We describe a method to administer a controlled, effective stressor to humans in the laboratory. The method combines the Trier Social Stress Test (TSST) and the Cold Pressor Test into a single, believable procedure called the Fear-Factor Stress Test (FFST). In the procedure, participants imagine auditioning for the reality television show Fear Factor. They stand before a video recorder and a panel of judges while (a) delivering a motivational speech, (b) performing a verbal arithmetic task, and (c) placing one hand into a bucket of ice water for up to 2 min. We measured subjective anxiety, heart rate, and salivary cortisol in three groups of young adults (n = 30 each, equal numbers of men and women): FFST, TSST, and Control (a placebo version of the FFST). Although the FFST and TSST groups were not distinguishable at the cortisol measure taken 5 min post-manipulation, at 35 min postmanipulation average cortisol levels in the TSST group had returned to baseline, whereas those in the FFST group continued to rise. The proportion of individual cortisol responders (≥ 2 nmol/l increase over baseline) in the TSST and FFST groups did not differ at the 5-min measure, but at the 35-min measure the FFST group contained significantly more responders. The findings indicate that the FFST induces a more robust and sustained cortisol response (which we assume is a marker of an HPA-axis response) than the TSST, and that it does so without increasing participant discomfort or incurring appreciably greater resource and time costs.