In the Face of Opportunities: Facial Structures of Scientists Shape Expectations of STEM Environments

Abstract

Impressions of role leaders provide information about anticipated opportunities in a role, and these perceptions can influence attitudes about science, technology, engineering, and mathematics (STEM) pathways. Specifically, the facial structures of role leaders influenced perceived affordances of working with that person, such as the availability of communal and agentic opportunities (e.g., mentorship; achievement). STEM faculty with trustworthy (relative to dominant) faces were seen as valuing communal goals (Studies 1–3), and in turn, perceived as affording both communal and agentic opportunities in their research groups (Studies 2–3b). These heightened goal opportunities aligned with perceptions that trustworthy-faced advisors would enact more group-supportive behaviors (Study 2). Consequently, students anticipated fairer treatment and reported greater interest in labs directed by trustworthy- than dominant-faced leaders (Studies 3a–4a), even when images were accompanied by explicit information about leaders’ collaborative behavior (Study 4b). The faces of leaders can thus function as the “face” of that role and the surrounding culture.

Keywords

goals communality face perception STEM organizational trust

Decisions to enter particular social roles pervade people’s lives: Should I declare this major? Should I apply for this job, or should I leave my current position? When making these decisions about role transitions, information about what the new opportunity will be like is essential, and individuals seek out signals about whether this new role fits their values and goals. Some of this information might be explicit and overt, such as a job description or the list of required courses for a major. Yet much of the information is not stated directly but instead must be gleaned by observation or experience: Do people work collaboratively in this department? Does leadership support the growth of people on their team? We contend that such inferences can be informed by subtle cues in the environment—even as subtle as the facial structure of people in leadership. To answer questions about whether a particular role will afford a range of goal opportunities, decision makers use inferences about the motivation of leadership in that role, and in turn, projections about what their own experience in that role will be like.

These inferences about goal opportunities in a role are particularly pertinent in understanding how individuals navigate science, technology, engineering, and mathematics (STEM) pathways. Needs to elevate U.S. student engagement in science and technology are acute. In recent years, U.S. universities produced fewer science and technology bachelor degrees compared with universities of other nations, such as China (National Science Board, 2020). Gains in the number of U.S. STEM degrees awarded can also benefit individual degree holders because the STEM workforce provides high starting salaries and employment opportunities (Pew Research Center, 2018). Understanding the cognitive and motivational factors that inhibit entry into STEM pathways can thus provide benefits both to society and to individuals.

A known deterrent to STEM entry is the belief that STEM careers, relative to other careers, fail to afford communal goals, especially among communally oriented students (Diekman et al., 2010). Communal goals include fundamental motives for relatedness and prosociality. These communal opportunities are perceived as lacking in STEM pathways, even as agentic opportunities (i.e., self-oriented opportunities) are perceived as available (Diekman et al., 2020). Given the consensual perception of fewer communal opportunities in STEM fields, highly communally oriented individuals report less interest in STEM careers. Consequently, highlighting communal opportunities in STEM can heighten interest in these roles. For example, underrepresented minority students in biosciences reported greater motivation in science when they were embedded in research groups where peers believed that science affords prosocial goals (Thoman et al., 2017). College students who reported engaging in communal experiences in STEM reported perceptions that STEM careers provided communal opportunities, aligning with greater interest in STEM (Brown et al., 2018). The presence and benefit of these communal STEM microcultures thus lead to the question of how students might form inferences about the affordances of different STEM experiences.

Extant research finds that students form attitudes about entering fields based on cues available in the environment (Murphy et al., 2007). For example, the presence of objects that signify “geeky masculinity” (e.g., Star Wars/Star Trek items and video games) diminishes women’s sense of belonging or interest in computer science (Cheryan et al., 2009). Perceived affordances of STEM fields can also come from cues about particular environments, such as descriptions of scientists’ collaborative behavior (e.g., Diekman et al., 2011), the service-learning content of course descriptions (Belanger et al., 2017), or prosocially oriented descriptions of research (Brown et al., 2015). As individuals consider entering roles, they scan the surrounding environment for cues about what opportunities that role will offer. We propose that even the faces of potential STEM faculty advisors can embody messages about the communal or agentic opportunities afforded by being a member of their lab.

We propose that leaders within roles serve as the “face” of that role, and as such provide a route to further inferences about the nature of that occupation. Viewing people in occupational contexts leads to fairly consistent inferences about the occupational title (Child, 1936; Gahagan, 1933). Indeed, facial cues influence the attributions people make about others, such as forming judgments about selecting a leader (Olivola et al., 2014; Todorov et al., 2015). Extending this evidence, we propose that such inferences about leaders within a context will also inform beliefs about the experiences in the context itself. For example, do new graduate students believe science is a cutthroat endeavor or one in which scientists work toward a common goal of shared knowledge? The scientific exemplars that students see likely shape their beliefs about the scientist’s role. If true, the implications are important. This finding would support the premise that beliefs about roles and beliefs about role occupants are linked. Roles can shape inferences about individuals and individuals can shape inferences about role experiences. Furthermore, these inferences about the role can shape interest in entering the role.

Here, we investigate how inferences about leaders relate to inferences about role opportunities by focusing on signals from leaders’ faces. The faces of scientists may provide cues that shape expectations of ones’ role experiences. Scientists who are perceived as attractive garner more interest in their research (Bi et al., 2020; Gheorghiu et al., 2017, 2020). Building upon this evidence that impressions of science relate to scientists’ facial features, we ask the novel question of how the facial structure of scientists shapes expectations of experiences in STEM roles (and, in turn, interest in entering such roles). Next, we detail how beliefs about what the role will be like might be cued from facial signals of trustworthiness and dominance.

Trustworthiness and Dominance

We posit that a role occupant’s facial signals of trustworthiness and dominance provide cues not simply about that individual but about the broader role. We build these predictions based on literature about inferences of individuals, given that this is the majority of the existing evidence. The two dimensions of trustworthiness and dominance account for approximately 80% of the variance in judgments of faces (Oosterhof & Todorov, 2008) and are inversely related (Kleisner et al., 2013). Trustworthiness is often inferred from high inner eyebrows and pronounced cheekbones (Todorov et al., 2008), and dominance is often inferred from low brows, large chin, and a more angular face (Keating, 1985; Zebrowitz & Collins, 1997). Facial judgments of trustworthiness are associated with perceived warmth and predict approach responses, whereas facial judgments of dominance are associated with perceived power, strength, and signal capability to act (Oosterhof & Todorov, 2009).

Perceivers often use these facial signals as an early cue in social interactions. Even brief visual perception can influence behavioral intentions (Ambady et al., 1999). For instance, facial judgments of trustworthiness and dominance of political candidates positively predicted intent to vote for those candidates (Little et al., 2012). During times of conflict, more dominant-faced people were preferred for leadership positions (Ferguson et al., 2019). Dominant faces are also motivationally salient, eliciting attention to highly dominant faces (Wang et al., 2016). Conversely, people with more trustworthy-faced partners anticipated cooperation and invested more in their partners (Stirrat & Perrett, 2010). In investment games, trustworthy-faced individuals were more often selected as partners for risky scenarios and decisions to cooperate (Krumhuber et al., 2007). Indeed, even when a partner’s past trustworthy behavior is known, facial trustworthiness predicted the decision to invest in a partner in the economic trust game (Rezlescu et al., 2012).

These trustworthiness and dominance judgments align with broader judgments about motives to benefit the self or others. These fundamental dimensions of agency and communality organize a wealth of other- and self-related information (Abele & Wojciszke, 2014). Impressions of communality reflect the positive or negative intentions of other people (aligning with trustworthiness), and impressions of agency generally reflect the capability of acting on those intentions (aligning with dominance; Fiske et al., 2002). Thus, we propose that trustworthy and dominant facial structures will lead to inferences of others’ goal orientations (that is, inferred goal endorsement). We hypothesize that role leaders with trustworthy facial structures will be perceived as endorsing communal goals (i.e., helping others), whereas role leaders with dominant facial structures will be perceived as endorsing agentic goals (i.e., advancing the self).

To an outsider considering entering a social role, such judgments of a leader’s motives can be critical information in forming beliefs or attitudes about the social role. The novel question guiding the current research is whether inferences move beyond judgments of a leader’s motives to form impressions of the roles that the leader constructs. We anticipate that perceptions of a leader’s motives will translate into perceptions of what the experience of working with that leader will be like. In other words, key impressions are formed not only about individual leaders but also about the environments that they occupy.

Cues About Affordances

Do inferences from a leader’s facial structure cue inferences about the context that the leader occupies? If so, then these signals can affect projections about the experience in a particular social role and decisions about whether to enter that role. We posit that a leader’s facial structure will shape impressions of the opportunities available in STEM environments. When considering the opportunities available in a role, the facial structure of a leader leads to inferences about what goals that person endorses, which align with expectations about what goals are afforded in the role.

We anticipate that inferences about goal affordances will follow one of the two patterns (see Figure 1). On one hand, perceived affordances might align directly with face-based inferences of a leader’s goal orientation. In this matching hypothesis, a trustworthy face may signal a communally oriented leader who would offer communal opportunities (but not agentic opportunities), whereas a dominant face may signal an agentically oriented leader who would offer agentic opportunities (but not communal opportunities). For example, believing a STEM faculty member values prosocial goals would foster the impression that working in that person’s research group will involve opportunities to benefit the community but not necessarily opportunities to achieve. Similarly, believing a STEM faculty member values agentic goals would foster inferences that working in this person’s research group will involve opportunities to achieve but perhaps not opportunities to benefit the community.

Figure 1.

Two proposed patterns of face to role inferences.

On the other hand, trustworthy facial structures might cue beliefs that a range of goals could be afforded because facial trustworthiness fosters inferences of the communal orientation to support others. In this support hypothesis (Figure 1, bottom panel), a communally oriented leader would foster impressions that the role offers a range of goal opportunities, both communal and agentic. Insofar as a communal leader seeks to support their students’ professional goals—both agentic and communal—inferring communality from a trustworthy (vs. dominant) faced leader may foster the impression that working in this person’s research group will involve a wide range of goal opportunities. Supporting this, students who prefer communal advisors also perceived the communal advisor to fulfill more communal and agentic goals (Fuesting & Diekman, 2017). To this point, considering STEM faculty who endorse growth mindsets foster perceptions that those instructors’ classrooms as affording both communal and agentic opportunities (Fuesting et al., 2019). Furthermore, successful leaders are expected to provide a structure that satisfies subordinates (Schriesheim et al., 1995). STEM faculty faces that signal communality may thus be perceived as structuring environments that support their students’ goals across a wide range of content. These inferred communal STEM faculty will be expected to behave in ways that benefit the group, which may foster expectations that this individual will support a range of goals valued by their subordinates.

Overview of the Current Research

The present work investigates whether perceivers infer affordances within a particular role from subtle variations in leaders’ facial structure. Using both naturalistically varying and manipulated facial structures, we test two competing hypotheses. A “matching hypothesis” suggests that trustworthy (vs. dominant) facial structures of leaders distinctly cue communal affordances, whereas dominant (vs. trustworthy) facial structures of leaders distinctly cue agentic affordances. Conversely, a “support hypothesis” predicts that trustworthy (vs. dominant) facial structures signal leaders’ communal goal endorsement, which leads to inferences of both communal and agentic affordances, because such leaders support their team members’ goal pursuits. According to this hypothesis, trustworthy faces compared with dominant faces would foster perceptions that the leader will support a wide range of communal and agentic goals.

As depicted in Figure 2, the current studies examine face-based expectations of goal opportunities.

Figure 2.

Conceptual progression of current studies.

In our Preliminary Study, we investigated whether face type (trustworthy vs. dominant faces) influences the inferred communal versus agentic orientation of a leader. Next, Studies 1a and 1b test the key hypothesis that perceivers infer role affordances from inferences about leaders, with findings consistent with the Support Hypothesis. Study 2 examines whether effects are mediated by expectations that the leader will engage differentially in group-supportive behaviors. Next, we examine downstream consequences and test boundary conditions: Studies 3a and 3b examine the effect of leader facial structure on organizational trust. Studies 4a and 4b test the ecological validity of the leader face effect: Does the face effect emerge among real STEM faculty and even in the presence of direct information about leaders’ collaborative behavior?

Preliminary Study: Facial Structures Signal Goal Endorsement

As an initial step, we tested the foundational assumption that the facial structure of role leaders influences perceptions of the goals they endorse. Put simply, do trustworthy (relative to dominant) faced leaders appear to value more communal and/or more agentic goals?

Here, 55 participants were recruited from Mechanical Turk (MTurk) and each compensated US$0.50 for their time. An additional 20 respondents were excluded for failed attention checks. Participants were mostly White (61.8%, N = 34) and male (58.2%, N = 32) with an average age of 36. For a repeated-measures 2 (face type) × 2 (target gender) × 2 (goal type) interaction (power = 0.80; α = .05, N = 55), sensitivity analyses indicated a detectable effect size of f = .14.¹

Experimental Task

The preliminary study employed manipulated facial structure stimuli, drawn from the Basel Face Database (BFD) developed by Walker and colleagues (2018). Targets included eight White target identities whose faces were manipulated to appear high in trustworthiness or high in dominance, for a total of 16 target faces (see Table 1, top panel). To avoid identity repetition, we created two counterbalancing blocks, each of which showed only the trustworthy or dominant version of a given face identity. Target stimuli included both male and female individuals; prior work found female leaders cued greater opportunities in a role and greater trust (Joshi & Diekman, 2021), but the effects of facial structure documented in the current article appear to operate independently of the target gender.

Table 1.

Facial Stimuli Examples.

Facial Structure Stimuli Across Studies	Trustworthy faces	Dominant faces
Manipulated Facial Structure Preliminary study; Studies 1a, 2, 3a–3b

Naturally Varying Facial Structure Studies 1b, 4b

Note. Manipulated facial stimuli were drawn from the Basel Face Database. Naturally varying facial stimuli for Studies 1b and 4b were drawn from the Chicago Face Database, and for Study 4a from a new sample of publicly available images of STEM faculty collected for this research (not pictured). STEM = science, technology, engineering, and mathematics.

Participants were randomly assigned to one of the two blocks each consisting of eight unique face identities in a 2 (target gender: female, male) × 2 (face type: trustworthy, dominant) within-subjects design. Faces in each block were presented in random order.

Participants viewed each face individually along with a short description (e.g., “Dr. X is a physicist[engineer] at the University of Y. He[She] went to the University of Z to earn his[her] doctoral degree in physics[engineering]”). Descriptions and faces were counterbalanced. Participants rated how much each target personally valued communal and agentic goals, along with other measures reported within each study.

Inferred communal goal endorsement items included Dr. X values helping society, helping other scientists and science in general, mentoring others, and attending to others’ needs. Inferred agentic goal endorsement items were Dr. X values attaining power as a scientist, gaining financial rewards, achievement, and self-promotion. Items were adapted from Diekman et al. (2011) and were averaged within goal type to form indices (αs > .90).

Results

Initial analyses did not detect variation by face block; reported analyses collapsed across blocks. Data were submitted to a 2 (face type: trustworthy, dominant) × 2 (target gender: female, male) × 2 (goal: communal, agentic) repeated-measures analysis of variance (ANOVA).²

As presented in Figure 3 (left panel), the predicted Face Type × Goal interaction emerged, F(1, 54) = 6.13, p = .02, $η_{p}^{2} = . 10$ , 95% confidence interval (CI): [.003, .26]. Leaders with trustworthy versus dominant facial structures were perceived as endorsing more communal goals, F(1, 54) = 7.58, p = .008, d = 0.75. In contrast, leaders with trustworthy and dominant facial structures were not perceived to differ significantly in endorsing agentic goals, F(1, 54) = 2.58, p = .11, d = 0.43. A STEM research advisor’s facial structure influenced perceptions of how much they valued communal goals but not agentic goals within the research role. Critically, the target gender did not systematically interact with facial structure. Thus, for both female and male STEM research advisors, their facial structure signaled their presumed communal orientation.³

Figure 3.

Inferred goal endorsement by face type.

Summary

The preliminary analysis indicates that variations in STEM mentors’ facial structure led to inferences that leaders endorsed different goals. More trustworthy facial structures cued the perception that the leader endorsed more communal goals, yet dominant facial structures did not signal greater agentic goal endorsement. Building on this evidence, we turn to the novel questions of this research: Does the facial structure of a leader function as a cue to perceivers about the content of roles?

Studies 1a and 1b: Investigating Support Versus Matching Hypotheses

Studies 1a and 1b examined the novel question of how leaders’ facial structures influence perceived opportunities for others to fulfill communal and agentic goals in that context. Specifically, we test two competing hypotheses: role leaders with trustworthy facial structures (a) cue both communal and agentic opportunities, reflecting that trustworthiness signals support of others’ goals, whether agentic or communal (“support hypothesis”) or (b) cue distinctly communal opportunities, suggesting a matching effect between a communal cue/communal opportunities or agentic cue/agentic opportunities (“matching hypothesis”). Study 1a employs stimuli with manipulated facial structure and Study 1b employs stimuli with naturalistically varying facial structure.

Method

Participants

In Study 1a, 92 MTurk participants were compensated US$1.00 for their time. Participants tended to identify as White (65.2%, N = 60) and male (57.6%, N = 53) with an average age of 35. An additional 28 participants were excluded due to failed attention checks. Sensitivity analyses indicated a detectable effect size of f = .10 for a repeated-measures 2 × 2 × 2 interaction (power= 0.80; α = .05, N = 92).

In Study 1b, 83 MTurk participants were compensated US$0.50 for their time. Participants tended to identify as White (72.3%, N = 60) and female (56.6%, N = 47) with an average age of 37. An additional 24 participants were excluded due to failed attention checks. Sensitivity analyses indicated a detectable effect size of f = .14 for a repeated-measures 2 × 2 interaction (power = 0.80; α = .05, N = 83).

Procedure

Study 1a participants imagined engaging in a research group led by the target individual and reported the likelihood of goal opportunities in that role (i.e., perceived goal affordances). Study 1a facial stimuli and goal endorsement measures were identical to the preliminary study.

Study 1b followed the same procedure with facial stimuli that varied naturalistically in trustworthiness versus dominance, drawn from the Chicago Face Database (CFD; Ma et al., 2015; see Table 1, bottom panel). Four male face images were selected to have high ratings on trustworthiness and low ratings on dominance, and four were selected for high dominance/low trustworthiness, based on CFD trait data. Using face as the unit of analysis, the group of trustworthy faces was rated higher in trustworthiness than the group of dominant faces, t(6) = 6.25, p = .001, and the group of dominant faces was rated higher in dominance than the trustworthy faces, t(6) = −5.00, p = .002.

Measures

Inferred goal endorsement

Items were identical to the preliminary study (agentic αs ≥ .95; communal αs ≥ .95).

Perceived goal affordances

Participants imagined working in a research group led by the target individual and the opportunities available in that faculty member’s lab. Participants rated how much working with this person would provide communal goal opportunities (e.g., work with or collaborate with others; conduct research that benefits others; form connections with others; increase your affiliation with others; αs ≥ .95) and agentic goal opportunities (i.e., competence, skills, deeper understanding of the course or research materials, and success; αs > .96; adapted from Fuesting & Diekman, 2017). Items were rated from 1 (Not at all) to 7 (Extremely) and averaged to form indices within each goal type.

Results

For Study 1a, independent t tests revealed no significant differences between face blocks (ps > .30); reported analyses collapsed across blocks. The core analysis was a 2 (face type) × 2 (goal) repeated-measures ANOVA, with target gender as a between-subjects factor in Study 1a (1b included only male faces).

Perceived goal affordances

Results were consistent with the support hypothesis: Trustworthy faces (M_{Study 1a =} 5.39, SD = 0.79; M_{Study 1b =} 5.29, SD = 0.78) compared with dominant faces (M_{Study 1a =} 5.21, SD = 0.91; M_{Study 1b =} 5.01, SD = 0.92) were perceived as affording both communal and agentic goal opportunities working with that person, Study 1a: F(1, 91) = 15.80, p < .001, $η_{p}^{2} = . 15$ , 95% CI [.04, .28]; Study 1b: F(1, 82) = 17.34, p < .001, $η_{p}^{2} = . 18$ , 95% CI [.05, .31]. The main effect of facial structure was not moderated by goal affordance type in either study: Study 1a: F(1, 91) < .001, p > .99; Study 1b: F(1, 82) = 0.15, p > .70. In sum, trustworthy faces convey perceptions of both communal and agentic goal opportunities more than dominant faces do.

Consistent with STEM domain stereotypes, targets were perceived to afford more agentic opportunities (M_{Study 1a =} 5.41, SD = 0.83; M_{Study 1b =} 5.31, SD = 0.83) than communal opportunities (M_{Study 1a =} 5.20, SD = 0.89; M_{Study 1b =} 4.98, SD = 0.84), Study 1a: F(1, 91) = 17.52, p < .001, $η_{p}^{2} = . 16$ , 95% CI [.05, .29]; Study 1b: F(1, 82) = 36.45, p < .001, $η_{p}^{2} = . 31$ , 95% CI [.15, .44]. In Study 1a, female targets (M = 5.40, SD = .88) compared with male targets (M = 5.20, SD = 0.84) were perceived as affording more goal opportunities, F(1, 91) = 17.42, p < .001, $η_{p}^{2} = . 16$ , 95% CI [.05, .29].

Face type, goal endorsement, and perceived goal opportunities

We tested inferred communal goal endorsement as a potential mediator between face type and perceived affordances using within-subjects mediational analyses (Montoya & Hayes, 2017). As shown in Figure 4, the indirect effect of face type through inferred communal goal endorsement occurred for both communal goal affordances and agentic goal affordances. These patterns are consistent with a causal sequence in which trustworthy versus dominant faces of faculty signal that they endorse communal goals, which aligns with anticipating both communal and agentic opportunities for lab members.

Figure 4.

Mediational models of face type, inferred goal endorsement, and goal affordances (studies 1a and 1b).

Discussion

Through subtle variations in the facial structure of a role occupant, participants inferred what opportunities would be present in a role. Trustworthy faces compared with dominant faces signaled a leader’s presumed communal orientation, and in turn, that this leader’s lab would provide both agentic and communal goal opportunities—that it is a place both to develop skills and to connect with others. The data did not support a “matching” hypothesis in which trustworthy-faced leaders were thought to distinctly provide communal opportunities and dominant-faced leaders agentic opportunities. Instead, trustworthy-faced, relative to dominant-faced, leaders were anticipated to provide a range of goal opportunities, both communal and agentic. Furthermore, these facial structure effects were robust across different forms of facial stimuli—whether manipulated or naturalistic variation in facial structure, and whether female or male faces. Next, we examine the process by which the facial structures of leaders might cue a range of goal opportunities through suggesting different anticipated behaviors.

Study 2: Faces May Signal Goal Opportunities Through Group-Supportive Behaviors

We next turned to examine a potential mechanism for goal-to-role inferences via expectations about leaders’ behaviors. Put simply, why might trustworthy (relative to dominant) lab leaders appear to afford both communal and agentic goals? In line with the support hypothesis, we hypothesized that participants would judge leaders with trustworthy (relative to dominant) facial structures as more likely to dedicate time and effort to support students’ goals. In turn, that support might be seen to foster both agentic and communal goal opportunities. For example, STEM faculty who spend more time mentoring and meeting with their lab members may be inferred to provide their students social support and collaboration opportunity (i.e., communal affordances) as well as achievement or competence mentoring (i.e., agentic affordances). Thus, Study 2 examines whether supportive behaviors are expected of a trustworthy (vs. dominant) faced leader in a STEM lab and whether these behaviors are associated with heightened goal affordances.

Method

Participants

MTurk participants (N = 126) were compensated US$0.75 for their time. The sample was predominantly White (76.7%, N = 92) and male (55.4%, N = 67) with an average age of 35. An additional 23 participants were excluded for missing data. Sensitivity analyses indicated a detectable effect size of f = .10 for a repeated-measures 2 × 2 × 2 interaction (power = 0.80; α = .05, N = 126).

Procedure

Like prior studies, participants viewed STEM faculty and rated each target’s inferred goal endorsement, group-supportive behaviors, and the goal affordances of the lab.

Measures

Participants completed measures identical to prior studies for inferred goal endorsement (αs ≥ .96) and perceived goal affordances (αs ≥ .97).

Group-supportive behavior

For each face, participants were instructed to imagine working in a research group supervised by that person. Participants rated their agreement with the following statements (adapted from Dansky, 1996; Ellemers et al., 1998): This person makes many sacrifices as are necessary for the research group; spends more time mentoring the research group than on their own career; does not devote time to the research group (reverse-coded); meets with the research group often. Items were averaged (αs = .95).

Results

Independent t-tests revealed no significant differences between face blocks on any of the dependent measures (ps > .3); analyses reported below collapsed across blocks. The core analysis was a 2 (face type: trustworthy, dominant) × 2 (target gender) within-subjects ANOVA; analyses of affordances included goal type as a within-subjects factor.

Perceived goal affordances

Replicating Studies 1a and 1b, more goal opportunities were anticipated when working with faculty with trustworthy faces (M = 5.09, SD = 1.03) than dominant faces (M = 4.98, SD = 1.04), F(1, 125) = 6.91, p = .01, $η_{p}^{2} = . 05$ , 95% CI [.003, .14]. In addition, more goal opportunities were perceived when working with female faculty (M = 5.13, SD = 1.02) than male faculty (M = 4.94, SD = 1.05), F(1, 125) = 21.07, p < .001, $η_{p}^{2} = . 14$ , 95% CI [.05, .26].

Group-supportive behavior

As predicted, group-supportive behaviors were anticipated more from faculty with trustworthy faces (M = 4.93, SD = 0.88) than dominant faces (M = 4.79, SD = 0.91), F(1, 125) = 10.71, p = .001, $η_{p}^{2} = . 08$ , 95% CI [.01, .18].

Face type, goal affordances, endorsement, and group-supportive behaviors

As shown in Figure 5, we tested serial mediation models with inferred communal goal endorsement and group-supportive behavior mediating the face structure relationship with communal affordances (Panel A) and agentic affordances (Panel B). In both models, leaders with trustworthy faces (vs. dominant faces) were inferred to endorse more communal goals, and this inference was linked to beliefs that the leader would enact group-supportive behaviors. In turn, anticipating these behaviors predicted perceiving more goal opportunities in the research group. Indirect effects were significant for both communal goal affordances and agentic goal affordances. These data are consistent with a model in which subtle variations in facial structure initiate a sequence of cognitions about the broader goal opportunities facilitated or impeded by that leader.

Figure 5.

Inferred communal goal endorsement and group-supportive behaviors mediating the face structure relationship with communal affordances (Panel A) and agentic affordances (Panel B).

Discussion

Variability in facial structures led to differential anticipation of specific behaviors that support group processes. A leader with a trustworthy facial structure, compared with a dominant facial structure, was projected to support the research group through behaviors such as dedicating time and effort. Facial structure, above and beyond leader gender, led to inferences about the leader’s own goals and behaviors, which in turn shaped beliefs about what this lab role will be like. We then turned to examine the consequences of face to role judgments: Is facial structure a cue to fairness?

Studies 3a and 3b: Faces Signal Fairness Through Communal Opportunities

Studies 1a and 2 found that faculty with trustworthy (compared with dominant) facial structures were perceived as leading labs that afford both communal and agentic goals. Next, we examine an additional downstream consequence: namely, a sense of fairness within that context. That is, do STEM faculty with trustworthy faces lead people to anticipate that the organization will value them (i.e., organizational trust)? Organizational trust relates to forming relationships with other members to accomplish shared goals (Fukuyama, 1996) and consists of positive expectations of the intents of organizational members (Shockley-Zalabak et al., 2000). Studies 3a and 3b examined whether leader facial structure signals to the perceiver that they will receive fair treatment and whether that signal particularly operates through perceptions of communal goal affordances. The perception of communal opportunity may especially align with beliefs that others in the organization have positive intent and will treat others fairly. This pattern would be consistent with evidence that perceptions of organizational trust in STEM labs and companies were predicted by anticipated communal affordances but not agentic affordances (Joshi & Diekman, 2021). Similarly, the current studies examine whether facial structure cues expectations of fairness in STEM through perceived communal affordances.

Moreover, signals of fairness in STEM labs may matter especially for students in the STEM pathway. For instance, STEM majors may be especially likely to weigh the facial cues of STEM faculty. We examined whether STEM majors compared with non-STEM majors especially value facial cues to fairness in STEM labs by combining Study 3a and 3b data sets for sufficient power. This analysis was suggested and preregistered after data collection; other aspects of this study were not preregistered (https://osf.io/qpx36).

Method

Participants

Study 3a included 121 MTurk participants, compensated US$0.50 for their time. Participants tended to identify as White (64.5%, N = 78) and female (59.5%, N = 72) with an average age of 39 years. An additional 24 participants were excluded due to failed attention checks. Sensitivity analyses indicated a detectable effect size of f = .10 for a repeated-measures 2 × 2 interaction (power = 0.80; α = .05, N = 121).

Study 3b recruited 156 participants from the introductory psychology pool at a Midwestern university for partial course credit. Participants tended to identify as White (65.4%, N = 102) and female (50.6%, N = 79) with an average age of 19 years. An additional three participants were excluded due to failed attention checks. Sensitivity analyses indicated a detectable effect size of f = .10 for a repeated-measure 2 × 2 interaction (power = 0.80; α = .05, N = 156).

Analysis preregistration plan

As specified in the analysis preregistration plan for the combined analyses, we aggregated St

udy 3a and 3b data sets to attain sufficient power to examine effects by STEM major/nonmajor status. We sought to detect a medium-sized effect for a 2 (face type) × 2 (goal type) × 2 (major type) mixed-model experimental design. This power analysis revealed a minimum sample size of 100 people for a within-between interaction to detect the minimum effect of f = .33; α = .05, power = 0.80. The aggregated data set included 61 STEM and 183 non-STEM majors.

Procedure

Like prior studies, participants viewed brief information about each leader, imagined working in that lab, and provided ratings of the leader and the lab. Studies 3a and 3b presented four male faces from the Basel Face Database.

Measures

Perceived goal affordances

These items were identical to prior studies (α’s>.90).

Organizational trust

Participants completed a measure of organizational trust (adapted from Emerson & Murphy, 2015; Purdie-Vaughns et al., 2008). Participants rated their agreement with the following statements: I would be treated fairly by Dr. X in this research lab; I would trust Dr. X to treat me fairly; I would be willing to put in extra effort if Dr. X asked me to; My values and the values of this research lab are very similar; I would like to work at a research lab that has similar values as those of this research lab. Items were rated on scales from 1 (Strongly Disagree) to 7 (Strongly Agree) and averaged to form an index (αs > .95).

Results

Independent t-tests revealed no systematic differences between face blocks (ps > .2); analyses collapsed across blocks.⁴ Within each measure, we first present analyses performed within each study; we then test differences by STEM major following preregistered analyses conducted on a combined data set.

Perceived goal affordances

Data from Studies 3a and 3b were separately submitted to 2 (face type: trustworthy, dominant) × 2 (goal: communal, agentic) repeated-measures ANOVAs. Replicating previous studies, trustworthy faces (M_{Study 3a} = 5.06, SD = 0.97; M_{Study 3b} = 5.24, SD = 0.86) were perceived to afford more opportunities than dominant faces (M_{Study 3a} = 4.83, SD = 1.13; M_{Study 3b} = 4.97, SD = 0.98), Study 3a: F(1, 120) = 10.13, p = .002, $η_{p}^{2} = . 08$ , 95% CI [.01, .18]; Study 3b: F(1, 155) = 22.64, p < .001, $η_{p}^{2} = . 13$ , 95% CI [.05, .23]. Consistent with STEM stereotypes, participants perceived more agentic goal opportunities (M_{Study 3a} = 5.05, SD = 1.03; M_{Study 3b} = 5.29, SD = 0.91) than communal goal opportunities (M_{Study 3a} = 4.83, SD = 0.96; M_{Study 3b} = 4.92, SD = 0.86), Study 3a: F(1, 120) = 24.03, p < .001, $η_{p}^{2} = . 17$ , 95% CI [.06, .28]; Study 3b: F(1, 155) = 87.32, p < .001, $η_{p}^{2} = . 36$ , 95% CI [.24, .46].

The main effect of face type was moderated by goal, as reflected in the Face Type × Goal interaction, Study 3a: F(1, 120) = 7.10, p = .009, $η_{p}^{2} = . 06$ , 95% CI [.003, .15]; Study 3b: F(1, 155) = 8.14, p = .005, $η_{p}^{2} = . 05$ , 95% CI [.005, .13]. More communal opportunities were perceived when working with faculty with trustworthy faces (M_{Study 3a} = 4.99, SD = 0.99; M_{Study 3b} = 5.10, SD = 0.91) than dominant faces (M_{Study 3a} = 4.68, SD = 1.13; M_{Study 3b} = 4.73, SD = 1.03), Study 3a: F(1, 120) = 14.47, p < .001, d = 0.70; Study 3b: F(1, 155) = 25.16, p < .001, d = 0.81. Significant but weaker effects were detected for agentic goal opportunities: More agentic opportunities were perceived when working with faculty with trustworthy faces (M_{Study 3a} = 5.13, SD = 1.04; M_{Study 3b} = 5.38, SD = 0.92) than dominant faces (M_{Study 3a} = 4.97, SD = 1.18; M_{Study 3b} = 5.20, SD = 1.03), Study 3a: F(1, 120) = 4.40, p = .038, d = 0.38; Study 3b: F(1, 155) = 9.35, p = .003, d = 0.49. The pattern of the interaction is consistent with the support hypothesis: Trustworthy faces signaled higher levels of both communal and agentic goal opportunities, although more strongly for communal opportunities.

STEM major moderation analysis

As planned, the combined data set was submitted to a 2 (face type: trustworthy, dominant) × 2 (goal: communal, agentic) × 2 (major type: STEM, non-STEM) mixed-model ANOVA. Analyses revealed no main effect of major type, F(1, 242) = 0.10, p = .751, $η_{p}^{2} < . 01$ , 95% CI [0, .02]: STEM majors (M = 5.02, SD = 1.80) did not differ from non-STEM majors (M = 5.06, SD = 1.03) in perceived affordances. STEM major status did not significantly interact with face type (Face Type × Major interaction: F(1, 242) = 1.04, p = .309, $η_{p}^{2} < . 01$ , 95% CI [0, .04]) or with goal type (Goal Type × Major interaction: F(1, 242) = 0.03, p = .873, $η_{p}^{2} < . 01$ , 95% CI [0, .01]. The face type effect remained, F(1, 242) = 26.86, p < .001, $η_{p}^{2} = . 10$ , 95% CI [.04, .17]). Overall, these analyses do not provide support for the idea that STEM majors and non-STEM majors perceive different cues from the facial structure of potential advisors (see Table 2 for means and standard deviations).

Table 2.

Perceptions by STEM and Non-STEM Major.

Outcomes	Trustworthy faces				Dominant faces
	STEM majors		Non-STEM majors		STEM majors		Non-STEM majors
	M	SD	M	SD	M	SD	M	SD
Agentic Affordances	5.28	1.14	5.28	1.97	5.07	2.19	5.13	1.27
Communal Affordances	5.09	1.84	5.06	2.19	4.63	2.12	4.80	1.23
Organizational Trust	4.99	2.03	4.85	1.17	4.55	2.17	4.62	1.25

Note. Ratings were made on scales from 1 to 7. STEM = science, technology, engineering, and mathematics.

Organizational trust

One-way repeated-measures ANOVAs detected greater organizational trust when considering working with STEM faculty with trustworthy faces than dominant faces, Study 3a: M_{trustworthy =} 4.95, SD = 1.10; M_{dominant =} 4.63, SD = 1.23; F(1, 115) = 21.99, p < .001, $η_{p}^{2} = . 16$ ; Study 3b: M_trustworthy = 4.85, SD = 0.97; M_dominant = 4.54, SD = 1.03, F(1, 155) = 17.43, p < .001, $η_{p}^{2} = . 10$ , 95% CI [.03, .20].

STEM major moderation analysis

We examined effects of STEM major on organizational in the combined data set (see Table 2). The 2 (face type) × 2 (major type) mixed-model ANOVA revealed that the main effect of major was not significant (M_STEM = 4.77, SD = 1.24; M_non-STEM = 4.74, SD = 1.11), F(1, 242) = 0.06, p = .813, $η_{p}^{2} < . 01$ , 95% CI [0, .02]. STEM major status did not moderate the effect of facial structure on organizational trust, Face Type × Major Type interaction: F(1, 242) = 2.73, p = .100, $η_{p}^{2} = . 01$ , 95% CI [0, .05]. Yet, the main effect of face type remained intact: M_trustworthy = 4.92, SD = 1.01; M_dominant = 4.58, SD = 1.08, F(1, 242) = 29.56, p < .001, $η_{p}^{2} = . 11$ , 95% CI [.05, .19].

Face type, communal goal affordances, and organizational trust

Is organizational trust related to perceived communal opportunities, agentic opportunities, or both? In a parallel mediation model (see Figure 6), trustworthy faces (vs. dominant faces) signaled organizational trust through an indirect effect of communal opportunities, but not agentic opportunities. These models suggest that inferred communal opportunities might play a distinctive role in anticipated organizational trust.

Figure 6.

Face type cues organizational trust through communal affordances in Study 3a (Panel A) and Study 3b (Panel B).

Discussion

Trustworthy faces, compared with dominant faces, of faculty leaders can signal organizational trust, and the evidence suggests that inferences occur similarly for both STEM and non-STEM majors. Trustworthy faces, compared with dominant faces, led to inferences of more goal opportunities in a research group setting; perceived communal opportunities uniquely fostered organizational trust in these physics or engineering labs. Whether respondents are recruited from the community, from a college sample, or are in or outside of a STEM pathway, similar patterns emerge: Facial signals of trustworthiness relative to dominance do not just contribute to impressions of the individual but also inform trust in the broader organization (e.g., Joshi & Diekman, 2021).

Studies 4a and 4b: Examining the Face Effect Under Ecologically Valid Conditions

The preceding experiments show that leaders’ trustworthy vs. dominant facial structures can be cues to the roles they occupy. In Study 4a, we address whether facial trustworthiness is still important when perceivers view ecologically valid stimuli (real-world STEM faculty) and whether facial trustworthiness (relative to dominance) exerts effects over and above perceived warmth. Specifically, Study 4a assessed whether the perceived trustworthiness of STEM faculty drawn from publicly available images, versus their perceived dominance, uniquely predicted STEM interest. In addition, we measured perceived warmth and competence to examine whether role inferences were merely due to facial positivity. Study 4a compiled a new stimulus set that overcame some limitations of prior stimulus sets: The images were ecologically valid because they were drawn from publicly available profiles of actual STEM faculty members, and the number of unique identities included was larger than used in prior studies. Thus, would the effects of advisors’ facial trustworthiness on STEM interest occur even among the faces of actual STEM faculty, and would such effects obtain even when controlling for a positivity halo (i.e., warmth)?

In Study 4b, we extend these questions (again with the faces of actual STEM faculty) and investigate whether the effects of facial structural cues emerge even in the context of other direct information about a faculty member valuing communal goals, such as explicit communal statements from STEM faculty. We see this point as particularly important. Inferences from faces are interpreted within contexts (e.g., Barden et al., 2004). Do the effects of facial cues survive even in the presence of other direct expressions of values? Prior evidence has documented that direct descriptions of exemplar behavior influence perceived opportunities in roles; for example, scientists described as collaborative compared with independent are perceived as offering more communal opportunities to advisees (Fuesting & Diekman, 2017). However, such subtle effects may be eliminated in the presence of clear individuating information about the STEM faculty themselves (Crawford et al., 2011). Here, we ask whether facial structure cues role opportunities, even alongside direct endorsements of communality.

Thus, Study 4b examined whether subtle variations in potential advisors’ facial structure influence role impressions alongside more direct information. Facial cues influence impressions even in the presence of explicit behavioral information: For instance, when participants have both facial cues and valenced behavioral descriptions of novel individuals, participants’ attitudes about individuals more strongly reflect facial signals (McConnell et al., 2008). Therefore, Studies 4a and 4b allow us to address whether variation in the trustworthy versus dominant facial structures of STEM faculty alongside other cues predicts respondents’ interest in working with them.

Methods

Participants

Study 4a participants (N = 60) from the introductory psychology pool at a midwestern university completed the study online and received partial course credit. An additional 9 participants were excluded due to failed attention checks. Participants tended to identify as White (71.7%, N = 43) and female (66.7%, N = 40) with an average age of 19. Sensitivity analyses indicated a detectable effect size of f² = 0.24, for a multi-linear regression with five predictors (80% power; α = .05, N = 60).

Study 4b participants (N = 95) from the introductory psychology pool at a midwestern university completed the study online and received partial course credit. Participants tended to identify as White (66.3%, N = 63) and female (64.2%, N = 61) with an average age of 19. An additional 11 participants were excluded due to failed attention checks. Sensitivity analyses indicated a detectable effect size of f = .10 for a repeated-measures 2 × 2 × 2 interaction (power = 0.80; α = .05, N = 95).

Experimental stimuli and procedure

For Study 4a, a pair of research assistants independently collected a total of 57 images of male faculty from a list of prestigious top-tier STEM universities. The assistants selected images in which faculty were presented as smiling and facing forward. Each of these selected individuals was an active STEM researcher and was visually identified as White. Each image was converted to grayscale (images available upon request). Participants were presented with 30 images randomly selected from a total of 57 images. They rated each target’s trustworthiness and dominance, along with interest in working with the target.

Study 4b participants viewed 12 websites presenting facial images of White male faculty members (from Study 1b) in an “about me” page with an image and a text description (see supplemental materials). New to this study was the manipulation of text depicting collaborative behavior, leading to a 2 (text description: collaborative, neutral) × 2 (face type: trustworthy, dominant) within-subjects design.

The text information presented collaborative or neutral descriptions of each faculty member’s research. The neutral description was the same as used in prior studies (e.g., PhD; current institution). The collaborative descriptors were one-sentence statements adapted from Diekman and colleagues (e.g., I like thinking hours on end with my lab group about ways to improve how research is conducted; 2011). A pretest with an independent online sample (N = 40) recruited from introductory psychology confirmed that collaborative descriptors were rated as providing more communal opportunity than the neutral descriptors, F(1, 35) = 75.17, p < .001, d = 2.94; conditions did not differ in agentic opportunity, F(1, 35) = 0.83, p = .37, d = 0.31.

Measures

Inferred characteristics

On scales from 1 to 7, Study 4a participants were asked to what extent each target appeared trustworthy, dominant, warm, and competent. They also estimated the age of each target as a control variable using a scale from 0 to 100.

Organizational trust

The measure of organizational trust was identical to Study 3 (α = .96).

Role interest

Study 4a participants reported their role interest by responding to the question: “If you were looking for a job, how interested would you be in applying to work with this person?” Study 4b participants reported their interest in working with each leader (adapted from Belanger et al., 2020) on a two-item scale including If you were looking for a job, how likely would it be for you to apply to be a research assistant in Dr. X’s lab?; If you were looking for a job, how interested would you be in becoming a research assistant in Dr. X’s lab? Items were averaged (α = .96). In each study, items were rated on scales from 1 (Not at all) to 7 (Extremely).

Results: Study 4a

Data were analyzed in a multilinear regression with perceived trustworthiness, dominance, warmth, and competence predicting interest in working with faculty members (see Table 3). The model included the estimated age of faculty members as a covariate. The results with this ecologically valid stimulus set replicated those found with more tightly controlled stimuli: Perceived trustworthiness was strongly related to students’ interest in working with faculty members. This effect held over and above perceived dominance, warmth, competence, and estimated age. As expected, perceived dominance was not related to students’ interest in working with faculty members. These data suggest that perceived trustworthiness signaled by the face is linked to students’ interest in working with different faculty members, and this effect is not reducible to perceived warmth.

Table 3.

Face-Based Trait Inferences on Role Interest.

Inferred characteristics	Role interest
Inferred characteristics	B	p
Trustworthiness	.543	.007
Dominance	–.076	.482
Warmth	.235	.076
Competence	–.001	.997
Age	–.034	.124

Note. The table reports standardized βs.

Results: Study 4b

Study 4b data were submitted to 2 (face structure) × 2 (text description) repeated-measures ANOVAs.

Organizational trust

Trustworthy faces (M = 4.97, SD = 0.81) signaled greater organizational trust than did dominant faces (M = 4.50, SD = 1.04), F(1, 94) = 68.07, p < .001, $η_{p}^{2} = . 42$ , 95% CI [.27,.55]. This effect was qualified by the presence of collaborative text, as shown in the Face Type × Text interaction, F(1, 94) = 31.29, p < .001, $η_{p}^{2} = . 25$ , 95% CI [.11, .38] (see Figure 7A). In both text conditions, trustworthy-faced leaders signaled greater organizational trust than dominant-faced leaders; however, the facial structure effect was attenuated by more than half when accompanied by the collaborative text: control condition—F(1, 94) = 79.64, p < .001, d = 1.84; collaborative condition—F(1, 94) = 8.37, p = .005, d = 0.60. Even in the presence of descriptions of collaborative behavior, facial structure conveys signals.

Figure 7.

Effects of face type and text description on organizational trust (panel A) and role interest (panel B).

Role interest

Students reported greater interest in working with advisors with trustworthy faces (M = 4.31, SD = 1.15) than dominant faces (M = 3.70, SD = 1.20), F(1, 94) = 80.68, p < .001, $η_{p}^{2} = . 46$ , 95% CI [.31, .57]. This effect was moderated by text description, as reflected in the Face Type × Text Description interaction, F(1, 94) = 53.72, p < .001, $η_{p}^{2} = . 36$ , 95% CI [.21, .49] (see Figure 7B). Again, in both text conditions, trustworthy-faced leaders fostered more role interest than dominant-faced leaders. However, with collaborative text, the facial structure effect is attenuated by more than half: control condition—F(1, 94) = 105.45, p < .001, d = 2.12; collaborative condition—F(1, 94) = 5.77, p = .02, d = 0.50.

Discussion

Studies 4a and 4b provide three important contributions. First, both studies demonstrate that cues from facial structure occur for faces of actual STEM faculty. Second, the relationship between facial trustworthiness and positivity toward a STEM advisor occurred not just when controlling for facial dominance but also when controlling for other potential predictors of positivity, such as warmth. Third, and perhaps most importantly, the effects occur even in the presence of direct information about a leader’s behavior, speaking to the power of leaders’ facial cues on judgments about roles.

General Discussion

Expectations of role content are inferred from subtle facial variations in occupants of the role. The cues gleaned from leaders’ facial structure influence perceptions of the nature of the role: What opportunities are present in working with this person? Specifically, facial structures signaling trustworthiness compared with dominance fostered beliefs that a faculty member would be communally oriented, and communally-oriented faculty were perceived to offer a wider array of both agentic and communal opportunities for their teams. Perceived goal opportunities were related to important outcomes, such as anticipated organizational trust and interest in joining a STEM lab. In this way, the faces of leaders can shape perceptions of social structure.

Theoretical Implications

A classic social psychological principle is that impressions of individuals follow from their immersion in a particular situation; the current work extends this principle to illuminate how impressions of individuals influence impressions of roles. These findings move beyond the knowledge that context moderates facial structure effects to establish that facial structure can cue beliefs about surrounding context. Established work shows that social roles moderate the inferences drawn from facial stimuli (e.g., Barden et al., 2004) or that context moderates evaluation of facial signals (preferences for leaders’ facial structures vary depending on times of war or peace; Little et al., 2007). Here we extend this knowledge to consider how face structure influences judgments not only of the target but also of the role. This approach extends the breadth of face perception literature by demonstrating meaningful consequences beyond an individual to the role and allows for new questions to be investigated. For example, what features of individual impression formation most strongly form impressions of roles? Do role impressions form spontaneously from person impressions? What person and situation factors promote or inhibit the formation of role impressions?

Practical Implications

These findings matter particularly in the context of STEM fields and for the question of how outsiders to STEM might form impressions about experiences within STEM. Existing research has established that scientists’ actual facial structure can influence professional impressions (e.g., perceived research quality; Gheorghiu et al., 2017). Yet the documentation of the face-to-role inference suggests a farther-reaching effect—that scientists’ facial structures shift perceptions of the role and, in turn, students’ attitudes about entering that role.

STEM fields (particularly physics and engineering) continue to be male-dominated; the current research provides insight into how men might transmit (knowingly or not) information about STEM fields. In the current research, trustworthy-faced male faculty fostered organizational trust and lab interest. The signaling capacity within male faces matters because men are the majority in many STEM fields, and thus they provide mentorship to the next generation of students. Understanding cues to trustworthiness through facial expression, verbal statements, or supportive behaviors can help minimize disparities. For example, female engineering students with male peer mentors (relative to female peer mentors) reported greater anxiety and worse outcomes (Dennehy & Dasgupta, 2017). Understanding the psychological processes involved in establishing trustworthiness can help to mitigate these disparate effects.

Limitations and Future Directions

These studies provide ample avenues for future research. The current work focused on anticipated, rather than actual, role experience, and thus cannot speak to the accuracy of these impressions. Do the facial signals of a leader convey signals about an individual’s later experience in that leader’s organization? Prior evidence supports this possibility. For example, students’ impressions of faculty teaching effectiveness based on brief exposures (as little as 10 s of silent video) corresponded to student ratings after a semester in the faculty member’s classroom (Ambady & Rosenthal, 1993). But such inferences are not wholly accurate. Although perceivers may expect attractive company CEOs to perform better (Rule & Ambady, 2008), there is no clear relationship between CEO appearance and company performance (Hopp et al., 2020; Stoker et al., 2016). The processes underlying the relationship from leader facial cues to organizational outcomes deserve research attention: Facial cues might lead others (coworkers, subordinates) to treat the leader in ways that confirm the expectancy derived from the facial cue. For instance, individuals whose faces cue dominance may be treated by others with greater deference and provided more leadership opportunities. Alternately, such leader impressions may affect who is chosen to lead but not their actual leadership performance (e.g., Stoker et al., 2016). Whether early impressions of a role leader relate to organizational experiences and outcomes is a promising question for future research.

The finding that facial structure provides a cue to organizational trust is of particular interest in considering how to broaden representation in STEM. Students from underrepresented backgrounds can particularly benefit from highlighting communal opportunities within STEM (Thoman et al., 2017). In the current research, analyses of Study 3a and 3b data found that URM students anticipated less organizational trust than represented students. However, signals of organizational trust from facial structure did not differ significantly for represented and underrepresented students or for female and male students (see supplemental materials). Extracting motivational inferences from advisor facial structure appears to function similarly for students from majority and minority groups within STEM.

Yet, these data cannot speak to how the trust signals extracted from faces may be weighted differently by minoritized versus majoritized students in subsequent decisions. Individuals who are uncertain about entering a role may be especially likely to weigh cues to trust. As such, minoritized students might especially act on role information inferred from the facial structure of role leaders. Entering new roles entails risks, and that risk is elevated when individuals have less information—for example when they do not have family and friends with experience. For instance, first-generation college students may seek out cues to belonging when navigating higher education environments (e.g., Stephens et al., 2012). The current studies provide foundational evidence that the facial structure of STEM role leaders informs beliefs about role content; future research needs to examine for whom and under what conditions these signals matter most.

Conclusion

Impressions of the social structure can form from inferences based on subtle variations in the facial structure of role leaders. These initial impressions of role leaders—whether encountered in face-to-face interaction or through media representations—can provide cues about what entering a new role will be like. Who enters one role and leaves another may be guided, in part, by subtle cues emitted by leaders in those roles.

Supplemental Material

sj-docx-1-psp-10.1177_01461672221077801 – Supplemental material for In the Face of Opportunities: Facial Structures of Scientists Shape Expectations of STEM Environments

Supplemental material, sj-docx-1-psp-10.1177_01461672221077801 for In the Face of Opportunities: Facial Structures of Scientists Shape Expectations of STEM Environments by Mansi P. Joshi, E. Paige Lloyd, Amanda B. Diekman and Kurt Hugenberg in Personality and Social Psychology Bulletin

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: These studies were partially supported by NIH grant T32HD007475-26 to Mansi Joshi, NSF grant 1232364 to Amanda Diekman, and a PEO Scholar Dissertation Award to E. Paige Lloyd.

ORCID iDs

Mansi P. Joshi

E. Paige Lloyd

Supplemental Material

Supplemental material is available online with this article.

Notes

References

Abele

A. E.

Wojciszke

(2014). Communal and agentic content in social cognition: A dual perspective model. In Olson

J. M.

Zanna

M. P.

(Eds.), Advances in experimental social psychology (Vol. 50, pp. 195–255). Academic Press. https://doi.org/10.1016/B978-0-12-800284-1.00004-7

Ambady

Hallahan

Conner

(1999). Accuracy of judgments of sexual orientation from thin slices of behavior. Journal of Personality and Social Psychology, 77(3), 538–547. https://doi.org/10.1037/0022-3514.77.3.538

Ambady

Rosenthal

(1993). Half a minute: Predicting teacher evaluations from thin slices of nonverbal behavior and physical attractiveness. Journal of Personality and Social Psychology, 64(3), 431–441. https://doi.org/10.1037/0022-3514.64.3.431

Barden

Maddux

W. W.

Petty

R. E.

Brewer

M. B.

(2004). Contextual moderation of racial bias: The impact of social roles on controlled and automatically activated attitudes. Journal of Personality and Social Psychology, 87(1), 5–22. https://doi.org/10.1037/0022-3514.87.1.5

Belanger

A. L.

Diekman

A. B.

Steinberg

(2017). Leveraging communal experiences in the curriculum: Increasing positivity toward engineering by changing stereotypic expectations. Journal of Applied Social Psychology, 47, 305–319. https://doi.org/10.1111/jasp.12438

Belanger

A. L.

Joshi

M. P.

Fuesting

M. A.

Weisgram

E. S.

Claypool

H. M.

Diekman

A. B.

(2020). Putting belonging in context: Communal affordances signal belonging in STEM. Personality and Social Psychology Bulletin, 46(8), 1186–1204. https://doi.org/10.1177/0146167219897181

Chan

H. F.

Torgler

(2020). “Beauty” premium for social scientists but “unattractiveness” premium for natural scientists in the public speaking market. Humanities and Social Sciences Communications, 7(1), 1–9. https://doi.org/10.1057/s41599-020-00608-6

Brown

E. R.

Steinberg

Diekman

A. B.

(2018). Is the lone scientist an American dream? Perceived communal opportunities in STEM offer a pathway to closing U.S.–Asia gaps in interest and positivity. Social Psychological and Personality Science, 9(1), 11–23. https://doi.org/10.1177/1948550617703173

Brown

E. R.

Thoman

D. B.

Smith

J. L.

Diekman

A. B.

(2015). Closing the communal gap: The importance of communal affordances in science career motivation. Journal of Applied Social Psychology, 45, 662–673.

10.

Cheryan

Plaut

V. C.

Davies

P. G.

Steele

C. M.

(2009). Ambient belonging: How stereotypical cues impact gender participation in computer science. Journal of Personality and Social Psychology, 97(6), 1045–1060.

11.

Child

(1936). The judging of occupations from printed photographs. The Journal of Social Psychology, 7(1), 117–118. https://doi.org/10.1080/00224545.1936.9921650

12.

Crawford

J. T.

Jussim

Madon

Cain

T. R.

Stevens

S. T.

(2011). The use of stereotypes and individuating information in political person perception. Personality and Social Psychology Bulletin, 37(4), 529–542. https://doi.org/10.1177/0146167211399473

13.

Dansky

K. H.

(1996). The effect of group mentoring on career outcomes. Group & Organization Management, 21(1), 5–21. https://doi.org/10.1177/1059601196211002

14.

Dennehy

Dasgupta

(2017). Female peer mentors early in college increase women’s positive academic experiences and retention in engineering. Proceedings of the National Academy of Sciences, 114, 5964–5969.

15.

Diekman

A. B.

Brown

E. R.

Johnston

A. M.

Clark

E. K.

(2010). Seeking congruity between goals and roles: A new look at why women opt out of science, technology, engineering, and mathematics careers. Psychological Science, 21(8), 1051–1057. https://doi.org/10.1177/0956797610377342

16.

Diekman

A. B.

Clark

E. K.

Johnston

A. M.

Brown

E. R.

Steinberg

(2011). Malleability in communal goals and beliefs influences attraction to STEM careers. Journal of Personality and Social Psychology, 101, 902–918.

17.

Diekman

A. B.

Joshi

M. P.

Benson-Greenwald

T. M.

(2020). Goal congruity theory: Navigating the social structure to fulfill goals. In Gawronski

(Ed.), Advances in experimental social psychology (Vol. 62, pp. 189–244). Academic Press. https://doi.org/10.1016/bs.aesp.2020.04.003

18.

Ellemers

de Gilder

van den Heuvel

(1998). Career-oriented versus team-oriented commitment and behavior at work. Journal of Applied Psychology, 83(5), 717–730. https://doi.org/10.1037/0021-9010.83.5.717

19.

Emerson

K. T. U.

Murphy

M. C.

(2015). A company I can trust? Organizational lay theories moderate stereotype threat for women. Personality and Social Psychology Bulletin, 41(2), 295–307. https://doi.org/10.1177/0146167214564969

20.

Ferguson

H. S.

Owen

Hahn

A. C.

Torrance

DeBruine

L. M.

Jones

B. C.

(2019). Context-specific effects of facial dominance and trustworthiness on hypothetical leadership decisions. PLOS ONE, 14(7), Article e0214261. https://doi.org/10.1371/journal.pone.0214261

21.

Fiske

S. T.

Cuddy

A. J. C.

Glick

(2002). A model of (often mixed) stereotype content. Journal of Personality and Social Psychology, 82, 878–902.

22.

Fuesting

M. A.

Diekman

A. B.

(2017). Not by success alone: Role models provide pathways to communal opportunities in STEM. Personality and Social Psychology Bulletin, 43(2), 163–176. https://doi.org/10.1177/0146167216678857

23.

Fuesting

M. A.

Diekman

A. B.

Boucher

K. L.

Murphy

M. C.

Manson

D. L.

Safer

B. L.

(2019). Growing STEM: Perceived faculty mindset as an indicator of communal affordances in STEM. Journal of Personality and Social Psychology, 117(2), 260–281. https://doi.org/10.1037/pspa0000154

24.

Fukuyama

(1996). Trust: The social virtues and the creation of prosperity (1. Free Press paperback ed.). Free Press.

25.

Gahagan

(1933). Judgments of occupations from printed photographs. The Journal of Social Psychology, 4(1), 128–134. https://doi.org/10.1080/00224545.1933.9921562

26.

Gheorghiu

A. I.

Callan

M. J.

Skylark

W. J.

(2017). Facial appearance affects science communication. Proceedings of the National Academy of Sciences, 114(23), 5970–5975. https://doi.org/10.1073/pnas.1620542114

27.

Gheorghiu

A. I.

Callan

M. J.

Skylark

W. J.

(2020). A thin slice of science communication: Are people’s evaluations of TED talks predicted by superficial impressions of the speakers? Social Psychological and Personality Science, 11(1), 117–125. https://doi.org/10.1177/1948550618810896

28.

Hopp

Wentzel

Rose

(2020). Chief executive officers’ appearance predicts company performance, or does it? A replication study and extension focusing on CEO successions. The Leadership Quarterly. Advance online publication. https://doi.org/10.1016/j.leaqua.2020.101437

29.

Joshi

M. P.

Diekman

A. B.

(2021). My fair lady? Inferring organizational trust from the mere presence of women in leadership roles. Personality and Social Psychology Bulletin. https://doi.org/10.1177/01461672211035957

30.

Keating

C. F.

(1985). Gender and the physiognomy of dominance and attractiveness. Social Psychology Quarterly, 48(1), 61–70. https://doi.org/10.2307/3033782

31.

Kleisner

Priplatova

Frost

Flegr

(2013). Trustworthy-looking face meets brown eyes. PLOS ONE, 8(1), Article e53285. https://doi.org/10.1371/journal.pone.0053285

32.

Krumhuber

Manstead

A. S. R.

Cosker

Marshall

Rosin

P. L.

Kappas

(2007). Facial dynamics as indicators of trustworthiness and cooperative behavior. Emotion, 7(4), 730–735. https://doi.org/10.1037/1528-3542.7.4.730

33.

Little

A. C.

Burriss

R. P.

Jones

B. C.

Roberts

S. C.

(2007). Facial appearance affects voting decisions. Evolution and Human Behavior, 28(1), 18–27. https://doi.org/10.1016/j.evolhumbehav.2006.09.002

34.

Little

A. C.

Roberts

S. C.

Jones

B. C.

Debruine

L. M.

(2012). The perception of attractiveness and trustworthiness in male faces affects hypothetical voting decisions differently in wartime and peacetime scenarios. Quarterly Journal of Experimental Psychology (2006), 65(10), 2018–2032. https://doi.org/10.1080/17470218.2012.677048

35.

D. S.

Correll

Wittenbrink

(2015). The Chicago face database: A free stimulus set of faces and norming data. Behavior Research Methods, 47(4), 1122–1135. https://doi.org/10.3758/s13428-014-0532-5

36.

McConnell

A. R.

Rydell

R. J.

Strain

L. M.

Mackie

D. M.

(2008). Forming implicit and explicit attitudes toward individuals: Social group association cues. Journal of Personality and Social Psychology, 94(5), 792–807. https://doi.org/10.1037/0022-3514.94.5.792

37.

Montoya

A. K.

Hayes

A. F.

(2017). Two-condition within-participant statistical mediation analysis: A path-analytic framework. Psychological Methods, 22(1), 6–27. https://doi.org/10.1037/met0000086

38.

Murphy

M. C.

Steele

C. M.

Gross

J. J.

(2007). Signaling threat: How situational cues affect women in math, science, and engineering settings. Psychological Science, 18(10), 879–885. https://doi.org/10.1111/j.1467-9280.2007.01995.x

39.

National Science Board. (2020). Science and engineering indicators 2020: The state of science and engineering (NSB-2020-1). National Science Foundation. https://ncses.nsf.gov/pubs/nsb20201/

40.

Olivola

C. Y.

Funk

Todorov

(2014). Social attributions from faces bias human choices. Trends in Cognitive Sciences, 18(11), 566–570. https://doi.org/10.1016/j.tics.2014.09.007

41.

Oosterhof

N. N.

Todorov

(2008). The functional basis of face evaluation. Proceedings of the National Academy of Sciences, 105(32), 11087–11092. https://doi.org/10.1073/pnas.0805664105

42.

Oosterhof

N. N.

Todorov

(2009). Shared perceptual basis of emotional expressions and trustworthiness impressions from faces. Emotion, 9(1), 128–133. https://doi.org/10.1037/a0014520

43.

Pew Research Center. (2018). Women and men in STEM often at odds over workplace equity. https://www.pewsocialtrends.org/2018/01/09/women-and-men-in-stem-often-at-odds-over-workplace-equity/

44.

Purdie-Vaughns

Steele

C. M.

Davies

P. G.

Ditlmann

Crosby

J. R.

(2008). Social identity contingencies: How diversity cues signal threat or safety for African Americans in mainstream institutions. Journal of Personality and Social Psychology, 94(4), 615–630. https://doi.org/10.1037/0022-3514.94.4.615

45.

Rezlescu

Duchaine

Olivola

C. Y.

Chater

(2012). Unfakeable facial configurations affect strategic choices in trust games with or without information about past behavior. PLOS ONE, 7(3), Article e34293. https://doi.org/10.1371/journal.pone.0034293

46.

Rule

N. O.

Ambady

(2008). The face of success: Inferences from chief executive officers’ appearance predict company profits. Psychological Science, 19(2), 109–111. https://doi.org/10.1111/j.1467-9280.2008.02054.x

47.

Schriesheim

C. A.

Cogliser

C. C.

Neider

L. L.

(1995). Is it “trustworthy”? A multiple-levels-of-analysis reexamination of an ohio state leadership study, with implications for future research. The Leadership Quarterly, 6(2), 111–145. https://doi.org/10.1016/1048-9843(95)90031-4

48.

Shockley-Zalabak

Ellis

Winograd

(2000). Organizational trust: What it means, why it matters. Organization Development Journal; Chesterland, 18(4), 35–48.

49.

Stephens

N. M.

Markus

H. R.

Fryberg

S. A.

(2012). Social class disparities in health and education: Reducing inequality by applying a sociocultural self model of behavior. Psychological Review, 119(4), 723–744. https://doi.org/10.1037/a0029028

50.

Stirrat

Perrett

D. I.

(2010). Valid facial cues to cooperation and trust: Male facial width and trustworthiness. Psychological Science, 21(3), 349–354. https://doi.org/10.1177/0956797610362647

51.

Stoker

J. I.

Garretsen

Spreeuwers

L. J.

(2016). The facial appearance of CEOs: Faces signal selection but not performance. PLOS ONE, 11(7), Article e0159950. https://doi.org/10.1371/journal.pone.0159950

52.

Thoman

D. B.

Muragishi

G. A.

Smith

J. L.

(2017). Research microcultures as socialization contexts for underrepresented science students. Psychological Science, 28(6), 760–773. https://doi.org/10.1177/0956797617694865

53.

Todorov

Baron

S. G.

Oosterhof

N. N.

(2008). Evaluating face trustworthiness: A model based approach. Social Cognitive and Affective Neuroscience, 3(2), 119–127. https://doi.org/10.1093/scan/nsn009

54.

Todorov

Olivola

C. Y.

Dotsch

Mende-Siedlecki

(2015). Social attributions from faces: Determinants, consequences, accuracy, and functional significance. Annual Review of Psychology, 66(1), 519–545. https://doi.org/10.1146/annurev-psych-113011-143831

55.

Walker

Schönborn

Greifeneder

Vetter

(2018). The basel face database: A validated set of photographs reflecting systematic differences in Big Two and Big Five personality dimensions. PLOS ONE, 13(3), e0193190. https://doi.org/10.1371/journal.pone.0193190

56.

Wang

Hahn

A. C.

DeBruine

L. M.

Jones

B. C.

(2016). The motivational salience of faces is related to both their valence and dominance. PLOS ONE, 11(8), Article e0161114. https://doi.org/10.1371/journal.pone.0161114

57.

Zebrowitz

L. A.

Collins

M. A.

(1997). Accurate social perception at zero acquaintance: The affordances of a Gibsonian approach. Personality and Social Psychology Review, 1(3), 204–223. https://doi.org/10.1207/s15327957pspr0103_2

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.37 MB