Program Evaluation — Systematic Assessment of Program Merit and Worth

Academic research aims to generate generalizable knowledge and tests theory; it is judged primarily by methodological rigor and contribution to a discipline. Program evaluation aims to inform a specific decision about a specific program; it is judged by utility, feasibility, and the accuracy of its evaluative conclusions. Evaluators must balance rigor with practicality and must explicitly render a judgment — not just report findings — whereas researchers typically do not.

A process (implementation) evaluation asks whether the program is being delivered as designed and is reaching the intended population — it is appropriate at any stage but especially valuable in the first one to two years of a program. An outcome evaluation asks whether the program is producing the intended changes — it requires that the program be sufficiently mature and consistently delivered. In most cases, process evaluation should precede or accompany outcome evaluation; a clean null outcome result is uninterpretable without knowing whether the program was actually implemented.

Not necessarily, but attribution of effects to the program requires some form of comparison. Randomized control trials provide the strongest causal evidence but are not always feasible. Quasi-experimental designs — matched comparison groups, regression discontinuity, interrupted time series — can provide reasonable causal estimates. For formative or process evaluations, or when causal claims are not the primary purpose, simpler pre-post or descriptive designs are often sufficient and appropriate.

External evaluators bring independence and credibility, which is important for accountability-focused evaluations. Internal evaluators have greater program knowledge and are better positioned for ongoing formative evaluation and learning. Many programs use a combination: an internal evaluation team for continuous monitoring and improvement, with periodic external evaluations for summative accountability judgments. The choice should be driven by the evaluation's primary purpose and the need for independence versus deep contextual knowledge.

A logic model (also called a theory of change) is a diagram or narrative that maps the program's intended causal chain from inputs and activities through outputs to short-, medium-, and long-term outcomes. It is not legally required but is considered best practice by most funders and evaluation standards. Without an explicit logic model, evaluators cannot specify what outcomes to measure, at what time points, or how to interpret findings — making the evaluation much harder to design and more likely to produce uninterpretable results.