A literature review is one of the most fundamental components of academic research and one of the most misunderstood. Whether you are writing a thesis, preparing a journal article, or beginning a new research project, mastering the literature review is essential. This guide covers everything you need to know: what a literature review is, how to search and synthesize the literature, the different types, and how to structure one effectively.
What Is a Literature Review?
A literature review is a critical analysis and synthesis of existing published material on a particular topic. It does more than simply summarize sources; it evaluates and connects them, drawing out themes, contradictions, methodological strengths and weaknesses, and gaps in current knowledge.
In the words of the University of Toronto’s Writing Advice centre, a literature review conveys “what knowledge and ideas have been established on a topic, and what their strengths and weaknesses are.” It must be defined by a guiding concept) your research objective, research question, or argumentative thesis) rather than being a descriptive list of materials.
What a Literature Review Is NOT
It is equally important to understand what a literature review is not:
- Not an annotated bibliography: an annotated bibliography summarizes each source individually with no synthesis between them
- Not an essay or research paper: it does not prove or develop a central argument from scratch
- Not a list of summaries: it goes beyond “Source A says X, Source B says Y”
- Not a report: it does not present findings from primary data collection
Literature Review vs Background Section: Key Differences
| Aspect | Literature Review | Background (in a Research Proposal/Research Paper) |
| Primary Goal | Comprehensively synthesize and critically evaluate existing knowledge on a topic | Set the context and justify why the specific study being proposed/reported is needed |
| Scope | Broad — covers the full landscape of relevant research, including differing schools of thought, methodologies, and findings | Narrow and selective — includes only the sources needed to frame the specific research gap |
| Depth of Coverage | Exhaustive within its defined boundaries; aims to leave no major relevant work unaddressed | Curated; deliberately excludes tangential work even if relevant to the broader field |
| Structure | Often organized thematically, chronologically, or methodologically, with its own internal logic | Usually organized narrowly and linearly: general context → narrowing focus → the specific gap → how this study addresses it |
| Level of Critical Analysis | High — evaluates strengths, weaknesses, contradictions, and methodological quality across studies | Lower — critical only where needed to justify the gap; not an evaluation of the entire field |
| Length | Can range from several pages (as a chapter/section) to an entire standalone paper | Typically short — a page or two, sometimes just a few paragraphs |
| Standalone Value | Can exist independently as a publishable contribution (e.g., a review article) | Cannot stand alone; only meaningful as part of a larger proposal or paper |
| End Point | Concludes by identifying gaps, debates, or open questions in the field generally | Concludes by directly leading into the specific research question, hypothesis, or objective of the study |
| Reader’s Takeaway | “Here is the state of knowledge on this topic, and here’s where the field stands overall” | “Here is why this particular study needs to happen” |
| Typical Location | A dedicated chapter/section (in a thesis) or its own article (as a review paper) | A subsection near the start of the Introduction in a proposal, grant application, or research paper |
Core distinction: a literature review maps an entire field to show what is known and unknown broadly, while a background section is a brief, targeted narrative that uses only as much of the literature as necessary to justify one specific piece of new research. In practice, a background section often reads like a compressed, purpose-built extract of a larger literature review.
Why Is a Literature Review Important?
A well-conducted literature review serves multiple interconnected purposes:
| Purpose | Description |
| Contextualises the research | Places your study within the existing body of knowledge |
| Identifies knowledge gaps | Highlights where research is absent, contradictory, or limited |
| Establishes credibility | Demonstrates your familiarity with the field |
| Informs theory and methods | Helps justify your theoretical framework and methodological choices |
| Guides research questions | Shapes and sharpens what you choose to investigate |
| Avoids duplication | Ensures your work builds on, rather than repeats, prior studies |
| Informs practice | For practitioners, review articles offer state-of-the-art evidence to guide decisions |
Types of Literature Reviews
The table below shows the common types of literature reviews:
| Type | Core Purpose | Approach | Best Used When |
| Narrative Review | Summarize and synthesize a broad topic | Flexible, thematic, qualitative | Providing background context; exploring a broad topic |
| Systematic Review | Answer a specific question with minimal bias | Highly structured, protocol-driven, exhaustive search | Aggregating evidence for evidence-based practice |
| Meta-Analysis | Statistically combine results from multiple studies | Quantitative synthesis using statistical methods | Sufficient homogeneous studies exist |
| Scoping Review | Map the extent and nature of a topic area | Comprehensive but exploratory | Emerging topics; assessing feasibility of a full systematic review |
| Rapid Review | Answer a specific question under tight time constraints | Systematic-style methods, streamlined or abbreviated to save time | Policy or clinical decisions need evidence quickly |
| Descriptive/Mapping Review | Identify patterns and trends in the literature | Systematic search + frequency analysis | Understanding publication trends in a field |
| Critical Review | Critically evaluate and interpret existing literature | Reflective, appraisal-focused | Identifying methodological weaknesses in a body of work |
| Integrative Review | Synthesize findings across studies with different methodologies | Combines quantitative, qualitative, and mixed-methods evidence | Building a fuller picture of a complex, multifaceted topic |
| Realist Review | Understand how and why complex interventions work | Theory-driven, context-sensitive | Complex social or policy interventions |
| Umbrella Review | Synthesize findings from multiple systematic reviews | Tertiary synthesis | Broad research topics with multiple existing systematic reviews |
A Closer Look at the Most Common Types
Narrative Reviews
Narrative reviews are the most traditional form. They offer flexibility in how sources are searched and synthesized, and are particularly useful for providing context in research papers or introducing students and practitioners to a topic. However, they can be subjective if not carefully conducted.
Systematic Reviews
Systematic reviews apply explicit, reproducible methods to minimize bias. They typically require a pre-registered protocol, multiple independent reviewers, and exhaustive database searches. They are the gold standard in health and medical research.
Scoping Reviews
Scoping reviews are ideal when a topic is emerging or when you want to assess how much literature exists before committing to a full systematic review. They include grey literature and aim to be as comprehensive as possible.
Meta-analyses
A meta-analysis statistically combines data from multiple studies to generate a pooled effect size, providing more precise estimates of relationships or intervention effects than any single study can offer.
Rapid Reviews
A rapid review follows a systematic review’s basic logic but compresses the timeline, often to a few weeks rather than months. To achieve this, researchers streamline certain steps: limiting the number of databases searched, restricting results by date or language, skipping dual independent screening, or having a single reviewer extract data with a second checking only a sample. The trade-off is a higher risk of bias or missed studies in exchange for speed. Rapid reviews are common when policymakers, health systems, or organizations need evidence-based guidance quickly, such as during an emerging public health situation, and are typically framed as a starting point rather than a definitive answer, sometimes flagged for later upgrading to a full systematic review.
Umbrella Reviews
An umbrella review, sometimes called a “review of reviews,” sits a level above systematic reviews: instead of synthesizing individual primary studies, it synthesizes findings across multiple existing systematic reviews and meta-analyses on related questions within a broader topic. This makes umbrella reviews especially useful when a field already has several systematic reviews, but a bird’s-eye view is needed to compare their conclusions, identify where reviews agree or conflict, and assess the overall quality and consistency of the evidence base. They are common in medicine, where dozens of systematic reviews on related interventions or outcomes may accumulate over time, making a tertiary synthesis valuable for guiding clinical or policy decisions.
Integrative Reviews
An integrative review draws together and synthesizes findings from studies with diverse methodologies, quantitative, qualitative, and mixed-methods, within a single review. Unlike a systematic review, which usually restricts itself to one methodological type to allow direct comparison, an integrative review deliberately combines different kinds of evidence to build a more complete picture of a topic. This makes it well suited to complex, multifaceted phenomena in fields like nursing, education, and management, where relevant insights come from experimental studies, observational research, and qualitative accounts alike. The analytical work is demanding, since the reviewer must find a coherent way to compare and combine evidence that was generated using very different methods.
Realist Reviews
A realist review asks not simply whether an intervention works, but how, why, for whom, and under what circumstances it works or fails. Rooted in realist philosophy of science, this approach is theory-driven: reviewers develop and refine explanatory theories about the underlying mechanisms that produce outcomes, and test these theories against the evidence in light of the contexts in which interventions are implemented. This makes realist reviews particularly well suited to complex social, educational, or policy interventions, where the same program can succeed in one setting and fail in another because of contextual factors. Rather than producing a single pooled verdict, a realist review typically yields context-mechanism-outcome configurations that explain the conditions under which an intervention is likely to succeed.
How to Choose the Right Type of Literature Review
| Question to ask yourself | If yes / applies | Suggested review type(s) |
| Do I just need broad background/context on a topic, without a rigid methodology? | Yes | Narrative review |
| Do I need to answer one specific, focused research question with minimal bias? | Yes → check timeline below | Systematic review, meta-analysis, or rapid review |
| Do I have only a few weeks, not months, to complete this? | Yes | Rapid review |
| Do the included studies report comparable quantitative data that can be statistically pooled? | Yes | Meta-analysis |
| No, but I still need a structured, protocol-driven approach | Yes | Systematic review |
| Is the topic new or emerging, and I’m not yet sure how much literature exists? | Yes | Scoping review |
| Do I need to compare conclusions across several existing systematic reviews on related topics? | Yes | Umbrella review |
| Does my evidence base include a mix of quantitative, qualitative, and mixed-methods studies I need to combine? | Yes | Integrative review |
| Do I need to explain how, why, for whom, or under what conditions a complex intervention works (not just whether it works)? | Yes | Realist review |
| Does my field/journal have a strong methodological convention (e.g., PRISMA in health sciences)? | Yes | Follow that convention regardless of the above — field norms override general fit |
| Do I have a registered protocol or plan to register one (e.g., PROSPERO)? | Yes | Signals systematic review or meta-analysis; rapid/narrative reviews are rarely pre-registered |
| Am I a single reviewer with no second person to independently screen/extract data? | Yes | Leans toward rapid or narrative review; systematic reviews typically require dual review for rigor |
| Do I need the review to be publishable as a standalone paper (not just a background section)? | Yes | Any of the above except a research proposal’s background section |
How to use this: work down the left column in order. The first “yes” that applies is usually your primary indicator. If two rows point to different types (e.g., timeline says rapid, but your field’s norms say systematic), prioritize field norms and stated publication requirements over timeline convenience, since a rapid review may need to be explicitly justified or flagged as a limitation in more traditional fields.
How to Search the Literature: A Step-by-Step Guide
Effective literature searching is a skill. A poorly executed search can leave crucial sources undiscovered or flood you with irrelevant material. Follow these steps systematically.
Step 1: Define Your Research Question
A well-formulated question shapes every later step of a review, from search terms to inclusion criteria, so choosing the right framework matters.
PICO
PICO (Population, Intervention, Comparison, Outcome) is the standard in health and clinical sciences. It works well for intervention-effectiveness questions, e.g., “In adults with type 2 diabetes (P), does metformin (I) compared to lifestyle intervention alone (C) reduce HbA1c levels (O)?” Variants like PICOS (adding Study design) or PICOT (adding Time) refine it further for systematic reviews.
SPIDER
SPIDER (Sample, Phenomenon of Interest, Design, Evaluation, Research type) is better suited to qualitative and mixed-methods research, where there’s no discrete “intervention” to test. It focuses instead on the experience or phenomenon being studied, e.g., “What are new mothers’ (S) experiences of postnatal anxiety (PI) as explored through interviews (D), regarding coping strategies (E), in qualitative studies (R)?”
PEO
PEO (Population, Exposure, Outcome) fits observational and exposure-based questions common in public health and epidemiology, where there’s no intervention being tested, just a naturally occurring exposure, e.g., “Among adolescents (P), does social media use (E) affect sleep quality (O)?”
Choosing the wrong framework often produces overly broad or oddly structured questions, so it’s worth matching the framework to the review’s underlying design before drafting a question.
Step 2: Choose Your Databases
Different fields have different key databases. Use multiple databases to maximize coverage:
| Field | Recommended Databases |
| Medicine & Health | PubMed/MEDLINE, Cochrane Library, CINAHL, Embase |
| Social Sciences | PsycINFO, Sociological Abstracts, ERIC |
| General/Multidisciplinary | Web of Science, Scopus, Google Scholar, JSTOR |
| Engineering & Technology | IEEE Xplore, ACM Digital Library |
| Humanities | JSTOR, Project MUSE, MLA International Bibliography |
Step 3: Develop Your Search Strategy
- Identify keywords: Use synonyms and related terms (e.g., “adolescent,” “teenager,” “youth”)
- Use Boolean operators: AND narrows your search; OR broadens it; NOT excludes terms
- Apply truncation and wildcards: e.g., “therap*” captures therapy, therapies, therapist
- Use subject headings: Controlled vocabulary like MeSH terms (in PubMed) improves precision
- Set date filters: Limit to a relevant time period based on your topic’s history
- Document your search: Record exactly what you searched and when — this is essential for transparency and reproducibility
Database-specific syntax
- PubMed uses MeSH (Medical Subject Headings) terms and field tags like [tiab] for title/abstract.
- Scopus and Web of Science support advanced field codes (TITLE-ABS-KEY, TS=) and citation-tracking features.
- Google Scholar has limited Boolean support and no field-specific tagging, so it’s typically used to supplement, not replace, database searches.
Step 4: Screen and Select Sources
Once you have your initial results, apply inclusion and exclusion criteria:
- Inclusion criteria might specify: peer-reviewed articles, published after a certain year, studies conducted in a specific population, written in a specific language
- Exclusion criteria might eliminate: conference abstracts, grey literature (unless doing a scoping review), studies below a quality threshold
For systematic reviews, screening should be done independently by at least two reviewers, with a clear process for resolving disagreements.
Step 5: Assess Source Quality
Not all published research is of equal quality. Ask yourself:
- Is it peer-reviewed?
- Is the methodology clearly described and appropriate?
- Is the sample size adequate?
- Are the conclusions supported by the data?
- Are there conflicts of interest?
- For qualitative studies: is the researcher’s positionality addressed?
Step 6: Supplement Your Search
- Backward searching: check the reference lists of key papers you have already identified
- Forward searching: use Google Scholar or Web of Science to find papers that have since cited a key article
- Grey literature: government reports, theses, and preprints may be relevant, particularly for scoping reviews
How to Synthesize the Literature
Synthesis is the intellectual core of a literature review. It is what distinguishes a literature review from an annotated bibliography.
Difference between Summary, Synthesis, and Evaluation of Literature
| Activity | What It Means | Example |
| Summary | Recapping the key findings of a single source | “Smith (2020) found that exercise reduced anxiety scores by 30% in adults.” |
| Synthesis | Re-organising and connecting information across multiple sources to reveal relationships | “While Smith (2020) and Jones (2019) both found that exercise reduced anxiety, they differ on whether aerobic or resistance training is more effective.” |
| Evaluation | Assessing the worth, quality, or significance of a source | “However, Smith’s (2020) study is limited by its small, homogeneous sample, which may restrict generalizability.” |
A strong literature review combines all three activities throughout.
Narrative Synthesis vs. Statistical Pooling
These are the two main approaches to combining findings once studies are selected, and the choice depends on how comparable the included studies are.
Statistical Pooling
Statistical pooling (used in meta-analyses) combines numerical results from multiple studies into a single, quantitative effect size, using statistical models such as fixed-effect or random-effects models. This requires the studies to be sufficiently homogeneous, i.e., similar populations, interventions, and outcome measures, so that combining their data produces a meaningful, precise estimate rather than an apples-to-oranges average.
Narrative synthesis
Narrative synthesis is used when studies are too heterogeneous in design, population, or outcome measures to pool statistically. Instead, findings are summarized and compared in text, often organized by theme, outcome, or study characteristic, with tables used to structure comparisons. It’s inherently more subjective than statistical pooling since it relies on the reviewer’s interpretation and grouping of results rather than a standardized formula.
When to choose narrative synthesis vs statistical pooling?
| Factor | Choose Narrative Synthesis | Choose Statistical Pooling (Meta-Analysis) |
| Study homogeneity | Studies vary substantially in population, intervention, or outcome measures | Studies are sufficiently similar in design, population, and outcomes |
| Outcome measures | Outcomes measured inconsistently across studies (different scales, tools, or definitions) | Outcomes measured consistently or convertible to a common metric |
| Study designs included | Mix of qualitative, quantitative, and/or mixed-methods studies | Predominantly quantitative studies with comparable numerical data |
| Number of studies | Very few studies, or too few per subgroup to pool meaningfully | Enough studies (typically several) to generate a statistically stable pooled estimate |
| Statistical heterogeneity | High heterogeneity (e.g., high I² statistic) that would make pooling misleading | Low to moderate heterogeneity, or heterogeneity that can be explained/adjusted for |
| Goal of synthesis | Understanding patterns, context, or a range of findings | Producing one precise, generalizable effect size |
| Data availability | Missing or incomplete raw/summary data from studies | Full or sufficient summary statistics available from each study |
The two aren’t mutually exclusive: a review may narratively synthesize findings from one subset of studies while pooling data statistically from a more homogeneous subset. Choosing narrative synthesis doesn’t lower a review’s rigor; it’s simply the appropriate choice when statistical pooling would misleadingly force incompatible data into a single number.
Strategies for Effective Synthesis
- Group sources thematically rather than discussing one paper at a time. Cluster sources that address the same concept or finding
- Compare and contrast. Identify where researchers agree, disagree, or use conflicting methods
- Track developments over time. Note how thinking on a topic has evolved across decades
- Identify patterns. Look for recurring findings, dominant methodological approaches, or shared theoretical frameworks
- Highlight gaps: point out where has research been sparse, inconsistent, or entirely absent
- Build a synthesis matrix: a table with sources as rows and key themes or variables as columns helps you visualize connections and gaps before you start writing
Synthesis Matrix Example
| Source | Population | Methodology | Key Finding | Limitations |
| Smith (2020) | Adults 18–40 | RCT | Exercise ↓ anxiety by 30% | Small sample |
| Jones (2019) | Older adults | Cohort study | Resistance training most effective | No control group |
| Patel (2021) | Adolescents | Within-subjects design | Exercise improved self-efficacy | Self-report measures |
What Synthesis Is NOT
- Stringing quotes together from different papers
- Summarizing each paper in a separate paragraph, one after the other
- Paraphrasing without analysis: true synthesis draws connections the original authors themselves may not have made explicit
Quality/Risk-of-Bias Appraisal Tools for Literature Reviews
Once studies are selected, their methodological quality must be appraised before their findings are trusted or synthesized, and the right tool depends on study design.
- CASP (Critical Appraisal Skills Programme) checklists: offer a set of design-specific checklists (for RCTs, cohort studies, qualitative research, systematic reviews, etc.), each posing structured questions about validity, results, and relevance. They’re widely used for their simplicity and accessibility, especially by newer reviewers.
- GRADE (Grading of Recommendations Assessment, Development and Evaluation): doesn’t just appraise individual studies but rates the overall certainty of evidence for each outcome across a body of studies, considering factors like risk of bias, inconsistency, imprecision, indirectness, and publication bias. It’s the standard for producing clinical practice guidelines.
- Cochrane Risk of Bias tool (currently RoB 2): is specific to randomized controlled trials, assessing bias across defined domains: randomization process, deviations from intended interventions, missing outcome data, outcome measurement, and selective reporting.
- MMAT (Mixed Methods Appraisal Tool): is designed for reviews combining qualitative, quantitative, and mixed-methods studies within one synthesis, since single-design tools can’t meaningfully appraise all three. It includes separate criteria sets depending on each study’s design.
- JBI (Joanna Briggs Institute) checklists: provide a broad suite of critical appraisal tools covering diverse study types, including prevalence studies, case reports, and qualitative research, and are commonly used in scoping and systematic reviews, particularly in nursing and allied health fields.
Choosing a mismatched tool, e.g., applying an RCT-specific tool to observational studies, undermines the appraisal’s validity, so matching the tool to both the study design and review type is essential.
Common Pitfalls in Literature Reviews and How to Avoid Them
Even well-intentioned literature reviews can fall short due to a handful of recurring mistakes. Being aware of these early can save significant rework later.
Overreliance on abstracts instead of full texts
Abstracts are compressed summaries and often omit crucial methodological details, limitations, or nuanced findings. Relying on them alone risks misrepresenting a study’s actual scope or conclusions, and can lead to including studies that don’t actually meet inclusion criteria once the full text is examined. Abstracts are useful for initial screening, but decisions about inclusion, quality appraisal, and data extraction should always be based on the full text.
Cherry-picking studies that confirm a hypothesis (confirmation bias)
It’s tempting, consciously or not, to give more weight to studies that support a preferred conclusion while downplaying or excluding contradictory evidence. This undermines the credibility of the entire review. A rigorous review reports the full range of findings, including conflicting or null results, and explains discrepancies rather than glossing over them. Predefined, transparent inclusion/exclusion criteria (ideally set before searching, as in a registered protocol) help guard against this.
Outdated sources / neglecting recency
Citing only older, “foundational” work while ignoring recent developments can make a review feel disconnected from the current state of a field, especially in fast-moving disciplines like technology or medicine. While seminal older studies often deserve inclusion for context, reviewers should explicitly search for and incorporate recent literature, and periodically re-run searches if the review process spans many months.
Poor synthesis (a “list of summaries” rather than an integrated argument)
One of the most common structural weaknesses: presenting a string of “Study A found X. Study B found Y. Study C found Z.” without ever connecting these findings into a coherent argument. A strong literature review identifies patterns, contradictions, and gaps across studies, and organizes discussion thematically or conceptually rather than study-by-study. The goal is synthesis, not summary.
Ignoring grey literature where relevant
Grey literature (conference proceedings, theses, government reports, preprints, clinical trial registries) can be essential for topics prone to publication bias, since studies with null or negative results are less likely to appear in peer-reviewed journals. Excluding grey literature entirely, without acknowledging the decision, can skew a review toward inflated effect sizes or overly positive conclusions. Whether to include it should be a deliberate, justified methodological choice, not an oversight.
Citation/reference management errors
Misattributed quotes, incorrect page numbers, duplicate entries, or inconsistent citation formatting undermine a review’s credibility and can misdirect readers trying to verify claims. These errors compound quickly in reviews citing dozens or hundreds of sources. Using reference management software (e.g., Zotero, EndNote, Mendeley) from the outset, and periodically auditing the reference list against the in-text citations, helps catch these issues before submission.
Literature Review for a Thesis or Dissertation: Special Considerations
A thesis or dissertation literature review carries unique expectations compared to a standalone review article.
Purpose Within the Thesis
The literature review in a thesis serves to:
- Establish the theoretical and conceptual framework you will work within
- Justify your research questions by showing what remains unknown or contested
- Demonstrate to examiners that you have a thorough, critical grasp of the field
- Provide the scholarly context that validates your chosen methodology
- Show original critical thinking (not just coverage, but insight)
Key Differences from a Standalone Review
| Feature | Dissertation Literature Review | Standalone Review Article |
| Primary purpose | Grounds your original research | Is research in itself |
| Perspective | Written from your position as a researcher-in-progress | Addresses the broader scholarly community |
| Coverage | Selective but rigorous, must justify your study | May be exhaustive (systematic) or broadly representative (narrative) |
| Critical voice | Your analytical stance should be present throughout | Voice depends on review type |
| Evolution | May be revised as empirical work progresses | Fixed once published |
Common Pitfalls for Thesis Writers
- Over-describing and under-analysing: spending too much time recounting what studies found and too little evaluating and connecting them
- Losing the thread: failing to link the discussion back to your own research question
- Neglecting seminal works: skipping foundational theories or landmark studies in your field
- No critical voice: presenting the literature as if all sources are equally valid and beyond debate
- Outdated coverage: not including recent publications. Always check for literature from the last 3–5 years before submission
How to Structure a Standalone Literature Review Article
A standalone literature review (published as a journal article) follows a recognizable structure, the IMRAD format. Review articles are among the most widely cited in academic literature, because they synthesize knowledge that others need as a starting point for their own research.
1. Title and Abstract
- The title should clearly signal the topic, scope, and review type (e.g., “A Systematic Review of…”; “Trends in… A Narrative Review”)
- The abstract should summarize the purpose, scope, methods (for systematic reviews), key findings, and conclusions
2. Introduction
The introduction should:
- Introduce the broader subject and establish its significance
- Narrow to your specific topic and the research question driving the review
- Define the scope: what time period, disciplines, and study types are included and why
- State the type of review being conducted
- Explain what the review contributes and how it is organized (a forecasting statement)
3. Methods (Required for Systematic, Scoping, and Descriptive Reviews)
For rigorous review types, a methods section must include:
- Search strategy: databases, date ranges, and search terms used
- Inclusion and exclusion criteria: clearly stated, ideally in a table
- Screening process: number of reviewers and how conflicts were resolved
- Quality appraisal: tools or criteria used to assess study quality
- Data extraction: what information was extracted and how
- A PRISMA flow diagram is standard for systematic reviews (Preferred Reporting Items for Systematic Reviews and Meta-Analyses)
4. Body / Results
Organize the results section or body according to one of these approaches:
| Organizational Approach | Description | Best For |
| Thematic | Groups sources by topic, concept, or finding | Most reviews; emphasizes intellectual connections |
| Chronological | Traces how knowledge has evolved over time | When historical development is important |
| Methodological | Groups by the methods used in primary studies | When comparing research approaches is the main goal |
| Theoretical | Groups by theoretical framework or school of thought | Theory-heavy disciplines like sociology or education |
Within each section of the body:
- Provide background and definitions as needed
- Group and synthesize related findings. Do not move source by source
- Compare and contrast studies, and evaluate their methodological quality
- Note contradictions and debates
- Write topic sentences that express the analytical point of the paragraph, not just the name of a researcher
5. Discussion
Where present, the discussion section:
- Draws together patterns and conclusions from across the body
- Addresses the implications of knowledge gaps for the field
- Evaluates the overall strength and consistency of the evidence base
- Suggests specific directions for future research
6. Conclusion
The conclusion should:
- Summarize the major contributions of the reviewed literature
- Reiterate what is known, what is contested, and what remains unknown
- Propose concrete avenues for future inquiry
- Avoid introducing new material not covered in the body
7. References
- Maintain a consistent citation style throughout (APA, MLA, Chicago, Vancouver, etc.)
- Cite both direct quotations and paraphrased material
- Use citation management software (Zotero, Mendeley, EndNote) to stay organized
Frequently Asked Questions
How many sources should a literature review include?
There is no universal rule. The appropriate number depends on the scope of your topic, the level of your work, and the type of review. A thesis chapter might include 40–80 sources; a short undergraduate paper might require 10–20; a comprehensive systematic review in medicine can include hundreds of primary studies. The key criterion is not quantity but coverage: have you engaged with the key theories, landmark studies, and recent developments relevant to your question? Check with your supervisor or your institution’s guidelines if you are unsure.
Can I include grey literature such as government reports, policy documents, or dissertations?
Yes, and in some cases you should. Grey literature can be valuable, particularly for scoping reviews or when the academic literature on a topic is sparse. Apply the same critical appraisal standards you would to any source: consider the authority of the producing organisation, the transparency of the methodology, and any potential for bias. Including grey literature is also important for systematic reviews because it helps reduce publication bias (the tendency for studies with positive or significant findings to be published more readily than those with null results).
How do I handle conflicting findings across sources?
Conflicting findings are a feature of the literature, not a problem to be hidden. They are often the most intellectually valuable thing a literature review can surface. Rather than ignoring contradictions, explore possible reasons for them: differences in study populations, methodologies, measurement tools, sample sizes, or time periods. Acknowledge where consensus is strong and where genuine debate exists. This kind of nuanced analysis signals to your reader that you understand the field deeply.
Should I include sources that contradict my thesis or position?
Absolutely. Engaging critically with contradictory evidence is a mark of scholarly integrity. Selectively citing only sources that support your position (sometimes called “cherry-picking”) is a serious academic flaw. A strong literature review acknowledges the full landscape of evidence, including studies that challenge your perspective, and explains why, on balance, the evidence or argumentation supports your position. As the University of Toronto advises, always ask yourself whether you have cited and discussed studies contrary to your own perspective.
What is a qualitative systematic review and meta-synthesis?
A qualitative systematic review is a structured, transparent method for identifying, appraising, and synthesizing findings from multiple qualitative studies on a specific research question. Unlike traditional narrative reviews, it follows an explicit, reproducible protocol for searching, selecting, and evaluating studies, reducing bias and improving rigor.
Meta-synthesis is the interpretive process at its core. Rather than statistically pooling data (as meta-analysis does with quantitative studies), it integrates findings from qualitative research to generate new, higher-order understandings that transcend individual studies. Researchers compare and combine themes, concepts, and interpretations across studies to produce fresh insights.
Common approaches include meta-ethnography (translating studies into one another), thematic synthesis, and grounded theory synthesis. The goal is not simply to summarize but to reinterpret and build richer, more comprehensive theoretical or conceptual understandings of a phenomenon—for example, patients’ lived experiences of chronic illness. This enhances the transferability and practical relevance of qualitative evidence.
When should I update my literature review?
Treat a literature review as a living document during the research process. For a thesis, update it whenever significant new publications appear, and always check for recent work before final submission. For published standalone reviews, the shelf life depends on the pace of the field: in fast-moving areas like biomedicine or artificial intelligence, reviews may become partially outdated within 2–3 years. As a general rule, if your database search was conducted more than 12–18 months ago in an active field, it is worth running updated searches before finalizing your manuscript.
How do I make sure I’m not citing retracted studies or papers from low-quality, predatory journals?
To avoid citing retracted papers or work from predatory journals, start with a few manual checks: look up retraction status through the Retraction Watch Database or Crossref’s metadata, and confirm journals appear in legitimate indexes like DOAJ, Scopus, or Web of Science. Cross-referencing questionable titles against Cabell’s Predatory Reports or the archived Beall’s List also helps, as does watching for red flags like unclear peer review, suspiciously fast publication, vague editorial boards, or aggressive solicitation emails.
But checking every reference by hand is slow and error-prone, especially since retraction status can change after you’ve already cited a paper. That’s where Paperpal’s Reference Checker (https://paperpal.com/tools/reference-checker) helps. It automatically scans your full reference list and flags retracted papers, citations from predatory or low-quality journals, broken or inaccurate references, and formatting inconsistencies.
Running your manuscript through it before submission is worth the few minutes it takes, since even one problematic citation can undermine an otherwise solid paper.
Can I cite AI in a literature review?
Yes, but with care. You must distinguish between using AI as a source vs using AI to actually perform some of the study tasks (e.g., search or screen papers).
Citing AI as a source (references)
- If you’re referencing an AI-generated output itself as content (e.g., quoting or discussing something an AI produced), cite AI output like an Internet or non-standard source, with tool name, developer, version, and access date.
- AI outputs aren’t peer-reviewed or independently verifiable, so they generally shouldn’t be cited as evidence or scholarly sources within the review’s argument, only as what they are (AI-generated text)..
Disclosing AI assistance (process)
- Separate from citations, goes in Methods or an AI disclosure statement.
- State which tasks AI helped perform: search term generation, summarizing papers, identifying themes/gaps, drafting or editing prose, formatting references.
- Note the tool/version used and confirm you verified outputs (accuracy of summaries, correct attributions, no fabricated citations).
- Check your target journal’s or institution’s AI-disclosure policy. Norms vary more for narrative/literature reviews than for systematic reviews (less standardized than PRISMA).
Rule of thumb: cite AI only if referencing its output as content; disclose AI whenever it assisted your review workflow.
General rules for using AI in systematic reviews:
- Don’t outsource judgment: AI can help find, summarize, or organize sources, but inclusion decisions, critical appraisal, and synthesis should remain your own.
- Verify everything: Check AI-generated summaries, quotes, and especially citations against original sources; AI can fabricate or misattribute references.
- Disclose your use: Note where/how AI was used (search, drafting, editing) in your methods or acknowledgments, per journal/institutional policy.
- Cite AI content only when referencing it as content: Not as a substitute for citing the primary literature it was trained on or summarized.
- Maintain a paper trail: Keep records of AI queries/outputs used, so your process is transparent and reproducible if questioned.
References
- Cantero, C. (2019). How to Write a Literature Review. San José State University Writing Center. https://www.sjsu.edu/writingcenter/docs/handouts/Literature%20Reviews.pdf
- Alex, D. (2026). What is a Literature Review? How to Write It (with Examples). Paperpal Blog. https://paperpal.com/blog/academic-writing-guides/what-is-a-literature-review-how-to-write-it-with-examples
- Paré, G., & Kitsiou, S. (2017). Chapter 9: Methods for Literature Reviews. In F. Lau & C. Kuziemsky (Eds.), Handbook of eHealth Evaluation: An Evidence-based Approach. University of Victoria. https://www.ncbi.nlm.nih.gov/books/NBK481583/
- Taylor, D. (n.d.). The Literature Review: A Few Tips on Conducting It. Health Sciences Writing Centre, University of Toronto. https://advice.writing.utoronto.ca/types-of-writing/literature-review/
- Institute for Academic Development, University of Edinburgh. (n.d.). Literature Review. https://institute-academic-development.ed.ac.uk/study-hub/learning-resources/literature-review
This article was originally published on April 3, 2024, and updated on June 3, 2026.
Comment