1. Define the Purpose, Review Type and Scope
Before searching, clarify what the literature review must accomplish inside your dissertation. In many projects, it establishes the theoretical and empirical context for a later methodology and results chapter. In others, the literature review is itself the principal method, as in a systematic, scoping or integrative review.
Your review type affects the question, search depth, screening process, documentation and writing structure. Do not label a review “systematic” simply because it is organized. Systematic reviews require a preplanned and reproducible method for identifying, selecting, appraising and synthesizing evidence. PRISMA 2020 is primarily a reporting guideline for systematic reviews; it should be used when it fits the research design, not forced onto every narrative dissertation review.
| Review approach | Typical purpose | Expected process |
|---|---|---|
| Narrative or traditional review | Explain, compare and critique a body of scholarship around a focused problem. | Transparent but flexible searching, deliberate selection and thematic synthesis. |
| Systematic review | Answer a tightly defined question through comprehensive and reproducible evidence synthesis. | Protocol, explicit eligibility criteria, documented searches, screening, appraisal and reporting. |
| Scoping review | Map the extent, characteristics and gaps of a broad or emerging evidence base. | Structured search and screening with emphasis on coverage, concepts and research gaps. |
| Integrative review | Combine diverse theoretical and empirical literature to develop broader understanding. | Explicit selection, comparison and synthesis across different study types. |
Define five boundaries before searching
- The central research problem and the purpose of reviewing prior scholarship.
- The population, setting, discipline, concept, intervention or phenomenon of interest.
- The date range, language, geography and publication types that may be included.
- The level of evidence or study designs relevant to the question.
- The chapter word count, dissertation timetable and available database access.
Students who are unsure what their brief requires can begin with an assignment clarity review before committing time to a search process that may not fit the assessment.
2. Turn the Topic Into a Focused Review Question
A topic identifies an area; a review question determines what evidence you need. “Artificial intelligence in education” is a topic. “How does generative AI-supported feedback influence self-regulated writing among first-year university students?” is a question that suggests concepts, populations and relationships to investigate.
Start with a scoping search to learn the vocabulary used by researchers. Notice subject headings, alternative spellings, related theories, common measures and recurring debates. Then narrow the question until it is broad enough to locate evidence but focused enough to support a coherent synthesis.
A focused question should clarify
- What phenomenon, relationship, intervention or experience is being examined?
- Which population, context, discipline or setting matters?
- What kind of evidence would answer the question?
- Which boundaries are defensible rather than merely convenient?
- What will the review help the dissertation establish or justify?
If nearly every paper in the discipline appears relevant, the question is not yet doing enough analytical work. Narrow the relationship, context, population, time period or evidence type.
3. Build a Search Strategy Before Opening Multiple Databases
A search strategy is a planned method for retrieving relevant evidence. It should combine the main concepts in your question with synonyms, spelling variants, abbreviations and database-specific subject headings. Search terms that work well in one database may need to be adapted for another.
Use OR to connect alternatives within the same concept and AND to connect different concepts. Quotation marks may preserve exact phrases, while truncation symbols can capture word variations where the database supports them. Record every search string, database, date and limit that contributes to the final evidence set.
Search-strategy quality checks
- Each major concept from the review question appears in the search plan.
- Synonyms reflect the terminology used in the discipline, not only everyday language.
- Subject headings are tested alongside keywords where databases provide them.
- Limits such as date, language and document type are justified and documented.
- The search retrieves known relevant papers without producing an impossible volume of noise.
4. Choose Databases and Source Types That Fit the Discipline
No single database covers every field equally well. Use your university library’s subject guides to identify disciplinary databases, then add a multidisciplinary database where appropriate. Google Scholar can help with citation tracking and discovery, but it should not be the only search environment for a rigorous dissertation review.
Source selection should follow the review question. Peer-reviewed journal articles may provide current empirical evidence, while books can be essential for theory and historical context. Theses, conference papers, government reports, professional guidance and other grey literature may reveal emerging evidence or reduce publication bias in some review designs.
Use subject-specific and multidisciplinary databases selected for your discipline.
Review reference lists and later papers that cite important studies.
Include reports, theses or policy sources when the review question justifies them.
Do not download sources without organizing them. Export accurate records into a reference manager such as Zotero, then check titles, authors, publication dates, journal details and DOIs. A reference manager reduces repetitive formatting work, but it does not guarantee correct metadata.
5. Screen Sources for Relevance and Evaluate Their Quality
Screening prevents the review from becoming a collection of everything you happened to find. Create inclusion and exclusion criteria before final selection. Criteria may involve population, setting, methodology, publication date, language, evidence type, geographical scope or direct relevance to the review question.
For a structured review, screening commonly begins with titles and abstracts before moving to full-text assessment. Record the reason for excluding a full-text source when methodological transparency is required. In a narrative review, the process may be more flexible, but the selection logic should still be explainable.
Evaluate more than the journal name
- Does the source directly address the review question or only mention the topic?
- Are the design, sample, measures and analysis appropriate for the claim being made?
- Are limitations, uncertainty and alternative explanations acknowledged?
- Is the evidence current enough for the aspect of the topic under discussion?
- Does the source contribute a theory, method, finding, contradiction or gap worth synthesizing?
A methodologically strong article may be irrelevant to your question, while a highly relevant source may contain serious limitations. Record both judgments instead of collapsing them into a single impression.
6. Create an Evidence Matrix and Take Structured Notes
An evidence matrix turns reading into comparable data. Instead of storing unrelated annotations across dozens of PDFs, use consistent columns to record information that matters to the review question. The matrix may be a spreadsheet, table or structured reference-manager note.
Useful evidence-matrix fields
- Full citation and unique source identifier.
- Purpose, research question or theoretical perspective.
- Population, setting, sample and study design.
- Methods, measures, data sources and analytical approach.
- Main findings relevant to your review—not every result in the paper.
- Limitations, quality concerns and contextual boundaries.
- Theme codes, agreements, contradictions and possible research gaps.
- Page numbers for quotations or claims that may be cited later.
Write notes in your own words and distinguish direct quotations clearly. This reduces accidental patchwriting and makes later synthesis easier. The University of Sydney’s guidance on evidence extraction similarly recommends deciding which data elements are needed and using a consistent extraction form.
| Source | Method and context | Relevant finding | Limitation | Theme code |
|---|---|---|---|---|
| Author A, 2024 | Survey of first-year students | Feedback frequency related to confidence | Single institution; self-report data | Feedback and self-efficacy |
| Author B, 2025 | Qualitative interviews | Students valued explanatory tutor dialogue | Small volunteer sample | Relational support |
| Author C, 2023 | Controlled tutoring intervention | Skill gains depended on active practice | Short follow-up period | Guided practice |
7. Build Themes From Patterns, Contradictions and Gaps
Themes are analytical categories that help answer the review question. They should emerge from repeated relationships across sources, not from arbitrary labels created before reading. A theme may capture a shared mechanism, a recurring outcome, a methodological divide, a contextual difference or an unresolved debate.
Begin by coding relevant observations in the evidence matrix. Combine related codes, test whether each group is internally coherent and compare it with competing interpretations. Strong themes usually contain more than agreement; they reveal where evidence changes across populations, methods, settings or theoretical assumptions.
Questions that strengthen a theme
- Which sources support this pattern, and which sources challenge it?
- Does the pattern change by population, context, method or measurement?
- Is the theme descriptive, or does it explain why findings differ?
- What evidence is missing, underrepresented or methodologically weak?
- How does the theme contribute to the dissertation’s research problem?
8. Synthesize the Evidence and Write a Critical Argument
Synthesis means combining sources to explain a pattern, debate or gap. Avoid the “one paragraph, one article” structure in which each study is summarized and abandoned. Instead, begin paragraphs with an analytical claim, bring several relevant sources into conversation and explain what their agreements or differences mean.
A literature-review paragraph often moves through five functions: establish the theme, present converging evidence, introduce a contradiction or boundary, evaluate the explanation and connect the insight to the review question. Not every paragraph needs every function, but the writing should consistently show relationships across sources.
Weak summary versus stronger synthesis
Source-by-source summary
“Smith found that feedback improved confidence. Jones studied online tutors. Patel reported positive student perceptions.”
Problem: The sources appear next to each other, but the reader is not shown how the findings relate.
Thematic synthesis
“Across survey, interview and intervention studies, feedback appears most strongly associated with confidence when students are required to explain and apply the guidance rather than receive corrections passively.”
Strength: The sentence identifies a cross-study relationship and a possible mechanism.
Critical qualification
“However, most studies rely on short-term self-report measures, making it difficult to determine whether confidence persists or transfers to independent assessment.”
Strength: The limitation applies across the evidence base rather than attacking one source in isolation.
Research-gap connection
“This creates a need for longitudinal research examining whether tutor-supported confidence predicts later independent performance.”
Strength: The gap follows logically from the synthesis and helps justify the dissertation.
Organize the chapter around themes, theories, methods, chronology or debates—whichever structure best answers the review question. Use clear transitions to show why the discussion moves from one section to the next. Students can request academic draft review and feedback for prioritized comments on structure, synthesis and clarity while keeping authorship of the dissertation.
9. Cite, Verify and Audit Every Piece of Evidence
Citation accuracy begins during reading, not at the end. Record page numbers with notes, preserve full bibliographic metadata and distinguish your paraphrase from direct wording. Every claim should be supported by the source cited, and every source cited in the text should appear in the reference list.
Use the required style guide rather than relying on a generator’s default output. Reference managers can organize sources and produce citations, but imported records may contain capitalization errors, missing issue numbers, incorrect publication types or incomplete author lists. Crossref Metadata Search can help verify DOI and publication details for many scholarly works.
Run a four-part citation audit
- Claim audit: open the source and confirm that it supports the exact statement.
- Location audit: add page or section details where the citation style or quotation requires them.
- Metadata audit: verify authors, year, title, journal, volume, issue, pages and DOI.
- Match audit: ensure every in-text citation has a reference entry and every reference entry is used.
Software can format and store records, but only the writer can confirm that the cited source supports the interpretation presented in the dissertation.
For a structured review of in-text citations, reference-list consistency and source metadata, use reference validation and citation support.
10. Dissertation Literature Review Final Checklist
- The review question, purpose and scope are clear and consistent.
- The chosen review approach matches the dissertation methodology and instructions.
- Search terms, databases, dates and limits are documented appropriately.
- Inclusion and exclusion decisions follow explicit criteria.
- Source quality and relevance have been evaluated separately.
- The evidence matrix records comparable details from every included source.
- Themes are supported by several sources and tested against contradictory evidence.
- Paragraphs synthesize relationships rather than listing studies one by one.
- Research gaps follow logically from the reviewed evidence.
- Every claim, quotation, citation, reference entry and DOI has been checked.
11. Frequently Asked Questions
How many sources should a dissertation literature review include?
There is no universal number. The appropriate evidence base depends on the discipline, degree level, question, review type, word count and availability of research. Aim for sufficient coverage of the important theories, methods, findings and debates rather than an arbitrary total.
How long should the literature review chapter be?
Follow your department’s dissertation guidance and supervisor’s expectations. The chapter must be long enough to establish context, synthesize evidence, evaluate limitations and justify the research gap without becoming a general history of the entire field.
Can Google Scholar be used for a dissertation literature review?
Google Scholar is useful for discovery and citation tracking, but it should normally be combined with suitable library databases. Record where each source was found and use database-specific tools when the review requires a reproducible search.
Should every dissertation literature review use PRISMA?
No. PRISMA 2020 is a reporting guideline designed primarily for systematic reviews. Use it when the methodology requires systematic reporting and follow any discipline-specific guidance. A narrative review still needs transparency, but it may not require a PRISMA flow diagram.
What is the difference between summarizing and synthesizing sources?
Summary reports what an individual source says. Synthesis combines several sources to explain agreements, contradictions, patterns, mechanisms, contextual differences or gaps that answer the review question.
When should I start writing the literature review?
Begin drafting once you have a focused question, a workable search and enough structured notes to identify preliminary themes. Continue searching and refining as needed, but do not wait until every possible source has been read before testing the chapter structure.
References and Further Reading
The process in this guide is informed by current university-library guidance, official reporting standards, citation guidance and research-management documentation.
- Cornell University Library — Demystifying the Literature Review
- University of Sydney Library — Systematic Reviews
- University of Sydney Library — Search Strategies
- University of Sydney Library — Screening Results
- University of Sydney Library — Extracting and Appraising Data
- University of Sydney Library — Analysing and Interpreting Data
- PRISMA — PRISMA 2020 Statement and Checklists
- APA Style — Basic Principles of Citation
- Zotero Documentation — Quick Start Guide
- Crossref — Metadata Search for Scholarly Works
