What a literature review is
A literature review is not a list of summaries. It synthesises research — connecting, comparing and evaluating sources — to show what is known, how well it is known, and what is still missing. Reviews differ mainly in how sources are found, whether their quality is formally appraised and how results are combined.
Types compared
| Type | Search | Quality appraisal | Synthesis | Typical time |
|---|---|---|---|---|
| Narrative | Flexible, author-selected | Optional, informal | Thematic narrative | Weeks |
| Critical | Flexible, purposeful | Central — drives the argument | Argument / new interpretation | Weeks |
| Systematic | Comprehensive, pre-specified | Formal risk-of-bias tools | Narrative, tabular or statistical | Months |
| Scoping | Comprehensive, documented | Usually not | Map / chart of evidence | Months |
| Meta-analysis | As systematic review | Formal | Statistical pooling | Months |
| Rapid | Systematic but limited | Simplified | Narrative or tabular | Weeks |
Narrative (traditional) review
A readable overview of a topic, selected and summarised by the author.
What it is
A narrative review summarises and discusses what is known about a topic. The author decides which sources to include, and organises them into themes or a story of how the field has developed. It is the most common type in essays, dissertation chapters and textbook-style articles.
What makes it different
There is no fixed search protocol, so it is quick and flexible — but it is open to selection bias, and another author could reach a different picture. Its value depends on breadth of reading and clear, honest synthesis.
Use it when: To introduce a topic, give background for a study, or cover a broad area where a formal review is not needed.
Step by step
- 1
Define the topic and scope
Write one sentence saying what the review covers — and what it does not (time period, population, discipline). - 2
Search widely
Use your library database, Google Scholar and the reference lists of key papers. Keep a note of the search terms you used. - 3
Read and take structured notes
For each source record: aim, method, sample, key findings, limitations and how it relates to others. - 4
Group sources into themes
Organise by idea, debate or development over time — not one source per paragraph. - 5
Write thematically
Introduce each theme, compare what different sources say, and note agreements, disagreements and gaps. - 6
Conclude with the gap
Summarise what is known, what is uncertain, and what this means for your question or study.
Critical literature review
Goes beyond summary to evaluate the literature and build an argument.
What it is
A critical review examines the literature in depth to judge its quality, expose weaknesses or contradictions, and often to propose a new interpretation, model or direction. It is usually structured around an argument rather than a list of findings.
What makes it different
Like a narrative review, it does not require a systematic search. What sets it apart is evaluation and argument: every section asks “how good is this evidence and what does it mean?” — exactly the PEER approach from the critical writing guide. Its output is often a conceptual contribution, not just a summary.
Use it when: For most undergraduate and master’s assignments that ask you to ‘critically review’ or ‘critically discuss’ a literature.
Step by step
- 1
Frame a clear review question
Make it arguable: for example, “How convincing is the evidence that X influences Y?” - 2
Search and select key literature
Prioritise influential, recent and methodologically varied sources. Record why each was included. - 3
Appraise each source
Judge design, sample, measures, bias and how far the conclusions follow from the data. A checklist such as CASP can help. - 4
Identify patterns and tensions
Where do studies agree or conflict? Are the conflicts explained by methods, samples or definitions? - 5
Build your argument
Decide your overall position, then structure sections by issue, each using PEER points. - 6
Draw conclusions and implications
State what the literature can and cannot support, where it is weakest, and what research or practice should do next.
Systematic review
A transparent, reproducible search and appraisal of all relevant studies on a focused question.
What it is
A systematic review answers a tightly focused question by searching for all relevant studies using a pre-planned method, selecting them against explicit criteria, assessing their risk of bias, and synthesising the results.
What makes it different
Every decision is written down in advance and reported, so another team could repeat it. It normally involves two reviewers screening independently, a documented search of several databases, and a flow diagram showing how many studies were found and excluded. It is slow (often many months) but much less prone to bias.
Use it when: When you need the most reliable answer to a specific question, such as the effect of an intervention, often for practice or policy.
Step by step
- 1
Write a focused question
Use a framework: PICO (Population, Intervention, Comparison, Outcome) for interventions, or SPIDER for qualitative questions. - 2
Write and register a protocol
Set out the question, eligibility criteria, search strategy and methods before you start. Health-related reviews can be registered on PROSPERO. - 3
Build a comprehensive search
Combine keywords and subject headings with AND/OR across several databases (for example PsycINFO, MEDLINE, Web of Science). Consider grey literature and trial registers. A subject librarian can help. - 4
Screen in two stages
Titles and abstracts first, then full texts, against your criteria. Ideally two people screen independently and resolve disagreements. - 5
Extract data
Use a standard form: study details, participants, methods, outcomes and results. - 6
Assess risk of bias
Use a recognised tool suited to the study design — for example Cochrane RoB 2 for randomised trials, ROBINS-I for non-randomised studies, or CASP checklists. - 7
Synthesise
Combine results narratively, in tables, or statistically through meta-analysis if studies are similar enough. - 8
Report transparently
Follow the PRISMA 2020 reporting guideline, including the flow diagram, and discuss certainty of evidence and limitations.
Scoping review
Maps the size, range and nature of the literature on a broad question.
What it is
A scoping review uses a systematic search to map what research exists on a topic — what has been studied, how, and where the gaps are. It answers “what is out there?” rather than “what is the effect?”.
What makes it different
It is systematic in its searching but usually does not appraise quality or combine results statistically. Questions are broader, and the output is often a chart or table describing the evidence base.
Use it when: For new or diverse fields, to clarify concepts, or to check whether a full systematic review is feasible.
Step by step
- 1
Identify the research question
Often framed with PCC: Population, Concept, Context. - 2
Identify relevant studies
Search several databases and sources with a documented strategy. - 3
Select studies
Apply inclusion criteria; these may be refined as you learn about the field. - 4
Chart the data
Extract key information (aims, methods, populations, findings) into a charting table. - 5
Collate, summarise and report
Describe the extent and nature of the evidence and identify gaps. Report using PRISMA-ScR. - 6
Consult stakeholders (optional)
Ask practitioners or service users whether your map matches their experience.
Meta-analysis
Statistically combines results from comparable studies into one overall estimate.
What it is
A meta-analysis pools quantitative results (effect sizes) from several studies to estimate an overall effect and how much it varies between studies. It is usually carried out as part of a systematic review.
What makes it different
It is the only type that produces a pooled numerical answer with a confidence interval. It needs studies that are similar enough in design, measures and outcomes — otherwise combining them can mislead (“combining apples and oranges”).
Use it when: When a systematic review finds enough comparable quantitative studies to combine.
Step by step
- 1
Complete the systematic review steps
A meta-analysis is only as good as the search, selection and bias assessment behind it. - 2
Choose an effect size
For example, standardised mean difference (Cohen’s d / Hedges’ g), odds ratio or correlation, and calculate it for each study. - 3
Choose a model
A random-effects model assumes true effects differ between studies and is usually more realistic than a fixed-effect model in behavioural science. - 4
Pool and display results
Use dedicated software (for example R’s metafor package, or RevMan) and show a forest plot. - 5
Assess heterogeneity
Report statistics such as I² and tau², and explore sources using subgroup analysis or meta-regression if planned. - 6
Check for small-study and publication bias
Use funnel plots and related tests, recognising their limits with few studies. - 7
Rate certainty
Frameworks such as GRADE help you judge how confident readers can be in the pooled estimate.
Rapid review
A streamlined systematic review completed quickly by simplifying some steps.
What it is
A rapid review follows systematic review methods but deliberately cuts corners to deliver results in weeks rather than months — for example to inform an urgent decision.
What makes it different
Its shortcuts — fewer databases, date or language limits, a single screener — must be stated openly, because they increase the risk of missing studies or bias.
Use it when: When a decision-maker needs evidence quickly and a full systematic review is not possible in time.
Step by step
- 1
Agree the question and deadline
Clarify with the person who needs the answer exactly what decision it will inform. - 2
Decide and record your shortcuts
For example: limit to the last ten years, English language, two databases. - 3
Search and screen
Often one reviewer screens, with a second checking a sample. - 4
Extract and appraise efficiently
Use a short extraction form and a brief quality appraisal. - 5
Synthesise and report limitations
Present findings clearly and state how the shortcuts may affect confidence in the conclusions.
Tips for every review
- Keep a search log: databases, dates, search terms and number of results.
- Use reference management software (for example Zotero, Mendeley or EndNote) from day one.
- Organise by theme or issue, not by author.
- Prefer primary research and peer-reviewed sources; be clear when you cite a secondary summary.
- Evaluate as you go — note limitations while reading, not afterwards.
- Only cite sources you have read yourself, and check every reference is accurate.
- End with the gap: what the literature does not yet answer, and why it matters.