How to write a literature review that actually synthesizes
On this page
- Know what kind of review you are writing
- Step 1: Define the scope
- Step 2: Search systematically, even if informally
- Step 3: Read for what you need, and take structured notes
- Step 4: Choose an organizing principle
- Step 5: Write synthesis paragraphs
- Step 6: Make the gap explicit
- Citing well
- Common mistakes
- A simple outline you can adapt
A literature review is not a list of summaries. It is an argument about the state of knowledge on a topic: what is established, what is contested, what is missing, and why your study is the logical next step. Reviewers and supervisors can tell within a page whether a review synthesizes or merely reports, and the difference often decides whether the rest of the paper is taken seriously.
This guide covers the process from search to final draft, with paragraph templates you can adapt.
Know what kind of review you are writing
Before searching, be clear about the review's purpose:
- A review section inside an empirical paper (or thesis chapter) builds the case for your specific study. It is selective and ends with your hypotheses or research questions.
- A standalone narrative review maps a field for readers, often organized by themes or debates.
- A systematic review follows a pre-registered, reproducible search and screening protocol (for example, PRISMA) and reports exactly how studies were selected.
Most students are writing the first kind. That means you do not need every paper ever written; you need the papers that define the concepts, the main findings and the gap.
Step 1: Define the scope
Write down your topic as a question, list the core concepts, and decide on boundaries: time period, populations, settings, methods and languages. For a study on remote work and burnout, the core concepts might be remote-work intensity, burnout or exhaustion, and boundary management, limited to knowledge workers and studies from the last ten to fifteen years plus the key theoretical sources.
Scope protects you from the most common trap: reading endlessly without writing.
Step 2: Search systematically, even if informally
Use a scholarly database relevant to your field (Web of Science, Scopus, PubMed, PsycINFO, ERIC, Google Scholar or an open index such as OpenAlex). Combine synonyms with OR and concepts with AND, and keep a simple log of the searches you ran. Then:
- start with two or three recent, highly cited reviews to learn the landscape
- follow references backward (what they cite) and forward (who cites them)
- prioritize peer-reviewed work, and note where evidence comes only from preprints or grey literature
Record each source in a reference manager (Zotero, Mendeley, EndNote) from day one. Rebuilding a reference list at the end is slow and error-prone.
Step 3: Read for what you need, and take structured notes
You do not need to read every paper cover to cover. For most sources, read the abstract, introduction, the key results table and the discussion. For the handful of studies that are central to your argument, read the full text carefully.
A synthesis matrix makes the next step much easier. Create a table with one row per source and columns such as:
- citation
- context and sample
- design and measures
- main finding
- limitations
- how it relates to your question
Patterns that are invisible in a pile of PDFs become obvious in a matrix: three studies use the same measure, two find opposite effects, nobody has studied a particular population.
Step 4: Choose an organizing principle
Organize the review around ideas, not authors. Common structures:
- Thematic: each subsection covers one concept, mechanism or debate.
- Variable-by-variable: common in quantitative papers: the outcome, the main predictor, then each mediator or moderator, ending with how they connect.
- Methodological: grouping studies by approach when methods explain differing results.
- Chronological: useful only when the field's development is itself the point.
Draft your subheadings before writing paragraphs. Each one should make a claim or name a debate, not just a topic.
Step 5: Write synthesis paragraphs
The core skill is the synthesis paragraph. A reliable pattern:
- Topic sentence making a claim about the literature: "Evidence on whether remote work increases exhaustion is mixed."
- Evidence from several sources, grouped by what they show: "Several cross-sectional surveys report higher exhaustion among fully remote employees (Author, Year; Author, Year), whereas longitudinal studies find the association weakens after the first months (Author, Year)."
- Analysis that explains the pattern: differences in samples, measures, designs or contexts.
- Link to your study: "These inconsistencies suggest that the effect depends on how employees manage work–home boundaries."
Notice that the paragraph is about the idea, and sources appear as evidence in parentheses. If most of your sentences begin with "Smith (2020) found…", you are summarizing, not synthesizing.
Useful moves include comparing ("Consistent with…", "In contrast to…"), qualifying ("However, these studies relied on…"), and evaluating ("A limitation shared by this work is…").
Step 6: Make the gap explicit
End the review (or each major subsection) by stating what remains unknown and why it matters. Strong gap statements are precise:
- a population gap: the effect is established for one group but not another
- a methodological gap: findings rely on cross-sectional self-reports
- a theoretical gap: two frameworks predict different outcomes and have not been tested against each other
- a contextual gap: post-2020 work arrangements differ from pre-pandemic telework
Then state your research questions or hypotheses. In a proposal, this is where the review hands over to the design; see how to write a research proposal.
Citing well
Cite every claim that is not your own reasoning or common knowledge, prefer primary sources over secondhand citations, and keep citation style consistent. When you cannot find a source for a claim, either find one or soften and flag the claim; never attach a citation you have not checked. Our citation basics guide explains author-year formats and common errors.
Common mistakes
- The annotated bibliography: one paragraph per paper, no comparison.
- No critical voice: reporting findings without evaluating their quality.
- Outdated or unbalanced sources: ignoring recent work or only citing studies that support your view.
- Over-quoting: paraphrase and synthesize; quote only when the exact wording matters.
- A gap that does not match the study: the gap you identify must be the one your design addresses.
A simple outline you can adapt
- Opening paragraph: the scope of the review and how it is organized.
- Theme or variable 1: what is established, where findings conflict.
- Theme or variable 2: same pattern.
- How the themes connect (often the theoretical framework).
- Summary of gaps and the resulting research questions or hypotheses.
Once the outline and synthesis matrix exist, drafting goes much faster. If you are stuck at the blank page, see our guide on turning notes into a first draft.