You’ve got forty saved PDFs, a folder of half-read journal articles, and a document titled ‘Literature Review’ that’s been blank for longer than you’d like to admit. Turning a pile of sources into a coherent chapter is the part almost nobody explains properly, most guides describe what a literature review should achieve without showing the actual mechanics of getting there.
This guide covers how to write a literature review that argues rather than lists: what the chapter is actually for, how to choose between a thematic, chronological or methodological structure, a synthesis matrix technique for organising your sources before you draft a single paragraph, how to write about disagreement between sources without sounding unsure of your own position, how to identify and state the gap your own research fills, and realistic length expectations at undergraduate, Masters and PhD level.
How to Write a Literature Review: What It’s For (and Three Things It Isn’t)
A literature review positions your own research within an existing body of work. It shows an examiner what’s already established, where credible sources genuinely disagree, and precisely where your own study fits into that landscape, rather than existing as a separate, self-contained chapter of reading notes.
It’s worth being just as clear about what a literature review isn’t, since these three substitutes are the most common way a review under-performs. It isn’t a summary of sources taken one at a time, since that reports what each source says without ever putting sources in conversation with each other. It isn’t an annotated bibliography, since that evaluates each source individually rather than showing how sources relate to one another across the whole body of work.
And it isn’t a background section, since background describes context, while a review has to critically engage with the literature itself, weighing claims against each other rather than simply setting the scene.
Where it sits in a dissertation matters too, since that placement shapes what the chapter needs to do. It usually comes after the introduction and before the methodology, which means it has to justify the research question the introduction has just raised and set up the methodological choices the next chapter is about to defend, rather than functioning as a standalone piece of reading about the topic in general.
It’s also worth distinguishing this kind of narrative literature review from a systematic review, since the two get confused. A systematic review follows a pre-registered, exhaustive search protocol and is itself a research method, most common in health and clinical fields.
What this guide covers, the kind almost every undergraduate, Masters and PhD dissertation includes, is a narrative review, built around your own judgement of which sources matter and how they relate, rather than a fixed, replicable search procedure. If your specific assignment asks for a systematic review, check your module handbook, since the two follow genuinely different rules.
There’s a genuinely useful test for telling the difference. Could your literature review be reordered into an alphabetical list of the sources it covers, with nothing lost? If yes, it’s a summary rather than a review, because the argument in a genuine review runs across sources thematically, not through them one after another in whatever order you happened to read them.
University of Kent’s guide to writing a literature review is a useful companion to this section, covering the preparing and assessing stage that comes before you start structuring anything.
Thematic, Chronological or Methodological: Choosing Your Structure
Three structures cover almost every literature review, and choosing between them is a decision worth making deliberately rather than defaulting to whichever one you saw in a previous student’s dissertation.
Thematic organises the review by idea, concept or sub-question rather than by individual source. This is the most common structure for a standard dissertation-length review, and it’s the best fit when the literature naturally clusters around a small number of recurring debates or themes rather than a single narrative arc.
Chronological organises the review by how thinking on the topic developed over time. It’s the best fit when the field has gone through a clear shift, a landmark study, a change in dominant method, or a genuine paradigm change, and that development is itself part of what you’re arguing, not just background colour.
Methodological organises the review by how the underlying studies were conducted, qualitative against quantitative, lab-based against field-based, and so on. It’s the best fit when method is the actual variable your own research addresses, for example if your own study compares findings that used different measurement approaches to the same underlying question.
The practical decision rule is to look at your own research question first, then ask which structure would let you make that specific argument most directly, rather than picking a structure early and forcing the literature to fit it afterwards. In practice, most strong reviews mix approaches: thematic overall, with a chronological or methodological thread running within individual themes, rather than one structure applied with no exceptions from the first paragraph to the last.
Newcastle University’s guide to structuring a literature review covers a fourth option worth knowing, a funnel structure that moves from broad context down to the specific studies closest to your own question. It’s a useful variant on thematic organisation for a review that needs to establish a wide field before narrowing in, rather than a genuinely separate fourth category.
A concrete example makes the choice easier to see in practice. A review of remote working and productivity, where the literature clusters into distinct debates around measurement, role type and management style, suits a thematic structure built around those three clusters. A review of a field that has visibly shifted its dominant explanation over twenty years, attention-deficit research moving from a purely behavioural framing towards a neurodevelopmental one, for instance, suits a chronological structure that lets the shift itself carry part of the argument.
Building a Synthesis Matrix Before You Write a Single Paragraph
A synthesis matrix is a simple table, one row per source and one column per theme, recurring finding or key variable, filled in before you write a single paragraph of the actual review. It sounds like extra work before the writing starts, and it is, but it’s the single step that most reliably prevents a review reading as a list of summaries.
Building one in practice starts with listing your themes across the top of the table, based on the recurring ideas you’ve noticed across your reading so far. Then list your sources down the side, and for each source, fill in each relevant cell with a short note on what that specific source says about that specific theme, not a restatement of the source as a whole.
Once the table is filled in, the real value shows up when you read it column by column rather than row by row. Reading horizontally across a single theme column shows you at a glance which sources agree with each other, which disagree, and which don’t address that theme at all, which is exactly the synthesis a strong literature review needs and a source-by-source summary structurally can’t produce.
This is genuinely the step most students skip, going straight from reading to drafting because the matrix feels like an extra task rather than part of the actual writing. It’s usually the reason a first draft reads as a list of individually accurate summaries rather than an argument that moves across sources, so treat the hour spent building it as writing time, not preparation time.
A short worked example makes this concrete. Say your themes are ‘measurement method’, ‘sample population’ and ‘reported effect size’, and one of your rows is a 2019 study. The cells for that row might read ‘self-reported survey’ under measurement method, ‘undergraduate students only’ under sample population, and ‘small but statistically significant’ under reported effect size.
Reading down the reported-effect-size column afterwards, across every row, shows you immediately whether studies using self-reported measures tend to report smaller effects than studies using observed measures. That’s precisely the kind of pattern a paragraph-by-paragraph summary would leave you to spot by memory alone, several weeks and dozens of sources later, rather than seeing it laid out in a single column.
University of Westminster’s guide to literature review structure covers the same synthesis-over-summary principle from the writing side, worth reading once your matrix is built and you’re moving into drafting.
Writing About Disagreement in the Literature Without Sounding Unsure
Disagreement between sources is often treated as an inconvenience to smooth over, when it’s usually the opposite: a field where credible sources genuinely conflict is exactly where your own research has the clearest gap to address.
The practical structure for handling it is straightforward. Name both positions clearly, state the strongest evidence behind each one, then either take a position of your own with stated reasons, or explicitly note that the question remains genuinely open in the field, rather than quietly reporting one side’s summary and leaving the other out.
A few phrases hold two conflicting sources in the same sentence without hedging awkwardly. ‘While Smith found X, Jones’s findings suggest otherwise, largely because the two studies measured Y differently’ is one useful pattern. ‘This disagreement may reflect a methodological difference between the two studies rather than a genuine contradiction in the underlying phenomenon’ is another.
There are two common failures worth naming directly. The first is presenting only the source that supports your own eventual argument and quietly leaving out the one that complicates it. The second is presenting both sides fairly but never stating which is better supported, which reads as balanced description rather than the critical engagement a literature review is actually there to demonstrate.
Taking a position doesn’t mean declaring one study simply wrong. It can mean something narrower and more defensible, for example noting that a larger, more recent study used a more representative sample than the smaller study it contradicts, so its finding should carry more weight without dismissing the earlier study’s contribution entirely. That kind of qualified judgement, weighing rather than picking a winner outright, is usually closer to what the evidence actually supports and reads as considerably more credible to an examiner than an absolute claim either way.
Identifying and Articulating the Gap Your Work Fills
A genuine gap usually shows up as one of a handful of recognisable signals: a specific question nobody has directly tested, an established finding nobody has applied to a particular population or context, a method nobody has used on this specific question, or a disagreement in the field nobody has actually tried to resolve.
The sentence structure that states a gap defensibly follows a clear pattern: name what’s already established, name what’s specifically missing from that picture, then state how your research addresses that particular absence. This is considerably stronger than the vaguer ‘more research is needed in this area’, which tells an examiner nothing about what your own study specifically contributes.
It’s worth being careful not to overclaim here. A gap should be scoped to what your own study can actually address, rather than framed as though an entire field has somehow been overlooked when a handful of closely adjacent studies already exist and simply haven’t answered your exact question.
Put concretely, a defensible gap statement might read: ‘existing studies have established that remote working affects reported productivity for individual-task roles, but none have examined whether this effect holds for roles requiring frequent, unscheduled collaboration, which this study addresses directly.’ Notice what that sentence does: it credits the existing literature accurately, names a specific, bounded absence rather than a general one, and states exactly how the study in front of the examiner closes it.
This section is the direct bridge between the literature review and the rest of your dissertation, worth getting genuinely specific about rather than generic, since a vague gap statement is one of the more common criticisms examiners raise at Masters and PhD level.
How Long Should a Literature Review Be? UG, Masters and PhD Expectations
Length expectations vary meaningfully by level, and treating a single fixed word count as universal is one of the more common ways students either under-develop or badly over-write this chapter.
| Level | Typical literature review length | Typical total dissertation length |
|---|---|---|
| Undergraduate | 2,000–4,000 words | 8,000–12,000 words |
| Masters | 4,000–6,000 words | 15,000–20,000 words |
| PhD | 8,000–12,000 words or more, sometimes split across more than one chapter | 60,000–100,000 words |
These figures are realistic starting expectations rather than fixed rules. They vary by department, discipline and the volume of directly relevant literature that actually exists for your specific topic, so confirm the expected range against your own handbook and supervisor before treating any number here as a target to hit.
The practical point underneath these figures matters more than the numbers themselves. Length should track how much genuinely relevant literature exists and how much synthesis it needs, not a fixed proportion applied mechanically regardless of topic. A narrow, well-defined topic with a small but tightly focused body of literature can produce a strong review at the lower end of its range, while a broad or contested topic may genuinely need the upper end to do the synthesis justice.
Some PhD programmes, particularly in the sciences, now use an article-format thesis, where the literature is reviewed more briefly in an introductory chapter and then again, more narrowly, at the start of each published-paper chapter, rather than in one long standalone review. If your programme uses this model, the length expectations above apply to the introductory chapter’s review rather than to a single chapter carrying the full literature on its own, so check with your supervisor which format your department expects before assuming the traditional single-chapter model applies.
Before You Submit
The practical sequence, in order, is worth more than any single tip in this guide: build the synthesis matrix before drafting, choose the structure that fits your actual argument rather than forcing the literature into a default shape, then write disagreement and your own gap explicitly rather than leaving either implied.
Consistent referencing matters more in this chapter than almost anywhere else in a dissertation, simply because of how many sources a literature review draws on in one place. Our Harvard referencing guide is worth a pass against your reference list once the chapter is otherwise finished.
Most of what separates a strong literature review from a weak one isn’t the reading, it’s whether that reading gets synthesised on the page rather than simply reported. If you’ve built the matrix and you’re still not sure whether a section synthesises or just summarises, a second, trained read specifically checking for that distinction usually catches what’s genuinely hard to see in your own writing, particularly after weeks spent inside the same sources.
None of this requires reading more sources than you already have. It just requires putting the sources you’ve already read in deliberate conversation with each other on the page, rather than letting them sit next to one another in the order you happened to find them.