How to Build an Evidence Trail in Programme Evaluation

Evaluation reports often compress months of research into a handful of findings and recommendations. A strong evidence trail preserves the route between those conclusions and the material that produced them.

Begin with the evaluation questions. Each major question should be linked to relevant evidence sources rather than treated as a heading added only during reporting.

Sources may include interviews, monitoring data, documents, observations, workshops and administrative records. The evaluation record should make clear which sources inform which questions and where evidence is thin.

Qualitative coding should remain connected to the source material. If a finding concerns implementation barriers, reviewers should be able to identify the relevant passages, cases and analytical categories rather than relying solely on a narrative summary.

Triangulation should not mean simply collecting several types of evidence. The evaluator needs to show whether sources converge, complement or contradict one another.

For example, programme monitoring may indicate high attendance while interviews suggest participation was irregular for a particular subgroup. The contradiction is not a problem to erase; it may be a finding about how the indicator behaves.

Recommendations need a trail too. A recommendation should be linked to the problem it addresses and to the evidence supporting that diagnosis. This reduces the risk of generic recommendations appearing at the end of a report without a clear analytical basis.

A simple evaluation evidence matrix can connect:

• evaluation question;
• source;
• key evidence;
• analytical interpretation;
• contradictory evidence;
• finding;
• recommendation.

The matrix does not replace narrative analysis. It helps reviewers examine whether the narrative is grounded.

Evidence trails are also valuable after the evaluation. Programme staff, funders or later evaluators may need to understand why a conclusion was reached long after the original consultants have moved on.

Where data are sensitive, traceability can remain internal rather than public. The objective is not to publish every transcript. It is to preserve a defensible analytical record within appropriate access controls.