Writing a literature review that does more than summarise

Many literature reviews read like an annotated bibliography: Author A found this, Author B argued that, Author C studied something similar. Each paragraph is accurate, yet the reader finishes with no clear picture of what the field knows, where it disagrees, or why your study is needed. Examiners and reviewers notice this immediately.

A good literature review is an argument built from sources. It organises existing work around ideas, judges its strengths and limits, and leads the reader to the gap your research addresses. Here is how to get there.

Know which kind of review you are writing

The word “review” covers several different tasks, and it helps to be clear which one you are doing:

• A review chapter or section within a thesis or article. Its purpose is to set up your own study. It is selective and should end by pointing to your research question.
• A stand-alone narrative review. A published article that synthesises a body of work and offers an interpretation or framework.
• A systematic review. A review that follows a pre-specified, reproducible protocol for searching, screening and appraising studies. These have formal reporting standards (PRISMA is widely used), and are a method in their own right.
• Scoping and integrative reviews. Ways of mapping a broad or emerging area, often with less emphasis on appraising quality.

Most of the advice below applies to the first two types. If you are attempting a systematic review, follow the relevant guidelines from the start rather than retrofitting them.

Search with a plan

A review is only as good as the search behind it. Before you start reading:

1. Write down your question in one or two sentences. This decides what is in and out.
2. List search terms, including synonyms, regional terms and older terminology. Indian and international literature sometimes use different words for the same concept.
3. Use more than one database. Combine a general tool such as Google Scholar with subject databases and, where you have access, Scopus or Web of Science.
4. Follow citations both ways. Check the reference lists of key papers (backward) and see who has cited them since (forward).
5. Keep a search log. Note databases, dates and search strings. You will be glad of it when an examiner asks how you found your sources.

Be careful about where your sources come from. Papers in questionable outlets can carry weak or unreviewed claims. Our guide on how to spot a predatory journal is a useful filter.

Read actively and take structured notes

Summaries in your notes quickly become summaries in your draft. Instead, record the same few things for every source, ideally in a spreadsheet or reference manager:

• The question or claim.
• Method, data and context (country, period, sample).
• Main findings or argument.
• Strengths and limitations, in your own judgement.
• How it relates to other sources: agrees with, contradicts, extends.
• Possible themes or categories it fits under.

The last two columns are where your review starts to take shape. When you can sort your sources by theme, method or position, you have the bones of a structure.

Organise by ideas, not by authors

The single biggest improvement most reviews need is to change the organising unit from the source to the idea. Common structures include:

• Thematic: grouping work around key concepts, debates or variables.
• Methodological: comparing what different approaches have shown and missed.
• Chronological: tracing how thinking developed, useful when the history itself matters.
• Theoretical: setting out competing frameworks and their implications.

Whichever you choose, each section should make a point in its first sentence and use sources as evidence for it. Compare these two openings:

Weak: “Author A (2019) studied microfinance in one state. Author B (2021) studied microfinance in another state.”

Stronger: “Studies of microfinance in India broadly agree on its effect on household consumption but diverge on its effect on women’s decision-making, partly because they measure empowerment in different ways.”

(The examples are illustrative, not real studies.) The second version tells the reader what the literature says as a whole, and sets up a question.

Synthesise and evaluate

Synthesis means combining sources to say something none of them says alone. Useful moves include:

• Identifying patterns of agreement and explaining them.
• Identifying disagreements and suggesting why they exist (different methods, contexts, definitions).
• Pointing out what has been studied a lot and what has been neglected.
• Judging the quality of evidence, not just its quantity. Ten small studies with the same flaw do not add up to strong evidence.

Use reporting verbs that carry meaning. “Argues”, “demonstrates”, “assumes”, “challenges” and “overlooks” tell the reader what you think of a source; “says” and “states” do not.

Lead to the gap

A review within a thesis or article should end by making the case for your study. Pull the threads together: what is known, what is unresolved, and the specific gap your research fills. A gap is not just “no one has studied this in my district”. Explain why that absence matters, for theory, policy or practice.

A revision checklist

• Does each section open with a claim rather than a citation?
• Are most paragraphs drawing on several sources rather than one?
• Have you evaluated sources, not only described them?
• Have you included recent work as well as foundational texts?
• Are disagreements in the literature acknowledged and explained?
• Does the review lead clearly to your research question?
• Is every source cited accurately and paraphrased in your own words?

How CollabCite fits in

Reviews are often stronger when written with someone who knows a neighbouring literature, and review articles are a natural first project for new collaborators. CollabCite helps researchers find co-authors by field and research interest, with a verified member ID for each person. Browse our other guides or join as a founding member.