The most common pitfall is measuring outputs, such as participants enrolled, instead of outcomes, such as participants who actually benefited, which can make an ineffective program look successful. A related but distinct exercise is cost-benefit analysis, which goes further and monetizes the evaluated outcomes against the program's cost.