A good literature review is arranged by theme, and it builds one argument that ends at your research gap. Still, examiners and research committees keep finding the same problem. Pages go like “Kumar (2019) studied X and found Y. Sharma (2020) studied Z and found W.” Writing like this is only an annotated bibliography. It proves that you have read. What it does not prove is that you understand what all that reading adds up to.
The method is simple enough. Search properly, sort what you find, and write a review that argues its way to your research gap. It is meant for the review chapter of a PhD thesis, which is mostly the longest chapter and the one you revise most. The same method works for the literature section of a paper or a synopsis also. Last reviewed September 2026.
What a literature review is for
There are three jobs a PhD literature review has to do.
- Show the examiner you know the field. That means the main theories and the key studies, the methods people are using, and the debates going on now.
- Evaluate that work. Some findings have good support and some are weak. Which ones? When studies conflict, why is that happening?
- Arrive at your gap. On the last page, the reader should already see the missing piece before you write it down.
Scholars miss the third job most of the time. Think of the review as an argument whose conclusion is your research problem. Each section has to take the reader a little closer to that problem. If you can remove a section and the argument stays just as strong, probably that section should go.
How to frame the gap and which words to use, that is a separate topic. Our guide to writing a research gap statement covers it.
Where to search for literature and how to keep track
First pick up a few recent review articles on your topic. They give you a quick map of the field, and going through their reference lists will save you weeks. After that, do a proper search in the main databases.
| Source | Best for | Watch out for |
|---|---|---|
| Scopus | Wide coverage in all disciplines, tracking citations, exporting many records at one time | You get access only if your institution has a subscription |
| Web of Science | Core journals that are highly cited, and following citation chains backward and forward | In many fields the coverage is narrower than Scopus, and access is through your library |
| Google Scholar | Locating one particular paper, free full-text copies, grey literature | Predatory and unreviewed material also comes up, and systematic searching is difficult |
| Subject databases | PubMed (medicine), IEEE Xplore (engineering), ERIC (education), EconLit (economics) and so on | Search syntax is different in each one |
| Shodhganga | Indian PhD theses on related topics, with data and questionnaires from the Indian context | A thesis is not a peer-reviewed publication, so cite it and give it weight keeping that in mind |
| e-ShodhSindhu and your library | Full text of subscribed journals through the INFLIBNET consortium, if your university is a member | How remote access is set up changes from university to university |
Build and record a search string
Write down your main concepts and then the synonyms for each one. Synonyms are joined with OR, and concepts are joined with AND. One example is ("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 how many results came. Your committee may ask how you found your sources, and with this record answering is simple. In a formal systematic review, keeping this record is compulsory.
Follow the citation trail
For each key paper, look back at what it cites, then forward at who cited it later. You can see “cited by” in Scopus, Web of Science and Google Scholar. Usually two or three rounds of this bring up core literature. When fresh searches keep giving you papers you already have, you are near saturation.
Filter for quality
Being indexed does not make a paper trustworthy. Check whether the journals you are depending on are genuine, and our guide to spotting predatory journals has the steps. Your key claims should stand mainly on peer-reviewed articles, and a conference abstract or a thesis is only extra support. Be careful also with papers where the methods section is so thin that you cannot judge the work. The questions to ask of each paper are listed in our guide on how to critique a research article.
Use a reference manager from day one
Many times a PhD review ends with 150 to 300 sources. Managing that many by hand in a Word file will not work. Zotero and Mendeley are free, both of them, and our Zotero vs Mendeley comparison will help you decide. Against every record, keep the PDF, your notes and a theme tag.
How to read papers and take notes for a literature review
You need not read every paper fully. For most, read the abstract, then figures and tables, then the conclusion. Keep full reading for the 30 to 50 studies that are closest to your question.
Now the most useful tool is a synthesis matrix. It is a spreadsheet with one row per study and columns for what you want to compare. Pick your columns as per your research questions.
| Study | Context and sample | Method | Main finding | Limitation | Theme |
|---|---|---|---|---|---|
| Author A (2021) | Urban Karnataka, 300 SMEs | Survey, SEM | Trust predicts adoption | Cross-sectional; urban only | Trust |
| Author B (2022) | Kenya, 45 interviews | Qualitative | Cost matters more than trust for micro-firms | Small, non-Indian sample | Cost |
| Author C (2023) | Five Indian states, panel data | Fixed-effects regression | Adoption rose after subsidy | No firm-level attitudes measured | Policy |
Once you have entered 40 or 50 papers, sort the sheet by theme. Read every trust paper together, then every cost paper, and so on. Now patterns start showing which you will never notice paper by paper. Maybe most of the researches are done in urban areas. Maybe most use cross-sectional surveys. Maybe the results on cost split by firm size. Your review is really about these patterns.
From the very first day, write your notes in your own words. A sentence copied into notes today can go into the thesis as it is, and that is plagiarism. After some months you will not remember which sentences were your own.
How to organise a literature review: themes, not authors
A review is commonly organised in one of four ways. The first way is used in most PhD theses, and sometimes it is mixed with one of the other three.
- Thematic. Each section covers one concept or variable, for example “trust and adoption”, “cost and adoption” or “policy interventions”. Mostly this is the right choice, since it follows your research questions.
- Methodological. Here studies are grouped by approach, like experimental studies, surveys and qualitative work. This is useful if part of your contribution is a method.
- Chronological. Here you trace how the thinking on the topic moved over the years. For a short section on a concept’s history it is fine. For a whole chapter it is rarely suitable.
- Theoretical. Grouping is by framework, such as TAM, UTAUT or diffusion of innovations, and under each framework come the studies that test it.
Many theses do well with the following structure. It opens with a short introduction on what the chapter covers and how you searched. Next, the chapter should discuss about the theoretical framework in its own section. After that come three to five thematic sections linked to your objectives. A study set in India or a local area needs a separate context section. The final section brings the gaps together and states your research problem. In many Indian universities this chapter is titled “Review of Literature”. For examiners the name is a small matter, they get bothered more when sections don’t lead into each other. For a complete 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
In synthesis, your paragraph makes a point and the studies serve as evidence for that point. See the two versions below. Both use 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 do not agree on whether trust or cost drives adoption, and the split seems 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.”
In the second version, the topic sentence makes a claim. Studies are grouped according to their findings. A reason is given for why they disagree, and the evidence is weighed. The same paragraph is also pointing towards a gap. (These citations are made up for illustration.)
Phrases that signal synthesis
Look for phrases like “consistent with”, “in contrast”, “both studies rely on”, “this finding has not been replicated in” and “a limitation shared by these studies is”. Maybe most sentences in your section are opening with an author’s name. Then change them so the idea comes first.
How long should it be?
For most PhD theses, the review chapter comes to roughly 20 to 30 per cent of the main text, commonly 30 to 60 pages. Field and university make a difference here. In pure mathematics or some engineering areas the review may be much shorter, and many times it sits inside the introduction itself. To know what is normal in your department, look at its recent theses on Shodhganga. How to search there is explained in our Shodhganga guide.
How to revise and update a literature review
Most scholars write the review early, in the first or second year. The examiner reads it three or more years after that. So keep time to update it before you submit to the university. Run your saved searches again for the years in between and add relevant papers. Also confirm no recent publication has filled your gap already. Examiners really check this.
Use this list as the last check before the chapter goes to your guide.
- Every section starts with a claim about the literature. No section starts with an author’s name.
- Most paragraphs cite two or more studies and say how those studies connect with each other.
- Where studies disagree, you have said so and given a possible reason for it.
- Important studies from the last three years are there. The older studies that everybody in your field cites are also there.
- You have read every source yourself, at least the abstract, methods and results. No citation is picked up from somebody else’s review.
- The last part of the review states the gap, and from the gap your objectives follow directly.
- Nothing in the review falls outside your research questions, even if the material is very interesting.
If you want a specialist in your field to comment on your draft chapter’s organisation and synthesis, that is what our literature review support does. We work on your draft and on your reading, and the review remains your own work.
Want a specialist to help with this in your own work? Literature Review Structuring and Feedback: a free first consultation, then a written scope and fixed quote.
Sources
FAQ
Questions scholars ask
How many references should a PhD literature review have?
There is no correct number. Many theses cite somewhere between 100 and 300 sources overall. Still, a focused review with 120 sources that are well chosen and well used is stronger than one that just lists 300. A low count by itself is not a problem for examiners. Missing an important study they do notice.
How recent should my sources be?
Keep the founding studies however old they are, and see that the last five years are properly covered. If nothing from the last two or three years is in your review, it looks like the search was never updated before submission.
Is a literature review the same as a systematic review?
No, they are different. Usually a thesis literature review is a narrative review, where you select studies and interpret them to build an argument. A systematic review works from a protocol fixed in advance. The search has to cover everything, and the inclusion criteria are decided beforehand. Then the whole process gets reported, mostly as per PRISMA. Some theses have a systematic review as one chapter or one paper. The PRISMA 2020 steps are set out in our systematic review guide.
Can I cite Shodhganga theses in my review?
Yes, you can. For Indian-context evidence they are many times the best source available. Cite them as unpublished doctoral theses. Where peer-reviewed work also exists on the same point, give that more weight.
Can I use AI tools to summarise papers for my review?
Only with a lot of care. AI summaries can report findings wrongly, and they can invent references which do not exist. If you use them at all, use them only for deciding what to read. A paper you have not opened yourself should never be cited. Also check what your university’s current policy says on AI use in theses.
