One Database vs. Multiple Databases for a Literature Review

If you are choosing between one database and several for a literature review, the practical answer is: start with one database to build your search vocabulary and eligibility criteria, then add sources until new searches stop changing your included set. Cochrane's search guidance points the same way — plan searches from eligibility criteria, use multiple sources, combine free-text terms with subject headings, document the process, deduplicate records, screen in stages, and rerun searches before publication. The decision is not about which option sounds more thorough; it is about where each option fails on your specific task, deadline, and access limits.

This article compares the two approaches on the same criteria — convenience, overlap, unique coverage, deduplication cost, and risk — using a concrete student deadline scenario, then gives a decision table and a combined workflow.

Set the task first. Imagine a master's student with three weeks to finish a course paper on a defined question, a rubric that rewards a documented search strategy, and library access to one major index plus some open sources. The supervisor's feedback on the draft says the search needs to be described more clearly and the reference list needs to show where records came from.

Hold that task constant and compare the two options on five criteria:

• Convenience: how much setup, learning, and time the option requires before you can screen results.
• Overlap: how many records the option returns that you would also find elsewhere.
• Unique coverage: what the option adds that a single source would miss.
• Deduplication cost: the effort of removing duplicate records and linking multiple reports of one study.
• Risk: what can go wrong for your deadline, rubric, or write-up if you choose that option.

Cochrane's guidance is useful here because it treats these as connected steps rather than separate choices: eligibility criteria drive the search, multiple sources are recommended, free-text and subject headings are combined, the process is documented, records are deduplicated, screening happens in stages, and multiple reports of one study are linked. PRISMA 2020 adds a reporting layer — a checklist, expanded explanations, an abstract checklist, and flow-diagram templates — for transparent reporting of search, selection, and synthesis methods.

A single database is the fastest way to start. You learn one interface, one set of filters, and one export format. For a student on a deadline, that convenience is real: you can move from question to a first screenable list quickly.

Where one database fails is coverage and defensibility. A single source cannot show you what it does not index. If your rubric rewards a documented, reproducible search, a one-database strategy gives you little to document beyond that one source. And if your supervisor asks why a relevant paper is missing, you have no second search to point to.

One database is also a weak base for deduplication planning. With only one source, you may not notice that the same study appears in multiple reports — Cochrane explicitly calls for linking multiple reports of one study, which is easier to see when you are comparing across sources.

Best use: the first pass, to build search terms and test eligibility criteria before you commit to a broader strategy.

Multiple databases address the coverage gap that a single source leaves open. Cochrane recommends using multiple sources, and combining free-text terms with subject headings — a practice that matters more when different sources use different indexing vocabularies.

The cost is convenience and deduplication. Each additional source adds a new interface, a new export, and a new set of duplicates. Cochrane's guidance to deduplicate records and screen in stages exists precisely because multi-source searching produces overlap you have to manage. If you add sources without a deduplication step, your screening time grows faster than your unique coverage.

Where multiple databases fail is when they are added without a plan. Searching five sources with inconsistent terms, no documentation, and no deduplication produces a larger pile, not a better review. The risk is not that multiple databases are wrong — it is that undisciplined multi-source searching consumes the deadline you were trying to protect.

Best use: the main search, once your terms and criteria are stable, with a documented deduplication and staged-screening step.

These three criteria move together. More sources usually mean more overlap and more unique coverage at the same time, and both increase deduplication cost. The useful question is not "how many databases?" but "how much unique coverage am I gaining per unit of deduplication effort?"

A practical way to see this on your own task: run your search in your first source, record the included set, then run the same search in a second source and note how many new included records appear. If the second source adds little and costs a full deduplication pass, you have evidence for stopping. If it adds records your first source missed, you have evidence for continuing.

Cochrane's staged screening and linking of multiple reports support this incremental approach: you are not trying to screen everything at once, and you are treating multiple reports of one study as one study rather than as separate findings.

Risk-based selection means choosing sources according to what could go wrong for your specific deliverable, not according to a fixed number.

• Rubric risk: if the rubric rewards a documented strategy, a single undocumented source is the higher risk. Document terms, sources, dates, and screening decisions.
• Coverage risk: if your question spans fields or regions, one source is the higher risk. Add sources until new included records stop appearing.
• Deadline risk: if time is short, adding sources without deduplication is the higher risk. Budget deduplication and staged screening before you add a third or fourth source.
• Reporting risk: if you are writing up search and selection methods, PRISMA 2020's checklist and flow-diagram templates give you a structure for transparent reporting.

NIH's rigor and reproducibility guidance reinforces the same posture: careful assessment of prior knowledge, strengths, weaknesses, knowledge gaps, transparent methods, validation, and reproducibility are core parts of rigorous research practice. Source selection is part of that method, not an afterthought.

Return to the master's student with three weeks, one major index, and supervisor feedback asking for a clearer search description.

One database fails if the paper's question spans literatures the single index covers unevenly, or if the rubric asks for a documented multi-source strategy. The student would finish on time but with a thin methods paragraph.

Multiple databases fail if the student adds sources without deduplication and staged screening. The deadline gets consumed by duplicate records, and the methods paragraph becomes a list of sources rather than a described process.

The failure mode is not the number of databases. It is choosing a number without matching it to the rubric, the question's breadth, and the time available for deduplication.

The combined workflow uses one database first and multiple databases second, with a stopping rule.

1. Define eligibility criteria before searching. Cochrane's guidance starts here because criteria drive terms and sources.
2. Build terms in one database. Combine free-text terms with subject headings, and note which terms retrieve relevant records. You can run and refine these searches through WisPaper's scholarly search interface.
3. Document the first search. Record the source, terms, date, and filters.
4. Add a second source and compare included sets. Note unique records and overlap. The public Scholar Search page explains the entry point for finding academic papers.
5. Deduplicate records and link multiple reports of one study, as Cochrane recommends.
6. Screen in stages — title/abstract first, then full text.
7. Stop adding sources when new searches stop changing your included set, or when the remaining time cannot cover another deduplication pass.
8. Rerun searches before submission to catch recent records.
9. Report the process using PRISMA 2020's checklist and flow-diagram templates if your paper reports a systematic search.

As records accumulate, saving and organizing them in My Library keeps the included set in one place. When you need to check whether a candidate record actually addresses your eligibility criteria, Library QA lets you ask questions based on papers uploaded to My Library or added from search results. If your review grows into a longer project with its own scope and timeline, the Research Projects area is where you can set that up.

For the student scenario, the concrete submission decision is: submit with a two-source documented search, not a one-source search and not an unbounded multi-source search. Use the first database to build and document terms, add a second source to test unique coverage, deduplicate and screen in stages, rerun before submission, and describe the process in the methods section. If the second source adds no new included records, say so — that is a documented decision, not a gap.

If the paper reports a systematic search, follow PRISMA 2020's checklist and flow-diagram templates so the search, selection, and synthesis methods are transparent. If the paper is a course paper with a lighter rubric, the same two-source logic still protects you: it gives you a methods paragraph you can defend and a stopping rule you can explain.