2 Planning and Conducting Literature Searches

Purpose and iterative approach

A literature search is a planned process for finding relevant publications in scholarly sources. A clear search strategy can make the process more efficient, help locate a wider range of useful material, and make it possible to explain how the publications were found.

The process is usually iterative: test a query, examine the results, and adjust the strategy as you learn more about the topic.

Plan around the research question

Begin with the research question and identify its main concepts. For example, a question about whether school-based exercise programs improve adolescents’ sleep can be divided into adolescents, exercise programs, and sleep.

Not every part of a question needs to become a search concept. Outcomes or comparisons may be described inconsistently in article records, so requiring those terms could exclude relevant work. Prioritize concepts likely to be clearly represented in titles, abstracts, or database indexing.

Before searching, decide what kinds of publications or evidence are needed, which disciplines, populations, places, or time periods are relevant, whether limits are justified by the research purpose, and how broad the search needs to be. A focused class assignment may need fewer sources than a systematic review.

Avoid arbitrary limits, such as restricting results to recent years without a reason, because they can hide relevant publications. For a comprehensive review, plan sources and limits in advance and align them with the review’s eligibility criteria.

Select databases and sources

Choose databases according to the subject and the kinds of material needed. A discipline-specific database may provide deeper coverage of its field, while a multidisciplinary database can help locate work across fields. For interdisciplinary questions, searching more than one relevant database can help find material that a single source does not index.

Learn each database’s coverage and search features, and adapt the strategy to its interface. A subject librarian can help identify appropriate databases and develop a robust strategy.

A library catalog is useful for books and other library holdings, whereas article databases index scholarly publications. Google Scholar and similar broad search tools can help with initial exploration or citation discovery, but should not automatically replace subject-relevant databases when a thorough search is needed.

Choose terms and build a query

For each concept, list alternative search terms, including synonyms, spelling variants, abbreviations, older or newer terminology, and related phrases. A subject thesaurus, database subject headings, and the vocabulary in relevant article titles and abstracts can help identify useful terms.

Many databases support both keywords—words in fields such as titles and abstracts—and controlled vocabulary, or standardized subject headings. Using both can improve coverage.

Boolean operators combine the terms and concepts in a query:

• OR joins alternatives within a concept and broadens the search, as in adolescent OR teenager.

• AND combines different concepts and narrows the search, as in adolescent AND sleep.

• NOT excludes a term, but should be used cautiously because relevant records may also contain the excluded word.

Use parentheses to group alternatives and quotation marks for exact phrases. Some databases support truncation or wildcards. For example, adolescen* may retrieve adolescent and adolescents, but symbols and behavior vary by database; check the database’s help guide before relying on a feature.

This example groups alternatives within each concept and requires all three concepts:

Adapt the query to the database because search fields, subject headings, phrase syntax, and truncation symbols are not identical everywhere.

Test and refine the search

Judge a query by examining a sample of its results, not only by its hit count. Check whether relevant known publications appear and whether the results generally fit the research question.

• Too few relevant results: Add synonyms, check spelling and subject headings, remove an overly specific concept, or broaden the date range if appropriate.

• Too many irrelevant results: Add a well-defined concept, search particular fields, or use a justified limit. Avoid narrowing the search so far that relevant work disappears.

• Familiar articles are missing: Check their keywords, abstracts, and assigned subject headings for terms to add.

A search strategy may need separate versions for different databases because their coverage and search rules differ. Track changes and test revised queries against the results.

Keep a search record

Record searches as they are conducted; memory alone is not enough to reproduce a search or explain how it changed. A search log can be kept in a spreadsheet or table.

For each search, include the database or source name, search date, exact search terms and operators, fields searched, limits or filters, and number of results. Notes can capture changes or other relevant observations. For example, one record might note an example database, a search of title and abstract fields with no limits, 126 results, and a plan to add “youth” and inspect subject headings.

Save useful citations and, when available, the database’s search history or exported search results. A clear record helps prevent repeated work, shows how decisions affected results, and makes it easier to update or report the search.