Probability Errors

Physicians and patients deciding on treatment options often need to estimate the probability of various outcomes (e.g., death, partial recovery, full recovery) in order to make effective decisions. However, human probability estimation is often fraught with errors that may interfere with making correct decisions. There are a number of probability errors that physicians ought to be aware of.

Human probability judgments often exhibit a rather basic bias: We overestimate low probabilities and underestimate high probabilities. For instance, people overestimate the likelihood of dying from rare diseases such as smallpox or botulism while underestimating the likelihood of dying from more common afflictions such as strokes or heart disease. One explanation for these biases might be that our probability estimates tend to regress toward the mean. That is, ...