For: educators, programme teams, schools and funders · Focus: distinguishing activity, output and outcome evidence · Evidence note: this Learning Lab synthesises external evaluation guidance and explains how Anhaya applies that distinction to its own reporting. It is not an evaluation of any individual Anhaya programme.
The problem starts with a very ordinary sentence
“We conducted a workshop for 200 students.”
That may be an important fact. It tells us something happened, at a certain scale. But it does not tell us what participants understood, whether they acquired a skill, whether they later used it, or whether the programme caused a longer-term change.
The distinction matters because educational organisations can accidentally move from what we did to what changed without collecting the evidence needed to support that second claim.
Five different questions
A useful way to examine a workshop is to separate five questions rather than collapse them into one.
1. Delivery — What did we do?
For example: one workshop was delivered; 200 students attended; four activities were completed.
2. Output — What did the activity produce?
For example: students completed a budgeting exercise, wrote an argument, produced a reflection, or participated in a structured discussion.
3. Immediate learning — What appears to have been learned?
This requires evidence suited to the learning claim: perhaps a short assessment, a scored task, a rubric, a before-and-after comparison, or another appropriate measure.
4. Application — Was the learning used later?
A participant demonstrating a skill after the session is a different claim from understanding it during the workshop. Follow-up evidence may be needed.
5. Longer-term outcome — What changed over time?
Claims about sustained behaviour, educational performance or broader impact require stronger designs and evidence. They should not be inferred from attendance, completion or satisfaction alone.
Why satisfaction is useful — and limited
Participant feedback can tell us whether a session felt relevant, clear, engaging or useful. That information matters for improving delivery. But a positive reaction is not automatically evidence that knowledge increased or behaviour changed.
OECD guidance on capacity-building evaluation discusses the widely used Kirkpatrick framework, which separates participant reaction, learning, behaviour and results. The OECD also cautions that these levels should not be treated as a simple hierarchy or assumed to correlate automatically, and notes limitations in relying heavily on questionnaires and tests.
Outputs are not inferior evidence
The answer is not to stop reporting attendance, sessions, resources or completed work. These are legitimate measures when labelled accurately.
A school needs to know whether a programme was delivered. A programme team needs to know whether participants completed the intended activities. A funder may need to know reach. The mistake is not measuring outputs. The mistake is renaming an output as an outcome.
The evidence should match the claim
UNICEF describes evaluation as a way to understand what works, for whom, under what circumstances and why. The OECD similarly emphasises assessment approaches suited to the results being examined, while the World Bank distinguishes inputs, processes, intermediate indicators and outcomes in results frameworks.
That leads to a simple discipline: decide what you want to be able to say, then ask what evidence would justify saying it.
If the claim is “200 students attended,” a reliable attendance record may be enough. If the claim is “students improved their ability to identify a weak argument,” the evidence requirement is different. If the claim is “the workshop changed students' behaviour six months later,” it is different again.
What Anhaya takes from this
Anhaya's public impact methodology deliberately separates reach and activity, outputs, and outcomes. The Foundation's verified headline figures describe the scale of work carried out; they are not presented as proof of learning or behavioural change.
For future evidence-led publishing, the practical rule is:
Report the strongest conclusion the evidence supports — not the strongest conclusion the programme would like to make.
That means a workshop can be worth documenting even when there is no outcome study. We can say what was delivered. We can show what participants produced. We can report bounded observations or attributed participant feedback. And where stronger learning evidence exists, we can report that separately with the method and limitations visible.
A five-minute evidence check
Before publishing a workshop result, ask:
What exactly is the claim? Is it about reach, completion, reaction, learning, application or longer-term change?
What evidence supports it? Attendance record, artefact, observation, participant report, assessment, rubric, follow-up or evaluation?
Does the evidence measure the thing named in the claim?
What can we not conclude?
Would the sentence still be accurate if the programme name were removed? This helps expose language that is promotional rather than evidential.
What this Learning Lab is not claiming
This article does not claim that workshops are ineffective, that every educational activity requires an impact evaluation, or that one evaluation model fits every programme. It argues for proportional evidence: use a method appropriate to the question, and keep the language within the boundaries of what that method can establish.
Sources & further reading
UNICEF Evaluation — What we do: Evaluations
OECD — Measuring the results of capacity building and locally led initiatives
World Bank — Results-Based Financing and Results in Education for All Children
Anhaya — How we report impact
Continue with Learning Lab
Next: How Do We Know When Learning Happened? → Explore Learning Lab