What Is an Evaluability Assessment?
An evaluability assessment asks which evaluation questions can be answered credibly, in time to inform a decision, and with feasible resources. It considers the program's logic and stage, available evidence, possible methods, access, ethics, skills and budget. Its result may be to proceed, narrow the questions, prepare missing evidence, or defer a particular claim.
Readiness depends on the question. An early program can be ready for an assessment of implementation even when outcomes have not had time to develop. A missing baseline can restrict measurement of change or a proposed impact design without ruling out every useful evaluation.
Start With the Intended Decision
Name the decision, its users and its deadline. Then specify what the evaluation needs to establish. For example: understand why referrals are delayed, describe who is reached, or assess whether the program caused a change. These require different evidence and methods. Use the evaluation questions to set scope before choosing a study.
The Magenta Book's evaluation-scoping guidance treats feasibility and design as iterative: available methods and resources may change which questions can be answered. Record that change explicitly in the agreed scope.
Review Each Question and Its Evidence
Use the following checks to prepare a question-specific assessment. Record the evidence inspected, unresolved points, responsible person and next decision.
| Check | What to Record |
|---|---|
| Purpose and use | The decision, intended users, deadline and question to be answered. |
| Program logic and stage | The relevant activities, expected changes, assumptions and competing interpretations. Is the expected change observable yet? |
| Existing evidence | Sources, definitions, coverage, period, quality, missingness and access conditions. Check the records themselves where permitted. |
| Design and inference | What the proposed method can establish, what it cannot, and how it will examine alternative explanations where causal claims are intended. |
| Feasible collection | Any additional evidence needed, respondent burden, consent, accessibility and safe access. |
| Resources and timing | Costed tasks, skills, staff availability, review time and whether results can arrive before the decision. |
A baseline is relevant when the question and design require a defensible starting point. Check whether suitable historical records exist and whether their definitions, coverage and timing permit comparison. Do not relabel a later observation as a pre-program baseline. A comparison group also needs a credible justification; its availability alone does not establish a valid counterfactual.
A Worked Readiness Decision
This fictional example illustrates the reasoning, not a standard study design.
A community referral program has six months of dated referral records but no pre-program baseline. The manager needs to decide next month whether to change the referral handoff. Staff disagree about where delays arise.
- Feasible question: Where do the recorded referrals stall? Check record completeness and stage definitions, analyze the elapsed time between recorded steps, and explore selected cases with appropriate consent. Limit conclusions to the records and cases examined. Gaps in dates may require a narrower analysis or additional prospective tracking.
- Unsupported question on this evidence alone: How much did the program improve health outcomes compared with no program? Referral records do not establish the outcomes or what would have happened without the intervention. Do not infer impact from completed referrals.
- Decision: Commission the bounded process inquiry if access, safeguards, skills and time are adequate. Record the causal question as unresolved. Assess whether a separate credible impact approach is feasible before promising an impact estimate; simply collecting a baseline now cannot recover the lost starting point.
The output is a usable commissioning decision with limits, rather than a blanket judgment that the whole program is ready or unready.
Turn Gaps Into Decisions
For each applicable gap, choose an action and assign its owner and review date:
- Proceed: Evidence and design support the agreed question; document remaining limitations.
- Proceed With a Narrower Scope: Revise the question or population and state the resulting limits on inference.
- Prepare First: Secure necessary access, clarify definitions, establish safeguards or collect information required by the selected design.
- Defer or Decline That Question: The available design, resources or timing cannot support the intended answer.
A "No" in a readiness check is a reason to assess the consequence. It is not automatically a ban on all evaluation. However, an unresolved safety, permission or validity requirement can block the affected activity. Do not average such a blocker away with other positive checks.
Plan Early and Revisit When Conditions Change
At program design, identify likely decisions and preserve evidence that may be impossible to reconstruct later. Schedule any needed baseline before the relevant exposure or change. Set evaluation timing around the question and when results can emerge, rather than a universal month of implementation.
Before issuing the evaluation Terms of Reference, confirm scope, evidence access, resources and unresolved conditions. Revisit the assessment if the program, decision deadline or available data change. The assessment can be brief for a narrow question or more involved for a complex design; no fixed duration fits every case.
Last updated October 27, 2026First published March 3, 20261 source checked
M&E Studio also does this work for clients.
The team that writes and reviews this library takes on M&E system design, evaluation quality reviews, AI advisory and workflow automation.
readiness assessment, program readiness, evaluation readiness