Transforming Undergrad Engineering Education through Adaptive Learning and Student Data Analytics

Autar Kaw
University of South Florida

Need: Because of the pandemic, courses taught in a face-to-face flipped format had to be transformed to the online environment. One such course was a junior-level course in Numerical Methods at the University of South Florida. The quality of the pre-class preparation by the students continued to be a challenge in the online environment as well. This challenge was addressed by developing and deploying pre-class lessons using an online adaptive platform that personalized instruction for individual-student needs so all could achieve mastery. Adaptive platforms monitor student progress and modify instruction based on their performance and interactions.

Guiding Question: The research questions investigated the impact of the adaptive lessons on the classroom environment, student perceptions and attitudes, and exam performance in a flipped online course. In the control group, the pre-class learning was done via YouTube video lectures and quizzes conducted through the learning management system. This one-size-fits-all approach was replaced in the experimental group by adaptive lessons using the RealizeIT platform.

Outcomes: There were increases in all seven of the College and University Classroom Environment Inventory dimensions in the flipped online classroom with adaptive learning modality, with a significant increase in the Innovation dimension. The innovation dimension, which measures perceptions of novel teaching practices or activities, increased with an effect size of d= 0.54 (p=0.007). Relative to student perceptions, there was a significant decrease in the proportion of students who experienced load, burden, or stressors in the online flipped classroom when adaptive learning was available versus not (p = 0.036).
Multiple-choice final examination and concept‐inventory results were slightly higher when the adaptive lessons were available versus not. The multiple-choice results had an effect size of d=0.08 (p=0.594), while the concept inventory had an effect size of d=0.14 (p=0.370). The most promising exam and concept inventory results were for Pell grant recipients and community college transfers, who achieved effect sizes as large as d=0.40 with the adaptive learning.

Broader Impacts: These results suggest that adaptive learning can improve the student’s affective and cognitive outcomes, especially for nontraditional students who transfer from community colleges and those from low socioeconomic backgrounds. As this research is continued in flipped classrooms across multiple schools, we will have more evidence of the impact. Adaptive learning can address gaps in prerequisite knowledge and the challenge of all students mastering pre-class learning for flipped instruction. Also, via data dashboards, adaptive learning platforms inform the instructor in real-time of students who are struggling on a particular subject matter. Such knowledge can suggest when one-on-one interventions such as tutoring and study habits are needed.


Renee Clark, University of Pittsburgh