A two-group randomized multivariate analysis of covariance (MANCOVA) was used to investigate the effects of cognitive-behavioral hypnosis in reducing test anxiety and improving academic performance in comparison to a Hawthorne control group. Subjects were enrolled in a rigorous introductory psychology course which covered an entire text in one quarter. In addition to randomization, two covariates were used as a statistical control for the selection threat to internal validity. The subjects were measured on two covariates and two dependent variables. The covariates were a midterm course grade for introductory psychology and a pretest on the Test Anxiety Inventory. There were a decrease in test anxiety and improvements in achievement for the hypnosis group. The treatment gains were maintained at a 6-week follow-up. The study suggests that cognitive-behavioral hypnosis is effective in reducing test anxiety and in improving academic performance. Contains 15 references. (Author/GLR)