Abstract
Self-efficacy is critical for academic success. Self-efficacy theory suggests that the social setting and problem format can influence self-efficacy. While robust studies assessing students’ self-efficacy and its relation to other academic outcomes have been undertaken, little is known about the relation between self-efficacy and competence on specific problems and how problem format and implementation type relate to self-efficacy. 94 conceptual questions in three formats (interpretation, ranking task, and multiple-choice) were developed through an iterative research-based effort. Ten university faculty implemented questions with their students in one of five types (closed-book exam, take-home exam, in-class quiz, homework, and group discussion). In total, 569 individual students responded to 2,006 questions. To capture item-specific self-efficacy (ISSE), students were asked to rate their confidence immediately after they had answered individual questions, on a 10-point scale. It was found that students with similar performance self-reported a wide variance of self-efficacy and that many student answers lacked correspondence between self-efficacy and performance. Results of multivariate analysis of variance (MANOVA) revealed that student performance differed according to question format and implementation. Students performed best on interpretation questions and they had the highest ISSE and percentage correctness for problems completed in groups. The findings support that interactive learning can contribute to student competence and self-efficacy.