Riskier for me or for others? The role of domain and probability in self-other differences, in risky decision-making

Abstract

In social and economic interactions, people often decide differently for others, as against for themselves, under situations involving risks. This sometimes leads to conflicts or contradictions. Although previous studies have explored such contradictions, the findings have been inconsistent. To reconcile these inconsistencies, this paper investigates the role played by the different domains and probabilities in the self-other differences under risk. Two groups of participants completed a gambling task combining different domains (gain vs. loss) and probabilities (small vs. large). One group made decisions for others and the other group made decisions for themselves. The results revealed a four pattern of discrepancy: the ones who made decisions for others were less risk-seeking than those who made decisions for themselves over the small probability gains. This was reversed over the large probability gains. Conversely, the participants who made decisions for others were more risk-seeking than those who made decisions for themselves over the small probability losses. The results were reversed over the large probability losses. These results reconcile the contradictory findings of the previous studies and suggest the significant role played by contextual factors in such discrepancies.

Keywords:

• Self-other difference
• risk preference
• domain
• probability

Notes

Acknowledgments

This work was supported by National Natural Science of Foundation of China [No. 71801193]; Humanities and Social Sciences Foundation of Ministry of Education of China [No. 18YJC630155; 16YJC840006]; National Natural Science of Foundation of Zhejiang [No. LY17G010009].

Notes

1 Following Tverskyand Kahneman (1992) and Sun et al. (2018), we defined p < .5 as small probability and p ≥ .5 as large probability. These studies demonstrated that the risk preference of p = .5 is like that of p > .5. The Spearman’s correlation coefficients between the risk preference of p = .5 and that of p > .5 is greater than 0.80 for both decision targets, in both the loss and the gain domains, in this paper. Thanks an anonymous reviewer for raising this issue.

2 We wonder whether the proportion of gains and losses would also influence the difference in risky decision-making for oneself and others. A 2 (decision target) × 2 (domain) × 2 (probability) × 2 (money amount: x_i ≤ ¥200 or x_i > ¥200) was conducted. The results showed insignificant interactions between the amount of money and other variables (ps >.05), which was consistent with the study of Vieider, Villegas-Palacio, Martinsson, and Mejia (2016), in which non-significant differences were found in the value functions between the decision-making for oneself and others.

3 Given the significant effect of “task effort” in this paper, we tested whether such task efforts would mediate the relationship between the decision target and the risk preference. We thus conducted mediation tests using a bootstrapping procedure (5,000 bootstrap samples; Hayes & Preacher, 2014). The results showed non-significant mediation effects of “task effort” (small probability gains: 95% CI [–.02, .01]; large probability gains: 95% CI [–.01, .01]; small probability losses: 95% CI [–.01, .02]; large probability losses: 95% CI [–.01, .01]). This suggests, that the self-other difference in risk preference in the four scenarios are unrelated to task efforts.

4 We thank an anonymous reviewer for raising this explanation.