Abstract
Why do individuals interpret the same information differently? We propose that individuals form beliefs following Bayes' Rule with one exception: when assessing the credibility of experts, they double-dip the data and use already-updated beliefs instead of their priors. This “pre-screening” mechanism explains why individuals jointly disagree about states of the world and the credibility of experts, why the ordering of signals and experts affects final beliefs, and when individuals over- or underreact to new information. In a trading game, pre-screening generates excessive speculation, bubbles, and crashes. Our theory provides a micro-foundation for why individuals disagree about how to interpret the same data.