Get fewer, more relevant results by learning about automatic mapping and checking the Search details for individual searches in PubMed.
Most databases have a search box where it is tempting to enter words or terms and hit "Go". We have Google to thank for that! Google searches content from all over the web and relies on algorithms to get the most relevant items to the top of the results list. To balance the volume of information available with the variability of individual searches, Google relies on more than 200 predetermined algorithms to direct searches. These algorithms depend on many factors, including the terms used on a website, your location, and web traffic (2019).
PubMed also relies on algorithms, or automatic term mapping, to retrieve search results. When you enter search terms, they are matched against Medical Subject Headings (MeSH), journal titles, and author/investigator names. If no match is found, every field describing each individual item in the database is searched.
Most of the time, automatic term mapping in PubMed works in a searcher’s favour and you get relevant results. Sometimes, however, it can lead you down a path that isn’t comprehensive or efficient. Some searches can be matched to MeSH or other criteria that don’t cover what you’re looking for. Say you’re looking for articles on computed tomography scans. Search “CT” in PubMed (1), set Sort by: Publication Date (2) and check the Search details (3). Note that PubMed is translating your search into results by referencing journal titles [Journal] and every single field of information in the database [All Fields].
This search retrieves articles about CT scans but also articles about communication theory and cancer therapy. You get a high number of results, but many aren’t relevant. By understanding how automatic mapping works and checking the Search details, you can get a more manageable number of results that are more relevant to your context.
When searching large, complex databases like PubMed, minimize confusion by testing each term you plan to search. Analyze the Search details for anything unexpected and revise your search terms accordingly. In the “CT” search above, there are a few ways to resolve this issue, depending on why you need the information. If you want to quickly gather a few relevant and current articles about CT scans, you can be more specific and search the complete phrase, “CT scan”. This simple adjustment completely changes the Search details.
You can make the search even more relevant by adding a second concept, such as “risk”. Do the Search details look reasonable? If so, add the concepts together using the AND operator to form a single search string: “CT scan AND risk”.
While Google does not provide search details, you can keep in mind a few things that affect Google’s algorithms. If you are logged into Google, your previous search history may influence the ranking and quality of your results. The location you are searching from may also be a factor. If you are in Halifax, you’ll get a different list of results than if you are in Toronto. To reduce bias as much as possible, log out of Google before searching. You can also try alternative web search engines that do not track your activity, such as DuckDuckGo or Yippy.
It’s important to think critically about how your searches are being translated, so that the results you are choosing from are as bias-free as possible. Thoughtful searching takes practice. If you want to learn more, sign up for one of our Learning Opportunities or consult with a library and information professional.
Librarian Educator, Education & Outreach
Dickson Building, Central Zone
0 Comments.