The American screenwriter and film director Billy Wilder referred to the perfect vision of hindsight; “hindsight is always twenty-twenty”. But hindsight may also distort and prejudice our judgment of past events in a way that it is difficult to compensate for. A 28 year old paper, part of a doctoral dissertation submitted to the Hebrew University of Jerusalem, has been reprinted in Quality and Safety in Health Care under the heading “Classic Paper”.

The three experiments described in this paper, together with a fourth referred to briefly, show that thinking that we know the outcome of a series of events, even if what we believe about the outcome is false, tends to make us consider that that series of events almost inevitably led to that outcome and that we and others should have had the foresight to have predicted it. This tendency has been called “creeping determinism”.

In experiment 1, subjects (non-medical university students) were each asked to read one of four 150 word passages. Two of the passages each described a different series of historical events and two were clinical case descriptions. For each series of events four possible outcomes were provided, one of them the actual outcome. Subjects were divided into five groups in each case; one group given no information about actual outcome, one told the true outcome, and the other three groups each told that a different one of the three false outcomes was the actual outcome. They were asked to assign a probability for each outcome in the light of information given in the passage, the four probabilities to add up to 100%. In each case subjects given an outcome assigned a greater probability to that outcome than did subjects given no information about outcome. They also tended to assess the relevance of each piece of information provided in their passage according to the outcome they had been given. Thus the concept of creeping determinism was supported.

In experiment 2, subjects were asked to try to ignore what they had been told about outcome in assessing outcome probabilities. Again they tended to give a higher probability to the outcome they had been given and assessed the relevance of items of information accordingly. In experiment 3, subjects were asked to respond as they thought other students given no outcome data might respond. They again were biased towards the outcome they had been given. In the fourth experiment subjects were asked to predict the outcomes of a forthcoming event. They were later, after the event, asked to restate their predictions. Their memories were biased towards what they believed had proved to be correct predictions.

Interpretation of past events may be biased by belief (true or false) about the outcome of those events. This may lead to unjust criticism of self or others for reacting inappropriately to events that, in hindsight, appear to have led to an inevitable conclusion. It may also cloud judgments of history in general.