Learning Data Analytics: Providing Actionable Insights to Increase College Student Success

• In Event: Faculty Development to Obtain Student Success: Learning Across National Science Foundation Grant Projects

Abstract

Higher education is increasingly moving toward data-informed decision making, driven by internal needs, external pressures, and emerging technologies. Institutions are facing increasing pressure to provide proof of learning, teaching practices are moving to an increasingly individualized and student-focused learning models and innovative technologies are allowing for greater mining of student data (Amey, 1999; Bichsel, 2012). The capacity to mine and store educational data created a proliferation of dynamic educational technology tools, many of which are incorporating educational big data, known as learning analytics (LA) (Ali, Hatala, Gašević, & Jovanović, 2012; Bichsel, 2012; Norris & Baer, 2013; Peña-Ayala, 2014). The data generated through LA tools have the potential to provide greater institutional insight, informed decision making, and institutional effectiveness (Bichsel, 2012; Dahlstrom, Brooks, & Bichsel, 2014; Norris & Baer, 2013).
Among the LA tools introduced to academia in the last decade are learning management systems (LMS) and early warning systems (EWS), which have begun to incorporate predictive algorithms based on a multitude of student and institutional data points. Many faculty and staff use these teaching and advising systems, with Blackboard, Moodle and EAB Student Success Collaborative being among the most popular (Dahlstrom et al., 2014). These tools hold potential to inform multiple levels in institutions by creating responsive feedback mechanisms that shape data-informed decision making (Baker & Yacef, 2009; Ben-Naim, Bain & Marcus, 2009; Bichsel, 2012; MacFadyen & Dawson, 2012; Mazza & Dimitrova, 2007).
The promise LA tools provide is one way for higher education institutions to respond “to internal and external pressures for accountability in higher education, especially in the areas of improved learning outcomes and student outcomes” (Norris & Baer, 2013, p. 11) from a conceivably more informed, dynamic, and efficient perspective. Despite the potential value of LA tools, there are barriers and challenges (e.g., lack of interest, knowledge, time, training, resources, incentives, institutional readiness, institutional commitment, etc.) related to broad adoption and use of these tools by the individuals within the organization who will ultimately decide to use them (Arnold, Lonn, & Pistilli, 2014; Bichsel, 2012; Norris & Baer, 2013).
The PIs will share findings from multiple focus groups with academic advisors, faculty, and students on challenges and opportunities for using LA tools. Results from the advisor and faculty focus groups indicate that a lack of alignment between LA tools and teaching/advising philosophies; a lack of accuracy, quality and trustworthiness of data related to tools; and institutional factors (lack of communication and training and bifurcated or closed systems), are among the strongest barriers for faculty and advisor adoption and use of LA tools. Preliminary results from the student focus groups indicate that organizational context and commitment, including structure, policies, processes, and leadership impact individual decisions to trust and adopt learning analytics tools. Further, results indicate the importance of a comprehensive, inclusive and well-communicated implementation plan for maximum adoption by users. Recommendations to improve institutional structures and norms to promote use of learning analytic will be reviewed. Finally, the authors will present concerns related to ethics and privacy.

Authors

• Jaime R. Lester, George Mason University

• Carrie Klein, George Mason University