A literature review is not a queue of article summaries. It is an argument about what a body of research collectively shows, where it disagrees, how strong the evidence is and what remains unresolved.
The difficult part is usually the transition from reading to writing. You may have dozens of highlighted PDFs and pages of notes but no clear idea how to turn them into paragraphs. This guide shows how to move from individual sources to themes, synthesis and a defensible research gap.
If your notes are still scattered, start with the literature review matrix template. If you are spending too long on every paper, use the three-pass reading method first.
What a literature review needs to do
Monash University describes literature reviewing as summarising, analysing, evaluating and synthesising relevant research. Those are different jobs:
| Task | Question you are answering |
|---|---|
| Summarise | What did this source do and find? |
| Analyse | How was the argument or evidence constructed? |
| Evaluate | How convincing, useful or limited is it for this question? |
| Synthesise | How does it relate to the other sources? |
The review becomes useful when the final column—synthesis—drives the structure.
1. Define the question your review serves
You cannot decide which literature matters until you know what the review is trying to establish. Write a provisional review question or purpose statement. It can change, but it needs enough focus to guide inclusion.
For example, “AI in higher education” is a subject. “What does current evidence show about the effects of generative-AI feedback on undergraduate revision behaviour?” gives you a reason to include some papers and exclude others.
If your question is still sprawling, use the research-question narrowing guide.
2. Set boundaries before the literature sets them for you
Decide which types of evidence are relevant:
- date range, if justified;
- population or setting;
- discipline;
- study designs;
- languages you can assess;
- published and grey literature;
- outcomes or concepts;
- geographical scope.
A narrative literature review does not require the protocol of a systematic review, but your choices should still be explainable. “These were the first papers I found” is not a defensible boundary.
3. Read for comparison
As you read, capture the same core fields for each source. Useful columns include:
- research question;
- sample or material;
- context;
- method;
- key finding;
- limitation;
- theme;
- relationship to other studies;
- how you might use the source.
The goal is not to produce a database for its own sake. The columns should make meaningful comparisons possible.
4. Group the literature by ideas, not authors
A weak draft often follows the order in which papers were read:
Smith found X. Jones found Y. Patel found Z. Nguyen also studied X.
The reader has to do the synthesis themselves.
A stronger structure groups studies around questions such as:
- Where do findings agree?
- Which results conflict?
- What explanations compete?
- How have methods changed?
- Which populations or settings dominate?
- What does one research tradition overlook?
The Monash guide to the literature-review process describes synthesis as connecting and positioning sources against one another to identify themes, trends, agreement and disagreement.
5. Choose a structure that fits the field
There is no single correct literature-review structure. Common approaches include:
| Structure | Useful when |
|---|---|
| Thematic | The literature clusters around recurring concepts or problems. |
| Chronological | The development of ideas or methods over time matters. |
| Methodological | Different methods produce different kinds of evidence. |
| Theoretical | Competing frameworks interpret the same phenomenon differently. |
| Debate or controversy | The field contains clearly opposed positions or inconsistent findings. |
| Broad to narrow | You need to establish a wider field before focusing on your specific gap. |
Monash’s structuring guidance lists several of these patterns and stresses that the structure should follow the field and argument.
6. Build paragraphs around a claim
Before writing a paragraph, complete this sentence:
The point of this paragraph is that…
Then select the sources needed to establish, complicate or challenge that point.
A practical synthesis paragraph structure
- Claim: state the pattern or issue.
- Evidence: bring several relevant studies into conversation.
- Comparison: explain agreement, difference or methodological reason for variation.
- Evaluation: identify what the evidence can and cannot support.
- Implication: show what this means for your question.
Worked example: summary versus synthesis
The following studies are fictional.
Summary-heavy version:
Lopez (2024) surveyed 800 students and found frequent AI use. Chen (2025) interviewed 30 students and found that students used AI for feedback. Ahmed (2025) analysed platform logs and found high use before assessments.
Synthesis version:
Across self-report, interview and behavioural data, AI use appears to increase around assessment tasks. However, the studies measure “use” differently: Lopez counts reported frequency, Chen examines students’ stated purposes, while Ahmed records platform activity without establishing what students did with the output. Together, the studies support the prevalence of assessment-related use more strongly than they explain its effect on revision behaviour.
The second paragraph does more than compress three summaries. It tells the reader what the evidence collectively establishes and where the limit remains.
7. Evaluate evidence proportionately
Critical writing does not mean finding something negative to say about every paper. Evaluate what matters for the claim you are making.
Ask:
- Does the design answer the stated question?
- How was the sample selected?
- Are measures valid for the concept?
- Could alternative explanations account for the result?
- Is the context important for generalisation?
- Are conclusions stronger than the evidence?
- Do later studies replicate, qualify or contradict the finding?
A limitation only belongs in your review when it changes how the reader should interpret the evidence.
8. Identify the gap without manufacturing one
Your literature review should make the gap emerge from the comparison. Useful gaps include unresolved contradictions, weak evidence, a missing mechanism, poor external validity or an important context where assumptions have not been tested.
Use the companion guide on how to find a research gap before you write an extravagant sentence claiming that “no research exists.”
9. Write the introduction after the structure is clearer
Your introduction normally needs to establish:
- the scope of the review;
- why the topic matters;
- the question or purpose;
- important boundaries;
- how the review is organised.
You can draft it early, but revisit it after the body. Your final structure often changes what the introduction needs to promise.
10. Make the conclusion do intellectual work
A literature-review conclusion should not merely say “more research is needed.” It should summarise the strongest patterns, state where confidence is limited, identify the most important unresolved issue and show how your research question follows.
If the review supports a doctoral proposal, the next logical step is the PhD research proposal guide.
A literature-review drafting template
Review purpose/question:
Scope and boundaries:
Theme 1:
Main pattern:
Key supporting sources:
Contradictory evidence:
Methodological issue:
What this theme establishes:
Theme 2:
Main pattern:
Key supporting sources:
Contradictory evidence:
Methodological issue:
What this theme establishes:
Theme 3:
Main pattern:
Key supporting sources:
Contradictory evidence:
Methodological issue:
What this theme establishes:
Across themes:
What is well established:
What remains uncertain:
Most defensible research gap:
How my question follows:Final checklist
- The review has a clear purpose rather than a broad topic.
- Sources are grouped by ideas, evidence or debate—not simply by author.
- Most paragraphs contain more than one source where comparison is appropriate.
- Claims distinguish strong evidence from tentative evidence.
- Limitations are connected to interpretation.
- The structure makes the research gap understandable before you state it.
- The conclusion explains what follows from the review.
Once the draft exists, use the literature review matrix as a reverse check: if an important source appears in the matrix but has no role in the argument, decide whether the source or the draft needs to change. The aim is not to mention everything you read. It is to build the clearest evidence-based account of the question you are trying to answer.