Although their numbers have been slowly rising, women continue to be underrepresented in leadership roles throughout the world (e.g., Catalyst, 2022; World Economic Forum, 2020). For example, only 30% of board directors and 6% of Chief Executive Officers at the top U.S. companies are women (Catalyst, 2022). Women also remain underrepresented in political leadership, with the global average share of women in ministerial positions in 2022 at 16.1% (World Economic Forum, 2020). A confluence of factors may explain these gender gaps, including discrimination (Eagly & Heilman, 2016), structural inequalities (England et al., 2020), and gender differences in leadership ambition (Netchaeva et al., 2022). This latter factor, which has received relatively less attention than the others, is the focus of the current investigation. We adopt a developmental perspective on this phenomenon, seeking to examine the origins of gender differences in leadership aspirations. Whereas considerable attention has been devoted to understanding how adults think about leadership—including, in the case of women, the perceived social costs of behaving in leader-like ways—the developmental origins of these attitudes remain underexamined (e.g., Heck et al., 2021). In three studies, we investigate young children’s interest in leader roles and their beliefs about the social rewards and costs of leadership, with particular attention to potential gender differences.

Gender Gaps in Leadership Ambition

Overall, adult women appear less interested than men in a wide range of leadership roles (Netchaeva et al., 2022). For example, Sheppard (2018) found that undergraduate women were less interested than men in elite leadership positions and had fewer positive associations with those roles. Similarly, Schneider et al. (2016) found that women expressed less interest in running for political office (see also Pate & Fox, 2018). More generally, research has documented gender gaps in the desirability of social power—the authority to control and allocate resources, make decisions, and influence one’s own and others’ outcomes—which is inherent in most leadership roles (Magee & Frasier, 2014). For instance, women find power less appealing than men and are less motivated to obtain it (Schuh et al., 2014).

Two related processes have been proposed to jointly explain these gender gaps in interest in leadership. The first one involves gender differences in the anticipation (and the actual experience) of social support and cooperation from other people. According to lack of fit theory (Heilman, 1983, 2001) and role congruity theory (Eagly & Karau, 2002), the perceived lack of fit, or the incongruity, between the traditional female role (which prescribes communality—e.g., being sensitive and considerate to others’ needs) and leader-like attributes (which entail agency—e.g., being dominant and assertive) limits women’s opportunities. Specifically, women must work harder than men to be considered effective leaders and encounter disapproval when they enact dominant behaviors entailed in leader roles (Williams & Tiedens, 2016). Female leaders are often disliked by others (e.g., Heilman & Okimoto, 2007) and undermined by subordinates—especially men (Netchaeva et al., 2015; Vial et al., 2018). In the context of politics, Fox and Lawless (2010) found a remarkable lack of support for women to run for office (see also Bauer et al., 2022; Okimoto & Brescoll, 2010). The anticipation of a lack of support or disapproval from others can be a powerful disincentive (Rudman & Fairchild, 2004; Tomasello, 2014). A lack of social support is particularly discouraging for would-be leaders, who must garner support from others or risk having their authority questioned (e.g., Levi et al., 2009). Accordingly, women’s anticipation of resistance from others may dissuade them from behaving assertively (Moss-Racusin & Rudman, 2010), and may generally reduce their sense of self-efficacy as leaders and their interest in leadership (Fox & Lawless, 2011; Schneider et al., 2016; Sheppard, 2018). For example, Fisk and Overton (2019) found that women anticipated harsher sanctions than men for failing as leaders, which in turn predicted women’s lower interest in a leader role.

A second, related reason that may jointly explain adult gender gaps in leadership ambition is articulated by goal congruity theory, which posits that individuals are motivated to align their behavior with the demands of their social roles, including those dictated by their gender (Diekman & Eagly, 2008). People are drawn to gender-congruent roles, whereas gender-incongruent roles are disfavored, partly because fulfilling gender-role congruent goals can be psychologically rewarding by itself (Witt & Wood, 2010), and partly because of a desire to avoid the potential disapproval of engaging in gender-incongruent behavior (e.g., Tomasello, 2014). Thus, the interests and preferences of women and men tend to diverge in ways that are consistent with gender role-congruent goals (Brown & Diekman, 2010; Diekman et al., 2010). From a young age, girls more so than boys are socialized to endorse and exhibit communal attributes (Chaplin et al., 2005; Hibbard & Buhrmester, 1998), and adult women consistently report more communal, other-oriented goals than men (Hsu et al., 2021; Witt & Wood, 2010). In contrast, men tend to espouse more agentic, self-oriented goals (Evans & Diekman, 2009). Given that leadership is associated with traits such as dominance and power (Koenig et al., 2011), leader roles afford men the opportunity to fulfill their gender-stereotypic (agentic) goals. This is not the case for women; indeed, leadership may be perceived to clash with stereotypically-feminine (communal) goals. Even though scholars recognize the value of communality in leaders (Gerzema & D’Antonio, 2017), agency-related traits and behaviors continue to be seen as the hallmark of leadership, whereas communality is viewed as desirable, yet peripheral, in leaders (e.g., Koenig et al., 2011; Vial & Napier, 2018). Thus, the perceived incongruence between leader roles and the traditional female role may reduce women’s interest in leadership (Schneider et al., 2016).

In the current studies, we investigate the developmental roots of the well-documented gender gaps in leadership ambition (Netchaeva et al., 2022). Understanding how children think of leadership could provide important insights to address the origins of these gender gaps. Drawing from lack of fit and role congruity theories (Eagly & Karau, 2002; Heilman, 1983, 2001), we expect that children would anticipate negative reactions to girls who display a will to lead in interpersonal interactions. Moreover, drawing from goal congruity theory (Diekman & Eagly, 2008), given that girls are socialized into peer norms that tend to discourage leader-like behaviors (Hibbard & Buhrmester, 1998; Martin & Fabes, 2021; Sebanc et al., 2003), we expect that girls may anticipate negative reactions to anyone who displays this kind of behavior. The prospect of lack of support may in turn reduce girls’ interest in leadership. Many girls may instead develop other interests that are more congruent with the traditional female role—that is, roles that clearly afford communal goals. Over time, girls’ avoidance of leader-like roles and behaviors could result in a gender gap in leadership ambition among adults. We shed light on these possibilities by investigating children’s anticipation of social support for leaders, children’s own interest and self-efficacy in leader roles, and the influence of contextual cues (specifically, communal goal affordances and role models) on these cognitive and motivational variables. We also explore differences based on children’s racial/ethnic background. The potential utility of adopting a developmental lens to understand gender inequality in leadership has been recognized before (Caleo & Halim, 2021; Diekman et al., 2021; Heck et al., 2021; Martin & Fabes, 2021). Yet, little research has been devoted to understanding young children’s leadership cognitions with relation to gender.

Children’s Beliefs About Leadership

There is a growing literature devoted to understanding children’s generic concepts of leadership and related constructs such as power and status. This research has shown, for example, that infants expect leaders (but not non-leaders) to rectify a fairness transgression within the group (Stavans & Baillargeon, 2019). Children between 3 and 4 years of age associate having authority with dictating rules (Zhao & Kushnir, 2018), with non-verbal cues such as expansive body postures (Terrizzi et al., 2019), and with making and imposing decisions on others (Charafeddine et al., 2015). Children of this age already seem able to recognize not only “malevolent” power but also benevolent power (i.e., a more communal kind of authority) (Gülgöz & Gelman, 2017). Gülgöz and Gelman (2017) found that, by age 5, children conceived of social power in a similarly multifaceted way as adults, encompassing dimensions such as controlling resources, achieving goals, granting permission, setting norms, and giving orders. In sum, children’s concepts of power and leadership resemble those of adults from an early age.

Do Children See Leadership as Stereotypically Masculine?

Given that children endorse general gender stereotypes of agency and communality already by 5 years of age (Banse et al., 2010), it seems reasonable to expect that they would similarly endorse gender stereotypes of leadership. After all, children encounter stereotypes equating power with men in media and other cultural artifacts with high frequency (e.g., in animated children’s films; Aley & Hahn, 2020). And yet, whereas adults clearly view leadership as male-typed (e.g., Koenig et al., 2011; Vial & Napier, 2018), children’s beliefs are not so clear-cut, raising the possibility that leadership interest might be more similar among boys and girls than among adult men and women.

Some research suggests that even very young children expect leaders to possess stereotypically-masculine (rather than -feminine) attributes. For example, Terrizzi et al. (2019) found that 3-year-old children associated masculine facial features with interpersonal authority. Moreover, like adults, children as young as 5 and 6 years of age seem to associate men and boys (rather than women or girls) with high-status, high-power, or high-authority roles (e.g., Bigler et al., 2008; Charafeddine et al., 2020; Gülgöz, 2015; Liben et al., 2001; Mandalaywala et al., 2020; Neff et al., 2007). For example, Bos et al. (2022) found that 6- to 12-year-old children were more likely to draw a man (vs. a woman) when asked to draw a political leader, a tendency that grew among girls as they aged.

In contrast, other research suggests that children’s leadership cognitions are less decisively masculine, and may incorporate communal attributes to some extent. To illustrate, Patterson et al. (2019) found that children between the ages of 5 and 11 rated stereotypically-feminine qualities (e.g., “being gentle”) as more important for being president than stereotypically-masculine qualities (e.g., “being competitive”). Relatedly, Bos et al. (2022) found that younger children used stereotypically-masculine and -feminine traits with the same frequency to describe political leaders, whereas older children used stereotypically-masculine (vs. -feminine) traits more often. Researchers have also found that children can be sensitive to the communal obligations inherent in some leadership roles, expecting leaders to contribute more than non-leaders to joint projects (Stavans & Diesendruck, 2021) and to refrain from abusing their power over others (Reyes-Jaquez & Koenig, 2021).

Similarly, past investigations showed that children may associate gender with some dimensions of social power but not with others. Specifically, Gülgöz (2015) found that 3- to 9-year-old children associated boys with the authority to control resources and grant permissions, but they had no systematic gender associations when power was operationalized as the authority to give orders or set norms. It is also possible that children are more consistently biased toward the gender ingroup (rather than male targets) when attributing power or assigning high-power roles. For instance, Ayman-Nolley and Ayman (2005) observed own-gender biases among elementary school-aged children asked to draw a leader. Reyes-Jaquez and Koenig (2022) similarly found own-gender biases in 6- to 10-year-old children asked to select an adult leader.

These mixed findings indicate that children’s beliefs about gender, gender-stereotypical attributes (agency and communality), and leadership are less clear-cut than adults’ beliefs. Thus, it remains an empirical question whether the gender gaps in leadership interest that have been found in adults would also emerge among young children, or whether young girls may instead be as interested in leader roles as young boys.

Are Girls Less Interested in Leadership Than Boys?

A gender gap in interest in leadership (Netchaeva et al., 2022) seems to be already in place in adolescence. For example, Fox and Lawless (2014) identified large gender gaps in political leadership ambitions in a nationally representative sample of U.S. adolescents. However, evidence for gender gaps in leadership interest in younger children is more mixed: Whereas Bos et al. (2022) found that 6- to 12-year-old girls expressed lower levels of political ambition than boys, neither Bigler et al. (2008) nor Patterson et al. (2019) found gender differences in this age range in children’s interest in being president. Similarly, outside the context of politics, Reyes-Jaquez and Koenig (2022) found no gender gap in this age range in children’s interest in being in charge of controlling resources (i.e., allocating candy) in a group setting. Given the inconsistent findings about children’s beliefs about gender and leadership and about children’s own leadership aspirations, more research is necessary to understand whether young girls are less interested overall in leader roles compared to boys—and if so, why.

The Current Investigation

We conducted three studies to investigate children’s gendered beliefs about and interest in leadership. These studies build on each other to examine children’s (1) anticipation of social support for and cooperation with leaders (Studies 1–3), (2) their own interest in leadership roles (Studies 2 and 3), (3) their anticipated self-efficacy as leaders (Studies 2 and 3), and (4) whether these beliefs and motivations are sensitive to contextual features that mark the leadership role as more (or less) communal and gender neutral (Study 3). Additionally, we explored (5) differences across age and race/ethnicity in children’s leadership cognitions and motivation to lead (Studies 1–3). Although all three studies investigate gender differences in beliefs about and interest in leadership, Study 1 focused on children’s perceptions of other leaders, whereas Studies 2 and 3 focused on children’s perceptions of themselves as leaders. Each study adopts a different vantage point on the phenomenon of interest, providing a nuanced portrait of children’s gendered leadership cognitions. Importantly, whereas most studies have examined children’s attitudes about adult leaders (e.g., Ayman-Nolley & Ayman, 2005; Bigler et al., 2008; Bos et al., 2022; Neff et al., 2007; Patterson et al., 2019; Terrizzi et al., 2019), our research adds to a growing literature focused on children’s attitudes about peer leaders (Charafeddine et al., 2020; Gülgöz, 2015; Mandalaywala et al., 2020; Reyes-Jaquez & Koenig, 2022). These peer contexts may be especially relevant to understanding the development of gender gaps in leadership ambition, as boys and girls may be socialized into peer norms that differ in their emphasis on and encouragement of leader-like behaviors (e.g., Martin & Fabes, 2021).

(1) Perceptions of the Social Rewards and Costs of Leadership

First, we investigated how children think about the social rewards and costs of leadership—namely, how much peer support and cooperation they expect were they (or another child) to take on a leadership role (Studies 1–3). The focus on these cognitions is motivated by the key role of perceived social support and cooperation from others in explaining adult women’s interest in leadership and enactment of leader-like behaviors (e.g., Brescoll, 2011). Importantly, even 3- to 4-year-old children actively seek to manage what others think about them (Asaba & Gweon, 2022) and are highly sensitive to the possibility of social disapproval (e.g., Tomasello, 2014). Social sanctions for not following peer gender norms are particularly salient to children from a young age (Hibbard & Buhrmester, 1998; Martin & Ruble, 2010). In the context of leadership and gender, one study showed that preschool girls were more accepting of low-status (i.e., low-ranked) girls than of high-status girls, whereas the reversed pattern was observed among boys (Sebanc et al., 2003). These gendered patterns of peer support as a function of social rank may lead girls more so than boys to generally expect children who behave in dominant leader-like ways to incur social backlash and may also discourage girls from enacting these assertive behaviors themselves (Martin & Fabes, 2021). However, children’s expectations of social support for or cooperation with leaders have only been examined in the context of politics (e.g., Bigler et al., 2008; Patterson et al., 2019), and it remains unknown how children think about the social costs and rewards of leadership in peer contexts. We addressed this gap by investigating how much children expected that others would support and cooperate with girl and boy leaders in a group activity (Study 1) and how much support and cooperation they expected for themselves were they to become the leader (Studies 2 and 3).

(2) Interest in Leader Roles

Second, we investigated gender differences in children’s interest in a leader role (Studies 2 and 3). Past studies found gender gaps in children’s interest in political leadership (e.g., becoming president one day; Bos et al., 2022). However, there are distinctive features of politics that do not apply to leadership roles more generally. For instance, as discussed by Dolan and Lawless (2021), perceptions of political leaders tend to be negative—an attitude that may not generalize to other kinds of leaders. Moreover, whereas political leadership may be relatively removed from children’s everyday lives, children may have direct experience with leader roles in some capacity (e.g., organizing play or activities with other children) and may be regularly exposed to adults in leader roles outside the realm of politics (e.g., obeying authorities such as teachers and parents). For these reasons, it is important to investigate children’s perceptions of leadership and possible gender gaps in interest in leader roles in age-appropriate contexts.

To our knowledge, the only study that investigated gender differences in interest in a more general leader role that was potentially relevant to children’s lives (i.e., being in control of allocating candy) did not find a gender gap (Reyes-Jaquez & Koenig, 2022), raising the possibility that leadership interest does not differ by gender at an early age. However, in that study, the leader role was embedded in a dyadic situation involving an adult, and it remains unknown how children think of leadership in peer group contexts, where the pressures and incentives to conform to gender-group norms may be stronger (e.g., Martin & Fabes, 2021).

In the current research, we build on these past studies and extend their scope by investigating children’s interest in leader roles that were embedded in a variety of peer-group activities (e.g., building a sandcastle, playing a group game)—a naturalistic context for leadership behavior in childhood. Although communality is important for leaders (e.g., Gerzema & D’Antonio, 2017), our studies emphasized the high-agency aspects of leadership, which are viewed as core or defining characteristics of leaders (Koenig et al., 2011; Vial & Napier, 2018), and which overlap particularly strongly with stereotypically-male traits while being at odds with the traditional female role (Eagly et al., 2020). Specifically, whereas leadership roles can be attained in multiple ways (e.g., by democratic vote), we focused on children’s attitudes toward leader roles that were claimed assertively (e.g., announcing, “I will be in charge!”) because this is the context in which we would expect gender norms to be most influential—for example, girls in particular may anticipate low levels of peer support for leaders in these types of contexts, whereas boys may be generally more receptive to the kinds of dominance displays associated with taking on a leader role (e.g., Hibbard & Buhrmester, 1998).

(3) Self-Efficacy in Leader Roles

Third, in addition to children’s anticipated support for leaders and their own interest in a leader role, we also examined their self-efficacy as leaders—that is, their confidence in and expectations about how well they would perform as leaders (Studies 2 and 3). Self-efficacy is a pillar of motivation (Bandura, 1977; Eccles & Wigfield, 2020; Wigfield & Eccles, 2000) that partly explains adult gender gaps in leadership ambition (Fox & Lawless, 2011; Schneider et al., 2016; Sheppard, 2018). Self-efficacy is shaped by, among other factors, the expectation of support from others. In the context of leadership in particular, a leader’s capacity to lead effectively requires cooperation from followers (Levi et al., 2009; Tyler, 2002). If girls were generally discouraged within their peer groups from enacting leader-like behaviors, as reviewed earlier, they would be unlikely to develop confidence in their ability to lead effectively (see Shapiro et al., 2015). In contrast, boys may be socialized to value leadership skills more, and boy peer groups may offer more opportunities to build a sense of self-efficacy in leader-like roles (Martin & Fabes, 2021). Thus, given that a sense of self-efficacy about one’s own leadership skills may be an important precursor to viewing oneself as a leader (e.g., Eccles & Wigfield, 2020), we investigated whether gender differences emerged in children’s leadership self-efficacy.

(4) Sensitivity of Children’s Leadership Motivation to Contextual Cues

From an intervention perspective, it would be valuable to understand the conditions under which leadership is most appealing to children—and, perhaps, the conditions under which boys and girls find leadership equally attractive. Exploring this issue was the fourth goal of our research. We examined experimentally what promotes or hinders children’s motivation to lead (Study 3). Specifically, we investigated whether children’s (and particularly girls’) interest in leadership may be shaped by two kinds of contextual cues: (a) framing leadership in more communal ways, and (b) the presence of girl leaders as peer role models.

Communality as a Contextual Cue

Although leadership is usually stereotyped as highly dominant and self-oriented (Koenig et al., 2011), it can also entail communal and other-oriented components (e.g., serving one’s community). Regardless of how they are appointed, leaders can behave in highly prosocial ways toward others, making decisions that promote the welfare of subordinates (see Vial & Cowgill, 2022). Indeed, prosocial behaviors are recognized as key features of effective leaders (e.g., Gartzia & van Knippenberg, 2016; Gerzema & D’Antonio, 2017). Consistent with the goal congruity perspective (Diekman & Eagly, 2008), adult gender gaps in interest in leader roles are reduced when the communal and other-oriented aspects of leadership are emphasized (Schneider et al., 2016). Emphasizing communal goals afforded by other male-dominated roles (e.g., science careers) has similarly shown to close gender gaps in interest and to generally boost interest among both women and men (e.g., Belanger et al., 2020). Given that gender differences in the endorsement of communal goals and values are already present at age 6 (Block et al., 2018), scholars have proposed that framing leadership in communal terms may encourage girls to want to become leaders (e.g., Diekman et al., 2021). However, we are unaware of any studies that have examined this idea empirically, as we did in the current research, by manipulating a leader role to be more (or less) communal.

Role Models as a Contextual Cue

The second contextual cue that we examined was the presence of girl leaders as peer role models. Children are encouraged by and imitate the behaviors of their peers, who act as role models (e.g., Bussey & Bandura, 1999). For these reasons, scholars have proposed that exposing girls to relatable female leaders could increase their interest in leadership (e.g., Heck et al., 2021). Indeed, relatable role models can increase middle-school girls’ interest in other male-dominated areas (e.g., O’Brien et al., 2017). The presence of women and girls in leadership roles may counter the stereotypic view that leadership is “for men” or “for boys” and signal that leadership is compatible with the traditional female role. Consistent with this view, research has found higher leadership aspirations among young women who had had extensive exposure to female leaders (i.e., who had attended female-only colleges; Dasgupta & Asgari, 2004). Similarly, Goodwin et al. (2020) showed that women were more interested in joining a leadership group with a higher (vs. lower) proportion of women. However, it remains an open question whether the presence of own-gender peers (i.e., other girls) in leader roles would increase girls’ own interest in leadership. We addressed this gap by manipulating perceptions of the proportion of boys and girls in leader roles.

(5) Age and Race/Ethnicity as Moderators of Leadership Cognitions and Motivation

Research suggests that concepts of social power and leadership may be somewhat tenuous before age 5 (e.g., Gülgöz, 2015); thus, we focused our investigation on children between the ages of 5 and 10—an age range that is common in other studies on children’s leadership cognitions in relation to gender (e.g., Bigler et al., 2008; Reyes-Jaquez & Koenig, 2022). As past investigations have found an increase with age in children’s tendency to associate leadership with males and masculinity (e.g., Charafeddine et al., 2020; Mandalaywala et al., 2020; Neff et al., 2007), we explored the possibility that any male bias or gender gaps in interest in leadership in our studies may similarly become accentuated with age. This may be especially the case among girls, whose internalization of traditionally female standards and peer norms may strengthen between 5 and 10 years of age, potentially leading to more negative attitudes toward leadership over time.

Moreover, we took advantage of the diversity of our samples in terms of race/ethnicity (39.3% identified as White, on average, across the three studies) to explore the possibility that children of color may differ in their attitudes toward leadership from White children and to examine whether these patterns might further depend on child gender. Although our studies were not explicitly designed to take an intersectional approach, which highlights the uniqueness of interlocking social identities such as gender and race (e.g., Lei & Rhodes, 2021), our findings may serve as a foundation on which future work can build on. Whereas research has found that children sometimes associate power and status with members of dominant (vs. subordinate) racial/ethnic groups (Bigler et al., 2003, 2008; Liben et al., 2001), to our knowledge, no studies on gender gaps in interest in leadership have taken an intersectional perspective, either with children or with adults (e.g., Netchaeva et al., 2022; Pate & Fox, 2018; Schneider et al., 2016; Schuh et al., 2014; Sheppard, 2018). However, some studies suggest that lack of fit or incongruity perceptions for leadership roles depend not only on gender but also on race/ethnicity (e.g., Galinsky et al., 2013; Livingston et al., 2012; Rosette & Tost, 2010; Rosette et al., 2016), and the mismatch between agentic leadership and female stereotypes may be weaker for women and girls of color relative to White women and girls. For instance, Black women (compared to White women) are more likely to be characterized with attributes that align with dominant agency (e.g., confident, assertive, aggressive, strong, dominant, not subservient) (Ghavami & Peplau, 2013; Rosette et al., 2016). Thus, taking on a leadership role may constitute less of a proscriptive role violation for women (and girls) of color relative to White women. If so, it is possible that gender gaps in interest in leadership may be less apparent among children of color.

Transparency and Openness

The three studies, which were not preregistered, were approved by the institutional review board at New York University (IRB-FY2016-1163). Across studies, we report how we determined sample size, all data exclusions, all manipulations, and all measures. The full materials, raw data, analytic syntax, and supplementary analyses for the three studies are available on the Open Science Framework (OSF): https://osf.io/h684j/.

Study 1

In Study 1, we examined children’s anticipation of social support for children who claim a leader role in the context of a group activity. Given that girls’ peer groups tend to discourage assertive and dominant behaviors (e.g., Sebanc et al., 2003), we expected that girls may anticipate negative reactions to anyone who displays this kind of behavior, consistent with goal congruity theory (Diekman & Eagly, 2008). Moreover, drawing from lack of fit and role congruity theories (Eagly & Karau, 2002; Heilman, 1983, 2001), we also tested the possibility that children might be biased against girl leaders, as adults have been found to be biased against women leaders (e.g., Brescoll, 2011; Heilman & Okimoto, 2007). That is, we also examined whether children anticipated less social support for girl (vs. boy) leaders.

Method

Participants

Participants were 99 children in a large city in the Northeastern U.S. between the ages of 5 and 10 years (50 boys, 49 girls; M = 7.84 years, SD = 1.71 years, range = 4.59–10.54 years), who participated in the study at their schools (n = 37), in children’s museums (n = 52), or in a university laboratory (n = 10). We sampled children systematically by age, taking care to include similar numbers of boys and girls in three age “bins”: 5–6, 7–8, and 9–10 years. Children were 34.3% White, 15.2% Hispanic or Latinx, 11.1% Asian or Pacific Islander, 3.0% Black, 18.2% Multiracial or Multiethnic, and 6.1% Other; 12.1% of parents did not report their child’s race/ethnicity. The sample size for each age group by child gender and race/ethnicity is reported in Table S1 in the Supplement on https://osf.io/h684j/.

The total sample size was determined a priori based on several considerations such as (a) ensuring balance by leader gender and participant gender within each age group and (b) counterbalancing story order, group size, and gender composition of the peer groups described in the stimuli (see Procedure and Measures). A sensitivity analysis conducted with G*Power (Faul et al., 2007) indicated that n = 99 was sufficient to detect medium-sized differences (Cohen’s d = 0.56) by participant gender (male; female) and leader gender (boy; girl) on a two-tailed independent-samples t-test, assuming power = 80% and alpha = .05. Although, as detailed in the Analytic Strategy, our actual analyses in Study 1 consisted of mixed-effects models, we report a sensitivity analysis based on a t test both because (1) it is likely conservative—mixed-effects models are generally more powerful because they take advantage of more of the data than t tests—and because (2) sensitivity analyses for mixed-effects models are (still) dauntingly complex to perform.

Procedure and Measures

We presented each child with four stories set in four different contexts (at the beach, at school, at summer camp, or at the park) in counterbalanced order. Each story described a group of children engaging in an activity together (e.g., building a sandcastle) and a child within the group who claimed a leader role (i.e., “the leader”) by announcing, “I will be in charge of [activity]; I will make decisions and tell everyone else what to do.” Thus, across stories, leaders claimed their role in a highly assertive manner that was at odds with the traditional female role (Eagly et al., 2020). The gender of the leader varied between participants: For half of them, the leader in each of the four stories was male, whereas for the other half, the leader in each of the four stories was female. The full script for each story is available in the Supplement on https://osf.io/h684j/ (p. 3). Images of all leaders and other visuals are also available on https://osf.io/h684j/. We pre-tested these materials by asking a sample of adults (n = 52) on Amazon Mechanical Turk (MTurk) to rate each leader on perceived age, intelligence, warmth, and physical attractiveness. These judgements were unrelated to children’s responses in Study 1 (see Supplement on https://osf.io/h684j/, Table S4). Each story featured a different group of children, counterbalanced within-subjects along two dimensions (a) size (2 vs. 10 children in addition to the leader) and (b) gender composition (same gender as the leader vs. mixed gender). These factors did not influence the results (see Supplement on https://osf.io/h684j/, Table S2).

After each story, we asked children four questions in counterbalanced order to gauge their anticipation of social support for the leader from other children in the story. Here and in Studies 2 and 3, our measure of anticipated social support was conceptually similar to adult measures of “fear of backlash” (e.g., Moss-Racusin & Rudman, 2010); however, we employed primarily positively-worded items to facilitate children’s comprehension. Our chosen label for the measure (“anticipated social support”) reflects this adaptation. The questions were as follows: (a) “Would they think he/she is nice, or not?”, (b) “Would they think he/she is bossy, or not?”, (c) “Would they want to be friends with him/her, or not?”, and (d) “Would they like him/her more, or less?” Children responded in two steps: an initial yes/no response (e.g., “Yes, they would think she’s nice”), followed by a two-point scale (e.g., “Sort of nice? Or really nice?”). Responses ranged from 1 (e.g., really not nice) to 4 (e.g., really nice). After reverse-scoring the “bossy” item, we averaged answers to the four questions to form a measure of anticipated social support for leaders (α = .85); higher numbers indicate stronger anticipation of social support.

After the four stories, children were told that “it is okay for any child to step up to be in charge,” and were offered a small prize (e.g., a sticker). For exploratory purposes, at the end of the session we also asked a subset of children (n = 33; 17 boys and 16 girls) to indicate why they believed the four leaders in the stories said they would be in charge, and whether they (the participants) would like to be in charge like the children in the stories. A descriptive summary of the responses is available in the Supplement on https://osf.io/h684j/ (p. 7 and Tables S5 and S6).

Analytic Strategy

To examine whether participant gender and leader gender influenced responses to the four stories, we first conducted a mixed-effects linear regression on anticipated social support with leader gender condition (0 = male leaders, 1 = female leaders), participant gender (0 = boys, 1 = girls), participant age in years (with 2+ decimal precision; continuous), and their interactions as predictors, including crossed random intercepts for participant and story setting. We then examined whether participant race/ethnicity (0 = children of color, 1 = White children) moderated any of the effects by including this variable and all interactions in the model (n = 87). Finally, we evaluated whether the effects of leader gender condition and/or participant gender varied as a function of (a) group size and gender composition and (b) leader characteristics. These factors did not change the results appreciably (see Supplement on https://osf.io/h684j/, Tables S2 and S4). In all models, predictors were mean-centered.

Across Studies 1–3, all analyses were conducted in Stata 16 (StataCorp, 2019). In all studies, we standardized the dependent variables before including them in the models. Thus, the coefficients for dichotomous variables can be interpreted similarly to a Cohen’s d: They represent the fraction of a standard deviation by which the dependent variable changes in response to a shift from one level of a dichotomous predictor to the other. The mixed-effects models were computed with the mixed command; follow-up tests on these models were computed with the margins command. Here and in subsequent studies, we report observed (rather than predicted) means.

Results

The analysis revealed a significant main effect of participant gender, such that girls anticipated lower social support (M = 1.78, SD = 0.66) than boys (M = 2.08, SD = 0.81), β = −0.42, SE = 0.16, p = .011 (see Fig. 1). The main effect of leader gender condition was not significant, β = 0.05, SE = 0.16, p = .75, nor was there a significant two-way interaction between participant gender and leader gender, β = 0.64, SE = 0.33, p = .050. The age coefficient was not significant, β = −0.03, SE = 0.05, p = .53, and there were no interactions between participant age and the other predictors, all ps > .10.

Fig. 1
figure 1

Anticipated Social Support for Child Leaders Among Girls and Boys in Study 1. Note. Each box extends from the first to the third quartile of the data (Q1 and Q3, respectively; the “interquartile range”). The whiskers extend up to 1.5 × the interquartile range below Q3 and above Q1. Within each box plot, the solid line in the middle represents the median and the diamond represents the mean

Full size image

An additional model that included children’s race/ethnicity revealed no significant difference between White children and children of color in terms of the social support they anticipated for the child leader, β = 0.31, SE = 0.18, p = .093. However, we did observe an interaction between children’s race/ethnicity and leader gender, β = 0.75, SE = 0.37, p = .043, such that White children (n = 48) anticipated more social support for female (vs. male) leaders, β = 0.61, SE = 0.29, p = .034, whereas children of color (n = 39) did not, β = −0.08, SE = 0.22, p = .73. There were no other significant interactions between race/ethnicity and leader gender, participant gender, or age on anticipated social support, all ps ≥ .093 (full results in Table S3 in the Supplement on https://osf.io/h684j/).

Discussion

The results of Study 1 suggest that girls expect child leaders to receive less social support from peers compared to boys, regardless of leader gender and context (i.e., the setting of the activity and the size or gender composition of the group; see Supplement on https://osf.io/h684j/, Table S2). It is possible that girls in Study 1 expected less social support for leaders than boys because, across stories, claiming the leader role represented a highly assertive behavior, which is at odds with the traditional female role (Eagly et al., 2020). Alternatively, girls may have expected leaders to be less supported because girls’ peer groups tend to discourage dominant behaviors, whereas boys view them more favorably (e.g., Sebanc et al., 2003). The participant gender difference in anticipated social support for leaders was not moderated by children’s age or race/ethnicity, but we did find that White children expected girl leaders to elicit more social support than boy leaders, whereas children of color did not differentiate between girl and boy leaders. Although speculative, it is possible that White children in Study 1 may have been more attuned to leader gender differences than children of color given that the leaders in the stories were all White—a limitation we return to in the General Discussion. Even so, it is interesting to note the contrast between the fact that participating girls expected lower social support for leaders, while the girls in the stories were expected to receive more support, at least by White children. The latter finding is discrepant with most of the adult literature on leadership (Eagly & Heilman, 2016).

If girls expect child leaders to be relatively unsupported by their peers (compared to boys), then they might be more reluctant than boys to act as leaders. We examined this possibility in Study 2, building on Study 1 by again measuring participating children’s anticipation of social support from peers, this time in reference to themselves in a leader role.

Study 2

In contrast to Study 1, which focused on children’s perceptions of other leaders, in Study 2 we investigated children’s perceptions of themselves as leaders—in particular, whether girls express less interest than boys in being a leader, whether they expect lower levels of social support (similar to Study 1) and cooperation from other children, and whether they exhibit less self-efficacy regarding their ability to lead—which, we reasoned, might be connected to children’s expectation of social support and cooperation from others (e.g., Tyler, 2002).

Method

Participants

Participants were 149 children in a large city in the Northeastern U.S. between the ages of 5 and 10 years (77 boys, 72 girls; M = 7.99 years, SD = 1.66 years, range = 4.86–10.96 years), who took part in the study at their schools (n = 37), in children’s museums (n = 94), or in a university laboratory (n = 18). The sample size was determined a priori to ensure balance by participant gender and age. We again sampled children systematically by age, taking care to include similar numbers of boys and girls in three age “bins”: 5–6, 7–8, and 9–10 years. The sample size for each age group by child gender and race/ethnicity is reported in Table S1 in the Supplement on https://osf.io/h684j/. We excluded children who answered attention check questions incorrectly (n = 4; see Procedure). A sensitivity analysis conducted via simulation (50,000 iterations) with the paramtest package (Hughes, 2017) in R version 4.2.2 (R Core Team, 2021) indicated that a sample of 149 children was sufficient to detect small-to-medium gender differences in interest in leadership (|β|s ≥ 0.46) with 80% power in a linear regression structured as described in Analytic Strategy. Children were 31.5% White, 14.8% Hispanic or Latinx, 8.7% Asian or Pacific Islander, 6.0% Black, 11.4% Multiracial or Multiethnic, and 6.0% Other; 21.5% of parents did not report their child’s race/ethnicity.

Procedure

We presented children with a novel game (the “Zarky Game”) and told them that they would have an opportunity to play this game in the future with other children of their age. We described the game as one in which children play together and, although no leader is necessary, one player may decide to be in charge. We specified that having a leader was not mandatory and that claiming the leader role was entirely optional; thus, as in Study 1, claiming the leader role in Study 2 constituted a highly assertive behavior at odds with the traditional female role (Eagly et al., 2020):

In the Zarky Game, children play together, and they don’t need anyone to be in charge. But sometimes one of the kids in the game wants to be the Zarky Boss. At the beginning of the game, all of the kids are sitting down, see? If a kid stands up and says, ‘I will be the Zarky Boss,’ then this kid becomes the Zarky Boss. The Zarky Boss is then in charge of the game, makes decisions, and tells everyone else what to do. The kids who stay seated and who do not stand up are the followers. They do whatever the Zarky Boss says.

After hearing the description of the game, children were reminded that they would play the game in the future with other children their own age and were then asked a series of questions to measure their interest in the leader role. These questions were followed by three counterbalanced blocks of questions that measured children’s expectations if they were the game leader, including their anticipation of (a) social support (similar to Study 1) and (b) cooperation from other children in the game, and (c) their sense of self-efficacy as game leader. Within each block, question order was counterbalanced. For exploratory purposes, we tested the same constructs with respect to other children in the leader role (a boy and a girl). A description of these additional measures and the results is provided in the Supplement on https://osf.io/h684j/ (pp. 14–17).

At the end of the session, children answered two attention check questions to confirm their understanding of the leader role (e.g., “Does the Zarky Boss tell the other kids what to do? Or does the Zarky Boss do what the other kids say?”). We excluded children who answered both questions incorrectly (n = 4). Finally, children were told that “it is okay for any child to step up to be in charge” and were offered a small prize (e.g., a sticker).

Measures

Interest in the Leader Role

First, children were asked, “Would you like to be the Zarky Boss? Or would you not like to be the Zarky Boss?” After the child responded, the experimenter followed up by asking whether they would “sort of” or “really” like/not like to be the Zarky Boss. These responses were coded on a 4-point scale (1 = really not like to, 4 = really like to). Next, children were asked to explain the reasoning behind their decision (open-ended). We coded these responses for exploratory purposes; a descriptive summary of the results can be found on pp. 10–12 and Table S7 in the Supplement on https://osf.io/h684j/. Finally, we asked children to choose between being a follower (coded as 0) and being the Zarky Boss (coded as 1). We standardized children’s responses to the first and last questions (r = .70, p < .001) and averaged them into a single measure. Higher numbers indicate more interest in being the leader.

Anticipated Social Support

To gauge how much social support children anticipated receiving as would-be game leaders, we adapted three items from Study 1 to the game context. The questions began with the stem, “After you said you would be in charge of the game, would the other children…”: (a) “like you more? Or like you less?”; (b) “want to be friends with you? Or not want to be friends with you?”; and (c) “think you are nice? Yes? Or no?” Responses were coded on a 4-point scale (e.g., 1 = like me a lot less, 4 = like me a lot more), and were averaged into a single measure (α = .70); higher numbers indicate stronger anticipation of social support from other children.

Anticipated Cooperation

Here and in Study 3, our measure of anticipated cooperation was inspired by the adult literature on leadership, which suggests that subordinate cooperation reflects and shapes the legitimacy of authority figures and differs based on leader gender (e.g., Tyler, 2002; Vial et al., 2018; see also Vial et al., 2016). Three questions gauged how much participants expected the other children to cooperate with them as would-be leaders. The questions began with the stem, “After you said you would be in charge of the game, would the other children…”: (a) “want to play Zarky more? Or want to play less?”; (b) “do what you say? Or not do what you say?”; and (c) “pay attention to you? Or not pay attention to you?” Responses were coded on a 4-point scale (e.g., 1 = really not do what I say, 4 = really do what I say), and were averaged into a single measure (α = .75); higher numbers indicate stronger anticipation of cooperation from other children.

Self-Efficacy as Leader

Three questions evaluated children’s self-efficacy as a would-be leader: (a) “How good do you think you’d be as the Zarky Boss? Would you be good at it? Or would you not be good at it?”; (b) “How good do you think you’d be at telling other kids what to do? Would you be good at it? Or would you not be good at it?”; and (c) “How well would the other children do at Zarky with you as Zarky boss? Would they do well? Or not so well?” Responses were coded on a 4-point scale (e.g., 1 = really not well, 4 = really well), and were averaged into a single measure (α = .63).

Analytic Strategy

For each dependent variable, we first conducted a linear regression with participant gender (0 = boys, 1 = girls), participant age in years (with 2 + decimal precision; continuous), and their interaction as predictors. Then, for each dependent variable, we examined whether participant race/ethnicity (0 = children of color, 1 = White children) moderated any of the effects by including this variable and all interactions in the model (n = 117). When appropriate, we used a Bonferroni-corrected alpha level to account for multiple comparisons. In all models, predictors were mean-centered.

Results

Means, standard deviations, and bivariate correlations can be found in Table 1. Anticipated social support and cooperation from others were positively related to each other, r(146) = .42, p < .001, and both were positively related with self-efficacy as a leader (rs ranging from .34 to .44, ps < .001). These three variables were positively related with interest in the leader role (rs ranging from .24 to .52, ps < .01).

Table 1 Means, Standard Deviations, and Bivariate Correlations in Study 2
Full size table

Interest in the Leader Role

Girls’ interest in the leader position (M = −0.13, SD = 0.98) did not differ significantly from boys’ interest (M = 0.12, SD = 1.01), β = −0.24, SE = 0.16, p = .14. The coefficient for participant age was not significant either, β = −0.07, SE = 0.05, p = .13, nor was there a significant gender × age interaction, β = 0.14, SE = 0.10, p = .16.

When participant race/ethnicity was entered into the model, results revealed a significant three-way interaction with participant gender and age, β = 0.55, SE = 0.23, p = .018. A sensitivity analysis conducted via simulation (50,000 iterations) indicated that n = 117 was sufficient to detect a three-way interaction effect as small as |β|= 0.69 with 80% power in this linear regression model (see https://osf.io/h684j/ for R code). Thus, our sample was slightly underpowered to detect the observed interaction. As illustrated in Fig. 2, which shows interest in the leader role as a function of age and gender separately for children of color (Panel A) and White children (Panel B), the three-way interaction emerged because interest in the leader position increased with age for White girls, whereas it decreased with age for all other groups (i.e., White boys, girls of color, and boys of color). None of the slopes were significant after adjusting for multiple comparisons, ps > .09.

Fig. 2
figure 2

Interest in Leader Role as a Function of Age and Gender Among a Children of Color and b White Children in Study 2

Full size image

Anticipated Social Support

The basic model revealed no significant gender difference in anticipated social support from other children in the game, β = 0.07, SE = 0.16, p = .67, no significant age differences, β = −0.07, SE = 0.05, p = .18, and no interaction between gender and age, β = 0.10, SE = 0.10, p = .33. The model with participant race/ethnicity revealed no main effect for this variable, β = 0.15, SE = 0.19, p = .44, and no interactions with the other predictors, ps > .18.

Anticipated Cooperation

Results revealed no significant gender difference in anticipated cooperation from other children in the game, β = −0.07, SE = 0.16, p = .68, no significant age differences, β = 0.06, SE = 0.05, p = .24, and no significant interaction between gender and age, β = 0.02, SE = 0.10, p = .83. Adding participant race/ethnicity to the model revealed no significant difference between White children and children of color, β = 0.35, SE = 0.19, p = .066. No other coefficients were significant, ps > .09.

Self-Efficacy as Leader

The basic model revealed no significant gender difference in sense of self-efficacy as leader, β = 0.03, SE = 0.16, p = .87, no significant age differences, β = −0.06, SE = 0.05, p = .27, and no interaction between gender and age, β = 0.03, SE = 0.10, p = .77. Similarly, no coefficients were significant in the model that included the race/ethnicity variable, ps > .11.

Discussion

The results of Study 2 provide no evidence of gender gaps in interest in leadership in children and no support for the idea that leadership interest among girls is particularly low. In fact, the results suggest a growing interest in leadership among White girls from ages 5 to 10, and a waning interest among all other groups (i.e., White boys, girls and boys of color). This finding indicates that the gender gaps in interest in leadership found among adults may emerge at a later developmental stage, presumably in adolescence (Bos et al., 2022; Fox & Lawless, 2014). We return to these possibilities in the General Discussion. Although the results involving children’s racial/ethnic background should be interpreted with caution given the small sample sizes for any particular age group × race/ethnicity combination, it is noteworthy that, as shown in Fig. 2, the relationship between age and interest in the leader role appears to differ for White girls compared to White boys but shows no apparent gender difference in children of color. These results underscore the value of investigating intersecting identities in the study of gender gaps in leadership ambition (e.g., Heck et al., 2021).

We found little evidence in the current study for gender gaps in the anticipation of social support or cooperation from peers when children imagined themselves in a leader role. This result is in tension with the findings of Study 1, in which girls expected less social support than boys for other children in leader roles. We return to this inconsistency in the General Discussion. Importantly, in Study 2, we also found little evidence for a gender gap in children’s sense of self-efficacy, which suggests that any gender differences in interest in leader roles at this young age may not be due to girls’ lack of confidence in their leadership abilities.

In our final study, we build on these findings by investigating the possibility that children’s expectations for themselves in a leader role (and their interest in it) may be sensitive to contextual cues that mark the leader role as more or less communal and gender-neutral.

Study 3

The goal of Study 3 was to build on the results of Study 2 to test whether children’s interest in a leader role (particularly among girls) is sensitive to (a) whether the leader role is framed in a way that aligns with communality and (b) whether girls are explicitly represented among leaders (thus serving as peer role models for participating girls). To accomplish this goal, we used the same procedure as in Study 2, with two important changes. First, we manipulated the description of the leader role. Drawing from goal congruity theory (Diekman & Eagly, 2008), we reasoned that emphasizing the communal aspects of the leader role might increase children’s interest in that role and their anticipation of support and cooperation from others, and potentially close gender gaps in children’s interest in leadership. Thus, in one condition, the leader role was described as communal—specifically, helpful to others. In the other condition, the leader role was described exactly as in Study 2—that is, in highly agentic terms. Second, we manipulated the presence of girl leaders, which allowed us to test the possibility that exposure to female role models could increase girls’ interest in leadership (e.g., Heck et al., 2021). In one condition, most past leaders were boys, which we assumed to be closest to children’s baseline impressions of leadership, given that leadership stereotypes tend to overlap with male stereotypes (Koenig et al., 2011), that leaders around the world tend to be men (World Economic Forum, 2020), and that children see these patterns reflected in media from an early age (Aley & Hahn, 2020). In the other condition, an equal number of past leaders were boys and girls.

We expected children—girls, in particular—to be more interested in the leader role and to anticipate more support and cooperation from others when the leader role was framed in communal (vs. agentic) terms and when girl leaders were well-represented (vs. a minority). We expected these two factors to have an additive effect on children’s responses; we did not expect them to interact.

Method

Participants

Participants were 244 children in a large city in the Northeastern U.S. between the ages of 5 and 10 years (121 boys, 123 girls; M = 8.01 years, SD = 1.64 years, range = 5.00–11.02 years), who participated in the study at their schools (n = 72), in children’s museums (n = 46), in a university laboratory (n = 21), or online via Zoom (n = 105). The sample size was determined a priori based on several considerations: We sampled children systematically by age as in Studies 1 and 2 and tried to ensure gender balance within each age group (5–6, 7–8, and 9–10); we also counterbalanced a number of variables described in the Procedure and Measures section. The sample size for each age group by child gender and race/ethnicity is reported in Table S1 in the Supplement on https://osf.io/h684j/. An additional 26 children were tested but excluded because they answered attention check questions incorrectly (see Procedure and Measures). A sensitivity analysis conducted via simulation (50,000 iterations) indicated that a sample of 244 children was sufficient to detect effects of the manipulations on children’s leadership interest as small as |β| = 0.37 with 80% power in a linear regression structured as described in Analytic Strategy. Children were 52.1% White, 14.8% Hispanic or Latinx, 8.2% Asian or Pacific Islander, 6.2% Black, 9.4% Multiracial or Multiethnic, and 0.8% Other; 8.6% of parents did not report their child’s race/ethnicity.

Procedure and Measures

We presented children with the same novel game from Study 2, using an identical script, but we modified the description of the game to manipulate (a) the perception of the leader role as communal or agentic, and (b) the presence of girl peer leaders as role models (i.e., whether girls were well-represented or a minority among past leaders). The Appendix contains the full script used to describe the game. At the end of the session, children were told that “it is okay for any child to step up to be in charge” (as in Study 2) and were offered a small prize (e.g., a sticker).

Agentic vs. Communal Leader Framing Manipulation

We randomly assigned half of the children to an “Agentic Leader” framing condition in which the description of the leader role was identical to that in Study 2. The other half of the children were assigned to a “Communal Leader” framing condition in which the experimenter emphasized that having a leader in the game was “very helpful” (see Appendix). We reasoned that, although claiming the leader role may generally constitute an assertive behavior, doing so in a context in which it is explicitly helpful to have a leader (vs. a context in which helpfulness was not mentioned) would represent a more communal behavior. To help children encode the information about the leader role, the experimenter asked them a memory check question (Communal Leader condition: “Is it helpful to have a Zarky Boss? Or is it not helpful?”; Agentic Leader condition: “Does the Zarky game need a boss? Or does it not need a boss?”). Regardless of whether the child responded correctly, the experimenter restated the correct answer before continuing.

Role Model Manipulation

We manipulated the presence of girl leaders as peer role models by showing participants a list of children who had ostensibly played the game the week before and had claimed the leader role (see Appendix). The list contained twelve children including their names, ages, and a picture, with their gender clearly marked by the colors blue (for boys) and pink (for girls). We randomly assigned half of the participating children to a “majority male” condition in which girl leaders were poorly represented (i.e., only 25% of past leaders were girls). We assumed that this condition was closest to a baseline condition in which information about past leaders was not made explicit (e.g., Study 2), given male stereotypes about leadership (Koenig et al., 2011), the fact that most leaders in the world are men (World Economic Forum, 2020), and that children are exposed to these patterns in media from an early age (e.g., Aley & Hahn, 2020). The other half of the children were assigned to a “gender-balanced” condition in which girl and boy leaders were equally represented. To help children encode the gender composition of past leaders, the experimenter asked them to count the number of boys and girls. If the child answered incorrectly, the experimenter counted aloud to arrive at the correct response. Next, the experimenter asked children a memory check question: “Of the kids who were Zarky Boss last week, were they mostly boys? Or were they mostly girls? Or about the same number of boys and girls?” If the child answered incorrectly, the experimenter produced the list again and counted the boys and girls one more time before proceeding.

Measures

After describing the game, the experimenter reminded participating children that they would play the game in the future with other children their own age and asked the same questions from Study 2 to measure their interest in the leader role (r = .72) and their expectations as would-be game leaders, including their anticipation of (a) social support (α = .74) and (b) cooperation (α = .69) from other children, and (c) their sense of self-efficacy as game leaders (α = .79). We counterbalanced the order of the three blocks of questions and of the questions within each block. We also explored children’s expectations for leaders of their own gender group. A full description of these additional measures and an analysis of children’s responses is reported in the Supplement on https://osf.io/h684j/ (pp. 18–22).

Next, children answered four questions that gauged their attentiveness and recall of the key manipulations. The first two questions were identical to the questions that had already been asked earlier in the session to help children retain the information about the manipulations: (a) whether the leaders of the previous week had been mostly boys, mostly girls, or about the same number of boys and girls, and (b) whether or not the game needed a leader (in the “Agentic Leader” condition) or whether or not it was helpful to have a leader (in the “Communal Leader” condition). We excluded children who answered either question incorrectly when asked this second time (n = 15). Lastly, to gauge recall of what the leader role entailed (i.e., telling other kids what to do), we asked all children the two attention check questions from Study 2; we excluded children who answered both questions incorrectly (n = 11).

Analytic Strategy

We conducted a linear regression model for each dependent variable with participant gender (0 = boys, 1 = girls), participant age in years (with 2 + decimal precision; continuous), leader framing condition (0 = agentic leader, 1 = communal leader), and role model condition (0 = majority male, 1 = gender-balanced) as predictors (all mean-centered), and all interactions. Then, as in Studies 1 and 2, we tested the same models with the addition of participant race/ethnicity (0 = children of color, 1 = White children) (n = 223) to examine if this variable moderated any of the effects. When appropriate, we use a Bonferroni-corrected alpha level to account for multiple comparisons.

Results

Means, standard deviations, and bivariate correlations can be found in Table 2. As in Study 2, anticipated social support and cooperation from others in Study 3 were positively related to each other, r(242) = .62, p < .001, and both were positively related with self-efficacy as a leader (rs ranging from .47 to .48, ps < .001). These three variables were positively related with interest in the leader role (rs ranging from .18 to .52, ps < .01).

Table 2 Means, Standard Deviations, and Bivariate Correlations in Study 3
Full size table

Interest in the Leader Role

Interest in the leader role did not differ significantly between girls (M = 0.09, SD = 0.98) and boys (M = −0.09, SD = 1.01), β = 0.17, SE = 0.13, p = .18. Children were significantly more interested in the leader role in the communal leader framing condition (M = 0.20, SD = 0.95) compared to the agentic leader framing condition (M = −0.19, SD = 1.01), β = 0.38, SE = 0.13, p = .003 (see Fig. 3). The two-way interaction between participant gender and leader framing condition was not significant, β = −0.07, SE = 0.25, p = .77. Surprisingly, children expressed significantly more interest in the leader role in the majority-male role model condition (M = 0.15, SD = 0.99) compared to the gender-balanced role model condition (M = −0.13, SD = 1.00), β = −0.28, SE = 0.13, p = .031; this effect did not interact with participant gender, β = 0.04, SE = 0.25, p = .89. The coefficient for participant age was not significant, β = −0.02, SE = 0.04, p = .52, nor were there any other significant coefficients in the model, ps > .21.

Fig. 3
figure 3

Interest in Leader Role as a Function of Participant Gender and a Leader Framing Condition and b Role Model Condition in Study 3

Full size image

Adding participant race/ethnicity to the model revealed no main effect of this variable, β = −0.02, SE = 0.15, p = .91. However, there was a significant three-way interaction with participant gender and leader framing condition, β = 1.30, SE = 0.59, p = .028. A sensitivity analysis conducted via simulation (50,000 iterations) indicated that a sample of 223 children was sufficient to detect a three-way interaction effect as small as |β|= 1.68 with 80% power in this linear regression model (see https://osf.io/h684j/ for R code). Thus, our sample was slightly underpowered to detect the observed interaction. As seen in Fig. 4, which shows interest in the leader role as a function of leader framing condition and gender separately for children of color (Panel A) and White children (Panel B), the effect of leader framing condition was particularly strong among boys of color, who reported significantly more interest in the leader role in the communal (vs. agentic) leader framing condition, β = 0.91, SE = 0.35, p = .037 (Bonferroni-corrected). For all other groups (i.e., girls of color, White boys, and White girls), the effect of leader framing condition was not significant after adjusting for multiple comparisons, ps > .14. Participant race/ethnicity did not moderate any other effects, ps > .11.

Fig. 4
figure 4

Interest in Leader Role as a Function of Leader Framing Condition and Participant Gender Among a Children of Color and b White Children in Study 3

Full size image

Anticipated Social Support

Girls anticipated significantly more social support (M = 3.17, SD = 0.67) than boys (M = 2.99, SD = 0.71), β = 0.25, SE = 0.12, p = .041. Age was negatively associated with anticipated social support, β = −0.11, SE = 0.04, p = .003. As with interest in the leader role, there was a significant effect of leader framing condition, β = 0.42, SE = 0.12, p = .001, such that children anticipated more social support in the communal leader condition (M = 3.22, SD = 0.60) compared to the agentic leader condition (M = 2.94, SD = 0.75). No other coefficients were significant, ps > .08.

Results were similar when we included participant race/ethnicity in the regression model. The main effect of this variable was not significant, β = 0.01, SE = 0.14, p = .92, and race/ethnicity did not interact significantly with any other predictors, ps > .13.

Anticipated Cooperation

Similar to the results for anticipated social support, girls anticipated significantly more cooperation from other children (M = 3.34, SD = 0.63) than boys (M = 3.16, SD = 0.72), β = 0.27, SE = 0.13, p = .033. Additionally, children anticipated significantly more cooperation in the communal leader condition (M = 3.43, SD = 0.53) compared to the agentic leader condition (M = 3.08, SE = 0.75), β = 0.52, SE = 0.13, p < .001. No other coefficients were significant, ps > .34.

Results revealed a non-significant main effect of participant race/ethnicity when this variable was included in the model, β = −0.002, SE = 0.15, p = .99. Race/ethnicity did not interact significantly with any other predictors, ps > .059.

Self-Efficacy as Leader

Children’s self-efficacy was significantly higher in the communal leader condition (M = 3.29, SD = 0.67) compared to the agentic leader condition (M = 3.01, SD = 0.82), β = 0.37, SE = 0.13, p = .004. No other coefficients were significant, ps ≥ .066.

Adding participant race/ethnicity to the model revealed a non-significant main effect of this variable, β = 0.07, SE = 0.15, p = .64. However, there was a significant race/ethnicity × leader framing condition interaction, β = −0.59, SE = 0.30, p = .048. Framing the leader role as communal had a substantial positive effect on self-efficacy among children of color (n = 127, 67 boys and 60 girls), β = 0.77, SE = 0.23, p = .001, but not among White children (n = 96, 44 boys and 52 girls), β = 0.17, SE = 0.18, p = .37. There were no other significant interactions with participant race/ethnicity, ps > .12.

Discussion

The results of Study 3 showed that interest in a leader role was stronger when this role was more compatible with communality—a pattern that emerged for all children, not only girls, similar to past findings among adults (e.g., Belanger et al., 2020). The leader framing manipulation had reliable effects across dependent variables, such that children were more interested in the leader role in the communal leader condition compared to the agentic leader condition (which resembled the leader role in Study 2), and they anticipated stronger social support and cooperation as would-be leaders, as well as higher self-efficacy. As a whole, these findings suggest that many children, regardless of gender, may benefit from learning about the communal aspects of leadership.

As in Study 2, some of these positive effects interacted in nuanced ways with participant demographics. Specifically, framing the leader role as communal had a much stronger effect on interest in leadership among boys of color than among girls of color or White children. Moreover, children of color (regardless of gender) anticipated significantly higher self-efficacy as leaders when the leader role was framed as communal, whereas this manipulation did not increase self-efficacy for White children. Although these patterns should be interpreted with caution given the small sample sizes for any particular gender × race/ethnicity combination, they again highlight the relevance of intersectional perspectives for a comprehensive understanding of leadership cognitions, not only in adults (e.g., Rosette et al., 2016) but also in children (Heck et al., 2021; Lei & Rhodes, 2021).

It is also worth noting that, as in Study 2, we did not find a gender gap in interest in leadership in Study 3, regardless of leader framing condition. These findings suggest that, if gender gaps in interest in leadership exist among 5- to 10-year-old children, they may be less robust than adult gender gaps.

Although we have focused on the positive effects of framing leadership in communal terms, the patterns we observed could conceivably be due to the detrimental effects of framing leadership in exclusively agentic terms. Given that the latter framing is closer to adults’ view of leadership (e.g., Koenig et al., 2011), we conceptualized the agentic leader framing condition as a baseline. However, given mixed findings on children’s cognitions about gender and leadership (Bos et al., 2022), it seems possible that children in Study 3 might have reacted negatively to the agentic framing condition instead of (or in addition to) reacting positively to the communal framing condition.

The role model manipulation had no impact on children’s anticipation of social support or cooperation from others or on their self-efficacy as leaders. Surprisingly, children were overall less interested in the leader role when previous leaders were gender-balanced (vs. mostly male). It is possible that a variety of beliefs about the leader role—or the game in which it was embedded—may have shifted as a function of this manipulation (e.g., the perceived status of the leader role, inferences about the level of physical activity involved in the game, expectations of how fun the game might be), in ways that influenced interest in the leader role in the opposite direction than we had expected. We return to this possibility in the General Discussion.

Older children in Study 3 anticipated lower social support than younger children, even though age was not associated with self-efficacy or interest in the leader role. We also found some gender differences that emerged regardless of how the leader role was framed and regardless of the presence of girl peer role models. Namely, girls anticipated more social support and stronger cooperation from other children than boys across conditions. Like the results of Study 2, these results are in contrast with those of Study 1, in which girls’ expectations for other children in a leader role were less positive than boys’ expectations. We speculate about these discrepancies in the General Discussion.

General Discussion

Whereas considerable attention has been devoted to understanding how adults think about leadership, the developmental origins of these attitudes remain underexamined. Adopting a developmental perspective has the potential to increase current knowledge about the roots of gender inequality in leadership positions (e.g., Heck et al., 2021). With this goal in mind, we conducted three studies to investigate children’s beliefs about leadership and gender. Although these studies explored this phenomenon from different perspectives (e.g., first vs. third person), they nevertheless complemented each other and provided a comprehensive examination of children’s gendered leadership cognitions across a wide age span: We examined 5- to 10-year-old children’s expectations for other leaders (Study 1) and for themselves as leaders (Studies 2 and 3). We examined children’s interest in leadership (Studies 2 and 3), as well as a series of associated constructs that are implicated in adult gender gaps in leadership, including self-efficacy (Studies 2 and 3; e.g., Fox & Lawless, 2011) and the anticipation of social support (Studies 1–3) and cooperation from others (Studies 2 and 3; e.g., Brescoll, 2011; Vial et al., 2016, 2018). We also examined whether children’s beliefs about leadership and gender are sensitive to contextual features that mark the leadership role as more (or less) communal (Study 3), in line with past investigations with adults (e.g., Belanger et al., 2020; Schneider et al., 2016). The results of the three studies, though nuanced and complex, indicate that gender gaps in children’s expectations for other leaders may start early, but girls themselves are no less interested in taking on a leadership role than boys, and emphasizing the communal aspects of leadership can increase interest in leadership among some children. The results—particularly those in Studies 2 and 3—also underscore the value of taking an intersectional perspective to understand children’s cognitions, in this case related to gender and leadership (e.g., Heck et al., 2021).

Gender Differences in Children’s Reasoning About and Interest in Leadership

We found some indication that girls may be more sensitive than boys to the possibility that children may suffer social repercussions for taking on a leader role. In Study 1, girls expected lower social support than boys for other leaders, regardless of leader gender and across a variety of naturalistic scenarios relevant to children’s everyday lives (e.g., playing at the park). However, we found no evidence of gender gaps in children’s own interest in a leadership role, contrasting with robust evidence among adult samples (Netchaeva et al., 2022). In fact, in Study 2, we found that interest in leadership appeared to increase with age among White girls (whereas interest did not change or decreased with age for other groups of children). Furthermore, we found no evidence suggesting a lower sense of self-efficacy in leadership among girls compared to boys—if anything, some of our results indicate that self-efficacy may be stronger in girls (Study 3), in contrast to the patterns that have been documented among adults (e.g., Fox & Lawless, 2011). Moreover, whereas girls were more likely to anticipate low social support for other children who took on a leader role (Study 1), we found no clear gender differences in children’s anticipation of how much others would support them or cooperate with them were they to be the leader (Study 2). In fact, in Study 3 we found that girls were more optimistic than boys when faced with the prospect of occupying a leader role—they anticipated more social support and more cooperation from others. As a whole, these patterns indicate that young girls between the ages of 5 and 10 may not anticipate the negative repercussions for behaving in leader-like ways that adult women often fear (e.g., Moss-Racusin & Rudman, 2010) and encounter (e.g., Eagly & Heilman, 2016), and that gender gaps in leadership interest may solidify later in life, perhaps in adolescence (e.g., Bos et al, 2022; Fox & Lawless, 2014). Thus, our results point to a complex developmental trajectory of gendered leadership cognitions, and to the possibility that middle childhood might be a key life stage to intervene to sustain girls’ interest in leadership. More optimistically, our results may also point to changing societal attitudes in the U.S. with respect to women’s role in leadership—changes that may now be visible only in the attitudes of the younger generations but will eventually translate into more equitable representation in leadership positions.

Importantly, children’s leadership cognitions and motivation to lead were sensitive to the way that leadership was framed. In Study 3, children anticipated more social support and cooperation from others and exhibited a stronger sense of self-efficacy as leaders when the leader role was framed in a way that would afford communal goals (e.g., helping others) compared to a more self-oriented, agentic framing. Children in Study 3 also expressed more interest in leadership when the leader role was framed as communal, an effect that was particularly evident among boys of color. Although more research is necessary to adjudicate whether emphasizing agency has a detrimental effect or whether emphasizing communality has a positive impact (or both), the findings in Study 3 clearly indicate that children find leadership more appealing when the leader role is aligned with communal (vs. agentic) goals.

Surprisingly, children in Study 3 expressed more interest in the leader role when other children who had been leaders were primarily boys than when boys and girls were equally represented among past leaders. One possible explanation for this unexpected finding is that the gender composition of previous leaders may have shaped children’s perceptions of the leader role in unanticipated ways. For example, in past studies, children associated higher prestige, status, or power with male-dominated roles compared to female-dominated roles, even when those roles were novel (Liben et al., 2001; Weisgram et al., 2010). Thus, it is possible that children in our study may have seen the leader role as more prestigious and therefore more appealing when most (vs. only half) of the previous leaders were boys. If so, these findings would suggest that the pro-male bias in perceptions of status and prestige that has been identified in children’s thinking about adult roles may begin very early and extend more generally to children’s activities—an intriguing notion that may be fruitful to examine in future research. However, other explanations seem plausible here as well. For instance, the gender composition of previous leaders may have shaped children’s inferences about what the game itself is: Perhaps children inferred that the game entails higher levels of physical activity or is easier when it is primarily boys who engage in it. Future investigations may explore these different psychological processes to shed light on the reasons why emphasizing male representation might make leader roles more appealing to children (including girls) at young ages. Ultimately, our results indicate that female role models may not always be effective in boosting girls’ interest in leadership, although additional research is necessary to draw stronger conclusions (see also Gladstone & Cimpian, 2021).

Age and Racial/Ethnic Differences in Children’s Reasoning About and Interest in Leadership

Children’s leadership cognitions and motivation to lead varied with age and race/ethnicity in complex ways. To start, our results indicate that children’s expectations about leadership may become less positive as they grow older. Compared to younger children, older children in Study 3 anticipated that others would support them less if they were the leader, suggesting that, with age, children may generally become wary of the potential social costs of leadership. This finding may inform interventions designed to increase leadership ambition in children regardless of gender. Interestingly, we found some evidence in Study 2 that some girls (i.e., White girls) may become more interested in leadership over time in the 5- to 10-year-old age span. Together with past findings showing gender gaps in leadership ambition among adolescents (Bos et al, 2022; Fox & Lawless, 2014) and adults (Netchaeva et al., 2022), these findings suggest that, perhaps at some point in middle school, girls’ leadership cognitions may begin to change, with girls gradually adopting the belief that leadership is not “for them.” It is possible that these changes may be due to an accumulation of negative personal experiences of backlash for behaving in explicitly dominant ways (e.g., Okimoto & Brescoll, 2010; Williams & Tiedens, 2016) and/or to continued exposure to cultural stereotypes equating leadership with boys and men (Aley & Hahn, 2020; Koenig et al., 2011), but additional research is necessary to investigate these possibilities.

We also explored whether children of color may differ from White children in their attitudes toward leadership, and whether these variations might further depend on child gender, given that the mismatch between agentic leadership and female stereotypes seems weaker for women and girls of color relative to White women and girls (e.g., Ghavami & Peplau, 2013; Rosette et al., 2016). A few results are worth highlighting on this topic. In Study 1, White children anticipated stronger social support for girl leaders than boy leaders, whereas children of color did not differentiate. (Notably, the hypothetical girl and boy leaders in Study 1 were all White themselves.) In Study 2, interest in the leader role appeared to increase with age for White girls, but not for White boys or for children of color. In Study 3, framing a leader role in communal terms boosted interest in the leader role particularly strongly among boys of color, and it increased self-efficacy as leaders among boys and girls of color but not among White children. Considered together, these findings may suggest that (a) early leadership cognitions and attitudes (at this point in time and in this region of the U.S.) tend to be particularly favorable toward and among White girls, which is also why (b) it is other groups (particularly, boys of color) who stand to benefit most from interventions to improve these cognitions and attitudes.

However, the overall picture that emerges from these findings is not necessarily as cohesive as stated above, and it is also noteworthy that our studies were not designed to investigate the moderating role of racial/ethnic background: Our sample sizes were often too small to allow strong conclusions about intersectional effects. Thus, caution is needed in interpreting the present findings. Nevertheless, our studies suggest that children from different racial/ethnic backgrounds vary in how they think about gender and leadership, and they underscore the need for additional research into children’s attitudes toward leadership from an intersectional perspective, as has been proposed by others (e.g., Heck et al., 2021). Future investigations can build on these preliminary intersectional findings, further documenting how leadership cognitions may vary along different axes of identity and developing theory to articulate when and why these differences may emerge.

Strengths and Contributions

Our findings add to a growing literature focused on understanding how children think of leadership and related constructs (e.g., Charafeddine et al., 2015; Gülgöz & Gelman, 2017; Reifen-Tagar & Cimpian, 2022; Terrizzi et al., 2019; Zhao & Kushnir, 2018), and—more directly—they contribute to recent scholarship on children’s beliefs about gender and leadership (e.g., Bos et al., 2022; Charafeddine et al., 2020; Mandalaywala et al., 2020; Neff et al., 2007; Patterson et al., 2019; Reyes-Jaquez & Koenig, 2022). Whereas previous studies have mostly examined children’s attitudes about adults who were in positions of authority or who occupied leadership roles (e.g., Ayman-Nolley & Ayman, 2005; Bigler et al., 2008; Bos et al., 2022; Neff et al., 2007; Patterson et al., 2019; Terrizzi et al., 2019), our studies focused on children’s beliefs and attitudes about leadership in peer contexts (Charafeddine et al., 2020; Gülgöz, 2015; Mandalaywala et al., 2020; Reyes-Jaquez & Koenig, 2022), where gender-group norms may be most salient (e.g., Martin & Fabes, 2021).

Importantly, our results add to growing evidence against gender gaps in leadership interest among young children (e.g., Bigler et al., 2008; Patterson et al., 2019; Reyes-Jaquez & Koenig, 2022; cf. Bos et al., 2022). We found little evidence of gender gaps in leadership interest in peer contexts even when leadership roles were framed in agentic terms and were thus inconsistent with stereotypical expectations for girls’ and women’s behavior. As discussed previously, these results suggest that—at least in the region of the U.S. where this research was conducted—gender gaps in leadership ambition may begin to take shape beyond the age range examined here, in adolescence. This possibility is worth investigating further.

Unlike most developmental research on this topic, we framed our investigation drawing from theories that have proven to be valuable tools for explaining adult gender gaps in leadership, including goal congruity theory (Diekman & Eagly, 2008) and lack of fit and role congruity theories (Eagly & Karau, 2002; Heilman, 1983, 2001). This approach inspired us to examine specific constructs relevant to leadership that, to our knowledge, had not been investigated in young children before, including the expectation of social support and cooperation from peers. Our studies demonstrate that it is possible to examine these constructs in age-appropriate ways, revealing that children consider the social costs of leadership (i.e., whether they will be supported by others) from an early age and that these beliefs are malleable: They change in predictable ways based on the context of leadership, such that children anticipate more support from others when leadership is framed in communal ways. Future investigations may employ similar methodologies to expand psychologists’ current understanding of the developmental origins of leadership cognitions. However, it should be noted that many of our findings stand in contrast with predictions drawn from adult-based theories. For example, we found that children in Study 1 did not expect more negative reactions to girl (vs. boy) leaders; that girls and boys were similarly interested in leadership roles, even when the roles were framed in agentic terms (Studies 2 and 3); and that girls overall appeared more optimistic than boys about others’ social support were they to be leaders. These patterns suggest that children at this young age may not be attuned to the potential incongruity between the female gender role and leadership roles, an intriguing possibility that highlights the need for further theoretical elaboration of existing adult theories on gender and leadership from a developmental perspective.

Finally, we note that our studies provide a comprehensive investigation of children’s leadership cognitions by approaching this topic from different vantage points—specifically, how children think of other leaders and how they think of themselves as leaders. The inclusion of these two different perspectives revealed mixed patterns that are worth pursuing further in future research. We found that girls expected other leaders to face negative social repercussions (Study 1), and yet, they did not expect any less social support as leaders relative to boys (Studies 2 and 3). It is possible that these initially-disjointed beliefs may start to align later on in development, perhaps as girls accumulate personal experiences with leader-like behavior (and others’ reactions to it). But regardless of the reason for these inconsistencies, they highlight the value of collecting data from different perspectives (e.g., self vs. others), an approach that future studies may find it useful to adopt as well.

Limitations and Future Research Directions

A limitation of the present studies is the narrow focus on a specific aspect of leadership—the ability to give orders, which is only one dimension of a multifaceted construct. A leader’s power and social influence involves more than issuing directives: For instance, they also have the freedom to make decisions independently from others (Gülgöz & Gelman, 2017). Future research may examine gender gaps in interest in leader roles across these different dimensions of leadership to arrive at a more complete understanding of the origins of gender differences in leadership ambition.

Although we identified age-related differences in children’s views of leadership (e.g., as they grow older, children’s leadership cognitions may become more negative), the age range in our studies did not cover the full span of childhood. Future research extending this line of work to older age groups (e.g., 11–15-year-olds) might be helpful to more precisely identify the point at which girls’ leadership cognitions and motivations begin to resemble those of adult women—for example, exhibiting uncertainty in their leadership abilities or expecting low levels of support and cooperation from others. Longitudinal studies are also necessary to draw stronger conclusions about developmental changes and to potentially illuminate the ebb and flow of children’s attitudes toward leadership.

Although we took advantage of the diversity of our samples to investigate racial/ethnic differences in leadership cognitions and attitudes, our studies were not designed from an intersectional perspective, nor were they sufficiently powered to detect nuanced intersectional patterns. Another clear limitation is that we treated children of color as a monolithic group. Although combining multiple racial/ethnic backgrounds into a single group allowed us to maximize statistical power, children from these different backgrounds may have meaningfully different leadership-related experiences (e.g., availability of role models among prominent leaders), which may in turn affect their gendered leadership cognitions and attitudes. Additionally, the stimuli in our studies included only White children, but it is possible that reactions to child leaders might depend not only on leader gender but also on leader race, as is the case for reactions to adult leaders (e.g., Rosette et al., 2016). These limitations notwithstanding, we hope that our findings pave the way for additional research on intersectional leadership cognitions in children. This work is necessary to fully understand when, why, and for whom leadership is or is not attractive (Heck et al., 2021), much as is the case among adults (e.g., Rosette et al., 2016).

Practice Implications

Our findings highlight the potential value in early interventions seeking to nurture children’s leadership ambitions. As such, they have practical implications for educators and parents seeking to encourage children (including young girls) to pursue leadership roles in the future. The present studies suggest that gendered leadership cognitions are unstable and malleable between the ages of 5 and 10, and that the way in which leadership is discussed around children (e.g., in the classroom, in educational materials, in children’s media, or at home) may influence children’s attitudes toward leader roles, with potentially meaningful consequences for gender equality in the long run (e.g., Martin & Fabes, 2021). Specifically, emphasizing the communal aspects of leadership, which are known to characterize effective leaders (e.g., Gartzia & van Knippenberg, 2016; Gerzema & D’Antonio, 2017), could be a powerful way to encourage children to actively seek out roles that allow them to develop confidence and interest in leadership. Such interventions may also curtail the development of negative social expectations about leadership as children age, expectations that we observed in Study 3. Thus, our findings underscore the importance of challenging the common view that agency is central to leadership and communality is peripheral or secondary (Koenig et al., 2011; Vial & Napier, 2018). It is particularly important to change this view among educators and creators of educational and entertainment media targeting children, so that they may cultivate more communal leadership cognitions in younger generations.

Conclusion

Our studies illustrate the utility of adopting a developmental lens to study leadership cognitions and gender gaps in interest in leadership. Taken together, our results indicate that gender gaps in leadership interest, if they still exist in the cultural context where this research was conducted, may not begin until after 10 years of age. Although young girls seem to expect worse social outcomes for other leaders relative to boys, they also appear more optimistic than boys about their own prospects as leaders. Nevertheless, we found evidence that children’s expectations about leaders and leadership generally become more negative with age. Framing leadership in more communal, altruistic ways could be an effective way to encourage children’s participation in activities that allow them to develop leadership skills, regardless of their gender or racial/ethnic background.