Qualitative and quantitative evidence are complementary, not substitutes
Aliases: mixed methods integration · evidence complementarity · methodological fit
What it is
Qualitative and quantitative evidence are not substitutes because evidence must match the inference. A recurring theme in a small interview sample does not estimate population prevalence, while a stable association in large-scale telemetry does not by itself explain the meaning of action. The approaches can address different layers of one phenomenon and can connect in mixed methods. Neither should carry a claim unsupported by its design merely because its evidence is larger, more vivid, or easier to present.
Why it happens
The approaches compress information differently. Quantitative analysis reduces cases to comparable variables to gain scale, precision, and modeling; qualitative analysis preserves sequence, language, and context to gain meaning and mechanism clues. More of the same data cannot restore information removed by that compression: additional clicks do not contain uninstrumented motives, and additional interviews do not create a probability-sampling denominator. Complementarity arises when one output changes the design or interpretation of another, not merely when two findings point in the same direction.
Studying it
A mixed-methods design specifies its connection: qualitative material can generate constructs and items before quantitative estimation; telemetry can identify anomalous groups for process-focused interviews; parallel strands can be integrated in a joint display that examines convergence, divergence, and silence. Each strand retains its unit of analysis and quality criteria. Integration maps claims to evidence and asks whether disagreement reflects non-equivalent measures, different samples, different time scales, or a theory requiring revision rather than averaging conflict away.
Where it stops holding
Non-substitutability is not the rule that qualitative work only explains and quantitative work only counts. Qualitative comparison can test theoretical boundaries, and quantitative experiments can investigate mechanisms when their added designs and assumptions support those aims. One method may be sufficient for a narrow, low-risk question; mixing methods for symmetry adds no value. Apparent convergence across different populations or settings is not mutual confirmation until construct equivalence is established.
Related
- Same group: Q1.03.1 Qualitative inquiry examines meaning and process · Q1.03.2 Quantitative inquiry estimates magnitude and uncertainty
- Adjacent: Q1.10 Three-dimensional classification of research methods · Q4.10 Generalizability of research findings
- Search terms:
mixed methods integration·joint display·methodological fit