Literature searching explained

A search strategy is an organised structure of key terms used to search a database. The search strategy combines the key concepts of your search question in order to retrieve accurate results.

Your search strategy will account for all:

• possible search terms;
• keywords and phrases;
• truncated and wildcard variations of search terms; and
• subject headings (where applicable).

Each database works differently so you need to adapt your search strategy for each database. You may wish to develop a number of separate search strategies if your research covers several different areas.

It is a good idea to test your strategies and refine them after you have reviewed the search results.

How a search strategy looks in practice

Take a look at this example literature search in PsycINFO (PDF) about self-esteem.

The example shows the subject heading and keyword searches that have been carried out for each concept within our research question and how they have been combined using Boolean operators. It also shows where keyword techniques like truncation, wildcards and adjacency searching have been used.

Search strategy techniques

The next sections show some techniques you can use to develop your search strategy.

Skip straight to:

• Choosing search terms
• Searching with keywords
• Searching for exact phrases
• Using truncated and wildcard searches
• Searching with subject headings
• Using Boolean logic
• Citation searching

Choose search terms

Concepts can be expressed in different ways eg “self-esteem” might be referred to as “self-worth”. Your aim is to consider each of your concepts and come up with a list of the different ways they could be expressed.

To find alternative keywords or phrases for your concepts try the following:

• Use a thesaurus to identify synonyms.
• Search for your concepts on a search engine like Google Scholar, scanning the results for alternative words and phrases.
• Examine relevant abstracts or articles for alternative words, phrases and subject headings (if the database uses subject headings).

When you've done this, you should have lists of words and phrases for each concept as in this completed PICO model (PDF) or this example concept map (PDF).

Watch this short video on how to plan a search and generating good search terms:

As you search and scan articles and abstracts, you may discover different key terms to enhance your search strategy.

Using truncation and wildcards can save you time and effort by finding alternative keywords.

Search with keywords

Keywords are free text words and phrases. Database search strategies use a combination of free text and subject headings (where applicable).

A keyword search usually looks for your search terms in the title and abstract of a reference. You may wish to search in title fields only if you want a small number of specific results.

Some databases will find the exact word or phrase, so make sure your spelling is accurate or you will miss references.

Search for the exact phrase

If you want words to appear next to each other in an exact phrase, use quotation marks, eg “self-esteem”.

Phrase searching decreases the number of results you get and makes your results more relevant. Most databases allow you to search for phrases, but check the database guide if you are unsure.

Truncation and wildcard searches

You can use truncated and wildcard searches to find variations of your search term. Truncation is useful for finding singular and plural forms of words and variant endings.

Many databases use an asterisk (*) as their truncation symbol. Check the database help section if you are not sure which symbol to use. For example, “therap*” will find therapy, therapies, therapist or therapists.

A wildcard finds variant spellings of words. Use it to search for a single character, or no character. Check the database help section to see which symbol to use as a wildcard.

Wildcards are useful for finding British and American spellings, for example: “behavio?r” in Medline will find both behaviour and behavior.

There are sometimes different symbols to find a variable single character. For example, in the Medline database, “wom#n” will find woman and also women.

Watch this video to learn more about searching using truncation and wildcards:

Use adjacency searching for more accurate results

You can specify how close two words appear together in your search strategy. This can make your results more relevant; generally the closer two words appear to each other, the closer the relationship is between them.

Commands for adjacency searching differ among databases, so make sure you consult database guides.

In OvidSP databases (like Medline), searching for “physician ADJ3 relationship” will find both physician and relationship within two major words of each other, in any order. This finds more papers than "physician relationship".

Using this adjacency retrieves papers with phrases like "physician patient relationship", "patient physician relationship", "relationship of the physician to the patient" and so on.

Searching with subject headings

Database subject headings are controlled vocabulary terms that a database uses to describe what an article is about.

Watch this video to learn more about using subject headings in database searching:

Using appropriate subject headings enhances your search and will help you to find more results on your topic. This is because subject headings find articles according to their subject, even if the article does not use your chosen key words.

You should combine both subject headings and keywords in your search strategy for each of the concepts you identify. This is particularly important if you are undertaking a systematic review or an in-depth piece of work

Subject headings may vary between databases, so you need to investigate each database separately to find the subject headings they use. For example, for Medline you can use MeSH (Medical Subject Headings) and for Embase you can use the EMTREE thesaurus.

SEARCH TIP: In Ovid databases, search for a known key paper by title, select the "complete reference" button to see which subject headings the database indexers have given that article, and consider adding relevant ones to your own search strategy.

Use Boolean logic to combine search terms

Boolean operators (AND, OR and NOT) allow you to try different combinations of search terms or subject headings.

Databases often show Boolean operators as buttons or drop-down menus that you can click to combine your search terms or results.

The main Boolean operators are:

• OR
• AND
• NOT

OR is used to find articles that mention either of the topics you search for.

AND is used to find articles that mention both of the searched topics.

NOT excludes a search term or concept. It should be used with caution as you may inadvertently exclude relevant references.

For example, searching for “self-esteem NOT eating disorders” finds articles that mention self-esteem but removes any articles that mention eating disorders.

Watch this video to learn more about how to combine keywords efficiently to find relevant sources:

Citation searching

Citation searching is a method to find articles that have been cited by other publications.

Use citation searching (or cited reference searching) to:

• find out whether articles have been cited by other authors
• find more recent papers on the same or similar subject
• discover how a known idea or innovation has been confirmed, applied, improved, extended, or corrected
• help make your literature review more comprehensive.

You can use cited reference searching in:

• OvidSP databases
• Google Scholar
• Web of Science
• Scopus

Cited reference searching can complement your literature search. However be careful not to just look at papers that have been cited in isolation. A robust literature search is also needed to limit publication bias.

Watch this video to learn more about citation search: