Cognitive and conditioning models of cue-controlled relaxation as well as the relative efficacy of applied relaxation and cue-controlled relaxation in alleviating test anxiety were investigated. Test anxious subjects (N=18) received either applied relaxation, cognitive cue-controlled relaxation, or conditioned cue-controlled relaxation. Between group analyses revealed no differences among the conditions on any of the behavioral and self-report measures thus supporting the conditioning model of outcome effects. Within group comparisons indicated highly significant pre-post changes, with applied relaxation subjects demonstrating the greatest improvement. (Author)