We used correlation and spectral analyses to investigate the cognitive structures and processes producing biased judgments. We used 5 different sets of driving problems to exemplify problems that trigger biases, specifically: (1) underestimation of the impact of occasional slow speeds on mean speed judgments, (2) overestimation of braking capacity after a speed increase, (3) the time saving bias (overestimation of the time saved by increasing a high speed further, and underestimation of time saved when increasing a low speed), (4) underestimation of increase of fatal accident risk when speed is increased, and (5) underestimation of the increase of stopping distance when speed is increased. The results verified the predicted biases. A correlation analysis found no strong links between biases; only accident risk and stopping distance biases were correlated significantly. Spectral analysis of judgments was used to identify different decision rules. Most participants were consistent in their use of a single rule within a problem set with the same bias. The participants used difference, average, weighed average and ratio rules, all producing biased judgments. Among the rules, difference rules were used most frequently across the different biases. We found no personal consistency in the rules used across problem sets. The complexity of rules varied across problem sets for most participants.

Risk attitudes implied by valuations of risk-increasing assets depart markedly from those implied by valuations of risk-reducing assets. For instance, many are unwilling to pay the expected value for a risky asset or for its perfect hedge. Although nearly every theory of risk preference (and logic) demands a negative correlation between valuations of bets and hedges, we observe positive correlations. This inconsistency is difficult to expunge.

Kenneth R. Hammond (1917–2015) made several major contributions to the science of human judgment and decision making. As a student of Egon Brunswik, he kept Brunswik’s legacy alive – advancing his theory of probabilistic functionalism and championing his method of representative design. Hammond pioneered the use of Brunswik’s lens model as a framework for studying how individuals use information from the task environment to make clinical judgments, which was the precursor to much ‘policy capturing’ and ‘judgment analysis’ research. Hammond introduced the lens model equation to the study of judgment processes, and used this to measure the utility of different forms of feedback in multiple-cue probability learning. He extended the scope of analysis to contexts in which individuals interact with one another – introducing the interpersonal learning and interpersonal conflict paradigms. Hammond developed social judgment theory which provided a comprehensive quantitative approach for describing and improving judgment processes. He proposed cognitive continuum theory which states that quasi-rationality is an important middle-ground between intuition and analysis and that cognitive performance is dictated by the match between task properties and mode of cognition. Throughout his career, Hammond moved easily from basic laboratory work to applied settings, where he resolved policy disputes, and in doing so, he pointed to the dichotomy between theories of correspondence and coherence. In this paper, we present Hammond’s legacy to a new generation of judgment and decision making scholars.

The transitivity axiom is common to nearly all descriptive and normative utility theories of choice under risk. Contrary to both intuition and common assumption, the little-known ’Steinhaus-Trybula paradox’ shows the relation ’stochastically greater than’ will not always be transitive, in contradiction of Weak Stochastic Transitivity. We bespoke-design pairs of lotteries inspired by the paradox, over which individual preferences might cycle. We run an experiment to look for evidence of cycles, and violations of expansion/contraction consistency between choice sets. Even after considering possible stochastic but transitive explanations, we show that cycles can be the modal preference pattern over these simple lotteries, and we find systematic violations of expansion/contraction consistency.

Individuals avoid objects that have been in physical contact with morally offensive or disgusting entities. This has been called negative magical contagion, an implicit belief in the transmission of essence by physical contact. Alternatively, individuals may avoid a negatively contaminated object because: 1) the object is a strong reminder of the original contagion source (association account); or 2) the act of interacting with the object signals specific information about the self (social communication account). We report that: 1) people often prefer to interact with an entity that they believe is more associated with a negative source rather than an entity that is less associated but has made physical contact with the same negative source; 2) while an associative account requires that contact enhances association, a study of memory for visual pairings of objects indicates that when objects are touching, their associative link (recall) is no greater than when they are in proximity; and 3) subjects continue to show aversion to (prefer to wear gloves to handle) an object that contacted a negative entity even if they are handling the object in order to physically destroy it, hence strongly signaling their rejection of that object. Association and social communication are at best partial accounts for contagion effects.

Predictions of magnitudes (costs, durations, environmental events) are often given as uncertainty intervals (ranges). When are such forecasts judged to be correct? We report results of four experiments showing that forecasted ranges of expected natural events (floods and volcanic eruptions) are perceived as accurate when an observed magnitude falls inside or at the boundary of the range, with little regard to its position relative to the “most likely” (central) estimate. All outcomes that fell inside a wide interval were perceived as equally well captured by the forecast, whereas identical outcomes falling outside a narrow range were deemed to be incorrectly predicted, in proportion to the magnitude of deviation. In these studies, ranges function as categories, with boundaries distinguishing between right or wrong predictions, even for outcome distributions that are acknowledged as continuous, and for boundaries that are arbitrarily defined (for instance, when the narrow prediction interval is defined as capturing 50 percent and the wide 90 percent of all potential outcomes). However, the boundary effect is affected by label. When the upper limit of a range is described as a value that “can” occur (Experiment 5), outcomes both below and beyond this value were regarded as consistent with the forecast.

We consider the recently-developed “surprisingly popular” method for aggregating decisions across a group of people (Prelec, Seung and McCoy, 2017). The method has shown impressive performance in a range of decision-making situations, but typically for situations in which the correct answer is already established. We consider the ability of the surprisingly popular method to make predictions in a situation where the correct answer does not exist at the time people are asked to make decisions. Specifically, we tested its ability to predict the winners of the 256 US National Football League (NFL) games in the 2017–2018 season. Each of these predictions used participants who self-rated as “extremely knowledgeable” about the NFL, drawn from a set of 100 participants recruited through Amazon Mechanical Turk (AMT). We compare the accuracy and calibration of the surprisingly popular method to a variety of alternatives: the mode and confidence-weighted predictions of the expert AMT participants, the individual and aggregated predictions of media experts, and a statistical Elo method based on the performance histories of the NFL teams. Our results are exploratory, and need replication, but we find that the surprisingly popular method outperforms all of these alternatives, and has reasonable calibration properties relating the confidence of its predictions to the accuracy of those predictions.

It has been argued that individuals behave according to a threshold level of concern decision rule when considering protection against risk: if the perceived probability of the risk is below a threshold level, then the likelihood of the risk is treated as zero and protection is deemed unnecessary. Little is known about the determinants of this threshold nor about whether individual thresholds are related to risk specific emotions like worry and regret. We study threshold probabilities and factors that influence these in the context of flood insurance decision making. Based on data collected from 1,041 Dutch homeowners, we find that on average the threshold level of concern for flood insurance demand is negatively related to the expected regret an individual might feel about not purchasing flood insurance if a flood occurs, as well as to worry about flooding.

Most decision-making models describing individual differences in heuristics and biases tasks build on the assumption that reasoners produce a first incorrect answer in a quick, automatic way which they may or may not override later and that the advantage of high capacity reasoners arises from this late correction mechanism. To investigate this assumption, we developed a mouse-tracking analysis technique to capture individuals’ first answers and subsequent thinking dynamics. Across two denominator neglect task experiments, we observed that individuals initially move the mouse cursor towards the correct answer option in a substantial number of cases suggesting that reasoners may not always produce an incorrect answer first. Furthermore, we observed that, compared to low capacity reasoners, high capacity individuals revise their first answer more frequently if it is incorrect and make fewer changes if it is correct. However, we did not find evidence that high capacity individuals produce correct initial answers more frequently. Consistent with the predictions of previous decision-making models, these results suggest that in the denominator neglect task the capacity-normativity relationship arises after the initial response is formulated. The present work demonstrates how the analysis of mouse trajectories can be utilized to investigate individual differences in decision-making and help us better apprehend the dynamics of thinking behind decision biases.

Two studies replicate the anomaly identified by Frederick, Meyer and Levis (2015) and Frederick, Levis, Malliaris and Meyer (2018). People show typical risk averse behavior by valuing risk below the focal lottery’s expected value, but they do not bid above its expected value for the hedge that eliminates the risk. Following the authors, we conduct finer analyses by separating participants into two groups – “experts” who understand that acquiring the hedge makes winning certain versus “novices” who do not understand the winning implications of acquiring the hedge. We find that (1) “experts” are more inclined to purchase the hedge compared to the “novices” and (2) unlike the “novices,” they value the hedge significantly more than the risk instrument, but only if they are given the risk instrument free of charge. However, even there, the hedge valuations are significantly less than the lottery’s expected value suggesting that the anomaly described in Frederick et al. (2015, 2018) is robust and likely to affect the way our discipline conceptualizes and models risk behavior.

In many sports contests, the equilibrium requires players to randomize across repeated rounds, i.e., exhibit no temporal predictability. Such sports data present a window into the (in)efficiency of random sequence generation in a natural competitive environment, where the decision makers (tennis players) are both highly experienced and incentivized compared to laboratory studies. I resolve a long-standing debate about whether professional players’ tennis serve directions are serially independent (Hsu, Huang & Tang, 2007) or not (Walker & Wooders, 2001) using a new dataset that is two orders of magnitude larger than those studies. I examine both between- and within-player determinants of the degree of serial (in)dependence. Evidence of the existence of significant serial dependence across serves is presented, even among players ranked Number 1 in the world. Furthermore, significant heterogeneity was found with respect to the strength of serial dependence and also its sign. A novel finding is that Number 1 and Number 2 ranked players tend to under-alternate on average, whereas in line with previous findings, the lower-ranked the players, the greater their tendency to over-alternate. Within-player analyses show that high-ranked players do not condition their randomization behavior on their opponent’s ranking. However, the under-alternation of top players would be consistent with a best-response to beliefs that the population of opponents over-alternates on average. Finally, the degree of observed serial dependence is not systematically related to other match variables proxying for match difficulty, fatigue, and psychological pressure.

The representativeness heuristic suggests that similarity judgments provide a basis for judgments of likelihood. We use Tversky’s (1977) contrast model of similarity to design tests of this underlying mechanism. If similarity is used to judge likelihood, factors that are known to affect similarity should also affect judgments of likelihood. In two experiments, we manipulated two such factors described in the contrast model of similarity: the nature of the task and context effects. In a between-subject design, respondents assessed either similarity of fictive citizens of 15th century Florence, or the likelihood that they belonged to the same family. The factors that affected similarity also affected the likelihood judgments. These results support the assumption that similarity is an important contributor to judgments of likelihood.

Redundant or excessive information can sometimes lead people to lean on it unnecessarily. Certain experimental designs can sometimes bias results in the researcher’s favor. And, sometimes, interesting effects are too small to be studied, practically, or are simply zero. We believe a confluence of these factors led to a recent paper (Isaac & Brough, 2014, JCR). This initial paper proposed a new means by which probability judgments can be led astray: the category size bias, by which an individual event coming from a large category is judged more likely to occur than an event coming from a small one. Our work shows that this effect may be due to instructional and mechanical confounds, rather than interesting psychology. We present eleven studies with over ten times the sample size of the original in support of our conclusion: We replicate three of the five original studies and reduce or eliminate the effect by resolving these methodological issues, even significantly reversing the bias in one case (Study 6). Studies 7–8c suggest the remaining two studies are false positives. We conclude with a discussion of the subtleties of instruction wording, the difficulties of correcting the record, and the importance of replication and open science.

A new theory of decision-making under risk, the Opportunity-Threat Theory is proposed. Analysis of risk into opportunity and threat components allows description of behavior as a combination of opportunity seeking and threat aversion. Expected utility is a special case of this model. The final evaluation is an integration of the impacts of opportunity and threat with this expectation. The model can account for basic results as well as several “new paradoxes” that refuted cumulative prospect theory in favor of configural weight models. The discussion notes similarities and differences of this model to the configural weight TAX model, which can also account for the new paradoxes.

Adaptive actors must be able to use probabilities as decision weights. In a computerized multi-attribute task, the authors examined the decisions of children (5–6 years, n = 44; 9–10 y., n = 39) and adults (21–22 y., n = 31) in an environment that fosters the application of a weighted-additive strategy that uses probabilities as weights (WADD: choose option with highest sum of probability-value products). Applying a Bayesian outcome-based strategy classification procedure from adult research, we identified the utilization of WADD and several other strategies (lexicographic, equal weight, naïve Bayes, guessing, and saturated model) on the individual level. As expected based on theory, the prevalence of WADD-users in adults was high. In contrast, no preschoolers could be classified as users of probability-sensitive strategies. Nearly one-third of third-graders used probability-sensitive strategies.

True and Error Theory (TET) provides a method to separate the variability of behavior into components due to changing true policy and to random error. TET is a testable theory that can serve as a statistical model, allowing one to evaluate substantive theories as nested, special cases. TET is more accurate descriptively and has theoretical advantages over previous approaches. This paper presents a freely available computer program in R that can be used to fit and evaluate both TET and substantive theories that are special cases of it. The program performs Monte Carlo simulations to generate distributions of test statistics and bootstrapping to provide confidence intervals on parameter estimates. Use of the program is illustrated by a reanalysis of previously published data testing whether what appeared to be violations of Expected Utility (EU) theory (Allais paradoxes) by previous methods might actually be consistent with EU theory.

This study analyzes the relative accuracy of experts, polls, and the so-called ‘fundamentals’ in predicting the popular vote in the four U.S. presidential elections from 2004 to 2016. Although the majority (62%) of 452 expert forecasts correctly predicted the directional error of polls, the typical expert’s vote share forecast was 7% (of the error) less accurate than a simple polling average from the same day. The results further suggest that experts follow the polls and do not sufficiently harness information incorporated in the fundamentals. Combining expert forecasts and polls with a fundamentals-based reference class forecast reduced the error of experts and polls by 24% and 19%, respectively. The findings demonstrate the benefits of combining forecasts and the effectiveness of taking the outside view for debiasing expert judgment.

We estimate the effects of repeated exposure to the Cognitive Reflection Test (CRT) by examining 14,053 MTurk subjects who took the test up to 25 times. In contrast with inferences drawn from self-reported prior exposure to the CRT, we find that prior exposure usually fails to improve scores. On average, respondents get only 0.024 additional items correct per exposure, and this small increase is driven entirely by the minority of subjects who continue to spend time reflecting on the items. Moreover, later scores retain the predictive validity of earlier scores, even when they differ, because initial success and later improvement appear to measure the same thing.

Whether there are gender differences in lying has been largely debated in the past decade. Previous studies found mixed results. To shed light on this topic, here I report a meta-analysis of 8,728 distinct observations, collected in 65 Sender-Receiver game treatments, by 14 research groups. Following previous work and theoretical considerations, I distinguish three types of lies: black lies, which benefit the liar at a cost for another person; altruistic white lies, which benefit another person at a cost for the liar; and Pareto white lies, which benefit both the liar and another person. The results show that: males are significantly more likely than females to tell black lies(N=4,173); males are significantly more likely than females to tell altruistic white (N=2,940); and results are inconclusive in the case of Pareto white lies(N=1,615). Furthermore, gender differences in telling altruistic white lies are significantly stronger than in the other two cases.

As individuals from different nations increasingly interact with each other, research on national in-group favoritism becomes particularly vital. In a cross-national, large-scale study (N = 915) including representative samples from four Latin American nations (Chile, Peru, Colombia, Venezuela) and the USA, we explore differences regarding nationality-based in-group favoritism. In-group favoritism is assessed through differences in prosocial behavior toward persons from the own nation as compared to persons from other nations in fully incentivized one-shot dictator games. We find strong evidence for national in-group favoritism for the overall sample, but also significant differences among national subsamples. Latin Americans show more national in-group favoritism compared to US Americans (interacting with Latin Americans). While US Americans mainly follow an equal split norm (for both in- and out-group interactions), Latin Americans do so only in in-group interactions. The magnitude of in-group favoritism increases with social distance toward the out-group.