01 · The Question
What might disappear when you search only one database?
You search a major research database, retrieve hundreds or perhaps thousands of records, and begin screening. It feels comprehensive. But the size of the result set does not tell you what the database does not contain.
Bibliographic databases differ in the journals and other sources they cover, the disciplines they emphasize, the geographic and linguistic literature they represent, the records they index, and the search functionality they provide. JBI consequently states that there is no single source for evidence synthesis and recommends choosing databases that are relevant to both the topic and, where applicable, geography.
The practical question is therefore not simply whether you searched a good database. It is whether the sources you searched collectively give the relevant literature a reasonable opportunity to be found.
03 · What You Need to Know
Why does database choice change the literature you find?
A database is a window onto the literature, not the literature itself
Researchers sometimes talk about searching “the literature” when what they have actually searched is one representation of it. A bibliographic database contains records selected and organized according to its own scope and indexing practices. Another database may represent overlapping literature, but the overlap will rarely be perfect.
This distinction matters because a search can be technically excellent and still miss relevant evidence if the database itself does not adequately cover where that evidence appears.
Search strategy
Determines which records your query can retrieve from an information source.
Database coverage
Determines which records are available in that information source for your strategy to retrieve in the first place.
You cannot solve a coverage problem simply by adding more keywords. If a relevant record is absent from the database, no ingenious Boolean expression will summon it into existence.
Different databases may represent different disciplinary worlds
The appropriate databases depend substantially on the research question. JBI gives straightforward examples: CINAHL may be appropriate for nursing questions, PsycINFO for psychology, and LILACS when literature from Latin America and the Caribbean is particularly relevant. It also emphasizes that database selection should reflect topic and geography rather than an arbitrary target number.
This becomes especially consequential for interdisciplinary questions. Consider research on university students' use of artificial intelligence. Depending on the question, relevant work might arise from education, psychology, computer science, information science, communication, or other fields. Searching only the database most familiar to an education researcher could privilege one disciplinary conversation while making another much less visible.
The same issue can occur geographically. If the phenomenon being studied has a strong regional dimension, databases with relevant regional coverage may add evidence that internationally dominant databases represent less completely. JBI explicitly recommends considering geography when selecting information sources.
More databases are not automatically better
The corrective to searching one database is not to search every database your university happens to subscribe to. Database selection should be defensible, not ceremonial.
Two databases with very similar coverage may produce large numbers of duplicate records while contributing relatively little unique evidence. A specialized database, by contrast, may contribute a smaller number of records but expose a research tradition poorly represented elsewhere.
JBI does not prescribe a universal number of databases. Its current guidance notes that most evidence syntheses search at least three, while emphasizing that the relevant databases should be selected according to the review question. That observation should not be converted into a rule that “three databases equals comprehensive.”
Watch Out
Do not choose databases by counting them. Three poorly chosen databases do not become a comprehensive search merely because three sounds methodologically respectable.
Think in terms of complementary coverage
A useful database earns its place because it has a plausible contribution to make. That contribution might come from disciplinary specialization, geographic coverage, particular publication types, indexing practices, or access to a body of literature not adequately represented elsewhere.
Before adding a database, ask what you expect it to contribute. Before excluding one, ask what literature becomes harder to discover without it.
| Question to ask |
Why it matters |
| Which disciplines contribute to this research question? |
A database centered on only one discipline may underrepresent relevant work from another. |
| Is geography important to the phenomenon? |
Regional databases may expose literature inadequately represented in broader international sources. |
| What types of evidence am I trying to locate? |
Different information sources may be needed for journal articles, trials, theses, reports, conference material, or other evidence. |
| What does each additional database contribute? |
This helps distinguish complementary searching from unnecessary duplication. |
| Can I explain why each database was included? |
A defensible search should be driven by the research question rather than habit or convenience. |
Searching another database usually requires translation, not copying and pasting
Moving a search from one database to another is not necessarily a matter of pasting the same query into a different search box. Databases and platforms can differ in controlled vocabularies, field codes, operators, truncation conventions, proximity searching, and other syntax.
Cochrane's technical guidance illustrates this even for different interfaces providing access to the same underlying database: search syntax can vary across platforms, and interface differences can affect how a strategy must be expressed.
JBI accordingly describes translating the search strategy from the primary database into the other relevant databases rather than simply reproducing one query unchanged. It also recommends documenting the platform, search terms, languages, limits, and search dates.
This makes database selection and search vocabulary related but distinct problems. One concerns where you search. The other concerns how you express the concepts you want those sources to retrieve.
Bibliographic databases are not the only information sources
For some review questions, searching multiple bibliographic databases is still not enough. JBI's guidance for systematic reviews identifies a wider range of potential information sources, including trial registers, relevant organizational websites, specific journals, regulatory agencies, contacts with researchers, and grey-literature sources.
Supplementary approaches may also include backward and forward citation searching, handsearching relevant journals, conference searching, thesis searching, and contacting researchers.
Whether those approaches are necessary depends on the purpose of the review. A student looking for several strong sources for a class paper does not necessarily require the search architecture of a systematic review. A systematic review claiming comprehensive identification of eligible evidence faces a much higher standard.
Your research question should determine the search burden
There is a continuum between exploratory searching and formal evidence synthesis. At one end, you may simply be orienting yourself to an unfamiliar field. At the other, your methodology may require a reproducible attempt to identify all eligible studies.
The stronger your claim of comprehensiveness, the stronger the justification needed for the sources you searched and those you omitted.
For a systematic review, JBI recommends searching the widest reasonable collection of information sources appropriate to the question. For scoping reviews, it similarly calls for an explicit, comprehensive, and transparent search designed to identify the available evidence sources.
Additional databases can also teach you about the field
Searching another database is not useful only because it may add records. It can reveal terminology, authors, journals, disciplines, and conceptual traditions that were not apparent from the first source. Those discoveries can improve subsequent searches.
This iterative process is one reason literature searching is rarely a single query followed by passive screening. What you find should sometimes change how you search next.
The deeper objective remains finding the literature that materially matters, not accumulating the largest possible pile of records.
04 · A Practical Example
What can a second disciplinary database reveal?
Hypothetical Example
An educational technology question that crosses disciplinary boundaries
Suppose a researcher is examining how university students regulate their use of generative AI while completing academic tasks. Because the study is situated in higher education, the researcher begins with an education-focused database and retrieves a substantial collection of relevant studies.
During screening, however, several papers refer to constructs developed in cognitive and educational psychology. The researcher therefore searches an appropriate psychology-focused database using a translated version of the strategy. That search retrieves studies examining closely related regulatory behaviors using terminology that was uncommon in the education literature.
A multidisciplinary citation database then reveals papers from human-computer interaction that conceptualize aspects of the same behavior differently. Not every new record belongs in the review, but the additional sources expose relevant intellectual traditions that the first database alone did not adequately represent.
First database
The researcher retrieves a substantial education literature and identifies the dominant terminology used in that field.
Coverage question
Relevant papers indicate that psychology and human-computer interaction also contribute to the problem.
Complementary searches
The strategy is adapted to appropriate databases representing those literatures.
Result
The researcher discovers relevant studies, terminology, and theoretical perspectives that were weakly represented in the initial search.
Interpretation
The value of the additional databases lies in what they contribute to coverage, not simply in increasing the record count.
06 · What This Means for You
How should you decide which databases to search?
Begin with the intellectual territory of the question rather than a favorite search platform. Identify the disciplines, populations, geographic contexts, and evidence types that could plausibly contain relevant research. Then choose information sources that collectively represent that territory.
A simple decision framework
If your question sits mainly within one well-defined discipline and your purpose is exploratory
A strong discipline-appropriate database may be a reasonable starting point, while recognizing that the search is not necessarily comprehensive.
If your question crosses disciplinary boundaries
Identify databases that represent the major contributing disciplines rather than relying only on the database most familiar to your own field.
If geographic or regional evidence is important
Consider information sources with appropriate regional coverage rather than assuming internationally dominant databases represent that literature adequately.
If you are conducting a systematic or similarly comprehensive evidence synthesis
Predefine and justify multiple appropriate information sources, translate the strategy for each, and document the searches transparently.
If important evidence may exist outside conventional journal databases
Consider whether registries, grey literature, citation searching, organizational sources, or other supplementary approaches are necessary for the review question.
Do not mistake access limitations for methodological reasoning. If an ideal database is unavailable, state the limitation rather than quietly pretending that another source is equivalent. Conversely, do not add databases solely to make the methods section look impressive.
Once the sources are selected, the next challenge is making sure each search can recognize the many ways authors may describe the concepts you care about. That requires deliberately searching beyond a single set of keywords.
07 · A Quick Checklist
Have you chosen enough information sources for your question?
Before finalizing your database search, check:
I can explain why each selected database is relevant to my research question.
I have considered every major discipline that plausibly contributes evidence to the question.
I have considered whether geographic or regional coverage matters for this topic.
I have not assumed that a large result count means the literature has been comprehensively covered.
I understand what each additional database is expected to contribute beyond the others.
I have translated the search appropriately for each database or platform rather than assuming identical syntax will work everywhere.
Where comprehensive retrieval is required, I have considered relevant information sources beyond bibliographic databases.
I have recorded the database, platform, search date, strategy, and relevant limits so the search can be transparently reported.
