We call attention to an important, but overlooked finding in research reported by Longoni, Bonezzi and Morewedge (2019). Longoni et al. claim that people always prefer a human to an artificially intelligent (AI) medical provider. We show that this was only the case when the historical performance of the human and AI providers was equal.  When the AI is known to outperform the human, their data showed a clear preference for the automated provider. We provide additional statistical analyses of their data to support this claim.

Information search is key to making decision from experience: exploration permits people to learn about the statistical properties of choice options and thus to become aware of rare but potentially momentous decision consequences. This registered report investigates whether and how people differ when making decisions from experience in isolation versus under competitive pressure, which may have important implications for choice performance in different types of choice environments: in “kind” environments without any rare and extreme events, frugal search is sufficient to identify advantageous options. Conversely, in “wicked” environments with skewed outcome distributions, rare but important events will tend to be missed in frugal search. One theoretical view is that competitive pressure encourages efficiency and may thereby boost adaptive search in different environments. An alternative and more pessimistic view is that competitive pressure triggers agency-related concerns, leading to minimal search irrespective of the choice environment, and hence to inferior choice performance. Using a sampling game, the present study (N = 277) found that solitary search was not adaptive to different choice environments (M = 14 samples), leading to a high choice performance in a kind and in a moderately wicked environment, but somewhat lower performance in an extremely wicked environment. Competitive pressure substantially reduced search irrespective of the choice environment (M = 4 samples), thus negatively affecting overall choice performance. Yet, from the perspective of a cost-benefit framework, frugal search may be efficient under competitive pressure. In sum, this report extends research on decisions from experience by adopting an ecological perspective (i.e., systematically varying different choice environments) and by introducing a cost-benefit framework to evaluate solitary and competitive search — with the latter constituting a challenging problem for people in an increasingly connected world.

In the light of the potential negative consequences of dishonest behaviors for individuals and societies, researchers from different disciplines have aimed to investigate situation and person factors shaping the occurrence and extent of such behaviors. The present study investigates the roles of a situation factor, the baseline probability of observing a favorable outcome, and a person factor, trait Honesty-Humility from the HEXACO Model of Personality, in shaping dishonest behavior. Next to main effects, a person-situation interaction between these factors was tested. Across three studies with 5,297 participants overall, we find that a higher baseline probability of observing a favorable outcome and lower levels in Honesty-Humility are linked to more dishonest behavior, whereas there was no strong evidence for an interaction between these factors. By testing the assumed effects in two different cheating paradigms, this study additionally allows to disentangle previously found effects of (a) the distance between an observed and the favorable outcome and (b) the baseline probability of observing a favorable outcome.

There is a lively debate about the effect of maximizing and satisficing tendencies on well-being. The question is, whether maximizing and satisficing have an adaptive or maladaptive effect on well-being. There are also issues regarding the conceptualization and measurement of maximizing and satisficing tendencies. In a sample of 514 subjects from the general population in Slovakia, a two-component model of maximizing was examined. Satisficing tendency was measured as a separate construct. The results show the usefulness of a two-component model (maximizing as a strategy and maximizing as a goal) in measuring maximizing tendency. Maximizing as a strategy (measured as alternative search) turned out to be maladaptive (positively related to depression and negatively related to happiness), whereas maximizing as a goal (measured as high standards) had no maladaptive effect (no relation with well-being). In addition, the two components were differently associated with personality factors, which strengthens the need to distinguish between them. However, the satisficing tendency measured separately from maximizing tendency was not related to anything which raises a question about the conceptualization and validity of this tendency. The results of the current study, therefore, indicate that the (mal)adaptive effect of these tendencies depends on their conceptualization as well as on how these tendencies are measured, and also on their different relationship with personality factors. However, results also point to the importance of considering the cultural context that may have an effect on the relationship between maximizing and well-being. Therefore, the results may vary due to different cultures.

Previous research has shown that short-term changes in blood glucose influence our preferences and may affect decisions about risk as well. However, consensus is lacking about whether and how blood glucose influences decision making under risk, and we conduct two experiments and a meta-analysis to examine this question in detail. In Study 1, using a pecuniary valuation method, we find no effect of blood glucose on willingness to pay for risky products that may act as allergens. In Study 2, using risky gambles, we find that low levels of blood glucose increase risk taking for food and to a lesser degree for non-food rewards. Combining our own and previous findings in a meta-analysis, we show that low levels of blood glucose on average increase risk taking about food. Low blood glucose does not increase risk taking about non-food rewards although this is subject to heterogeneity. Overall, our studies suggest that low blood glucose increases our willingness to gamble on how much food we can get, but not our willingness to eat food that can harm us. Our findings are best explained by the energy budget rule.

The rapid spread of COVID-19 has emphasized the need for effective health communications to coordinate individual behavior and mitigate disease transmission. Facing a pandemic, individuals may be driven to adopt public health recommendations based on both self-interested desires to protect oneself and prosocial desires to protect others. Although messages can be framed around either, existing research from the social sciences has offered mixed evidence regarding their relative efficacy. Informing this dialogue, in the current study we report on the findings of a field experiment (N = 25,580) conducted March 21–22 on Facebook during the critical initial weeks of the COVID-19 outbreak in the United States. We observed that ad messages using a distant prosocial frame (“protect your community”) were in fact significantly less effective than those using a self-focused frame (“protect yourself”) in eliciting clickthroughs to official CDC recommendations. However, ad messages with a close prosocial frame (“protect your loved ones”) were equally effective as the self-focused frame. These findings catalog the differential efficacy of ad messaging strategies during the early stages of the coronavirus pandemic.

In Longoni et al. (2019), we examine how algorithm aversion influences utilization of healthcare delivered by human and artificial intelligence providers. Pezzo and Beckstead’s (2020) commentary asks whether resistance to medical AI takes the form of a noncompensatory decision strategy, in which a single attribute determines provider choice, or whether resistance to medical AI is one of several attributes considered in a compensatory decision strategy. We clarify that our paper both claims and finds that, all else equal, resistance to medical AI is one of several attributes (e.g., cost and performance) influencing healthcare utilization decisions. In other words, resistance to medical AI is a consequential input to compensatory decisions regarding healthcare utilization and provider choice decisions, not a noncompensatory decision strategy. People do not always reject healthcare provided by AI, and our article makes no claim that they do.

During pregnancy, a variety of psychological and physical changes occur in women, which may have different impacts on risk decision-making involving different processes systems. Based on the dual-process theories of decision-making, using the Columbia Card Task (CCT) as the experimental paradigm, which can trigger deliberative versus affective decision-makings respectively, this study recruited 240 pregnant women and non-pregnant women aged 20-40 as the experimental group and control group respectively, investigated how pregnancy impacted on women’s risk decision-making, as well as the possible roles played by a series of psychological factors (impulsivity; sensation seeking; emotional state) and physiological factors (gestational age; human Chorionic Gonadotropin, hCG; progesterone) in the above process. The results were as follows: (a) Compared with non-pregnant women, pregnant women tended to choose fewer cards, indicating a higher risk aversion consistent with a more conservative strategy, both in cold and hot CCTs; in both cold and hot CCTs, compared with pregnant women in the second trimester of pregnancy, pregnant women in the first and the third trimesters of pregnancy had a higher risk aversion tendency. (b) Pregnant women had lower levels of all dimensions of sensation seeking than did non-pregnant women, pregnant women in the third trimester of pregnancy had lower levels of Disinhibition (DIS) and Boredom Susceptibility (BS) of sensation seeking than pregnant women in the first and the second trimesters of pregnancy, but there was no significant difference in levels of emotional state or impulsivity between pregnant woman and non-pregnant women. (c) DIS of sensation seeking played a fully mediating role in the impact of pregnancy on hot CCT performance. (d) Both hCG and progesterone levels were negatively correlated with pregnant women’s hot CCT performances. (e) Positive emotion played a partial mediating role in the effect of progesterone on hot CCT performance of pregnant women.

We clarify two points made in our commentary (Pezzo & Beckstead, 2020, this issue) on a recent paper by Longoni, Bonezzi, and Morewedge (2019). In both Experiments 1 and 4 from their paper, it is not possible to determine whether accuracy can compensate for algorithm aversion. Experiments 3A-C, however, do show a strong effect of accuracy such that AI that is superior to a human provider is embraced by patients. Many papers, including Longoni et al. tend to minimize the role of this compensatory process, apparently because it seems obvious to the authors (Longoni, Bonezzi, Morewedge, 2020, this issue). Such minimization, however, can lead to (mis)citations in which research that clearly demonstrates a compensatory role of AI accuracy is cited as non-compensatory.

There are many ways to measure how people manage risk when they make decisions. A standard approach is to measure risk propensity using self-report questionnaires. An alternative approach is to use decision-making tasks that involve risk and uncertainty, and apply cognitive models of task behavior to infer parameters that measure people’s risk propensity. We report the results of a within-participants experiment that used three questionnaires and four decision-making tasks. The questionnaires are the Risk Propensity Scale, the Risk Taking Index, and the Domain Specific Risk Taking Scale. The decision-making tasks are the Balloon Analogue Risk Task, the preferential choice gambling task, the optimal stopping problem, and the bandit problem. We analyze the relationships between the risk measures and cognitive parameters using Bayesian inferences about the patterns of correlation, and using a novel cognitive latent variable modeling approach. The results show that people’s risk propensity is generally consistent within different conditions for each of the decision-making tasks. There is, however, little evidence that the way people manage risk generalizes across the tasks, or that it corresponds to the questionnaire measures.

Butler and Pogrebna (2018) devised triples of three-branch gambles theorized to violate transitivity of preference according to a most probable winner model. According to this model, a person chooses the option that has the higher probability to yield a better outcome than the other alternative. They tested 11 triples with 100 participants and found cases that appeared to violate weak stochastic transitivity and the triangle inequality. But tests of weak stochastic transitivity and the triangle inequality do not provide a proper method to compare transitive and intransitive models that allow mixtures of preference patterns and random errors. Those older methods can yield false conclusions regarding transitivity, for example, if different participants have different true preferences or if different choice problems have different rates of error. This paper reanalyzes their data using a true and error (TE) model, which does not require these unrealistic assumptions, and which provides estimates of the incidence of transitive and intransitive behavior in a mixture. Reanalysis indicated that 3 of the 11 triples showed convincing evidence of violations of transitivity in the opposite direction of the predictions of the most probable winner model. Further, these and other triples showed other significant violations of the most probable winner model. Despite some violations of the true and error model, the data of Butler and Pogrebna appear to contradict not only transitive utility models but they also refute the most probable winner model as a descriptive theory of choice behavior.

True and Error Theory (TET) is a modern latent variable modeling approach for analyzing sets of preferences held by people. Individual True and Error Theory (iTET) allows researchers to estimate the proportion of the time an individual truly holds a particular underlying set of preferences without assuming complete response independence in a repeated measures experimental design. iTET is thus suitable for investigating research questions such as whether an individual ever is truly intransitive in their preferences (i.e., they prefer a to b, b to c, and c to a). While current iTET analysis methods provide the means of investigating such questions they require a lot of data to achieve satisfactory power for hypothesis tests of interest. This paper overviews the performance and shortcomings of the current analysis methods in efficiently using data, while providing new analysis methods that offer substantial gains in power and efficiency.

Birnbaum (2020) reanalyses the data from Butler and Pogrebna (2018) using his ‘true and error’ test of choice patterns. His results generally support the evidence we presented in that paper. Here we reiterate the reasons for our agnosticism as to the direction any cycles might take, even though the paradox that motivated our study takes a ‘probable winner’ direction. We conclude by returning to the potential significance of predictably intransitive preferences for decision theory generally.

The Markov True and Error (MARTER) model (Birnbaum & Wan, 2020) has three components: a risky decision making model with one or more parameters, a Markov model that describes stochastic variation of parameters over time, and a true and error (TE) model that describes probabilistic relations between true preferences and overt responses. In this study, we simulated data according to 57 generating models that either did or did not satisfy the assumptions of the True and Error fitting model, that either did or did not satisfy the error independence assumptions, that either did or did not satisfy transitivity, and that had various patterns of error rates. A key assumption in the TE fitting model is that a person’s true preferences do not change in the short time within a session; that is, preference reversals between two responses by the same person to two presentations of the same choice problem in the same brief session are due to random error. In a set of 48 simulations, data generating models either satisfied this assumption or they implemented a systematic violation, in which true preferences could change within sessions. We used the true and error (TE) fitting model to analyze the simulated data, and we found that it did a good job of distinguishing transitive from intransitive models and in estimating parameters not only when the generating model satisfied the model assumptions, but also when model assumptions were violated in this way. When the generating model violated the assumptions, statistical tests of the TE fitting models correctly detected the violations. Even when the data contained violations of the TE model, the parameter estimates representing probabilities of true preference patterns were surprisingly accurate, except for error rates, which were inflated by model violations. In a second set of simulations, the generating model either had error rates that were or were not independent of true preferences and transitivity either was or was not satisfied. It was found again that the TE analysis was able to detect the violations of the fitting model, and the analysis correctly identified whether the data had been generated by a transitive or intransitive process; however, in this case, estimated incidence of a preference pattern was reduced if that preference pattern had a higher error rate. Overall, the violations could be detected and did not affect the ability of the TE analysis to discriminate between transitive and intransitive processes.

The CNI model of moral decision-making is a formal model that quantifies (1) sensitivity to consequences, (2) sensitivity to moral norms, and (3) general preference for inaction versus action in responses to moral dilemmas. Based on a critique of the CNI model’s conceptual assumptions, properties of the moral dilemmas for research using the CNI model, and the robustness of findings obtained with the CNI model against changes in model specifications, Baron and Goodwin (2020) dismissed the CNI model as a valid approach to study moral dilemma judgments. Here, we respond to their critique, showing that Baron and Goodwin’s dismissal of the CNI model is based on: (1) misunderstandings of key aspects of the model; (2) a conceptually problematic conflation of behavioral effects and explanatory mental constructs; (3) arguments that are inconsistent with empirical evidence; and (4) reanalyses that supposedly show inconsistent findings resulting from changes in model specifications, although the reported reanalyses did not actually use the CNI model and proper analyses with the CNI model yield consistent findings across model specifications. Although Baron and Goodwin’s critique reveals a need for greater precision in the description of the three model parameters and for greater attention to properties of individual dilemmas, the available evidence indicates that the CNI model is a valid, robust, and empirically sound approach to gaining deeper insights into the determinants of moral dilemma judgments, overcoming major limitations of the traditional approach that pits moral norms against consequences for the greater good (e.g., trolley dilemma).

Presidential elections can be forecast using information from political and economic conditions, polls, and a statistical model of changes in public opinion over time. However, these “knowns” about how to make a good presidential election forecast come with many unknowns due to the challenges of evaluating forecast calibration and communication. We highlight how incentives may shape forecasts, and particularly forecast uncertainty, in light of calibration challenges. We illustrate these challenges in creating, communicating, and evaluating election predictions, using the Economist and Fivethirtyeight forecasts of the 2020 election as examples, and offer recommendations for forecasters and scholars.