A literature search strategy is a planned, transparent series of decisions and actions that find, select, appraise and record the evidence needed to answer a specific question. It requires (1) a clear question and scope, (2) selection of appropriate databases and other sources, (3) translation of the question into keywords and controlled vocabulary, (4) search construction with Boolean operators and limits, (5) systematic screening and appraisal, and (6) detailed records that enable reproducibility and audit. Well-documented strategies reduce bias, enable replication and support trustworthy synthesis; they are not a substitute for methodological or ethical oversight when live research or sensitive data are involved — consult institutional ethics and supervisory support for such work.
Why a planned literature search strategy matters
A planned strategy ensures that searching is systematic rather than ad hoc. It:
- Clarifies the question and scope so results are relevant.
- Reduces selection bias by documenting choices (databases, dates, languages).
- Facilitates reproducibility and auditability of the search.
- Supports robust appraisal and transparent reporting (for example in reviews, theses or policy briefs).
Good practice in search planning aligns with standards for evidence synthesis and reproducible research (see Barroga and Matanguihan, 2022; Gionfriddo et al., 2024).
Define purpose and scope
Begin by specifying the purpose of the search: is it a scoping search to map literature, a systematic search for a review, a rapid evidence check, or background reading for a proposal? The scope determines the depth and breadth of the strategy, resources required and the acceptable trade-offs between sensitivity (finding as many relevant items as possible) and specificity (excluding irrelevant items).
Key elements to define:
- Question framed (PICO, PCC, PICo or other suitable structure).
- Population, concepts, comparator(s) and outcomes where relevant.
- Study designs to include (e.g. RCTs, cohort studies, qualitative).
- Time limits, language limits and geographical limits, if any.
- Types of sources (peer‑reviewed literature, preprints, grey literature).
For quantitative research planning and reporting, consult guidance on study design and reporting to align search aims with statistical considerations (USC Libraries, 2026).
Choosing databases and other sources
Match sources to the disciplinary scope of your question:
- Health and medicine: MEDLINE/PubMed, Embase, CINAHL, Cochrane Library.
- Social sciences and education: PsycINFO, ERIC, Scopus, Web of Science.
- Multidisciplinary: Scopus, Web of Science, Google Scholar (for breadth).
- Grey literature: institutional repositories, preprint servers, clinical trial registers.
Table 1 gives a simple selection matrix to help choose where to search.
Table 1 — Database selection matrix
| Domain | Strengths | Typical coverage | When to include |
| PubMed/MEDLINE | High-quality indexing, MeSH controlled vocabulary | Biomedical journals, clinical studies | Clinical and biomedical questions |
| Embase | Extensive European and pharmacological coverage, Emtree terms | Drug studies, conference abstracts | Pharmacology, drug safety |
| CINAHL | Nursing and allied health focus | Nursing, rehabilitation | Clinical practice and allied health |
| PsycINFO | Psychological and behavioural literature | Mental health, behaviour studies | Psychology and behaviour-focused questions |
| Scopus / Web of Science | Broad multidisciplinary indexing; citation tracking | Cross-discipline coverage | Citation analysis, broad searches |
| Grey literature sources | Trials registers, theses, reports | Non-peer-reviewed, reports, preprints | Minimise publication bias |
No single database covers everything. Use at least two complementary sources for most comprehensive searches.
Developing keywords and controlled vocabulary
Translate each element of your question into:
- Free-text keywords: synonyms, abbreviations, spelling variants (e.g. “behaviour/behavior”, “older adult/elderly”).
- Controlled vocabulary: MeSH (PubMed), Emtree (Embase), CINAHL Headings — these standardise subject indexing.
Table 2 illustrates mapping of a sample concept.
Table 2 — Example: mapping a concept to search terms
| Concept | Controlled vocabulary | Free-text keywords and variants |
| Older adults | MeSH: Aged; Emtree: elderly | older adult*, elderly, “older people”, “senior* citizen*” |
| Physical activity | MeSH: Exercise; Emtree: physical activity | physical activity, exercise, walking, “green exercise” |
| Mental wellbeing | MeSH: Mental Health; Emtree terms | mental health, wellbeing, “psychological wellbeing”, depression, anxiety |
Best practice: test controlled terms and free-text words iteratively, and document each version. Note that controlled vocabulary changes over time; record the date of the search and version of the database.
Constructing Boolean searches and limits
Combine terms with Boolean operators:
- OR to combine synonyms and increase sensitivity.
- AND to intersect different concepts and increase specificity.
- NOT sparingly and with caution (can exclude relevant studies).
Use truncation and phrase searching where supported. Explicitly document filters and limits applied (date ranges, languages, human/animal, study design filters) because these materially affect results and reproducibility.
Example (conceptual): (“older adult*” OR elderly OR “senior citizen*”) AND (exercise OR “physical activity”) AND (“mental health” OR wellbeing OR depression)
Record the exact syntax used for each database (different databases require different syntax for truncation, adjacency, field limits).
Screening and selection
Screening is typically two-stage:
- Title and abstract screening against inclusion/exclusion criteria.
- Full-text screening for final eligibility.
Use piloted screening forms and, where possible, dual screening (two independent reviewers) with a mechanism for resolving conflicts (third reviewer or consensus meetings). Keep a log of reasons for exclusion at full-text stage to support transparent reporting and any future audit or replication.
When defining screening rules, a review of dissertation methodological rigour failures can help identify avoidable inconsistencies between question, criteria and inference.
Source appraisal and critical evaluation
Appraise included studies for relevance and quality using appropriate tools (for example, Cochrane Risk of Bias for RCTs, CASP for qualitative studies, Newcastle–Ottawa Scale for observational studies). Appraisal should be proportionate to the planned synthesis — critical appraisal informs weighting of evidence rather than simple exclusion (except where methodological flaws make findings uninterpretable).
Be explicit about appraisal criteria and report them in any write-up. When synthesising evidence, reflect the appraisal in interpretation and recommendations.
Recording, audit trail and reporting
Keeping a transparent, reproducible record is central to a literature search strategy. Essential items to record:
- Exact search strings for each database and date run.
- Database platforms and interface (e.g., Ovid MEDLINE, EBSCOhost CINAHL).
- Version of controlled vocabulary (e.g. MeSH 2025 update).
- Number of hits retrieved per search run and screened.
- Screening decisions and reasons for exclusion.
- Full bibliographic records and where full text was obtained.
A practical search record template reduces errors and aids reproducibility. Example fields: database, platform, date run, search string, hits, deduplicated hits, screening notes, download location.
For practical pointers on quantitative writing, sampling and data handling that intersect with search planning, consult discipline-level guidance on quantitative methods (USC Libraries, 2026). Consistent citation records can be supported by this guide to academic citation styles.
Fictional planning application
Project title: “Green walking and mental wellbeing in community-dwelling adults aged 65 and over (2000–2025): a scoping literature search”
Purpose: To map the volume and types of evidence addressing associations between walking in green spaces and mental wellbeing in older adults, to inform a subsequent systematic review proposal.
PICO-style frame:
- Population: Community-dwelling adults aged 65+
- Intervention/exposure: Walking in green spaces (parks, woodland)
- Comparator: Non-green walking, inactivity or other activities
- Outcomes: Mental wellbeing measures (validated scales), mood, depression, anxiety
- Study designs: Observational, interventional, qualitative studies (to capture experience)
Planned databases:
- PubMed/MEDLINE (MeSH: Aged; Parks; Walking)
- Embase (Emtree equivalents)
- PsycINFO (behavioural and wellbeing measures)
- Scopus (citation breadth)
- ClinicalTrials.gov for ongoing trials
- Select grey literature (local public health reports)
Example search approach on PubMed:
- Controlled vocabulary: Aged[MeSH] AND Parks[MeSH] OR Walking[MeSH]
- Free-text: (“older adult*” OR elderly) AND (“green walk*” OR “walking in green” OR park* OR woodland) AND (“mental health” OR wellbeing OR depression OR anxiety)
- Limits: 2000–2025, English, humans
Screening plan:
- Two independent reviewers for title/abstract; pilot of 100 records to refine inclusion criteria.
- Full-text screening with reasons recorded for exclusions.
- Data extraction template to include study design, setting, sample, exposure definition, outcome measures, key results and quality appraisal.
Expected outputs:
- PRISMA flow diagram for searches and screening.
- Tabulated map of evidence by study design and outcomes.
- Narrative synthesis with identified gaps to inform a systematic review protocol.
Note: This is a fictional planning application for demonstration only. For live research involving people or sensitive data, follow institutional ethics guidance and supervisory support.
Practical sequence
A pragmatic stepwise sequence to implement a literature search strategy:
- Clarify the question and scope; choose an appropriate question framework (PICO, PICo, PCC).
- Pilot keywords and controlled terms for each concept; create a working thesaurus.
- Select 3–5 core databases and 1–2 grey literature sources relevant to the topic.
- Translate search into each database’s syntax; document every string and date run.
- Run searches, export bibliographic records, and deduplicate.
- Pilot title/abstract screening with a subset to test inclusion criteria.
- Conduct full screening with pre-defined forms and record reasons for exclusion.
- Appraise included sources using appropriate quality tools.
- Extract data into a structured database and create summary tables.
- Record the entire process in a searchable audit trail and prepare a reproducible methods appendix for reporting.
This sequence supports transparency and reproducibility; consider preserving search histories and database exports in an institutional repository or secure project folder.
Search documentation and review
A defensible search record includes the purpose, databases and platforms, date searched, exact strings, filters, screening rationale and any citation searching. This record supports later review, clarification and responsible updating; it also prevents a reader from treating a convenient first-page result set as comprehensive evidence.
| Record element | Why it matters |
| Search string | Makes concept choices visible. |
| Database/platform | Explains coverage and functions. |
| Date and limits | Shows the currency and boundaries. |
| Screening decision | Connects inclusion to the question. |
Critical limitations and safeguards
Limitations
- Coverage bias: No collection is exhaustive; database indexing and publisher choices affect what is discoverable.
- Publication and language bias: Published, English-language and positive-result biases can skew synthesis.
- Indexing lag and evolving controlled vocabulary: Terms change over time; older works may be mismatched.
- Searcher error and variability: Search construction and screening may introduce human error or subjective judgment.
- Resource constraints: Time and database access limit comprehensiveness.
Safeguards
- Use multiple databases and grey literature sources to reduce coverage and publication bias.
- Keep full, dated records of search strings, platforms and controlled vocabulary versions to help replication.
- Use independent dual screening and transparent conflict resolution procedures.
- Pilot and refine search and screening tools before full implementation.
- Clearly report limitations in any public-facing output, including potential biases and the dates and scope of the search.
Ethics and data protection
- This guidance is general educational information only. For searches that will involve recruitment, participant data or the handling of sensitive datasets, seek institutional ethics and supervisory support before collecting or processing data.
- Store personal data (if any) according to institutional data protection policies and legal requirements (for example UK GDPR).
FAQs
Q1: How many databases should I search?
A1: There is no fixed number. Aim for at least two complementary databases (one subject-specific and one multidisciplinary) for most topics; include additional specialised databases and grey literature sources when the question requires comprehensiveness. Document the rationale for chosen sources.
Q2: When should I use controlled vocabulary versus free-text searching?
A2: Use both. Controlled vocabulary captures subject-indexed records that use standardised terms, while free-text picks up new terms, phrases, brand names and terms not yet indexed. Combining both improves sensitivity.
Q3: Do I need to report the exact search strings?
A3: Yes — reporting exact search strings, date ranges and platform details is essential for reproducibility and transparency. Any published synthesis should include the full search history, ideally as an appendix.
Q4: How do I manage duplicate records and multiple versions of the same study?
A4: Use reference management software to detect duplicates after exporting from different databases. For multiple versions (preprint and published article), select the version most appropriate for your synthesis (usually the peer-reviewed final version) but document all versions located and justify choices.
For related decision support, use academic citation styles and dissertation-methodology support.
References
Barroga, E. and Matanguihan, G.J. (2022) ‘A practical guide to writing quantitative and qualitative research questions and hypotheses in scholarly articles’, Journal of Korean Medical Science, 37(16), e121. Available at: https://doi.org/10.3346/jkms.2022.37.e121 (Accessed: 26 August 2026).
Gionfriddo, M.R., McClendon, C., Nolfi, D.A., Kalarchian, M.A. and Covvey, J.R. (2024) ‘Back to the basics: Guidance for designing good literature searches’, Research in Social and Administrative Pharmacy, 20(4), pp. 463–468. Available at: https://doi.org/10.1016/j.sapharm.2024.01.009 (Accessed: 26 August 2026).
USC Libraries (2026) ‘Quantitative methods’, Organizing Your Social Sciences Research Paper. Available at: https://libguides.usc.edu/writingguide/quantitative (Accessed: 26 August 2026).