Implementation and Evaluation of Impact on Student Learning of an Automated Platform to Score and Provide Feedback on Constructed-Response Problems in Chemistry

Introductory undergraduate chemistry courses are gatekeepers for STEM careers, particularly for underrepresented students. Findings from implementation of a novel platform for STEM problem-solving that automatically scores and gives students feedback for enhanced learning are presented. The platform uses classic AI with recursive algorithms and computer algebra processing. The platform tracks and scores the students’ solution pathway and displays the scores and descriptions for each correct step on the rubric as feedback for students to use on their next attempt. Teachers track student progress on a dashboard. Descriptions of the platform, student learning and perceptions, and differences from status quo solving and scoring are provided. 286 students used the platform. Scores increased with a large effect size. The platform distinguishes correctness of solutions beyond capabilities of human graders of students’ handwritten work. The types of errors students can make differ from manually solved problems.