A research synthesis covers a topic with five subareas. Two had robust source data (~15 strong sources each). Three had limited sources (1–2 sources, of varying quality). The current synthesis presents all five subareas in identical prose with identical apparent confidence — a reader can't tell which findings are well-supported and which rest on a single weak source.
What's the right pattern?
Why did you pick that answer? Two or three sentences. The act of articulating it is what builds the judgment — not the click that follows.
Coverage annotation is what distinguishes a credible synthesis from a confident-sounding one. A reader of a research report deserves to know not just what was found but how well-established it is. Marking the well-supported and the thinly-supported areas explicitly — with the source counts, the methodological notes, the gaps where evidence was unavailable — is honest, defensible, and useful. It also prevents a downstream decision-maker from acting on a single-source finding as though it were robust.
Dropping the under-supported areas hides their existence. The decision-maker doesn't know there are subareas where the evidence was thin — they assume the report is complete on what it covers. Suppression is worse than honest annotation.
Sometimes the sources genuinely don't exist, and you can search forever without filling the gap. The synthesis still has to ship; the right response is to surface the gap, not delay indefinitely.
A generic "some sections may be less complete" doesn't help the reader figure out which sections — they have to do the work the synthesis was supposed to do. Specific annotation is dramatically more useful than generic hedging.