Boolean operators for literature searching are the connecting words, chiefly AND, OR, and NOT, that tell a database how to combine your search terms. OR broadens a search by retrieving records containing any of the listed synonyms, AND narrows it by requiring every concept to be present, and NOT excludes records containing a term. Used together with truncation and phrase searching, they turn a list of words into a precise, reproducible query.
Why the order of combination decides your results
The single most common search error is mixing AND and OR on the same line without grouping them. Databases process operators in a set order, so an ungrouped query can silently retrieve the wrong set. The fix is the OR-within, AND-between structure used in every well-built systematic review search strategy: gather each concept’s synonyms on their own lines joined by OR, then combine the finished concept lines with AND. This keeps each concept intact and makes the logic auditable line by line.
OR: casting the net within a concept
Authors describe the same idea in many ways, so within a single concept you join every synonym, spelling variant, and related phrase with OR. A record only needs to match one of them to be retrieved. This is where most of a search’s sensitivity comes from, and where mining known papers and subject heading thesauri pays off, because a missed synonym is a silently missed study.
AND: requiring every concept
Once each concept block is built, you join the blocks with AND so only records touching all your concepts survive. Every additional AND narrows the result, which is why systematic searches keep the number of concepts small, usually the population and the intervention. The full reasoning for that restraint sits in our guide to running a systematic literature search.
NOT: the operator to use sparingly
NOT removes every record containing a term, which sounds tidy but is dangerous. A relevant study that happens to mention the excluded term in passing disappears with no trace. For that reason, experienced searchers avoid NOT in sensitive systematic searches and prefer to handle irrelevant records during title and abstract screening instead, where exclusions are visible and reversible.
Beyond AND, OR, and NOT
Truncation and wildcards
A truncation symbol, often an asterisk, retrieves all word endings from a stem, so one term covers a singular, a plural, and related forms at once. Wildcards stand in for a single character to catch spelling variants between regions. Both reduce the number of OR terms you have to type while keeping the search sensitive.
Phrase and proximity operators
Quotation marks force a database to match an exact phrase, while proximity operators retrieve records where two terms appear within a set number of words of each other. Proximity searching is more flexible than a fixed phrase and more precise than a plain AND, which makes it useful for tightly linked concepts.
How operator precedence changes your results
Databases evaluate operators in a fixed order rather than strictly left to right, and most process AND before OR unless you intervene. That single rule is behind the most damaging silent error in searching. Consider a line meant to find studies of children OR adolescents AND anxiety. Without grouping, the database reads it as “children, OR (adolescents AND anxiety)”, so every record mentioning children is retrieved regardless of anxiety, and the result balloons with irrelevant hits. The fix is explicit parentheses: write (children OR adolescents) AND anxiety, which forces the synonym group to resolve first. The practical safeguard is the line-numbered, one-concept-per-line method, where each OR block is built and stored as its own numbered line and the blocks are combined by number, so precedence can never bite you.
A worked example with truncation and proximity
Putting the pieces together, a single intervention concept for a search on cognitive behavioural therapy might read: (“cognitive behavio*r* therapy” OR CBT OR (cognitive ADJ3 therap*)). Here the truncation in therap* captures therapy and therapies, the wildcard in behavio*r* catches both the British and American spelling, and the proximity operator ADJ3 retrieves records where “cognitive” appears within three words of a form of “therapy”, catching phrasings a fixed quotation-mark phrase would miss. One well-built line like this often replaces a dozen separate OR terms while staying just as sensitive, which is exactly the economy a strong controlled-vocabulary and free-text block aims for.
Carrying logic between databases
The operators are conceptually the same everywhere, but the exact syntax, truncation symbols, and field codes differ by platform, so a strategy must be translated rather than copied when you move between databases for systematic reviews. Truncation is an asterisk on most platforms but not all; the proximity operator might be ADJ in one interface, NEAR or N in another, and the number that follows it may count words differently. Field tags shift too, so a title-and-abstract restriction written one way in your lead database needs rewriting elsewhere. Whatever the platform, you record every operator and line, as covered in documenting and reporting your search, and have the logic checked through PRESS peer review, which exists largely to catch precedence and translation slips.
Operator mistakes that recur in submitted reviews
Three Boolean errors turn up again and again, and all of them are invisible unless someone audits the search:
- Ungrouped AND and OR on one line, which lets precedence retrieve the wrong set, as in the children-and-anxiety example above.
- Over-truncation, where a stem cut too short, such as
cell*, pulls in unrelated words and floods the search with noise. - NOT used to tidy results, which silently discards relevant studies that mention the excluded term only in passing; handle these during full-text screening instead.
Getting the combination logic right is what makes the difference between a search that reflects the literature and one that reflects an accident of operator order. If you would rather hand the build and audit to an information specialist, our search design and peer review service produces a logic-checked strategy ready to run across every database.