Boolean Operators: How to Use AND, OR and NOT to Search for Research Papers

Website image

Dr. Rajni Garg

Executive Lead, Research & Content Strategy

7 min read

Boolean Operators: How to Use AND, OR and NOT to Search for Research Papers
Boolean Operators: How to Use AND, OR and NOT to Search for Research Papers

Boolean operators are the words AND, OR and NOT, which you type between search terms to tell a database how to combine them. To search for research papers, join synonyms for the same idea with OR, join different ideas with AND, and use NOT only when you are sure the excluded term never appears in a paper you want.

This guide covers what each operator does, the tools that go with them, how to build a search string, and how PubMed, Scopus and Google Scholar differ.

TL;DR

  • AND narrows: every result must contain all of your terms.

  • OR broadens: a result can contain any of your terms, so use it for synonyms.

  • NOT excludes: it drops every result that mentions a term, including some you may want.

  • Put each OR group in parentheses, and check each database’s help page, because databases do not all read operators the same way.

What are Boolean operators?

Boolean operators are logical connectors that tell a database how your terms relate: AND means all of them, OR means any of them, and NOT means leave this one out. Most research databases support them, though each has its own rules for typing them.

Venn diagrams showing that AND returns only the overlap of two terms, OR returns everything in either term, and NOT returns the first term with the second removed.

How do AND, OR and NOT work?

Say you are writing a paper on remote work for an organizational psychology course.

AND

AND means every result must contain all of your terms.

"remote work" AND productivity

You get papers that mention both remote work and productivity, and nothing that mentions only one. Each extra AND shrinks the results, so use it to focus a search that returns too much.

OR

OR means a result can contain any of your terms.

"remote work" OR telework OR telecommuting

Authors describe the same idea in different words, and OR catches them all. Each term you add widens the results, so use OR for synonyms, spelling variants and abbreviations, not for separate ideas.

NOT

NOT drops any result that contains a term.

"remote work" NOT "hybrid work"

You get remote work papers that never mention hybrid work. That sounds tidy, but it also removes papers that compare the two, or that mention hybrid work once. More on that below.

Operator

Use it for

What it does to your results

Example

AND

Joining different ideas

Fewer, more focused results

"remote work" AND productivity

OR

Joining synonyms for one idea

More, broader results

telework OR telecommuting

NOT

Excluding a term

Fewer results, with a risk of losing relevant papers

"remote work" NOT "hybrid work"

How do parentheses, quotation marks and truncation work?

These are not operators, but they control how a database reads your search.

Parentheses

Parentheses group terms into one unit. Compare:

"remote work" OR telework AND productivity

("remote work" OR telework) AND productivity

The second says exactly what you mean: either term for remote work, plus productivity. The first is ambiguous, and databases settle it differently. PubMed processes searches from left to right, while Scopus has processed OR before AND. A database that processes AND first would read it as remote work on its own, or telework with productivity. Put every OR group in parentheses and you never have to guess.

Quotation marks

Quotation marks turn words into a phrase. Typing remote work inside quotes finds the two words side by side, while the same words without quotes can match records that use remote and work anywhere, even far apart.

Truncation and wildcards

Truncation finds different endings of a word. The symbol is usually an asterisk, so this finds adolescent, adolescents and adolescence:

adolescen*

Choose the stem with care, because cat* also finds category and catalyst. Some databases also have a single-character wildcard. In Scopus, organi?ation finds both organisation and organization.

How do you build a search string from a research question?

Here is the method, step by step, for an environmental science question: How does urban tree cover affect summer temperatures in cities?

  1. Write your research question in one sentence.

  2. Pick out the two or three main concepts: here, urban trees and heat. Skip words such as “how” and “affect.”

  3. List synonyms for each concept. Skim the titles, abstracts and keywords of a few relevant papers to see which words authors use, including spelling variants and plurals.

  4. Join the synonyms in each concept with OR, inside parentheses.

  5. Join the concept groups with AND.

  6. Run the search, read the first 20 or so results, and adjust. If a paper you expected is missing, add the words it uses.

For this question, your concept list might look like this:

Urban trees

Heat

"street tree*"

"urban heat island"

"urban forest*"

"land surface temperature"

"tree canopy"

"heat mitigation"

"urban greening"

cooling

Put together, the search reads:

("street tree*" OR "urban forest*" OR "tree canopy" OR "urban greening") AND ("urban heat island" OR "land surface temperature" OR "heat mitigation" OR cooling)

Then test it. If cooling pulls in papers about air conditioning, swap it for “cooling effect.” There is no group for cities, because several terms already say urban and another AND group would mostly cut results.

Diagram of a research question turned into a search string, with synonyms joined by OR inside each concept and the two concepts joined by AND.

If you would rather start from a draft, the AnswerThis search string builder turns a topic or PICO question into an editable search for 12 search interfaces, including PubMed, Scopus, ERIC and Google Scholar. It groups synonyms for each idea with OR, joins the ideas with AND, and uses NOT only with care, since excluding records can remove relevant studies. Paste the draft into your database, run it there and review the results. You get 3 free runs per browser.

AnswerThis search string builder set to Scopus, with an example topic description entered and the Generate search string button below it.

How do Boolean operators differ between databases?

The logic of AND, OR and NOT is the same everywhere, but the syntax is not, so check each database’s help page before you reuse a search. If you are still choosing where to search, our guide on how to find scholarly articles covers the main databases.

PubMed

PubMed covers biomedical literature, so it suits a health question, such as whether physical activity improves sleep in older adults:

(exercise[mh] OR exercise[tiab] OR "physical activity"[tiab]) AND (sleep[mh] OR sleep[tiab] OR insomnia[tiab]) AND (aged[mh] OR "older adults"[tiab])

The square brackets hold search field tags. [tiab] searches the title and abstract. [mh] searches MeSH Terms, the controlled vocabulary NLM uses to index PubMed citations, so it finds papers indexed under a subject even when the authors use other words. Using both covers what one field misses. The help page lists other tags too, such as Text Words [tw].

Other rules from PubMed’s help page:

  • Type AND, OR and NOT in uppercase.

  • Field tags turn off automatic term mapping, PubMed’s process for matching your words to MeSH headings, and limit the search to the term you typed.

  • Truncation needs at least four characters before the asterisk, and it also turns off automatic term mapping.

  • Quotes around a phrase are ignored if the phrase is not in PubMed’s phrase index.

Our PubMed search string builder drafts this kind of search from a topic or PICO question, tagging free-text synonyms with [tiab] and candidate MeSH headings with [mh]. Paste the draft into PubMed to run it. A verified MeSH heading only confirms that the heading exists, so read its scope and narrower terms before you keep it.

Scopus

Scopus uses field codes in place of tags. According to Elsevier’s Scopus search guide, TITLE-ABS-KEY searches article titles, abstracts and keywords together:

TITLE-ABS-KEY(("street tree*" OR "urban forest*" OR "tree canopy") AND ("urban heat island" OR "land surface temperature"))

The guide adds three rules. Scopus writes NOT as AND NOT, which must go at the end of a search. It lists the processing order as OR, then AND, then AND NOT, and says that order is being updated in 2026 to align with industry standards, so parentheses matter here. And double quotes give a loose phrase that ignores punctuation and includes plurals, while braces, as in {urban heat island}, give an exact phrase.

Google Scholar

Google Scholar is built for quick searches. Its help page documents only a few search features: the author: operator, quotation marks around a paper’s title, and an advanced search by author, title and publication that can also limit by date. It does not document OR, the minus sign, wildcards or a limit on query length. Keep Scholar searches short, and add a subject database for a thorough search.

Why does NOT need care?

NOT is the easiest way to lose good papers without noticing. It removes every record that contains the term anywhere in the fields you search, even in passing. Take an education example:

bullying NOT cyberbullying

This drops papers about cyberbullying, as you wanted. It also drops a study of playground bullying whose abstract mentions cyberbullying once, and every study that compares the two. Before you add NOT, try these:

  • Preview the loss. Run the same search with AND in place of NOT. Those results are exactly what NOT would remove, so skim them first.

  • Use a more specific phrase, such as “face-to-face bullying,” instead of excluding.

  • Use the database’s filters for date, language or document type.

  • Remove the few off-topic papers by hand when you screen your results.

If you still need NOT, save it for a term that cannot appear in a paper you want.

What are the most common mistakes?

  • Joining synonyms with AND. A search for telework AND telecommuting asks for papers that use both words, and most do not.

  • Mixing AND and OR without parentheses, which leaves the database to decide the order.

  • Pasting curly quotation marks. Word processors curl straight quotes, and Scopus support warns that curly quotes can cause a query parsing error. Draft your searches in a plain text editor.

  • Reusing one database’s syntax in another. PubMed tags such as [tiab] are not Scopus syntax, so rewrite the fields for each database.

FAQs

What are the three Boolean operators?

AND, OR and NOT. AND narrows a search to results that contain all of your terms, OR widens it to results that contain any of them, and NOT removes results that contain a term.

Do Boolean operators have to be in capital letters?

In PubMed, yes: its help page asks you to enter operators in uppercase. Other databases differ, so typing AND, OR and NOT in capitals is a safe habit everywhere.

Does OR give more results than AND?

Yes. OR accepts any of your terms, so each one you add brings in more results. AND requires all of them, so each one you add leaves fewer.

Can I use the same search string in every database?

You can keep the same concepts and OR groups, but not the exact syntax. Field tags, phrase rules, truncation and operator order all differ, so adjust it using each database’s help page.

Does Google Scholar support Boolean operators?

Its help page does not document OR, NOT or the minus sign. It does describe the author: operator, quotation marks around a title and an advanced search, so keep Scholar searches short and check the results.

Ready to try your own question? Describe your topic in the search string builder to get an editable draft, then run it in your database and check the results yourself.