When performing an evaluation involves assessing multiple items of information about a target, evaluators can choose between two opposing strategies: reviewing all target information at once and forming an overall evaluation (a holistic approach) or reviewing each item of information independently and combining these independent reviews into an overall evaluation (a disaggregate approach). In four incentivized, pre-registered, online studies (N = 1,708), we compared the accuracy of these opposing approaches and people’s beliefs about which approach yields more accurate estimates. Across two experiments using different evaluative domains, we demonstrated that people make more accurate estimates using a disaggregate (vs. holistic) approach. However, a survey revealed that people anticipate greater accuracy with a holistic approach. Critically, we found that a brief training on the “wisdom of crowds” increases people’s preference for a disaggregate approach. We contribute to research on the wisdom of crowds and intuitive preferences for information and averaging.