Working School & college skill
Synthesize research evidence
Combine diverse research into a claim-by-claim synthesis that preserves disagreement, quality differences, and uncertainty.
Synthesis is not a stack of summaries. Organize evidence around questions and explain why the body of evidence supports each conclusion.
When to use
- Use after source collection for a paper, brief, review, recommendation, or evidence map.
- Use when studies differ in population, definition, method, period, or result.
Procedure
- Restate the research question, decision context, scope, and acceptable evidence.
- Build an evidence matrix with claims or subquestions as rows and sources as columns.
- Normalize only genuinely comparable concepts, measures, populations, and time periods.
- Weight evidence by directness, methodological quality, relevance, independence, and precision.
- Group consistent findings, conflicts, boundary conditions, gaps, and plausible explanations.
- Test whether apparently independent sources share data, authors, or an upstream report.
- Draft each finding as claim, supporting evidence, counterevidence, uncertainty, and implication.
- Calibrate language: distinguish establishes, supports, suggests, is consistent with, and is unknown.
- Verify every material sentence against the matrix and identify what new evidence would change the conclusion.
Failure plan
- Do not vote-count studies or let one vivid source outweigh a stronger body of evidence.
- Do not average incompatible measures or erase meaningful heterogeneity.
- If evidence is sparse or biased, synthesize the limitation rather than manufacture a conclusion.
- Keep factual uncertainty separate from value judgments and decision preferences.
--- name: synthesize-research-evidence category: school description: Combine diverse research into a claim-by-claim synthesis that preserves disagreement, quality differences, and uncertainty. Use when moving from collected sources to findings. --- # synthesize-research-evidence Synthesis is not a stack of summaries. Organize evidence around questions and explain why the body of evidence supports each conclusion. ## When to use - Use after source collection for a paper, brief, review, recommendation, or evidence map. - Use when studies differ in population, definition, method, period, or result. ## Procedure 1. Restate the research question, decision context, scope, and acceptable evidence. 2. Build an evidence matrix with claims or subquestions as rows and sources as columns. 3. Normalize only genuinely comparable concepts, measures, populations, and time periods. 4. Weight evidence by directness, methodological quality, relevance, independence, and precision. 5. Group consistent findings, conflicts, boundary conditions, gaps, and plausible explanations. 6. Test whether apparently independent sources share data, authors, or an upstream report. 7. Draft each finding as claim, supporting evidence, counterevidence, uncertainty, and implication. 8. Calibrate language: distinguish establishes, supports, suggests, is consistent with, and is unknown. 9. Verify every material sentence against the matrix and identify what new evidence would change the conclusion. ## Failure plan - Do not vote-count studies or let one vivid source outweigh a stronger body of evidence. - Do not average incompatible measures or erase meaningful heterogeneity. - If evidence is sparse or biased, synthesize the limitation rather than manufacture a conclusion. - Keep factual uncertainty separate from value judgments and decision preferences. ## Done - Findings are organized by question, not source order - Support, conflict, quality, dependence, applicability, and uncertainty are explicit - Every material conclusion is traceable and no claim is stronger than the evidence body