Table of Contents
What a Literature Review Must Do
A literature review is a critical synthesis of existing scholarship that maps the terrain of a research area, identifies what is known and contested, and, in its most important function, establishes the intellectual gap that your research fills. It is not a catalogue of what others have said; it is an argument about what the field does and does not yet understand.
A well-constructed literature review accomplishes four things simultaneously:
- Demonstrates your command of the field and its key debates
- Positions your work within an existing scholarly conversation
- Justifies your theoretical framework and methodological choices
- Establishes the gap or problem your research addresses
A literature review structured as "Smith (2018) found X. Jones (2019) argued Y. Williams (2020) showed Z." is not a review. It is an annotated bibliography with transition words. Each paragraph of a literature review should advance a synthetic claim about the field, not describe individual studies sequentially.
Review Types
| Type | Purpose | Defining characteristic | Common disciplines |
|---|---|---|---|
| Systematic review | Exhaustively map all evidence on a defined question | Explicit, reproducible search and selection protocol | Health sciences, education, psychology |
| Narrative review | Synthesise and interpret a body of scholarship | Interpretive, expert-guided selection | Humanities, social sciences, theoretical papers |
| Scoping review | Map the nature and extent of evidence on a broad topic | Breadth over depth; no quality appraisal | Emerging research areas, policy contexts |
| Meta-analysis | Statistically pool findings from multiple quantitative studies | Effect size calculation across studies | Clinical research, psychology, economics |
| Integrative review | Synthesise both empirical and theoretical literature | Allows mixed methods studies; theory-building oriented | Nursing, education, management |
Systematic Review Methodology
A systematic review is defined by methodological rigour and reproducibility. Its core requirement is that the search and selection process is transparent enough to be repeated by another researcher. Key components:
- Protocol registration: systematic reviews should be registered on PROSPERO (health sciences) or OSF (social sciences) before data collection begins
- Explicit inclusion/exclusion criteria: defined before searching, covering publication date, language, study design, population, and setting
- Multiple database searching: at minimum two or three major databases (e.g., MEDLINE + CINAHL + PsycINFO for health; Web of Science + Scopus + SSRN for social sciences)
- Dual-reviewer screening: title/abstract and full-text screening conducted by two independent reviewers; disagreements resolved through discussion or third-party arbitration
- Quality appraisal: using validated tools (CASP checklists, Newcastle-Ottawa Scale, GRADE framework)
PRISMA Protocol
PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) provides a standardised framework for reporting the search and selection process. The PRISMA flowchart is now required by most health and social science journals for systematic reviews.
Screening: Records after duplicates removed (n = 2,614) → Records screened by title/abstract (n = 2,614) → Records excluded (n = 2,438) = Records for full-text review (n = 176)
Eligibility: Full-text articles assessed for eligibility (n = 176) → Full-text articles excluded with reasons (n = 131, e.g. wrong population n=58, wrong outcome n=43, not peer-reviewed n=30) = Studies included in synthesis (n = 45)
Included: Studies included in qualitative synthesis (n = 45); studies included in meta-analysis (n = 23)
Boolean Search Strategy
A robust search strategy uses Boolean operators to combine search terms systematically across databases:
- AND: narrows the search (both terms must appear):
resilience AND adolescents - OR: broadens the search (either term may appear):
resilience OR psychological hardiness OR stress resistance - NOT: excludes terms:
resilience NOT ecological NOT engineering - Truncation (*): retrieves all word endings:
adolescen*finds adolescent, adolescents, adolescence - Phrase searching (""): matches exact phrases:
"academic achievement" - Wildcards (?): replaces single characters:
wom?nfinds woman, women
(resilience OR "psychological hardiness" OR "stress resistance") AND (adolescen* OR "young people" OR teenagers) AND (school* OR "academic setting" OR educational) AND (intervention* OR programme* OR training)This string retrieves studies on resilience interventions for adolescents in educational settings. Excluding ecological and engineering uses of "resilience" through the NOT operator applied at the screening stage rather than in the string (to avoid missing interdisciplinary work).
Critical Reading at Scale
A literature review for a dissertation may require reading 60–150 sources. A three-pass system manages this efficiently:
- Pass 1. Triage: Read title, abstract, and conclusion (3–4 minutes). Decide: include, exclude, or uncertain. Assign to reading priority tier.
- Pass 2. Extraction: For included sources, extract: theoretical framework, methodology, sample, key findings, limitations. Record in a synthesis matrix.
- Pass 3. Deep reading: For the 15–25 sources central to your argument, read in full with critical annotations. Note: what does this claim? Is the evidence adequate? What assumptions does it make?
The Synthesis Matrix
A synthesis matrix is a table with sources as rows and analytical dimensions as columns. It transforms individual notes into a visual map of the field, revealing patterns, agreements, contradictions, and gaps that are invisible when reading source by source.
| Source | Theoretical frame | Method | Key claim | Population | Limitations |
|---|---|---|---|---|---|
| Smith (2019) | Social capital theory | Survey, n=450 | Peer networks predict resilience | UK 16–18 | Self-report bias |
| Jones (2020) | Self-determination theory | Interviews, n=24 | Autonomy support mediates academic outcomes | US HE students | Small sample |
| Williams (2021) | Ecological systems | Mixed methods | Institutional context moderates individual resilience | Sub-Saharan Africa | Limited generalisability |
Once populated, the matrix reveals: which theoretical frameworks dominate, which are absent; where geographic or demographic gaps exist; which claims are replicated and which rest on a single study.
Organising Principles
The two main organising principles for literature reviews are thematic and chronological. Thematic organisation is strongly preferred at graduate and doctoral level:
- Thematic: Groups sources by conceptual similarity. Each section addresses a facet of the research question. More analytical; allows genuine synthesis across authors and periods.
- Chronological: Traces the development of a field over time. Useful when intellectual history is the argument; often degenerates into annotated bibliography if not carefully structured.
- Methodological: Groups studies by method type. Used in systematic reviews where comparing evidence quality across methods is the purpose.
Writing the Synthesis
Each paragraph of a synthesised literature review should make a claim about the field, not describe a source. The structure is: claim → evidence from multiple sources → analysis of what the evidence reveals.
Common Literature Review Failures
| Failure | Symptom | Fix |
|---|---|---|
| Descriptive not synthetic | Paragraphs start with an author's name, not a claim | Restructure: claim first, then evidence from multiple sources |
| Undirected coverage | Everything tangentially related is included | Structure every section around a dimension of your research question |
| Neglecting contradictions | Only sources that support your thesis are engaged | Literature reviews must document contradictions and explain them |
| No gap articulated | Review ends without identifying what is missing | The final section must explicitly state the gap your research addresses |
| Grey literature ignored | Policy documents, reports, and practice literature absent | In applied fields, grey literature is legitimate evidence. Exclude it only if justified |
| Out-of-date sources | Most citations are more than 10 years old in a fast-moving field | Check publication dates; supplement with recent systematic reviews for each sub-topic |