People systematically underestimate the extent to which dissimilar others are interested in socially connecting, a phenomenon we refer to as dissimilarity pessimism. Across six experiments, in the lab (N = 795), online (N = 2,681), and in the field (N = 457), we document this pessimism across multiple kinds of dissimilarity, including age, race, personality, job type, and sociocultural background. We explore two psychological processes that appear associated with these effects: greater uncertainty about how interested others are in connecting, relative to one's own interest in connecting; and a tendency for the thoughts that spontaneously come to mind about an interaction to skew negative when imagining how a partner might perceive them. We also find that people's overly pessimistic beliefs are associated with their choice of whom to approach, identifying a previously underappreciated antecedent of "homophily," people's tendency to associate with similar others. In other words, people may avoid dissimilar others not because they themselves are uninterested or because they lack the opportunity-the two explanations most commonly offered in the literature-but because they mistakenly believe dissimilar others are uninterested in them. Finally, we conduct a field experiment in collaboration with Mystery Minds, an organization dedicated to connecting employees across differences, and find that this pessimism emerges before naturally occurring workplace conversations, even among people who had opted into meeting dissimilar colleagues and personally welcomed the interaction. Together, these findings advance our understanding of metaperception, intergroup interaction, homophily, and the social cognition of conversation. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

Self-enhancement is considered a hallmark of grandiose narcissism, yet recent applications of condition-based regression analysis have questioned whether narcissistic admiration, the dimension of grandiose narcissism reflecting assertive self-enhancement, is reliably linked to self-views that overestimate one's attributes. We reconcile these divergent accounts by integrating a process-oriented perspective on self-view formation with condition-based regression analysis, which we use as a diagnostic tool to differentiate two distinct forms of self-enhancement: unanchored self-enhancement (i.e., generally positive self-views) and anchored self-enhancement (i.e., overestimation of self-views relative to objective benchmarks). Across four studies, we examined how actual cognitive ability (truth force) and narcissistic admiration (bias force) jointly influence self-views through cue-based inference and how these processes generate the distinct forms of self-enhancement effects in condition-based regression analysis. In Study 1, a simulation, we illustrated that the relation between narcissistic admiration and anchored inflation effects depends on degree of self-knowledge: Paradoxically, some accuracy is required for anchored self-enhancement to emerge. In Study 2 (N = 1,035), we experimentally manipulated self-knowledge, showing that greater accuracy strengthens the association between narcissistic admiration and self-enhancement. In Study 3 (N = 454), we longitudinally demonstrated that narcissists' self-views rise over time through biased cue-based inference. Finally, in Study 4 (N = 4,950), an individual participant data meta-analysis, we found that, under typical levels of self-knowledge, narcissistic admiration is related to anchored self-enhancement of cognitive ability. Cumulatively, these findings reestablish narcissism as the prototypical self-enhancing trait while clarifying the cognitive conditions under which the distinct forms of self-enhancement are detectable. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

While trait adjectives may be regarded as convenient summary labels for varieties of behavioral expressions with common themes and outcomes, in turn, behaviors typical of a person may be considered as expressions of personality dispositions. In fact, in daily life, behavioral descriptions are way more frequent than the more abstract adjectival descriptors. For the present study, behavioral, verb-based data sets collected in four languages were used for comparison. The data sets each consisted of ratings from 400 participants on 543 Dutch personality verbs, 1,247 participants on 414 Polish personality verbs, 951 participants on 289 Czech personality verbs, and 512 participants on 319 Serbian personality verbs. Principal component analysis-based structures (of different dimensionalities) were compared pairwise across the languages, making use of congruence coefficients. Moreover, using simultaneous component analysis, the verb-based structure common to the four languages, based on a shared data set with ratings from 3,110 participants on 324 personality verbs, was analyzed. Three- and six-component structures based on the common set are discussed in some detail, both showing portions of the Big Five but with no reference to the Intellect-Imagination factor. It is concluded that verb-based behavioral items, in comparison to adjective- and noun-based items, give rise to a diluted and reduced trait-semantic space. To arrive at a full trait-semantic representation, all linguistic categories should be exploited in a joint operation. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

The relationship between narcissism and aggression has received considerable research attention, with the main conclusion being that individuals with higher levels of narcissism tend to exhibit more aggressive behavior. However, most evidence comes from cross-sectional and experimental studies using between-person analyses and relies heavily on Western convenience samples, leaving the temporal dynamics and within-person processes largely unexplored. This preregistered study examined four large-scale data sets of Chinese undergraduates and offenders: two ambulatory assessment samples (427 undergraduates and 775 offenders; 19,331 and 42,820 observations, respectively) and two long-term longitudinal samples (N = 1,404, 969, 822, 613 for undergraduates and 2,064, 1,716, 1,609, 1,393 for offenders, each with four assessments over a year). Results indicated that narcissism and aggression prospectively predicted one another at the momentary level, but not the long-term level. These reciprocal effects were largely replicated across undergraduate and offender samples. Person-level narcissism and aggression moderated several of these dynamic associations, with the most robust pattern showing that higher person-level aggression consistently strengthened the effects of narcissistic states on subsequent aggression among offenders. Overall, this study provides, to our knowledge, the first within-person evidence for dynamic reciprocal associations between narcissism and aggression, while also offering stronger support for the robustness, cultural generalizability, and population representativeness of these processes. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

Loneliness is not only a stable dispositional trait but also a momentary state that can fluctuate within individuals. However, empirical studies on how contextual factors relate to state loneliness remain scarce. This study investigated how state loneliness varies depending on social company, everyday activities, physical locations, and the duration of exposure to these contextual factors. Drawing on preregistered and exploratory analyses, we examined data from the Day Reconstruction Method module of the German Socio-Economic Panel Innovation Sample, collected across four waves between 2012 and 2015. The representative sample comprised 22,031 episodes nested within 2,508 participants (52.6% female, age range 17-95 years, Mage = 50.0, SDage = 18.9). Multilevel logistic regression models showed that being alone consistently predicted higher odds of state loneliness, while the presence of others showed opposing associations. Participants were more likely to feel lonely when at home than when at work or elsewhere. Watching television was linked to a higher probability of reporting state loneliness, whereas self-care/health and media use showed a negative relationship. For some activities, this relationship was moderated by the duration of the activities: Reporting state loneliness was more likely during prolonged episodes of care work but less likely during prolonged episodes of social contact and television consumption. Overall, these findings contribute to a more robust, nuanced understanding of the contextual and temporal features tied to momentary feelings of loneliness, underscoring the need for dynamic, time-sensitive approaches in both research and interventions addressing everyday experiences of loneliness. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

Across six experiments, we examined how sensitivity to change varies with protected values (PVs). In Experiment 1, vegetarians were less appreciative than meat-eaters of large (but only partial) reductions in other people's meat consumption, but they were more appreciative than meat-eaters of smaller but absolute reductions (i.e., reduction to zero meat consumption). In Experiment 2, a similar pattern emerged in political groups: Republicans, who value immigration control more than Democrats, were relatively less sensitive to large but partial reductions in illegal immigration than to its elimination; the reverse pattern was observed for gun possession. In Experiment 3, we manipulated participants' values regarding child labor. Individuals who generated arguments in defense of child labor held less protected values and were comparatively more sensitive to change in child labor numbers. Further studies showed that protected values make people focus on the outcome (i.e., the present state) more than on the trajectory of change (i.e., the difference from the past; Experiment 4) and that they inspire stricter thresholds for what is considered acceptable (Experiment 5). Experiment 6 shows that protected values cause a heightened sensitivity to absolute changes to zero, rather than insensitivity to changes overall. These findings indicate that protected values amplify sensitivity to absolute changes to zero, compared to partial changes-even larger ones-that fall short of zero. The results have theoretical implications for theories of progress perception and quantity (in)sensitivity. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

Reports an error in "People in historically rice-farming areas are less happy and socially compare more than people in wheat-farming areas" by Cheol-Sung Lee, Thomas Talhelm and Xiawei Dong (Journal of Personality and Social Psychology, 2023[May], Vol 124[5], 935-957; see record 2023-14674-001). In the article (https://doi.org/10.1037/pspa0000324), the tables report the range of values for urbanization. Tables 5 and 7 report a range of 16-88. The range should be 23-88. The analyses use the correct urbanization variable throughout. In Table 6 and Table S8, the word "log" should appear for the GDP variable. All analyses in the article (including these tables) use log GDP per capita. In the Happiness subsection of Study 3, the original article stated that the Supplemental Materials have the full wordings of the happiness items used in Study 3. However, the full questionnaire is available on the Open Science Framework (OSF) rather than in the Supplemental Materials: https://osf.io/7bnjm/?view_only=4211ec8265e84a6a88c9ec9ec6e1fb55. As a point of clarification, Studies 1 and 2 used income and occupational status as key predictor variables. Therefore, the analysis did not include respondents who were missing data for these variables. The article did not state this explicitly. To be clear, analyses in Studies 1 and 2 included respondents with positive income and valid occupational information: In the Samples subsection, 3,780 participants in Study 1 (World Values Survey) and 12,265 participants in Study 2 (China Family Panel Study). (The following abstract of the original article appeared in record 2023-14674-001.) Using two nationally representative surveys, we find that people in China's historically rice-farming areas are less happy than people in wheat areas. This is a puzzle because the rice area is more interdependent, and relationships are an important predictor of happiness. We explore how the interdependence of historical rice farming may have paradoxically undermined happiness by creating more social comparison than wheat farming. We build a framework in which rice farming leads to social comparison, which makes people unhappy (especially people who are worse off). If people in rice areas socially compare more, then people's happiness in rice areas should be more closely related to markers of social status like income. In two studies, national survey data show that income, self-reported social status, and occupational status predict people's happiness twice as strongly in rice areas than wheat areas. In Study 3, we use a unique natural experiment comparing two nearby state farms that effectively randomly assigned people to farm rice or wheat. The rice farmers socially compare more, and farmers who socially compare more are less happy. If interdependence breeds social comparison and erodes happiness, it could help explain the paradox of why the interdependent cultures of East Asia are less happy than similarly wealthy cultures. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

When making inferences about others, people often need to integrate multiple types of information, including the person's social categories (e.g., occupation) as well as other "individuating" information (e.g., their behaviors). The current work reexamines this integration process to understand when and why it produces biases involving overreliance on category information. We identify two key challenges in identifying such biases, and develop a novel Bayesian modeling approach to overcome them. As a first step in applying this approach, the current work examined a set of novel predictions based on viewing the continuum model of impression formation in light of this Bayesian approach. Specifically, a series of six studies tested whether occupation categories might be overused in general, or especially in conditions thought to reduce effortful processing (i.e., greater cognitive load or information consistency). Results showed no evidence of category overuse, nor of shifts in this direction, in any of these conditions (though some ambiguity remained about the load manipulation's effectiveness). This provides initial evidence against the idea that intrinsic differences in how these categories are processed or represented will lead to their overuse. A seventh study provided the first direct evidence of a case where categories were overused: when contextual factors (i.e., background goals) made them the category especially relevant, while processing resources were limited. More broadly, the current work developed the theoretical and methodological foundations for identifying category overuse that stems from biased inference processes and demonstrated the power of this approach for understanding when and why such biases occur. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

Recent research has examined the state dynamics of narcissistic grandiosity and vulnerability and of positive and negative affect, yet few studies have directly linked the two conceptually similar constructs or addressed their conceptual overlap, especially beyond Western samples. In this study, we used a newly collected Chinese sample (N = 594, observations = 26,802) and an existing American data set (N = 862, observations = 36,631) to investigate the dynamics of narcissism and affect in daily life. Using dynamic structural equation modeling, we characterized these dynamics through key indicators-mean, variability, contemporaneous correlations, autoregressive effects, and cross-lagged effects-and examined their intercorrelations and associations with dispositional variables. Results showed that indicators within each dynamic were closely interrelated and substantially linked to dispositional traits. The dynamics of narcissism and affect exhibited moderate similarity with grandiosity corresponding to positive affect and vulnerability to negative affect. The two dynamics mainly differed in the associations between their two components: Grandiosity and vulnerability were generally more positively correlated than positive and negative affect, which were somewhat inversely related. These patterns generalized well across Chinese and American samples, although Chinese participants showed a stronger cross-lagged effect of vulnerability on subsequent grandiosity. Additionally, grandiosity and positive affect showed more adaptive external associations with trait variables, whereas variability in affect and narcissism showed more maladaptive external associations in the Chinese sample. This study bridges narcissism and affect research, advances understanding of their dynamics, and enhances the generalizability of findings by examining them across two national contexts. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

Equality is widely endorsed in principle, yet equality-enhancing policies are routinely opposed as if one group's gain is another's loss. We propose that group status fundamentally shapes motivation to enhance equality-above and beyond ideological beliefs, threat-based reactance, personal gain, or underlying cognitive biases-determining whether individuals see equality-enhancing policies as harmful. Across five preregistered lab and representative field studies (N = 5,116), we find that advantaged-but not disadvantaged-group members systematically perceived equality policies as harmful to advantaged group resource access, even when policies explicitly preserve or improve access. Effects persisted when non-zero-sum policies were jointly evaluated with explicitly zero-sum policies, suggesting a mechanism other than cognitive zero-sum biases (Study 2). Independent of participants' particular group membership, both advantaged societal status (Study 3) and contextual resource advantage (Study 4) predicted lower motivation to enhance equality between two groups, which in turn predicted perceived resource access. Experimentally manipulating motivation to enhance equality or to preserve the status quo directly affected both advantaged and disadvantaged participants' perceptions of policy harm (Study 5). Equality may be elusive not only because of cognitive zero-sum biases, threat reactance, and vocal antagonists, but also because the social motivations imbued by one's hierarchical position dictate whether equality itself is seen as necessarily zero-sum. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

Influencing others through conversation is fundamental to social interactions. With the rise of generative artificial intelligence (GenAI), people are increasingly relying on conversational AI for guidance on consumption and lifestyle choices, from picking a cuisine to choosing a travel destination. However, can conversational AI exert enough persuasive influence to serve as a viable substitute for human opinions? Across five experiments involving real conversations, participants discussed an opinion-based topic (e.g., whether it is worth subscribing to more than one streaming service) with a conversation partner framed as either an AI or a human. Unbeknownst to them, all participants actually conversed with a large language model prompted to try to change their minds on the topic. We found that participants showed greater attitude change toward their partner's counterattitudinal position (i.e., became more persuaded) when their partner was framed as an AI (vs. a human). We demonstrate that this occurs because people tend to view AI as inherently less persuadable than humans. As a result, they become less argumentative with an AI-framed (vs. human-framed) partner, and this reduced argumentativeness, in turn, predicts greater attitude change. Consistent with this mechanism, the effect is mediated by linguistic markers of reduced argumentativeness and attenuated when people have a goal to persuade their partner, when an AI-framed partner is described as persuadable, and when the interaction format is constrained (i.e., when argumentation is precluded and thus the partner is structurally unpersuadable). These findings advance understanding of how persuasion unfolds in conversations with AI. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

Powerholders can be the object of great respect and admiration, but they may also be distrusted and "canceled" in the court of public opinion. This raises the question of whether power is an asset or a detriment when one is implicated in wrongdoing. Different theoretical perspectives informed competing hypotheses, suggesting power may either decrease (shielding hypothesis) or increase (liability hypothesis) judgments of culpability. Here, we report four main studies and 10 supplemental studies that addressed this conundrum, using diverse transgressions, methods, and samples. In Study 1, participants judged higher power individuals to be likelier sources of deceitful die-rolling reports than lower power individuals, and they trusted them less in an incentivized behavioral trust game. In Study 2, participants rated higher power people as more probable perpetrators of financial fraud than lower power people, and this effect was mediated by perceived impunity (i.e., powerholders seemingly believing they can get away with anything). Study 3 was an archival study of 97,628 forum posts about diverse transgressions where there was ambiguity about who was the culprit, and users rendered judgments. We found that higher power posters were about twice as likely to be deemed guilty than lower power posters. Finally, Study 4 excluded generic power saliency as an alternative explanation and replicated the mediation via perceived impunity. These findings indicate that, besides many perks, power comes with perils in the form of ascribed culpability. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

Amid rising political polarization, derogatory rhetoric targeting social groups has become common and highly contested across party lines. These partisan divides are often attributed to disagreement about whether bias has occurred, and theories of social norms, harm-based moral judgment, and cognitive consistency suggest that recognizing bias should track moral condemnation. Yet other theories suggest that recognition and condemnation are psychologically distinct. We propose that, even when partisans similarly recognize explicit bias ("yeah, it's biased"), they may diverge in moral response ("but …")-depending on the target's advantaged or disadvantaged status. Across six prestudies (N = 1,197), eight preregistered experiments (N = 7,145), and 16 supplemental studies, including two preregistered experiments (N = 1,029), we examined this "yeah, but" phenomenon. For speech that partisans equivalently recognized as biased based on race and gender, Democrats condemned bias against disadvantaged targets more than Republicans; Republicans condemned bias against advantaged targets more than Democrats-patterns that extended to endorsing sanctions (Study 1). Partisans invoked the same justifications-invulnerability to harm and honesty-but selectively so, depending on the target (Study 2). These patterns emerged for biases based on socioeconomic status and religion (Studies 3-4), behaviors (Study 5), and partisanship (which lacks a stable status hierarchy; Study 6). However, a status-juxtaposition intervention (judging bias against both advantaged and disadvantaged groups) narrowed Democrats' within-party asymmetry (Studies 7a-b). By disentangling bias recognition from bias condemnation, we show that moral polarization may persist even when Democrats and Republicans both "see" bias and invoke the same moral concerns. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

The present work evaluated the predictive model for national life satisfaction from the World Happiness Report. Drawing on national data from 165 countries spanning 19 annual surveys (2005-2023), between- and within-nation variation in national life satisfaction and six predictors were examined: log gross domestic product per capita, healthy life expectancy at birth, and social perceptions concerning generosity, support, freedom, and corruption. The pooled regression model employed by the World Happiness Report explained 75% of the variance in national life satisfaction. In contrast, two-level models generated for the present analyses explained 84% of the between-nation variance and either 25% (fixed effects model) or 67% (random effects model) of the within-nation variance. Nation-specific regression models generated for the present analyses explained between 19% and 97% of the variance within individual countries. The statistical significance, magnitude, and sign/direction for each of the six predictors varied widely depending on the analytic model and level of analysis. Few countries were characterized by a pattern of predictive effects that matched the pooled regression model. Thus, rather than suggesting a single set of results that generalizes around the world, the predictive model for national life satisfaction appears to differ by country. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

Artificial intelligence (AI) systems increasingly serve as advisors, evaluators, and decision-makers, yet little is known about how people perceive AI as a social entity. Across five primary studies and eight supplementary studies, we show that people assign racial identities to AI systems, overwhelmingly perceiving them as White-even though these systems provide no visual, vocal, or identity cues from which race might ordinarily be inferred. Using forced-choice, open-ended, and implicit reverse-correlation measures, Studies 1 and 2 demonstrate that AI is explicitly and implicitly associated with Whiteness across diverse samples. Study 3 extends these findings beyond the United States, showing that AI is perceived as White in Japan and India. Study 4 examines the implications of AI racialization, showing that perceiving an AI system as more White is associated with greater trust in and persuasiveness of its recommendations. Supplementary studies identified two potential mechanisms: stereotype spillover linking intelligence with Whiteness and ecosystem-based inferences based on beliefs about who creates, trains, and uses AI. Building on this account, Study 5 provides causal evidence by isolating a key ecosystem cue-the racial composition of AI training data. Participants assigned less cognitively demanding tasks to AI systems described as trained on Black and Latino data than to otherwise identical systems trained on White or unspecified data. Together, these findings show that AI is not perceived as socially neutral but instead acquires racial meanings associated with credibility, authority, and capability, demonstrating how social categories shape perceptions of novel technological entities beyond their underlying algorithmic properties. (PsycInfo Database Record (c) 2026 APA, all rights reserved).