The effects of presenting test items in random order or in a sequence parallel to the order of presentation were studied by testing 92 undergraduates in an introductory psychology course at Emporia State University (Kansas). Two test forms, sequential (S) and random (R), of multiple-choice questions were prepared for four 1-hour examinations administered to three sections of classes using the same text. The students were randomly assigned to receive the four examinations according to the following orders: SRRS, SSRR, SRSR, RRSS, RSSR, and RSRS. A two-factor, split-plot analysis of variance was used to analyze the scores obtained. There was no instance in which a sequential test resulted in a higher score than a random test. It appears that instructors can use the random-order option offered by current test bank software with confidence that this will not have a negative impact on student performance. (SLD)