Personality traits and the HEXACO model

Personality refers to relatively stable individual differences in patterns of thinking, feeling, and behaving that influence how people respond to their social world (Roberts, Reference Roberts2009). These traits are shaped by both biological and environmental factors and tend to remain stable throughout the life course (e.g., Briley and Tucker-Drob, Reference Briley and Tucker-Drob2014). Personality psychologists have developed various frameworks for assessing individual differences in personality, ranging from models that focus on specific traits – such as the Dark Triad (e.g., narcissism) discussed below – to more comprehensive models capturing the full spectrum of individual differences. Among the latter, the Big Five is the most widely known model, proposing that human personality varies across five core dimensions: openness to experience, conscientiousness, extraversion, agreeableness, and neuroticism (Costa Jr and McCrae, Reference Costa and McCrae1992). Over the past two decades, the HEXACO model (Ashton and Lee, Reference Ashton and Lee2007) has gained prominence. It posits that the main dimensions on which people differ are honesty-humility (H), emotionality (E), extraversion (X), agreeableness (A) conscientiousness (C) and openness to experience (O). Like the Big Five, HEXACO is derived from lexical analysis and includes the same five traits, though its definitions of agreeableness and emotionality slightly differ from Big Five agreeableness and neuroticism.Footnote ¹ Crucially, however, it adds a sixth dimension – honesty-humility (Lee and Ashton, Reference Lee and Ashton2014) – which makes it particularly relevant for examining virtue-related traits of (aspiring) political candidates.

More broadly, our study contributes to the political psychology literature that has successfully used personality traits to predict key political outcomes, including ideological preferences (e.g., Gerber et al., Reference Gerber2010), voting behavior (e.g., Bakker, Rooduijn, and Schumacher, Reference Bakker, Rooduijn and Schumacher2016), and political participation (e.g., Mondak et al., Reference Mondak2010). Particularly relevant to our focus is the growing body of research showing that elected officials systematically differ from the general population in terms of personality (e.g., Schumacher and Zettler, Reference Schumacher and Zettler2019; Weinberg, Reference Weinberg2021). This raises important questions about a potential personality-based selection funnel into political office.

Honesty-humility and the dark triad

Much of the existing literature on nascent political ambition that we further review below centers on the Dark Triad traits – narcissism, Machiavellianism, and psychopathy (e.g., Blais, Pruysers, and Chen, Reference Blais, Pruysers and Chen2019; Peterson and Palmer, Reference Peterson and Palmer2022). Meta-analytic evidence shows that the shared core of these traits overlaps substantially with the low pole of honesty-humility in the HEXACO model used in our study (Hodson et al., Reference Hodson2018).

While we acknowledge this conceptual proximity, focusing on honesty-humility and the HEXACO model provides several advantages. First, honesty-humility offers a more parsimonious and theoretically coherent framework. It is a single personality trait, consisting of four underlying facets: sincerity, fairness, modesty, and greed avoidance. The Dark Triad, in turn, consists of three distinct constructs (Ashton and Lee, Reference Ashton and Lee2007). Second, unlike the Dark Triad, honesty-humility offers a bipolar structure, capturing not just socially aversive traits but also positive dispositions that may be equally relevant in understanding who chooses to enter politics. Third, compared to the Dark Triad, honesty-humility is part of a larger structural model capturing the entire personality domain (Hodson et al., Reference Hodson2018, 127). This will allow us to assess the relevance of honesty-humility as compared to other dimensions of human personality like extraversion. Importantly, HEXACO’s dimensional structure has been validated across many cultures (e.g., Ashton et al., Reference Ashton2004).

Extant research on personality and ambition

Research on personality and political ambition can be grouped into two main strands. The first relies on comprehensive personality frameworks to explain political ambition domains (e.g., Allen and Cutts, Reference Allen and Cutts2018; Dynes, Hassell, and Miles, Reference Dynes, Hassell and Miles2019; Scott and Medeiros, Reference Scott and Medeiros2020). Yet, with one notable exception (Blais, Pruysers, and Chen Reference Blais, Pruysers and Chen2019), these studies focus on the Big Five model, which does not contain honesty-humility.

The second strand does focus on specific virtue-related personality traits, mostly the Dark Triad, showing that these traits are associated with higher nascent political ambition in both student and adult populations in the U.S. and Canada (Blais and Pruysers Reference Blais and Pruysers2017; Peterson and Palmer Reference Peterson and Palmer2022). The positive relationship between Machiavellianism and narcissism and nascent ambition finds widespread empirical support (Blais and Pruysers, Reference Blais and Pruysers2017; Peterson and Palmer, Reference Peterson and Palmer2021, Reference Peterson and Palmer2022). Some studies also report that psychopathy correlates with nascent ambition (Blais, Pruysers, and Chen, Reference Blais, Pruysers and Chen2019; Peterson and Palmer, Reference Peterson and Palmer2021). Traits such as empathy have also been found to correlate positively with political ambition (Clifford, Kirkland, and Simas Reference Clifford, Kirkland and Simas2019). Yet, these studies focus on traits in isolation, thereby not capturing the entire human personality domain. This raises the question of how important virtue-related traits are – in this research captured by honesty-humility – compared to traits less clearly related to virtue. To our best knowledge, only one prior study has delved into this, revealing a negative correlation between honesty-humility and nascent ambition within a sample of Canadian citizens (Blais, Pruysers, and Chen Reference Blais, Pruysers and Chen2019). We examine whether this result extends to different countries with different political systems.

Finally, the psychological antecedents of actual candidacy remain understudied. Dispositional empathy and narcissism are positively related to running for office among US citizens (Clifford, Kirkland, and Simas Reference Clifford, Kirkland and Simas2019; Sendinc and Hatemi Reference Sendinc and Hatemi2023). However, can these results be generalized outside of the US? Also, these studies focus only on these specific traits, which raises the question of how important virtue versus non-virtue-related traits are in the recruitment funnel. Schumacher and Zettler (Reference Schumacher and Zettler2019) – so far the only candidate recruitment study using the HEXACO model – found that candidates for the Danish parliament self-report higher honesty-humility than the general population. It, however, remains unclear whether this pattern reflects self-selection (nascent ambition) or recruitment. We endeavor to uncover these different aspects of the recruitment process, relying on the comprehensive HEXACO personality model.

Theoretical framework: situation-trait-activation

Much of the research on personality and political ambition remains inductive (e.g., Dietrich et al., Reference Dietrich2012; Blais and Pruysers, Reference Blais and Pruysers2017; Scott and Medeiros, Reference Scott and Medeiros2020) or loosely links trait descriptions to the demands of political work. Allen and Cuts (Reference Allen and Cutts2018, p. 78), for instance, argue that ‘people high in extraversion [.] are outgoing and enjoy social interaction and [.] therefore […] the plethora of interpersonal encounters involved in running for political office will not deter them’. This reasoning implicitly links personality traits to characteristics of the political environment, but without a formal theoretical model.

To fill this void, we rely on perspectives proposing that both personality and the environment affect political behavior (Mondak et al., Reference Mondak2010; Arceneaux et al., Reference Arceneaux2025). Specifically, we utilize the situation-trait-activation theory developed to understand the rise of leaders within organizations (De Vries, Reference De Vries2018), a theory that applies, but is certainly not exclusive, to political organizations such as parties. This theory consists of two complementary mechanisms: situation-activation and trait-activation. Situation-activation refers to individuals being drawn to environments they perceive as compatible with their personality traits. In our context, this helps explain nascent ambition – the initial interest in political office. Trait-activation, in turn, refers to how specific situations or environments provide cues as to which latent personality traits an individual should express. Trait-activation also implies that political parties (and other organizations responsible for recruiting candidates) may select or recruit individuals who possess traits aligned with what they view as necessary for political success. Thus, our framework captures two stages of the recruitment funnel: individuals with certain traits self-select into politics (situation-activation), but whether they are recruited also depends on the traits deemed relevant by recruiters (trait-activation).

Our framework also aligns with the functional matching paradigm, which posits that individuals are more likely to be selected – and to succeed – when their traits match the perceived demands of the role (Tett and Burnett, Reference Tett and Burnett2003).

Hypotheses

Below, we link situation-activation to nascent ambition and trait-activation to candidate recruitment. For recruitment, we argue that a candidate’s honesty-humility poses an intuitive dilemma to recruiters, resulting in two competing sets of hypotheses about being asked to and actually running.

We begin with perhaps the more straightforward aspect of the link between honesty-humility and office holding, which is its relation to nascent ambition. Politics is often characterized as a domain of competition, conflict, and manipulation (Peterson and Palmer Reference Peterson and Palmer2022). From a situation-activation perspective, such an environment may particularly attract those low on honesty-humility. To recap, this trait comprises four facets: sincerity, fairness, greed avoidance, and modesty (Ashton and Lee, Reference Ashton and Lee2007). First, individuals low in sincerity are more comfortable with manipulation and self-preservation strategies (Ashton, Lee, and De Vries, Reference Ashton, Lee and De Vries2014). This could make activities like campaigning more attractive, where impression management is often key. Second, low fairness reflects a willingness to exploit others or society at large (ibid.). This may also involve grey areas of fairness, like selectively revealing information, bending procedural roles, or supporting allies over more qualified candidates. Such individuals may be drawn to politics precisely because political settings often contain such opportunities. Third, low greed avoidance aligns with the pursuit of power, status, and material gain (ibid.). Politics offers prominent rewards – power, visibility, and sometimes financial benefits – that may be particularly attractive to such individuals. Fourth, low modesty marks an inflated self-confidence and sense of entitlement (ibid.). Since nascent ambition often includes self-assessed qualifications and a belief in one’s potential to hold office (Fox and Lawless, Reference Fox and Lawless2005), low modesty may directly reinforce the perception that one is suited to political leadership. Last, the different facets also work in concert, where, for instance, manipulation and exploitation (low sincerity and fairness) are justified by an individual’s feelings of entitlement (low modesty) (Ashton and Lee, Reference Ashton and Lee2007).

Based on situation-activation, we may thus expect that individuals low on honesty-humility will be more attracted to a political career, implying higher scores on the two nascent ambition dimensions, i.e., having thought about running and deeming themselves qualified to run (see Fox and Lawless, Reference Fox and Lawless2005). This negative relationship between honesty-humility and nascent ambition is expressed in H1.1.

Turning to candidate recruitment, our data cannot distinguish between three distinct mechanisms: recruiter intent, individuals’ efforts at self-promotion, and – in case of candidacy – the propensity to accept invitations. Still, we interpret the pipeline to political recruitment as shaped by all three. Hence, certain personalities may be more likely to self-promote and accept invitations, but recruiters may (consciously or not) also have their own preferences for particular personalities.

Regarding the role of recruiters, we shift from situation-activation to trait-activation. As said, political recruiters may evaluate whether individuals possess the personality traits needed to successfully run for office (Tett and Burnett Reference Tett and Burnett2003; De Vries Reference De Vries2018). However, how being asked and actually running relates to honesty-humility is still poorly understood. Hence, we formulate two competing hypotheses: H.1.2 and H.1.3. High honesty-humility can be considered a character strength related to honesty, altruism, and justice (Lenzi, Reference Lenzi2023). Hence, those excelling in civic virtue could be perceived as better suited to advance the common good. Consequently, they may be asked the most to run for office by political parties or other organizations responsible for recruiting candidates.

Turning to parties as gatekeepers more specifically, given the correlation between low honesty-humility and unethical behavior (Ashton and Lee, Reference Ashton and Lee2008; De Vries et al., Reference De Vries2017), they have good reasons to avoid recruiting those low on this trait by not asking them to run and not selecting them to run if they step forward. Voters prefer politicians higher on honesty-humility than themselves (Aichholzer and Willmann, Reference Aichholzer and Willmann2020). Other research has also shown that voters generally find candidates less palatable if they score higher on the Dark Core, reflecting a candidate’s average score across the Dark Triad traits (Nai and Maier, Reference Nai and Maier2023). Moreover, the related concept of integrity was found to be an important marker of candidate quality (Mondak, Reference Mondak1995; Franchino and Zucchini, Reference Franchino and Zucchini2015). Awareness of this – whether implicit or explicit – should incentivize parties to seek candidates high on honesty-humility. They could use signaling (e.g., emphasize the societal relevance of political work), screening (e.g., integrity tests, recruiting from their rank and file), and commitment devices (e.g., loyalty pledges) to select such individuals (Ceva and Ferretti, Reference Ceva and Ferretti2021). Hence, even though those lower on honesty-humility may have more ambition (H1.1), based on this reasoning, people higher on honesty-humility might eventually be asked and selected to run for office.

All of the above notwithstanding, people low on honesty-humility might still be seen as stereotypically more fit for office. First, low honesty-humility is associated with facets such as low sincerity and low modesty, which could predispose candidate recruiters observing such individuals to assume – and voice – that they will do well in political life: they may be perceived as better able to deal with the conflict and manipulation inherent to politics and to shine in campaigns (also see Peterson and Palmer, Reference Peterson and Palmer2021). Second, low humility – and especially low levels of modesty, one of its constituent facets – mirror self-confidence, and, similar to the argument outlined in Fox and Lawless (Reference Fox and Lawless2014), can also reinforce validation of one’s leadership qualities and precipitate encouragement to run by others. Third, despite the socially undesirable qualities associated with low honesty-humility, those involved in candidate recruitment might nevertheless see such traits as indicative of potentially effective (albeit lacking in scruples) future politicians, who might be more likely to serve the party well. This implies that trait-activation cues may not always function in a normatively desirable way: environments may reinforce or reward traits like low honesty-humility. Suggestive evidence on this is found in the literature documenting how men are far more frequently encouraged to run for office by party operatives compared to equally qualified women, often owing to reliance on gendered stereotypes of fitness for office that map directly onto those encompassing honesty-humility traits (Niven, Reference Niven1998; Crowder-Meyer, Reference Crowder-Meyer2013).

Finally, specifically relating to actually running for office, those low on honesty-humility may more favorably respond to invitations to run for office, either because they are more likely to espouse narcissistic tendencies (correlated with low honesty-humility) or because they are attracted to environments that enable or even encourage manipulative and dishonest conduct (also see Preece, Stoddard, and Fisher, Reference Preece, Stoddard and Fisher2016, for related types of differential responsiveness to recruitment messages; Pruysers and Blais, Reference Pruysers and Blais2017, Reference Pruysers and Blais2019). This alternative line of reasoning is supported by a study from the US reporting that individuals who self-report higher levels of grandiose narcissism – a trait partly correlated to honesty-humility – are more likely to run for office (Sendinc and Hatemi, Reference Sendinc and Hatemi2023). Hence, our competing hypothesis (H1.3) is as follows:

These hypotheses form our study’s core. However, since our instrument measures additional traits as part of a conceptual whole, we also formulate hypotheses for three other personality traits where prior research supports directional expectations.

Starting with emotionality, which captures fearfulness and anxiety, the logic of situation-activation suggests that individuals high on this trait will find politics – an unpredictable and high-pressure environment (Weinberg, Reference Weinberg and Weinberg2012) – unappealing. Extant research confirms a negative correlation between emotionality and nascent ambition (Blais, Pruysers, and Chen, Reference Blais, Pruysers and Chen2019). Research on neuroticism, a closely related Big Five trait, also generally shows a negative association with nascent ambition (Allen and Cutts, Reference Allen and Cutts2018; Scott and Medeiros, Reference Scott and Medeiros2020; but see Dynes, Hassell, and Miles, Reference Dynes, Hassell and Miles2019) and other forms of political participation (Bromme, Rothmund, and Azevedo, Reference Bromme, Rothmund and Azevedo2022).

As regards recruitment, emotionality has a negative premium among voters (Aichholzer and Willmann, Reference Aichholzer and Willmann2020), making it less likely that parties or peers will encourage them or select them to run. In terms of trait-activation, political environments do not seem to offer cues that support or reward emotionality. Prior research also reveals that these people are less likely to respond to recruitment appeals (Pruysers and Blais, Reference Pruysers and Blais2019).

Turning to extraversion, activities like campaigning and interacting with voters, mass media, and other public officials are more likely to attract the energetic and socially bold (situation-activation). This is echoed by empirical evidence of extraversion positively correlating with nascent ambition (Blais and Pruysers, Reference Blais and Pruysers2017; Allen and Cutts, Reference Allen and Cutts2018; Blais, Pruysers, and Chen, Reference Blais, Pruysers and Chen2019; Dynes, Hassell, and Miles, Reference Dynes, Hassell and Miles2019; Scott and Medeiros, Reference Scott and Medeiros2020) and other types of political participation (Bromme, Rothmund, and Azevedo, Reference Bromme, Rothmund and Azevedo2022). Voters also prefer extraverted candidates (Aichholzer and Willmann, Reference Aichholzer and Willmann2020), suggesting that recruiters may hold a preference for extraverted people (trait-activation).

Last, people high on openness tend to seek novelty, intellectual stimulation, and new experiences (Ashton, Lee, and De Vries, Reference Ashton, Lee and De Vries2014). From a situation-activation perspective, politics – characterized by complexity and variety – may be particularly attractive to them. Empirical studies generally support a positive relationship between openness and nascent ambition (also see Allen and Cutts Reference Allen and Cutts2018; Blais and Pruysers Reference Blais and Pruysers2017; Scott and Medeiros Reference Scott and Medeiros2020; Dynes, Hassell, and Miles Reference Dynes, Hassell and Miles2019; but see Blais, Pruysers, and Chen Reference Blais, Pruysers and Chen2019) or engagement in other types of political activities (Bromme, Rothmund, and Azevedo, Reference Bromme, Rothmund and Azevedo2022). Regarding recruitment, voters tend to prefer politicians who are higher than themselves on openness (Aichholzer and Willmann, Reference Aichholzer and Willmann2020). Political environments may therefore provide trait-relevant cues that both invite and reward the expression of openness. Consequently, those high on this trait may also be perceived as suited for it and thus more likely to be encouraged or recruited (trait-activation).

Regarding the remaining traits, agreeableness and conscientiousness, the link with our dependent variables is less clear-cut. Agreeable people – gentle, comforting, and sympathetic – may be deterred by the incivility and conflict in politics (Allen and Cutts, Reference Allen and Cutts2018; Dynes, Hassell, and Miles, Reference Dynes, Hassell and Miles2019). Some studies indeed support that agreeableness negatively correlates with nascent ambition (Allen and Cutts, Reference Allen and Cutts2018; Dynes, Hassell, and Miles, Reference Dynes, Hassell and Miles2019; but see Blais, Pruysers, and Chen, Reference Blais, Pruysers and Chen2019; Scott and Medeiros, Reference Scott and Medeiros2020). Conversely, political roles also require collaboration and compromise (Flick Witzig and Vatter, Reference Flick Witzig and Vatter2024). Agreeable people might find this more enjoyable and be deemed better up to this task. These counteracting logics could explain why several studies report null findings regarding the association between agreeableness and nascent ambition (Blais and Pruysers, Reference Blais and Pruysers2017; Blais, Pruysers, and Chen, Reference Blais, Pruysers and Chen2019).

Formulating unequivocal expectations on conscientiousness, nascent ambition, and political recruitment is equally difficult. Conscientious individuals tend to have a strong sense of purpose, discipline, and long-term planning. These qualities may align with certain features of political work, such as the pursuit of policy goals. Simultaneously, however, conscientious people exhibit high risk aversion. Given the unpredictability of political careers, public scrutiny, and the likelihood of electoral failure, conscientious individuals may perceive politics as an excessively risky environment. These contrasting expectations are mirrored in empirical findings: some studies report a negative correlation with political ambition (Allen and Cutts, Reference Allen and Cutts2018; Dynes, Hassell, and Miles, Reference Dynes, Hassell and Miles2019), others find the opposite (Blais, Pruysers, and Chen, Reference Blais, Pruysers and Chen2019), or no significant relationship at all (Blais and Pruysers, Reference Blais and Pruysers2017; Scott and Medeiros, Reference Scott and Medeiros2020).

Given these mixed findings, we do not advance hypotheses for either agreeableness or conscientiousness. The data will reveal how these traits relate to nascent ambition and recruitment.

Data

To test our hypotheses, we conducted online surveys between February and May 2022 as part of the POLPOP project.Footnote ² Within this larger project, five countries – Canada (N = 2312), Denmark (N = 2311), Israel (N = 2745), the Netherlands (N = 1942), and Switzerland (N = 2300) – fielded personality questions. These countries differ in their type of democracy (majoritarian versus proportional representation systems), emphasis placed on individual candidates (district competition versus party list), and degree of inter-party competition (Lijphart, Reference Lijphart1985). This variation may shape the types of personalities politics attracts, consistent with situation-trait-activation theory. For example, recent research suggests that Switzerland’s consensual democracy values agreeableness (Flick Witzig and Vatter, Reference Flick Witzig and Vatter2024). While the number of countries is too small to formally test hypotheses on institutional effects, the aforementioned study does suggest that testing our hypotheses across different national contexts is needed to explore potential heterogeneity in personality effects (also see Weinschenk, Reference Weinschenk2017). Simultaneously, some traits – especially honesty-humility – may operate similarly across systems, given the competitive and strategic nature of politics in general.

Survey respondents were recruited through the sample provider Dynata, using online non-probability samples. This approach offers practical advantages: cost-efficiency, timely cross-national fielding, and respondent diversity. Prior research suggests that quota-based non-probability samples can yield results comparable to probability-based samples in political behavior research (Ansolabehere and Schaffner, Reference Ansolabehere and Schaffner2014). Quotas were set for gender, age, and education. Since some groups (i.e., those with lower educational attainment and older individuals) were still underrepresented in our sample, throughout we apply probability weights on age, gender, education, and party choice in the latest elections to accurately approximate the electorates in the countries studied. As further detailed in the online Supporting Information (SI), due to missing data on individual personality items (see Section 3) and other independent variables like income (Section 2.2), the observation numbers in our core models (see notes in Figure 2) are somewhat lower than the original N by country reported at the top of this section. In the SI’s Section 3, we also provide other relevant details on the sample (response rates, fraudulent responses, etcetera).

Figure 1.

Distribution of HEXACO personality traits across countries.

Notes: The red dots below the distributions denote the average. NL = Netherlands, IL = Israel, DK = Denmark, CH = Switzerland, CA = Canada.

Figure 2.

Evaluating the effects from the logistic regressions.

Notes: Values represent the change in the predicted probability of the outcome when a continuous independent variable increases along its interquartile range. For nominal variables (i.e., gender and foreign-born), we depict the difference in predicted probability when comparing the focal category against the reference category. Country fixed effects were included. 90% (inside brackets) and 95% (outside brackets) CI. Estimates based on the logistic regressions displayed in Table A4 of the SI. N is 7958, 7959, 7958, and 8483.

Variables

Our first dependent variable captures whether someone ever thought about running for political office (1. Never, 2. It crossed my mind, 3. Many times). Given the small number in the top category (225 cases, 2.25 percent), like the state-of-the-art (e.g., Blais and Pruysers, Reference Blais and Pruysers2017; Blais, Pruysers, and Chen, Reference Blais, Pruysers and Chen2019), we recoded this into a binary variable (1. Many times, or it crossed my mind; 0. Never). Our second dependent variable captures whether respondents deem themselves qualified to run. Originally measured on a four-point scale (1. Very qualified, 2. Qualified, 3. Somewhat qualified, 4. Not at all qualified), following extant research (e.g., Blais, Pruysers, and Chen, Reference Blais, Pruysers and Chen2019), we recoded it into a binary variable: those that deem themselves at least somewhat qualified (categories 1–3) versus those who do not (category 4). These two dependent variables reflect the two dimensions of nascent ambition (Fox and Lawless, Reference Fox and Lawless2005; Blais, Pruysers, and Chen, Reference Blais, Pruysers and Chen2019). Although binary coding is standard, we confirm in the SI (Section 2.1) that results are robust using the original scales.

Our remaining two dependent variables relate to political recruitment: whether respondents have ever been asked to run (1 = Yes, 0 = No) and actually ran for office (1 = Yes, 0 = No). With 599 and 773 positive cases, respectively, we meet the threshold of ten positive events per predictor (e.g., Vittinghoff and McCulloch, Reference Vittinghoff and McCulloch2007).

Our main independent variables constitute the HEXACO personality traits. To preserve survey space, we had to rely on the Brief HEXACO Inventory (BHI), consisting of 24 items: 4 per HEXACO domain (De Vries, Reference De Vries2013; also see Schumacher and Zettler, Reference Schumacher and Zettler2019). Though it takes only 2–3 minutes to complete, its alpha reliability is relatively low, as with most short personality measures. In Section 1.2 of the SI, we indeed report lower alpha reliabilities than Blais and Pruysers (Reference Blais and Pruysers2017), who relied on the 192-item HEXACO-PI in their nascent ambition study. That said, the validity loss is only modest because of its relatively high test-retest reliability, self-other agreement, and convergent validity with full-length measures (De Vries, Reference De Vries2013). Respondents had to indicate on a scale ranging from 1 (strongly disagree) to 7 (strongly agree) whether they agreed with a particular item. We reversed the scales for statements where agreement entails lower scores on a HEXACO trait. In the SI (Section 1.2), we list the 24 items. Confirmatory factor analysis reveals all items to have a positive factor loading. We calculated the respondents’ mean score over the four items belonging to each HEXACO domain.Footnote ³

Figure 1 displays the distribution of the HEXACO traits across countries. As should be, all traits are roughly normally distributed (Ashton and Lee, Reference Ashton and Lee2009). The slight left skew in honesty-humility may indicate some social desirability bias (Dunlop et al., Reference Dunlop2020). Nonetheless, the mean is only slightly higher than that of the arguably less virtue-related trait of extraversion (5.1 vs. 5), and the distribution remains approximately normal. We therefore find no evidence of serious bias – consistent with previous research showing that in an anonymous research context like ours, respondents provide trustful answers to honesty-humility inventories (Zettler et al., Reference Zettler2015).

We also include controls. First, given the observed gender gap in politics, where women are less likely to have nascent ambition (Fox and Lawless, Reference Fox and Lawless2005), less likely to be recruited to run for office (Fox and Lawless, Reference Fox and Lawless2010), and less likely to positively respond to any recruitment efforts (Preece and Stoddard, Reference Preece and Stoddard2015), we control for gender (female and non-binary versus male). Second, we control for age (e.g., Peterson and Palmer, Reference Peterson and Palmer2022). To gauge possible non-linear effects, we also include its square. Third, we consider whether the respondent was born in the country sampled (1. Yes, 0. No). Fourth, we control for income (recoded to ten deciles) (e.g., Fox and Lawless, Reference Fox and Lawless2005; Blais, Pruysers, and Chen, Reference Blais, Pruysers and Chen2019; Peterson and Palmer, Reference Peterson and Palmer2022). Fifth, we consider education (1. Primary, 2. Secondary, 3. Higher non-university, 4. University), as it is related to political participation (e.g., Persson, Reference Persson2015; Blais, Pruysers, and Chen, Reference Blais, Pruysers and Chen2019). Sixth, we control for political interest (ranging from 1. Not at all interested to 5. Very much interested) (e.g., Fox and Lawless, Reference Fox and Lawless2005; Blais, Pruysers, and Chen, Reference Blais, Pruysers and Chen2019). Seventh, we include the respondents’ ideological extremity (Peterson and Palmer, Reference Peterson and Palmer2022). Here, we subtracted the respondents’ self-placements (1. Left, 10. Right) from the scale’s midpoint. We adopted these controls based on nascent ambition studies. How they influence our other dependent variables remains an open question.

Estimation

All dependent variables are binary, so we specify logistic regressions. To account for country differences in political ambition and recruitment, we include country fixed effects. We also cluster the standard errors at the level of countries, as respondents from the same country might not be considered independent observations.