Integrating industry-crowdsourced projects in university capstone courses: A comparative study using parametric statistics and sentiment analysis

Abstract

This study explores integrating industry-crowdsourced projects within capstone courses of a 4-year Bachelor of Science program at an accredited American university. A unique business consulting model was developed for the final year course, aligning students with 16-weeks industry projects that reflected their academic goals and the program’s learning objectives. The study aimed to evaluate the efficacy of this pedagogical approach compared to traditional capstone courses. This evaluation involved collecting data from grading systems and anonymous course surveys. A novel aspect of the research design was the synergetic combination of nonparametric and parametric statistical techniques with modern machine learning (ML) algorithms to analyse the students’ grades, survey comments and third-party course opinion comments. Additionally, independent third-party course ratings were examined to triangulate the results. Findings revealed that while the academic performance in the industry-crowdsourced capstone course mirrored that of the traditional course, the industry-crowdsourced variant elicited significantly more positive responses in course surveys. Furthermore, ML sentiment analysis of comments from third-party forums indicated a stronger positive reception for the industry-crowdsourced course over the traditional approach.