How to write a literature review

What a literature review is for

A PhD literature review has three jobs:

1. Show the examiner you know the field. The main theories, the key studies, the methods people use, and the current debates.
2. Evaluate that work. Which findings are well supported, which are weak, where studies conflict and why.
3. Arrive at your gap. By the last page, a reader should see for themselves what is missing, before you say it.

The third job is the one scholars most often miss. The review is an argument whose conclusion is your research problem. Every section should move the reader closer to it. If a section could be deleted without weakening that argument, it probably should be.

How the gap itself is framed and worded is a subject of its own, and our guide to writing a research gap statement covers it in detail.

Where to search for literature and how to keep track

Start with a few recent review articles on your topic. They map the field quickly and their reference lists save you weeks. Then search the main databases properly.

Build and record a search string

List your main concepts, then synonyms for each. Combine synonyms with OR and concepts with AND. For example: ("digital payment" OR UPI OR "mobile payment") AND (adoption OR "intention to use") AND (SME OR "small business" OR retailer). Run it in each database and note the date, the database, the exact string and the number of results. Your committee may ask how you found your sources, and a record makes the answer easy. For a formal systematic review, this record is mandatory.

Follow the citation trail

For each key paper, look backwards at what it cites and forwards at who has cited it since (Scopus, Web of Science and Google Scholar all show “cited by”). Two or three rounds of this usually find the core literature. When new searches keep returning papers you already have, you are close to saturation.

Filter for quality

Not everything indexed is reliable. Check that the journals you rely on are genuine; our guide to spotting predatory journals explains how. Prefer peer-reviewed articles over conference abstracts and theses for key claims, and be cautious with any paper whose methods section is too thin to evaluate. Our guide on how to critique a research article lists the questions to ask of each paper.

Use a reference manager from day one

A PhD review often ends up with 150 to 300 sources. Keeping them in a Word file by hand does not work. Zotero and Mendeley are both free; our Zotero vs Mendeley comparison helps you pick one. Store the PDF, your notes and a theme tag against each record.

How to read papers and take notes for a literature review

You do not need to read every paper in full. For most, read the abstract, then the figures and tables, then the conclusion. Read the whole paper only for the 30 to 50 studies closest to your question.

The single most useful tool at this stage is a synthesis matrix: a spreadsheet with one row per study and columns for the things you want to compare. Choose columns that match your research questions.

Once you have 40 or 50 rows, sort by the theme column and read down it. Patterns appear that you cannot see paper by paper: most studies are urban, most use cross-sectional surveys, results on cost are split by firm size. Those patterns are what your review is really about.

Write your notes in your own words from the start. Copying sentences into your notes is how unintentional plagiarism enters a thesis, because months later you will not remember which sentences were yours.

How to organise a literature review: themes, not authors

There are four common ways to organise a review. Most PhD theses use the first, sometimes combined with one of the others.

• Thematic. Sections by concept or variable: “trust and adoption”, “cost and adoption”, “policy interventions”. This is usually the best choice because it lines up with your research questions.
• Methodological. Grouped by approach: experimental studies, surveys, qualitative work. Useful when your contribution is partly a method.
• Chronological. How thinking developed over time. Useful for a short section on the history of a concept, rarely for a whole chapter.
• Theoretical. Grouped by framework, such as TAM, UTAUT or diffusion of innovations, then the studies that test each.

A structure that works for many theses is: a short introduction saying what the chapter covers and how you searched; a section on the theoretical framework; three to five thematic sections tied to your objectives; a section on the Indian or local context if your study is set there; and a final section that summarises the gaps and states your research problem. Many Indian universities title this chapter “Review of Literature”. The name matters less than the logic. For a full sample outline and annotated example passages, see literature review examples and format; for the framework section, see theoretical vs conceptual framework.

How to write synthesis instead of summary

Synthesis means your paragraph makes a point, and the studies are the evidence for it. Compare these two versions of the same material.

Summary: “Rao (2021) found that trust influences adoption among SMEs in Bengaluru. Otieno (2022) found that cost is more important than trust in Kenya. Singh (2023) showed adoption increased after a government subsidy.”

Synthesis: “Studies disagree on whether trust or cost drives adoption, and the split appears to follow firm size. Survey studies of established SMEs find trust to be the stronger predictor (Rao, 2021), while qualitative work with micro-enterprises points to cost (Otieno, 2022). The rise in adoption after India’s subsidy (Singh, 2023) supports the cost explanation, but that study measured no attitudes, so it cannot separate the two.”

The second version has a topic sentence that states a claim, it groups studies by what they found, it offers a reason for the disagreement, and it evaluates the evidence. That paragraph also points towards a gap. (The citations above are invented for illustration.)

Phrases that signal synthesis

“Consistent with”, “in contrast”, “both studies rely on”, “this finding has not been replicated in”, “a limitation shared by these studies is”. If most of your sentences start with an author’s name, rewrite them to start with the idea.

How long should it be?

In most PhD theses the review chapter runs to roughly 20 to 30 per cent of the main text, commonly 30 to 60 pages. That varies by field and university; a thesis in pure mathematics or some engineering areas may have a much shorter review, often inside the introduction. Look at recent theses from your department on Shodhganga to see local practice; our Shodhganga guide explains how to search it.

How to revise and update a literature review

Your review will be written early, usually in the first or second year, and read by examiners three or more years later. Plan to update it before submission. Rerun your saved searches for the years since, add what is relevant, and check that nothing published recently has already filled your gap. Examiners do check.

Before you call the chapter finished, check it against this list:

• Each section opens with a claim about the literature, not with the name of an author.
• Most paragraphs cite more than one study, and say how those studies relate to each other.
• You have stated where studies disagree and offered a reason for the disagreement.
• Key studies from the last three years are included, and so are the older studies everyone in your field cites.
• Every source is one you have read, at least the abstract, methods and results, rather than a citation lifted from someone else’s review.
• The review ends by stating the gap, and the gap leads straight to your objectives.
• Nothing in the review is irrelevant to your research questions, however interesting it is.

For a specialist in your field to read your draft chapter and comment on its organisation and synthesis, our literature review support covers exactly that. We work on your draft and your reading; the review remains your work.