Integration should begin at design stage
Quantitative and qualitative tools should not be designed separately and combined only at the end. The evaluation matrix should show how each method contributes to each evaluation question. This creates a clear logic for triangulation.
Use qualitative evidence to explain patterns
Survey results can show whether an outcome changed or how responses differ by group. Qualitative data can help explain why those differences exist, how beneficiaries experienced the intervention and what contextual factors shaped results.
Do not let quotes stand alone
Quotes are valuable when they illustrate a finding, but they should not replace analysis. A strong report introduces the finding, uses a short quote as evidence and then interprets what the quote contributes to the argument.
Look for convergence and divergence
Triangulation is not only about confirming the same result across sources. Differences between survey data, interviews and documents can be analytically important. They may reveal implementation gaps, group differences, measurement problems or contested interpretations.
Present findings around questions, not methods
A common reporting mistake is to put all survey results in one section and all qualitative findings in another. A more useful structure presents integrated findings under each evaluation question or theme.
Sequence methods around the evaluation questions
Mixed methods can be implemented concurrently or sequentially. A survey may identify an unexpected pattern that is explored through later interviews, or early qualitative work may help refine survey questions and response categories. The sequence should be chosen because it helps answer the evaluation questions, not simply because both methods are listed in the TOR.
The inception phase should describe where integration will occur: sampling, tool design, data collection, analysis, interpretation or all of these stages.
Use joint displays and comparison tables where useful
Analysts can make integration more transparent by placing quantitative and qualitative findings side by side. A joint display may show an indicator result, relevant qualitative explanation, documentary evidence and the resulting interpretation for each evaluation question.
This approach makes contradictions visible and reduces the risk that one evidence stream dominates the report simply because it is easier to summarise.
Preserve subgroup differences
An overall percentage can hide important differences by gender, location, livelihood, age or programme exposure. Qualitative evidence can help interpret why those differences appear, but the analysis should avoid assuming that one participant quote explains an entire subgroup.
Where sample sizes allow, quantitative disaggregation and purposive qualitative comparison can work together to produce more credible explanations of variation.
Write one integrated conclusion
The final judgement should reflect the combined weight of evidence rather than separate quantitative and qualitative conclusions. Analysts should state where sources converge, where they differ and which interpretation is best supported.
This integrated reasoning is especially important for OECD-DAC criteria, contribution analysis and cross-cutting issues, where no single data source is usually sufficient on its own.