The credibility of unsupervised online exams is an ongoing concern in higher education. Proctoring, in the form of physical or remote supervision, has been the main mechanism for maintaining academic integrity. However, both forms of proctoring are expensive and inconvenient. Several researchers have examined security mechanisms as a substitute for proctoring and obtained mixed results. This article describes a quasi-experimental study, the main goal of which was to examine the effectiveness of nonbiometric security mechanisms. The security mechanisms were selected based on the taxonomy of cheating reduction techniques rooted in the fraud triangle theory. The security mechanisms were considered effective if the scores were equivalent or lower on the unproctored exams. Two one-sided dependent t tests were used to test for equivalence of scores on two sets of proctored and unproctored exams in face-to-face (N = 704), hybrid (N = 91), and online (N = 55) introductory statistics community college courses. In the first set, the proctored exam was followed by the unproctored exam; in the second set, the order was reversed. In the first set, the scores on proctored and unproctored exams were equivalent in face-to-face and online groups, but students in the hybrid group had significantly lower scores on the unproctored exam. In the second set, the students’ scores were lower on the unproctored exam in all groups. The study’s results suggest that the used security mechanisms were effective.