RESEARCH DOSSIER
HEXACO and Alternative Personality Trait Models
Abstract. This fundamental research monograph critically examines the HEXACO six-factor model in comparison with major alternative trait systems, including the Five-Factor Model, Eysenck's PEN, reinforcement sensitivity theory, Cloninger's temperament–character model, Zuckerman's Alternative Five, the interpersonal circumplex, and maladaptive trait taxonomies. Its questions concern how dimensions are discovered and defined, whether factor structures reproduce across methods and populations, which criteria support differential predictive validity, and how far biological or clinical explanations can be inferred from descriptive scores. The evidence base includes original lexical studies, psychometric investigations, self–observer comparisons, laboratory behavioral experiments, cross-cultural tests, and published meta-analyses. Evidence particularly supports the usefulness of Honesty–Humility for predicting exploitation-related criteria while revealing limits of cross-cultural measurement, factor correspondence, and individual inference. This study is a critical narrative synthesis, not a new experiment, preregistered systematic review, or meta-analysis.
Keywords: HEXACO; Honesty–Humility; personality trait models; Big Five; Five-Factor Model; Eysenck PEN; reinforcement sensitivity theory; BIS/BAS; Cloninger temperament and character; Alternative Five; interpersonal circumplex; factor analysis; construct validity; cross-cultural psychometrics; observer reports; incremental validity; workplace deviance; dark traits; Artificial Era
The Research Object and the Problem of Competing Trait Models
Personality trait models are organized representations of recurrent differences between people in characteristic patterns of experience, motivation, and behavior. A dimension is useful when an assessment can identify variation reliably and connect that variation with information from another method, occasion, or outcome. HEXACO and alternative trait models differ in the processes they emphasize, the behaviors they treat as representative, and the statistical structures they impose on observations. The research object of this dossier is the comparative scientific status of those model architectures: what they describe, where they overlap, what empirical observations distinguish them, and which explanatory claims are warranted Cronbach & Meehl (1955) McCrae & John (1992).
The six-factor HEXACO structure receives the most detailed analysis because it makes specific competing predictions about fairness, exploitation, emotional vulnerability, anger regulation, and social cooperation. Eysenck's PEN model, Gray's reinforcement sensitivity theory, Cloninger's temperament–character framework, Zuckerman's Alternative Five, and the interpersonal circumplex enter as independently motivated comparison systems. Maladaptive personality trait measures receive a bounded analysis because their selection of pathological content addresses a different target population and measurement purpose. The objective is model comparison, including theoretical vocabulary, instruments, factor structure, criterion validity, generalizability, and risk of overinterpretation Ashton & Lee (2007) Eysenck (1990) Corr (2004) Cloninger et al. (1993) Zuckerman et al. (1993) Krueger et al. (2012).
This work is a critical narrative research synthesis drawing on identified original investigations, instrument-development studies, theory papers, cross-cultural samples, and published meta-analyses. It does not claim preregistered comprehensive retrieval, risk-of-bias scoring for every relevant experiment, independent reanalysis of raw data, or pooled new effect sizes. Empirical statements identify the corresponding published study and retain the boundaries of its design. Strong claims about universal structure, neurobiology, clinical translation, or prediction are evaluated by asking which observable evidence would discriminate alternatives rather than by treating the popularity of a taxonomy as confirmation Cronbach & Meehl (1955) Ashton & Lee (2025).
The Scientific History of Trait Taxonomies
The history of structural personality science is inseparable from decisions about which behaviors and words should enter an inventory. Lexical research assumes that recurrent, socially meaningful differences are represented to some extent in ordinary personality-descriptive language. Statistical factors are subsequently estimated from covariation among the selected terms. The language sample, translation, rating target, number of retained factors, and rotation method all influence the resulting structure. Modern research therefore treats a factor model as an empirical abstraction from specified measurements and communities, not a timeless natural inventory of human attributes McCrae & John (1992) Lee & Ashton (2008).
Five-factor approaches consolidated a descriptive vocabulary centered on Extraversion, Agreeableness, Conscientiousness, Neuroticism, and Openness to Experience. McCrae and John (1992) synthesized lexical and questionnaire evidence for this hierarchy, while six-factor research subsequently identified additional recurrent content relating to sincerity, fairness, modesty, and avoidance of exploitation. Ashton and Lee (2005, 2007) developed a systematic case for Honesty–Humility and for reconceptualizing the content of Agreeableness and Emotionality. The scientific dispute concerns the factor space generated by an explicit sampling procedure, not whether five or six is an intrinsically preferable number McCrae & John (1992) Ashton & Lee (2005) Ashton & Lee (2007).
Other lines of research followed different routes. Eysenck foregrounded broad biologically oriented dimensions, Gray reformulated motivational sensitivity around responses to reinforcement and conflict, Cloninger combined temperament and self-concept character dimensions, and Wiggins structured interpersonal behavior around agency and communion. Their dates and conceptual claims mark genuinely distinct research programs. Treating all of them as mere alternative spellings for the Big Five would erase differences in experimental commitments, construct boundaries, and criterion selection Eysenck (1990) Corr (2004) Cloninger et al. (1993) Wiggins (1979).
What Makes a Trait Model Scientifically Informative
A trait model makes two separable claims. A structural claim proposes a relatively parsimonious organization of individual differences; an external-validity claim proposes relationships with observations not used to derive that organization. Replication of a questionnaire factor pattern supports the first claim under the sampled conditions. Prediction of independent choices, peer reports, or prospective behavior supports the second. Explanatory causal claims require further experimental and longitudinal evidence. This distinction is central to construct validation because a model can summarize covariation accurately yet leave the processes producing it substantially unknown Cronbach & Meehl (1955) Lee & Ashton (2004).
Comparative tests should match models on measured breadth, item count, reliability, criterion relevance, and validation design. A six-domain inventory can appear superior if it samples exploitation more directly than the five-domain competitor; this is a substantive gain in content coverage but does not automatically prove a superior global ontology. Conversely, a five-factor measure with appropriately selected lower-order facets can capture some criterion variance attributed to a separate domain. Ashton and Lee's (2008) comparison explicitly considered whether a proxy Honesty–Humility score from Five-Factor facets would reduce predictive differences Ashton & Lee (2008).
Researchers must also specify the level of the construct: domain, facet, item nuance, motivational mechanism, interpersonal vector, or maladaptive trait. Apparent cross-model disagreement may arise when an entire broad domain is compared with one narrowly focused scale. Evidence of measurement invariance, stable factor interpretation, and incremental validity is more informative when compared at the same level of abstraction. This methodological discipline keeps the model question answerable through future data rather than definitional advocacy Ion et al. (2017) Lee & Ashton (2018) Krueger et al. (2012).
Lexical Evidence and the Emergence of Six Factors
The HEXACO proposal arises from lexical factor-analytic results in multiple languages, followed by a theoretical interpretation of the resulting dimensions. Lee and Ashton (2008) examined 449 English adjectives selected by frequency of personality-descriptive use and compared questionnaire marker correlations with adjective factors. They also studied similarities between factor descriptions and indigenous lexical summaries across twelve languages. Convergent patterns supported recovery of six broad conceptual dimensions in those materials, including Honesty–Humility Lee & Ashton (2008).
That program supplies evidence about the descriptive social vocabulary available in the sampled linguistic traditions. It does not prove that every population will yield the same solution or that all language-specific dimensions can be reduced to six. Vocabulary coverage, interviewer comprehension, social desirability, factor extraction, and translation equivalence can change the results. The analysis gains authority from cross-language convergence while remaining limited to the included languages and measurement contexts. Direct indigenous trait discovery and translated questionnaires test related but distinct propositions about universality Lee & Ashton (2008) Ion et al. (2017) Gurven et al. (2013).
Six factors also constitute a choice of analytic resolution. Changing extraction criteria can expose finer-grained facets or merge broad regions of the space. Competing rotations may move evaluative and antagonistic content between factors while leaving many item correlations similar. The appropriate scientific question is whether independent data identify stable factor meaning, adequate convergent and discriminant associations, and useful outcome prediction. The lexical program is therefore a powerful beginning for construct investigation, followed by stringent tests rather than a terminal demonstration Lee & Ashton (2004) Ashton & Lee (2007) Cronbach & Meehl (1955).
The Six HEXACO Domains and Their Psychological Content
Honesty–Humility encompasses sincerity, fairness, greed avoidance, and modesty. Emotionality includes fearfulness, anxiety, dependence, and sentimentality. Extraversion includes social self-esteem, boldness, sociability, and liveliness. Agreeableness emphasizes forgiveness, gentleness, flexibility, and patience; Conscientiousness includes organization, diligence, perfectionism, and prudence; Openness includes aesthetic appreciation, inquisitiveness, creativity, and unconventionality. These descriptions correspond to the operationalized HEXACO inventory rather than to a claim that every colloquial use of a domain name has the same boundaries Lee & Ashton (2004) Ashton & Lee (2009) Lee & Ashton (2018).
Each domain captures a distribution of questionnaire and informant responses; it is not a unitary biological substance. A person's domain score aggregates responses to distinct behaviors and attitudes, so two people with comparable totals may differ at the facet level. This issue is especially important for the comparison of HEXACO Emotionality with Big Five Neuroticism and HEXACO Agreeableness with Big Five Agreeableness. Identical labels do not guarantee shared item content, and different labels do not necessarily imply independent factors Ashton et al. (2014) McCrae & John (1992).
The inventory includes an interstitial Altruism versus Antagonism scale in some forms, which is informative because not all socially consequential content sits neatly inside one factor. Lee and Ashton (2006) assessed additional scales and an observer form in a community sample; this extended the earlier student self-report work. A well-specified six-factor structure can therefore acknowledge meaningful cross-domain locations and facet heterogeneity without abandoning its broad taxonomy Lee & Ashton (2006) Ashton & Lee (2025).
Honesty–Humility as a Distinctive Domain
Honesty–Humility concerns willingness to engage with others without manipulating, exploiting, or claiming disproportionate status and material advantage. Its importance follows from measurable differences in sincere and fair conduct rather than from a simple assertion of moral worth. Ashton and Lee (2005) showed why conventional five-factor Agreeableness did not exhaust the same conceptual space, and their subsequent theoretical account emphasized reciprocal cooperation and the distribution of social benefits. This interpretation offers hypotheses about partner choice and temptation, with evidence most direct where behavior is independently observed Ashton & Lee (2005) Ashton & Lee (2007).
A study of psychopathy, Machiavellianism, and narcissism reported substantial negative correlations of all three questionnaire traits with Honesty–Humility in its sample. Lee and Ashton (2005) thus demonstrated descriptive convergence with measures of antagonistic personality content. Those correlations do not establish clinical diagnoses, biological causes, or individual misconduct. They do indicate that a dimension combining exploitativeness and entitlement can organize variance that may be dispersed across parts of Five-Factor Agreeableness and other facets Lee & Ashton (2005).
The domain's explanatory usefulness depends on making predictions beyond shared self-report wording. Tests must distinguish fair behavior in situations with real incentives from the endorsement of socially approved claims. They should also distinguish low Honesty–Humility from contextual self-protection, justified distrust, and culturally distinctive norms of modesty. A validated score becomes a probabilistic predictor under specified conditions, not a moral verdict about an individual. The strongest evaluation connects independent ratings and behavior with appropriate controls Hilbig & Zettler (2015) Ashton & Lee (2025).
Emotionality, Neuroticism, Fear, and Sentimentality
HEXACO Emotionality and Five-Factor Neuroticism overlap in anxiety-related content yet partition their remaining components differently. The HEXACO domain brings fearfulness, emotional dependence, and sentimentality together, while irritability and anger are chiefly represented in the negative pole of HEXACO Agreeableness. Five-Factor Neuroticism commonly combines anxiety, vulnerability, anger, and related negative affect. Consequently, translating scores between frameworks by simply relabeling the axis would systematically mischaracterize some profiles Ashton et al. (2014) McCrae & John (1992).
This reallocation has theoretical consequences. Fearful attachment to significant others and empathic concern may coexist with relatively little anger; a propensity for hostile reaction can coexist with low fearfulness. Correlations among such experiences are observable, but a model must decide whether they should be organized by general negative affect, prosocial affiliation, or differentiated defensive and affiliative processes. Ashton, Lee, and de Vries (2014) review this distinction and its implication for prediction. That paper provides a theoretical integration of empirical results, not an experimental proof of a unique emotional mechanism Ashton et al. (2014).
When a clinical or health variable is associated with a domain, the measured items determine what the association means. A study using HEXACO Emotionality has not necessarily tested the same hypothesis as one using Neuroticism. Crosswalk analyses therefore require common samples, aligned facets, and independent criterion measurement. Neither anxiety scores nor high sensitivity to threat can be read as a psychiatric diagnosis from a personality inventory alone Ashton et al. (2014) Krueger et al. (2012).
HEXACO Agreeableness and Regulation of Anger
HEXACO Agreeableness is oriented toward tolerance, forgiveness, and cooperative response after provocation. Its low pole contains anger, resentment, and retaliation; its higher pole emphasizes willingness to refrain from harsh responses when treated unfairly. Honesty–Humility, by contrast, concerns willingness to refrain from exploiting others when doing so could be advantageous. This is a theoretically important difference between resisting temptation to exploit and resisting retaliation after perceived exploitation Ashton et al. (2014).
The separation matters for interpersonal predictions. Someone may be reluctant to exploit others yet respond strongly to insult, or be patient under insult while using strategic manipulation in advantageous situations. Those combinations are possible within the six-factor arrangement and could be lost when both tendencies are collapsed into a general pleasantness score. Correlational discrimination must nevertheless be tested through multiple raters and behavioral tasks rather than inferred from hypothetical examples alone Ashton & Lee (2008) Ashton & Lee (2025).
A relational research design should vary competitive pressure, provocation, anonymity, and opportunities to repair relationships. It should measure reactions over time and distinguish emotion, expressed anger, and choice of action. The HEXACO formulation then yields testable contrasts: exploitation should relate especially to Honesty–Humility, retaliatory anger especially to Agreeableness, with context and their interaction shaping observed behavior. Such propositions remain conditional empirical hypotheses even when the domains are reliably measured Ashton & Lee (2007) Ashton et al. (2014).
Extraversion, Conscientiousness, and Openness Across Models
Extraversion, Conscientiousness, and Openness retain broadly recognizable content between HEXACO and Big Five systems, including sociability, planful self-regulation, and interest in ideas and novelty. Their similarity is an important empirical constraint on claims of radical separation between the models. However, the inventories select and weight facets differently; correlations between broad scores depend on available items and respondents. Psychometric convergence therefore identifies overlapping descriptive variance, while criterion comparisons can still favor one instrument for particular research questions Lee & Ashton (2018) McCrae & John (1992).
Extraversion also overlaps with approach tendencies in reinforcement sensitivity theory, and Conscientiousness can be related to restraint and regulatory organization. The overlap is partial. A person may pursue incentives strongly while preferring solitary rather than social activities; another may maintain diligent routines without the same physiological sensitivity to anticipated reward. A cross-model comparison should test these possibilities using parallel questionnaires and independent incentive tasks instead of treating social energy as synonymous with reward sensitivity Carver & White (1994) Corr (2004).
Openness raises distinctive issues in cross-cultural and observer research because its verbal and aesthetic indicators may depend on access to education and shared cultural activities. Ashton and Lee's (2025) review also notes that self and observer reports of Openness can differ systematically. Broad structural correlations are therefore interpretable only after considering content opportunity, observer access, and comparable item meaning. Such constraints apply equally to the five- and six-factor traditions Ashton & Lee (2025) Ion et al. (2017).
The HEXACO Personality Inventory and Its Revision
The original HEXACO Personality Inventory was developed to operationalize six factors identified in lexical research. Lee and Ashton (2004) reported a validation study of more than 400 respondents, with 24 facet-level scales and evidence for internal consistency, expected factor structure, and convergent validity. The design made the theoretical taxonomy measurable, while remaining initially concentrated in self-report data. A scale's validation, however, is a continuing program across populations, occasions, observers, and intended interpretations, not a permanent property guaranteed by its title Lee & Ashton (2004).
The HEXACO-PI-R and abbreviated forms address different practical constraints. Ashton and Lee (2009) described a 60-item version allocating ten items to each domain, aiming at broad coverage with reduced respondent burden. Lee and Ashton (2018) examined the 100-item measure using an online self-report sample of 100,318 and a student self/observer sample of 2,868. These are substantial measurement datasets, but their sampling arrangements and respondent recruitment are part of the evidentiary interpretation; size does not in itself confer national population representativeness Ashton & Lee (2009) Lee & Ashton (2018).
The longer instrument supports richer facet investigation at a higher time cost, whereas a shorter form can attenuate lower-order precision. For empirical applications researchers should document the exact form, language, scoring key, missing-data rule, target population, and intended level of inference. The same numerical label cannot be compared across translations, versions, or age groups until measurement equivalence and scale functioning have been examined Lee & Ashton (2018) Ion et al. (2017).
Self-Report, Observer Reports, and Multimethod Convergence
A self-report inventory asks respondents to summarize their own characteristic tendencies. Informant ratings can introduce observations inaccessible to introspection but are affected by relationship context, opportunity to observe, and assumed similarity. Lee and Ashton (2006) used observer reports as a validation step; Ashton and Lee (2025) reviewed evidence for convergence across longer Big Five and HEXACO instruments. In their synthesis, agreement tended to be somewhat lower for cooperation-related traits than for several other domains, an important boundary for socially evaluative constructs Lee & Ashton (2006) Ashton & Lee (2025).
The self–observer comparison is not a referendum on which observer possesses the true personality. An acquaintance may witness conduct under cooperative or conflictual circumstances unknown to the target's self-assessment. A participant may understand private motives while presenting a favorable narrative. Convergence across independent sources supports the trait's generalization beyond one channel; disagreement can motivate analysis of situational specificity and measurement method. Designs should collect multiple informed raters, specify their familiarity, and account for nested dependencies among ratings Ashton & Lee (2025).
Social desirability becomes especially salient when measuring sincerity, humility, or cheating. Anonymous item responses may still reflect self-deception, reputation management, and differences in normative interpretation. Experimental honesty tasks, observed workplace actions, and prospective peer reports offer more demanding tests of construct validity. Yet they also contain incentives, opportunities, and situational constraints that must be modeled. No single behavior supplies a pure, error-free indicator of an enduring moral disposition Hilbig & Zettler (2015) Cronbach & Meehl (1955).
Cross-Cultural Replication and Measurement Equivalence
Cross-language lexical convergence is a central foundation of the HEXACO framework. Lee and Ashton (2008) reported correspondence between six-factor descriptions and results from personality vocabularies in multiple linguistic settings. However, this evidence should be separated from direct tests of whether a particular questionnaire preserves item meaning, domain loadings, and scale thresholds after translation. A general conceptual resemblance of factor labels does not establish metric or scalar invariance Lee & Ashton (2008) Ion et al. (2017).
Ion and colleagues (2017) investigated HEXACO structure in samples from India, Indonesia, Oman, Romania, and Thailand, using exploratory structural equation modeling. Their total sample exceeded 1,600 people, and the abstract reports replication in some of the studied cultures. That qualified finding should be preserved: it provides a stronger basis than assuming identical structure everywhere, while demonstrating that cross-cultural equivalence remains an empirical question. Cross-national comparisons of means demand particularly careful examination of response style and scalar functioning Ion et al. (2017).
Research among Tsimane participants in the Bolivian Amazon found that a standard translated Big Five measure failed to reproduce the expected five-factor structure robustly in that context. Gurven and colleagues (2013) tested an alternative cultural setting, revealing how literacy, interpretation, and ecological specialization can constrain imported taxonomies. Their result is directly about the Five-Factor instrument employed and cannot be turned into a claim that HEXACO is thereby universally superior. The appropriate response is locally grounded measurement, transparent adaptation, and replication Gurven et al. (2013) Ion et al. (2017).
The Big Five and Five-Factor Model as Comparison Baselines
Five-factor trait frameworks constitute the most important comparison baseline because they are extensively developed, broadly familiar, and supported by several decades of cross-instrument evidence. The model described by McCrae and John (1992) organizes variation into Extraversion, Agreeableness, Conscientiousness, Neuroticism, and Openness. Its empirical strengths include parsimonious description and substantial predictive work; its limitations include differential item coverage, contextual generalization, and ambiguous facet allocation. Comparisons should use the actual scales administered rather than an abstract claim that a model of five dimensions necessarily omits all Honesty–Humility content McCrae & John (1992) Ashton & Lee (2005).
Ashton and Lee (2008) examined criteria related to Honesty–Humility with self- and observer ratings in three samples exceeding 200 participants each. They reported prediction advantages for HEXACO in that criterion family and considered proxy representation within Five-Factor facet scores. This strengthens the conclusion that deliberate inclusion of exploitation-avoidance indicators matters. It does not establish advantage across every outcome, because some outcomes are more appropriately predicted by other factors or by situational variables Ashton & Lee (2008).
The choice between five and six factors is partly a choice of content allocation and research purpose. For a study of occupational integrity, social reciprocity, or dark traits, six-domain Honesty–Humility may be directly relevant. For broad continuity with longitudinal datasets collected using established Five-Factor inventories, a five-factor analysis may provide stronger comparability. Scientific practice can retain multiple interoperable maps while specifying their correspondence and the limits of score transformation Pletzer et al. (2019) McCrae & John (1992).
Eysenck's PEN Model and the Biological Tradition
Eysenck's three major dimensions are Psychoticism, Extraversion, and Neuroticism. His program aimed to connect broad individual differences with biological and environmental processes and treated inherited contributions as part of a research explanation. In a 1990 synthesis Eysenck considered genetic and environmental contributions to these dimensions. The historical significance lies in the ambition to connect questionnaire dimensions to underlying mechanisms, not in a demonstrated one-to-one mapping from each questionnaire score to a single neurotransmitter or neural circuit Eysenck (1990).
The meaning of Psychoticism requires special attention. It is a personality dimension in this framework, with content involving tough-mindedness, impulsivity, and antagonistic characteristics; it is not a diagnosis of psychosis. Comparisons with HEXACO Honesty–Humility or Agreeableness can identify overlapping content without declaring numerical equivalence. Zuckerman and colleagues' (1993) joint comparison of the Big Three, Big Five, and Alternative Five found considerable common structure in Extraversion and Neuroticism, while placements of impulsivity, antagonism, and openness were more complex Zuckerman et al. (1993).
A biological interpretation of PEN must be evaluated against independent physiological measures, longitudinal prediction, and reproducible findings under realistic environmental controls. Factor correlations alone cannot establish a biological substrate or a specific developmental origin. The model remains important as an historically consequential alternative that foregrounds psychobiological questions distinct from descriptive lexical coverage Eysenck (1990) Corr (2004).
Reinforcement Sensitivity Theory and Motivational Systems
Gray's reinforcement sensitivity research, subsequently revised, models personality-relevant differences through responsiveness to rewarding, punishing, and conflict-associated conditions. Contemporary accounts distinguish a behavioral approach system associated with appetitive incentive processes, a fight–flight–freeze system in defensive responding, and a behavioral inhibition system especially concerned with goal conflict and anxiety. Corr's (2004) historical and theoretical analysis traces revisions of the model from Eysenckian dimensions toward neurobehavioral processes. The framework therefore targets motivational mechanism more directly than a lexical taxonomy Corr (2004).
The distinction between anxiety and fear is theoretically consequential in revised reinforcement sensitivity theory. Threat-related tendencies cannot be collapsed into a single general score without deciding whether an experiment concerns avoidance of aversive stimuli, conflict between approach and avoidance, or responses to inescapable threat. Reward pursuit likewise has components of anticipation, effort, and pleasure. Experimental manipulations should isolate these situations and test whether measured sensitivities predict behavior under the predicted contingencies Corr (2004) Torrubia et al. (2001).
HEXACO and reinforcement sensitivity theory can complement one another in a multimethod model. Extraversion may covary with approach tendencies, and Emotionality with defensive sensitivity, but the mapping is neither exact nor causally established merely by scale correlation. Theoretical progress requires manipulations of incentive and conflict and mediation or longitudinal analyses linking mechanisms with stable questionnaire tendencies. Specific behavioral mechanisms remain hypotheses to test, not properties automatically established by factor names Carver & White (1994) Corr (2004).
BIS/BAS and Sensitivity-to-Reinforcement Instruments
Carver and White (1994) developed BIS/BAS self-report scales and included experimental studies involving anticipation of reward and punishment. The scales predicted differential affective responses under those laboratory conditions, demonstrating an early attempt to link questionnaire variation with experimental contingencies. Later theoretical revisions require care because the original BIS scale was developed within an earlier conception of the system; its label alone does not guarantee selective measurement of the revised conflict-based BIS Carver & White (1994) Corr (2004).
Torrubia, Ávila, Moltó, and Caseras (2001) developed the Sensitivity to Punishment and Sensitivity to Reward Questionnaire, examining correlations with anxiety, impulsivity, and Eysenckian measures. Their work illustrates how distinct instruments can operationalize partly overlapping motivational tendencies while yielding different item distributions. Agreement between questionnaires is one piece of evidence; discriminant predictions from behavioral tasks offer a stronger test of whether their proposed systems are functionally distinct Torrubia et al. (2001).
A research comparison should therefore report the theoretical edition of reinforcement sensitivity theory under examination, the measure actually used, and whether the outcome is a self-description, physiological response, or observable choice. Without this distinction, apparent support for a neural mechanism could be no more than covariance among self-reported adjectives. Multimethod triangulation and temporal separation of predictor and criterion reduce some of these interpretive risks Corr (2004) Carver & White (1994).
Cloninger's Temperament and Character Model
Cloninger, Svrakic, and Przybeck (1993) proposed a psychobiological model combining four temperament dimensions—Novelty Seeking, Harm Avoidance, Reward Dependence, and Persistence—with three character dimensions—Self-Directedness, Cooperativeness, and Self-Transcendence. Their original report described a framework for development and measurement and included general-population data supporting its seven-dimensional structure. The separation of temperament from character represents a theoretical account of habit-related dispositions and self-concept development, not a direct decomposition of questionnaire scores into immutable genetic traits and purely acquired moral identity Cloninger et al. (1993).
Unlike HEXACO, the temperament–character model puts identity, goals, and self-interpretation at the center of part of its dimensional system. It is thus well suited to research questions concerning self-regulation and meanings attached to action. Nonetheless, proposals about biological pathways require their own genetic, neurophysiological, and experimental validation. Naming a factor Novelty Seeking or Harm Avoidance expresses a psychological interpretation; it does not uniquely identify a neurochemical cause Cloninger et al. (1993) Cronbach & Meehl (1955).
Comparisons need to determine which level of explanation is appropriate. HEXACO Conscientiousness may overlap with practical organization relevant to Self-Directedness, and HEXACO Emotionality may overlap with aspects of Harm Avoidance. Such associations do not make the systems interchangeable, because the character dimensions include adaptive and evaluative self-concepts outside the narrower lexical domain structure. The correct analysis tests criterion specificity and validates translations at facet rather than domain-name level Cloninger et al. (1993) Ashton et al. (2014).
Zuckerman's Alternative Five and Impulsive Sensation Seeking
Zuckerman, Kuhlman, Joireman, Teta, and Kraft (1993) directly compared three structural models using questionnaire factor analysis. Their Alternative Five emphasized dimensions including Impulsive Sensation Seeking, Neuroticism–Anxiety, Aggression–Hostility, Activity, and Sociability. Extraversion and Neuroticism had recognizable counterparts across major models, while impulsivity-related and antagonism-related indicators redistributed across dimensions. The empirical comparison helps reveal how supposedly general factors depend on the behavioral content supplied for analysis Zuckerman et al. (1993).
Sensation seeking is especially important for differentiating social energy, preference for stimulation, reward responsiveness, and risk-taking. These processes may be correlated yet distinguishable in decisions under uncertainty and in approach to novel environments. A high score on a sensation-seeking scale does not establish low conscientiousness, antisocial intent, or a psychiatric condition. Experimental measures of exploration and risky choice should specify available incentives, perceived probability, and opportunity for harm Zuckerman et al. (1993) Corr (2004).
The Alternative Five illustrates the limits of purely numerical comparisons. A factor model that isolates impulsive sensation seeking may explain a targeted behavioral criterion more directly than a broader factor arrangement, even if the latter summarizes a wider collection of ordinary trait adjectives. Scientific adequacy is always indexed to construct coverage and the intended prediction task; there is no one-dimensional winner across every purpose Zuckerman et al. (1993) Ashton & Lee (2007).
The Interpersonal Circumplex and Relational Personality
The interpersonal circumplex organizes recurring interpersonal behavior around two broad axes: agency or dominance, and communion or warmth. Positions around the circle represent blends of those dimensions, allowing an interpersonal style to be described through both orientation and intensity. Wiggins (1979) developed a taxonomy of trait-descriptive interpersonal terms, providing an early empirical foundation for a circumplex rather than an arbitrary collection of independent labels. This geometry addresses the relational expression of personality, including interpersonal complementarity and conflict Wiggins (1979).
Markey and Markey (2009) developed a brief interpersonal circumplex instrument, the IPIP-IPC, making the representation usable in measurement research. Circumplex validity requires more than finding two correlated scales: the pattern of relations among angular positions should approximate the predicted circular ordering and magnitude. A broad factor taxonomy such as HEXACO can be mapped onto the circumplex for interpersonal criteria, but domain scores also contain components unrelated to agency and warmth Markey & Markey (2009).
The circumplex is particularly relevant to cooperation, dominance, social threat, and dyadic interaction. Honesty–Humility and Agreeableness may relate to aspects of warmth or antagonism, yet they distinguish exploitation from anger tolerance in ways that a broad communion axis may compress. Conversely, a circumplex can preserve the interpersonal position of a behavior more directly than a general Honesty–Humility score. The models thus support different levels of description in the same encounter Ashton et al. (2014) Wiggins (1979).
Maladaptive Traits, Psychoticism, and the Clinical Boundary
Models devised for general-population descriptive variation and those devised for clinically salient maladaptive traits have different sampling aims. Krueger and colleagues (2012) developed a provisional DSM-5-oriented personality trait inventory incorporating 25 lower-order traits under five broad domains: Negative Affect, Detachment, Antagonism, Disinhibition, and Psychoticism. Its item construction and iterative psychometric development drew on clinically relevant difficulties, so comparisons with HEXACO must acknowledge selection for dysfunction rather than the ordinary full range of trait language Krueger et al. (2012).
The word Psychoticism appears in both Eysenck's PEN tradition and the maladaptive five-domain model, but these constructs have different operational definitions and clinical implications. Eysenck's broad historical personality dimension is not interchangeable with the psychoticism domain of the pathological inventory, and neither alone supplies a diagnosis of a psychotic disorder. The same principle applies to psychopathy and dark-trait measures, which require separate diagnostic and forensic evidence before they support clinical classification Eysenck (1990) Krueger et al. (2012) Lee & Ashton (2005).
Clinical translation should examine impairment, duration, context, distress, and functional consequences using an appropriate validated assessment process. Trait scores can contribute to probabilistic description and to research on risk, but norms, validity evidence, and clinical judgment remain essential. A theoretical correspondence between low Honesty–Humility and exploitative behavior cannot determine an individual's disorder or treatment need. This dossier analyzes model structure and measurement rather than prescribing diagnosis or intervention Krueger et al. (2012) Cronbach & Meehl (1955).
Dark Traits and the Question of a Common Antagonistic Core
Dark Triad research describes Machiavellianism, narcissism, and psychopathy as partly overlapping constructs characterized by different profiles of manipulation, entitlement, callousness, and related interpersonal tendencies. Lee and Ashton (2005) found negative relationships between their assessed Dark Triad scales and Honesty–Humility in a particular questionnaire sample. That result is relevant to construct convergence: a broad low-Honesty–Humility factor captures shared morally and socially consequential variance across otherwise distinct labels Lee & Ashton (2005).
Moshagen, Hilbig, and Zettler (2018) proposed a common dark factor and presented evidence across four studies involving more than 2,500 participants. This research offers another route to organizing antagonistic dispositional variance, while leaving important questions of explanatory level: a statistical general factor may summarize common tendencies without identifying their unique causal source. The fact that different measures yield correlated profiles is compatible with several accounts involving shared item meaning, incentive responsiveness, or social context Moshagen et al. (2018).
Comparative tests should include criterion behavior with variable real incentives, independent measurement of power or opportunity, and distinct trait components. Honesty–Humility can provide a broad predictor, while narrower dark traits may retain incremental validity for specialized outcomes. Neither taxonomy should be transformed into a comprehensive measure of ethical character. Outcomes involving exploitation, hostility, and calculated deception have multiple determinants beyond personality scale scores Hilbig & Zettler (2015) Moshagen et al. (2018).
Behavioral Experiments on Dishonesty and Cooperation
Behavioral experiments provide an important test of whether questionnaire descriptions of honesty and cooperation generalize to decisions in which a participant can benefit from misrepresentation. Hilbig and Zettler (2015) reported six experiments investigating dishonest behavior and the explanatory role of Honesty–Humility. The study provides convergent experimental evidence for trait-associated variation across manipulations, while interpretation still depends on incentive structures, opportunity, anonymity, and the operational definition of dishonest action Hilbig & Zettler (2015).
Such experiments are not equivalent to randomized interventions changing a personality trait itself. They can test predictions of an association between a premeasured disposition and behavior under controlled circumstances, and they can experimentally manipulate situational conditions. A causal inference that an enduring trait produces behavior by a particular mechanism needs additional evidence, such as process measures or repeated prospective designs. Situational constraints may reduce or amplify predictive differences, making trait-by-context interaction scientifically central Hilbig & Zettler (2015) Cronbach & Meehl (1955).
A stronger program includes preregistered predictions, transparent exclusion rules, multiple tasks, independent observers, and replication across laboratories and cultures. Researchers should also measure whether an apparent honesty effect is better explained by social desirability, comprehension of instructions, tolerance of uncertainty, or differential stakes. This procedure distinguishes an interpretable robust trait–behavior relation from an effect limited to one method and its vocabulary Ashton & Lee (2008) Thielmann et al. (2020).
Prosocial Behavior and Social-Dilemma Evidence
Prosocial behavior includes helping, cooperation, generosity, and decisions that affect others' outcomes. The relation between personality and prosocial action is therefore conditional on the kind of interdependence, potential exploitation, and personal cost being studied. Thielmann, Spadaro, and Balliet (2020) integrated experimental social-dilemma findings in a theoretical framework and meta-analysis. Their work supports treating personality as a relevant predictor while separating different social motives and decision environments Thielmann et al. (2020).
HEXACO Honesty–Humility is theoretically well placed to predict fairness and resistance to exploitation, whereas Agreeableness is more centrally implicated in anger regulation and forgiveness. However, cooperation in a repeated relationship may reflect reciprocity, reputation, expected future gains, cultural norms, and task design. A laboratory trust-game result does not automatically generalize to every form of real-world altruism. Studies must document incentives, partner information, order of decisions, and whether responses are experimentally independent Ashton et al. (2014) Thielmann et al. (2020).
Prosocial criteria can also be selected in ways that favor the conceptual domain of one model. This is scientifically acceptable when criterion selection is theoretically justified, but comparisons should include both targeted outcomes and broad cross-domain batteries. Incremental validity must be estimated with adequate sample size and out-of-sample procedures; otherwise performance advantages may reflect overfitting to the chosen study or predictor combination Ashton & Lee (2008) Thielmann et al. (2020).
Workplace Integrity and Counterproductive Behavior
Organizational research is an informative context because rule breaking, interpersonal harm, and counterproductive work behavior are important independent outcomes. Pletzer and colleagues (2019) synthesized studies comparing Big Five and HEXACO domains in relation to workplace deviance. Their meta-analysis identified Honesty–Humility as a major correlate and reported greater explained variance for the HEXACO domain set than for the Big Five set under the included synthesis. This finding provides direct evidence for criterion differentiation in one consequential outcome family Pletzer et al. (2019).
The published comparison reported approximately 31.97% explained variance for HEXACO domains and 19.05% for Big Five domains under its analytic approach. These values belong to that meta-analysis and should not be extrapolated as universal prediction rates or as prospective causal effects. Differences across instruments, employment settings, deviance measures, sample selection, and shared-method reporting may influence the comparison. A robust organizational application requires independent criterion sources and validation in the actual role and jurisdiction Pletzer et al. (2019).
Employment decisions also raise fairness, privacy, and adverse-impact questions that trait validity alone cannot settle. Even a well-replicated group-level correlation cannot justify diagnosing dishonesty from an individual inventory response. The scientific use is to estimate probabilistic associations under specific measurement conditions, evaluate incremental value beyond established predictors, and examine whether model performance transports to new occupational samples Ashton & Lee (2025) Pletzer et al. (2019).
Factor Analysis, Rotation, and Structural Competition
Exploratory factor analysis estimates patterns of shared variation without fixing every parameter in advance, whereas confirmatory models evaluate a specified loading and covariance structure. Exploratory structural equation modeling can combine flexible cross-loadings with multigroup structural tests. In the HEXACO cross-cultural study by Ion and colleagues (2017), this approach was presented as useful for examining equivalence across populations. Model fit is one criterion; factor interpretability, independent replication, and criterion behavior remain necessary Ion et al. (2017).
The number of extracted factors may change with item pool composition, communality, sample heterogeneity, estimation method, and rotation. A fixed simple-structure solution can assign complex traits to a predominant factor while ignoring meaningful secondary loadings. Researchers should report analytic decisions before interpreting domain differences as natural boundaries. Comparing solutions at five, six, or other resolutions is most informative when each is tested on fresh data using equivalent indicator sets Lee & Ashton (2008) McCrae & John (1992).
Structural comparisons should include factor congruence, reliability of domain and facet scores, differential item functioning, and the stability of external correlations. An extra factor may be statistically recoverable yet offer little practical value for a particular target criterion; alternatively, a narrower content factor may add substantial explanatory specificity. The central result is an evidence-based map of which dimensions generalize, not a ritual choice of a preferred factor count Ashton & Lee (2007) Ion et al. (2017).
Incremental Validity, Prediction, and Generalization
Incremental validity asks whether one model predicts an independently measured outcome after accounting for an appropriate competing representation. The comparison should control for predictor reliability and response burden and disclose whether the outcome was chosen because one model explicitly samples its content. Ashton and Lee (2008) compared prediction of Honesty–Humility-relevant criteria, while Pletzer and colleagues (2019) examined workplace deviance across a broader meta-analytic literature. These are informative evidence streams with different methods and generalization boundaries Ashton & Lee (2008) Pletzer et al. (2019).
A fair comparison may report prediction by each complete inventory, by an equal-length item set, and by a small number of theoretically targeted facets. Cross-validated prediction on held-out observations should be preferred to an uncorrected within-sample maximum when many alternative models are tested. Confidence intervals, sample weighting, and the relationship between baseline rate and practical benefit matter for applied decisions. Predictive differences are not automatically clinically or ethically meaningful simply because they are statistically detectable Cronbach & Meehl (1955) Thielmann et al. (2020).
Repeated prospective measurement strengthens inference about stability of prediction and whether outcome information precedes or follows trait measurement. An association in the same questionnaire session may reflect common method variance and overlapping wording. Experimental task measures create opportunities to test behavior more directly, though their construct validity and test–retest reliability must also be established. A defensible synthesis reports strengths and weaknesses separately for lexical replication, psychometric convergence, field prediction, and experimental behavioral validation Hilbig & Zettler (2015) Ashton & Lee (2025).
Development, Stability, and the Limits of Biological Explanation
Trait rankings often display temporal consistency, but stability is a statistical property of repeated distributions rather than a proof that individuals cannot change. The HEXACO domains are intended to capture relatively enduring tendencies while allowing changes in context, developmental priorities, and social roles. Trait-model comparison should therefore include longitudinal rank-order association, mean-level trends, and within-person variability, each of which answers a different question. The studies reviewed here primarily establish factor measurement and criterion relations; they do not by themselves establish a comprehensive HEXACO developmental trajectory Lee & Ashton (2018) Ashton & Lee (2025).
Biological accounts associated with PEN, reinforcement sensitivity theory, and Cloninger propose additional causal mechanisms. Their plausibility does not remove the need for genetic, neurophysiological, and experimental evidence. Eysenck's synthesis considered heritable and environmental components; Corr's review connected motivational sensitivities with neurobehavioral systems; Cloninger proposed differentiated temperament and character. These should be evaluated as particular explanatory hypotheses with their own methods, rather than treated as stronger solely because they employ biological vocabulary Eysenck (1990) Corr (2004) Cloninger et al. (1993).
A mature integrative model would connect repeated observable choices and reports to identifiable processing differences under controlled incentives, follow these links over developmental time, and test moderation by environments. Cross-sectional questionnaire covariance cannot independently establish neurochemical pathways, developmental origins, or a direct causal chain to life outcomes. This principle applies to every alternative trait program considered here Cronbach & Meehl (1955) Carver & White (1994).
Measurement Bias, Sampling, and Reproducibility
Personality research often recruits university students, online volunteers, convenience samples, or people who can complete lengthy questionnaires. The large HEXACO-100 validation samples provide strong evidence of scale functioning among those respondents, while population transport requires further sampling and invariance checks. A respondent who chooses a long online personality assessment may differ systematically from the relevant occupational, rural, older, or clinical population. Precise sample counts cannot substitute for information about selection and representativeness Lee & Ashton (2018) Ion et al. (2017).
Common sources of bias include acquiescence, extreme-response style, socially desirable reporting, translation artifacts, criterion contamination, and investigator flexibility in factor extraction. Self–other agreement can be informative, but acquaintances' ratings also reflect shared environments and selective observation. Ashton and Lee (2025) demonstrate why source independence matters when evaluating socially sensitive constructs. Researchers should prespecify factor identification strategies, report item and facet statistics, and retain access to the analytic code when permitted Ashton & Lee (2025) Lee & Ashton (2006).
Independent replication has several levels: recovery of factor content, reproducibility of scale reliability, agreement among observers, stability of outcome relationships, and convergence across cultural settings. A successful replication in one domain does not license every theoretical interpretation of that domain. Ion and colleagues' (2017) qualified cross-cultural findings illustrate why generalization must be stated at the same level as the evidence, including unsuccessful recovery and disagreement across groups Ion et al. (2017) Gurven et al. (2013).
Model Comparison as an Organized Evidence Synthesis
The evidence considered in this monograph differentiates at least five scientific purposes. HEXACO supplies a relatively broad lexical taxonomy with targeted content for exploitation and fairness; the Five-Factor Model supplies a mature five-domain descriptive baseline; PEN proposes a compact psychobiological orientation; reinforcement sensitivity theory specifies approach, defensive responding, and conflict processes; Cloninger's framework combines temperament with character; and the interpersonal circumplex maps relationship style. The distinction arises from the published research designs, not solely from their verbal descriptions Ashton & Lee (2007) McCrae & John (1992) Eysenck (1990) Corr (2004) Cloninger et al. (1993) Wiggins (1979).
A rigorous comparison follows a sequence: first verify construct definitions and measurement coverage; next evaluate internal and cross-method structure; then measure criterion associations with comparable controls; finally test independent replication and applicability to new populations. No single step substitutes for another. Lexical recovery does not itself prove a specific neural system, a psychobiological theory does not itself guarantee clinical validity, and a high predictive association does not by itself warrant consequential individual classification Cronbach & Meehl (1955) Ion et al. (2017) Krueger et al. (2012).
The strongest convergence concerns the existence of structured, replicable individual differences and the usefulness of quantitative traits in predicting relevant outcomes. Substantive disagreement remains about domain granularity, distribution of anger and fear, exploitation-related content, relationship between self-concept and temperament, generalizability beyond intensively studied populations, and the neural mechanisms underlying dimensions. The appropriate synthesis preserves these as separate research questions instead of declaring an overall winner without specifying the criteria Ashton et al. (2014) Pletzer et al. (2019) Ashton & Lee (2025).
Research Priorities in the Artificial Era
Digital environments create new opportunities to evaluate trait predictions across varied behavioral contexts, including interaction with recommender systems, automated persuasion, synthetic conversational partners, and digital traces of cooperation. Such records require explicit consent, privacy controls, defensible criterion construction, and tests of whether platform affordances select for particular behavior. Responses produced under algorithmic incentives may differ from those observed in conventional questionnaires. A research program can compare HEXACO's fairness-related predictions with incentive-based accounts in controlled human–artificial interaction studies Ashton & Lee (2008) Corr (2004).
Research about artificial agents must specify the unit of analysis carefully. Human attributions of dominance, warmth, sincerity, or personality to an artificial system are psychological observations about users and interaction contexts. Claims that a system itself possesses enduring internal traits require operational definitions, reproducible tests across tasks and model versions, and evidence of cross-context functional continuity. A questionnaire written for human self-report is not automatically a valid measure of artificial psychological organization. Comparisons should distinguish human perceptions, system outputs, and independently validated mechanisms Cronbach & Meehl (1955) Markey & Markey (2009).
The most productive extensions test whether trait and motivational models generalize to situations involving artificial partners and digitally mediated institutions, rather than importing model labels without measurement equivalence. Within the Artificial Era horizon, cross-system research can investigate perceived social trust, response to incentives, prosocial choices, and human–artificial cooperation under controlled conditions. Such proposals are prospective interdisciplinary research questions and carry no claim of an already established artificial personality taxonomy Thielmann et al. (2020) Ashton & Lee (2025).
Open Questions and a Comparative Research Agenda
Several problems remain open. The first concerns whether HEXACO's sixth factor retains stable meaning across languages with very different descriptive vocabularies and among non-WEIRD populations. A second concerns the extent to which carefully constructed Five-Factor facets can match its prediction of honesty-relevant outcomes. A third concerns the distinction between a stable statistical trait and its causal processes. Each question requires independently collected evidence and clearly specified contrasts rather than extended argument over preferred terminology Ion et al. (2017) Ashton & Lee (2008) Gurven et al. (2013).
A fourth problem concerns selection of criterion domains. Research programs may give disproportionate weight to fairness or exploitation while neglecting domains where competing models could be more informative. A fifth concerns multimethod validity in morally loaded contexts, where informant biases and demand characteristics affect outcomes. A sixth concerns the boundary between descriptive traits and clinical dysfunction, where conceptual overlap is easily mistaken for diagnostic equivalence. Cross-validated prediction, representative sampling, preregistered model selection, and cross-cultural invariance testing will be decisive Ashton & Lee (2025) Pletzer et al. (2019) Krueger et al. (2012).
Mechanistic theories face a distinct agenda: experiments that dissociate reward pursuit, fear, anxiety during goal conflict, habitual impulsivity, and socially regulated anger. Direct behavioral and physiological measurements can test whether broad trait scores relate to functional processes through stable and generalizable pathways. The results should inform revisions of all models, including the leading six-factor proposal. This open-ended approach treats personality science as cumulative empirical comparison anchored to specific constructs and observations Corr (2004) Carver & White (1994) Cloninger et al. (1993).
Conclusion
HEXACO and alternative personality trait models describe overlapping but differently organized aspects of stable variation among people. The six-domain HEXACO framework makes a substantive contribution by representing Honesty–Humility and reallocating some anger, fear, and affiliation content relative to customary Five-Factor dimensions. Its evidence includes cross-language lexical research, psychometric validation, observer reports, laboratory examinations of dishonest behavior, and synthesis of consequential criterion outcomes. The strongest conclusions concern measurable descriptive structure and associations with specific classes of behavior Lee & Ashton (2008) Lee & Ashton (2018) Hilbig & Zettler (2015) Pletzer et al. (2019).
PEN, reinforcement sensitivity theory, Cloninger's temperament–character model, Zuckerman's Alternative Five, and the interpersonal circumplex retain independent value because they select different explanatory targets: broad biological hypotheses, reward and defensive processes, the formation of character, stimulation and impulsivity, and the relational geometry of behavior. They can be evaluated together through clearly defined measurement and prediction studies while respecting their conceptual differences. Their biological and clinical interpretations require evidence beyond questionnaire factor recovery Eysenck (1990) Corr (2004) Cloninger et al. (1993) Zuckerman et al. (1993) Wiggins (1979).
The cumulative finding is methodological as well as substantive: a model is scientifically useful when its constructs are measured reliably, its structure replicates, its predictions generalize, and its limits remain explicit. The research agenda demands equitable sampling across languages and populations, credible independent criteria, longitudinal evidence, and direct tests of motivational mechanisms. Future human–artificial relations research may extend the domain of inquiry, provided that each cross-system comparison defines and validates its own constructs. A single authorized canonical Research Dossier can integrate this evidence without replacing the separate Science and Publication studies Cronbach & Meehl (1955) Ashton & Lee (2025).
References
- Ashton, M. C., & Lee, K. (2005). Honesty-Humility, the Big Five, and the Five-Factor Model. Journal of Personality, 73(5), 1321–1353. https://doi.org/10.1111/j.1467-6494.2005.00351.x
- Ashton, M. C., & Lee, K. (2007). Empirical, theoretical, and practical advantages of the HEXACO model of personality structure. Personality and Social Psychology Review, 11(2), 150–166. https://doi.org/10.1177/1088868306294907
- Ashton, M. C., & Lee, K. (2008). The prediction of Honesty–Humility-related criteria by the HEXACO and Five-Factor Models of personality. Journal of Research in Personality, 42(5), 1216–1228. https://doi.org/10.1016/j.jrp.2008.03.006
- Ashton, M. C., & Lee, K. (2009). The HEXACO-60: A short measure of the major dimensions of personality. Journal of Personality Assessment, 91(4), 340–345. https://doi.org/10.1080/00223890902935878
- Ashton, M. C., & Lee, K. (2025). Self- and observer reports of personality. Annual Review of Psychology, 76, 771–795. https://doi.org/10.1146/annurev-psych-020124-115044
- Ashton, M. C., Lee, K., & de Vries, R. E. (2014). The HEXACO Honesty-Humility, Agreeableness, and Emotionality factors: A review of research and theory. Personality and Social Psychology Review, 18(2), 139–152. https://doi.org/10.1177/1088868314523838
- Carver, C. S., & White, T. L. (1994). Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: The BIS/BAS Scales. Journal of Personality and Social Psychology, 67(2), 319–333. https://doi.org/10.1037/0022-3514.67.2.319
- Cloninger, C. R., Svrakic, D. M., & Przybeck, T. R. (1993). A psychobiological model of temperament and character. Archives of General Psychiatry, 50(12), 975–990. https://doi.org/10.1001/archpsyc.1993.01820240059008
- Corr, P. J. (2004). Reinforcement sensitivity theory and personality. Neuroscience & Biobehavioral Reviews, 28(3), 317–332. https://doi.org/10.1016/j.neubiorev.2004.01.005
- Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52(4), 281–302. https://doi.org/10.1037/h0040957
- Eysenck, H. J. (1990). Genetic and environmental contributions to individual differences: The three major dimensions of personality. Journal of Personality, 58(1), 245–261. https://doi.org/10.1111/j.1467-6494.1990.tb00915.x
- Gurven, M., von Rueden, C., Massenkoff, M., Kaplan, H., & Lero Vie, M. (2013). How universal is the Big Five? Testing the five-factor model of personality variation among forager-farmers in the Bolivian Amazon. Journal of Personality and Social Psychology, 104(2), 354–370. https://doi.org/10.1037/a0030841
- Hilbig, B. E., & Zettler, I. (2015). When the cat's away, some mice will play: A basic trait account of dishonest behavior. Journal of Research in Personality, 57, 72–88. https://doi.org/10.1016/j.jrp.2015.04.003
- Ion, A., Iliescu, D., Aldhafri, S., Rana, N., Ratanadilok, K., Widyanti, A., & Nedelcea, C. (2017). A cross-cultural analysis of personality structure through the lens of the HEXACO model. Journal of Personality Assessment, 99(1), 25–34. https://doi.org/10.1080/00223891.2016.1187155
- Krueger, R. F., Derringer, J., Markon, K. E., Watson, D., & Skodol, A. E. (2012). Initial construction of a maladaptive personality trait model and inventory for DSM-5. Psychological Medicine, 42(9), 1879–1890. https://doi.org/10.1017/S0033291711002674
- Lee, K., & Ashton, M. C. (2004). Psychometric properties of the HEXACO Personality Inventory. Multivariate Behavioral Research, 39(2), 329–358. https://doi.org/10.1207/s15327906mbr3902_8
- Lee, K., & Ashton, M. C. (2005). Psychopathy, Machiavellianism, and narcissism in the Five-Factor Model and the HEXACO model of personality structure. Personality and Individual Differences, 38(7), 1571–1582. https://doi.org/10.1016/j.paid.2004.09.016
- Lee, K., & Ashton, M. C. (2006). Further assessment of the HEXACO Personality Inventory: Two new facet scales and an observer report form. Psychological Assessment, 18(2), 182–191. https://doi.org/10.1037/1040-3590.18.2.182
- Lee, K., & Ashton, M. C. (2008). The HEXACO personality factors in the indigenous personality lexicons of English and 11 other languages. Journal of Personality, 76(5), 1001–1054. https://doi.org/10.1111/j.1467-6494.2008.00512.x
- Lee, K., & Ashton, M. C. (2018). Psychometric properties of the HEXACO-100. Assessment, 25(5), 543–556. https://doi.org/10.1177/1073191116659134
- Markey, P. M., & Markey, C. N. (2009). A brief assessment of the interpersonal circumplex: The IPIP-IPC. Assessment, 16(4), 352–361. https://doi.org/10.1177/1073191109340382
- McCrae, R. R., & John, O. P. (1992). An introduction to the five-factor model and its applications. Journal of Personality, 60(2), 175–215. https://doi.org/10.1111/j.1467-6494.1992.tb00970.x
- Moshagen, M., Hilbig, B. E., & Zettler, I. (2018). The dark core of personality. Psychological Review, 125(5), 656–688. https://doi.org/10.1037/rev0000111
- Pletzer, J. L., Bentvelzen, M., Oostrom, J. K., & de Vries, R. E. (2019). A meta-analysis of the relations between personality and workplace deviance: Big Five versus HEXACO. Journal of Vocational Behavior, 112, 369–383. https://doi.org/10.1016/j.jvb.2019.04.004
- Thielmann, I., Spadaro, G., & Balliet, D. (2020). Personality and prosocial behavior: A theoretical framework and meta-analysis. Psychological Bulletin, 146(1), 30–90. https://doi.org/10.1037/bul0000217
- Torrubia, R., Ávila, C., Moltó, J., & Caseras, X. (2001). The Sensitivity to Punishment and Sensitivity to Reward Questionnaire (SPSRQ) as a measure of Gray's anxiety and impulsivity dimensions. Personality and Individual Differences, 31(6), 837–862. https://doi.org/10.1016/S0191-8869(00)00183-5
- Wiggins, J. S. (1979). A psychological taxonomy of trait-descriptive terms: The interpersonal domain. Journal of Personality and Social Psychology, 37(3), 395–412. https://doi.org/10.1037/0022-3514.37.3.395
- Zuckerman, M., Kuhlman, D. M., Joireman, J., Teta, P., & Kraft, M. (1993). A comparison of three structural models for personality: The Big Three, the Big Five, and the Alternative Five. Journal of Personality and Social Psychology, 65(4), 757–768. https://doi.org/10.1037/0022-3514.65.4.757