A narrative review tells the story of a field: what is known, what is debated, and where the gaps lie, organised by an author who knows the literature well. Unlike a systematic review, it does not try to find and appraise every study on a question; instead, it selects and interprets the literature to build an argument about the state of knowledge. Done well, it is one of the most useful articles a field can publish, because it saves every reader the work of assembling that picture alone.
This guide explains how to write a narrative review for a journal: how it differs from other review types, how to define a question and select literature transparently, how to organise and synthesise the material, and how to handle disagreement, limitations, and review. If you are planning a review that interprets a field rather than exhaustively auditing it, this is the form to learn.
A narrative review is a scholarly article that surveys and interprets existing literature on a topic, organised around themes, questions, or an argument, rather than around a formal search and appraisal protocol. The author selects studies that matter for the argument, explains how they fit together, and draws conclusions about what is established, what is contested, and what remains unknown. The value lies in the synthesis: a good narrative review says something no single study says on its own.
Narrative reviews are common in fields where knowledge is broad, contested, or developing too fast for a formal protocol to capture well, and where expert interpretation is itself part of the contribution. They are also the form most often written by researchers invited for their expertise. That invitation is not a shortcut: the review is still peer reviewed, and reviewers will judge whether the selection is fair, the interpretation is supported, and the gaps are honestly stated.
The three main review types answer different needs. A systematic review starts with a narrow, pre-specified question, searches exhaustively, appraises studies against set criteria, and often synthesises results quantitatively. A scoping review maps the extent and nature of literature on a broad topic, showing what exists and where the gaps are, without usually judging study quality in depth.
A narrative review sits between breadth and judgement: it covers a topic broadly, but it interprets rather than audits. It does not claim to have found every study, and it does not hide its selectivity; instead, it explains the basis on which literature was chosen and lets the quality of the argument carry the paper. Choosing the wrong type is a common cause of rejection, so check the target journal's definitions and recent examples before you commit. Our guide on how to publish a research paper covers how article types fit into journal selection.
A narrative review needs a question sharp enough to organise the literature, but broad enough to justify a review rather than a single study. Start by stating the question in one sentence: what body of knowledge are you reviewing, and what will readers understand after reading your review that they did not understand before? If the question could be answered by one study, it is too narrow; if it covers a whole discipline, it is too broad.
Test the question against the literature you already know. Can you name the major themes, the key debates, and the turning points? Can you sketch the structure of the review before you search in earnest? If the answer is no, read more before writing, or narrow the question. A review written around a question the author has not yet thought through tends to become a list of summaries, which is the weakest form of the genre.
A narrative review does not need an exhaustive search, but it does need a defensible one. Search the major databases in your field, follow citations from key papers, and include recent work as well as foundational studies. Keep notes on where you searched, what terms you used, and roughly how you decided what to include, because reviewers may ask, and because your own notes will keep the selection honest.
State your selection basis in the paper, briefly but plainly: the scope, the time period, the kinds of studies included, and any deliberate exclusions. You do not need a flow diagram, but you do need to show that the selection was guided by the question, not by convenience or by favouring studies that support a preferred conclusion.
The organisation is the argument. Three structures cover most narrative reviews. A thematic structure groups the literature by topic or mechanism, which suits fields with distinct lines of work. A chronological structure follows how understanding developed over time, which suits topics with clear turning points. A debate structure organises the literature around competing explanations, which suits contested questions.
Choose the structure that best serves your question, and state it early so readers can follow it. Within each section, move from established findings to newer or contested ones, and end each major section with a short synthesis: what this body of work shows, and what it leaves open.
The introduction should do three things quickly: explain why the topic matters, show that existing literature needs synthesis, and state the review's question and scope. Explain what the review covers and what it deliberately leaves out, and why. A scope statement protects you twice: it tells readers what to expect, and it gives reviewers a fair basis for judging whether the coverage is adequate for the stated aim.
Be specific about scope. Name the field, the population or setting where relevant, the time period, and the kinds of studies considered. If the review focuses on one region, species, or method, say so and justify it. A precise scope lets the review be judged on its own terms.
Synthesis is what separates a review from an annotated bibliography. For each theme, compare studies with each other: where do they agree, where do they differ, and why? Differences in population, method, measurement, or setting often explain apparent contradictions, and naming those reasons is more valuable than listing results side by side.
Use synthesis to build toward conclusions. After presenting a body of work, state what it establishes, what remains uncertain, and what the uncertainty implies for practice or future research. Where evidence is strong, say so and show why; where it is weak or mixed, say that too. A review that concludes every section with 'more research is needed' without saying what research, and why, has not finished the job of synthesis.
Disagreement in the literature is material for a review, not an embarrassment. Present competing findings fairly, with their evidence, and explain the sources of disagreement: different methods, different populations, different definitions, or genuinely unresolved questions. Do not resolve a real debate by citing only one side, and do not manufacture consensus where none exists. Readers trust a review that shows its working on contested points.
Name the gaps explicitly, and distinguish kinds of gaps: questions nobody has studied, questions studied only in one setting, and questions where studies exist but disagree. Each kind points to different future work. End the review with a clear account of what is known, what is contested, and what should be studied next, ordered by importance rather than by the order in which topics happened to appear.
Tables and figures earn their place in a narrative review by making complex material graspable. A summary table comparing key studies — population, method, main finding, limitation — lets readers see patterns across a body of work at a glance. A conceptual figure showing how themes relate can carry an argument that paragraphs alone would take pages to make.
Keep display items disciplined. A table should be readable without the whole paper: clear title, defined abbreviations, and only the columns the argument needs. A figure should be legible at publication size, with a caption stating what it shows. If a table merely repeats the text, cut one of them. Check every entry in a summary table against its source study, since errors in tables are easy to make and hard for readers to spot.
A narrative review must state its own limitations, and the main one is selectivity. Because the author chose the literature, the review may reflect the author's perspective, and readers should know that. State the selection basis, acknowledge relevant work you could not cover, and avoid claiming comprehensiveness you do not have. Honesty about selectivity is a strength, not an admission of failure.
Other limitations deserve mention: the time period covered, the languages or databases searched, and any field where evidence is thin. If the review reflects one theoretical perspective, say so. Reviewers judge a narrative review partly on whether its limitations are recognised; a review that claims to be definitive while quietly omitting a whole line of work will be judged more harshly than one that states its bounds plainly.
Submit as the journal's review article type, with a cover letter stating the review's question, scope, and why it fits the journal. Say whether the review was invited, if that applies, and confirm the usual declarations. Keep the manuscript's structure visible: clear headings, a stated scope, and synthesis at the end of each major section help editors and reviewers follow the argument.
Reviewers of narrative reviews often challenge selection and interpretation: a missing body of work, an overconfident conclusion, or a debate presented one-sidedly. Reply point by point, stating what you changed and where, and where you respectfully disagree, why, with evidence. Adding a missing section or softening a conclusion usually strengthens the paper; defending an omission without reason rarely does.
Journals define review types differently, so check the target journal's own guidance. The table below gives the usual pattern.
| Review type | Question | Literature search | Synthesis | Best for |
|---|---|---|---|---|
| Narrative review | Broad, interpretive | Selective, question-guided | Thematic interpretation | Interpreting a field and its debates |
| Systematic review | Narrow, pre-specified | Exhaustive, protocol-driven | Appraisal and often meta-analysis | Answering a precise question definitively |
| Scoping review | Broad, mapping | Broad, systematic mapping | Mapping extent and gaps | Showing what literature exists |
| Literature review chapter | Set by a thesis or course | Guided by the project | Background for a study | Supporting a dissertation or proposal |
Narrative reviews fail most often by summarising instead of synthesising: a sequence of study summaries with no comparison, no argument, and no conclusion about the field as a whole. Other common faults are an undefined question, an unstated selection basis, one-sided coverage of a debate, and conclusions that outrun the evidence presented.
A further mistake is mistaking length for coverage. A long review that wanders across a field without a question or structure is weaker than a shorter review with a sharp question, a stated scope, and a synthesis in every section. Plan the structure before writing, synthesise within sections, and let a colleague who knows the field check whether the review is fair before you submit.
Use this checklist on the final version, before submission:
These authoritative external resources explain the standards and tools mentioned in this guide:
Is a narrative review peer reviewed?
Yes, at reputable journals. Narrative reviews go through peer review like other articles, and reviewers judge the fairness of the selection, the quality of the synthesis, and whether the conclusions follow from the literature presented.
Does a narrative review need a methods section?
It needs a transparent account of how literature was searched and selected, even though it does not follow a formal protocol. State the scope, databases, time period, and inclusion basis briefly, so readers can judge the selection fairly.
How many sources should a narrative review cite?
There is no fixed number. The review should cite the studies its argument needs: foundational work, key recent studies, and work representing each side of a debate. Coverage should match the stated scope, not a target count.
Can students write a narrative review?
Yes, with guidance. A student review needs a sharp question, a stated scope, and a mentor's judgement on selection and fairness. Reviews by authors still learning a field benefit most from an expert co-author or reviewer before submission.
How is a narrative review different from a literature review chapter?
A thesis literature review supports one study, usually by showing the gap it fills. A journal narrative review stands alone: it interprets a field for readers who will not read the author's other work, and its conclusions about the field are the contribution.
Can a narrative review express the author's opinion?
It can present the author's interpretation, provided the interpretation is argued from the literature and competing views are presented fairly. An interpretation supported by evidence is scholarship; an opinion asserted without evidence, or while omitting contrary work, is not.