How to Write a Literature Review: From First Search to a Critical, Structured Chapter
I have a topic and a folder of PDFs, but I still do not know how to write a literature review from them. Where does it start? How do I search without drowning in papers? How many sources are enough, and when can I stop reading? What turns a review into critical analysis instead of a list of summaries? And how do I shape my notes into a chapter with a clear structure?
The short answer: plan a focused search, read selectively, record each source in a matrix, group the findings by theme, and write about the themes instead of about single papers. Open with a brief scope statement, compare and evaluate the evidence in the body, and finish by showing the gap your own study will address.
The steps below cover the narrative review that forms a dissertation or thesis chapter. The same method works for a standalone review assignment. Systematic and scoping reviews follow stricter protocols, so confirm with your supervisor which type your department expects. Handbooks differ between universities, and your own handbook overrides anything written here.
What a Dissertation Literature Review Has to Do
A literature review traces what is already known about your topic and shows where your own study fits. The Academic Writing Hub at Oxford Lifelong Learning describes it as a section that follows the development of a research area, so that your work has a foundation. It also stresses that description alone is not enough. A strong review adds value by grouping sources in fresh ways, connecting one study to another and weighing which evidence is stronger.
In practice, the chapter has two jobs. First, it places your investigation in context and justifies your approach. Second, it supplies evidence you return to when you explain your findings, and students tend to forget that second job. The University of Edinburgh’s academic skills team sums up the aim well: you are building “an argument, not a library”.
Where the review sits depends on your department. It usually follows the introduction and leads into the theoretical framework or methodology. Some fields fold it into the introduction instead, and others use a stand-alone chapter titled Background, Theoretical Framework or Research Context. Keep the background in your dissertation introduction short, because the review is where the critical comparison happens.
Which Type of Literature Review Does Your Dissertation Need?
Most dissertation chapters use a narrative, or traditional, review: a selective and critical account of current thinking. A systematic review is different. It is a research methodology with a standardised, repeatable search and a reporting structure set by your discipline. If your handbook asks for a systematic or scoping review, follow the protocol and reporting guideline your department names. The matrix and synthesis steps below still help, but the search and reporting rules become much stricter.
Literature Review or Annotated Bibliography? Spot the Difference Early
Many students hand in something halfway between the two. An annotated bibliography lists sources one by one, each with a short summary and evaluation, usually in alphabetical order. A literature review groups sources by theme or question and compares them, so one study can appear in several sections.
| Annotated bibliography | Literature review | |
|---|---|---|
| Organised by | Source, often alphabetically | Theme, debate or question |
| Treats each source | One at a time | In relation to other sources, sometimes more than once |
| Main job | Describe and assess a single source | Compare sources and build an argument |
| Ends with | The last entry | A summary of the evidence and the gap your study addresses |
Annotated notes are still useful, and many students write them first as raw material. The mistake is to turn that list into paragraphs and stop there. Without comparison and judgement, the result is a summary of studies, not a review of the evidence.
How to Write a Literature Review in Five Steps
The five steps run in order, although you will loop back. A search often reveals a new theme, and a draft often exposes a missing source.
Step 1: Plan Your Literature Search Before You Read Anything
Start with questions, not papers. What has already been researched on your topic? Which sub-areas matter? Which neighbouring research might also be relevant? Write your first thoughts as a list or spidergram, because the University of Reading’s study advice guide warns that unfocused reading wastes time. If the topic itself is still loose, narrow it first with our advice on choosing a research topic. A tight question also keeps the search short, and the guide on how to turn your idea into a research question shows how to build one.
Next, list the main concepts in your question and add synonyms and related terms for each. Add to the list whenever you meet a new term in your reading. The example below uses an invented topic: peer mentoring and first-year student retention.
| Concept | Search terms to combine with OR |
|---|---|
| Intervention | “peer mentor*”, “peer support”, “student mentoring” |
| Outcome | retention, persistence, attrition, dropout |
| Population | “first-year”, undergraduate*, “higher education” |
Join the concepts with AND, and the synonyms inside each concept with OR. For example: (“peer mentor*” OR “peer support”) AND (retention OR persistence OR attrition) AND (“first-year” OR undergraduate*). The asterisk is a truncation symbol, but symbols and quotation rules differ between databases, so check each database’s help before you rely on a string.
Search library databases first. Reading’s guide points out that a general web search returns thousands of results with few authoritative ones, although Google Scholar can help alongside the databases. Journal articles usually carry the most up-to-date research, so expect to cite them. Review articles are handy for spotting sources, but go back to the original studies and form your own analysis. Then follow the reference lists of your best sources, because works that keep appearing are probably standard texts in your field. Postgraduates may also need theses, conference papers and grey literature.
Finally, keep a simple search log. Record the database, the date, your search terms and the number of results. The University of Strathclyde’s dissertation support guide includes an example search-management spreadsheet and a synthesis matrix template. A log also lets you repeat or update a search later without starting again.
Step 2: Choose What to Read and Know When to Stop
You cannot read everything, and you do not need to include everything you read. Sheffield’s study skills team stresses quality over quantity: keep the sources that are most relevant, important or influential. Reading advises giving recent research priority while still acknowledging the standard texts. Older studies that started a field but have since been overtaken may need only a few sentences.
Screen each source with a few questions:
- What problem or question does the author address?
- How are the key concepts defined?
- Which theory, model or method does the study use?
- What are the results, and how far do they support the conclusions?
- Does the study confirm, extend or challenge other work?
- What are its strengths and weaknesses?
Credibility matters too. Prefer peer-reviewed journals, academic books, research reports and official publications over general websites.
Stopping is a judgement call, but Reading’s advice on when to stop reading gives three practical signals. First, you keep meeting the same viewpoints and little new information appears. Second, you can show you have covered both the major standard texts and the most recent research. Third, your planned word count for the section is nearly filled. Therefore, set a word budget for each section before you read deeply, and plan backwards from your deadline.
Postgraduate researchers face a slightly different test. Once the review has given you enough to justify your design and objectives, it may be time to move on to data collection, while still keeping a regular slot for reading. A dissertation timeline planner can help you protect that slot.
Step 3: Build a Literature Matrix, Then a Synthesis Matrix
A matrix turns a pile of notes into something you can compare. Sheffield recommends two spreadsheets. The first is a literature matrix, with one row per source and column titles chosen to suit your aims. The second is a synthesis matrix, which sorts sources by theme and adds a synthesis column where your own comparison and critical voice go.
Useful columns for the first matrix include:
- The question or aim of the study
- Sample or context
- Method and measures
- Main finding
- Limitation
- Theme it belongs to
- How it relates to your own study
Once several sources sit under one theme, ask four questions. Do the sources build on each other, perhaps over time? Do they contradict each other and reveal a debate? Do they point to an area or gap that deserves attention? Do they help complete a bigger picture? The answers become your synthesis column. The table below shows the idea with invented studies, so do not cite them.
| Source (invented) | Context and method | Finding | Limitation |
|---|---|---|---|
| Study A | One university, survey | Mentored students reported a stronger sense of belonging | Self-reported, no comparison group |
| Study B | Three universities, administrative records | Small retention difference after mentoring | Mentoring intensity not recorded |
| Study C | One university, interviews | Trained mentors felt more confident | Small sample |
Synthesis note: Evidence on belonging is positive but self-reported, and the retention effect looks small once records are used. No study compares mentor training or contact time, so both are possible gaps.
Step 4: Choose a Structure That Fits Your Question
Structure is where the matrix pays off, because the themes you spotted often suggest the headings. Several organising patterns exist, and you can combine them.
| Structure | Works best when | Watch out for |
|---|---|---|
| Thematic | You can group studies by concept or issue, which suits most subjects | Themes so broad that they hide the argument |
| Chronological | The order of events matters or a concept changed over time | A list of who found what, and when |
| Methodological | Approaches differ, such as qualitative against quantitative | Comparing methods while ignoring findings |
| Theoretical or conceptual | You need to compare models or definitions to build a framework | Theory with no empirical evidence |
| Combined | You want overall themes, each discussed in time order | Switching pattern midway without telling the reader |
Whichever you pick, move from general to specific. Reading describes an inverted triangle: briefly show the breadth of the field, narrow to studies that overlap with yours, then spend most time on the research closest to your question. Disciplines differ, too. Science reviews often concentrate on recent studies, while humanities reviews may take a longer historical view, so ask your supervisor what suits your subject.
Oxford offers a second way to see the same movement. Writers first set out settled knowledge, then show a gap or problem, and then describe recent research that tries to address it. The gap move is often short, sometimes only a paragraph or two, while the other two moves take more space.
A thematic outline for the peer mentoring example might run like this:
- Introduction and scope
- Why first-year students leave university (broad context)
- Transition and belonging in the first year (narrower)
- Peer mentoring models and how they differ (closer)
- Evidence on mentor training and contact time (closest)
- The gap and how the study responds
- Conclusion
Step 5: Write the Introduction, Body and Conclusion
Like an essay, the review has an introduction, a body and a conclusion. What each part contains depends on the job the review has to do in your dissertation.
The Introduction
The Royal Literary Fund’s structure advice says the introduction should define the topic and its context, explain why you are reviewing the literature, outline the order of sections and state the scope. The QUT cite|write guide adds that you can state your selection criteria, such as source types and date range.
A scope statement can be a single sentence. For example: “Research on university retention is broad. Because this study concerns first-year students who join peer mentoring, work on postgraduate persistence appears only where it clarifies a concept.” A sentence like that tells the reader what you left out and why.
The Body
Give each theme its own section or run of paragraphs. A reliable paragraph moves in four beats: a topic sentence that names the idea, two or three sources compared, your evaluation, and a closing sentence that ties the point back to your question. Transitions such as “whereas”, “similarly” and “in contrast” show the relationships you found in the matrix. Each section can then end with a short overview of the theme, which is where your synthesis column does its work.
The Conclusion
The conclusion is not a second introduction. Summarise the main agreements and disagreements, judge how strong the evidence is, name the gap or limitation, and connect it to your aims. If your study tests predictions, turning the gap into testable hypotheses is the natural next move. Likewise, an aims and objectives generator can draft objectives that match the gap you found.
Writing a Literature Review Critically: From Description to Synthesis
Feedback such as “too descriptive” or “summary, not analysis” usually means the text reports what sources say without judging or connecting them, as Southampton’s feedback decoder explains. Reading draws the same line. Descriptive writing lists who found what, whereas analytical writing weighs the studies, shows their limits and justifies your aims.
Here is the difference in practice, using invented placeholders. First, the descriptive version:
Author A (Year) surveyed students at one university and found higher belonging among mentored students. Author B (Year) used administrative records at three universities and found a small retention difference. Author C (Year) interviewed trained mentors.
Now the analytical version of the same material:
Mentoring appears to raise belonging, but the evidence is thinner for retention. Author A (Year) reports higher belonging in a single-university survey, whereas Author B (Year) finds only a small retention difference in records from three universities. Because Author A relies on self-report and Author B does not measure mentoring intensity, neither study explains why effects differ. Author C (Year) hints that training may matter, which is the question the present study takes up.
Notice what changed. The second version opens with a claim, compares sources, evaluates their methods and points to a gap. The first only reports.
To keep your own voice clear, ask what evidence convinced you and why. Edinburgh’s literature review top tips encourage you to show your stance, so a short evaluative sentence at the end of each theme earns its place. Also cite every paraphrase as well as every quotation, and keep one referencing style throughout, as San José State University’s Writing Center reminds students. Writing your notes in your own words from the start makes later paraphrasing a genuine rewrite instead of a reshuffle of the original wording.
How Long Should a Literature Review Be, and How Many Sources Do You Need?
No universal figure exists. The right length depends on your assessment brief, your level and your total word count, so your handbook and marking criteria decide. Online rules of thumb, such as a fixed share of the total word count, are only rough starting points unless your department confirms them.
The source count varies in the same way. QUT’s writing guide says the number can range from about five sources at first-year undergraduate level to more than fifty for a thesis, and it tells students to check with their lecturer. Therefore, ask your supervisor what range they expect, then judge sources by relevance instead of volume.
A practical method is to take your dissertation’s word limit, split it across chapters with a word count calculator, and give each theme in the review its own target. Reading recommends this kind of per-section target to avoid over-reading or over-writing one topic.
Common Literature Review Mistakes and How to Fix Them
Weak reviews fail in predictable ways. Each mistake below has a quick fix.
- Listing who did what. Group studies by theme and compare them inside each theme.
- Skipping quality checks. Add at least one evaluative comment on relevance or method for each source or cluster of sources.
- Having no logical thread. End each section with one sentence on what it means for your question, and state the gap outright.
- Hiding disagreement. Include conflicting findings and suggest why they differ, for example in sample, method or date.
- Waiting until reading feels finished. Draft one section after a few key texts, then fill the gaps you notice.
- Leaning on one review article. Use it to find the originals, then read and judge them yourself.
- Referencing slips. Provide a reference list of sources you cited, not a bibliography of everything you read, and make sure text and list match.
- Never revisiting the chapter. Return to it once your results arrive, as the next section explains.
For a research-based checklist, Justus Randolph’s 2009 guide to the dissertation literature review covers a taxonomy of reviews, the steps, common mistakes and a framework for self-evaluation.
Revisit Your Literature Review When You Reach the Discussion
The review does its second job after you have results. Your discussion should return to the literature and ask whether your findings agree with earlier work, and why they might differ, as Reading’s notes on developing the review explain. Reading compares the two chapters to an hourglass: the review narrows toward your study, then the discussion widens again to the wider field.
Because of this, expect to redraft. Studies that looked central in the first draft may matter less once your findings arrive, and new sources may deserve a place. Long projects make this obvious, since a PhD review is closer to a work in progress than a finished chapter. If you are shaping the discussion now, help with the discussion chapter can check that the two chapters speak to each other.
Literature Review Checklist Before You Submit
- The chapter shows why the topic matters.
- Sources are credible and include both key earlier work and recent research.
- The scope and selection criteria are stated.
- Sections follow a logical order, from general to specific.
- Inside each section, sources are compared instead of listed.
- The strengths and limits of the evidence are noted.
- Disagreements in the literature are acknowledged.
- A gap or niche is named and tied to your aims.
- Every citation, including paraphrases, appears in the reference list in one style, and the list matches the text.
- The conclusion points back to your research question.
Quick Answers to Common Literature Review Questions
Can I start writing before I finish reading?
Yes. Draft one section after you have read a few key texts, then see what is missing. Edinburgh’s guidance notes that the review is usually one of the first tasks after choosing a topic, and Reading notes that long projects redraft it as the work develops.
Do I have to include studies that disagree with my argument?
Yes. Showing controversy and conflicting evidence is part of the job, and a review that hides it looks one-sided. Explain why findings might differ, then say what that means for your question.
Can review articles replace reading the original studies?
No. Review articles help you find sources and see how a field is organised, but your analysis must rest on the original texts you have read yourself.
Is a systematic review the same as a literature review chapter?
No. A systematic review is a research methodology with a standardised search and reporting structure. A literature review chapter is usually a narrative, selective and critical account. Ask your supervisor which one your programme expects.
Can AI tools summarise papers for my literature review?
They can speed up skimming, but a summary is not analysis, and a tool cannot judge which studies matter for your question. Open and read every paper before you cite it, and confirm that each reference really exists. Universities set their own rules on AI use and change them often. Position correct as of October 2026; check your university’s current academic integrity policy before you use any tool.
Your Next Step
Open a spreadsheet today and enter the five most relevant sources you already have. Add the columns from Step 3 and look for a first theme. If you wrote a research plan earlier, its source list is a ready-made starting point, and our guide to the dissertation proposal shows how that early review fits.
Want a second pair of eyes on structure, feedback or a model review to learn from? Literature review support is available. Any model work is for reference and learning, and should be used in line with your institution’s policy.