Self-interested behavior has been well identified in dictator games: people allocate little or nothing to serve their interests better. Although it has long been recognized that beliefs about others are critical in interpersonal interactions, the mechanism by which self-serving beliefs result in self-serving behavior remains unclear. To explore the relationship among the conflict of interest, belief formation, and the resulting behavior, this study designed a three-stage reciprocity game with belief elicitation and regret option and conducted a laboratory experiment using university students as the subject pool. By asking recipients to predict the allocators' decisions and incentivizing the prediction, we discover that people rationalize their own self-serving bias by believing that others are similarly self-interested and then returning the 'favor' in the name of fairness. Additionally, this study examined post-decision regret by the reciprocators, and the asymmetry between upward and downward adjustments indicates a self-serving pattern: Reciprocators who under-estimated the allocated amount are less likely to make an upward adjustment compared to the likelihood of downward adjustment made by those who over-estimated the allocated amount. This study contributes to the literature on self-serving bias by better understanding the presence of self-serving beliefs and self-serving behavior.
Article
Self-serving Regret: Experimental Evidence from a Reciprocity Game
https://doi.org/10.21203/rs.3.rs-2067703/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
self-serving
regret
reciprocity game
experiment
Subjects in various experiments indicate a conflict between acting in their self-interest and upholding moral standards1–3. Although people incline to be fair and committed to the moral rule, they deviate from such ideals when they can excuse it4,5. It suggests that people value the appearance of fairness over their actual adherence to moral rules6,7 because people could benefit from maintaining a positive social self-image6,8−10.
How can people maintain their positive self-image while behaving selfishly? A large body of evidence suggests that people allocate resources by trading off their material self-interest with their desire to be perceived as fair3,6,11. According to this view, people tend to interpret fairness in self-serving ways and adopt self-serving beliefs to justify their selfish behavior12–14. A variety of motivated reasoning and tactics contribute to the formation of self-serving beliefs. People process information in motivated ways to serve their self-interest, such as: taking advantage of uncertainty, behaving more selfishly when the impact on others is ambiguous, and deliberately avoiding information to reduce altruism3,7,15−17. Some people also tell lies to maintain a positive self-image18,19. People in different power situations prefer to see fairness in ways that benefit themselves the most20. Moreover, people tend to misremember the extent of their selfishness, rationalize their immoral acts and thereby protect their moral self-image21. As the formation of self-serving beliefs has been well researched, a gap remains unclear: whether self-serving beliefs about others serve as a mechanism via which individuals might act selfishly and still feel moral. One possibility is that people may convince themselves that others are likewise selfish and immoral, so they have a legitimate motive to behave self-interestedly while retaining the appearance that they treat others fairly.
In this study, we designed a multi-stage reciprocity game with belief elicitation to test this possibility. Based on Ben-Ner et al.'s setting, the first stage is a conventional dictator game (the allocators are unaware of the multi-stage structure), and the recipients (reciprocators) in the first stage become allocators in the second stage22. We modified this setting by adding a belief elicitation stage. The reciprocators were asked to estimate their anonymous partners' decisions, revealing their beliefs about others' morality. Some estimates were monetarily incentivized to induce their true beliefs. If one does not adopt self-serving beliefs, there should be little difference between the incentivized and un-incentivized estimates. Instead, if they purposefully consider others selfish and immoral, such biases should be reflected in estimates.
Another key focus of this work concerns post-decision behavior. According to extensive research, regret occurs when a decision outcome is worse than it would have been if people chose a different option23. Numerous studies have examined the psychological and economic factors that lead to regret in decision-making. According to the decision justification theory (DJT) proposed by Connolly and Zeelenberg23, the two primary components of decision-related regret are the comparative evaluation of the outcome and the feeling of self-blame for making a poor decision. Most studies focus on the former component, mainly contributing to post-purchase mentality24, risky financial decision-making25, and individual healthcare decisions26. While these studies focused on regret due to personal practical loss, regret may also occur in social and interpersonal contexts where decisions put the interests of others at risk. Evidence shows that violating fairness standards may immediately induce negative emotions such as guilt and regret27,28. However, when choosing for others as opposed to oneself, individuals are less mindful of regret (and are more strongly motivated by larger risks)29. Additionally, ego depletion strengthens feelings of internal struggle and decision regret30. Previous studies also investigated the impact of anticipated regret and regret aversion on decision-making. When people feel more responsible for the decision outcomes31,32, or there is uncertainty28,33, goal attainment34, and descriptive norms and expectation of others29, the decision maker tend to be more scrutinized and avoid regret.
As we predicted, people endorse self-serving beliefs to maximize their interests. It is unclear if people regret when their motivated thinking (i.e., self-serving beliefs) is revealed. In this study, we gave participants a regret option. We are particularly interested in how people would alter their previous decisions, in which direction, and to what extent. We predicted that when people find their beliefs about others are biased and self-serving (e.g., believing others to be stingy, but they and not), they would be a little more generous than they had been. In contrast, people will regret their generosity if they are overly giving while others fall short of their expectations (e.g., believing others to be generous, but they are actually stingy). Thus, we aim to probe into this regret pattern and its mentality.
Our study involves 8 Treatments, each with 2 or 3 stages. In all Treatments, participants made a series of decisions in which they chose how to allocate tokens between themselves and a particular anonymous partner. In the first stage, allocators are endowed with ten tokens and are instructed to play a conventional dictator game with a recipient (reciprocator). The allocators reported contributing an average of 3.09 tokens (N = 128, M = 3.09, SD = 1.98, Min = 0, Max = 8), consistent with the literature indicating that dictators allocate 20 percent to 30 percent of the endowment5. The reciprocators then returned the favor by making the same allocation decision without knowing the number that the allocators had given to them. They have allocated a mean of 2.68 tokens. We also asked some reciprocators to estimate the number and gave a bonus for more accurate estimates. Later, we informed the reciprocators of the number of tokens that allocators had initially given to them. In addition, we reminded them of their allocations and estimates. Given the information, reciprocators have the option of adjusting their earlier allocations.
Reciprocity and Motivated Self-serving Beliefs
We searched for evidence that the self-serving bias is motivated by the self-serving beliefs of others' morality (e.g., the belief that others are selfish). We examined this possibility by probing the subjective beliefs of reciprocators regarding the number of tokens allocated to them. As the reciprocators were to reciprocate back the favor, we also manipulated the sequences of estimation and allocation of the reciprocators to see whether the timing of allocations affects estimations or vice versa. In Treatments 2 and 4, the reciprocators estimated before they made allocations, whereas, in Treatment 3, the sequence was inverted.
In addition, we introduced monetary incentives to motivate reciprocators' estimates to be accurate. In Treatments 2I, 3I and 4I, we offered monetary incentives that scaled directly with estimating accuracy. We anticipated that incentives would elicit reciprocators' true beliefs of the allocators. We hypothesized that those reciprocators who think allocators to be fair and generous (i.e., those who estimated the allocators to give them a fair number of tokens) would reciprocate this generosity (i.e., return a fair number of tokens to the allocators). In contrast, those who believe others to be selfish (i.e., those who estimated the allocators to give them nothing or few tokens) would do the same to them (i.e., return nothing or few tokens to the allocators).
To test this hypothesis, we dotted each pair of estimates and allocations for each reciprocator (N = 64). As predicted, despite the outliers, we found a strong, positive, linear correlation between estimates and allocations (see Fig. 1). The linear regression analysis (P < 0.001) also supports this conclusion (see Supplementary Information). This finding supports the idea that when individuals consider others to be giving, they will reciprocate by acting generously towards them.
Result 1 (Reciprocity): Reciprocators allocate more to those allocators they believe would do the same to them, indicating a substantial pattern of reciprocity.
Another main issue we focused on is whether self-interested reciprocators would endorse motivated, self-serving beliefs. We hypothesized that regardless of how the reciprocators genuinely perceive the allocators, they would prefer to distort their moral beliefs of them (i.e., believe that the allocators are more self-interested than they truly are). This belief would justify their selfish behavior, provide a justification for themselves behaving selfishly (e.g., allocate few or nothing), and eventually increase their payoffs. If our hypothesis is correct, the incentivized estimates should be closer to the actual number of tokens distributed and greater than those without incentives. Moreover, the estimates without incentives should be closer to the number returned by the reciprocators.
Across all Treatments, the allocations and the estimations were non-normally distributed (W = 0.84-89, p < 0.001; Shapiro - Wilk test). Thus, we report non-parametric statistics for key comparisons. All reported p-values are two-tailed.
We pooled the allocations of all the allocators (N = 128) across all treatments, as their allocations were not affected by other factors. To test our hypothesis, we compared the un-incentivized estimates of the reciprocators (E) with the actual number of allocations (AA). We found no significant difference between these two groups of data (W = 4230.5, p = 0.95; Mann–Whitney U test). Similarly, we also found no significant difference between the incentivized estimates and the actual number of allocations (W = 4195, p = 0.86; Mann–Whitney U test). Nevertheless, even if the reciprocators show no significant self-serving beliefs on the statistical level, we compared the mean value of these data, which would still partially support the hypothesis of self-serving beliefs (see Fig. 2). The average number of tokens the reciprocators gave (AR) was 2.63 (N = 64; SD = 2.20). The mean value of the reciprocators' un-incentivized estimates (E) is 2.84 tokens (N = 32; SD = 2.07), whereas the mean value of the incentivized estimates is 3.13 tokens (N = 32; SD = 1.74). We can see that the estimates with incentives were more accurate on average and higher than those without incentives. Moreover, the mean estimate without incentives was closer to the number of tokens returned by the reciprocators. These results are in accordance with our hypothesis.
In Treatments 2 and 4, the reciprocators made the estimations first and then the allocations. To rule out the possibility that the order between estimation and allocation could affect the reciprocators' elicited beliefs, in Treatment 3, the reciprocators made the allocation decisions first, followed by estimations. We found no difference in the reciprocators' estimations across Treatments 2, 3, and 4 (W = 258.5, p = 0.83; Mann–Whitney U test), indicating that the order of estimations and allocations has little effect on the reciprocators' beliefs regarding the amount allocated by the allocators.
Result 2 (Self-serving Beliefs): The order between estimation and allocation does not significantly affect the reciprocators' elicited beliefs on the allocated amount by allocators. Reciprocators' estimations are insignificantly lower in un-incentivized treatments than in incentivized treatments, providing weak evidence for self-serving beliefs.
Self-serving Regret
To test whether people regret when they realize their beliefs about others are biased, we allowed reciprocators to reallocate. We predicted that the reciprocators would make different adjustments under different circumstances: compared to the actual number given by the allocators, if the reciprocators had under-estimated, they would either slightly upward their previous allocation or make no adjustments; if they had over-estimated, they would downward their previous allocations without hesitation; if their estimations were accurate, they would either choose not to make adjustments or downward their allocations. Thus, we predicted that the reciprocators who under-estimated the allocated amount were less likely to make an upward adjustment compared to the likelihood of a downward adjustment made by those who over-estimated the allocated amount. In addition, the average range of the upward adjustments should be less than that of the downward adjustments. To test this hypothesis, we examined the direction and the intensity of reciprocators' regret (i.e., discrepancies between each reciprocator's estimations and reallocations). If our hypothesis is correct, then such post-decision regret should tend to be self-serving: there is an asymmetry between upward and downward adjustments.
We divided the reciprocators into an under-estimated group (N = 31), an over-estimated group (N = 21), and an accurate-estimated group (N = 12). Regarding the direction of the adjustments, 61.3 percent made upward adjustments for the under-estimation group. Among over-estimators, 80.9 percent made downward adjustments, which outnumbers the former by more than 33 percent. Even with a precise estimation, 16.7 percent of the reciprocators still chose the downward alternative (see Fig. 3). Regarding the range of adjustments, on average, reciprocators who over-estimated opted to decrease 1.9 tokens compared to their initial allocations. In contrast, those who under-estimated increased 1.41 tokes, and those who estimated accurately decreased 0.33 tokens (see Fig. 4). Such a self-serving pattern aligned with our hypothesis: reciprocators who over-estimated (i.e., thinking too good of others) chose to make downward adjustments without hesitation, while those under-estimated (i.e., thinking too bad of others) tended to make slightly upward adjustments.
Still, there is a possibility that reciprocators regret because of the differences in allocations between them and the allocators in the first two stages, which directly affect their payoffs. We compared the allocation differences and examined how they adjusted in reallocation. Our predictions are similar to the above. Compared to the actual number allocated by the allocators, if the reciprocators had under-allocated, they would either slightly upward their previous allocation or make no adjustments; if they had over-allocated, they would downward their previous allocations; if their allocations are the same, they would either choose not to make adjustments or downward their allocations. We also divided the reciprocators into three groups: the under-allocated group (N = 40), the over-allocated group (N = 25), and the same-allocated group (N = 15). Regarding the direction of the adjustments, as for the under-allocated group, 60 percent made upward adjustments. Among over-allocators, 92 percent made downward adjustments. One-fifth of the reciprocators still chose the downward alternative for the same-allocated group (see Fig. 3). Regarding the range of adjustments, on average, reciprocators who over-allocated decreased 2.24 tokens compared to their initial allocations, while those who under-allocated increased 1.50 tokens, and those who allocated the same as the allocators decreased 0.53 tokens (see Fig. 4).
Result 3 (Self-serving Regret): Reciprocators adjust their allocations more in the over-estimated group than in the under-estimated group and also more in the over-allocated group than in the under-allocated group, in terms of both extensive margin and intensive margin, exhibiting a significant pattern of self-serving regret.
In this study, we investigated two aspects of self-serving bias: how people formed their self-serving beliefs about others and how they regret when their beliefs about others are proven to be self-serving. Across all 8 Treatments, we observed self-serving bias in both motivated beliefs and post-decision regret. Belief-wise, people reciprocate with those they perceive to be generous or benefit others, yet selfish people tend to form self-serving beliefs to rationalize their self-interested behavior. They intentionally believe that others are selfish or unkind so that they can treat others similarly. This finding is consistent with the literature that people avoid altruistic actions by distorting beliefs about others' altruism35; Regret-wise, we observed post-decision regret by the reciprocators. There is an asymmetry between upward and downward adjustments in both extensive and intensive margins, indicating a self-serving pattern. Specifically, when people behave more altruistically but find it not worthwhile, such as overestimating another's moral character, they immediately regret it and will not hesitate to penalize the other. However, when people purposely distort their moral beliefs of others (i.e., believing others to be less moral than they are) and when they find that the beliefs are biased, they will only experience mild regret and slightly compensate for their self-serving beliefs. We also found it interesting that even if people do not hold self-serving beliefs about others, they nevertheless tend to reduce altruism and serve their own interests when given a regret option.
In line with previous studies, we found the tension between the desire to hold a positive self-image and actual behavior, which is resolved by developing subjective, self-serving beliefs regarding the morality of others12–14, 35. Our findings are also consistent with earlier research indicating that individuals engage in "moral wriggling," i.e., they find justifications for acting self-interestedly without making it apparent to themselves or others7. In terms of the method, we extended Ben-Ner et al.'s structure. We used a multi-stage reciprocity game and not only investigated reciprocity between participants but also elicited the moral beliefs of the reciprocators regarding the allocators, in contrast to the methods in the literature in which allocators were asked to assess how the recipients would accept the allocation35. This finding may help us to understand the mentality behind selfish or unkind behavior. In daily life, consider an employee who exerts low effort because he believes his boss does not value him. In a broader economic and political context, a country could accuse other countries of having high CO2 emissions to justify its high CO2 emissions.
More crucially, this study focuses on post-decision regret and examines self-serving bias in regret. As the causes of regret (particularly intrapersonal regret) have been extensively studied, we investigated how individuals regret and its intensity in interpersonal contexts. This research fills a gap in the literature on self-serving bias and interpersonal regret. The observed self-serving regret extends the findings of Corbett, et al. 36 that individuals who are kinder and more prosocial are more likely to experience interpersonal regret than those who choose not to be. In contrast to Corbett et al.’s study, which focused on regret in children, we found substantial variations in regret levels among adults. One possible explanation for the self-serving pattern in regret is that prosocial persons are more likely to feel regret because others let them down, but those who show less prosocial tendencies only aim to maintain a favorable self-image and thus are less likely to experience regret. This finding has important implications in interpersonal contexts. In order to avert stronger feelings of regret, people might choose not to be prosocial in the first place. People might be more likely to avoid altruism by distorting their moral beliefs about others.
However, our experiment has the following limitations. We did not find evident evidence for self-serving beliefs, most likely due to the three factors below. One is that the number of participants was restricted. We had only 16 reciprocators in each Treatment, and 32 participants participated in estimations, so we could only collect a small amount of data for analysis. The second is that the monetary incentives were not so appealing that even a 30 percent deviation in estimation would not affect participants' payoffs. Thus, the participants' incentivized estimates during the belief elicitation stage were not significantly different from those of the non-incentivized estimates. So, we were unable to obtain the most accurate beliefs about others. Another one lies in that 84 percent of our participants were female. As previous studies have proven that female is significantly more generous than men37, and that people expect women to be more altruistic than men38, this could influence the accuracy of our data. Therefore, future research studying self-serving bias in beliefs could improve these flaws, which could help better understand the formation of self-serving bias.
Data source and experimental setting. The data was generated from our laboratory experiments conducted in the Behavioral and Experimental Social Science Laboratory (BESSL) at Beijing Foreign Studies University (BFSU). The study was approved by the Ethical Review Committee of the Behavioral and Experimental Social Science Laboratory at Beijing Foreign Studies University (Reference No. 2022L004). The experiments were programmed by z-Tree39, and the experiments were implemented in accordance with the relevant guidelines and regulations of Beijing Foreign Studies University for studies with human participants. Before conducting the experiments, we obtained informed consent from all participants.
In all, we had 8 treatments. 256 participants participated in our experiments (214 females, 42 males; mean age = 21.0), and 32 participants in each Treatment. All participants were graduate and undergraduate students from BFSU.
At the beginning of each Treatment, participants were randomly assigned to the roles of allocators or reciprocators and were subsequently matched in pairs. In the instructions, the allocator was labeled as "Player 1" and the reciprocator as "Player 2". Participants were informed that they remained in their roles throughout the whole Treatment. In Stage 1, the allocator's task across all treatments was to decide 10 tokens to be distributed between themselves and the reciprocator. Each of them was informed that his/her decision would influence the payoff of himself/herself and his/her partner. The allocators were also informed that whatever their decision, the reciprocator could only accept it. They were instructed that this stage was executed once.
In Treatment 1, when the allocators made their decisions, we informed their partners (reciprocators) of this decision. Now the reciprocators were endowed with 10 tokens, and they were instructed to assign these tokens between themselves and the allocators. They were informed that whatever their decision was, the allocators can only accept it. They were instructed that this stage was executed once. Similar to Treatment 1, in Treatment 5, when it came to the reciprocators making allocation decisions, they were not informed of the decision made by the allocators in Stage 1.
For Stage 2, in Treatment 2, we asked the reciprocators to estimate the decisions that the allocators had made in Stage 1. Then the reciprocators were instructed to make the allocation decision without knowing the allocator's previous decisions. The only variation between Treatment 2 and 3 lies in the reversed sequence of reciprocators' estimations and allocations. However, the difference between Treatments 2 and 4 was that in the latter, when the reciprocators made allocations, they were already informed of the allocators' decisions.
For Treatments 2, 3, and 4, we added a supplementary treatment for each, naming treatment 2I (Incentivized), 3I, and 4I. The procedures of these three incentivized treatments had no difference from the un-incentivized ones, only that we gave monetary incentives to the reciprocators' estimates. We incentivized accuracy with a quadratic scoring rule:
$$\begin{array}{c}Monetary Incentives \left(in RMB\right)= 4 \ast \left(1-\left[{\left(\frac{E-A}{10}\right)}^{2}\right]\right)\left(1.\right)\end{array}$$
where \(E\) means the estimates of the reciprocators and \(A\) means the allocations of the allocators.
For stage 3, in Treatments 2, 2I, 3, 3I, and 5, we added a reallocation stage where reciprocators could choose to either change their previous decision or not. On the screen, we gave the information on the allocators' decisions, the reciprocators' first decision, and their estimations. The reciprocators should make the reallocation decision based on this information. The Table 1 below shows the detailed information about each Treatment.
Table 1 Detailed Information about each Treatment
Data availability
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Acknowledgments
For helpful comments, we thank seminar audience at TP-MEET at Tsinghua University and conference participants at the International Conference on "Game Theory and Applications". Jianing Chen, Lang Gan, Jiaxun Dai and Jing Ren provided excellent assistance in the process of the experimental implementation.
Author contributions
J.Z. and Z.L. designed the experiment. Z.L. and X.S. implemented the experiment and collected the data. All the authors analysed the data and wrote the manuscript. The authors had equal contributions to this work.
Funding
This research was funded by the Fundamental Research Funds for the Central Universities (2020JJ029, 2022ZX015) and Research Achievements of "Double First-class" Major Project of Beijing Foreign Studies University (2022SYLZD001), National Natural Science Foundation of China (72073080 and 71873074), National Social Science Foundation (20&ZD057).
Competing interests
The authors declare no competing interests.
- Rodriguez-Lara, I. & Moreno-Garrido, L. Self-interest and fairness: self-serving choices of justice principles. Exp. Econ. 15, 158–175 (2011).
- Kassas, B. & Palma, M. A. Self-serving biases in social norm compliance. J. Econ. Behav. Organ. 159, 388–408 (2019).
- Gino, F., Norton, M. I. & Weber, R. A. Motivated Bayesians: Feeling Moral While Acting Egoistically. J. Econ. Perspect. 30, 189–212 (2016).
- Edele, A., Dziobek, I. & Keller, M. Explaining altruistic sharing in the dictator game: The role of affective empathy, cognitive empathy, and justice sensitivity. Learn. Individ. Differ. 24, 96–102 (2013).
- Engel, C. Dictator games: a meta study. Exp. Econ. 14, 583–610 (2011).
- Bénabou, R. & Tirole, J. Incentives and Prosocial Behavior. Am. Econ. Rev. 96, 1652–1678 (2006).
- Dana, J., Weber, R. A. & Kuang, J. X. Exploiting moral wiggle room: experiments demonstrating an illusory preference for fairness. Econ. Theory 33, 67–80 (2006).
- Akerlof, G. A. & Dickens, W. T. The Economic Consequences of Cognitive Dissonance. Am. Econ. Rev. 72, 307–319 (1982).
- Bénabou, R. Groupthink: Collective Delusions in Organizations and Markets. Rev. Econ. Stud. 80, 429–462 (2013).
- Kunda, Z. Motivated inference: Self-serving generation and evaluation of causal theories. J. Pers. Soc. Psychol. 53, 636–647 (1987).
- Rode, J. & Le Menestrel, M. The influence of decision power on distributive fairness. J. Econ. Behav. Organ. 79, 246–255 (2011).
- Shalvi, S., Gino, F., Barkan, R. & Ayal, S. Self-Serving Justifications: Doing Wrong and Feeling Moral. Curr. Dir. Psychol. Sci. 24, 125–130 (2015).
- Bénabou, R. The Economics of Motivated Beliefs. Rev. Econ. Polit. 125, 665–685 (2015).
- Bénabou, R. & Tirole, J. Mindful Economics: The Production, Consumption, and Value of Beliefs. J. Econ. Perspect. 30, 141–164 (2016).
- Feiler, L. Testing models of information avoidance with binary choice dictator games. J. Econ. Psychol. 45, 253–267 (2014).
- Hamman, J. R., Loewenstein, G. & Weber, R. A. Self-Interest through Delegation: An Additional Rationale for the Principal-Agent Relationship. Am. Econ. Rev. 100, 1826–1846 (2010).
- Kappes, A. et al. Uncertainty about the impact of social decisions increases prosocial behaviour. Nat. Hum. Behav. 2, 573–580 (2018).
- Fischbacher, U. & Föllmi-Heusi, F. Lies in Disguise-an Experimental Study on Cheating. J. Eur. Econ. Assoc. 11, 525–547 (2013).
- Heger, S. A., Slonim, R. & Tausch, F. Self-serving dishonesty: The role of confidence in driving dishonesty. J. Risk. Uncertain. 64, 235–250 (2022).
- Lois, G. & Riedl, A. Interplay between different forms of power and meritocratic considerations shapes fairness perceptions. Sci. Rep. 12, 11428 (2022).
- Carlson, R. W., Marechal, M. A., Oud, B., Fehr, E. & Crockett, M. J. Motivated misremembering of selfish decisions. Nat. Commun. 11, 2100 (2020).
- Ben-Ner, A., Putterman, L., Kong, F. & Magan, D. Reciprocity in a two-part dictator game. J. Econ. Behav. Organ. 53, 333–352 (2004).
- Connolly, T., & Zeelenberg, M. Regret in Decision Making. Curr. Dir. Psychol. Sci. 11, 212–216 (2002).
- Sarwar, M. A., Awang, Z. & Habib, M. D. Consumer Purchase Regret: A Systematic Review. J. Int. Soc. Res. 9 (2019).
- C.H. Lin, W.H. Huang & Zeelenberg, M. Multiple reference points in investor regret. J. Econ. Psychol. 27, 781–792 (2006).
- Brewer, N. T., DeFrank, J. T. & Gilkey, M. B. Anticipated regret and health behavior: A meta-analysis. Health. Psychol. 35, 1264–1275 (2016).
- McGraw, K. M. Guilt following transgression: an attribution of responsibility approach. J. Pers. Soc. Psychol. 53, 247–256 (1987).
- Wiwad, D. & Aknin, L. B. Motives matter: The emotional consequences of recalled self- and other- focused prosocial acts. Motiv. Emot. 41, 730–740 (2017).
- Kumano, S., Hamilton, A. & Bahrami, B. The role of anticipated regret in choosing for others. Sci. Rep. 11, 12257 (2021).
- Osgood, J. M. Ego-depletion increases selfish decision making, but may also increase self-conflict and regret about those decisions. J. Soc. Psycho.l 159, 417–430 (2019).
- Zeelenberg, M., van Dijk, W. W. & Manstead, A. S. R. Reconsidering the Relation between Regret and Responsibility. Organ. Behav. Hum. Decis. Process. 74, 254–272 (1998).
- Nicolle, A., Bach, D. R., Frith, C. & Dolan, R. J. Amygdala involvement in self-blame regret. Soc. Neurosci. 6, 178–189 (2011).
- Gabillon, E. When choosing is painful: Anticipated regret and psychological opportunity cost. J. Econ. Behav. Organ. 178, 644–659 (2020).
- Kamiya, A. S. M., Zeelenberg, M. & Hernandez, J. M. d. C. Regulating regret via decreasing goal level: Comparing maximizers and satisficers. Pers. Individ. Differ. 178 (2021).
- Di Tella, R., Perez-Truglia, R., Babino, A. & Sigman, M. Conveniently Upset: Avoiding Altruism by Distorting Beliefs about Others' Altruism. Am. Econ. Rev. 105, 3416–3442 (2015).
- Corbett, B., Feeney, A. & McCormack, T. Interpersonal regret and prosocial risk taking in children. Cogn. Dev. 58 (2021).
- Doñate-Buendía, A., García-Gallego, A. & Petrović, M. Gender and other moderators of giving in the dictator game: A meta-analysis. J. Econ. Behav. Organ. 198, 280–301 (2022).
- Brañas-Garza, P., Capraro, V. & Rascón-Ramírez, E. Gender differences in altruism on Mechanical Turk: Expectations and actual behaviour. Econ. Lett. 170, 19–23 (2018).
- Fischbacher, U. z-Tree: Zurich toolbox for ready-made economic experiments. Exp. Econ. 10, 171–178 (2007).
No competing interests reported.