A literature review is a structured survey and critical synthesis of the existing scholarship on a defined topic. It maps what is already known, evaluates the strengths and weaknesses of that work, identifies gaps and debates, and positions new research within the field. Unlike a summary, it organises and interprets sources rather than listing them.
Key Takeaways
- A literature review synthesises and evaluates existing research; it is not a list of summaries.
- The right type (narrative, scoping, systematic, integrative, umbrella, rapid) depends on your question, timeline, and field.
- A strong review is built around themes, debates, and gaps, not one source per paragraph.
- Reporting standards such as PRISMA 2020 apply to systematic and scoping reviews.
- AI tools can speed up discovery and screening, but citations must be verified against the original papers.
What is a literature review?
A literature review is an organised, critical account of the published work on a topic that summarises, compares, and evaluates that work to establish what is known, where scholars disagree, and what remains unanswered. It can be a standalone review article or a section within a thesis, dissertation, or research paper.
Two distinctions matter from the start.
Summary vs. synthesis. A summary restates what individual sources say. A synthesis reorganises that material around themes, methods, or debates so that the sources speak to one another. A review built mostly of summaries reads like a catalogue; a review built on synthesis reads like an argument about the state of the field. University writing centers consistently identify this as the line between a weak and a strong review.
What a literature review is not. It is not a book review (a judgement of one work), not an annotated bibliography (a list of sources with notes), and not original empirical research — a review is a secondary source that does not report new experiments or data.
Why literature reviews matter
A good review does practical work. It demonstrates command of the field, justifies your research question, prevents duplication, reveals methods and theories you can build on, and exposes the gap your study will fill. For readers who are new to a topic, a well-made review is also an efficient entry point into a body of knowledge.
Types of literature reviews
There is no single “literature review.” Methodologists have catalogued many distinct review types — a widely cited typology by Grant and Booth (2009) identified 14 — but a handful account for most academic work. The choice depends on your research question, time and resources, required comprehensiveness, and discipline.
| Review type | Core purpose | Method rigour | Typical use | Watch-outs |
|---|---|---|---|---|
| Narrative (traditional) | Broad, flexible overview of a topic | Lower / flexible | Coursework, thesis intro, humanities & social sciences | Risk of bias; state scope and key terms clearly |
| Scoping | Map the size and range of evidence; find gaps | Structured; follows a protocol | Emerging fields; deciding whether a systematic review is warranted | Maps rather than appraises findings |
| Systematic | Answer one focused question using all eligible evidence | High; explicit, reproducible methods | Clinical questions; evidence-based practice | Time- and resource-intensive |
| Integrative | Combine diverse methodologies into new frameworks | Moderate-high | Theory development; nursing, management | Synthesising mixed methods is demanding |
| Umbrella (review of reviews) | Synthesise existing systematic reviews/meta-analyses | High | Broad, well-studied topics with many reviews | Requires existing high-quality reviews |
| Rapid | A streamlined systematic review under time pressure | Moderate (some steps simplified) | Policy and fast-moving health questions | Transparency about what was simplified |
How to choose. Broad understanding or a thesis chapter → narrative. Mapping an under-explored area → scoping. A precise, answerable question with available evidence → systematic. Mixed evidence and theory-building → integrative. Many systematic reviews already exist → umbrella. Urgent decision, limited time → rapid. No type is inherently superior; the best choice is the one that fits the question.
How to write a literature review: 7 steps
- Define your question and scope. Clarify the assignment or aim, set boundaries (time period, populations, disciplines), and decide which review type fits. A narrower question makes the workload manageable.
- Build a search strategy. List keywords, synonyms, and broader/narrower terms. Search scholarly databases appropriate to your field (for example Google Scholar, PubMed, Scopus, Web of Science, JSTOR, PsycINFO). Record your searches so they are reproducible.
- Screen and select sources. Read titles and abstracts first to triage. Prioritise peer-reviewed, recent, and seminal works. Apply consistent inclusion/exclusion criteria.
- Read critically and take structured notes. For each source capture the question, method, key findings, limitations, and how it relates to other studies. Track full citation details as you go.
- Identify themes, patterns, debates, and gaps. Look for recurring concepts, methodological trends, points of disagreement, pivotal studies, and what is missing. These become your organising structure.
- Outline and synthesise. Group sources under themes or sub-questions — not one paragraph per author. Each paragraph should make a point supported by several sources.
- Write, cite, and revise. Draft introduction, body, and conclusion; cite consistently (e.g., APA 7th); then revise for flow, balance, and accuracy. Have someone read it before you finalise.
How to structure a literature review
A review usually has three parts:
- Introduction: define the topic and scope, state the purpose and organising concept, and explain what is and is not covered.
- Body: the synthesis, organised by one of the common patterns below.
- Conclusion: summarise the main trends, restate the gap your work addresses, and connect the topic to the wider field.
Common organisation patterns.
- Thematic — grouped by recurring themes or concepts (most common and usually strongest).
- Chronological — traces how the field developed over time.
- Methodological — groups studies by the methods they used.
- By debate / theoretical — organised around competing positions or frameworks.
A useful diagnostic: if every paragraph begins with a researcher’s name, you are probably describing rather than synthesising.
How to synthesise sources
Synthesis means making sources talk to each other. A synthesis matrix is the simplest tool: a grid with sources as rows and themes as columns. Filling it in reveals where studies agree, conflict, or leave gaps — and those rows-into-columns become your paragraphs.
Descriptive vs. analytical writing (mini-example).
- Descriptive (weaker): “Smith found X. Jones found Y. Lee found Z.”
- Analytical (stronger): “Although early work reported X, more recent studies converge on Y; the remaining disagreement centres on Z, which no study has tested directly.”
The analytical version groups findings, signals relationships with linking words (although, however, converge), and ends by naming the gap — the move that turns a summary into a review.
Example of a literature review
Below is a short, original worked example built around a single theme so you can see the moves in practice. The topic is the effect of remote work on employee productivity. The author names and years are illustrative placeholders to demonstrate form — they are not real studies and must not be cited.
Step 1 — The research question. How does remote work affect employee productivity, and what explains the conflicting findings?
Step 2 — A synthesis matrix. Before writing, the sources are arranged by theme so patterns become visible:
| Source (illustrative) | Finding on productivity | Key factor emphasised | Method |
|---|---|---|---|
| Author A (2019) | Increases | Fewer interruptions, autonomy | Survey |
| Author B (2021) | Increases, then plateaus | Home setup, role type | Longitudinal |
| Author C (2021) | Decreases | Weaker team communication | Case study |
| Author D (2023) | Mixed | Depends on task complexity | Mixed-methods |
The matrix already reveals the structure: studies agree productivity changes, disagree on direction, and the disagreement seems to hinge on mediating factors (autonomy, communication, task type).
Step 3a — Descriptive version (weaker).
Author A (2019) found that remote work increased productivity. Author B (2021) also found an increase but said it levelled off. Author C (2021) found a decrease. Author D (2023) found mixed results.
This is a list. Each sentence stands alone, the sources do not interact, and no gap emerges.
Step 3b — Synthesised version (stronger, annotated).
Early evidence suggested that remote work raises productivity, largely by reducing interruptions and increasing autonomy (Author A, 2019). [Opens with a theme and groups the “increase” camp.] More recent work complicates this picture: while productivity gains appear in the short term, they may plateau as the novelty fades and depend heavily on the employee’s home setup and role (Author B, 2021). [Uses “complicates” to signal tension and introduces mediating factors.] A competing strand reports declines, attributing them to weaker team communication and coordination costs (Author C, 2021). [Names the opposing camp and its mechanism.] These apparently contradictory results may be reconciled by task complexity: remote work appears to benefit independent, focused tasks but to hinder highly interdependent ones (Author D, 2023). [Synthesises both camps under one explanatory factor.] Taken together, the literature has moved from asking whether remote work helps to asking under what conditions it does — yet few studies isolate task type and communication structure simultaneously, leaving this interaction underexplored. [States the trend and names the precise gap the new study can fill.]
Why the second version works. It is organised by a theme (not one paragraph per author), it makes the sources argue with each other, it uses linking language to signal relationships, and it ends by naming a specific, researchable gap — exactly the function a literature review is meant to serve.
Real examples to study. To see full, published reviews, look for articles whose titles include “: A Literature Review” or “: A Systematic Review” in databases such as Google Scholar, PubMed, or your library, and read how they handle introduction, thematic structure, and the concluding gap statement. Publicly available teaching examples are also collected by university writing centers and guides such as Scribbr.
Literature review vs. systematic review vs. annotated bibliography
| Feature | Literature (narrative) review | Systematic review | Annotated bibliography |
|---|---|---|---|
| Goal | Broad overview and synthesis | Answer one focused question | List and describe sources |
| Method | Flexible | Explicit, reproducible protocol | None required |
| Comprehensiveness | Selective | Exhaustive within criteria | Selective |
| Output | Integrated narrative | Structured findings, often a meta-analysis | Entries with short notes |
| Synthesis across sources | Yes | Yes | No |
Reporting standards and quality
For more rigorous reviews, reporting standards improve transparency and reproducibility. The PRISMA 2020 statement is the standard for systematic reviews, and the PRISMA extension for scoping reviews (PRISMA-ScR, 2018) provides a checklist for scoping reviews. Even for a narrative review, borrowing PRISMA’s habits — documenting your search and selection — strengthens credibility.
AI and digital tools for literature reviews
AI and specialised research tools have become a normal part of the workflow, but they help most when matched to a specific stage rather than used to “write the review.” A risk-aware breakdown:
- Discovery: Semantic Scholar, Research Rabbit, and Connected Papers find related papers and map citation networks by similarity rather than keywords.
- Screening (systematic reviews): dedicated screening tools (e.g., Rayyan) and structured-extraction tools (e.g., Elicit) speed up triage of large sets.
- Evidence Q&A: Consensus surfaces what multiple papers conclude on a focused question.
- Citation context: Scite shows whether later work supported or contradicted a study.
- Reference management: Zotero, Mendeley, and EndNote organise sources and format citations.
The non-negotiable caveat. General-purpose chatbots can fabricate citations, misattribute findings, or oversimplify field-specific concepts. Any AI output must be verified against the original paper — confirm the DOI, authors, and that the source actually says what is claimed. Use AI to accelerate the mechanical work; keep the judgement, interpretation, and final synthesis human. (Vendor surveys report rising AI adoption among researchers; treat such figures as vendor-reported rather than independently verified.)
Common mistakes and how to avoid them
- Summarising instead of synthesising — organise by theme, not by source.
- One paragraph per author — a sign of description; group studies that speak to the same point.
- No clear organising concept — anchor the review to your research question.
- Ignoring disagreement — include studies and findings that complicate the picture.
- Outdated or thin sourcing — in fast-moving fields, recency matters; in the humanities, foundational works still count.
- Citation errors and accidental plagiarism — track sources carefully and cite consistently.
- Trusting AI-generated references — verify every one.
How long should a literature review be?
There is no fixed length. It depends on the assignment, the breadth of the question, and how much relevant literature exists. A coursework review may be a few pages; a dissertation chapter can run tens of pages. Comprehensiveness within your scope matters more than word count — always follow your supervisor’s or journal’s guidelines.
Conclusion
A literature review is less a chore than a thinking tool: done well, it turns scattered studies into a clear map of a field and a justification for your own work. Choose the review type that fits your question, build the structure around themes and gaps rather than individual authors, synthesise instead of summarise, and use AI tools to accelerate — never replace — your judgement. The result is a review that informs your readers and earns their trust.
FAQ
What is a literature review in simple terms? It is an organised, critical overview of the existing research on a topic that synthesises what is known, evaluates it, and identifies gaps — not just a list of summaries.
What are the main types of literature reviews? The most common are narrative, scoping, systematic, integrative, umbrella, and rapid reviews. The right one depends on your question, timeline, and field.
How is a literature review different from a systematic review? A narrative literature review gives a flexible, broad overview; a systematic review follows an explicit, reproducible protocol to answer one focused question as comprehensively as possible.
Is a literature review the same as an annotated bibliography? No. An annotated bibliography lists sources with short notes; a literature review synthesises sources around themes and arguments.
How many sources should a literature review include? There is no set number. Include enough to cover the key themes, debates, and seminal works in your scope — quality and coverage matter more than a target count.
How do I synthesise instead of summarise? Group sources by theme using a synthesis matrix, then write paragraphs that compare findings and signal relationships, ending each theme by naming the gap.
Can I use AI to write my literature review? AI can help find, screen, and organise papers, but it can fabricate citations and misread findings. Verify every source against the original and keep the analysis your own.
How long should a literature review be? It varies by assignment and topic breadth; follow your institution’s or journal’s guidance and prioritise completeness within your scope over a word count.
References
- Grant, M. J., & Booth, A. (2009). A typology of reviews: an analysis of 14 review types and associated methodologies. Health Information & Libraries Journal, 26(2), 91–108.
- Snyder, H. (2019). Literature review as a research methodology: An overview and guidelines. Journal of Business Research, 104, 333–339.
- Page, M. J., et al. (2021). The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ, 372, n71.
- Tricco, A. C., et al. (2018). PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation. Annals of Internal Medicine, 169(7), 467–473.
- Torraco, R. J. (2005). Writing Integrative Literature Reviews: Guidelines and Examples. Human Resource Development Review, 4(3), 356–367.
- University of North Carolina Writing Center — Literature Reviews (writingcenter.unc.edu).
- University of Toronto — The Literature Review: A Few Tips on Conducting It (advice.writing.utoronto.ca).
- University of Arizona Libraries — Conduct a literature review (lib.arizona.edu).
- Stanford Lane Medical Library / UNMC Libraries — Types of reviews guides.
