Research and sources
Narrow a research question to something answerable
A topic like "remote work" cannot be answered, but "which scheduling practices did these three published studies link to fewer missed handoffs" can. Start from the decision the answer will feed. Then set limits on population, period, place and source type until the evidence you can reach would settle the question.
A text to start with
My broad topic is [topic]. I need an answer by [date] to help me decide [decision]. I can spend about [hours] and can reach these sources: [source types]. Propose three narrower questions, each with a population, time period and source type. For each, say what would count as an answer and what it leaves out. Ask me anything you need before recommending one.
Example request. Change the details to fit your situation.Question scope sheet
Broad topic: [ ] Decision this supports: [ ] Decision owner and deadline: [ ] Time available: [ ] Draft question: [ ] Population or subject: [ ] Time period: [ ] Place or setting: [ ] Source types allowed: [ ] What counts as an answer: [ ] Excluded on purpose, with reason: [ ] Evidence needed: [ ] Evidence I cannot reach: [ ] Trial search result and adjustment: [ ] Final question: [ ]
Signs the question is still too large
Check whether a narrowed question still contains a word that hides several questions, such as "effective," "better" or "impact." Ask who the question applies to, which years it covers and what kind of evidence would end the search. A question that needs sources you cannot access belongs to a later task. Also check that narrowing has not changed the decision you started with. If you could not tell when you were finished, the scope is still open. Compare the evidence list against your available hours.
Ask another agent to check the result
Challenge this research question as too broad for [hours] of work and the sources listed. Point to vague terms, unbounded populations or periods, and evidence I cannot reach. Say what would stop me from knowing when I am done. Suggest one narrower version that keeps the decision it supports.
Test the question against real sources
Pick the narrowed question and run a short trial search. Note how many relevant sources turn up and whether they cover your population and period. If the count is far too high or too low, adjust one limit and search again. Write the final question, its exclusions and the evidence you will accept before you read in depth. Share it with whoever owns the decision.