How long should a literature review chapter be?

The number matters less than what makes it drift. Review chapters run long for one of three reasons, and each has a different fix, so diagnosing which one you have is more useful than a word target.

The first is coverage anxiety: the belief that the chapter must demonstrate you read everything. It must not. It must establish that your question is worth asking, and an examiner does not measure your reading by the reference count.

The published guidance says the same thing. Lingard calls the chapter that reproduces everything you read a "laundry list", and argues that a review should summarize only the part of the scholarly conversation needed to make your contribution informed and relevant (Lingard, 2018, Part I, pp. 47-48). Lingard also reports that in published medical education papers the review runs anywhere from a few paragraphs to a few pages. That is a journal figure and does not transfer to a thesis chapter, but it shows what sets the length, which is the space required to situate the contribution and name the gap (Lingard, 2018, Part I, pp. 47-48).

The second is structural. A chapter organized by author repeats itself, because five studies sharing a finding produce five paragraphs. Regrouping by claim collapses those into one paragraph with five citations, and reliably removes a quarter to a third of the words without losing a single idea.

The third is that the empirical work is behind. The review is the chapter you can always make progress on, so it absorbs the time that the analysis is not yet ready to take. If your review is at 30,000 words in year two, the length is a symptom rather than the problem.

Check your regulations too. Some institutions set an overall word limit that includes footnotes, and a long review chapter can leave you cutting analysis at the end, which is the worst place to lose words.

How many sources does that come to?

Typically 80 to 200 works cited in the chapter, depending on field and on how much of the literature is contested. The number follows from the argument rather than setting it, and a chapter with 400 references is usually citing much of it once, in passing, to show coverage.

Fields differ substantially. A theoretical chapter in the humanities may engage deeply with thirty works. A biomedical review may cite three hundred. Neither number tells you anything about the other field.

The maturity of the field changes the job more than the count does. Lingard contrasts a heavily studied area, such as professionalism in medical education, with a relatively understudied one. In the crowded area you have to argue explicitly that a gap still exists, which takes more words per claim even though you cite a smaller selection, and in both cases the review cannot include everything (Lingard, 2018, Part I, p. 48).

Pautasso puts the limit on the topic rather than on the reference list. If the topic is not defined narrowly enough, the relevant literature can run to thousands of publications, and a review of that is not useful to anyone (Pautasso, 2013, p. 1). A reference list that keeps growing usually means the question has not been bounded yet.

The more useful count is how many sources appear more than once. Works cited repeatedly are the ones your argument actually uses, and if that set is small the chapter is broad rather than deep.

What do I cut when the chapter is too long?

Any section that does not establish a claim your research question depends on. Also cut paragraph-per-author passages, which compress heavily when regrouped by claim, and background that a reader in your field already has. Those three cuts usually recover a third of the length.

Cut into a separate file rather than deleting. Some of it becomes a paper, some becomes an appendix, and some comes back when a reviewer asks a question you had already answered.

A stopping rule works better than a word target for deciding when to stop cutting. Lingard treats mapping the gap as the review's purpose, and warns that describing the known material too fully makes your own study look unnecessary (Lingard, 2018, Part I, p. 49). Pautasso sets a compatible bar, that the review should discuss the literature critically, identify methodological problems, and leave the reader with the field's main debates and open questions (Pautasso, 2013, p. 3). Once your chapter does those things, more pages are not justified by the fact that it is a PhD rather than an MA.

Do the cutting in a pass of its own, after the chapter is complete. Trying to write concisely and write at all in the same session slows both, and it is much easier to see what is unnecessary once the whole argument is on the page.