Reflecting on Article Keywords from Health Sciences Librarianship Journals: Creating a Taxonomy for Trend Analysis

Friday, April 29, 2022

4:00 PM – 4:15 PM CT

Presenter and Author(s)

Shalu Gillum, AHIP

Head of Public Services
Harriet F. Ginsburg Health Sciences Library University of Central Florida College of Medicine
Orlando, Florida

Slides

Author(s)

PH

Pamela R. Herring, AHIP

Electronic Resources Librarian
Harriet F. Ginsburg Health Sciences Library
Orlando, Florida
Deedra Walton, AHIP

Head of Electronic Resources
Harriet F. Ginsburg Health Sciences Library/University of Central Florida College of Medicine
Orlando, Florida
Nadine Dexter, EDD, MLS, DAHIP

Director Health Sciences Library UCF COM
Harriet F Ginsburg Health Sciences Library
Orlando FL, Florida
Terri Gotschall, AHIP

Scholarly Communications Librarian
University of Central Florida Health Sciences Library
Orlando, Florida
Natasha Williams, AHIP

User Services Librarian
Harriet F. Ginsburg Health Sciences Library
Orlando, Florida

Objectives: To create a medical library terms taxonomy in order to find trends in the medical library literature.

To explore trends in scholarly publishing by health sciences librarians in the last five years, we examined keywords from articles published in four health sciences library journals during this time. From our initial review we determined that: author-created keywords are not always based on a controlled vocabulary; MeSH subheadings do not include terms unique to librarianship; and Library of Congress taxonomy is too broad. Thus, it was necessary to create a taxonomy by grouping similar keywords into categories to accurately identify trends.

Methods: Citations from four health sciences library journals from 2016-2020 were exported from PubMed and EBSCO Academic Search Premier into EndNote X9. Keywords from all citations were exported from EndNote as a text file and then imported into Microsoft Excel. Pivot tables were used to determine the number of times each keyword was used and to aggregate identical keywords. A team of six librarians with a variety of skill sets and backgrounds reviewed each keyword to determine how to categorize and group terms to create a taxonomy.

Results: 8,806 keywords from the journals were downloaded into spreadsheets and aggregated. Pivot tables were used to combine identical terms resulting in 2,801 unique keywords which are currently being categorized to complete the taxonomy.

Conclusions: Choosing keywords for an article without the use of a controlled vocabulary allows for flexibility and customization. It may also result in better retrieval during natural language searching. However, what we discovered through this project is that choosing keywords without the use of a controlled vocabulary makes it difficult to conduct a keyword-based trend analysis.