Research Synthesis

---
name: research-synthesis
description: Synthesize user research into evidence-backed patterns, needs, mental models, barriers, opportunities, and product implications while preserving contradictory evidence and the limits of the study.
---
# Research Synthesis

Use after qualitative or mixed-method user research needs to be converted from raw observations into decision-ready findings.

## Procedure
1. Organize raw notes, transcripts, observations, task results, and artifacts by participant and research question while preserving provenance.
2. Code concrete evidence before writing themes, distinguishing observed behavior, participant statement, researcher interpretation, and quantitative signal.
3. Cluster recurring behaviors, needs, contexts, workarounds, barriers, motivations, and mental models without forcing every participant into one theme.
4. Preserve contradictions and meaningful outliers, especially where segments or contexts explain different behavior.
5. Assess confidence from method, sample, consistency, evidence quality, and alternative explanations rather than theme frequency alone.
6. Write findings as claims supported by representative evidence and state the scope they can reasonably generalize to.
7. Translate findings into product implications, opportunities, or open questions without pretending research owns prioritization or solution choice.
8. Store supporting evidence and synthesis artifacts so future researchers can audit, reuse, or challenge the conclusion.

## Decision rules
- A cluster of sticky notes is not a finding until the underlying evidence and claim are clear.
- Preserve negative and contradictory evidence instead of editing the story into false unanimity.
- Participant quotes illustrate evidence; they do not replace analysis.
- Separate user need from one participant's proposed solution.

## Quality gate
Synthesis is ready when findings trace to source evidence, confidence and scope are explicit, contradictory patterns and segments are represented, user needs are separated from solution requests, and product implications can be reviewed without losing the evidence underneath them.