Why chapters get sent back, the difference between summarising and synthesising, and how to make a gap out of a pile of reading.
A literature review argues; it does not report. Its job is to show what the field currently believes, where that account runs out, and why your question follows from that limit. The single most common reason a chapter gets returned is that it summarises studies one after another instead of organising them around ideas — and the fix is structural rather than a matter of writing better.
Most rejected literature review chapters share one shape. Paragraph one describes a study. Paragraph two describes another. Paragraph three describes a third. Each paragraph is accurate, well referenced and clearly written, and the chapter is still not doing its job.
That is a summary: organised by study, telling the reader what other people found. A synthesis is organised by idea, telling the reader what the field collectively believes, where it disagrees with itself, and what nobody has settled.
The tell is in the sentence openings. If most paragraphs begin with a name and a date — Smith (2019) found…, Patel and Osei (2021) argued… — the chapter is organised by study. If they begin with a claim, and the citations arrive as evidence for it, it is organised by idea.
Compare:
Summary: Smith (2019) found that workload predicted burnout in nurses. Patel and Osei (2021) found a similar relationship in teachers. Alvarez (2022) found no relationship in social workers.
Synthesis: Workload predicts burnout consistently in nursing and teaching, but the relationship does not appear in social work — a divergence that has been attributed to differences in autonomy rather than to workload itself, though this has not been tested directly.
Same three studies. The second says something, and the last clause is the beginning of a research gap.
The structural fix is to decide what the sections are about before deciding what goes in them. Sections should be named for ideas, debates or problems — never for authors, and rarely for chronology.
Ways of organising that work, depending on your field and question:
A practical test before you write: for each planned section, write one sentence saying what the reader will believe at the end of it that they did not believe at the start. If you cannot write that sentence, the section is a container for references rather than an argument.
A second test, for the whole chapter: does it end somewhere it could not have started? A review that arrives at your research question by necessity has worked. One that could have been assembled in any order has not.
The gap is what the entire chapter exists to establish, and it is the part examiners press hardest.
An absence is not a gap. “No study has examined this in my country” describes a hole in coverage; it gives no reason why filling it matters. The obvious follow-up — so what? — has no answer, and a thesis built on that footing is uncomfortable to defend.
A gap is a limit on what the field can currently explain, predict or decide. Ways one legitimately arises:
Each of those produces a consequence, and the consequence is what makes it a gap: because this is unsettled, the field cannot yet do X. If you can complete that sentence, you have a gap. If you can only say nobody has done it, you have a hole.
You will be asked how you found what you found, particularly if something obvious is missing.
Keep a record from the start: which databases, which search terms, which date ranges, and what you included and excluded. This takes minutes as you go and is close to impossible to reconstruct at the end — and for any review claiming to be systematic, it is not optional.
Practical points that improve coverage:
If your review is systematic, the requirements are stricter and specific: a protocol registered in advance, explicit inclusion and exclusion criteria, a documented search strategy, screening by more than one person where possible, a flow diagram accounting for every record, and formal appraisal of study quality. A narrative review does not need that apparatus — but do not describe a narrative review as systematic, because the terms carry expectations and using the wrong one invites a question you cannot answer.
“Be more critical” is the most common supervisor comment and the least specific. It usually means one of a small number of things.
What it does not mean is finding fault with everything. A chapter that dismisses each study in turn reads as posturing, and it also undermines the foundation you are about to build on. Criticism should be proportionate and should serve the argument: the limitations worth naming are those that matter for your question.
“I have read a hundred papers and cannot start writing.” Almost always a structure problem rather than a reading problem. Stop reading. Write the section headings as claims, then see which papers belong under each. The ones that fit nowhere are usually not needed.
“My chapter keeps being sent back.” If the comments concern writing but the same chapter keeps returning, the problem is usually the argument. Ask your supervisor directly what the chapter is supposed to establish — and whether it does.
“New papers keep appearing.” They will, throughout. Set a cut-off date for the main review, note it, and run one update search before submission. Chasing the literature indefinitely is a recognised way of not finishing.
“Everything has already been done.” Usually means the question is still too broad. Narrow it and the gaps appear — they exist at the level of specific mechanisms and conditions, not at the level of topics.
“How long should it be?” Your discipline and institution set expectations, and recent theses passed in your own department are the only reliable guide. It is finished when it has established your gap and justified your question, not when it reaches a word count.
A summary is organised by study and reports what each one found. A synthesis is organised by idea and says what the field collectively believes, where it disagrees, and what remains unsettled. The quickest diagnostic is how your paragraphs open: if most begin with an author and a date, the chapter is organised by study; if they begin with a claim and cite evidence for it, it is organised by idea.
Look for a limit on what the field can explain rather than an absence in its coverage. Contradictory findings nobody has reconciled, an assumption a body of work rests on untested, an established effect with no explained mechanism, or a design limitation preventing the question being answered. The test is whether you can complete the sentence “because this is unsettled, the field cannot yet…”. “Nobody has studied this here” does not complete it.
It usually means one of four things: you are reporting findings rather than judging the evidence behind them, you are not attending to what each study’s design could actually establish, you are accepting the field’s assumptions without examining them, or you are presenting disagreements neutrally rather than taking a position. It does not mean finding fault with everything — a chapter that dismisses each study in turn undermines the foundation you are about to build on.
A systematic review follows a pre-registered protocol with explicit inclusion criteria, a documented search strategy, screening by more than one person where possible, a flow diagram accounting for every record, and formal quality appraisal. A narrative review is a scholarly argument built from the literature without that apparatus. Both are legitimate — but do not call a narrative review systematic, because the term carries expectations you will be asked about.
There is no number, and your discipline’s recent theses are the only useful guide. Coverage is adequate when new searches keep returning material you already have. What matters far more than the count is whether the chapter establishes your gap and justifies your question — a focused review that argues well beats an exhaustive one that lists.
Earlier than feels comfortable. If you have read extensively and cannot begin, that is usually a structure problem rather than a knowledge gap. Write your section headings as claims, then see which papers belong under each — the ones that fit nowhere are generally not needed. Set a cut-off date for the main review and run one update search before submission.
If a review keeps being returned, the problem is usually the argument rather than the writing — and that is quicker to diagnose from outside. Send what you have and a PhD in your field will tell you what it is missing.
Leave an email or a WhatsApp number — whichever you prefer — and tick how we should reach you. We reach out within 30 minutes.
