Enabling Fairer Interpretations of Learning Analytics Data Using Multiple Visualizations: A Co-Design Study

• In Event: Saturday Roundtable Session 11:25 am
  In Roundtable Session: Online Learning Across Contexts (Table 1)

Abstract

In this paper, we report on the development of multiple visualizations to enable fairer interpretations of learning analytics data. Our work focuses on the iterative co-design and development of an instructor-facing learning analytics dashboard to assist instructors in quickly assessing online student engagement. We used a machine learning model to automatically classify student discourse. Our first set of developed visualizations provided accurate information, but we discovered that the visualizations may not fairly represent student engagement. Based on the first round of design and development, we describe the co-design processes of developing additional visualizations to support fair and accurate interpretation of data. This work adds to literature that addresses fairness and ethics in the use of artificial intelligence in education.

Authors

• Priya Sharma, Pennsylvania State University

• Mahir Akgun, Pennsylvania State University

• Qiyuan Li, Boston University