When members of a group agree publicly on controversial moral and political questions, this is widely taken as evidence of genuine consensus. However, publicly expressed judgments may not faithfully reflect private views, but arise from three sources: shared values, intrinsic conformity with prevailing views, and strategic conformity driven by public evaluation. To separate these mechanisms, we introduce a new moral-trilemma paradigm and adapt a transitivity-based experimental approach that allows us to infer participants' private judgments before social exposure and compare them with judgments made after observing majority views, with or without peer reward or punishment. Across twelve moral domains, participants shift toward majority views even when responses remain private, demonstrating intrinsic conformity. When judgments are evaluated and can elicit peer reward or punishment, conformity nearly doubles. Consequently, a substantial share of visible agreement reflects strategic preference falsification rather than genuine convergence. Observed consensus thus often overstates underlying agreement within groups, which may also amplify perceived polarization.

Large language models (LLMs) are poised to reshape the way individuals communicate and interact. While this form of AI has the potential to efficiently make many human decisions, there is limited understanding of how individuals will respond to its use in social interactions. In particular, it remains unclear how individuals interact with LLMs when the interaction has consequences for other people. Here, we report the results of a large-scale, preregistered online experiment ( n = 3 , 552 ) showing that human players' fairness, trust, trustworthiness, cooperation, and coordination in economic two-player games decrease when the decision of the interaction partner is taken over by ChatGPT. On the contrary, we observe no adverse reactions when individuals are uncertain whether they are interacting with a human or a LLM. At the same time, participants often delegate decisions to the LLM, especially when the model's involvement is not disclosed, and individuals have difficulty distinguishing between decisions made by humans and those made by AI.

How do people attribute responsibility when an outcome is not caused by an individual but results from a decision chain involving several people? We study this question in an experiment, in which five voters sequentially decide on how to distribute money between them and five recipients. The recipients can reward or punish each voter, which we use as measures of responsibility attribution. In the aggregate, we find that responsibility is attributed mostly according to the voters’ choices and the pivotality of the decision, but not for being the initial voter. On the individual level, we find substantial heterogeneity with three overall patterns: Little to no responsibility attribution, pivotality-driven, and focus on choices. These patterns are similar when praising voters for good outcomes and blaming voters for bad outcomes.

Repeated interactions provide an evolutionary explanation for one-shot human cooperation that is counterintuitive but orthodox 1 – 3 . Intergroup competition 4 – 7 provides an explanation that is intuitive but heterodox. Here, using models and a behavioural experiment, we show that neither mechanism reliably supports cooperation. Ambiguous reciprocity, a class of strategies that is generally ignored in models of reciprocal altruism, undermines cooperation under repeated interactions. This finding challenges repeated interactions as an evolutionary explanation for cooperation in general, which further challenges the claim that repeated interactions in the past can explain one-shot cooperation in the present. Intergroup competitions also do not reliably support cooperation because groups quickly become extremely similar, which limits scope for group selection. Moreover, even if groups vary, group competitions may generate little group selection for multiple reasons. Cooperative groups, for example, may tend to compete against each other 8 . Whereas repeated interactions and group competitions do not support cooperation by themselves, combining them triggers powerful synergies because group competitions constrain the corrosive effect of ambiguous reciprocity. Evolved strategies often consist of cooperative reciprocity with ingroup partners and uncooperative reciprocity with outgroup partners. Results from a behavioural experiment in Papua New Guinea fit exactly this pattern. They thus suggest neither an evolutionary history of repeated interactions without group competition nor a history of group competition without repeated interactions. Instead, our results suggest social motives that evolved under the joint influence of both mechanisms.

Generative artificial intelligence (AI) is poised to reshape the way individuals communicate and interact. While this form of AI has the potential to efficiently make numerous human decisions, there is limited understanding of how individuals respond to its use in social interaction. In particular, it remains unclear how individuals engage with algorithms when the interaction entails consequences for other people. Here, we report the results of a large-scale pre-registered online experiment (N = 3,552) indicating diminished fairness, trust, trustworthiness, cooperation, and coordination by human players in economic twoplayer games, when the decision of the interaction partner is taken over by ChatGPT. On the contrary, we observe no adverse welfare effects when individuals are uncertain about whether they are interacting with a human or generative AI. Therefore, the promotion of AI transparency, often suggested as a solution to mitigate the negative impacts of generative AI on society, shows a detrimental effect on welfare in our study. Concurrently, participants frequently delegate decisions to ChatGPT, particularly when the AI's involvement is undisclosed, and individuals struggle to discern between AI and human decisions.

We provide systematic insights on strategic conformist-as well as anti-conformist-behaviour in situations where people are evaluated, i.e., where an individual has to be selected for reward (e.g., promotion) or punishment (e.g., layoffs). To affect the probability of being selected, people may attempt to fit in or stand out in order to affect the chances of being noticed or liked by the evaluator. We investigate such strategic incentives for conformity or anti-conformity experimentally in three different domains: facts, taste and creativity. To distinguish conformity and anti-conformity from independence, we introduce a new experimental design that allows us to predict participants' independent choices based on transitivity. We find that the prospect of punishment increases conformity, while the prospect of reward reduces it. Anti-conformity emerges in the prospect of reward, but only under specific circumstances. Similarity-based selection (i.e., homophily) is much more important for the evaluators' decisions than salience. We also employ a theoretical approach to illustrate strategic key mechanisms of our experimental setting.

Gender segregation of occupations and entire industries is widespread. The segregation could be the result of perceived job-specific productivity differences between men and women. It could also result from the belief that homogeneous teams perform better or from in-group bias of male or female recruiters. We investigate these explanations in two samples: students and personnel managers. The subjects bet on the productivity of teams that are homogeneous with respect to gender at the outset and then either remain homogeneous or become diverse. The teams work on tasks that differ with respect to gender stereotypes. We obtain similar results in both samples. Women are picked more often for the stereotypically female task, and men are picked more often for the stereotypically male task. Subjects believe that gender-diverse teams perform better, especially in the task with complementarities, and they display an own-gender bias. Elicited expectations about the bets of others reveal that subjects expect the gender stereotypes of tasks but underestimate others’ bets on diversity. This paper was accepted by Yan Chen, behavioral economics and decision analysis. Funding: The authors acknowledge financial support from the Deutsche Forschungsgemeinschaft [Excellence Cluster EXC 2035/1 “The Politics of Inequality” (to U. Fischbacher), CRC TRR 190 “Rationality and Competition” (to D. Kübler), and the Excellence Cluster EXC 2055 “Contestations of the Liberal Script” (to D. Kübler)] and from Tamkeen [NYU Abu Dhabi Research Institute Award CG005 (to R. Stüber)]. Supplemental Material: The data files and online appendix are available at https://doi.org/10.1287/mnsc.2023.4943 .

Despite strong evidence for peer effects, little is known about how individuals balance intrinsic preferences and social learning in different choice environments. Using a combination of experiments and discrete choice modeling, we show that intrinsic preferences and social learning jointly influence participants' decisions, but their relative importance varies across choice tasks and environments. Intrinsic preferences guide participants' decisions in a subjective choice task, while social learning determines participants' decisions in a task with an objectively correct solution. A choice environment in which people expect to be rewarded for their choices reinforces the influence of intrinsic preferences, whereas an environment in which people expect to be punished for their choices reinforces conformist social learning. We use simulations to discuss the implications of these findings for the polarization of behavior.

Politics is increasingly driven by identity cleavages, which also affect the discussion about inequality and redistribution. Typically, redistribution is meant to reduce inequality, implying that redistribution nei- ther makes the rich richer nor the former poor the new rich. However, if identity affects redistribution, these limits might no longer be binding, and redistribution could further increase existing inequalities (making the rich richer) or reverse the income ordering to favor the once-poor (which can even be inequality increasing if redistribution is strong). In a laboratory experiment, we investigate redistribution via a novel smooth one-dimensional distribution mechanism that also allows for an increase or reversal of inequality. Decision-makers receive information about the recipients' political orientation, nationality, or seat number during the experiment, and we vary the structure and source of income inequality (in- come is either earned, random, or unfair). We find most choices of the decision-makers involve redistri- bution, with only 8 % of choices sticking with the status quo. While most redistribution choices reduce inequality, a larger share-(18 %)-increase inequality by making the rich richer, 13 % of choices reduce overall inequality but make the poor the new rich, and 9 % increase inequality by making the poor very rich. Thus, 40 % of decisions are redistributions that are typically unobserved in common redistribution designs. Ingroup favoritism is a strong motive for redistribution in general, and it is the most important motive for redistribution to increase or reverse inequality. Indeed, 85 % of the inequality-increasing or reversing decisions favor the ingroup. Complementary eye-tracking data show that decision-makers' attention to information about the recipients' groups and to poor outliers are related to higher levels of redistribution.(c) 2023 The Author(s). Published by Elsevier B.V.This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

While laboratory and field experiments are the major items in the toolbox of behavioral economists, household panel studies can complement them and expand their research potential. We introduce the German Socio-Economic Panel’s Innovation Sample (SOEP-IS), which offers researchers detailed panel data and the possibility to collect personalized experimental and survey data for free. We discuss what SOEP-IS can offer to behavioral economists and illustrate a set of design ideas with examples. Although we build our discussion on SOEP-IS, our purpose is to provide a guide that can be generalized to other household panel studies as well.

In many countries, the current vaccination rates are stagnating, to the extent that vaccine hesitancy-the delay or refusal to take recommended vaccinations-forms a major obstacle to ending the COVID-19 pandemic. This tendency is particularly concerning when observed among healthcare workers who are opinion leaders on medical matters for their patients and peers. Our study surveys 965 employees of two large Swiss hospitals and profiles vaccine-hesitant hospital employees using not only socio-demographic characteristics, but also a comprehensive set of standard behavioral preference measures: (i) Big-5 personality traits, (ii) risk-, time- and social preferences, and (iii) perceived prevailing social norms. Using multinomial probit models and linear probability models, we find that vaccine-hesitant hospital employees are less patient and less likely to perceive vaccination as the prevailing social norm-in addition to replicating previously published socio-demographic results. Our findings are robust to a range of model specifications, as well as individual and situational covariates. Our study thus offers actionable policy implications for tailoring public-health communications to vaccine-hesitant hospital employees.

We study caste-based discrimination and belief updating of employers, using a lab experiment in India. We vary the access to education of potential employees and test if the information about the education opportunity changes the extent of discrimination and belief updating. Overall, we do not find significant evidence for discrimination, irrespective of the information provided, although the non-upper caste employers discriminate in favour of their own group in the initial rounds. Over time, both the upper caste and non-upper caste employers update their beliefs about the true ability distribution and statistically discriminate in favour of upper caste employees.

People often have to judge the social motives of others, for example, to distinguish truly prosocial people from those merely trying to appear prosocial. Gaze can reveal the motives underlying social decisions, as decision-makers dedicate more attention to motive-relevant information. We extend the use of eye-tracking and apply it as a communication device by providing (real-time) eye-tracking information of one participant to another. We find that untrained observers can judge the prosociality of decision-makers from their eye-tracked gaze alone, but only if there are no strategic incentives to be chosen for a future interaction. When there are such strategic incentives, the cues of prosociality are invalidated, as both individualistic and prosocial decision-makers put effort into appearing more prosocial. Overall, we find that gaze carries information about a person's prosociality, but also that gaze is malleable and affected by strategic considerations.

Oral contraceptives (OC) and endogenous female sex hormones in naturally cycling women (NC) are related to a wide range of psychological variables (eg, cognition and affect). Little research on social behavior has been done. One study documented a tendency towards more prosocial behavior in NC than OC women, but the underlying neuroendocrine mechanisms remain unknown. The sex hormones progesterone and estradiol are potential candidates. We analyzed social decision-making and social behavior in 83 healthy women (38 OC and 45 NC) via the Social Value Orientation (SVO) and in real social interactions within a paradigm adapted from behavioral economics. We also measured empathy, and collected saliva samples to quantify the basal levels of estradiol and progesterone. Our analyses revealed higher levels of prosocial behavior and emotional empathy in NC than in OC women, a finding supported by higher levels of prosocial decisions in NC than OC women in the SVO. Regarding the underlying biological mechanisms, we detected lower progesterone levels in OC than NC women. Exploratory analyses revealed a negative correlation between progesterone and trust on the trend level. We found no correlations between estradiol and behavior. Our findings provide evidence that OC modulate social behavior and initial indications of a possible modulation by progesterone. Further research is needed to replicate our findings and extend them to other social behaviors.

Selfishness and fairness are important drivers of redistribution, but recently an additional motive got into focus. In heterogeneous societies, ingroup preferences can be an important determinant of redistribution decisions. In a laboratory experiment, we investigate the relative importance of the different motives. We create heterogeneity by providing subjects with information about a social group of recipients (nationality, minimal or political orientation), we manipulate how initial inequality is generated (earned, random or unfair) and the shape of the distribution. Further, we extend the redistribution mechanism to go beyond the limits of status quo and equality. We find ingroup favoritism to be the strongest motive; decision-makers almost exclusively use extreme forms of redistribution to favor members of their own social group. We complement the behavioral data with eye-tracking data, showing that attention to the social group information and to poor outliers are indicative of redistribution.

Human behavior can have effects on oneself and externalities on others. Mask wearing is such a behavior in the current pandemic. What motivates people to wear face masks in public when mask wearing is voluntary or not enforced? Which benefits should the policy makers rather emphasize in information campaigns-the reduced chances of getting the SARS-CoV-2 virus (benefits for oneself) or the reduced chances of transmitting the virus (benefits for others in the society)? In this paper, we link measured risk preferences and other-regarding preferences to mask wearing habits among 840 surveyed employees of two large Swiss hospitals. We find that the leading mask-wearing motivations change with age: While for older people, mask wearing habits are best explained by their self-regarding risk preferences, younger people are also motivated by other-regarding concerns. Our results are robust to different specifications including linear probability models, probit models and Lasso covariate selection models. Our findings thus allow drawing policy implications for effectively communicating public-health recommendations to frontline workers during the COVID-19 pandemic.

Charness and Dufwenberg's paper is an exemplary one in experimental economics because of its outstanding contributions in motivation, design, and interpretation. First, it addresses the very general question of how communication increases cooperation in a particularly interesting situation – in a trust game with a promise option. Second, the design is clean and has interesting features. It is based on a binary version of the trust game in which promises are easy to interpret, and the collection of higher-order beliefs provides a tight link to the theoretical model. Third, the use of a theoretical model as a basis for the experiment provides an interpretation that can be tested in much detail. Thus, the paper provides useful tools for experimental economics and it is influential because it initiated the investigation of promises and the role of guilt aversion.