About

Inclusivity and Bias in Metadata Research Intern Report

Posted on December 6, 2018 by

This is a guest blog post by Karen Koswara, School of Informatics, Computing, and Engineering, Indiana University.

I am Karen Koswara and I am the first Undergraduate Research Opportunities in Computing (UROC) student to be involved in IU Libraries’ Inclusivity and Bias in Metadata Research. As a new Informatics major, I want to explore as many areas as possible until I find a perfect field that I am interested in. Thus, I signed up for many projects through UROC. I remembered that there was a large number of projects from the School of Informatics, Computing, and Engineering (SICE) as listed in UROC’s sign-up page. I was encouraged to choose as many projects that interested me as possible as there was no guarantee that I will get my first preference. Among the many projects that I chose, I remember the word “metadata” in Inclusivity and Bias in library cataloging and Metadata was very appealing to me, thinking it would be amazing to start working with metadata. A few weeks later, I was signed up to work with Julie Hardesty and I had never been more excited. On our first meeting, I met up with Julie to clarify what I will be doing for the research. Julie explained to me that the Metadata Discussion Group never had students involved before and she was interested in exploring inclusivity and bias in metadata from a student point of view. The research would mainly be reading articles, discussing my thoughts and points of view, and look into ideas that will help the Metadata Discussion Group to move forward. I officially signed up for this research after that meeting.

Later in the research after having my second meeting with Julie, I realized many issues came from collections in other languages that Library of Congress Subject Headings (LCSH) classify too broadly, making those collections unreachable. Looking at this situation, it made me think to approach this issue by looking into different cataloging systems in different countries with different languages. My idea is that maybe other countries’ cataloging classification is more specific, and we can apply their classification terms to LCSH. I believe this is possible when I realize a language often borrows words from other languages. For example, in Indonesian language, many words are the same as Dutch as Indonesia was under the Dutch for about three centuries. Indonesian language also uses some English terms. For example, the word “transition” is “transisi” in Indonesian and they have the same meaning.

I first thought of the Chinese library because of the complex Chinese characters. The most widely used cataloging system in China is now known as CLC (Chinese Library Classification) also known as CCL (Classification for Chinese Library).[1] There are also many other classifications in China that are used such as LCPUC (Library Classification of the People’s University of China), LCCAS (Library Classification of the Chinese Academy of Sciences), and MSL (Library Classification for Medium and Small Libraries). Taiwan, Hong Kong, and Macau use a separate cataloging system known as New Classification for Chinese Library. Harvard has established their own cataloging system for Chinese collections known as Harvard-Yenching Classification System which many organizations in the United States adopted.

I also try to think of other organizations that might have their own library of some sort to keep track of their employees’ work such as National Geographic. I thought of National Geographic as I believe they would have some kind of cataloging system to keep track of their researchers’ and photographers’ work. Unfortunately, I was not able to find much since there is not much information available about this on the internet. I did find out that they do have some kind of library to keep images by their photographers.

As a first–time inexperienced researcher, I really hope this point of view helps leads to a new approach and a step forward. I have also learned that in doing this research, one has to look deep into oneself, and see how one portrays a word as well as what it means to them to avoid biases. I believe it would also help future researchers who are completely new to this research to look into a recent graduate student article titled “The Language of Cataloguing: Deconstructing and Decolonizing Systems of Organization in Libraries” by Crystal Vaughan at Dalhousie University in Nova Scotia, Canada to help understand the situation as it helped me very much.[2]

[1] Branum, Candise. (20 December 2017). “The Organization and Classification of Library Systems in China.” Candise Branum, MLS, https://candisebranum.wordpress.com/2014/04/25/the-organization-and-classification-of-library-systems-in-china/.

[2] Vaughan, Crystal. (2018). “The Language of Cataloguing: Deconstructing and Decolonizing Systems of Organization in Libraries.” Dalhousie Journal of Interdisciplinary Management 14, https://ojs.library.dal.ca/djim/article/view/7853.

Summary of May 7 Libraries In-house Institute session

Posted on May 22, 2018 by

The May 7, 2018 IU Libraries’ In-house Institute session, “Inclusivity and Bias in Metadata: Continuing the Conversation” began with a brief recap of two previous Metadata Discussion Group sessions this spring, Facts in Metadata and Bias in Metadata, followed by listening to the NPR CodeSwitch interview about implicit bias that discussed the one-day training that Starbucks employees will receive in response to a racial bias incident in Philadelphia. We then moved to a discussion of ways to address inclusion and bias in metadata standards. In discussing the fact that Library of Congress (LC) copy records are not richly described, a participant suggested leveraging consortiums or institutional affiliations in order to get catalog records.[1] On the other hand LC standards, while conservative and showing bias, have taken a long time to come about and be standardized so it is difficult to imagine re-doing all of that work.

Participants mused on the benefits of accepting the use of vocabularies other than those from LC. Such changes would require thorough and thoughtful planning, documentation, and change implementations in multiple systems. A major concept in this kind of re-thinking of the catalog records we create is to consider the marginalized communities that approach our catalog or are represented by items in our catalog.[2] Placing those communities in the center of our cataloging practice to ensure they can find themselves and what they need in our description ensures that the most marginalized are included in what we do, making all of our records that much more accessible to everyone.

The conversation then turned to translations and providing multiple language versions of cataloging records. All too often catalogers are presented with items about an unfamiliar subject, possibly in an unfamiliar language. Trying to offer translations and multiple language versions helps everyone find these items more easily but requires a lot more cataloging time and the tools are not always well-suited to show multiple language versions of cataloging records. Lack of subject matter familiarity seemed to be an even bigger obstacle. Presented with something about an unfamiliar subject, such as a place or culture, descriptive cataloging can end up too broad, sparse, and unhelpful, leading to catalog records without much detail and easily missed in normal end-user searching and browsing, especially for those end-users who are familiar with the subject. This has the effect of making small or specialized collections that much harder to find. One solution suggested by attending subject librarians was to enlist them in the cataloging process. Contacting a subject librarian with expertise on the subject presented in that item can help provide more detailed and relevant subject terms. It was also suggested that collection managers could help supply controlled vocabulary terms to enhance description or even provide terms that are not part of any controlled vocabulary if those sets of terms are not helpful to describe a collection. The big takeaway from this part of the conversation was that there did seem to be steps we could take in the cataloging process to improve inclusivity of better descriptive terms to provide catalog records that are more accessible.

We ended the session by discussing what next steps we could take to continue this conversation and make positive changes in our cataloging practices to provide more inclusive and accessible catalog records and combat our own implicit biases in our cataloging practice. Metadata Discussion Group sessions will continue to explore these topics. Implicit bias training and active bystander strategies training seem to hold a promise to at least increase awareness if not actually stop biased activity and behavior so we are hopeful that some kind of program can be made available within the Libraries.[3] Additionally, we can create a list of expertise in subject areas for catalogers to reference when encountering unfamiliar subjects to help enhance those records for better discoverability.

All of these ideas and others that we didn’t have time to discuss in this session are gathered in an online spreadsheet from this session: Inclusivity and Bias in Metadata – In-house Institute 2018. All are welcome to add more ideas on how to address inclusion and bias in metadata standards, our application of those standards, how to build transparency into our tools and processes, areas of improvement, and next steps that we can take as a library to provide relevant and inclusive information about our collections for better discoverability and access.

Stay tuned for future posts and scheduled events as we continue to think critically about our work as catalogers and metadata creators.

[1] Editorial note from Jennifer: IU Libraries is a long-time, productive member of the Program for Cooperative Cataloging and has in the past leveraged affiliations with the Big Ten Academic Alliance Library Initiatives in order to purchase ebooks and their associated catalog records.

[2] For additional and really helpful readings on the subject of inclusivity and bias in library practice, see the Digital Library Federation (DLF) Cultural Assessment Working Group’s Annotated Bibliography, particularly the readings in the sections on Collection Publicizing and Discoverability and Metadata and Description Practices: https://osf.io/94pgj/

[3] Jennifer Liss has started an open reading list within IUCAT on Implicit Bias: https://iucat.iu.edu/catalog/myfolders/folders/108239669

Next meeting: Toward Inclusive Metadata Practice

Posted on May 1, 2017 by

Illegal aliens. Sexual minorities. Ableism. In the attempt to describe an entire universe of things and ideas, librarians design thesauri, knowledge classification schemes, and metadata standards to help people find information. Because these knowledge organization schemes codify what can and cannot be said about a library resource or author, one would hope that these knowledge schemes err on the side of inclusivity and are transparent about inherent historical and cultural biases. Alas, mistakes have been made. How can we do better?

DATE: Monday, May 8th
TIME: 10:50-11:35 am
PLACE: Libraries’ In-house Institute, IMU Walnut Room
TOPIC: Toward Inclusive Metadata Practice
MODERATORS: Julie Hardesty & Jennifer Liss

Resources You Might Consult

There are simply too many great resources to choose from! We can post a bibliography after the session but for now, perhaps the following will provide a good entry into the topic:

  • Noble, Safiya. (2015 Oct. 26). Power, privilege and the imperative to act. DLF Forum 2015 keynote. Retrieved from http://hdl.handle.net/2429/55690 [Skip ahead to 24:50 in the recording]