Fifty Years of Hindsight Bias Research-Reflection on Fischhoff (1975)

Hindsight bias arises when people do not realize how extensively observing an event has changed their perception of the world. As a result, the event appears more likely than it actually was, in foresight. Underestimating how much one has to learn is a form of overconfidence that could, in the extreme, lead to a seemingly surprise-free past portending a surprise-full future. Fischhoff (1975) introduced tasks for studying the extent, causes, and consequences of the bias, along with initial evidence using historical vignettes. Subsequent studies have found the bias in a wide variety of experimental and real-world settings. Psychologists have linked the bias to research on cognitive, social, perceptual, and emotional processes. Other disciplines have implicated it in practical problems including clinical diagnosis, patent evaluation, legal adjudication, historical analysis, and safety engineering. Warning people about the bias has no discernible effect. Helping people to reconstruct past perspectives might help. After describing how the research program came about, the article briefly summarizes studies that the author especially likes and which have abundant references to studies in the diverse research areas that either study hindsight bias or use it to study other phenomena.