How might people revise their opinions on the basis of multiple pieces of advice? What sort of gains could be obtained from rules for using advice? In the present studies judges first provided their initial estimates for a series of questions; next they were presented with several (2, 4, or 8) opinions from an ecological pool of advisory estimates (Experiment 1), or with artificial advice (Experiment 2); finally they provided their revised estimates. Descriptive analyses of their revision process revealed that they egocentrically trimmed the opinion sets such that opinions distant from their own were greatly discounted. Normative analyses suggest that they gained substantially from the use of advice, though not optimally, due to their self-centered utilization of the advice. The results are discussed in connection with theories of belief revision and attitude change, with an emphasis on decision-makers' strategies for coping with conflicting opinions and the appropriateness of discounting distant or dissenting opinions. Prescriptive implications for the utilization of advice are also considered. (Contains 8 tables.)