Valuing Social Responsibility in the Era of Data Analytics: A Process Model for Effective Practice

Valuing Social Responsibility in the Era of Data Analytics: A Process Model for Effective Practice

Melissa R. Irvin (University of South Florida, USA)
DOI: 10.4018/978-1-7998-2177-9.ch011

Abstract

Higher education is increasingly interested in utilizing data analytics to support all aspects of university operations, including enrollment management and learning outcomes. Despite potential benefits to improve results and resource efficiency, the use of student information and the creation of predictive models is a potential minefield which could undermine larger higher educational missions tied to civic responsibility and social mobility. Questions remain as to the impacts of predictive modeling on underrepresented communities like students of color and differently abled students. Emerging research on similar fields of analytics, including predictive policing, provides a window into the ethical considerations that must be made to use data analytics responsibly. This chapter uses the construct of social responsibility to propose a process model for the responsible use of data analytics in colleges and universities derived from Carroll's Pyramid of Corporate Social Responsibility.
Chapter Preview
Top

Introduction

Big data is not going anywhere. Data-driven decision making plays an integral role in industries extending beyond higher education, including healthcare, marketing/sales, and social services programs (Ekowo & Palmer, 2016; Picciano, 2012; van Barneveld, Arnold, & Campbell, 2012). In higher education, growing pressures from various stakeholders to change and improve outcomes more quickly and effectively have many administrators investigating the promise of analytics to solve the unsolvable. Adelman (2006, p. 103) began a summary of longitudinal studies examining postsecondary student success with a recommendation to “fortify their institutional research capacities and integrate them more intimately with academic advising and course scheduling” to better support retention and graduation. Though it does not use the same terminology, this was a direct call to harness big data and use analytics to address higher education’s biggest challenges.

Modern advancements in technology have rapidly expanded the scope and capabilities of analytics. Institutions can mine vast pools of unused student data to identify trends that shed light on perennial student access and success problems (Yeado, Haycock, Johnstone, & Chaplot, 2014). Fledging data-informed reform movements need more reliable methods for tracking and storing large amounts of student information and key performance metrics using complex student management technologies (Yeado et al., 2014). This demand has driven an explosion of newly developed technology applications and software related to student data, tracking, and analytics. Over 120 vendors have developed a variety of data management, tracking, and communication tools in the hopes of providing a solution to help institutions reach their student success goals (Kalamkarian, Karp, & Ganga, 2017; Tyton Partners, 2015). Unfortunately, the spread of data analytics has not been accompanied by a framework to ensure university commitment to ethically driven practices.

Growing concerns about the use of big data within criminal justice provides insight into the dangers associated with the use of predictive analytics in social institutions. “Predictive policing” uses trends in large data sets to proactively identify areas and individuals at high risk for criminal activity (Kutnowski, 2017). An increasing amount of research provides early evidence that these models can be used to perpetuate existing biases against low-income and communities of color (Karppi, 2018; Kutnowski, 2017). These concerns are increasingly relevant for the use of predictive models for student success that indicate a student’s likelihood for persisting into the next semester or graduating with a degree. To ensure that these models are used for the benefit of students and reduce the likelihood of unintended harm, universities must consider their responsibilities as learning organizations, corporate citizens, and members of society at large. Can the concepts of corporate social responsibility provide a relevant and clear framework for building actionable guidelines for the ethical and effective use of predictive analytics?

This chapter, in examining Carroll’s (1991) Pyramid of Corporate Social Responsibility, aligns with concepts integral to understanding social responsibility within the context of higher education. This model will be used as a structural framework for examining the considerations that must be made to develop guidelines for the ethical use of data analytics. This chapter will also present a clear process for the use of predictive analytics in student success that builds on corporate social responsibility research, guidelines related to socially responsible practices, and the ethical use of data. This is an increasingly pressing issue as more institutions invest in analytics platforms, placing high expectations on the ability of algorithms to improve outcomes. Finally, this chapter will consider strategies to effectively involve different stakeholder communities to communicate expectations related to the use of analytics as well as to identify and manage risk.

Key Terms in this Chapter

University Social Responsibility (USR): The obligation of higher education institutions to promote and create knowledge and social norms in support of societal values and expectations.

Predictive Analytics: The use of large data sets and statistical analysis to uncover patterns and trends to predict outcomes and behaviors.

Data Mining: A comprehensive, in-depth exploration of large data sets.

Academic Analytics: The use of various analytic methods to support decision making in higher education institutions related to organizational management and fiscal responsibilities.

Data Warehousing: A system designed to store, protect, and maintain large sets of data.

Learning Analytics: The use of data to identify specific curricular or instructional strategies to support students in achieving identified learning outcomes.

Analytics: The use of large data sets and statistical analysis to answer questions and support decision making.

Complete Chapter List

Search this Book:
Reset