Boredom Across Activities, and Across the Year, within Reasoning Mind

Since the beginning of educational assessment, there has been interest in modeling more than just what a student knows. However, the ease of measuring relatively simple knowledge and skills through multiple-choice tests has led to a field that has largely focused on what is easy to measure. In recent years, there has been increasing interest in automated assessment of students in a broader range of contexts, and for a broader range of constructs, than traditional assessment achieves. In this paper, we present automated assessment that can infer boredom, a key noncognitive factor during student learning. We study this construct in the context of Reasoning Mind, a blended learning system that integrates a range of learning activities, embedding them into story-based curricula. We asses student boredom across the year using sensor-free automated detectors developed using a combination of quantitative field observations and data mining, validated to generalize across students and school contexts. We then apply these detectors to data from the entire cohort of 70,000 students who used Reasoning Mind during the 2012-2013 school year. We demonstrate the use of the detectors by investigating how student boredom while using the Reasoning Mind blended learning program changes over the course of the year, and how different objectives and activities result in more or less student boredom. We find that while there is essentially no relationship between student boredom and the time of year, it is strongly related to the objectives in the self-paced Reasoning Mind curriculum; in particular, boredom on Reasoning Mind objectives is bimodal in character, with some objectives inducing considerable boredom while others do not.