---
name: program-evaluation
description: Use when asked to evaluate, after implementation, whether a curriculum or educational program actually achieved its intended outcomes, as distinct from [[curriculum-design]], which designs the program beforehand, and [[training-needs-analysis]], which identifies a gap before any program exists.
---

# Program Evaluation

Program evaluation assesses, after a curriculum or educational program
has been running, whether it actually achieved the outcomes it was
designed for — using what data, compared against what original goals.
It comes after the fact: [[training-needs-analysis]] establishes the
gap a program should exist to close, [[curriculum-design]] builds the
program to close it, and program evaluation is the step that checks,
once the program has actually run, whether it did.

## Key components

- **Original goals** — the specific outcomes the program was designed
  to produce, retrieved from the original design rather than
  reconstructed after the fact from whatever data happens to be
  available.
- **Outcome data** — evidence of what actually happened: assessment
  results, completion rates, follow-on performance, gathered in a form
  that can be compared against the original goals rather than just
  described in general terms.
- **Comparison or baseline** — some basis for judging whether the
  outcome data represents success, whether that's a pre-program
  baseline, a comparison group, or a stated target set in advance.
- **Implementation fidelity check** — whether the program was actually
  delivered as designed, since a program that wasn't implemented as
  planned can't fairly be judged against outcomes it was never really
  given the chance to produce.
- **Findings and recommendation** — a clear judgment (continue, revise,
  discontinue) with the reasoning that led to it, not just a data
  summary left for someone else to interpret.

## Fidelity before outcome

A program evaluation that finds disappointing outcomes and stops there
is missing the question that actually determines what to do next: was
the program run as designed. A curriculum that looked strong on paper
can produce weak outcomes for two very different reasons — either the
design itself doesn't work, or the design was never actually delivered
as intended (teachers skipped components, pacing collapsed under real
constraints, the assessment used to measure outcomes wasn't the one
the design assumed) — and these call for opposite responses. Weak
outcomes from a faithfully implemented program are a signal to revise
or discontinue the design; weak outcomes from a program that drifted
from its own design during implementation are a signal to fix
implementation before concluding anything about the design at all.

Checking fidelity means gathering some evidence of what was actually
delivered — classroom observation, teacher self-report against the
program's own components, materials actually used — not just assuming
implementation matched the plan because the plan existed. Skipping
this step is how a genuinely sound program gets discontinued for a
delivery problem, or a genuinely broken one gets kept running because
nobody checked whether it was ever properly tried.

## Common pitfalls

- **No original goals to compare against** — evaluating a program
  against goals reconstructed after the fact, rather than what it was
  actually designed to achieve, makes almost any outcome look
  defensible.
- **Outcome data with no baseline** — reporting results with nothing to
  compare them to, so there's no way to tell whether the numbers
  represent success, failure, or business as usual.
- **Skipping the fidelity check** — judging a design by outcomes without
  first confirming the program was actually delivered as designed.
- **Evaluation timed too early** — measuring outcomes before the
  program has had enough time to produce them, especially for programs
  whose effects compound over multiple years.
- **Findings with no recommendation** — a data summary that describes
  what happened without stating what should happen next, leaving the
  actual decision to someone with less context than the evaluator had.

## Learn more

- [[curriculum-design]] for designing the program this evaluation
  assesses after the fact.
- [[curriculum-mapping]] for auditing coverage and alignment across
  courses, a different kind of check than whether a program met its
  outcomes.
- [[training-needs-analysis]] for confirming, before any program
  exists, what gap it needs to close — the goals a later evaluation
  will measure against.