To do a literature review, define a question, search widely, screen, synthesize, and write with clear structure.
A strong review shows what is known, what is missing, and where your project fits. You’re building a map of prior work, not a data dump. The process below works for thesis chapters, journal submissions, and capstone projects across fields.
Review Types At A Glance
Pick a review style that matches your aim and deadline. Here’s a quick picker to set scope and method.
| Type | Best For | Hallmark Methods |
|---|---|---|
| Narrative | Broad overviews and theory building | Flexible search, thematic grouping, critical commentary |
| Systematic | Answering a specific question with traceable steps | Protocol, database strings, inclusion rules, flow diagram |
| Scoping / Mapping | Sizing a field or clarifying concepts | Wide search, charting key features, gap spotting |
Conduct A Literature Review: Start Strong
This section lays out the full run from question to polished write-up. Use the steps as a checklist. Adjust depth to fit your course, journal, or committee rules.
1) Frame A Clear Question
State the main aim in one line. Narrow the who/what/where/when. In health and social science, a PICO or PICo frame can help. In humanities, name the lens, period, and corpus. Precision here saves weeks later.
2) Set Scope, Outcomes, And Boundaries
Define inclusion and exclusion rules before you search. Typical filters: population, setting, design, language, time span, and outcomes of interest. Write these in your notes so screening stays consistent.
3) Build Search Strings
List keywords and controlled terms (e.g., MeSH) for each concept. Use Boolean logic with AND/OR, phrase marks, and truncation. Draft strings once, then tune them per database quirks. Save every final string you run.
4) Choose Databases And Sources
Mix subject databases (like MEDLINE, PsycINFO, ERIC), cross-disciplinary tools (Scopus, Web of Science), and grey literature where relevant (theses, preprints, policy portals). Scan reference lists from key papers to catch strays.
5) Manage References From Day One
Pick a manager (Zotero, EndNote, Mendeley). Create folders for “Found,” “Screening,” “Included,” and “Excluded.” Attach PDFs, tag themes, and add short notes with study aims, design, sample, measures, and main findings.
6) Screen Titles And Abstracts
Apply your rules to titles and abstracts first. Work fast, but record reasons for removal in broad buckets such as “wrong topic,” “wrong design,” or “not primary data.” Move survivors to full-text screening.
7) Screen Full Text
Read each candidate with your rules in front of you. Log a short reason if you exclude at this stage. For systematic work, a flow diagram helps you track counts and decisions. The PRISMA 2020 flow diagram is the standard layout for that record-keeping.
8) Extract Core Data
Create a table that collects citation, setting, participants, methods, measures, outcomes, and caveats. Add columns for any moderators you care about (age band, sector, region). Keep it lean enough to fill, rich enough to compare.
9) Appraise Study Quality
Pick an appraisal tool that suits the designs you included. Rate risk of bias or strength of evidence with simple rubrics. Note design limits that could tilt findings. Mention these limits later when you weigh claims.
10) Cluster Themes And Build The Thread
Lay studies into clusters: methods, theories, populations, time periods, or result patterns. Look for points of agreement, disagreement, and blind spots. The thread of your narrative should follow a logic (chronology, concept, or method) that fits the field.
11) Synthesize
Bring studies into conversation. Compare aims, measures, and effect directions. Flag where contexts differ. If data allow, you may add a small meta-summary (counts of findings) or prepare the ground for a later meta-analysis.
12) Write With A Clear Structure
Use short sections that match your thread. Start each part with a signal line that states the claim of that section. Then bring in evidence from your set. End sections with a one-line takeaway that points forward.
Suggested Section Outline
Title And Opening
State the topic and audience. One to two short paragraphs that set the scene and define terms used in a specific way.
Methods (Transparent And Brief)
Report where you searched, date ranges, top strings, inclusion rules, and how many records made it into the set. If you used a structured approach, the PRISMA 2020 checklist gives a clean list of items to report for systematic work.
Results Of The Search
Give an overview of the set: number of studies, designs, settings, and date spread. If relevant, add a small figure or table in your thesis document to show clusters and counts.
Thematic Synthesis
Write one subsection per theme. Keep paragraphs tight. Compare claims, measures, and contexts. Point out gaps that matter to your question.
Implications And Gaps
Tie themes back to the question. Name practical takeaways for scholars or practitioners. Propose up to three next steps that follow directly from the gaps you found.
Write Crisp Paragraphs That Scan Well
Readers skim first. Use short paragraphs and clear topic lines. Keep quotes rare. When you paraphrase, carry the exact meaning and credit the source in line with your style guide (APA, MLA, Chicago, IEEE).
Source Quality: Fast Tests That Work
Run a quick credibility pass on every new source. Currency: date fits your field’s pace. Relevance: the study actually answers your aim. Authority: credible author or venue. Accuracy: methods and data are sound. Purpose: research, not a sales pitch. Many librarians teach this under the CRAAP shorthand; you can find primers through university guides.
Smart Search Habits
Chain From A Few “Seed” Papers
Start with two or three strong papers. Use “cited by” and “references” to branch forward and backward. This often surfaces clusters that raw keyword searches miss.
Balance Precision And Breadth
Begin wide across databases, then tighten with filters. Keep alerts running on your main strings until you submit. Late finds can go into a “recent additions” note if they don’t shift your claims.
Document Every Step
Keep a living note with dates, strings, and counts. This protects you against “did I run that already?” errors and helps you rerun the search later.
Search Log Template You Can Copy
Drop this into a spreadsheet or note. Fill it every session.
| Date | Source / Database | Query & Notes |
|---|---|---|
| 2025-10-22 | Scopus | “term A” AND “term B” AND (“context” OR “setting”); 512 hits; added filters: year >= 2015 |
| 2025-10-22 | MEDLINE | MeSH[Concept1] AND MeSH[Concept2]; 231 hits; excluded case reports |
| 2025-10-23 | Grey literature | Policy portal + thesis repository; 38 records; 6 kept after screen |
Writing Moves That Raise Clarity
Lead With Claims
Start each paragraph with the point you’re making. Bring evidence next. Close with why that point matters for your question.
Use Parallel Structure
When you compare studies, line up like with like: sample size vs. sample size, measure vs. measure, context vs. context. Readers can spot the differences instantly.
Blend Summary And Critique
Summarize only what you need, then add a short judgement about method or fit. Keep the tone neutral and fair.
Referencing Without Headaches
Pick your style early. Set the manager to that style. Clean each record as you go so your reference list builds itself. Check journal rules for title case, DOI format, and preprint handling before submission.
Ethics And Transparency
Cite datasets, preprints, and tools you used. If you used translation, report how. If you screened with a partner, mention how you resolved differences. Small notes like these raise reader trust.
Time Savers That Work
- Saved Searches: Store strings inside each database and turn on alerts.
- Tags: Tag by theme, method, and population so clusters appear fast.
- Templates: Reuse your extraction and screening notes for the next project.
- Short Writing Sprints: Draft mini-sections right after you finish a theme while it’s fresh.
Common Pitfalls And Easy Fixes
Scope Creep
When new threads appear, park them in a “future work” list. Protect the main aim you set at the start.
String Myopia
Relying on one platform or one phrasing hides papers. Vary strings and hop across databases.
Summary Without Synthesis
Paragraphs that only retell add little. Always add a line that shows the link to your question or the contrast with a nearby study.
Sample Mini-Protocol (You Can Adapt)
Question
“How do workplace schedule changes relate to burnout among hospital nurses?”
Inclusion Rules
Primary studies on hospital nurses; burnout measured with a validated tool; 2015-present; English; OECD countries. Exclude editorials and single-case anecdotes.
Databases
MEDLINE, CINAHL, Scopus, Web of Science; grey sources: thesis portals and government reports.
Strings (One Per Database)
Concepts: nurses; burnout; scheduling; hospital. Combine subject headings with keywords. Save final strings inside each platform.
Screening And Extraction
Two passes for titles/abstracts; one pass full-text; log removals by reason. Extract study details, measures, and outcome direction. Rate study quality with a simple rubric matched to design.
Reporting
Summarize the set with counts and a short theme map. Describe patterns across settings and measures. Name gaps and lay out next steps that fit those gaps.
Voice And Style Tips
Use plain verbs. Keep sentences under 20–24 words when you can. Vary rhythm to avoid monotony. Cut hedges and filler. Replace vague words with concrete ones. Where you must use terms of art, define them once, then keep usage consistent.
Where To Learn The Craft
If you want a concise writing primer from a trusted source, the Purdue Writing Lab has a clear guide on writing a literature review with examples of organization and tone. For structured review reporting, the PRISMA 2020 resources collect checklists and diagrams used by many journals.
One-Page Checklist
- Question: One clear line; scope defined.
- Rules: Inclusion and exclusion set in advance.
- Search: Strings built, saved, and adapted per database.
- Sources: Multiple databases plus reference chaining and grey where needed.
- Screen: Title/abstract pass, then full-text pass; reasons logged.
- Extract: Table of methods, measures, outcomes, and caveats.
- Appraise: Risk of bias or quality tool matched to design.
- Cluster: Themes decided; map of how studies group.
- Synthesize: Claims compared with context; gaps named.
- Write: Methods short and transparent; body arranged by a clear thread; tight paragraphs.
- Finish: Style guide set; references clean; links and figures labeled; date ranges stated.