How to write a review paper: format and example

A review paper, also called a review article, collects what has already been published on a topic and makes sense of it for other researchers. Editage puts the main difference from a research paper simply: a research paper "is based on original research", while a review paper "does not report original research". You do not run new experiments. Your material is other people's papers.

A review paper is also different from the literature review chapter of a thesis. That chapter exists to lead up to your own study. A review paper stands on its own, and it is written for researchers across the field, not for your examiners. If you want the basics first, read our guide on what a literature review is.

Literature reviews are in great demand in most scientific fields.

Before you write a word, find out which reviews already exist. In his "Ten Simple Rules for Writing a Literature Review", Marco Pautasso warns that "the chances are high that someone will already have published a literature review", if not on your exact topic, then on a related one.

1. Search for reviews only.In Scopus, open Document type on the left and choose Review. In PubMed, tick Review or Systematic Review under Article type. In Google Scholar, add the word "review" to your search. Scopus tags some book reviews the same way, so read the titles.
2. Note the date of each review.An old review in a fast-moving field may leave room for a new one.
3. Read their future research sections.Recent reviews often list what they could not cover, which can be your starting point.
4. Find your angle.Your review must add something: a narrower focus, newer studies, a comparison nobody has made, or a framework that ties the field together.

Sanjana found two recent reviews on biodiesel from waste cooking oil. Both covered the whole process, from collecting the oil to testing the fuel, and gave the catalyst, the substance that speeds up the reaction, only a few pages. Much of her own reading was about one kind of catalyst: those made from waste, such as eggshells and seashells. That was her angle.

Review papers come in a few main kinds, and each has its own rules. Editage lists three: the narrative review, the systematic review and the meta-analysis. The scoping review is a fourth.

Pautasso describes the main split this way: "while narrative reviews are qualitative, systematic reviews attempt to test a hypothesis based on the published evidence." Our guide to the types of literature review explains each kind in more detail.

For a systematic review, two names matter. PRISMA 2020, published by Matthew Page and colleagues in the BMJ in 2021, is a 27-item reporting checklist with a flow diagram that shows how many studies were found, screened and included. PROSPERO is an international register where health-related systematic reviews record their protocol, and its current guidance asks you to register before data extraction begins.

Sanjana chose a narrative review: her question was broad and her studies used very different methods. She would still describe her search, because SANRA, a scale that journal editors developed to judge narrative reviews, checks it too.

There is no single format for every journal, so always start with the guide for authors of the journal you are aiming at. Most review papers, though, follow a shape like this.

The main body is where reviews differ most. Pautasso notes that a review, like other papers, needs "a general introduction of the context and, toward the end, a recapitulation of the main points covered". What comes in between is up to you, but the headings should come from the ideas in the field, not from the list of papers.

Sanjana's new outline fit on one page. After the introduction and a short methods section came four themes: catalysts from eggshells, from seashells, from plant ash, and catalysts that wear out after a few uses. Then a comparison table, future directions and the conclusion.

With the outline settled, the writing follows a clear order.

1. Write your aim in one sentence.Sanjana's aim read: "This review compares catalysts made from waste materials for producing biodiesel from used cooking oil, and asks what limits their use at a larger scale."
2. Search, and search again.Pautasso's second rule is to "search and re-search the literature". Keep a record of the databases, search words and dates, so you can write the methods section later.
3. Fill a matrix as you read.One row per study, one column per question: the catalyst, its source material, the reaction conditions, the yield and the number of reuses. A literature review matrix turns a folder of PDFs into something you can compare.
4. Group the studies into themes.Read down the columns of the matrix. The patterns you see there become your headings.
5. Draft the main body first.Write the introduction and the abstract last, when you know what the review actually says.
6. Build the summary table.Keep the same columns for every row, so a reader can compare the key studies at a glance.
7. Ask for feedback.Pautasso's eighth rule is to make use of feedback. Give the draft to your guide and to one reader outside your group.

For the search step, Fiberarticle's Literature Reviewer finds papers across arXiv, OpenAlex, Semantic Scholar and Crossref and builds an evidence matrix you can export. Check what it finds against the papers themselves, because the judgement in a review has to be your own.

"Reviewing the literature is not stamp collecting," Pautasso writes. A good review "does not just summarize the literature, but discusses it critically." That was exactly what was wrong with Sanjana's 38 pages.

The second paragraph is hardly longer, yet it says far more: it joins the studies, names a pattern and ends with a gap. Be fair to every study, your own included. Pautasso's ninth rule is to include your own relevant research, but be objective about it.

Not every journal takes review papers, and many that do prefer to invite them. Editage notes that "review articles are typically sought out by the journal" and that "unrequested articles are not often published". Elsevier's Researcher Academy adds that at some journals it "might only be possible to submit a review article by invitation or prior consultation with an editor".

That does not shut students out. It means you should ask first. Many journals accept a short proposal, sometimes called a presubmission inquiry. Springer's journal Molecular Diversity, for example, asks for a proposal of two pages at most before an unsolicited review is submitted. A proposal usually covers:

• what the review will cover, in a short paragraph
• why it suits this journal and its readers
• how it differs from the reviews already published
• a list of key papers on the topic

Before you submit, test the draft the way an editor might. For a systematic review, go through the PRISMA checklist. For a narrative review, SANRA, published in 2019 by the editors Christopher Baethge, Sandra Goldbeck-Wood and Stephan Mertens, scores six things: why the review matters, its aims, the literature search, the referencing, the scientific reasoning and the presentation of relevant results. Then write a cover letter that says why your review is needed now. Our guide on writing a cover letter for journal submission shows how.

Sanjana emailed a proposal to the editor of a journal on renewable fuels. Two weeks later the reply came: send the full review. Dr. Gokhale read her new draft, now 22 pages built on four themes, and asked only one question: "Which theme will your own thesis answer?" She pointed to the last one, the catalysts that wear out.