Many university students suffer from test anxiety that is severe enough to impair performance. Given mixed efficacy results of previous cognitive-behavior therapy (CBT) trials and a theoretically driven rationale, an acceptance-based behavior therapy (ABBT) approach was compared to traditional CBT (i.e., Beckian cognitive therapy; CT) for the treatment of test anxiety. In this pilot study, 16 university students with test anxiety were randomly assigned to receive either a CT or ABBT 2-hr group workshop. The two treatments produced markedly different effects on test performance (measured by exam scores), with those receiving ABBT experiencing improvements in performance, whereas those receiving CT exhibited reduced performance. In addition, there was a suggestion that ABBT might have been more effective at reducing subjectively experienced test anxiety (i.e., a nonsignificant but medium-sized group by time interaction effect). Implications of these results for the treatment of test anxiety and for theoretical notions related to cognitive change strategies are discussed.