A cross-cultural study of hindsight bias and conditional probabilistic reasoning

Abstract

Hindsight bias is a mistaken belief that one could have predicted a given outcome once the outcome is known. Choi and Nisbett (2000) reported that Koreans showed stronger hindsight bias than Americans, and explained the results using the distinction between analytic cognition (Westerners) and holistic cognition (Easterners). The purpose of the present study was to see whether hindsight bias is stronger among Easterners than among Westerners using a probability judgement task, and to test an “explicit–implicit” hypothesis and a “rule-dialectics” hypothesis. We predicted that the implicit process is more active among Easterners to generate hindsight bias, and that Easterners are more dialectical thinkers, whereas Westerners are more rule-based thinkers. French, British, Japanese, and Korean participants were asked to make probabilistic judgements in a Good Samaritan scenario (Experiment 1) and in a scenario including conditional probabilistic judgement (Experiment 2). In both Experiments, we presume that the implicit revision of causal models is made just by being given unexpected outcome information, and that explicit revision is made by being asked to point out possible factors for an unexpected outcome. In the results Easterners showed greater hindsight bias generally and it was greater in the Good Samaritan scenario. We conclude that the reason why hindsight bias was lower among Westerners is primarily that they tried to follow a rule to suppress the bias.

Keywords:

• Hindsight bias
• Analytic cognition
• Holistic cognition
• Cultural difference
• Probabilistic reasoning.

Acknowledgments

This research has been supported by a CHORUS grant from the French Ministry of Research and the Japanese Society for the Promotion of Science and by a grant-in-aid from the Japanese Society for the Promotion of Science (No. 19202012; The project leader is Yukinori Takubo). We thank Mike Oaksford, Yukinori Takubo, Incheol Choi, Linden Ball, and Hartmut Blank for their comments on this research. We also thank Junichi Taniguchi and Masasi Hattori for their arrangement of running parts of this experiment. Furthermore, we thank Jonathan St. B. T. Evans, Shira Elqayam, and an anonymous reviewer for their comments on our draft paper.

Notes

¹The term consequent is used as just indicating “q” in a conditional. It does not mean “outcome” as in the literatures of hindsight bias.

²Although we needed to avoid excluding ethnic minority people in the UK when we ran Experiment 2, we did not ask the participants to divulge their ethnicity. Approximately 25% of the British participants were in minority ethnic groups, although many of them were born and grew up in the UK. We asked French, Japanese, and Korean participants to divulge their ethnicity, and eliminated their data from the analysis. We eliminated all the minority ethnic participants from the analysis in Experiment 1. For the reference of the pattern of British ethnic group, just the means of these groups in the UK of Experiment 1 are shown here. The means of initial estimated probability of ethnic British were 63.5 in the non-outcome condition (n = 8) and 70.6 in the outcome condition (n = 16). Those of final estimated probability were 40.4 in the non-outcome condition and 68.4 in the outcome condition. Each SD is in parenthesis. The pattern is almost the same as that of British in Experiment 1.

³It might be problematic if the majors of some Japanese participants are not psychology, because Koo and Choi (2005) showed the effect of academic training in Korean participants. However, what they demonstrated was that oriental medicine students are more likely to engage in holistic cognition than psychology students. Oriental medicine requires people to think more and more holistically. Therefore, in the case of our Japanese participants, we do not suspect serious differences in training between psychology, sociology, and environmental sciences.