Using Random Forests to Identify Factors of Student Motivation in a Project-Based Learning Course

Abstract

Grounded in motivation theory, the purpose of this research is to use random forest analysis to identify factors of motivation of students who participate in a project-based learning experience. Our research aims to answer: 1) How does project-based service learning affect students’ motivation as compared to conventional (non-service) project-based learning? 2) How are women affected differentially by project-based service-learning? The research, which began in 2011, was completed over a two-year period. The students and activities in Component Design, an existing junior-level course at the University of Colorado at Boulder served as the research focus. Specifically, project-based service-learning curriculum was implemented into a required design and build activity for Component Design students. Using a conventional design project as the control, how the context of project-based service learning affects aspects of student motivation was studied. This paper discusses the research design, theoretical framework, data analysis methods, and random forest results. Our findings indicate that students’ initial non-technical skills were the most important predictor of motivation in the conventional project-based learning experience and that students’ perceived value of the course and the project were the most important predictors of motivation in the project-based service-learning experience.