Assessing for Improvement: The Use of Artificial Intelligence to Uncover Potential Differential Impact of Assignments

• In Event: 77.051 - Student (and Robot?) Perceptions and Uses of Assessment

Abstract

The percentage of engineering and computer science degrees awarded to women is incredibly low, and engineering culture can be partly to blame. This study assessed the impact of three assignments created to promote diversity, equity, and inclusive teamwork in a computer coding class for engineering students. Using written reflections, a multi-method text analysis was performed using Artificial Intelligence, specifically semantic and topic analyses using natural language processing, and social network analysis to determine the underlying structure of response data, emotional engagement, and polarity and to explore the differential impact by gender. Results indicate men were more likely to be emotionally responsive from an outsider’s perspective while women engaged or disconnected from content depending on the context of any given task.

Authors

• Karen E. Rambo-Hernandez, West Virginia University

• Abhik Roy

• Melissa Morris, West Virginia University