Thesis Title

Hindsight Bias and Choice: Understanding Hindsight Bias and Perceived Control

Date of Graduation

Fall 2003


Master of Science in Psychology



Committee Chair

Harry Hom

Subject Categories



Hindsight bias, more commonly called the "knew-it-all-along" effect is an experience of certainty that one would have to be able to solve a problem or predict an event when the solution or outcome is already known (Fischhoff, 1975). It is one of many heuristics used in everyday information processing and decision making. Perceived control was defined by Langer (1975, p. 311) as "an expectancy of a personal success probability inappropriately higher than the objective probability would warrant." Some authors believe that the tendency to engage in illusions of control when making decisions results from a need to control which is salutary to mental health, well-being, and performance. In the proceeding study I attempted to link hindsight bias and control research using a direct control manipulation. I hypothesized that increasing one's sense of perceived control would lead to greater hindsight bias. Fifty-four participants, matched on ACT scores, participated. Using 20 pairs of general knowledge questions, participants were asked to choose from one of two questions of similar difficulty (or they were assigned a question to answer). Furthermore, particpants were provided with the outcome (hindsight) and asked to answer "as if" they did not know the solution; or they were not given the outcome and asked to supply one (foresight). A Likert-type scale was used ranging from 1 (very easy or very likely) to 10 (very difficult or very unlikely). Participants were asked to indicate how difficult the problem seemed and how likely they were to provide a correct response. There were two sessions of testing spaced one week apart. No significant interaction between hindsight bias and control was detected. The findings were consistent with past research in hindsight (there was a significant hindsight main effect) but not control (there was no significant main effect for control). Future research should include a control manipulation check.


© Maria Ciaramitaro