People see more of their biases in algorithms.

Algorithmic bias occurs when algorithms incorporate biases in the human decisions on which they are trained. We find that people see more of their biases (e.g., age, gender, race) in the decisions of algorithms than in their own decisions. Research participants saw more bias in the decisions of algorithms trained on their decisions than in their own decisions, even when those decisions were the same and participants were incentivized to reveal their true beliefs. By contrast, participants saw as much bias in the decisions of algorithms trained on their decisions as in the decisions of other participants and algorithms trained on the decisions of other participants. Cognitive psychological processes and motivated reasoning help explain why people see more of their biases in algorithms. Research participants most susceptible to bias blind spot were most likely to see more bias in algorithms than self. Participants were also more likely to perceive algorithms than themselves to have been influenced by irrelevant biasing attributes (e.g., race) but not by relevant attributes (e.g., user reviews). Because participants saw more of their biases in algorithms than themselves, they were more likely to make debiasing corrections to decisions attributed to an algorithm than to themselves. Our findings show that bias is more readily perceived in algorithms than in self and suggest how to use algorithms to reveal and correct biased human decisions.

Psychological ownership: Actors' and observers' perspectives.

Psychological ownership may be judged differently or similarly for self and others. Potential differences in how ownership is evaluated by actors and observers raise important questions about the concept of ownership (what is Mine, Ours, and Theirs) and how to resolve conflicting perceptions.

Correction, uncertainty, and anchoring effects.

We compare the predictions of two important proposals made by De Neys to findings in the anchoring effect literature. Evidence for an anchoring-and-adjustment heuristic supports his proposal that system 1 and system 2 are non-exclusive. The relationship between psychophysical noise and anchoring effects, however, challenges his proposal that epistemic uncertainty determines the involvement of system 2 corrective processes in judgment.

Preference for human, not algorithm aversion.

People sometimes exhibit a costly preference for humans relative to algorithms, which is often defined as a domain-general algorithm aversion. I propose it is instead driven by biased evaluations of the self and other humans, which occurs more narrowly in domains where identity is threatened and when evaluative criteria are ambiguous.

Why do people eat the same breakfast every day? Goals and circadian rhythms of variety seeking in meals.

People exhibit a circadian rhythm in the variety of foods they eat. Many people happily eat the same foods for breakfast day after day, yet seek more variety in the foods they eat for lunch and dinner. We identify psychological goals as a driver of this diurnal pattern of variety seeking, complementing other biological and cultural drivers. People are more likely to pursue hedonic goals for meals as the day progresses, which leads them to seek more variety for dinners and lunches than breakfasts. We find evidentiary support for our theory in studies with French and American participants (N = 4481) using diary data, event reconstruction methods, and experiments. Both endogenously and exogenously induced variation in hedonic goal activation modulates variety seeking in meals across days. Hedonic goal activation predicts variety seeking for meals when controlling for factors including time devoted to meal preparation and eating, the presence or absence of other people, and whether people ate a meal inside or outside their home. Goal activation also explain differences in time spent on meals, whereas increasing time spent on meals does not increase variety seeking. Finally, we observed that a similar increase in hedonic goal activation enacts a larger increase in variety seeking at breakfast than at lunch than at dinner, suggesting a diminishing marginal effect of hedonic goal activation on variety seeking.

Noise Increases Anchoring Effects.

We introduce a theoretical framework distinguishing between anchoring effects, anchoring bias, and judgmental noise: Anchoring effects require anchoring bias, but noise modulates their size. We tested this framework by manipulating stimulus magnitudes. As magnitudes increase, psychophysical noise due to scalar variability widens the perceived range of plausible values for the stimulus. This increased noise, in turn, increases the influence of anchoring bias on judgments. In 11 preregistered experiments (N = 3,552 adults), anchoring effects increased with stimulus magnitude for point estimates of familiar and novel stimuli (e.g., reservation prices for hotels and donuts, counts in dot arrays). Comparisons of relevant and irrelevant anchors showed that noise itself did not produce anchoring effects. Noise amplified anchoring bias. Our findings identify a stimulus feature predicting the size and replicability of anchoring effects-stimulus magnitude. More broadly, we show how to use psychophysical noise to test relationships between bias and noise in judgment under uncertainty.

Understanding, explaining, and utilizing medical artificial intelligence.

Medical artificial intelligence is cost-effective and scalable and often outperforms human providers, yet people are reluctant to use it. We show that resistance to the utilization of medical artificial intelligence is driven by both the subjective difficulty of understanding algorithms (the perception that they are a 'black box') and by an illusory subjective understanding of human medical decision-making. In five pre-registered experiments (1-3B: N = 2,699), we find that people exhibit an illusory understanding of human medical decision-making (study 1). This leads people to believe they better understand decisions made by human than algorithmic healthcare providers (studies 2A,B), which makes them more reluctant to utilize algorithmic than human providers (studies 3A,B). Fortunately, brief interventions that increase subjective understanding of algorithmic decision processes increase willingness to utilize algorithmic healthcare providers (studies 3A,B). A sixth study on Google Ads for an algorithmic skin cancer detection app finds that the effectiveness of such interventions generalizes to field settings (study 4: N = 14,013).

Psychological ownership: implicit and explicit.

Object ownership changes how people perceive objects and self through psychological ownership-the feeling that a thing is MINE. Psychological ownership usually tracks legal ownership, but the two can and do diverge. In this integrative review, I propose a dual-process model of psychological ownership. Antecedents of psychological ownership form self-object associations prompting an implicit inference of psychological ownership, which can then be accepted, corrected, or rejected by explicit judgments. The model explains cases where psychological ownership and legal ownership conflict and predicts psychological ownership felt in a variety of relationships between people and objects, including objects they legally own and use, objects they use but do not legally own, and objects they legally own but do not use.

Which social comparisons influence happiness with unequal pay?

We examine which social comparisons most affect happiness with pay that is unequally distributed (e.g., salaries and bonuses). We find that ensemble representation-attention to statistical properties of distributions such as their range and mean-makes the proximal extreme (i.e., the maximum or minimum) and distribution mean salient social comparison standards. Happiness with a salary or bonus is more affected by how it compares to the distribution mean and proximal extreme than by exemplar-based properties of the payment, like its comparison to the nearest payment or its distribution rank. This holds for randomly assigned and performance-based payments. Process studies demonstrate that ensemble representations lead people to spontaneously select these statistical properties of pay distributions as comparison standards. Exogenously increasing the salience of less extreme exemplars moderates the influence of the maximum on happiness with pay, but exogenously increasing the salience of the distribution maximum does not. As with other social comparison standards, top-down information moderates their selection. Happiness with a bonus payment is influenced by the largest payment made to others who solve the same math problems, for instance, but not by the largest payment made to others who solve different verbal problems. Our findings yield theoretical and practical insights about which members of groups are selected as social comparison standards, effects of relative income on happiness, and the attentional processes involved in ensemble representation. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Debiasing Training Improves Decision Making in the Field.

The primary objection to debiasing-training interventions is a lack of evidence that they improve decision making in field settings, where reminders of bias are absent. We gave graduate students in three professional programs (N = 290) a one-shot training intervention that reduces confirmation bias in laboratory experiments. Natural variance in the training schedule assigned participants to receive training before or after solving an unannounced business case modeled on the decision to launch the Space Shuttle Challenger. We used case solutions to surreptitiously measure participants' susceptibility to confirmation bias. Trained participants were 19% less likely to choose the inferior hypothesis-confirming solution than untrained participants. Analysis of case write-ups suggests that a reduction in confirmatory hypothesis testing accounts for their improved decision making in the case. The results provide promising evidence that debiasing-training effects transfer to field settings and can improve decision making in professional and private life.

Impact bias or underestimation? Outcome specifications predict the direction of affective forecasting errors.

Affective forecasts are used to anticipate the hedonic impact of future events and decide which events to pursue or avoid. We propose that because affective forecasters are more sensitive to outcome specifications of events than experiencers, the outcome specification values of an event, such as its duration, magnitude, probability, and psychological distance, can be used to predict the direction of affective forecasting errors: whether affective forecasters will overestimate or underestimate its hedonic impact. When specifications are positively correlated with the hedonic impact of an event, forecasters will overestimate the extent to which high specification values will intensify and low specification values will discount its impact. When outcome specifications are negatively correlated with its hedonic impact, forecasters will overestimate the extent to which low specification values will intensify and high specification values will discount its impact. These affective forecasting errors compound additively when multiple specifications are aligned in their impact: In Experiment 1, affective forecasters underestimated the hedonic impact of winning a smaller prize that they expected to win, and they overestimated the hedonic impact of winning a larger prize that they did not expect to win. In Experiment 2, affective forecasters underestimated the hedonic impact of a short unpleasant video about a temporally distant event, and they overestimated the hedonic impact of a long unpleasant video about a temporally near event. Experiments 3A and 3B showed that differences in the affect-richness of forecasted and experienced events underlie these differences in sensitivity to outcome specifications, therefore accounting for both the impact bias and its reversal. (PsycINFO Database Record

Overcorrection for Social-Categorization Information Moderates Impact Bias in Affective Forecasting.

Plural societies require individuals to forecast how others-both in-group and out-group members-will respond to gains and setbacks. Typically, correcting affective forecasts to include more relevant information improves their accuracy by reducing their extremity. In contrast, we found that providing affective forecasters with social-category information about their targets made their forecasts more extreme and therefore less accurate. In both political and sports contexts, forecasters across five experiments exhibited greater impact bias for both in-group and out-group members (e.g., a Democrat or Republican) than for unspecified targets when predicting experiencers' responses to positive and negative events. Inducing time pressure reduced the extremity of forecasts for group-labeled but not unspecified targets, which suggests that the increased impact bias was due to overcorrection for social-category information, not different intuitive predictions for identified targets. Finally, overcorrection was better accounted for by stereotypes than by spontaneous retrieval of extreme group exemplars.

More Similar but Less Satisfying: Comparing Preferences for and the Efficacy of Within- and Cross-Category Substitutes for Food.

When people cannot get what they want, they often satisfy their desire by consuming a substitute. Substitutes can originate from within the taxonomic category of the desired stimulus (i.e., within-category substitutes) or from a different taxonomic category that serves the same basic goal (i.e., cross-category substitutes). Both a store-brand chocolate (within-category substitute) and a granola bar (cross-category substitute), for example, can serve as substitutes for gourmet chocolate. Here, we found that people believe that within-category substitutes, which are more similar to desired stimuli, will more effectively satisfy their cravings than will cross-category substitutes (Experiments 1, 2a, and 2b). However, because within-category substitutes are more similar than cross-category substitutes to desired stimuli, they are more likely to evoke an unanticipated negative contrast effect. As a result, unless substitutes are equivalent in quality to the desired stimulus, cross-category substitutes more effectively satisfy cravings for the desired stimulus (Experiments 3 and 4).

Explanations of the endowment effect: an integrative review.

The endowment effect is the tendency for people who own a good to value it more than people who do not. Its economic impact is consequential. It creates market inefficiencies and irregularities in valuation such as differences between buyers and sellers, reluctance to trade, and mere ownership effects. Traditionally, the endowment effect has been attributed to loss aversion causing sellers of a good to value it more than buyers. New theories and findings--some inconsistent with loss aversion--suggest evolutionary, strategic, and more basic cognitive origins. In an integrative review, we propose that all three major instantiations of the endowment effect are attributable to exogenously and endogenously induced cognitive frames that bias which information is accessible during valuation.

Reliance on luck: identifying which achievement goals elicit superstitious behavior.

People often resort to superstitious behavior to facilitate goal achievement. We examined whether the specific type of achievement goal pursued influences the propensity to engage in superstitious behavior. Across six studies, we found that performance goals were more likely than learning goals to elicit superstitious behavior. Participants were more likely to engage in superstitious behavior at high than at low levels of chronic performance orientation, but superstitious behavior was not influenced by chronic learning orientation (Studies 1 and 2). Similarly, participants exhibited stronger preferences for lucky items when primed to pursue performance goals rather than learning goals (Studies 3 and 4). As uncertainty of goal achievement increased, superstitious behavior increased when participants pursued performance goals but not learning goals (Study 5). Finally, assignment to use a lucky (vs. unlucky) item resulted in greater confidence of achieving performance goals but not learning goals (Study 6).

Interference of the end: why recency bias in memory determines when a food is consumed again.

The results of three experiments reveal that memory for end enjoyment, rather than beginning enjoyment, of a pleasant gustatory experience determines how soon people desire to repeat that experience. We found that memory for end moments, when people are most satiated, interferes with memory for initial moments. Consequently, end moments are more influential than initial moments when people decide how long to wait until consuming a food again. The findings elucidate the role of memory in delay until repeated consumption, demonstrate how sensory-specific satiety and portion sizes influence future consumption, and suggest one process by which recency effects influence judgments and decisions based on past experiences.

The (perceived) meaning of spontaneous thoughts.

Spontaneous thoughts, the output of a broad category of uncontrolled and inaccessible higher order mental processes, arise frequently in everyday life. The seeming randomness by which spontaneous thoughts arise might give people good reason to dismiss them as meaningless. We suggest that it is precisely the lack of control over and access to the processes by which they arise that leads people to perceive spontaneous thoughts as revealing meaningful self-insight. Consequently, spontaneous thoughts potently influence judgment. A series of experiments provides evidence supporting two hypotheses. First, we hypothesize that the more a thought is perceived to be spontaneous, the more it is perceived to provide meaningful self-insight. Participants perceived more spontaneous kinds of thought (e.g., intuition) to reveal greater self-insight than did more controlled kinds of thought in Study 1 (e.g., deliberation). In Studies 2 and 3, participants perceived thoughts with the same content and target to reveal greater self-insight when spontaneously rather than deliberately generated (i.e., childhood memories and impressions formed). Second, we hypothesize that the greater self-insight attributed to thoughts that are (perceived to be) spontaneous leads those thoughts to more potently influence judgment. Participants felt more sexually attracted to an attractive person whom they thought of spontaneously than deliberately in Study 4, and reported their commitment to a current romantic relationship would be more affected by the spontaneous rather than deliberate recollection of a good or bad experience with their romantic partner in Study 5.

Does liking or wanting determine repeat consumption delay?

Does liking or wanting predict the delay between consumption episodes? Although these psychological processes are correlated, we find that memory for liking, rather than wanting, determines the number of days that pass until the consumption of a food is repeated. Experiment 1 found that liking (but not wanting) for a food at the end of a consumption experience predicted how many days passed until participants wanted to consume it again. Experiment 2 showed that mitigating the decrease in liking resulting from the repeated consumption of a food eliminates its effect on delay. Together, these findings suggest that end liking has a greater influence on when people will consume a food again in the future.