These four are often treated as a hierarchy of rigour, with narrative review at the bottom. They are better understood as different instruments for different questions.
Narrative literature review
Answers: what is currently understood about this topic?
Method: expert selection of relevant work, synthesised into an argument. No pre-specified search, no formal appraisal.
Use when: introducing a field, framing a thesis chapter, or writing an invited overview. Its strength is breadth and interpretation; its weakness is that selection cannot be checked.
Typical time: weeks.
Scoping review
Answers: what kind of evidence exists here, and how much?
Method: systematic search, but charting the literature rather than appraising quality or synthesising effect. Reported using PRISMA-ScR.
Use when: the field is unmapped, the question is broad, or you are deciding whether a systematic review is even feasible. Frequently the correct first step.
Typical time: three to six months.
Systematic review
Answers: what does the totality of evidence say about this specific question?
Method: registered protocol, exhaustive reproducible search, duplicate screening, formal risk-of-bias appraisal, structured synthesis. Reported using PRISMA 2020.
Use when: the question is narrow and answerable, and enough primary studies exist. See how to write one.
Typical time: six to twelve months with a team.
Meta-analysis
Answers: what is the pooled magnitude of this effect?
Method: a statistical technique applied within a systematic review, pooling comparable outcomes into a weighted estimate. See the difference between the two and how to conduct one.
Use when: included studies measure the same outcome comparably. When they do not, a narrative synthesis is the correct choice, not a failure.
Choosing
- Is your question narrow and specific? If not, scoping review.
- Do enough comparable primary studies exist? If unknown, scoping review first.
- Do you need to appraise quality, or only map the field? Appraisal means systematic.
- Do the included studies report the same outcome? If yes, add a meta-analysis.
- Do you have a team and six months? If not, be honest about which is achievable.