Most students write their first literature review like an annotated bibliography with transitions — one source per paragraph, summarized and then dropped. That is not a literature review; it is a reading log. A literature review is a structured synthesis of what is known on a question, organized by theme rather than by author, ending with the gap or tension that justifies whatever comes next (your study, your dissertation chapter, your argument). This guide covers how to scope the review, search systematically, synthesize rather than summarize, and write with a voice of your own running through the evidence. You will learn the thematic structure, how to handle disagreement in the field honestly, and the specific moves that make a review read as critical rather than passive.
PaperDraft is a writing assistant, not a paper generator — the draft is your starting point, not your submission. You are responsible for editing, verifying sources, and following your school's academic integrity policy.
What a literature review actually is
A literature review is a critical synthesis of the published scholarship on a specific question, organized thematically, ending with the current state of knowledge and its open questions. The purpose is not to demonstrate that you read — it is to give your reader a map of the field that identifies what is settled, what is contested, and what remains unasked.
Literature reviews take different forms. A standalone review (2,500 to 6,000 words) is assigned as a paper in its own right, often in upper-division or graduate coursework. A chapter review sits inside a thesis or dissertation, usually 6,000 to 12,000 words, and positions the study that follows. A systematic review follows a formal protocol (e.g., PRISMA) with explicit inclusion criteria and is common in health sciences. A scoping review maps a broader field without the systematic review's rigorous filtering.
It is not the same as an annotated bibliography, which lists and describes sources one at a time without the thematic synthesis. It is not the same as the background section of a research paper, which is briefer and positions one study. And it is not a history of the field — chronology is sometimes useful but rarely the best organizing principle.
Before you start
Three decisions shape everything that follows. Make them explicitly.
Decide whether this is standalone, chapter, or systematic
The type determines the length, the searching rigor required, and the voice. A standalone course review can be narrative and interpretive. A thesis chapter needs to justify a specific study. A systematic review follows a pre-registered protocol and is almost never what a coursework brief means unless it says "systematic review" explicitly.
Choose your citation style and stick to it
Literature reviews can run to 50 or more cited sources, so formatting matters. Psychology, education, and most social sciences use APA. Humanities commonly use MLA. Some fields (health, engineering) use their own numeric styles. Your field's convention probably dictates; when in doubt, confirm with the brief.
Secure database access before committing to a scope
A review's quality is bounded by the sources you can actually reach. Before you commit, confirm access to at least two relevant databases — Google Scholar for breadth, and a discipline-specific database (PubMed for health, PsycINFO for psychology, JSTOR for humanities, IEEE Xplore for engineering, Web of Science for interdisciplinary). If your library lacks the core database for your topic, narrow to something its catalog does cover.
Step-by-step: how to write a literature review
The procedure below is the difference between a reading log and a synthesis. Do not skip steps 2 and 4.
1. Define the review's scope and research question
State the question in one sentence. "What does the empirical literature since 2010 say about the relationship between social media use and adolescent depression in high-income countries?" — that sentence defines timeframe (since 2010), construct (social media use, adolescent depression), context (high-income countries), and source type (empirical). Everything outside those boundaries is not your review's problem.
Common mistake: a question so broad the review cannot finish ("What does the literature say about social media?"). Narrow until you can name boundaries on all axes.
Micro-example: "How have randomized controlled trials since 2015 evaluated cognitive behavioral therapy for adolescent anxiety, and what does the evidence suggest about effect size?"
2. Build a systematic search strategy
Pick your databases, then build search strings with keywords and their synonyms joined by Boolean operators. "(adolescent OR teen OR youth) AND ('social media' OR Instagram OR TikTok) AND (depression OR 'depressive symptoms')". Record the string, the database, and the date of the search. Save the searches so a reader could, in principle, reproduce them.
Common mistake: Googling around and citing what surfaces. You will miss the foundational work and the recent work at the same time.
3. Screen and select sources with explicit criteria
Decide in advance what qualifies. Peer-reviewed only? Studies with N ≥ 50? Published in English? 2015 onward? Write the criteria, then screen titles and abstracts against them. Exclude early and aggressively; it is easier to add a borderline source back than to prune a bloated review. Keep a simple count of sources screened, sources included, and sources excluded with reasons.
Common mistake: including a source because it is famous, not because it fits your criteria. Fame is not relevance.
4. Extract and synthesize into themes
Build a matrix — spreadsheet or table — with columns for author, year, research question, method, sample, key findings, limitations, and theme. Populate it as you read. Then look across the rows: which sources belong together? The clusters are your themes. Themes are the structure of the review. If your themes are just "studies from 2015," "studies from 2016," you have not synthesized — go back.
Common mistake: organizing by author ("Smith found X. Jones found Y. Brown found Z."). Organize by theme, then cluster sources within.
5. Draft the review around themes, not sources
Each H2 section is a theme. Within a theme, group sources that agree, then sources that diverge, and synthesize the disagreement. "Three studies (Smith, 2018; Jones, 2020; Lee, 2021) find a moderate positive association between X and Y, while two (Nguyen, 2019; Brown, 2022) find no effect; the divergence appears to track sample age, with the null-result studies using younger participants." That sentence is doing the work a literature review exists to do.
Common mistake: one paragraph per source. If each paragraph cites one author, the structure is wrong.
6. Identify the gap or tension your review surfaces
End the synthesis by naming what the literature does not resolve. "Research to date has focused on...; what is missing is...". The gap justifies your study if this is a thesis chapter, and it gives a standalone review a point. Without a stated gap, the review reads as descriptive rather than critical.
Common mistake: naming a "gap" that the literature has already filled. Read the most recent papers carefully — your gap should not be something published last year.
7. Revise for voice, balance, and citation density
Check three things on revision. First, your voice — are you synthesizing, or have you disappeared into the sources? Second, balance — is any theme starved for sources or drowning in them? Third, citation density — major claims should have at least one citation, and every citation should match a real source you read, not a source you saw cited by someone else.
Common mistake: citing sources at second hand. If you did not read Smith (1980), do not cite Smith (1980) — cite the source where you saw it discussed.
Stuck at the start? PaperDraft scaffolds a literature review draft — thesis, outline, opening sections — for you to revise. Start this paper — free.
Structure/outline template
For a 4,000-word standalone literature review, the following structure gives balanced proportions. Scale up for thesis chapters, scale down for coursework-length reviews.
Introduction (350–450 words). Name the topic and why it matters, state the review question, preview the themes, and name the inclusion criteria and scope. The reader should know in the first page what the review covers and what it does not.
Methods / search strategy (200–350 words, optional for non-systematic). The databases, search terms, date range, and inclusion criteria. In a coursework review this may be a short paragraph; in a systematic review it is a formal section.
Theme 1 (700–900 words). The first organizing theme. Synthesize the sources that address it, note convergence, note divergence. End with a one-sentence summary of where the literature stands on this theme.
Theme 2 (700–900 words). The second theme. Same structure. Transitions between themes should make the logical progression clear — usually moving from well-established to contested ground.
Theme 3 (500–800 words). The third theme, or a fourth if the scope requires it. Most coursework reviews hold to three or four themes; more fragments the synthesis.
Synthesis and gap (350–500 words). Step back. What do the themes collectively show? Where do they disagree? What is the gap — the unasked question, the unresolved tension, the methodological limitation across the field?
Conclusion (150–250 words). Restate the question, summarize the state of knowledge, and name the gap or implication. In a thesis chapter, this is where the study you are about to propose enters.
References (separate page). Every cited source, in the style your discipline requires.
Example excerpt
From a literature review on social media and adolescent depression, themed around sample age as a moderator.
Three longitudinal studies published since 2018 find moderate positive associations between social media use and depressive symptoms in adolescents. Twenge and Campbell (2018) report a dose-response relationship in a sample of 8th through 12th graders (N = 506,820), with heavy users (≥5 hours/day) reporting depressive symptoms at roughly twice the rate of light users (less than 1 hour/day). Kelly et al. (2019), using the UK Millennium Cohort, replicate the direction of the effect in 14-year-olds but with a smaller magnitude, and note that the association is stronger in girls. Two more recent studies, however, complicate the picture. Orben et al. (2022) find that the effect is statistically significant but small in magnitude (β = 0.12) when potential confounders are controlled, and Przybylski and Weinstein (2017) — working with a younger sample of 12- to 15-year-olds — find no reliable association at all. The apparent pattern is that studies drawing on older adolescents detect associations that studies of younger cohorts do not, suggesting sample age may be a moderator the field has not fully characterized.
Annotations: the paragraph names three sources that agree, two that diverge, and synthesizes the divergence as a moderator effect. The reader leaves the paragraph knowing what the literature shows and where it disagrees — the work of a review.
Common mistakes
- Author-by-author paragraphs. "Smith found X. Jones found Y." is a reading log, not a review. Cluster by theme.
- No named gap or contribution. A review that ends "more research is needed" has not done the work. Name the specific research that is needed and why.
- Unverified secondhand citations. If you cite Smith (1980) without reading Smith (1980), you are vulnerable to the source you saw it in having gotten it wrong. Either read the primary or cite the source as "(Smith, 1980, as cited in Jones, 2020)." This is the area where academic integrity lines are crossed most quietly; see our academic responsibility guide.
- Including sources because they are famous. If a famous paper does not fit your criteria, leave it out. Name-dropping weakens a review.
- No voice. A review where the writer never synthesizes, never evaluates, never pushes back — only reports — reads as passive. You are the synthesizer; speak.
How PaperDraft helps you start
PaperDraft is a writing assistant, not a paper generator — the draft is your starting point, not your submission. For literature reviews, it scaffolds the review question, proposes a thematic outline based on your topic, and drafts opening paragraphs in academic register with citation placeholders. You do the literature search (PaperDraft does not pretend to have read the field for you), then you populate the themes, verify every citation, and rewrite the synthesis in your voice. See our literature review landing for the drafting flow.
Frequently asked questions
How many sources should a literature review cite?
For a standalone undergraduate review, 20 to 40 sources is typical. A thesis chapter review usually cites 60 to 120. The brief sets the floor; the topic's maturity sets the ceiling. A review of a 30-year-old field needs more sources than a review of a field three years old.
How is a literature review different from an annotated bibliography?
An annotated bibliography lists sources individually with short descriptions. A literature review synthesizes sources into themes and argues about the state of the field. They sometimes accompany each other — the bibliography is the input, the review is the output. See our annotated bibliography guide for the input side.
Should I organize chronologically or thematically?
Thematically, almost always. Chronological organization makes sense only when the story of the field is genuinely a progression in time — for example, a review of methodological shifts. Even then, themes nested inside chronological periods usually read better than pure chronology.
Can I use AI to help with a literature review?
With careful disclosure, yes, for drafting and outlining. AI tools cannot do the literature search for you honestly — they hallucinate citations — and you must verify every source in your review against the original. See our AI disclosure guide for how to disclose AI assistance, and always check your course policy.
What citation style should I use for a literature review?
Your discipline's convention. Psychology, education, and social sciences use APA. Humanities typically use MLA. Health sciences often use Vancouver. The brief should specify; if it does not, match the field.
What does a "gap" in the literature look like?
A specific, defensible claim about what the literature has not yet addressed. "Studies have focused on X cohorts, but Y cohorts remain under-studied." "Effects have been documented in short-term designs, but long-term trajectories are unknown." A gap is not "more research is needed" — it is what research, specifically.