The distinction is regularly explained as a matter of rigour, as though a systematic review were a literature review done properly. That is not the line. The line is reproducibility: a systematic review is built so that somebody else, following what you wrote down, would select the same studies. A narrative review is a reasoned reading by a person with judgement, and it can be excellent without being reproducible. This page is about where that line falls and what it costs on each side.
The systematic review definition people quote, and what it commits you to
A systematic review answers a defined question using explicit, pre-specified methods to identify, select and appraise studies, and reports them well enough to be repeated. Each word in that is a commitment: a protocol written before you start, a search you can reproduce, criteria applied by more than one person, and counts at every stage. It is a different amount of work from a narrative review by roughly an order of magnitude, and it is not better for every question.
What a systematic literature review has to report
The databases and the exact searches with dates, the number of records each returned, the duplicates removed, the records screened and excluded at each stage with reasons, and the studies finally included. Those counts are what a flow diagram shows and they are not reconstructable after the fact, which is the single strongest argument for keeping the record as you go rather than at write-up. The screening funnel worksheet on this site works those numbers from your own figures.
When a narrative review is the right choice
When the question is broad, when the literature is heterogeneous enough that pooling would be meaningless, when you are mapping a field rather than answering a defined question, or when the point is to argue a position about how the field has developed. None of those is served by a protocol, and forcing a systematic shape onto them produces a worse piece of work with a better-looking method section.
The middle ground everyone actually occupies
Most thesis chapters are narrative reviews with some systematic habits: a documented search, stated inclusion criteria, a record of what was excluded. That is a good place to be, and it is worth being explicit about it rather than implying a rigour you did not apply. A paragraph saying what you searched and what you left out is honest and improves the chapter; calling it systematic when two people did not screen it is not.
Two reviewers, and why it matters more than it sounds
The requirement that decisions are made independently by two people is doing real work: it catches the drift a single reader develops over hundreds of abstracts, and it produces a disagreement rate you can report. If you cannot get a second screener for everything, doing a sample twice and reporting the agreement is far better than nothing, and it is the single cheapest way to move a review toward defensible.
What it costs to change your mind halfway
Deciding at month four that the narrative chapter should have been systematic means going back for the counts nobody kept: how many records each search returned, what was excluded and why. Those are not reconstructable, so in practice the choice is made once, at the start, and the honest way to keep the option open is to keep the record from day one even for a narrative review. Recording a search takes a minute; recovering it takes a re-run and an asterisk.
Questions people ask about systematic review vs literature review
Is a systematic review just a thorough literature review?
No. The difference is reproducibility rather than effort: a systematic review is written so somebody else following your method would select the same studies. A narrative review can be thorough, expert and excellent without being reproducible.
Can I call my chapter a systematic literature review?
Only if it was one: a protocol, a reproducible search, stated criteria, ideally two screeners, and reported counts at every stage. Claiming the label without the method is the thing examiners most reliably notice.
Which takes longer?
A systematic review, substantially, and the screening is where the time goes rather than the writing. The funnel worksheet on this site turns your own record count and screening rates into hours, which is worth doing before committing to the method.
What is a systematic review definition I can cite?
Use the method text your field uses rather than a dictionary line; in health and adjacent fields the Cochrane Handbook is the standard reference, and the PRISMA statement defines what has to be reported.
Do I need PRISMA for a narrative review?
Not formally, but borrowing the flow-diagram discipline costs little and makes any review more defensible. Recording what you searched and what you excluded is useful even when nobody will ask for it.
Can I convert a narrative review into a systematic one later?
Rarely, because the counts nobody kept cannot be reconstructed. If there is any chance the chapter becomes a publication, keep the search and exclusion record from the first day; it costs a minute per search and it is the whole difference.