Purpose & Goals

Weeding an academic library is a complex and labor-intensive process and poses challenges for designing a uniform approach that accommodates subject-specific review criteria and workflows. In 2022, subject librarians and Collection Services personnel at Western Washington University developed an ambitious and cyclical plan for reviewing the general collection of roughly 500,000 items every four years, something that had not systematically been done for decades. The Collection Management and Assessment (CMA) department was tasked with designing and facilitating the data-informed collection evaluation process. This poster explores how the CMA department collaborated with subject librarians and other Collections personnel to identify relevant data points to inform withdrawal decisions, methods for collecting and compiling data from multiple sources, and the evolution of decision support tools to respond to disciplinary needs and librarian preferences. Discussions about the relevance of different data points across disciplines and strategies for collection review between subject librarians and Collections personnel also led to increased understanding and alignment of Collection Services and subject librarian goals and priorities for the project.

Design & Methodology

The Collection Management and Assessment department conducted a literature review of large-scale weeding projects, focused on list-based weeding methods. The literature review revealed a gap in research comparing deselection criteria in Humanities subject areas to STEM subject areas, an area of particular importance at Western. Western’s collection review plan called for each of the six wings of the building to be assessed in six-month intervals. This ongoing assessment, coupled with swiftly shifting focus to different LC ranges in each wing, allowed for rapid prototyping of decision support tools based on user feedback to accommodate differences in subject area evaluation. To initiate the process, an ad hoc working group of Collections personnel convened to identify potential metrics and discuss how to collect the supporting data for a test set of 13K items. The group determined that the data they desired was housed in many different platforms (Alma, WorldCat, ClassWeb, and a bespoke reserves program) and the Collection Management and Assessment department worked to implement efficient and scalable procedures to collect and synthesize the information using Alma Analytics, OpenRefine, and Python scripts. For the first round of review, which included LC ranges R-Z, the CMA department piloted a review list built in Excel with a heavy emphasis on filters. This type of decision support tool worked well for STEM librarians who had clear criteria for creating withdrawal lists but did not support the more exploratory approach the Humanities librarians preferred. For the second round of review, which included LC ranges A-G, the CMA department designed an interactive Power BI report, with filters such as LC call number captions and subject headings, to create the opportunity for more qualitative and less rules-based filtering.

Findings

The first wing under review contained materials with the Library of Congress classifications of R-Z, with most materials in R, S, and T (Medicine, Agriculture, and Technology). CMA personnel developed spreadsheets containing bibliographic information, usage data, and electronic and resource-sharing availability for approximately 50k items, which the subject librarians used to review the materials in the wing. CMA personnel also created a tool for tracking retention decisions and what percentage of the collection had been reviewed. STEM librarians were tasked with most review responsibilities in this wing and relied heavily on quantitative criteria, rather than title level review, to make withdrawal recommendations. Based on their knowledge of the curriculum and disciplines, they had clear parameters that allowed them to easily filter the spreadsheet for outdated or low-use materials that are no longer necessary in our institutional context, making the spreadsheet an effective tool for their needs. Librarians in the Social Sciences and Humanities found that there are few clear criteria for identifying materials that are no longer needed in the collection and the spreadsheet did not accommodate a more exploratory approach to assessing the collection. As a result, CMA developed an interactive Power BI report that lets reviewers build their own reports and lists based on various criteria. This report also allows reviewers to build lists based on LC call number hierarchies and subject headings, which provides the librarians a more interdisciplinary lens through which to assess the collection than LC subclass alone. Throughout this process, the CMA department surfaced the need for different approaches to collection assessment and the importance of being flexible and responsive to varying strategies.

Action & Impact

In response to our findings, we have adjusted our decision support tools and print collection review workflow to optimize the process for both subject librarians and Collections personnel. Subject librarians from all disciplines are using the decision support tool built in Power BI in the latest round of collection review. Offering subject librarians the ability to access and use the collections data gathered by CMA in the way that best fits their review workflows has also allowed us to simplify the materials processing stage of the cycle. Rather than tracking withdrawal recommendations across multiple spreadsheets, CMA will collect lists of barcodes for withdrawal candidates at the end of the review period. Moving forward, an opportunity to revisit and iterate upon our decision support tools and workflows will be built into the project timeline. This responsive approach to project planning allows us to collaborate efficiently across library departments and coordinate each stage of the project, as we have the flexibility to adapt our workflows to accommodate changing circumstances.

Practical Implications & Value

This project adds to the research on subject-specific monograph assessment criteria and workflows and introduces strategies for collaborating on data-informed weeding between Collections personnel and subject librarians. The presenters will share their methodologies for collecting and integrating data from disparate systems into decision support tools to assist in making withdrawal recommendations, and the process of iteration based on reviewer feedback and disciplinary needs. They will discuss strategies they employed to collaborate across library departments, and opportunities they identified for reducing subject librarian workload related to the project. Viewers will come away with an understanding of the benefits and drawbacks of how two different decision support tools are used in the Humanities and STEM. The presenters will also provide a Power BI report of the decision support tool developed for the review of LC classes A-G.