Scientists have started to explore whether novel artificial intelligence (AI) tools based on large language models, such as GPT-4, could support the scientific peer review process. We sought to understand (i) whether AI versus human reviewers are able to distinguish between made-up AI-generated and human-written conference abstracts reporting on actual research, and (ii) how the quality assessments by AI versus human reviewers of the reported research correspond to each other. We conducted a large-scale field experiment during a medium-sized scientific conference, relying on 305 human-written and 20 AI-written abstracts that were reviewed either by AI or 217 human reviewers. The results show that human reviewers and GPTZero were better in discerning (AI vs. human) authorship than GPT-4. Regarding quality assessments, there was rather low agreement between both human–human and human–AI reviewer pairs, but AI reviewers were more aligned with human reviewers in classifying the very best abstracts. This indicates that AI could become a prescreening tool for scientific abstracts. The results are discussed with regard to the future development and use of AI tools during the scientific peer review process.

Past research suggested that assimilation (i.e., the tendency to exaggerate the similarity between one’s ideological position and the position of a preferred political object) is stronger than contrast (i.e., the tendency to exaggerate the dissimilarity between one’s ideological position and the position of a non-preferred political object). However, critiques of this research argue that this conclusion is unwarranted because it is biased toward assimilation appearing stronger than contrast. In the current study, we examine the ideological judgments of American presidential candidates between 1972 and 2020 and analyze all available subjects (in contrast to previous studies that relied primarily on data collected in the 1970s and 1980s and analyzed only subjects who actually voted), and show that, in these years, contrast was stronger than assimilation. We also show that during these years, there was very little change in assimilation but a substantial increase in contrast. We attribute this change to increased polarization among the American electorate.

What role does reasoning about moral principles play in people’s judgments about what is right or wrong? According to one view, reasoning usually plays little role. People tend to do what suits their self-interests and concoct moral reasons afterward to justify their own behavior. Thus, in this view, people are far more forgiving of their own violations than of others’ violations. According to a contrasting view, principled reasoning generally guides judgments and decisions about our own and others’ actions. This view predicts that people usually can, and do, articulate the principles that guide their moral judgments and decisions. The present research examined a phenomenon at the center of these debates: students’ evaluations of academic cheating. Across three studies, we used structured interviews and online surveys to examine first- and third-party judgments and reasoning about cheating events. Third-party scenarios were derived from students’ own accounts of cheating events and manipulated based on the reasons students provided. Findings supported the view that reasoning is central to evaluations of cheating. Participants articulated reasons consistent with their judgments about their own and others’ actions. The findings advance classic debates about reasoning in morality and exemplify a paradigm that can bring further advances.

To make profitable investment decisions, investors must know and understand their risks. They can learn about these risks in different ways. Evidence suggests that investors who learn from a ‘risk tool’ simulator perceive financial risk more accurately, feel more informed and confident, and thus take on more financial risk. We attempt a conceptual replication of these findings, exploring whether they extend from kind to wicked environments and to investors with some investment experience. We conducted four studies online, amounting to 3,804 participants, and observe that neither the risk tool nor any of the other risk communication interventions lead to different risk taking or to different subjective representations of risk relative to a control condition in which participants received no intervention. We evaluated several moderators, but none could explain the absence of an intervention effect, suggesting that the effect of risk tools and other interventions could be limited.

Research on advice taking has demonstrated a phenomenon of egocentric discounting: people weight their own estimates more than advice from others. However, this research is mostly conducted in highly controlled lab settings with low or no stakes. We used unique data from a game show on Norwegian television to investigate advice taking in a high stakes and highly public setting. Parallel to the standard procedure in judge–advisor systems studies, contestants give numerical estimates for several tasks and solicit advice (another estimate) from three different sources during the game. The average weight of advice was 0.58, indicating that contestants weighted advice more than their own estimates. Of potential predictors of weight of advice, we did not detect associations with the use of intuition (e.g., gut feeling, guessing) and advice source (family, celebrities, average of viewers from hometown), but own estimation success (the proportion of previous rounds won) was associated with less weight of advice. Solicitation of advice was associated with higher stakes. Together with the relatively high weight on advice, this suggests that participants considered the advice valuable. On average, estimates did not improve much after advice taking, and the potential for improvement by averaging estimates and advice was negligible. We discuss different factors that could contribute to these findings, including stakes, solicited versus unsolicited advice, task difficulty, and high public scrutiny. The results suggest that highly controlled lab studies may not give an accurate representation of advice taking in high stakes and highly public settings.

This study examined the sour grapes/sweet lemons rationalization through 2 conditions: ‘attainable’ (sweet lemons) and ‘unattainable’ (sour grapes), reflecting China’s 2019-nCoV vaccination strategy. The aim was to find ways to change people’s beliefs and preferences regarding vaccines by easing their safety concerns and encouraging more willingness to get vaccinated. An online survey was conducted from January 22 to 27, 2021, with 3,123 residents across 30 provinces and municipalities in the Chinese mainland. The direction of belief and preference changed in line with the sour grapes/sweet lemons rationalization. Using hypothetical and real contrasts, we compared those for whom the vaccine was relatively unattainable (‘sour grapes’ condition) with those who could get the vaccine easily (‘sweet lemons’). Whether the vaccine was attainable was determined in the early stage of the vaccine roll-out by membership in a select group of workers that was supposed to be vaccinated to the greatest extent possible, or, by being in the second stage when the vaccine was available to all. The attainable conditions demonstrated higher evaluation in vaccine safety, higher willingness to be vaccinated, and lower willingness to wait and see. Hence, we propose that the manipulation of vaccine attainability, which formed the basis of the application of sour grapes/sweet lemons rationalization, can be utilized as a means to manipulate the choice architecture to nudge individuals to ease vaccine safety concerns, reducing wait-and-see tendencies, and enhancing vaccination willingness. This approach can expedite universal vaccination and its associated benefits in future scenarios resembling the 2019-nCoV vaccine rollout.

Most research on intertemporal choice has examined choices between smaller, sooner gains and larger, later gains. A much smaller number of papers have examined intertemporal choices for losses. In this article, we explore whether mixed-sign choices with both gains and losses may better correlate with real-world behaviors. In two high-powered studies (pilot: N = 3,200; main study: N = 7,000), participants completed one of four normatively equivalent measures consisting of pure gain, pure loss, or mixed sign (Gain-Now-Loss-Later or Loss-Now-Gain-Later) intertemporal choices. Participants also self-reported a large number of demographic measures and real-world choice behaviors thought to be linked to intertemporal choice. The results indicate that (1) mixed-sign intertemporal choices yield more patient time preferences than pure-gain choices but less patient than pure-loss choices and (2) pure-gain intertemporal choices yield equivalent or superior predictive power across a range of real-world intertemporal choice behaviors.

The identified victim effect is the phenomenon in which people tend to contribute more to identified than to unidentified victims. Kogut and Ritov (Journal of Behavioral Decision Making, 18(3), 157–167, 2005) found that the identified victim effect was limited to a single victim and driven by empathic emotions. In a pre-registered experiment with an online U.S. American MTurk sample on CloudResearch (N = 2003), we conducted a close replication and extension of Experiment 2 from Kogut and Ritov (Journal of Behavioral Decision Making, 18(3), 157–167, 2005). The replication findings failed to provide empirical support for the identified single victim effect hypothesis since we found no evidence of differences in willingness to contribute when comparing a single identified victim to a single unidentified victim (η2p = .00, 90% CI [0.00, 0.00]), and no indication for the target article’s interaction between singularity and identifiability (original: η2p = .062, 90% CI [0.01, 0.15]; replication: η2p = .00, 90% CI [0.00, 0.00]). Extending the replication to conduct a conceptual replication of Kogut and Ritov (Organizational Behavior and Human Decision Processes, 104(2), 150–157, 2007), we investigated a boundary condition of the effect—group belonging. We found support for an ingroup bias in helping behaviors and indications for empathic emotions and perceived responsibility contributing to this effect. We discuss differences between our study and the target article and implications for the literature on the identified victim effect.

The literature on emotion and risk-taking is large and heterogeneous. Whereas some studies have found that positive emotions increase risk-taking and negative emotions increase risk aversion, others have found just the opposite. In this study, we investigated this question in the context of a risky decision-making task with embedded high-resolution sampling of participants’ subjective emotional valence. Across two large-scale experiments (N = 329 and 524), we consistently found evidence for a negative association between self-reported emotional valence and risk-taking behaviors. That is, more negative subjective affect was associated with increased risk-seeking, and more positive subjective affect was associated with increased risk aversion. This effect was evident both when we compared participants with different levels of mean emotional valence as well as when we considered within-participant emotional fluctuations over the course of the task. Prospect-theoretic computational modeling analyses suggested that both between- and within-participant effects were driven by an effect of emotional valence on the curvature of the subjective utility function (i.e., increased risk tolerance in more negative emotional states), as well as by an effect of within-person emotion fluctuations on loss aversion. We interpret findings in terms of a tendency for participants in negative emotional states to choose high-risk, high-reward options in an attempt to improve their emotional state.

Overconfidence plays a role in a large number of individual decision biases and has been considered a ‘meta-bias’ for this reason. However, since overconfidence is measured behaviorally with respect to particular tasks (in which performance varies across individuals), it is unclear whether people generally vary in terms of their general overconfidence. We investigated this issue using a novel measure: the Generalized Overconfidence Task (GOT). The GOT is a difficult perception test that asks participants to identify objects in fuzzy (‘adversarial’) images. Critically, participants’ estimated performance on the task is not related to their actual performance. Instead, variation in estimated performance, we argue, arises from generalized overconfidence, that is, people indicating a cognitive skill for which they have no basis. In a series of studies (total N = 1,293), the GOT was more predictive when looking at a broad range of behavioral outcomes than two other overestimation tasks (cognitive and numeracy) and did not display substantial overlap with conceptually related measures (Studies 1a and 1b). In Studies 2a and 2b, the GOT showed superior reliability in a test–retest design compared to the other overconfidence measures (i.e., cognitive and numeracy measures), particularly when collecting confidence ratings after each image and an estimated performance score. Finally, the GOT is a strong predictor of a host of behavioral outcomes, including conspiracy beliefs, bullshit receptivity, overclaiming, and the ability to discern news headlines.

In light of the growing threat of climate change and urgency of mitigation at the societal and individual level, an exponentially growing body of research has addressed how and what people think about climate change—ranging from basic judgments of truth and attitudes about risk to predictions of future outcomes. However, the field is also beset by a striking variety of items and scales used to measure climate change beliefs, with notable differences in content, untested structural assumptions, and unsatisfactory or unknown psychometric properties. In a series of four studies (total N = 2,678), scales for the assessment of climate change beliefs are developed that are comprehensive and balanced in content and psychometrically sound. The latent construct structure is tested, and evidence of high rank-order stability (1-year retest-reliability) and predictive validity (for policy preferences and actual behavior) provided.

In a questionnaire, participants judged the increase in SARS-CoV-2 virus exposure when moving closer to an infected person. Earlier studies have shown that the actual increase in virus exposure is underestimated and the present study replicated and extended these studies. The primary purpose was to investigate to what extent questionnaire judgments about hypothetical situations can predict judgments and actual behavior in real physical space. Participants responded to a questionnaire and the same participants also took part in a parallel study that was conducted in a room with a mannequin representing a virus infected person. The earlier reported bias in the perception of exposure as a function of distance to a virus source was replicated in the questionnaire and the physical laboratory study. A linear function connected median exposure judgments at the same distances from a virus source in the questionnaire and in the laboratory, R2 = 0.99. When asked to move to a distance that would give a prescribed exposure level, a linear function described the relationship between questionnaire distance judgments and moves to distances in the physical space, R2 = 0.95. We concluded that questionnaire data about perceived virus exposures are reliable indicators of real behavior. For health reasons, the significant underestimations of the steep increase of virus exposure during an approach to a virus source need to be stressed in communications to policy makers, the public, professionals working close to clients, nursing staff, and other care providers.

Many real-world intertemporal decisions involve a group of two or more individuals making consensual decisions through group collaboration. Here, we ask how group collaboration affects intertemporal choices. In two experiments, participants completed intertemporal choices individually first (the precollaboration phase). Then, participants were placed into groups of two or three and completed a similar intertemporal task, with the group arriving at a consensual decision on each trial (the group collaboration phase). Finally, participants once again completed the intertemporal choices individually (the postcollaboration phase). Results showed that after group collaboration, the delay discounting significantly decreased compared to before collaboration both at group level and at individual level. The effect of group collaboration on individual intertemporal choices was no longer discernible by 1 week later. Therefore, the current research demonstrates the effectiveness of group collaboration and provides a way to nudge both groups and individuals to make farsighted choices.

Research in behavioral decision-making has produced many models of decision under risk. To improve our understanding of choice under risk, it is essential to perform rigorous model comparisons over large sets of decision settings to find which models are most useful. Recently, such large-scale comparisons have produced conflicting conclusions: A variant of cumulative prospect theory (CPT) was the best model in a study by He, Analytis, and Bhatia (2022), whereas variants of the model BEAST were the best in two choice prediction competitions. This study delves into these contradictions to identify and explore the underlying reasons. We replicate and extend the analysis by He et al., this time incorporating BEAST, which was previously excluded because it cannot be analytically estimated. Our results show that while CPT excels in systematically hand-crafted tasks, BEAST—originally designed for broader decision-making contexts—matches or even surpasses CPT’s performance when choice tasks are randomly selected, and predictions are made for new, unknown decision makers. This success of BEAST, very different from classical decision models—as it does not assume, for example, subjective transformations of outcomes and probabilities—puts into question previous conclusions concerning the underlying psychological mechanisms of choice under risk. Our results challenge the field to expand beyond established evaluating techniques and highlight the importance of an inclusive approach toward nonanalytic models, like BEAST, to achieve more objective insights into decision-making behavior.

Most global inequality is between countries, but inequality perceptions have mostly been investigated within the country. Six studies (total N = 2656, 5 preregistered, 1 incentivized for accuracy, 1 with a sample representative of the USA) show that Westerners (U.S. American, British, and French participants) believe that developing and middle-income countries’ GDP per capita is much closer to developed countries’ than it actually is, and that people in developing and middle-income countries have higher rates of car ownership, larger houses, and eat out more frequently than they actually do, meaning that Westerners underestimate global inequality. This misperception is underpinned by a convergence illusion: the belief that over time, poorer countries have closed the economic gap with richer countries to a larger extent than they have. Further, overestimating GDP per capita is negatively correlated with support for aid to the target country and positively correlated with a country’s perceived military threat. We discuss implications for inequality perceptions and for global economic justice.