Start with the argument, not the reading list

Most weak literature reviews read like a sequence of book reports: one source, then another, then another, each summarised in its own paragraph with a sentence at the end along the lines of "this is relevant to my project". Markers spot this immediately, because it describes rather than analyses. A strong review is an argument about what the field currently knows, where it disagrees, and why that matters for the question you are asking.

Before you draft a single sentence, answer three questions in your notes. What areas of research does my project sit across? What do researchers broadly agree on? And what is still unresolved, disputed or untested? If you cannot answer the third question, you are not ready to write the review. Your job is to build a case, not to prove that you have read a lot of material.

Group your reading into themes

The quickest way to move from a pile of notes to a review is to build a simple table. Give it columns for the author and year, the method and sample, the key finding, the main limitation, and the theme it belongs to. The theme column does the heavy lifting: once you have twenty entries, patterns appear, and sources that fit no theme are often sources you can leave out.

Aim for three to five themes in a dissertation-length review, and one to three in a shorter assignment. Each theme becomes a subsection, and each subsection should be doing a job in your overall argument. Common ways to organise themes include:

• By theoretical approach — for example, studies framing a problem as structural compared with those framing it as individual.
• By population or setting — school-age compared with adult learners, urban compared with rural, UK compared with international.
• By method — experimental, survey-based, qualitative interview and ethnographic work, and what each can and cannot show.
• By competing explanation — the rival accounts researchers give for the same observed pattern.

Compare findings rather than listing them

Once your sources are grouped, the writing becomes a matter of putting them into conversation. This is where marks are won. Instead of "Study A found X. Study B found Y", write something like: "While study A reported X in a small undergraduate sample, study B's larger survey found the opposite, suggesting that the effect may depend on prior experience."

Useful moves to keep in your toolkit include:

• Agreement: "This pattern is echoed by…", "Two further studies reach the same conclusion…"
• Contrast: "By contrast…", "This finding sits uneasily with…"
• Explanation: "The discrepancy is likely to reflect differences in how engagement was measured…"
• Evaluation: "The sample was small and drawn from a single institution, so the finding should be treated as indicative."
• Synthesis: "Taken together, these studies suggest that…"

Look closely at why studies disagree. Different definitions, different measures, different contexts and different samples all explain more than the blunt claim that "researchers disagree". Naming the reason is a higher-order analytical move, and it strengthens your own methodology chapter later.

Use your review to locate the gap

The gap is not an afterthought you bolt on at the end. It should emerge from the themes you have just discussed: a pattern that holds in one context but has not been tested in yours, a contradiction nobody has resolved, a group that has been overlooked, a method that has not been applied to this question.

Be honest about scale. Claims that "no research exists" are rarely true and are easy to disprove. Prefer careful, defensible phrasing: "little is known about…", "the evidence remains mixed…", "this has yet to be examined in the context of…". Then make the link explicit. One or two sentences should carry you from the gap to your own aim, so the reader can see why your project is the logical next step.

Build paragraphs that move the argument forward

A reliable paragraph shape for review writing is: a topic sentence that names the theme, evidence from two or three sources, a comparison or evaluation, and a closing sentence that connects back to your research question. If a paragraph contains only one source, ask whether it belongs with another.

Signpost clearly, and only where the relationship is genuine. Words such as "similarly", "however", "more recently" and "taken together" tell the reader how to hold two findings at once. Remember too that the review is not a standalone piece: material you introduce here should be picked up in your methods and discussion, where you explain how your design responds to what the field already knows.

Common traps, and a note on drafting

A few habits reliably weaken otherwise good work:

• Long block quotations that stand in for your own analysis.
• Uncritical praise, where every source is described as "important" or "valuable".
• Ignoring evidence that complicates your argument, which readers notice.
• Chronological listing for its own sake, rather than because an idea genuinely developed over time.
• A sudden gap at the end that has not been set up by the themes above it.

Finally, treat the review as a living document. Draft it early to clarify your thinking, then return to it once your data or analysis is under way, because you will read your sources differently by then. Two or three passes are normal. Each time you revise, ask one question of every paragraph: does this help justify my research question, or is it simply showing that I have read something? Cut what does not earn its place, and the review will do the work it is meant to do.