How to do a literature review (2026 guide)

A literature review is one of the most important, and most misunderstood, parts of academic research. Done well, it doesn't just summarize what others have written. It maps the state of knowledge on your topic, identifies where the gaps are, and positions your own research within an ongoing scholarly conversation.

Done poorly, it becomes a list of summaries that tells your reader nothing about how the field has developed or why your research matters.

This guide walks you through the process step by step: from defining your research question to finding sources, evaluating them, organizing your reading, synthesizing what you've found, and writing a review that actually holds up.

What is a literature review?

A literature review is a critical survey of existing scholarly research on a specific topic. It identifies what has been studied, how it has been studied, what the findings show, where researchers disagree, and what remains to be explored.

It is not:

• An annotated bibliography (a list of summaries)
• A book report on individual sources
• A collection of quotes from papers you've read
• A simple description of what other researchers found

The key word is synthesis. A literature review weaves sources together to build an argument about the state of knowledge, showing patterns, contradictions, and gaps across the literature, not just describing each source in turn.

Types of literature review

The type of literature review you need depends on your research context. The most common types are:

Narrative or traditional review 

The most common type for dissertations and research papers. Covers the literature on a topic comprehensively but without a fixed methodology. Organized thematically or chronologically. Suitable for most academic writing contexts.

Systematic review

Follows a pre-registered, reproducible methodology to minimize bias. Typically uses PRISMA reporting standards. Common in health sciences, medicine, and social policy. Requires explicit inclusion and exclusion criteria and a documented search strategy.

Scoping review

Maps the breadth of literature on a topic without the depth of a systematic review. Useful when you want to understand what types of research exist on a topic before committing to a specific research question.

Meta-analysis

A quantitative synthesis of findings from multiple studies. Often accompanies a systematic review. Requires statistical expertise and is beyond the scope of most dissertation literature reviews.

Step 1: Define your research question

Every literature review starts with a question, not a topic.

"Climate change" is a topic. "How do coastal communities in low-income countries adapt to sea-level rise?" is a research question. The difference matters enormously: a topic has no boundaries, and a literature review without boundaries has no way to know when it's done.

A good research question is:

• Specific enough to be answerable
• Broad enough to have a body of literature around it
• Connected to a gap or debate in the existing research
• Relevant to your discipline and context

If you're struggling to define your research question, start with a preliminary search to understand the shape of the field. Read three to five recent papers on your topic and pay attention to what questions they raise, what they leave unanswered, and where they point to future research. That's often where your question lives.

Step 2: Search the literature

This is where most literature reviews go wrong, not because researchers search badly, but because they only use one type of search when two are needed.

Keyword search: the starting point

Begin with keyword searches in the databases relevant to your discipline. Useful starting points include:

• Google Scholar: broadest coverage, includes preprints and grey literature
• Scopus: strong for social sciences, sciences, and medicine
• Web of Science: stricter indexing, reliable for formal bibliometric analysis
• PubMed: health sciences and biomedical research
• PsycINFO: psychology and behavioral sciences
• JSTOR: humanities and social sciences

Use Boolean operators to sharpen your searches. AND narrows results (both terms must appear), OR expands them (either term), and NOT excludes a term. For example: "(climate adaptation OR climate resilience) AND coastal communities NOT urban" targets coastal research while excluding city-focused studies.

Build a list of synonyms and related terms before you search. Your research question uses specific vocabulary, but the papers you need might use different words, especially older papers, which may predate your current terminology.

Citation network search: finding what keyword search misses

Keyword search has a structural limitation: it can only find papers that use your exact words. This creates blind spots that affect almost every literature review built on keyword search alone.

The terminology gap means foundational papers from ten or twenty years ago may use completely different vocabulary than you're searching for. The connection gap means keyword search shows you papers in isolation, it can't show you that two papers you've already found both cite a third paper you haven't found yet, which might be the most important one for your review. And search algorithms optimize for recency, which can quietly bury older foundational papers that everyone in your field cites.

Citation network search solves these problems by navigating through connections between papers rather than through vocabulary.

In ResearchRabbit, once you've added one or more seed papers from your keyword search, you can:

• Select References to map what your seeds were built on, the foundational papers they cite, which often use older terminology your keyword search missed
• Select Citations to find what has been built on your seeds since they were published, the most recent work in your area
• Select Similar to find papers that are frequently co-cited alongside your seeds, papers the field treats as intellectually related, even if they use completely different terminology

Papers that appear repeatedly across different branches of this network are almost certainly foundational to your field, whether or not they appeared in your keyword search.

Document your search

Whether or not you're doing a systematic review, document what you searched, when you searched it, and what terms you used. This saves you from repeating searches and makes your methodology transparent to supervisors, reviewers, and future readers.

Step 3: Evaluate and screen your sources

Finding papers is only the first step. Not every paper you find belongs in your literature review, and not all of them deserve the same depth of attention.

Set inclusion and exclusion criteria

Before you start screening, decide what qualifies for inclusion. Common criteria include:

• Publication date range (e.g., last ten years, or all available literature)
• Language (most reviews are limited to English unless otherwise specified)
• Study type (peer-reviewed journal articles only, or including grey literature)
• Geographic scope if your research question is context-specific
• Relevance to your specific research question

Having these criteria in place before you start screening keeps your decisions consistent and defensible.

Screen efficiently

For each paper, assess relevance at three levels before committing to a full read:

Title and abstract (2-3 minutes): Is this paper addressing your research question? If the abstract doesn't suggest relevance, set it aside.

Introduction and conclusion (5-10 minutes): Does the paper's framing match your question? Does it contribute something you need?

Full read (30-60 minutes): Reserved for papers that are clearly central to your argument.

For a detailed guide to evaluating any paper quickly, including what to look for in the methods section and how to spot red flags, see how to evaluate a research paper quickly.

Step 4: Organize and manage your sources

A literature review involving dozens of papers across weeks or months of reading will collapse without a system for organization.

Use a reference manager

A reference manager stores your citations, generates formatted references automatically, and keeps notes attached to each paper. Zotero is free, open-source, and integrates directly with ResearchRabbit, making it the most practical choice for most researchers. Other options include Mendeley and EndNote.

Set up your reference manager before you start saving papers. Importing references you've already collected manually is far more time-consuming than capturing them automatically from the start.

Use ResearchRabbit collections and notes

In ResearchRabbit, organize papers into collections by theme or subtopic as you find them. A collection called "foundational, memory consolidation" tells you something when you come back to it three weeks later; a collection called "papers" tells you nothing.

Add a note to each paper as you read it, one sentence about what the paper contributes to your specific research question. ResearchRabbit's built-in notes keep this assessment attached to the paper itself, so you can see it alongside the abstract when you return.

Step 5: Read and take notes

The biggest time sink in a literature review isn't reading, it's rereading. Papers you read three weeks ago blur together. The way to avoid this is to take notes that will still make sense when you come back to them.

Match reading depth to paper type

Not every paper deserves the same depth of attention. Before opening any paper, decide how much time it warrants:

• Full read (30-60 minutes): Papers that are directly central to your research question, methodologically essential, or foundational to the field
• Focused skim (10-15 minutes): Papers useful for background or supporting context
• Quick scan (2-3 minutes): Papers you're assessing for relevance before deciding whether to read further

ResearchRabbit's citation map helps you prioritize before you start reading. Papers positioned high on the y-axis (many citations) that appear repeatedly across different branches of your network are almost always worth a full read.

Take notes that are evaluative, not descriptive

"This paper found that X" is less useful than "This paper found that X, which challenges my argument about Y" or "This paper's methodology is limited by Z, which I need to address when I cite it."

One evaluative sentence per paper, written immediately after reading, saves you twenty minutes of rereading later.

For a complete guide to efficient paper reading, see how to read research papers faster when doing a literature review.

Step 6: Synthesize, themes, not summaries

This is the step that separates a strong literature review from a weak one.

Synthesis means building an argument about what the literature shows, not describing each paper in turn. A literature review organized as "Jones (2015) found X. Smith (2018) found Y. Brown (2020) found Z." is not a literature review. It's a list of summaries.

Organize thematically, not by author

The body of your literature review should be organized around themes, debates, or methodological approaches in the literature, not around individual papers. Each section should

make a claim about what the literature shows in that area, then support the claim with evidence from multiple sources.

Look for patterns across sources

As you read, look for:

• Consensus: Where do most researchers agree?
• Contradiction: Where do findings conflict, and why?
• Evolution: How has thinking on this topic changed over time?
• Gaps: What hasn't been studied? What methods or populations are underrepresented?

These patterns become the backbone of your synthesis.

Start with an argument sentence

Each paragraph should open with a claim about what the literature shows in that area, not with a citation. "Research consistently shows that X" is a better opening than "Jones (2015) found that X." The second approach encourages listing; the first encourages synthesis.

Step 7: Write the review

With a structured outline based on your thematic synthesis, writing becomes significantly more manageable.

Structure

A standalone literature review or thesis chapter typically includes:

Introduction: States the topic, explains the scope and boundaries of your search, and previews the structure of the review. One to two paragraphs.

Body: Organized by theme, methodology, or chronology (thematic is usually strongest). Each section opens with an argument sentence, develops the claim with evidence from multiple sources, and ends with a transition that connects to the next theme.

Conclusion: Summarizes the state of the field, identifies the gaps your research addresses, and connects the review to your research question.

Writing tips

Use the active voice where possible, it's clearer and more direct than the passive. "Jones (2015) found" is stronger than "It was found by Jones (2015)."

Group papers that support the same point rather than discussing each one separately. "Several studies have documented X (Jones, 2015; Smith, 2018; Brown, 2020)" is more concise and more authoritative than three separate sentences.

Avoid starting every sentence with a citation. Sentences that begin with your own argument and then cite support are more readable and demonstrate synthesis rather than mere description.

Common mistakes

Starting to write before you've finished searching. Writing too early leads to a review built on an incomplete picture of the field. Finish your search, organize your reading, and outline your themes before you start drafting.

Relying only on keyword search. Keyword search misses papers with different terminology, older foundational work, and cross-disciplinary connections. Citation network search finds the papers your keyword search left behind.

Summarizing instead of synthesizing. A literature review is not a list of what each paper found. It's an argument about what the field shows. If your review could be reorganized in any order without losing meaning, it's probably describing rather than arguing.

Not using a reference manager. Managing citations manually across a long project leads to errors, duplicates, and missing references. Set up Zotero or another reference manager before you save your first paper.

Stopping too early. The most important papers in your field sometimes live two or three hops from your initial keyword search results. Citation network search, done iteratively, surfaces them. Don't stop at the first round of results.

Tools for literature reviews in 2026

ResearchRabbit, citation network discovery. Add seed papers and explore the network of connections around them, references (backward), citations (forward), and co-cited papers (similar). Finds papers that keyword search misses. Free, with RR+ adding advanced search filters including journal quality thresholds.

Zotero, reference management. Free, open-source, integrated with ResearchRabbit. Captures citations automatically from databases, generates formatted references, and lets you organize sources into collections with notes. The most practical choice for most researchers.

AI tools (Claude, ChatGPT, Gemini), useful for specific extraction tasks when used carefully. Effective prompts include "What is the methodology of this paper?", "What limitations do the authors acknowledge?", and "What claim does this paper make and what evidence supports it?" Not reliable for general summaries, which tend to flatten nuance and occasionally misrepresent findings. Never cite a paper based solely on an AI summary, verify the output against the actual paper before relying on it.

Google Scholar, broadest database coverage, free, includes preprints and grey literature. The best starting point for most searches.

Scopus and Web of Science, more selective indexing, better for formal bibliometric analysis. Both support advanced filtering and citation tracking.

FAQ

How long does a literature review take? 

It depends on the scope. A short literature review for a coursework assignment might take a week. A full literature review chapter for a PhD thesis can take several months of active work. The time is mostly spent reading and synthesizing, not searching, which is why efficient searching matters so much at the start.

How many sources do I need for a literature review? 

There's no universal number. A short coursework review might include 15-20 sources. A master's thesis literature review might include 40-80. A PhD chapter might draw on 100 or more. What matters is coverage, have you engaged with the key papers in your field and the major debates?, not hitting a specific number.

What's the difference between a literature review and a systematic review? A literature review is a critical survey of the field, organized around themes and arguments. A systematic review follows a pre-registered, reproducible methodology designed to minimize bias and is reported according to standards like PRISMA. Systematic reviews are more common in health sciences and are significantly more time-intensive. For a full comparison, see what is a systematic review.

Can I use AI to write a literature review? AI tools can help with specific tasks, extracting methodology from a paper, identifying limitations, generating a first outline. They're not reliable for synthesizing literature, evaluating sources, or producing text you can submit as your own work. The synthesis, argument, and critical evaluation in a literature review need to come from you.

How do I know when I've read enough? When new searches consistently surface papers you've already found, when you start seeing the same papers appearing repeatedly across different branches of your citation network, you've mapped the core of your literature. The signal isn't a specific number of papers; it's reaching a point where the network feels coherent and familiar.

What's the difference between a literature review and an annotated bibliography? An annotated bibliography is a list of sources with brief descriptions of each. A literature review synthesizes those sources into an argument about the state of knowledge on your topic. The annotated bibliography describes; the literature review argues.

‍