Data Management Practices in Academic Library Learning Analytics: A Critical Review

Abstract

INTRODUCTION Data handling in library learning analytics plays a pivotal role in protecting patron privacy, yet the landscape of data management by librarians is poorly understood. METHODS This critical review examines data-handling practices from 54 learning analytics studies in academic libraries and compares them against the NISO Consensus Principles on User's Digital Privacy in Library, Publisher, and Software-Provider Systems and data management best practices. RESULTS A number of the published research projects demonstrate inadequate data protection practices including incomplete anonymization, prolonged data retention, collection of a broad scope of sensitive information, lack of informed consent, and sharing of patron-identified information. DISCUSSION As with researchers more generally, libraries should improve their data management practices. No studies aligned with the NISO Principles in all evaluated areas, but several studies provide specific exemplars of good practice. CONCLUSION Libraries can better protect patron privacy by improving data management practices in learning analytics research.

A Comprehensive Primer to Library Learning Analytics Practices, Initiatives, and Privacy Issues

“…Briney found that only about 10 percent of published library LA studies mention data security practices. 133 These findings are echoed in Perry et al, where just over half of the survey respondents claim that they have security protections (physical and/ or digital) on LA data, and only a third secure data during transit. Additionally, fewer than half of respondents had a retention schedule for LA data, and a full 20 percent planned to hold onto these data "indefinitely."…”

“…116 In a recent piece, Jones, Rubel, and LeClere argue institutions are beholden to their students as "information fiduciaries," that they have a moral obligation to use LA to benefit students and respect the trust students place in their university. 117 Given these known ethical problems and quandaries related to LA, Asher et al argue that librarians must update their sense of threats to user privacy by taking into account algorith-mic bias, reidentification, widespread data trading and hoarding, and surveillance creep. 118 Briney homes in on research ethics and LA projects, noting that, even though institutional review boards (IRB) are responsible for identifying ethical issues, many miss violations of library privacy ethics.…”

Quantifying Library Engagement: Aligning Library, Institutional, and Student Success Data

“…22 In a recently published critical review of the literature about data-handling practices from 54 learning analytics studies in academic libraries, Briney found many examples of inadequate data management practices, including "extended data retention, a broad scope of data collection, insufficient anonymization, lack of informed consent, and sharing of patron-identified data." 23 Briney further recommended that researchers become more transparent in documenting and sharing how they have worked through legal requirements, policies, and practices related to data handling in learning analytics' projects. In this article, the authors describe how they incorporated the NISO Privacy Principles, the Code of Practice for Learning Analytics, and Briney's recommendations as they designed the present study.…”

Unethical Numbers? A Meta-analysis of Library Learning Analytics Studies

“…Datasets created for learning analytics activities are particularly vulnerable to re-identification, even after de-identification and anonymization techniques have been utilized. (Asher 2017;Briney forthcoming;Metcalf & Crawford 2016). This vulnerability potentially exposes these data's constituent individuals and populations to unintended disclosures and insufficiently considered reuse or misuse by unexpected actors, including commercial, governmental, or law enforcement interests, such as investigatory requests and subpoenas.…”