--- name: guide-research-report description: Guide to writing a decision-oriented research report. Covers question, recommendation, evidence, alternatives, limits and open decision. user-invocable: true metadata: kind: method origin: local pack: sdlc domain: development phase: cross-phase scope: generic invocation: both role: support --- # Research report guide A research report answers a bounded question and supports a decision. It is not a reading list, a diary of investigation or a claim of certainty the evidence does not support. ## Reader questions 1. What question was asked, and why does it matter now? 2. What is the recommendation, and what decision does it support? 3. What evidence was observed, from which sources and at which revision? 4. Which alternatives were considered? 5. What are the limits, assumptions and conflicting evidence? 6. What remains open, and who owns the decision? ## Minimum content - question and scope; - recommendation; - evidence: primary sources, probes, observations and revisions; - alternatives and tradeoffs; - limits, assumptions and contradictions; - open decision and owner; - references. ## Structure and quality - Put the recommendation and its confidence near the top. - Separate observed evidence from inference and opinion. - Cite primary sources; avoid citing a summary as if it were the source. - State what was not investigated. - Record the probe environment and revision when behavior was observed. - Compare alternatives on the dimensions that could change the decision. - Keep the report proportionate to the question. - If the evidence is insufficient, say so and name what would settle it. ## Common failures - a literature review with no recommendation; - a recommendation with no evidence or alternatives; - mixing observation with speculation; - outdated or secondary sources; - ignoring contradictory evidence; - a local probe presented as production behavior; - no owner for the open decision; - confident language for weak evidence. ## Review Read the report as the decision owner: - Can you state the recommendation and why? - Is the evidence traceable and current? - Are the alternatives fairly represented? - Are the limits and disagreements visible? - Is the next decision and owner clear? ## Handoff The decision owner needs the recommendation, evidence, alternatives, limits and open decision. If the research supports an existing change, link or summarize the conclusion there rather than creating a second narrative.