Scoping vs Systematic vs Narrative Review: Which Do You Need?

Three Reviews, Three Different Questions

Think of a city you do not know yet. You could draw a map of every road and district, so anyone can see how the place is laid out and where nobody has built anything. You could measure one junction precisely, counting every car, so a planner can decide whether to install a traffic light. Or you could tell the story of how the neighbourhood came to be the way it is, and argue about what it should become next.

Those are your three reviews. Scoping draws the map. Systematic measures the junction. Narrative tells the story and takes a position. All three are legitimate research activity, and none of them is a lesser version of another. They fail only when you pick one and then write another: a narrative review that pretends to be exhaustive, or a scoping review that quietly starts recommending a treatment.

Notice that the difference is not effort or word count. It is the shape of the question. A question beginning “what research exists on” points to scoping. A question beginning “does X improve Y” points to systematic. A question beginning “how should we understand” points to narrative. If you have not settled your research question yet, settle it first, because it chooses the review type for you.

Side by Side

Timelines here are the usual ranges, not rules; a scoping review of a small field can be quicker, and a systematic review with a large evidence base can run longer. For the two-way version of this comparison in more detail, see literature review vs systematic review.

The Narrative Review

This is the review almost every dissertation actually contains. You read widely, decide what matters, group the work into themes, and argue your way to the point where your own study becomes necessary. Selection is your judgement, and you defend it in prose rather than in a flow diagram.

Its reputation suffers because a bad one is easy to spot: a string of summaries, one paragraph per paper, no position taken. But that is a failure of execution, not of the method. A strong narrative review is the hardest of the three to write well, because nothing external tells you what to include, and the whole value sits in the synthesis. Summary reports what each study said. Synthesis says what the studies say together, and where they contradict each other.

Choose it when your review has to set up an empirical study, when the literature is heterogeneous enough that pooling results would be meaningless, or when your question is conceptual. In practice, the structure is thematic: organising by theme is what separates a narrative review from an annotated reading list.

The Scoping Review

A scoping review asks what is out there. It uses systematic machinery, documented search strings across named databases, stated inclusion criteria, a record of what was screened, but it points that machinery at a broad question and usually skips formal quality appraisal. The output is a map: how much research exists, what kinds of study designs, which populations, which outcomes, and where the blank spaces are.

Why would you not appraise quality? Because the question is not “is the evidence good enough to act on”. It is “what has been done at all”. Excluding weak studies would hide part of the map. That is also why a scoping review cannot conclude that an intervention works. It can conclude that eleven studies exist, nine of them in one country, none of them longitudinal, which is often the more useful finding for a student.

Choose it when the field is new or scattered, when your supervisor wants a defensible case that a research gap genuinely exists, or when you need to know whether a full systematic review is even feasible. Scoping reviews are reported against PRISMA-ScR, the scoping extension of the PRISMA statement, and the widely used methodological guidance comes from JBI. Expect to include a search log and a screening flow diagram in your appendices. Getting your database searches right matters more here than anywhere else, so read up on finding academic sources before you begin.

The Systematic Review

A systematic review is a study in its own right, and the literature is its dataset. You write a protocol before you search, ideally registered, specifying the question, the databases, the search strings, the eligibility criteria, and how disagreements between reviewers will be resolved. Then you follow it. The test is reproducibility: another researcher running your method next month should arrive at the same set of included studies.

Every included study gets appraised with a risk of bias tool, and findings are synthesised in a structured way. Where the studies are similar enough in design, population, and outcome measure, that synthesis may include a meta-analysis, which pools the results statistically. Where they are not similar enough, you synthesise narratively and say why pooling was inappropriate. Reporting follows PRISMA 2020, and in health research theCochrane Handbook is the usual reference for methods.

Choose it when your question is narrow and answerable, when a body of comparable studies exists, when you have a second screener, and when your timeline is measured in months rather than weeks. Two of those four conditions are the ones students underestimate. Screening several thousand abstracts alone is how a promising dissertation becomes an overdue one, and if the protocol is your method, it belongs in your methodology section, not tucked into the literature chapter.

A Decision Path (Five Questions)

Answer these in order and stop at the first one that decides it.

1. Has your department specified a type or a guideline? If the handbook names PRISMA, a protocol, or two screeners, the decision is made for you. Check the marking rubric before anything else, because rubrics are more specific than supervisors.
2. Is the review the whole project, or the run-up to one? If you are collecting your own data afterwards, the review is the run-up, and a narrative review is almost always what is wanted. If the review is the entire contribution, you need scoping or systematic.
3. Can you state your question in one sentence with a measurable outcome? “Does peer mentoring reduce first-year dropout” is systematic-shaped. “How has peer mentoring been studied” is scoping-shaped. If neither fits, your question is conceptual, which is narrative-shaped.
4. Do enough comparable studies exist? Run one quick search. Under about fifteen studies, or wildly different designs and outcome measures, a systematic review has little to synthesise; map the field with a scoping review instead.
5. What do you actually have: time, and a second pair of eyes? A solo systematic review on a three-month timeline is not a brave choice, it is a scoping review that will be marked as a failed systematic review. Pick the method your resources support and defend that choice in writing.

One Topic, Written Three Ways

Take a single subject: mobile learning apps and language retention. Here is what each review type does with it, and how different the finished claim looks.

Read those three claims again. Only one of them is a statement about effectiveness, and it is the one that paid for the right to make it with a protocol, two screeners, and appraisal. The scoping claim is about the shape of the evidence. The narrative claim is an interpretation. Match your claim to the work you actually did, and the review will hold up in a viva.

What Your Supervisor Probably Means

“Do a systematic literature review” is used loosely, and it causes a lot of wasted months. Sometimes it means a full systematic review with a protocol. Far more often, especially outside the health sciences, it means “be systematic about how you search, and show me your method”, which is a narrative review with a documented search. Those two readings differ by roughly a year of work, so it is worth thirty seconds of clarifying.

Ask it as a specific question rather than an open one. Something like: “For this chapter, do you expect a registered protocol, two screeners, and formal quality appraisal, or a thematic review with a documented search strategy?” The answer settles the review type, the timeline, and the word count in one go. Send it by email so you have it in writing.

It is also worth checking length expectations at the same time, since they differ sharply by type and by degree level. Our guide on how long a literature review should be gives the usual proportions.

Five Ways This Goes Wrong

• Deciding the type after the reading. A search you did not document cannot be turned into a systematic one later. You would have to search again, properly, and log it.
• Borrowing systematic vocabulary for a narrative review. Adding a PRISMA diagram to a selective review does not make it systematic. It makes the mismatch obvious.
• Drawing conclusions your method cannot support. A scoping review that ends by recommending an intervention has skipped the appraisal step that would justify the recommendation.
• A systematic question that is far too broad. “Does technology improve learning” is not answerable. Narrow the population, the intervention, the outcome, and the timeframe until the question fits one sentence.
• Choosing prestige over fit. Systematic reviews are not marked higher for being systematic. They are marked on whether the method fits the question and was carried out properly. Several of the other common literature review mistakes start with this one.

Before You Read Another Paper

Write your question on one line. Underneath it, write the review type you think it implies, and one sentence on why. Then send the clarifying email to your supervisor. That is twenty minutes of work, and it protects the next three months of it.

Whichever type you land on, start the log now: a row per source with author, year, design, population, and finding. A narrative review turns that log into themes. A scoping review turns it into a chart. A systematic review turns it into an evidence table. Same discipline, three destinations. If you want to see a finished chapter first, our literature review example walks through one end to end.