Start with the sentence you want to be able to say
There is a meaningful difference between “learners scored 8 out of 10 after the session” and “learners improved by three points.” The first is a post-programme performance statement. The second is a change statement. To make the second claim, you need relevant information about the starting point.
A baseline is simply evidence collected before the activity or intervention that is suitable for comparison with later evidence. It does not automatically make an evaluation rigorous. The before-and-after measures still need to examine the same construct in a sufficiently comparable way.
When a baseline matters
If your question is What could participants demonstrate at the end?, a well-designed post measure may be enough. If your question is Did participants improve?, you need evidence of where they began. If your question is Did our programme cause the improvement?, a baseline is necessary but usually not sufficient: other explanations for change also need to be considered.
A practical five-step check
- Name the intended change. Avoid measuring whatever is easiest merely because it produces a number.
- Choose a starting measure. It should reflect the same learning objective you plan to examine later.
- Keep the comparison interpretable. If the task, scoring rule or conditions change substantially, explain what can and cannot be compared.
- Record who completed each measure. A before-and-after comparison can become misleading when the groups are materially different.
- Match the language to the design. Report demonstrated performance, observed change or causal effect only when the evidence supports that level of claim.
What a baseline does not solve
A baseline does not tell you whether learning lasted, whether learners used the skill later, whether a self-report matches observed behaviour, or whether the programme rather than another factor produced the change. Those are different questions and require different evidence.
Use this rule
If the claim contains a word such as “improved”, “increased”, “reduced” or “changed”, ask: compared with what?
Sometimes the right answer is that no baseline exists. That does not make the work worthless. It means the report should describe what was actually observed instead of retrofitting a change claim.
Evidence boundary: This guide explains evaluation logic used by Anhaya's Learning Lab. It is not a claim that every educational activity requires a pre-test, nor that a pre/post design by itself establishes causation.