Research Synthesis Template: Evidence, Patterns, Contradictions and Decisions

A research synthesis template helps a team move from scattered interviews, observations, documents and measures to defensible patterns and decisions. Good synthesis keeps the chain from source to interpretation visible. It does not turn the loudest quote or most convenient theme into universal truth.

The template supports organization, not methodological validity or professional advice. Qualified researchers and domain owners must choose appropriate methods, consent, sampling, privacy controls and decision thresholds. Preserve uncertainty and avoid claims beyond the evidence collected.

Frame the synthesis question and scope

Write the decision or learning question the synthesis should inform. Define the population, product, service, journey stage, geography and time period in scope. State what is excluded. A question such as “Why do customers leave?” is too broad unless the relevant customer group and evidence window are clear.

Name the research owner, contributors, reviewers and decision owner. Separate who collected evidence, who interpreted it and who decides what to do. Record the intended use, because exploratory discovery, service improvement and high-stakes policy decisions require different rigor.

List known constraints, including recruitment gaps, inaccessible groups, missing operational data or translation limitations.

Copy this research synthesis template

RESEARCH SYNTHESIS WORKSPACE

Research question / decision to inform:
Population, context and period in scope:
Methods / sample / known limitations:
Research owner / reviewers / decision owner:
Consent, privacy and access controls:

EVIDENCE UNIT
Source ID / method / date / segment:
Observation or data point:
Context / exact location in source:
Researcher interpretation:
Code / confidence / limitation:

PATTERN REGISTER
Candidate pattern:
Supporting evidence IDs:
Contradicting evidence IDs:
Segments or conditions where it appears:
Alternative explanation:
Confidence and rationale:
Evidence gap / next research step:

DECISION BRIDGE
Supported implication:
Option or hypothesis:
Risk / assumption / guardrail:
Decision / owner / date:
Validation measure / review point:

Adapt the structure to the method and risk. Do not let a template create false comparability between sources or imply that an unsupported conclusion has been validated.

Build a controlled evidence inventory

Create one record for every source: interview, observation, survey, support ticket set, analytics extract, document or experiment. Include date, method, participant or segment code, collector, storage location and consent or usage restrictions. Use pseudonymous identifiers where direct identity is unnecessary.

Record source quality and limitations. A recalled anecdote, a verbatim transcript and a system measure are not interchangeable. Preserve original material in restricted storage and link to it rather than copying sensitive data into a broad synthesis board.

An audit evidence log template provides useful patterns for provenance, access and review, even when the research is not an audit.

Normalize observations without flattening context

Break source material into evidence units small enough to compare, while retaining context. Record what was observed or said separately from the researcher’s interpretation. Include the source location so another reviewer can check it.

Use consistent labels for segments, journey stages and research methods. Normalize obvious formatting differences, but do not rewrite participant language until it fits a preferred narrative. Translation and transcription uncertainty should remain visible.

Avoid treating frequency in a convenience sample as population prevalence. A repeated observation may be important, but its meaning depends on the sampling approach and question.

Code evidence with a reviewable approach

Begin with descriptive codes close to the evidence, then group them as understanding develops. Maintain a codebook with definitions, inclusion and exclusion examples, and changes over time. Allow new codes when evidence does not fit the initial frame.

For collaborative coding, compare a subset and discuss differences. The goal is not artificial unanimity; disagreement can expose ambiguous definitions or different assumptions. Record significant coding decisions.

Use a research debrief template for practical early clustering while keeping source IDs attached to every note.

Keep an uncoded or “does not fit” queue visible. Forcing every observation into an existing code can hide emerging evidence. Review that queue at defined intervals and decide whether to add a code, revise a definition or preserve the item as an exception.

When several researchers code the same material, record major disagreements and their resolution. Do not calculate agreement statistics unless the method and decision need them and a qualified researcher has defined the approach.

Develop patterns from multiple evidence units

A pattern should explain a meaningful relationship across evidence, not merely rename a pile of quotes. State the pattern, where it appears, conditions that shape it and which source IDs support it. Look for variation across segments and methods.

Distinguish observation, interpretation, hypothesis and implication. “Participants paused at checkout” is an observation; “they did not trust payment security” is an interpretation requiring further evidence. This separation prevents confident wording from substituting for support.

Use a public consultation report template as a structural example when the reviewed synthesis needs a clear audience-facing account of methods, themes, limitations and response.

Test each pattern against source diversity. Ten similar comments from one workshop may represent one shared context, not ten independent confirmations. Conversely, one observation from a critical edge case may deserve action even without frequency. Explain why the pattern matters rather than relying on counts alone.

Preserve contradictions and negative cases

Actively search for evidence that does not fit each candidate pattern. Record which sources contradict it and whether context, segment, timing or method could explain the difference. Do not hide negative cases to make a cleaner presentation.

Some contradictions are real and decision-relevant. New users may need guidance that experienced users find intrusive; buyers and daily users may value different outcomes. State the boundary conditions instead of averaging them into a vague middle.

Where the conflict remains unresolved, label it and propose targeted research. Uncertainty is a useful output when it prevents an overgeneralized decision.

Assess confidence and evidence gaps

Define a simple confidence approach appropriate to the work. Consider source diversity, directness, recency, consistency, sample coverage, method limitations and contradictory evidence. Avoid precise numerical confidence scores unless the method genuinely supports them.

Separate “not observed” from “does not exist.” Identify missing segments, inaccessible contexts and questions the study was not designed to answer. Prioritize gaps based on decision risk, not curiosity alone.

An UX research plan template can turn a material gap into a scoped follow-up with method, participants, ethics and decision use defined.

Maintain a limitations register throughout analysis rather than writing limitations at the end. Link each limitation to the claims it constrains. This makes it harder for a short summary to present a broad recommendation while the caution remains buried elsewhere.

Connect patterns to implications carefully

For each supported pattern, ask what it may mean for the decision and what assumption connects the two. Keep that reasoning visible. Evidence that users struggle with a step may support redesign, guidance, segmentation or further study; it does not dictate one solution automatically.

Prioritize implications with an explicit lens such as user harm, decision urgency, reversibility and evidence confidence. Keep business attractiveness separate from evidential strength. A desirable idea is not better supported simply because stakeholders prefer it.

Generate multiple options and identify tradeoffs, affected groups and possible unintended effects. Domain, legal, safety or accessibility specialists should review implications within their authority. Research evidence informs accountable judgment; it does not replace it.

Document consequential choices in a decision log template with rationale, dissent, owner and review date.

Review synthesis collaboratively

Invite researchers, domain experts and decision owners to challenge source coverage, code definitions, pattern boundaries and confidence. Include people close to the experience where appropriate, without exposing participant identity or asking stakeholders to override inconvenient evidence.

Prepare a review packet with the synthesis question, method summary, pattern cards, contradictions and limitations. Give reviewers a way to trace claims without granting unnecessary access to raw identifiable material. Record proposed edits and whether they correct evidence, refine interpretation or change a decision.

Distinguish factual corrections from interpretive disagreement. Preserve meaningful dissent and identify what evidence could resolve it. Do not let seniority determine which interpretation “wins” without examination.

Use meeting follow-up practices to capture reviewed decisions and research actions while keeping restricted source data outside the general summary.

Use AI and Kuno with accountable review

AI can help sort authorized text, suggest candidate codes and draft summaries, but it may omit context, merge distinct concepts or amplify prompt assumptions. Test outputs against source IDs, inspect negative cases and keep researchers accountable for every conclusion.

Research sessions and synthesis workshops may contain personal, confidential or proprietary information. Obtain required authorization and informed notice, minimize collection, restrict access and honor consent and retention terms.

For an authorized synthesis workshop, Kuno can help draft discussion notes and action items for human verification. It does not validate methods, infer participant intent or make research decisions. Explore Kuno

Never upload source material to an unapproved system merely for convenience.

Publish decisions and maintain traceability

Produce an audience-appropriate summary with the question, methods, limitations, patterns, contradictions, confidence and implications. Link claims to source IDs or controlled evidence views. Remove identifying details and avoid decorative quotes that expose participants without adding analytical value.

Tailor detail without changing the conclusion’s boundaries. A short leadership summary may omit process detail, but it must retain material uncertainty, excluded populations and unresolved contradictions. Keep a fuller research record available to authorized reviewers.

Plan maintenance at the point of publication. Name who can amend the synthesis, how new evidence is reviewed and when the decision owner will revisit assumptions. A synthesis that remains visible after its context expires should carry a clear date and scope warning.

Record what decision was made, by whom, under which assumptions and when it will be reviewed. Later outcomes should update the learning record without rewriting the original evidence. Version the synthesis when new research materially changes a pattern.

Keep synthesis discussions searchable without losing provenance. Kuno supports consented capture and draft follow-up; researchers verify evidence, protect participants and retain accountable judgment. See Kuno

FAQ

The frontmatter FAQ provides concise answers without repeating them in the body.