Right Stuff
Heroization and ironic funneling effects
Matthew Stanley, Steven Shepherd & Aaron Kay
Journal of Personality and Social Psychology, forthcoming
Abstract:
In recent years, much of the American public has venerated military veterans as heroes. Despite overwhelmingly positive public attitudes toward veterans, veterans have experienced higher rates of unemployment and underemployment than their nonveteran peers. The current research leverages theory and research on positive stereotypes to shed light on this seeming inconsistency between the heroization of veterans and their heightened rates of unemployment and underemployment. We conceptualize the hero label as a pervasive positive stereotype, and we employ complementary methods and designs (correlational, quasi-experimental, experimental, and mediational) to investigate the consequences and implications of attaching this label to military veterans. We then extend our theorizing to other heroized groups (e.g., firefighters, paramedics, teachers, and social workers). The results across studies suggest that heroization leads the American public to funnel heroized individuals and groups into a limited set of lower paying jobs, organizations, and careers associated with selflessness. This research not only offers insights into an important real-world problem but also offers a first experimental investigation of the consequences and implications of labeling a group of people as heroes.
Co-Writing with Opinionated Language Models Affects Users' Views
Maurice Jakesch et al.
Cornell University Working Paper, February 2023
Abstract:
If large language models like GPT-3 preferably produce a particular point of view, they may influence people's opinions on an unknown scale. This study investigates whether a language-model-powered writing assistant that generates some opinions more often than others impacts what users write - and what they think. In an online experiment, we asked participants (N=1,506) to write a post discussing whether social media is good for society. Treatment group participants used a language-model-powered writing assistant configured to argue that social media is good or bad for society. Participants then completed a social media attitude survey, and independent judges (N=500) evaluated the opinions expressed in their writing. Using the opinionated language model affected the opinions expressed in participants' writing and shifted their opinions in the subsequent attitude survey. We discuss the wider implications of our results and argue that the opinions built into AI language technologies need to be monitored and engineered more carefully.
Twelve weeks of self-control training does not reduce aggression
Joanne Beames, Eve Slavich & Thomas Denson
Aggressive Behavior, forthcoming
Abstract:
Self-control training (SCT) is a method of practicing self-controlled behavior in one domain that enhances self-controlled behavior in additional domains. We investigated whether 4 and 12 weeks of practicing self-control would improve control over aggressive behavior. Relative to the active control group, SCT did not reduce aggression regardless of the training duration. We also did not find supportive evidence to suggest that theoretically relevant variables mediated or moderated the effects of SCT on aggression over time. Bayesian analyses showed greater support for the null hypotheses than the alternative hypothesis. Our experiment casts doubt on the long-term effectiveness of using SCT for reducing reactive aggression. Additional research is necessary to identify the conditions under which SCT is most likely to facilitate control over aggressive behavior.
Cognitive foundations for helping and harming others: Making welfare tradeoffs in industrialized and small-scale societies
Andrew Delton et al.
Evolution and Human Behavior, forthcoming
Abstract:
For many abilities, such as vision or language, our conscious experience is one of simplicity: We open our eyes and the world appears; we open our mouths and grammatical sentences tumble out. Yet these abilities rely on immensely complex, unconscious computations. Is this also true of abilities related to cooperation or competition, like deciding whether to share food or spread gossip? We tested whether decisions like these are guided by precise psychological variables, called welfare tradeoff ratios. Welfare tradeoff ratios summarize information about multiple sources of social value (such as whether a specific other person is kin or is generous with the self) along with information about the situation (such as what's at stake or who else is watching). We evaluated these hypothesized variables in four societies: among college students in the USA and Argentina and among two groups of Amazonian forager-horticulturalists, the Shuar of Ecuador and the Tsimane of Bolivia (ns = 167, 131, 73, 23). In all societies people made a series of hypothetical decisions where they had to weigh help or harm for themselves versus others. We found strong evidence that people trade off their welfare for others with consistency -- a signature of decisions being guided by precise variables in the mind. We also found evidence in three of the societies that people discriminate among different categories of others in their welfare tradeoffs (e.g., friends versus acquaintances). Although most decisions about helping or harming feel simple and intuitive, they appear to be underwritten by precise computations.
Attributions of Trust and Trustworthiness
Rick Wilson & Catherine Eckel
Political Behavior, forthcoming
Abstract:
This study examines whether individuals can accurately predict trust and trustworthiness in others based on their appearance. Using photos and decisions from previous experimental trust games, subjects were asked to view the photos and guess the levels of trust and trustworthiness of the individuals depicted. The results show that subjects had little ability to accurately guess the trust and trustworthiness behavior of others. There is significant heterogeneity in the accuracy of guesses, and errors in guesses are systematically related to the observable characteristics of the photos. Subjects' guesses appear to be influenced by stereotypes based on the features seen in the photos, such as gender, skin color, or attractiveness. These findings suggest that individuals' beliefs that they can infer trust and trustworthiness from appearance are unfounded, and that efforts to reduce the impact of stereotypes on inferred trustworthiness may improve the efficiency of trust-based interactions.
"It's not about the money. It's about sending a message!" Avengers want offenders to understand the reason for revenge
Andras Molnar, Shereen Chaudhry & George Loewenstein
Organizational Behavior and Human Decision Processes, January 2023
Abstract:
While revenge is typically thought to serve utilitarian goals (deter future offenses) or as an end in itself (restore fairness, equate suffering), we test whether "belief-based" motives also shape revenge behavior. Across four studies-one observational, two hypothetical choice, and one real choice-we find evidence that avengers want the offender to understand why (and sometimes by whom) they are being punished, even when doing so cannot change the offender's future behavior. Avengers prefer punishments that allow them to communicate the reason to offenders, and they are willing to compromise on distributive justice to do so. Furthermore, avengers are less motivated to cause suffering if offenders remain ignorant of the reason. We explore reasons beyond deterrence that explain why avengers may care about what offenders believe, and also discuss the implications of these motives for organizations.
It Is Belief in Dualism, and Not Free Will, That Best Predicts Helping: A Conceptual Replication and Extension of Baumeister et al. (2009)
Oliver Genschow
Personality and Social Psychology Bulletin, forthcoming
Abstract:
Previous research found that experimentally reducing people's belief in free will affects social behaviors. However, more recent investigations could not replicate several findings in this literature. An explanation for the mixed findings is that free will beliefs are related to social behaviors on a correlational level, but experimental manipulations are not able to detect this relation. To test this interpretation, we conceptually replicated and extended a landmark study in the free will belief literature originally conducted by Baumeister et al. In five studies (total N = 1,467), we investigated whether belief in free will predicts helping behavior in comparison to other beliefs related to free will. Overall, our results support the original findings, as belief in free will correlated with helping behavior. However, the results also show that the best predictor of helping behavior is not belief in free will but belief in dualism. Theoretical implications are discussed.