Abstract
Leadership development programs traditionally emphasize the acquisition of leadership knowledge and skills, yet often fail to cultivate individuals’ internal leadership self-schemas. This gap is consequential because without strong leadership self-expectations, individuals may lack the internalized beliefs that shape their approach to leadership roles and opportunities, limiting the effectiveness of traditional development efforts. The present study investigates a mobile-based attribute conditioning intervention that aims to enhance individuals’ leader self-schema and self-expectations via Positive Implicit Leadership Theories (PILTs). In a randomized experimental design, 153 participants were assigned to either a positive leader self-schema conditioning group or a neutral control group using the smartphone application TAPPIT. Results revealed that the intervention significantly increased leader self-expectations, leadership self-efficacy, and social-normative motivation to lead, with leadership self-efficacy mediating the effects of leader self-expectations on motivation to lead. This study contributes to leadership development theory and practice by providing initial evidence that ethical, scalable, self-directed interventions can shape leadership self-cognition. By integrating the literature on implicit theories, associative conditioning, and self-fulfilling prophecy, the study opens new avenues for future scholarship and scalable practices.
Leadership development has long been a central focus in both research and practice, with traditional approaches emphasizing knowledge acquisition and skill application. Over the past three decades, leader development has been a major area of research (Day et al., 2014) and investment, with organizations spending an estimated $60 billion annually on leadership training (Yemiscigil et al., 2023). Most of these initiatives continue to rely on instructor-led trainings that focus primarily on knowledge and behavioral skills (Wentworth, 2016). A critical gap persists; most programs offer little direct support for enhancing leader mindsets and self-schemas, which are an essential yet overlooked element in achieving lasting leader development outcomes (Leung & Sy, 2018). Recent research highlights the necessity of not only building behavioral skills, but also cultivating internal leader mindsets and self-schemas, which are associated with leadership behaviors, relational dynamics, and effectiveness (e.g., Gottfredson & Reina, 2021; Haslam et al., 2022; Ibarra et al., 2014; Kouzes & Posner, 2019). Without reliable, scalable, and ethical methods for strengthening those beliefs, organizations may continue to invest heavily in training while missing a key driver of return on investment (Avolio et al., 2009).
The present research investigates a mobile application intervention designed to strengthen leadership self-schemas. Specifically, we examine whether a brief gamified attribute conditioning intervention that repeatedly pairs an individual's own image with positive leadership traits can strengthen leader-relevant self-cognitions and increase leadership self-efficacy and motivation to lead. Leadership development research widely recognizes that individuals’ self-views shape leader emergence, motivation, and developmental trajectories (Day & Harrison, 2007; Lord & Hall, 2005; Walker et al., 2024). However, most developmental programs focus primarily on knowledge acquisition or behavioral skill-building, with less attention to the cognitive structures through which individuals interpret leadership roles. Drawing on research on associative learning and attribute conditioning, which shows that repeated pairings can alter evaluative associations and self-relevant beliefs (De Houwer et al., 2001; Förderer & Unkelbach, 2015; Olson & Fazio, 2006), we examine whether a brief mobile intervention can influence leader-related self-cognitions. Although concerns about replicability have emerged for some social priming interventions, evaluative and associative conditioning effects remain well-documented when pairing procedures and exposure conditions are tightly controlled (Dai et al., 2023; De Houwer et al., 2001; Förderer & Unkelbach, 2015).
The study focuses on cognitive processes that theory suggests may precede self-fulfilling leadership dynamics. Prior literature has proposed links between self-expectations, intentions, and behavior (Armitage et al., 2015; Avolio et al., 2009; Eden & Ravid, 1982; Jussim & Harber, 2005; McNatt, 2000; Rosenthal & Rubin, 1978), although considerable variability remains in whether attitudinal change translates into behavioral enactment, including in organizational contexts (Conner & Norman, 2022; Enste & Altenhöner, 2021; Sheeran & Webb, 2016). We therefore do not examine downstream leadership behavior or performance in the present study. Instead, we investigate whether a scalable attribute conditioning intervention can strengthen leader-related self-cognitions that may contribute to leadership motivation.
We situate this approach conceptually within research on self-fulfilling prophecy and Galatea effects. Galatea processes describe situations in which individuals’ positive self-expectations influence subsequent outcomes (Eden & Kinnar, 1991; McNatt & Judge, 2004). The present research does not test these behavioral processes directly. Rather, it examines whether attribute conditioning can influence leader-relevant self-cognitions that represent an early stage in the theoretical pathway through which self-expectations may shape leadership motivation and behavior.
This research contributes to leadership development scholarship in three ways. First, it introduces attribute conditioning as a feasible intervention for shaping leader-related self-schemas. Although leadership research emphasizes the importance of leader identity and self-views, scalable methods for altering these cognitions remain limited. Second, the study demonstrates that a brief mobile intervention can influence multiple leader-related cognitions, including leader self-expectations, leadership self-efficacy, and motivation to lead. Third, the findings open a new line of inquiry on how digital micro-interventions may complement traditional leadership development programs by targeting cognitive foundations of leadership motivation.
Theoretical Background: Leader Identity, Self-Fulfilling Prophecy and Self-Expectations
Contemporary Leader Identity Literature
Contemporary scholarship emphasizes leader identity and self-views as foundational for leader emergence and effectiveness. Identity theorists argue that identity constitutes the deep structure that organizes leader development (Lord & Hall, 2005) and connects developmental processes to actual experience and practice (Day & Harrison, 2007). Social identity and identity leadership perspectives further clarify how shared social identities and prototypicality shape leader influence and follower responses (Haslam et al., 2022; Hogg, 2001; Hogg & van Knippenberg, 2003; van Knippenberg, 2023). Though this literature makes it clear that leader self-views are key mechanisms for leader development, there remains a gap between theory and practice. Many leader development programs emphasize skill rehearsal, behavioral feedback, and strategy, but the field has not yet produced a scalable set of techniques that reliably alter leader self-concept and self-expectations. We address this methodological gap by targeting leader-relevant self-perceptions through attribute conditioning, providing a feasible route for linking identity theory to practicable interventions (Day & Harrison, 2007; Hammond et al., 2017; Haslam et al., 2022; Lord & Hall, 2005).
Self-Expectations and Self-Fulfilling Prophecy
Self-fulfilling prophecy effects, where an individual's expectations alter real outcomes, are well-documented, including in the leadership literature (Eden & Kinnar, 1991; Eden & Zuk, 1995; Romney et al., 2024; Rosenthal & Jacobson, 1968; Rosenthal & Rubin, 1978; McNatt & Judge, 2004; van Ewijk et al., 2023; Wang et al., 2022; Whiteley et al., 2012). Much of this research has focused on Pygmalion leadership effects, where one's expectation of another (e.g., leaders’ expectations of followers) affects their behavior, in turn impacting outcomes (Eden, 1984, 2003; Rosenthal & Jacobson, 1968). According to a meta-analysis conducted by Avolio and colleagues (2009), Pygmalion leadership trainings have the largest impact on leader behavioral and cognitive outcomes (e.g., leader emergence, confidence, idea generation) compared to other training methods. However, Pygmalion leadership effects rely on others’ expectations to enhance leader behaviors. The literature remains more limited in its examination of self-expectations (Galatea effects) as a direct driver of leader development.
Galatea Leadership Effects
The Galatea effect refers to a type of self-fulfilling prophecy in which one's own positive self-expectations (as opposed to others’ expectations of them) lead to positive outcomes (e.g., Eden & Kinnar, 1991; Judge et al., 1998; McNatt & Judge, 2004; van Ewijk et al., 2023). The present study does not test Galatea effects directly. Instead, Galatea theory is presented as a conceptual framework that explains why leader self-expectations and self-cognitions may matter for leadership development and motivation. Galatea effects are impactful in various domains, including business (e.g., McNatt & Judge, 2004; van Ewijk et al., 2023; Watson et al., 2021), for enhancing motivation and goal pursuit (e.g., Eden & Kinnar, 1991; Lidor et al., 2021; McNatt & Judge, 2004). For example, a recent nine-country study of individuals in business and entrepreneurship courses found that participants’ pre-course expectations of entrepreneurial inspiration were positively associated with later entrepreneurial inspiration and post-course intentions (van Ewijk et al., 2023). Strengthening individuals’ self-expectations about their leadership capabilities may trigger self-fulfilling dynamics that foster leadership-relevant motivations and behaviors.
To test the first link of this chain (leader self-expectations enhancing self-efficacy and motivation), the present study evaluates a mobile, attribute-conditioning intervention designed to strengthen leader self-schemas. A challenge recognized in the leadership and Galatea literature is the need for interventions that may create self-fulfilling prophecy effects in applied settings (Eden et al., 2000; Eden & Ravid, 1982). A key unresolved gap is the need for ethical, reliable, and scalable methods to shape positive self-expectations without resorting to potentially deceptive practices (e.g., false feedback, such as expressing that someone has “high potential” or assigning them to a purportedly high-performing group; Eden et al., 2000). The present study utilizes attribute conditioning to strengthen leadership self-schemas, presenting a practical and readily implementable intervention for cognitive change.
Positive Implicit Leadership Theories (PILTs)
Implicit Leadership Theories (ILTs) are individuals’ cognitive schemas of the attributes that characterize leaders (Lord et al., 1984). These schemas can be categorized into positive leader prototypes (i.e., Sensitivity, Intelligence, Dynamism, and Dedication) and negative leader prototypes (i.e., Tyranny and Masculinity; Epitropaki & Martin, 2004). Individuals’ leadership self-schemas impact how they view themselves, behave in leadership roles, and develop self-expectations as leaders (Hannah et al., 2009). The present study investigates the use of individuals’ positive implicit leadership theories (PILTs) as proxies for their self-expectations as leaders. When PILTs are elevated, individuals’ leadership self-expectations may also be enhanced, as represented in the initial stage of the Galatea leadership process. Although research has established the relationship between ILTs and leader perceptions and behaviors (Epitropaki et al., 2013), there remains limited understanding of whether these cognitive schemas can be actively changed in a controlled, self-directed manner.
In the present research we distinguish three related but conceptually distinct constructs. First, implicit leadership theories refer to individuals’ cognitive prototypes of the attributes that characterize leaders (Lord et al., 1984). Second, leader self-schemas reflect the extent to which individuals associate those leadership attributes with themselves, shaping how individuals interpret and approach leadership roles (Lord & Hall, 2005). Third, leader self-expectations refer to individuals’ forward-looking beliefs about their capacity to enact leadership behaviors or perform effectively in leadership roles. The attribute conditioning intervention used in this study primarily targets leader self-schemas by strengthening associations between the self and positive leader attributes. Changes in leader self-schemas may in turn influence leader self-expectations and leadership motivation, consistent with identity-based models of leader development (Day & Harrison, 2007; Hannah et al., 2009). To address this gap, the following section introduces attribute conditioning, a type of associative learning, as a mechanism for strengthening associations between the self and positive leadership attributes. Rather than directly altering implicit leadership theories themselves, the intervention aims to increase the accessibility of leader-relevant attributes within individuals’ leader self-schemas. This distinction is important because the intervention operates by strengthening self-referential associations with leadership attributes rather than by altering individuals’ general beliefs about what leaders are like.
Elevating PILTs via Attribute Conditioning
Self-directed training or “positive activities” for self-improvement are common in positive psychology (e.g., Lyubomirsky & Layous, 2013; Sin & Lyubomirsky, 2009). We propose a parallel approach for organizational contexts to enhance individuals’ positive leadership self-schemas (i.e., PILTs). Because PILTs are shaped by repeated experiences and exposure to leader prototypes (Epitropaki et al., 2013; Lord et al., 2020), the present study utilizes a self-directed attribute conditioning intervention aimed at strengthening individuals’ positive leadership self-schemas through repeated exposure (e.g., Bosshard et al., 2019; Franklin et al., 2016; Olson & Fazio, 2006).
Associative conditioning is often used to alter individuals’ perceptions and ultimately influence their behaviors (e.g., De Houwer et al., 2001; Förderer & Unkelbach, 2015). One form of associative conditioning, attribute conditioning, pairs a target individual with various attributes through referential learning, or forming links between concepts in one's memory (Förderer & Unkelbach, 2016). Traits can be conditioned onto a target individual through repeated pairings. For example, when an athletic person is consistently paired with a neutral individual, observers are more likely to attribute athleticism to the neutral person over time (Förderer & Unkelbach, 2016). Applying this logic, we hypothesize that repeatedly pairing a photo of oneself with words empirically shown to represent PILTs (Epitropaki & Martin, 2004; Offermann et al., 1994) will lead individuals to associate these attributes with themselves, enhancing their leadership self-expectations. This approach offers a theoretically grounded, scalable, and experimentally tractable method for shaping leader self-schemas and studying leader self-concept processes.
Elevating PILTs to Enhance Leader Self-Expectations
Leader identity theorists emphasize the importance of accessible self-concept content for motivating leader behavior. Associative network theory also suggests that activating one concept (e.g., “leader”) can increase the accessibility of related traits (e.g., “dynamic,” “intelligent”), influencing subsequent cognition and behavior (Chen et al., 2021; Collins & Loftus, 1975; Dai et al., 2023; Fazio, 2001; Markus, 1977; Weingarten et al., 2016; Wheeler et al., 2007). Therefore, attribute conditioning can raise the accessibility of leadership related schemas and self-expectations by creating associative links, presenting a practical method for modifying cognitive content that supports leadership self-schemas, self-expectations, and in turn, motivation to lead.
This approach is intended as an initial, proximal lever that may facilitate longer term self-concept development when combined with experience and reflective practice. While the present study tests immediate changes in leadership-related cognition and motivation, determining whether these lead to longer-term self-schema shifts or observable leadership behavior will require longitudinal and field research. While attitudes and self-expectations can serve as antecedents of downstream intentions and behavior (e.g., Theory of Planned Behavior, Ajzen, 1991; Expectancy Effects, Rosenthal & Rubin, 1978), including in organizational contexts (e.g., Avolio et al., 2009; Eden & Ravid, 1982; Guillén et al., 2015; Kierein & Gold, 2000; McNatt & Judge, 2004), this study focuses on evaluating the efficacy of a brief mobile intervention for shaping self-expectations.
Self-Fulfilling Prophecy and Leadership Self-Efficacy
Self-efficacy is considered a key mechanism in the Galatea process (e.g., Eden & Kinnar, 1991; Eden & Zuk, 1995; McNatt & Judge, 2004). In work settings, Galatea interventions have been used to enhance workers’ self-efficacy, which in turn boosted performance (McNatt & Judge, 2004). We expect that enhancing individuals’ leader self-schemas will be associated with higher leadership self-efficacy, a domain-specific form of self-efficacy referring to individuals’ self-perceived capabilities for effective leadership behaviors (Chemers et al., 2000; Hannah et al., 2008; Kane et al., 2002; Ng et al., 2008). Outside feedback (from colleagues, mentors, or coaches) has been implicated as a key influence on leadership self-efficacy in formal training programs (Dwyer, 2019). Rather than using outside feedback, we utilize attribute conditioning as a potential self-directed way to enhance leader self-expectations, and therefore leadership self-efficacy. Specifically, we predict that when individuals internalize more PILTs through attribute conditioning, they will develop higher self-expectations as leaders, believing they possess the attributes necessary to lead effectively, which in turn will elevate their leadership self-efficacy.
Self-Fulfilling Prophecy and Motivation to Lead
Motivation is a crucial variable to examine because it can inform the extent to which individuals are willing to put in effort to fulfill leadership responsibilities or seek leadership opportunities (DeRue & Myers, 2014). Motivation to lead (Chan & Drasgow, 2001; McCormick et al., 2002), the desire to attain leadership roles and expend effort to fulfill leadership role requirements, includes three dimensions: social-normative motivation, affective-identity motivation, and non-calculative motivation. Social-normative motivation to lead reflects a sense of obligation or duty to lead, affective-identity motivation to lead reflects enjoyment of and identification with being a leader, and non-calculative motivation to lead reflects the degree to which individuals see leadership opportunities positively despite potential personal costs. Recent work suggests that these dimensions are only modestly correlated and have unique patterns of antecedents, so should be operationalized as distinct constructs (Badura et al., 2020).
Theoretically, social-normative and affective-identity motivations to lead may respond differently to attribute conditioning. Social-normative motivation is driven by external expectations, while affective-identity motivation comes from internal identification with leadership roles and personal enjoyment of leading (Chan & Drasgow, 2001; Badura et al., 2020). Conditioning individuals with positive leader schemas (PILTs) may enhance their sense of duty or obligation (social-normative motivation). However, it might have less impact on their intrinsic enjoyment or self-concept as leaders (affective-identity motivation), especially if individuals’ internal attitudes toward leadership are already well-established. By contrast, because non-calculative motivation is rooted in cost-benefit perceptions rather than self-concept, it is less likely to shift as a result of enhanced self-schemas.
Social-normative and affective-identity motivations are typically stronger predictors of leadership outcomes (e.g., leader emergence, leadership effectiveness) and are more strongly associated with agentic factors like self-efficacy and self-expectations (Badura et al., 2020). Given this, we predict that enhancing individuals’ PILTs will be positively associated with their social-normative and affective-identity motivations to lead, but not their non-calculative motivation.
PILTs, Leadership Self-Efficacy, and Motivation to Lead
Building on hypotheses 1–5, which investigate the impact of the attribute conditioning intervention on PILTs, leadership self-efficacy, and motivation to lead individually, we also expect that the relationships between these variables will be interconnected. Self-expectations enhance motivation (Eden, 1984), and self-efficacy is often identified as a mediating mechanism of this relationship (e.g., Eden & Kinnar, 1991; McNatt & Judge, 2004). There is consistent support for the role of self-efficacy in enhancing motivational outcomes in organizational settings (Anderson et al., 2008; Paglis, 2010; Paglis & Green, 2002). Activating PILTs via attribute conditioning is likely to enhance individuals’ leadership self-efficacy beliefs by reinforcing their internalized perceptions of possessing leadership-relevant attributes (Fiske & Taylor, 1991; Lord & Maher, 1993). Thus, to the extent that individuals perceive themselves as embodying positive leadership schemas (PILTs), they should experience greater leadership self-efficacy, which in turn increases their motivation to pursue and fulfill leadership roles. Specifically, we expect leadership self-efficacy to mediate the relationships between PILTs and both social-normative and affective-identity motivations.
The Present Study
The present study utilizes a mobile application to deliver an attribute conditioning intervention aimed at enhancing individuals’ positive leader self-expectations, self-efficacy, and motivation. We test whether pairing the self with positive leader attributes increases leader-relevant self-cognitions, including leadership self-efficacy and motivation to lead.
Methodological Contribution
This study validates the use of a practical attribute conditioning leadership intervention that is scalable, portable, and low cost: a gamified mobile application. Second, it advances attribute conditioning as a theoretically grounded technique for shifting leader relevant cognitions, connecting the growing literature on attribute conditioning with applied leader development practice (Förderer & Unkelbach, 2015; Unkelbach & Förderer, 2018).
Mobile application interventions have been utilized for evaluative conditioning (e.g., Conroy & Kim, 2021; Kosinski, 2019; McGregor et al., 2023), a process through which the positive or negative valence of one stimulus is transferred to a neutral or self-relevant stimulus through repeated pairings (e.g., Conroy & Kim, 2021; Kosinski, 2019; McGregor et al., 2023). Mobile conditioning interventions have been successfully applied in mental and physical health domains (Conroy & Kim, 2021; Kosinski, 2019), and, more recently, for the enhancement of follower self-concept through gamified attribute conditioning (Leung et al., 2025). This study is the first to utilize attribute conditioning for leader development. This method contributes to our understanding of one potential mechanism underlying individual self-concept change and motivation, offering a feasible approach to self-directed leader development.
Methods
Participants
We recruited 159 undergraduates from a large public university research pool. Participants averaged approximately 3.02 years of prior leadership experience, consistent with an emerging leader population that is an appropriate target for early-stage leader development interventions. Undergraduate samples are commonly used to study leader identity and leadership development, and emerging leaders in college contexts are often actively engaged in leadership and identity formation processes, making them a conceptually relevant population for interventions that shape nascent leader self-perceptions (Day et al., 2014; Dugan & Komives, 2010; Komives et al., 2006; Offermann et al., 1994).
After removing outliers, the sample (n = 153) consisted of 122 females (80%) and 31 males (20%) 1 . Ethnic identities represented were Asian American (41%, n = 62), Hispanic or Latino (35%, n = 54), Caucasian (10%, n = 15), African American (3%, n = 4), and Hawaiian/Pacific Islander (<1%, n = 1). Seventeen participants (11%) selected “other”. The mean age of the sample was 19.2 (SD = 2.3). Participants were randomly assigned to either the experimental (n = 76) or the control (n = 77) condition, and all participants received research credits for their participation.
Procedures
The study protocol was approved by the University Institutional Review Board. All participants provided informed consent and were informed of their right to withdraw at any time without penalty. To minimize potential influence on self-expectation, the mobile task was presented as a simple memory game, and participants received a full debrief explaining the study's purpose and safeguards. Upon study registration, participants completed a baseline (Time1) survey assessing key variables along with demographics and other measures used to obscure the key variables in the study 2 . Two weeks later, participants attended a laboratory session and were told they were participating in two unrelated tasks: a memory task to recall sequences of images and words, and a survey about self-expectation. The “memory task” was the attribute conditioning intervention, and the survey assessed participants’ Time2 reports of each of the variables of interest. For this study, participants were not told the true purpose of the intervention until after the study to avoid possible social desirability effects. Meta-analytic evidence suggests that associative conditioning effects are stronger when participants are aware of the purpose (Hofmann et al., 2010). Although future research could manipulate participant awareness to enhance effects, the present study deliberately minimizes awareness to conservatively test attribute conditioning as a mechanism for shaping self-expectations while maintaining internal validity.
Mobile Application Intervention for Attribute Conditioning
During the laboratory session, participants were provided with a smartphone preloaded with either the experimental or control version of a mobile game called TAPPIT (Figure 1) 3 . In the game, participants engaged in a memory task in which their own photograph was repeatedly paired with various attributes. Each stimulus was displayed for one second. They then completed recall prompts with three seconds to respond (e.g., “What was the third card?”), reinforcing the self-attribute pairings through active engagement with the stimulus sequence. Participants uploaded a photograph of themselves at the beginning of the session. They were instructed to maintain a neutral facial expression to minimize the activation of unintended emotional cues (e.g., nervousness or happiness). In the experimental condition, participants’ images were paired with positive leader attributes (PILTs; e.g., intelligent, dynamic, helpful, hardworking, energetic) from the validated Implicit Leadership Theories (ILTs) scale (Epitropaki & Martin, 2004, Table 1). The control condition paired their own images with neutral, non-leadership-related words (e.g., table, case) selected for semantic neutrality and lack of leadership connotations. Each participant played TAPPIT for approximately 20 min, which typically included around 50 trials. Immediately after the task, participants completed the Time2 survey.
TAPPIT gameplay.
Positive Implicit Leadership Theories.
Note. All items above were presented with the prompt to “indicate the extent to which each trait is characteristic of YOU as a leader.” These validated PILTs attributes are drawn from Offermann and colleagues (1994) and Epitropaki and Martin (2004).
Measures
Positive Implicit Leadership Theories (PILTs)
PILTs were measured with the positive attributes from the ILTs scale (Epitropaki & Martin, 2004). Participants were asked to rate on a scale of 1 to 10 the extent to which each item was characteristic of them as leaders. The PILTs scale consists of four positive dimensions (i.e., Sensitivity, Dedication, Dynamism, and Intelligence) with thirteen items: helpful, understanding, sincere, intelligent, educated, clever, knowledgeable, dedicated, motivated, hardworking, energetic, strong, and dynamic. Cronbach's alphas were .84 (Time1) and .91 (Time2). Omega coefficients were .85 (Time1) and .92 (Time2).
Leadership Self-Efficacy
Leadership self-efficacy was assessed using a five-item measure developed by Chan and Drasgow (2001). Participants rated on a seven-point scale the extent to which they disagree or agree with the following statements: “I feel confident that I can be an effective leader in most of the groups that I work with,” “I am not confident that I can lead others effectively,” “I have the ability to lead others to achieve any goal I set for them,” “Through my leadership, I can make any situation come out the way I intend it to,” and “I can use my leadership skills to deal effectively with any situation.” Cronbach's alphas were .86 (Time1) and .88 (Time2). Omega coefficients were .88 (Time1) and .91 (Time2).
Motivation to Lead
Social-Normative Motivation to Lead (SN-MTL)
Participants’ SN-MTL was measured with three items from Chan and Drasgow's (2001) motivation to lead scale. Participants were asked to rate on a seven-point scale the extent to which they disagree or agree with the following statements: “I feel I have a duty to lead others if I am asked,” “I agree to lead whenever I am asked or nominated by the other members,” and “I was taught to believe in the value of leading others.” Cronbach's alphas were .79 (Time1) and .80 (Time2). Omega coefficients were .81 (Time1) and .82 (Time2).
Affective-Identity Motivation to Lead (AFF-MTL)
Participants’ AFF-MTL was measured with three items from Chan and Drasgow's (2001) motivation to lead scale. Participants were asked to rate on a seven-point scale the extent to which they disagree or agree with the following statements: “Most of the time, I prefer being a leader rather than a follower when working in a group,” “I usually want to be the leader in the groups that I work in,” and “I have a tendency to take charge in most groups or teams that I work in.” Cronbach's alphas were .92 (Time1) and .91 (Time2). Omega coefficients were .92 (Time1) and .91 (Time2).
Non-Calculative Motivation to Lead (NC-MTL)
Participants’ NC-MTL was measured with three items from Chan and Drasgow's (2001) motivation to lead scale. Participants were asked to rate on a seven-point scale the extent to which they disagree or agree with the following statements: “I am only interested to lead a group if there are clear advantages for me,” “I would want to know what's in it for me if I am going to agree to lead a group,” and “I never expect to get more privileges if I agree to lead a group.” Cronbach's alphas were .60 (Time1) and .73 (Time2). Omega coefficients were .74 (Time1) and .81 (Time2).
Results
The means and standard deviations of PILTs, leadership self-efficacy, and motivation to lead for each condition at Time1 and Time2 are shown in Table 2. The correlations among all study variables are shown in Table 3. All results were analyzed using R version 4.4.2 (R Core Team, 2024).
Means and Standard Deviations of PILTs, Leadership Self-Efficacy, and Motivation to Lead by Condition.
Note. PILTs = Positive Implicit Leadership Theories; LSE = Leadership Self-Efficacy; AFF-MTL = Affective-Identity Motivation to Lead, SN-MTL = Social-Normative Motivation to Lead, NC-MTL = Non-Calculative Motivation to Lead.
Correlations among Variables.
Note. **p < .001 *p < .05; PILTs = Positive Implicit Leadership Theories; LSE = Leadership Self-Efficacy; SNMTL = Social-Normative Motivation to Lead; AffMTL = Affective-Identity Motivation to Lead; NCMTL = Non-Calculative Motivation to Lead.
Baseline Analysis
Baseline comparison of the control (n = 77) and experimental (n = 76) groups did not indicate any significant differences on the variables of interest at Time1, suggesting that random assignment was successful. There were no significant differences between the experimental and control groups on PILTs (t(144) = −.65, p = .519), leadership self-efficacy (t(147) = −.09, p = .926), social-normative motivation to lead (t(133) = .44, p = .662), affective-identity motivation to lead (t(150) = .09, p = .926), or non-calculative motivation to lead (t(151) = .48, p = .633).
Hypothesis Testing
PILTs
To examine the effect of the attribute conditioning intervention on participants’ PILTs, we regressed participants’ Time2 PILT ratings on their conditions (1 = experimental, 0 = control) while controlling for baseline (Time1) PILTs. The results of the multiple regression model were significant, F(2, 150) = 156.4, p < .001, Adj. R2 = .67 (see Table 4). When controlling for participants’ Time1 PILTs, experimental condition significantly predicted participants’ Time2 PILT ratings, b = .33, 95% CI = [.16, .50]. This effect was medium to large in magnitude (d = .61) 4 . These findings support hypothesis 1, suggesting that participants in the experimental condition (those who paired their photos with PILTs) rated themselves more positively as leaders than participants in the control version (those who paired their photos with neutral words).
Results of Multiple Regression Models.
Note. PILTs = Positive Implicit Leadership Theories; LSE = Leadership Self-Efficacy; SNMTL = Social-Normative Motivation to Lead; AffMTL = Affective-Identity Motivation to Lead; NCMTL = Non-Calculative Motivation to Lead.
Leadership Self-Efficacy
Next, we regressed Time2 leadership self-efficacy scores on condition, while controlling for baseline (Time1) leadership self-efficacy. The multiple regression model was significant, F(2, 150) = 171.37, p < .001, Adj. R2 = .69. After controlling for participants’ Time1 leadership self-efficacy, experimental condition significantly predicted Time2 leadership self-efficacy ratings, b = .21, 95% CI = [.01, .41]. This corresponds to a small standardized effect (d = .25) 5 . These findings support hypothesis 2, suggesting that participants in the experimental condition (those conditioned with PILTs) reported higher leadership self-efficacy than the control group. Of note, a sensitivity analysis (including the 6 participants who were previously excluded as outliers) indicated that the association between condition and leader-self efficacy was no longer significant, though still positive, b = 0.16, p = .12, 95% CI = [−.04, .36]. We therefore interpret support for Hypothesis 2 as provisional, warranting replication and larger samples to precisely estimate this relationship.
Motivation to Lead
We conducted multiple regression models regressing each component of motivation to lead (Chan & Drasgow, 2001) on experimental condition. The model for social-normative motivation to lead (SN-MTL) was significant and of moderate effect size, F(2, 150) = 108.14, p < .001, Adj. R2 = .59, d = .45. When controlling for participants’ Time1 SN-MTL, experimental condition was a significant predictor of participants’ Time2 SN-MTL scores, b = .35, 95% CI = [.12, .58] 6 . These findings support hypothesis 3, suggesting that participants in the experimental condition (conditioned with PILTs) report stronger SN-MTL than those in the control group.
Contrary to hypothesis 4, we did not find a significant relationship between experimental condition and Time2 affective-identity motivation to lead (AFF-MTL). While the overall regression model was significant, F(2,150) = 219.19, p < .001, Adj. R2 = .74, when controlling for Time1 AFF-MTL, experimental condition was not a significant predictor of Time2 scores, b = .05, p = .64. These results suggest that the degree to which participants enjoy being in leadership roles and see themselves as leaders did not differ significantly between the experimental and control conditions.
Lastly, the overall multiple regression model regressing Time2 non-calculative motivation to lead (NC-MTL) on experimental condition was significant, F(2,150) = 74.47, p < .001, Adj. R2 = .49. However, when controlling for Time1 NC-MTL, experimental condition did not significantly predict participants’ Time2 NC-MTL ratings. These findings align with hypothesis 5, suggesting that the degree to which participants viewed leadership opportunities positively despite potential costs or benefits did not differ between the experimental and control conditions.
Mediation Analyses: Relationships Between PILTs, Leadership Self-Efficacy, and Motivation to Lead
Following the theoretical basis of the first step of the Galatea model, we tested whether PILTs indirectly influenced SN-MTL and AFF-MTL through leadership self-efficacy. This ordering is theory driven: stronger association of the self with PILTs is expected to increase leadership self-efficacy, which in turn shapes intentions and motivation, as described in expectancy, social cognitive, and planned behavior accounts (Ajzen, 1991; Bandura, 1997; Eden & Ravid, 1982; Rosenthal & Jacobson, 1968). The analyses were computed using the R package bruceR, and specifically the “PROCESS” function (Bao, 2024), which follows the PROCESS macro (Hayes, 2013). Each model used bootstrapping procedures with 10,000 bootstrapped samples, and the full results can be seen in Table 5.
Results of Mediation Analysis.
Note. Bootstrap sample size = 10,000. Estimates are standardized. CI = confidence intervals. PILTs = Positive Implicit Leadership Theories; LSE = Leadership Self-Efficacy; SNMTL = Social-Normative Motivation to Lead; AffMTL = Affective-Identity Motivation to Lead.
Analytic Caveat for Mediation Analyses
These mediation analyses evaluate attitudinal pathways (PILTs → leadership self-efficacy → motivation to lead). Because we do not measure behavioral performance in this study, these analyses provide evidence for the proximal mechanism only and are not evidence of downstream behavioral change, although such attitude shifts may be mechanisms of downstream behavior (Avolio et al., 2009; Eden & Ravid, 1982; Kierein & Gold, 2000; McNatt, 2000; McNatt & Judge, 2004). In addition, although our randomized assignment allows causal inference for the immediate effect of the conditioning treatment on post-treatment attitudinal outcomes, mediation estimates require careful interpretation. Mediation analyses provide the strongest causal evidence when the mediator is experimentally manipulated or measured at a distinct time point from the outcome (Imai et al., 2010; Maxwell & Cole, 2007; VanderWeele, 2015). In this case, the mediator and outcome were assessed in a pre-test and a single post-test wave, so observed indirect effects should be read as evidence that changes in self-efficacy are associated with changes in motivation, but not as definitive evidence that self-efficacy causally impacted motivation via Galatea effect. We therefore present indirect effects as descriptive indices of a plausible pathway, though future designs must separate treatment, mediator, and outcome in time to provide a stronger test of the mediation pathway.
Social-Normative Motivation to Lead
We conducted three mediation analyses to examine the mediating effect of leadership self-efficacy on the relationship between PILTs and SN-MTL. In the full sample (Figure 2), PILTs significantly predicted leadership self-efficacy (path a = .46, p < . 001), which predicted SN-MTL (path b = .46, p < .001), yielding a significant indirect effect (ab = .21, p < .001). Because the direct effect of PILTs on SN-MTL was no longer statistically significant when leadership self-efficacy was included in the model (c′ = .06, p = .471), we consider this to be a partial mediation (Adj. R2 change = .16).
Time2 model with all participants. Leadership self-efficacy (LSE) as a mediator of PILTs and social-normative motivation to lead (SNMTL). The values represent standardized coefficients derived from a bootstrap procedure. *** p < .001
The second mediation model, which included the Time2 ratings of only the experimental condition, showed a similar pattern (Figure 3). PILTs significantly predicted leadership self-efficacy (path a = .48, p < .001), which predicted SN-MTL and yielded a significant indirect effect (path b = .45, p < .001; ab = .22, p = .005). Again, the direct effect of PILTs on SN-MTL was not statistically significant when leadership self-efficacy was included (c′ = .10, p = .363), indicating partial mediation (Adj. R2 change = .15).
Time2 model with experimental participants only. Leadership self-efficacy (LSE) as a mediator of PILTs and social-normative motivation to lead (SNMTL). The values represent standardized coefficients derived from a bootstrap procedure. *** p < .001, ** p < .01
The third mediation model included the Time2 ratings of only the control condition (Figure 4). There was not a significant direct or total effect between PILTs and SN-MTL (path c = .13, p = .287; c′ = −.06, p = .581). Yet, the path from PILTs to leadership self-efficacy (path a = .41, p < .001) and from leadership self-efficacy to SN-MTL (path b = .48, p < .001; ab = .20, p = .006) were both significant, indicating an indirect association operating through leadership self-efficacy. Current mediation theory recognizes that an indirect effect can exist in the absence of a significant total effect for several reasons, including limited power in subgroup analyses, opposing direct and indirect influences, or measurement error (Hayes, 2013; MacKinnon, 2008; Zhao et al., 2010). We therefore present the control group results as a suggestion of an indirect-only effect, but we interpret this result with caution. The overall mediation observed in the full sample appears to be driven primarily by the experimental condition, suggesting that the attribute conditioning manipulation strengthened the mediated pathway. Taken together, these findings provide conditional support for Hypothesis 6a while underscoring the need for longitudinal or mediator manipulation tests to more accurately assess the causal mediation pathway.
Time2 model with control participants only. Leadership self-efficacy (LSE) as a mediator of PILTs and social-normative motivation to lead (SNMTL). The values represent standardized coefficients derived from a bootstrap procedure. *** p < .001
Affective-Identity Motivation to Lead
Though multiple regression analyses did not support hypothesis 4 (individuals conditioned to associate themselves with PILTs will report higher AFF-MTL than those in the control condition), we investigated leadership self-efficacy as a mediator between PILTs and AFF-MTL at Time2 (Figure 5). For the full sample, PILTs significantly predicted leadership self-efficacy (path a = .46, p < .001), which in turn predicted AFF-MTL (path b = .61, p < .001; ab = .28, p < .001). When leadership self-efficacy was included in the model, the direct path became non-significant (c′ = .001, p = .981), again suggesting that leadership self-efficacy partially mediates the relationship between PILTs and AFF-MTL (Adj. R2 change = .29).
Time2 model with all participants. Leadership self-efficacy (LSE) as a mediator of PILTs and affective-identity motivation to lead (AffMTL). The values represent standardized coefficients derived from a bootstrap procedure. *** p < .001
The mediation model for the experimental group (Figure 6) also supported the hypothesis that participants’ Time2 PILTs significantly predicted leadership self-efficacy (path a = .48, p
Time2 model with experimental participants only. Leadership self-efficacy (LSE) as a mediator of PILTs and affective-identity motivation to lead (AffMTL). The values represent standardized coefficients derived from a bootstrap procedure. *** p < .001, **p < .01
Time2 model with control participants only. Leadership self-efficacy (LSE) as a mediator of PILTs and affective-identity motivation to lead (AffMTL). The values represent standardized coefficients derived from a bootstrap procedure. *** p < .001, **p < .01
Discussion
The present research provides evidence that a brief mobile attribute conditioning intervention can influence leader-related self-cognitions. Participants exposed to repeated pairings between their own image and positive leader attributes reported higher PILTs, greater leadership self-efficacy, and higher social-normative motivation to lead than participants in the control condition. These findings demonstrate that leader-related self-schemas can be influenced through a scalable digital intervention, challenging the idea that implicit leadership schemas are immutable and align with recent work showing that accessible self-relevant attributes can change with targeted interventions (Dasgupta, 2013; Epitropaki et al., 2013; Kurdi & Banaji, 2017; Mann et al., 2020; Offermann et al., 1994). Importantly, the present study does not test downstream leadership behavior or performance. Instead, it provides an initial empirical step toward understanding whether leader-relevant self-cognitions can be deliberately shaped through associative learning techniques.
The findings are theoretically grounded in expectancy theory, social cognitive accounts, and the Theory of Planned Behavior. Changes in expectations and attitudes have been shown to influence intentions and performance in both classic Pygmalion work and leadership research (Ajzen, 1991; Avolio et al., 2009; Rosenthal & Jacobson, 1968). In the present study we observed patterns consistent with an indirect pathway from elevated PILTs to motivation via increased leadership self-efficacy. However, because leadership self-efficacy and motivation were measured at the same posttest wave, these mediation results must be interpreted as provisional. Establishing causal mediation and durable change requires longitudinal dosing, temporal separation of measures, or experimental manipulation of the mediator in future work.
Hypothesis 4 was not supported. Affective-identity motivation to lead, which captures enjoyment of and identification with leadership roles, did not change significantly after a single session of conditioning. These results suggest that the degree to which participants enjoy being in leadership roles and see themselves as leaders did not differ significantly between the experimental and control conditions, as the majority of variance in Time2 scores are more strongly associated with baseline AFF-MTL. These baseline scores may suggest a potential ceiling effect and point to the need for longer or more identity-focused interventions to meaningfully shift AFF-MTL. Affective-identity motivation is closely tied to deeper self-concept and emotional meaning, which may help explain why short-term conditioning increased sense of duty and responsibility but did not substantially alter intrinsic enjoyment or self-identification with leadership. We therefore recommend testing longer or repeated dosing schedules, integrating reflective narrative exercises, and encouraging daily practice to allow gradual consolidation of affective related motivation. The mobile format is particularly well suited for distributed practice and booster sessions, which theory and evidence suggest support retention and transfer (Förderer & Unkelbach, 2015; Hofmann et al., 2010; Lyubomirsky & Layous, 2013).
Limitations
Although this experiment offers initial causal evidence that brief mobile attribute conditioning can shift leadership-relevant cognition and motivation, several limitations qualify our conclusions and indicate priorities for future work. First, the design tests immediate, proximal effects after a single session, so we cannot speak to durability or accumulation of change. Durability in related interventions typically depends on dosing, booster sessions, and contextual supports, and therefore longitudinal, repeated dosing studies are required to establish lasting self-schema change or behavioral impact (Förderer & Unkelbach, 2015; Rasera et al., 2022). The mobile, microlearning format of the intervention makes this a feasible future direction for studying how frequent practice, spaced repetition, and app-based boosters and reflective prompts lead to consolidation and transfer (Cepeda et al., 2006; Donker et al., 2013; Lyubomirsky & Layous, 2013).
Another important limitation concerns the possibility of demand characteristics. Because participants were repeatedly exposed to positively valenced leader attributes, some participants may have inferred the purpose of the intervention and adjusted their self-ratings accordingly. Though post-hoc interviews suggested that participants were unaware of the study's purpose, this possibility cannot be completely ruled out. Future studies could employ stronger masking procedures, indirect outcome measures, or incorporate behavioral tasks to further reduce potential demand effects.
Another limitation concerns the possibility that the observed effects reflect short-term activation rather than durable associative learning. Because outcomes were measured immediately after the intervention, the design cannot fully distinguish between transient priming, heightened accessibility of self-relevant attributes, and longer-term associative change. This distinction is important because many simple psychological interventions, including priming effects, have shown inconsistent replication in recent years (Mac Giolla et al., 2024; Open Science Collaboration, 2015). Replication across independent samples, longer follow-up intervals, and designs that test mechanisms such as blocking or extinction (Imai et al., 2010; Unkelbach & Förderer, 2018) will be necessary to determine whether attribute conditioning produces durable changes in leader self-schemas. At the same time, short-lived activation may still be practically valuable if it prompts leaders to engage in reflective or experiential practice that supports schema consolidation.
Another potential limitation is the possibility that the gamified task contributed to self-efficacy through enactive mastery experiences (e.g., success in the games enhanced individuals’ self-efficacy). However, because only the experimental group showed increases in leadership self-efficacy, the data suggest that pairing the self with leadership attributes was the key driver rather than game success alone.
Generalizability is limited by the emerging leader sample. Participants had an average of 3.02 years of leadership exposure, but field replications with practicing leaders and culturally diverse samples are needed to confirm external validity. Prior work suggests ILT structure is conceptually similar across students and adults (Offermann et al., 1994). Additionally, the cognitive mechanism targeted by attribute conditioning does not rely on extensive prior leadership experience; attribute accessibility can be shifted through repeated pairings even when practical leadership exposure is limited (Förderer & Unkelbach, 2015; Olson & Fazio, 2006). For transparency, we report participants’ prior leadership experience and used it in robustness checks; still, field-based replications with practicing organizational leaders are necessary to establish external validity further.
Finally, some measures showed limited variance or lower reliability and ceiling effects, and unobserved confounders or common method variance remain possible threats (Podsakoff et al., 2003; Spector et al., 2019). Future studies should use expanded item pools, multi-method outcomes, and robustness tests to improve sensitivity and causal inference. As a whole, the aforementioned limitations define a focused empirical agenda that would allow progress from proximal cognitive shifts to broader claims about self-concept consolidation and behavioral impact.
Directions for Future Study
The present findings should be interpreted in light of extensive evidence documenting gaps between intentions, self-cognition, and behavior. Recent work indicates that changes in attitudes and self-efficacy often translate to modest behavioral effects (e.g., Conner & Norman, 2022), and that simple psychological interventions may not reliably replicate or generalize across contexts (Camerer et al., 2018; Open Science Collaboration, 2015). Therefore, the present study should be viewed as an initial demonstration of short-term changes in leader-related self-cognitions rather than evidence of durable leadership development or behavioral change. This opens the door for many potential directions for future study.
Future work should explicitly address the heterogeneity of ILTs while testing the practical utility of attribute conditioning. Empirical work shows that ILT factor structure is broadly consistent across samples, yet endorsement of prototypes can vary by context, gender, and culture (Epitropaki et al., 2013; Epitropaki & Martin, 2004; Offermann et al. 1994). ILT measures may capture beliefs about leaders in general, prototypes of the typical leader, or self-schemas that reflect how an individual sees themself as a leader. Because our intervention pairs self-referent stimuli with leader attributes, it primarily targets self-relevant attribute accessibility rather than general beliefs about others. Future studies should therefore test whether conditioning effects differ for self-referent versus other-referent pairings, and whether tailoring attribute sets or permitting participants to choose their own attributes increases effectiveness.
Attitude and self-expectation shifts are theorized to be potential antecedents of intentions and behavior, but they are not sufficient to demonstrate enacted leadership (Ajzen, 1991; Avolio et al., 2009). Future research should include performance tasks, objective workplace metrics, peer and supervisor ratings, and ecological momentary assessments to capture downstream behavior change. Multi-method approaches will also reduce common method variance and improve causal inference.
Finally, researchers should evaluate cost effectiveness, uptake, fidelity, and ethical considerations of scaling an intervention that pairs the self with positive leader attributes. Pilot trials comparing TAPPIT integrated with reflection, coaching, or experiential exercises will clarify the pathways from proximal cognitive activation to sustained change in leadership self-schemas and observable leader behavior. Addressing these issues will determine whether brief mobile conditioning can become a robust, scalable component of leader development programs.
Practical Implications and Recommendations
The mobile delivery format of this intervention offers several practical benefits that address common barriers to scalable leader development. Because the application is built to be self-administered, organizations can deploy it at scale without large facilitator investments or lost work time for more structured trainings (Avolio et al., 2009). The format of the study sessions, brief modules delivered in a single 20-min session with roughly 50 trials, maps naturally onto busy leader schedules and supports distributed practice, which improves retention and transfer (Cepeda et al., 2006). The app format enables easy implementation of booster sessions and longer dosing schedules, an approach that has also proved useful for sustaining effects (Franklin et al., 2016; Rasera et al., 2022). When paired with structured reflection activities, the application may help leaders translate conditioned attribute associations into concrete intentions and behaviors, offering a pragmatic route to integrate cognitive conditioning with experiential and reflective approaches commonly used in leadership and self-development (Kolb, 1984; Lyubomirsky & Layous, 2013).
Organizations seeking to adopt TAPPIT may begin byembedding brief (3–5 min) modules within existing development sequences such as onboarding, emerging leader programs, or leader development trainings, potentially improving retention and transfer to practice (Johansson & Andersson, 2024). These concise exercises should be paired with structured reflection sessions in which participants link the app's conditioned attributes to real-world leadership challenges, fostering deeper self-awareness and schema integration (Lyubomirsky & Layous, 2013). To sustain these gains, periodic booster sessions within the app can refresh attribute associations and prevent training decay (Rasera et al., 2022). Finally, customizing the attribute word sets to mirror organizational values or inviting leaders to co-create their own target trait lists may enhance engagement and intrinsic motivation, which could yield up to a 40 percent increase in development ROI (Mikucki, 2023). Each of these recommendations should be empirically studied to identify the extent to which these strategies can help organizations achieve immediate improvements in motivation to lead while laying the groundwork for lasting gains in leader self-concept and self-efficacy.
Conclusion
The present study advances leadership development theory by providing initial evidence that leadership self-expectations can be enhanced through scalable, ethical, and self-directed mobile interventions. By introducing attribute conditioning as a potential tool for leader development, this work highlights a promising avenue for studying how mobile interventions may complement traditional leadership development approaches. This approach provides a pragmatic solution for organizations seeking accessible, cost-effective means of cultivating leadership capacity and motivation across a broad workforce. By illuminating a new pathway for leader self-development through mobile attribute conditioning, this research opens fertile ground for both theoretical refinement and innovative practice in leadership science.
Footnotes
Note
This article is based on the second author's doctoral dissertation.
Ethics Approval and Informed Consent
The study was approved by the University of California, Riverside IRB (approval # 12092) and renewed on 9/27/2024. All participants provided written informed consent prior to participating.
Author Contributions
Laura Ashlock: Formal Analysis, Writing – Original Draft
Alex Leung: Conceptualization, Investigation, Methodology, Writing - Original Draft
Thomas Sy: Methodology, Writing - Review & Editing, Supervision
Kashyap Panda: Methodology, Investigation, Writing – Review & Editing
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability
The datasets generated during and analyzed during the current study are available from the corresponding author on reasonable request.
1.
Participants who provided extreme data (i.e., Cook's distance beyond the threshold of 4/n) on two or more of the variables of interest were removed. This resulted in a total of 6 participants being excluded.
2.
In addition to the listed measures for each of the key study variables, participants also completed measures of implicit followership theories (Sy, 2010), personal mastery (Lachman & Weaver, 1998), life satisfaction (Riverside Life Satisfaction Scale; Margolis et al., 2019), self-uncertainty (Rast et al., 2012), subjective happiness (Lyubomirsky & Lepper, 1999), and self-esteem (Rosenberg Self-Esteem Scale; Rosenberg, 1965). These measures served to reduce social desirability by obscuring the key variables in the study.
3.
4.
A sensitivity analysis (including the 6 participants who were previously excluded as outliers) indicated that the effect of condition on PILTs remained significant, b = .35, 95% CI = [.17, .52]. The effect of condition on PILTs also remained significant when controlling for years of leadership experience, b = 0.33, 95% CI = [.16, .50].
5.
When controlling for years of leadership experience, the effect of condition on leadership self-efficacy remained significant, b = 0.22, 95% CI = [.03, .41].
6.
A sensitivity analysis (including the 6 participants who were previously excluded as outliers) indicated that the effect of condition on SN-MTL remained significant, b = .33, 95% CI = [.10, .56]. In addition, the effect of condition on SN-MTL remained significant when controlling for years of leadership experience, b = .35, 95% CI = [.13, .58].