Students ask for a number because a number is reassuring. Supervisors rarely give one, because the honest answer depends on how saturated the field is, not on a page count. A more useful question is: does your current source set let you write the argument you’re trying to make, without a reader noticing a hole?
Coverage, not count, is the actual target
A field with twelve well-known studies is fully covered at twelve. A field with three hundred papers on a broad topic is barely covered at fifty if your specific angle only appears in a handful of them. Count the sources that speak directly to your actual claim, not everything that shares your keywords — a large bibliography of tangential citations does not protect a thin argument.
A rough calibration by document type
For a course paper or literature-review chapter section, 15–30 closely relevant sources per major theme is typical. For a thesis or dissertation literature review, expect 60–150 depending on field maturity — humanities and interpretive fields often run higher due to fewer meta-analyses to lean on; fast-moving technical fields can run lower because recent work supersedes older work quickly. For a systematic review, the number is an output of your protocol, not an input you choose in advance.
Signs you have enough
You are seeing the same three or four studies cited by everyone new you find. New searches return sources you have already screened out. You can state, for each theme, what the “state of the field” claim is and name the two or three sources that would be cited if someone challenged it. When new sources stop changing your synthesis sentence, you have reached saturation for that theme, whatever the count says.
Signs you don’t
You cannot name the strongest counter-evidence to your own position. A key subtopic is represented by only one study. You are relying on secondary citations — citing what a review says a study found, rather than reading the study itself.
Use a coverage check, then read the edges
Make a small matrix with your central themes down one side and direct, contextual, and counter-evidence across the top. It quickly reveals a theme supported only by background material or a conclusion resting on one study. A coverage view in the research workspace can make that audit faster, but it should prompt another search or a closer reading, not become evidence of completeness on its own.