Data are presented which display the performance of students in an introductory physics course that was conducted in a competency-based, multiple-opportunity format. These data reveal features of student behavior that are reproducible. The specific costs, measured in instructor's time per student produced are shown to be related to the character of the examination questions and to the comprehensional modes of the students. Specific costs for single-trial and multiple-trial formats are compared and the evidence clearly shows that the multiple-trial opportunity has lower specific costs and a much higher yield of students at the full competency level.