Examining the Gap Between Students Self-reported Confidence and their Comprehension of Course Material

Previous studies have explored student self-reported confidence in several STEM fields. Most studies gather students’ self-reported confidence before and after an exam. In this study we focus solely on Computer Science students, and prompt them to self-report their confidence on each question during midterm and final exams. The goal of this research is to understand the gap between a student’s confidence and their mastering of a topic based on question performance. We also explore whether or not students can more accurately predict their confidence during the final exam compared to the midterm due to exposure of confidence prompts over the course of a term. We evaluated the exams to determine groupings based on question type. When conducting analysis we used Statistical Package for the Social Sciences (SPSS) to examine how confidence correlates with the correctness of the groupings. Our findings show that students more accurately reported their confidence on the final exam, compared to the midterm, for questions involving variables, functions, writing code and reading code. Furthermore, based on our analysis of the final exam, students were less accurate at reporting their confidence for questions involving loops and conditionals