Systematic, Scoping and Narrative Reviews: Choosing the Right Approach

Define what the review must achieve

A literature review can establish the state of knowledge, assess evidence on a focused question, map a broad field or develop an interpretive argument. These purposes require different approaches. Choose the review type before collecting a large folder of articles, then explain why it fits the intended contribution.

Begin with the reader's need. A programme leader deciding whether to adopt an intervention may need a focused synthesis of comparative evidence. A researcher exploring how a concept is defined across disciplines may need a broader map. A theoretical paper may need a carefully argued narrative synthesis that connects ideas rather than estimates an effect.

Understand the three approaches

A systematic review uses explicit methods to identify, select, assess and synthesise evidence for a defined question. The exact methods depend on the evidence and discipline; systematic does not automatically mean statistical pooling. PRISMA 2020 supplies reporting guidance and supporting tools for systematic reviews. ^[1]

A scoping review maps the extent, characteristics or concepts within a body of literature. PRISMA's scoping extension provides a reporting framework suited to this purpose. ^[2] A narrative review develops an organised account of a topic or argument, often with more flexible selection and interpretation. That flexibility still requires clarity about scope, sources and the basis of the synthesis.

Match the scope to available resources

Estimate the workload for searching, screening, obtaining full texts, extracting information, assessing studies and synthesising findings. Consider language expertise, database access and the need for more than one reviewer. A broad question may generate far more material than a student or small team can handle rigorously.

If resources are limited, narrow the question or state the methodological compromises. Do not label a convenient selection of papers as a systematic review. A transparent, bounded review can be useful, but readers need to know which evidence may have been missed and how that affects the conclusions. A review's label should describe its methods rather than serve as a quality badge.

Make decisions before screening

Prepare eligibility criteria covering relevant populations, concepts, settings, study types, dates and languages. Explain the reasons for boundaries. Build and test search terms with support from an information specialist where possible. Record the databases, search dates and exact strategies so the process can be understood and updated.

Create an extraction form that captures information needed for the question. Pilot it on several sources and resolve ambiguous fields. Decide how disagreements will be handled. Keep reasons for exclusion at the stage where they are meaningful. Any later changes to the protocol should be dated and justified rather than hidden in the final methods section.

Illustrative example: digital feedback in higher education

Imagine a team wants to understand digital feedback. If it asks whether a specific feedback intervention improves a defined writing outcome compared with usual practice, a systematic review may fit. If it asks what forms of digital feedback have been studied, in which disciplines and with which outcomes, a scoping review may be more appropriate.

If the purpose is to examine competing ideas about feedback and learner agency, a narrative review may offer a useful conceptual argument. These are different projects, not three interchangeable labels for the same search. The team should select the purpose first and avoid making an effectiveness claim from a review designed only to map the field.

Synthesis requires more than counting studies

Compare the quality, relevance, design and context of evidence. A list of positive and negative findings can be misleading when studies differ substantially. Explain whether differences arise from participants, intervention delivery, measurement or analytic choices. When pooling results, justify whether studies are sufficiently comparable and obtain appropriate statistical expertise.

For a narrative synthesis, show how themes or explanations were developed. For a scoping review, distinguish an observed concentration of publications from the importance of the underlying issue. For any review, identify uncertainty and avoid equating frequent publication with strong evidence.

Use a completion checklist

• The review question and chosen approach are consistent.
• Search sources, dates and boundaries are documented.
• Eligibility and screening decisions are traceable.
• Extraction fields support the intended synthesis.
• Study limitations influence interpretation where relevant.
• Conclusions remain within the review's purpose and coverage.
• Reporting guidance is used appropriately and acknowledged.

Finally, plan for updating. A review is a dated account of available evidence, not a permanent verdict. State when searching ended and identify developments that might require reconsideration. Readers can then judge whether the review remains suitable for their decision.