The fast-and-frugal heuristics approach to decision making under uncertainty advocated by Gigerenzer and colleagues (e.g., Gigerenzer & Goldstein, 1996) has achieved great popularity despite a relative lack of empirical validation. We report two experiments that examine the use of one particular heuristic-"take-the-best" (TTB). In both experiments the majority of participants adopted frugal strategies, but only one-third (33%) behaved in a manner completely consistent with TTB's search, stopping and decision rules. Furthermore, a significant proportion of participants in both experiments adopted a non-frugal strategy in which they accumulated more information than was predicted by TTB's stopping rule. The results provide an insight into the conditions under which different heuristics are used, and question the predictive power of the fast-and-frugal approach. (C) 2002 Elsevier Science (USA). All rights reserved.