Causal vs. Statistical Base Rates and Learning

Key Insight

Base rates are treated differently depending on whether they are perceived as statistical or causal. Statistical base rates are abstract facts about a population (e.g., 85% of cabs are Green) that do not directly explain an individual case, and these are often underweighted or neglected when specific information is available. In contrast, causal base rates imply a reason for individual events or traits (e.g., Green cabs are involved in 85% of accidents, suggesting reckless Green drivers) and are readily incorporated into judgments.

When base rates are framed causally, they are used to form stereotypes, which System 1 easily applies to individual cases. For example, knowing that Green cabs cause most accidents leads to a stereotype of 'Green recklessness,' which then influences judgments about individual Green drivers. This illustrates the mind's strong preference for information that fits into a coherent causal story, over 'mere statistical facts.'

This distinction also impacts learning. People often struggle to learn effectively from surprising statistical facts, especially if those facts contradict their existing beliefs; they tend to 'quietly exempt themselves' from the conclusions. However, surprising individual cases, particularly those that present a clear incongruity with expectations, are far more effective in changing beliefs because they force the mind to resolve the inconsistency and embed it into a new causal narrative.