My research in Knowledge Organization focuses on subject indexing, the multi-step process of determining what a resource is about (subject analysis) and then representing that aboutness in an indexing language (subject representation).
Research Focus and Approach
I am interested in comparing how different types of subject indexers do their work, characterizing how the indexes that they produce differ, and in identifying ways to combine the approaches and techniques of these different types of indexers. While every individual subject indexer has different experience, training, and motivations, I use a framework with four types of subject indexers in my research:
- Professional indexers, such as librarians and archivists
- Domain and folk experts, such as folk biologists and other academics and professionals
- Social and casual indexers, such as social taggers and ad hoc organizers of small and often personal information spaces
- Machine indexers, such as neural nets that classify and cluster information based on algorithms and training data
As I study these types of subject indexers, I look at the social and ethical implications of their work and at the practical effects of their subject indexing decisions on retrieval and information architecture. Relatively recent technological developments, like large digital collections and data sets, social tagging, and machine classification and clustering make these particularly rich areas of study.
I typically study subject indexing through case studies of specific subject indexers working with specific collections, focusing on the artifacts that they create—both the indexes themselves and other artifacts that reveal their processesses.
My research aims to answer three types of questions about subject indexing and subject indexers.
1. Comparison of Types of Subject Indexers
How do different types of subject indexers compare in their motivations, biases, tools, and processes? Researchers like Adler, Bates and Rowley, Kipp, and others have done good research in this space, but I think that we can learn more, especially about domain experts and machines, through direct comparison of the indexes that different subject indexers produce and notes about their processes and decision-making. We can also learn about these other subject indexers through studies that observe and describe their processes, tools, and outputs but do not explicitly compare them other types of subject indexers.
2. Hybrid Approaches to Subject Indexing in Digital Environments
Do subject indexers adopt tools and processes that were initially developed by other types of subject indexers? For example, many libraries have incorporated social tags into their catalogs (Maness, Furner). How did librarians learn about and evaluate that technology? What were the barriers to adoption? When is adopting tools and processes from other subject indexers most fruitful? Can we use this knowledge to build hybrid subject indexing approaches? I believe that these questions of adoption are especially interesting with the emergence of digital technology and very large digital collections. In particular, I want to emphasize the impact of social tagging (Golder and Huberman, Munk and Mork, and others), machine indexers (De Campos, et al., Golub, Mikolov et al., and others), and digital cataloging and indexing tools (Soergel) on subject indexing in this emergent digital environment.
3. Subject Indexing at Social, Cultural, and Ethical Intersections
How do the motivations, processes, and tools of different types of subject indexers intersect with social, cultural, and ethical aspects of subject indexing? Can folksonomies represent minority voices? Can machine indexers overcome biases in their training data? How do subject indexers promote social justice? How do the answers to that question change depending on the type of subject indexer? How can indexes reflect and honor both mainstream and folk wisdom? Because indexes codify the views of subject indexers in persistent and powerful ways, understanding how different subject indexers arrive at their views and representation choices is critical.
Selected Publications and Conference Proceedings
Holstrom, C. (2020). The Effects of Suggested Tags and Autocomplete Features on Social Tagging Behaviors. Proceedings of ASIS&T 2020, Pittsburgh, PA. Award: Best Student Paper
Holstrom, C. & J.T. Tennis. (2020). Visibility, Identity, and Personal Expression: Qualitative Case Studies of Social Tagging on MetaFilter. Proceedings of ISKO 2020, Aalborg, Denmark
Holstrom, C. (2019). Moving Towards an Actor-Based Model for Subject Indexing. Proceedings of NASKO 2019, Philadelphia, PA.
Holstrom, C. (2019). "Is This a Chapter Book?": Parent-Involved Categorization in a Kindergarten Classroom Book Collection. Cataloging & Classification Quarterly, 57(2).
Holstrom, C. (2018). Local Authorial Voice and Global Authorial Voice in Community-Authored Knowledge Organization Systems. SIG/CR Workshop at ASIS&T 2018. Vancouver, B.C.
Holstrom, C. (2018). Social Tagging: Organic and Retroactive Folksonomies. Proceedings of JCDL 2018. Fort Worth, TX.
For a full listing of my publications and conference proceedings, see my CV.