A Question the Evidence Can Answer

What it is for

How to turn a vague research topic into a question an evidence search can actually settle: one intervention, one outcome, one population, asked as yes/no, then repeated for the opposite claim and for recent years only. Use it before your first search rather than after five disappointing ones. It prevents reading an aggregate answer as a verdict on a question that was never asked.

The skill Copy

This kind of search answers questions, not topics. Shape the question before
you type it.

Write it with one intervention, one outcome, one population:
  "Does <specific thing> change <specific measurable outcome> in <who>?"

  Weak:   is intermittent fasting good
  Strong: does intermittent fasting reduce fasting blood glucose in adults
          with type 2 diabetes

Rules while searching:
1. Ask a yes/no question when you want the aggregated yes/no reading to mean
   anything. A how or why question gives you a reading list, which is useful,
   but it is not a balance of evidence and must not be quoted as one.
2. One variable per search. A question containing two interventions returns
   papers about each, and the summary blends studies that were never
   comparable.
3. Search the opposite claim as its own search. Papers reporting no effect
   are usually phrased in words your first query never used.
4. Use the literature's vocabulary, not the marketing one: the compound
   rather than the brand, the measured outcome rather than the slogan.
5. Narrow the population on purpose - age, condition, trained or untrained,
   animal or human. A result in mice sits in the same list as a result in
   adults and looks identical at a glance.
6. Re-run the same question restricted to recent years. If the balance flips,
   report that instead of averaging across decades.

Before you quote anything, write down the exact question you typed, the
filters you set, and the date. You will need to reproduce it, and the same
question typed loosely gives a different set next week.

The failure this prevents: quoting a confident-looking balance of evidence
that came from a question nobody actually studied.

How to install it Copy

1. Open the tool and start a new conversation, project or template.
2. Paste the text below wherever the tool keeps standing instructions
   — usually called system prompt, instructions, or persona.
3. Save it so it applies to every run rather than a single message.

Did it work for you?

Where it came from