Twenty-eight test-anxious Ss received either "cue-controlled relaxation therapy," an inert placebo treatment, or no treatment. Once before and twice after the experimental treatment period each S responded to three test-anxiety scales and to one measure of public-speaking anxiety. Once after the treatment, Ss rated the credibility of the procedure to which they had been exposed. Pre- and posttreatment scores on the measure of (untreated) speech anxiety and posttreatment ratings of credibility both showed that the cued relaxation and placebo procedures exerted equivalent experimental demand for improvement. Scores on all three test-anxiety measures showed that, under these conditions of equivalent demand,the placebo procedure reduced self-reports of test anxiety at least as much as did cue-controlled relaxation. Implications are discussed.