ID the key concepts of your question, in this case—
Come up with related terms for each key concept.
TIP! Use the UT Libraries Generate Search Terms tool to create search strategies like the one below.
In some databases, it can be helpful to use truncation to search for terms that begin with a word root.
will search for
osteoporosis, osteoporoses, osteoporotic, osteoporotically...
First, search each key concept separately (lines #1, #2, #3). Consult the database’s search history to see how many hits your terms produce. Big numbers are your friend. If there are too few hits for a key concept, add more synonyms or some broader terms. If there are too many hits, remove terms and/or use narrower terms.
Then combine search lines, as shown (lines #4, #5).
Searchers routinely ask some variation of these questions:
"How many search results is a good number to have?"
"I got 47 search results. Is that enough?"
"I have 2800 search results. Is that too many?"
The highly unsatisfying answer is, IT DEPENDS.
What are you working on? Systematic reviews and scoping reviews, for example, need sensitive searches--comprehensive searches that attempt, as near as possible, to identify all relevant articles. On the other hand, a class assignment on a topic may only require a specific search--a search that attempts to identify a set of highly relevant articles, and which will likely yield a smaller number of search results.
"When conducting a comprehensive search, it is critical to design a strategy that retrieves all potentially relevant articles. Experienced searchers know the power of using controlled vocabularies but also the frustration of not being able to pinpoint articles known to be relevant but missing from the initial retrieval set.
"Librarians have long analyzed Medical Subject Headings to design and refine searches. A MeSH analysis grid can help identify the problems in your search strategy by presenting the ways articles are indexed in the MEDLINE database in an easy-to-scan tabular format. Typically, each column in the grid represents an article, with identifying information of the article at the top of the column, such as the PMID, the author, and the year of publication. The MeSH terms are sorted and grouped alphabetically for ease of scanning. Librarians can then easily scan the grid and identify appropriate MesH terms, term variants, indexing consistency, and the reasons why some articles are retrieved and others are not. This inevitably leads to fresh iterations of the search strategy to include missing important terms.
"In addition to MeSH terms, author-assigned keywords, article titles, and abstracts can be included in the analysis grid."
This work is licensed under a Creative Commons Attribution-NonCommercial 2.0 Generic License.