Evidence Summary

Uneven Adherence to Professional Guidelines and Potential Ethnic Bias in Service Provision Evidenced in Virtual Reference Service Interactions

A Review of:

Hamer, S. (2021). Colour blind: Investigating the racial bias of virtual reference services in English academic libraries. The Journal of Academic Librarianship, 47(5), 102416. https://doi.org/10.1016/j.acalib.2021.102416

Reviewed by:

Scott Goldstein

Coordinator, Web Services & Library Technology

McGill University Library

Montréal, Québec, Canada

Email: scott.goldstein@mcgill.ca

Received: 1 Dec. 2021 Accepted: 19 Jan. 2022

2022 Goldstein. This is an Open Access article distributed under the terms of the Creative Commons‐Attribution‐Noncommercial‐Share Alike License 4.0 International (http://creativecommons.org/licenses/by-nc-sa/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly attributed, not used for commercial purposes, and, if transformed, the resulting work is redistributed under the same or similar license to this one.

DOI: 10.18438/eblip30085

Abstract

Objective – To investigate whether there is evidence for implicit ethnic bias in virtual reference service interactions.

Design – Email-based structured observation study.

Setting – Academic libraries in England.

Subjects – 158 email-based virtual reference service interactions from one of 24 academic libraries in England.

Methods – The study used a sample of 24 academic libraries across eight of the nine regions of England (excluding London). The body of the email message sent to each library consisted of one of five questions and was identical except for personalization to the institution. The first three questions were designed to be more likely to be answered in response to an unaffiliated user, and the last two questions were designed to be less likely to be answered in response to such a user. Each library received an email with each question from a different sender during each of five weeks, plus a repeat of question one in week six with slightly altered wording to serve as a control question. Emails were sent on randomized work days at different times of day. The messages were signed with one of six names representing the largest distinct ethnic population groups in England and Wales: Hazel Oakland (White British), Natasza Sakowicz (White Other), Zhao Jinghua (North Asian), Priya Chakrabarti (South Asian), Ebunoluwa Nweke (Black African), and Aaliyah Hajjar (Arab). All names were feminine and represented unaffiliated users. Email replies were coded according to a set of 27 characteristics based on the two most well-known professional guidelines for providing best practice reference services, namely, IFLA and RUSA.

Main Results – 133 out of 144 sent queries received a reply, of which 66 partially or fully answered the question. 158 total emails were received (since an email might receive multiple responses), and 67 of these partially or fully answered the question. Differences in how the librarian’s reply addressed the user were evident. Hazel was the only one never referred to by her full name, whereas Jinghua was the least likely to be referred to by her given name and most likely to be referred to by her full name or no name at all. Greeting phrases were used in most responses. About 20% of responses included a reiteration of the original request. Elements of the response which could be seen as promoting information literacy skills were provided in only 11% of responses. Natasza was the most likely to be referred to another source to answer her query, whereas Jinghua was least likely. Ebunoluwa was the least likely to receive a response to her query and least likely to have her question answered overall.

Conclusion – The findings point to some evidence of unequal service provision based on unconscious bias. In the aggregate, Ebunoluwa received the lowest quality of service, while Jinghua received the highest. There were several instances of inappropriately addressing the user, or what the author refers to as name-based microaggressions, and this was most common for Jinghua. The likeliest explanation is that many librarians are unfamiliar with the ordering of names traditionally found in East Asian cultures. The most noticeable result of the study is an overall lack of consistent adherence to professional guidelines. For instance, most queries received a reply within a reasonable timeframe, and greeting and closing phrases were included almost universally. However, other elements of the author’s rubric, such as those corresponding to clarity and information literacy, were not consistently applied. The results point to a greater need for librarians to follow best practice in virtual reference services. Furthermore, the author believes that best-practice guidelines must actively engage with anti-racist ideas to address the issues that were found in the study.

Commentary

Described by the author as an email-based structured observation study, the study might also be called a correspondence audit, which has been a staple in many of the social sciences since the 1960s but is relatively novel in library and information science. The study discussed here is an adaptation to a European context of Shachaf and Horowitz’s (2006) audit study of 23 US academic libraries, investigating whether racial or religious bias could be detected in email-based virtual reference services. They had found that, by manipulating the sender name of the unaffiliated user asking a reference question, quality of service was affected, with African-American and Arab users more likely to be ignored, given a longer wait, or answered with peremptory responses compared to White, Christian, Asian, Jewish, and Hispanic users. A follow-up study two years later with a larger sample size failed to replicate the previous results; however, the methods were not identical (Shachaf et al., 2008). Hamer largely follows the earlier study and adapts it for the ethnic and racial distribution of England. Given the wide adoption of virtual reference services in libraries, it is important to investigate whether discriminatory behaviour, even if unintentional, is observed in these interactions.

This study was evaluated using Glynn’s (2006) critical appraisal tool. The study sample was representative of the population of interest, namely academic libraries in England. Informed consent was not obtained, but for reasons mentioned in the article, this could be justifiable in a study of this kind to minimize participants ascertaining the true aim of the study and altering their behaviour. The methods were clearly outlined and draw on earlier scholarship. The biggest problem for the study is that the author gives no evidence to believe that the unequal service provision described—namely, similar but non-identical counts across a couple dozen characteristics, with no discernible pattern of one group outperforming others on some composite measure—represents a meaningful sign of ethnic bias. A sample size of 158 split across six groups is simply not large enough to make conclusions about group differences. The author acknowledges this as a limitation but sees it as a problem for generalizability rather than power, that is, the ability to detect an actual effect if one exists. It is inevitable that librarian responses will differ slightly across professional guidelines (as well as other dimensions) in a naturalistic setting. The question is, are they systematic differences, above and beyond natural variation, that show a clear pattern of discrimination across plausibly related measures? The author does not address that question, so claims of ethnic bias are premature.

Notwithstanding its shortcomings, this paper offers many interesting takeaways for librarians, especially in the midst of the COVID-19 pandemic with many libraries continuing to emphasize their virtual reference services. More attention needs to be paid to improving adherence to professional standards and best practices. This study did not examine whether the mixed service quality provided to users was the result of their being unaffiliated with the institution or perhaps general burnout from higher-than-usual queries. If it is the latter, perhaps certain “shortcuts,” such as the use of (partial) canned message replies or GIFs to illustrate information literacy concepts, would make responding to emails less burdensome. A further takeaway deals with correctly addressing users by their name. To what extent are librarians familiar with name ordering practices around the world? Are there best practices for when to use honorifics or how to avoid misgendering? This seems like an underexplored topic that merits consistent and easy-to-apply standards.

References

Glynn, L. (2006). A critical appraisal tool for library and information research. Library Hi Tech, 24(3), 387–399. https://doi.org/10.1108/07378830610692154

Shachaf, P., & Horowitz, S. (2006). Are virtual reference services color blind? Library & Information Science Research, 28(4), 501–520. https://doi.org/10.1016/j.lisr.2006.08.009

Shachaf, P., Oltmann, S. M., & Horowitz, S. M. (2008). Service equality in virtual reference. Journal of the American Society for Information Science and Technology, 59(4), 535–550. https://doi.org/10.1002/asi.20757