Student use of out-of-class time was measured for four years in the introductory second-semester calculus-based physics course at the University of Arkansas. Two versions of the course were presented during the time of the measurement. In both versions, the total out-of-class time a student invested in the course explained less than 1% of the variance in test average and in normalized gain on the Conceptual Survey in Electricity and Magnetism. The details of how students used their out-of-class time explained from 21% to 36% of the variance in test average and 19% to 37% of the variance in normalized gain. The amount of time spent on the course increased as the course was made more difficult, but less than would be expected based on the changes in the course. Students changed their allocation of out-of-class time based on their performance within the course as the course progressed.