Increasing numbers of universities are offering courses in online and hybrid formats. One challenge in online assessment is the maintenance of academic integrity. We present a thorough statistical analysis to uncover differences in student performance when online exams are administered in a proctored environment (i.e., in class) versus an unproctored environment (i.e., offsite). Controlling for student grade point average (GPA), no significant differences in mean overall course performance or exam performance between the two groups were found, nor were there any differences in the mean vectors of individual exam scores. The study reveals that the group taking online exams in the unproctored environment has significantly more variation in their performance results. In examining potential causes of the greater variation, analyses were performed to assess whether an increased level of possible cheating behavior could be observed from performance results for students in the unproctored section. No evidence of cheating behavior was found.