As Predictors of Implicit Leadership Theories

This study seeks to investigate how employed individuals form Implicit Leadership Theories (ILTs), or personal assumptions about the characteristics of a business leader. The existing literature consistently explains that employed individuals, through socialization and past experiences with leaders, develop ILTs. However, such studies have neglected to examine directly how one’s past leaders have affected the formation process, therefore leading to a hypothesis examining this formation process. As a competing hypothesis, subordinates’ personalities will also be assessed on the extent to which subordinates form ILTs congruent with their own personality and how having a leadership self-image affects this relationship. Lastly, it is also predicted that subordinates share a common element when assessing ILTs, such that they will rate participative leadership behaviors higher than any other leadership behaviors.

Personality and Past Experiences

As Predictors of Implicit Leadership Theories

Introduction

Employees’ perceptions, prior expectations, and cognitive prototypes regarding the leadership process have dominated part of the leadership literature (Foti & Lord, 1987; Kenney, Schwartz-Kenney, & Blascovich, 1996; Larson, 1982; Lord & Maher, 1993). Based on such literature, it has been determined that working individuals, through socialization and past experiences with leaders, develop Implicit Leadership Theories (ILTs), or “personal assumptions about the traits and abilities that characterize an ideal business leader” (Epitropaki & Martin, 2004 – PAGE NUMBER). ILTs stem from cognitive structures, or schemas, that specify traits and behaviors that followers believe an ideal leader should exhibit. They are stored in the memory, and when followers interact with a person in a leadership position, such schemas become stimulated (Kenney et al., 1996). These leadership schemas provide organizational members with a cognitive basis for understanding and responding to supervisor behavior, and they are essential elements of “organizational sense-making” (Poole, Gioia, & Gray, 1989; Weick, 1995).

The potential role of ILTs within organizational settings has been highlighted in the literature (Bass & Avolio, 1989; Epitropaki & Martin, 2004; Lord & Maher, 1993; Offermann, Kennedy, & Wirtz, 1994). Implicit leadership theories have been represented as a recognition-based approach to leadership (Lord, 1985). Based on this approach, employees compare their implicit leadership theories with their leader’s traits and behaviors (Calder, 1977). This matching process is expected to form the impressions employees hold for their leaders. Each individual develops a unique schema (which is sometimes referred to as a prototype) on what traits and behaviors constitute an ideal leader, specifically in the business world. A prototype is an “abstract conception of the most representative member or most widely shared features of a given cognitive category” (Phillips, 1984, p. 126). These leadership prototypes are said to be formed through exposure to social events, interpersonal interactions, and prior experiences with leaders (Epitropaki & Martin, 2004). Therefore, the general definition of ILTs has led to the belief that past experiences predict the formation of such theories.

Such prototypes are formed by role schemas (or “normative expectations”) that allow followers to understand and interpret leaders’ traits and behaviors (Calder, 1977). As such, it is proposed that leadership is a common label applied to the traits and behaviors congruent with the observer’s ILTs. When a person exhibits potential leadership characteristics (as defined by the ILT), subordinates assess the potential leader for distinctiveness (as in distinct leadership attributes and behaviors) regarding that behavior (Calder, 1977). If the behavior is distinguishable from the behaviors of other group members, then leadership may be attributed to the person.

Similarly, it has been suggested that the perception of leadership involves the activation of a “leader category” (Phillips & Lord, 1981). As such, a person’s attributes and/or behaviors are compared to “prototypical leader characteristics,” activating the leader category if there is a match between attributes and the prototypical attributes and behaviors. In this regard, Lord and his associates (e.g., Lord, 1985; Lord & Maher, 1993; Phillips & Lord, 1981) have contributed to the ILT field. On the basis of Rosch’s (1978) theory of cognitive categorization, such literature suggested a categorization theory to leadership and argued that leadership perceptions form a number of hierarchically organized schemas or cognitive categories, each of which is represented by a set of prototypes.

Additionally, once a person is categorized as a leader, “the activated leader prototype” causes followers to selectively attend to, encode, and retrieve schema-consistent information and to provide consistent information where such information does not exist” (Phillips & Lord, 1981 cited in Kenney, Blascovich, & Shaver, p.411). Moreover, Lord and others (Lord et al., 1984; Phillips & Lord, 1982) have found that people use categories to differentiate between leaders and non-leaders and between effective and ineffective leaders, which seems essential to the formation of ILTs.

Other cognitive categorization theories (e.g., Barsalou, 1985; Smith & Medin, 1981) however, argue that classification occurs as observers compare stimuli with ideals or specific examples stored in memory. An ideal includes characteristics that category members should exhibit if they are to serve as a representative to the category (Barsalou, 1985). As such, a category member’s prototypicality increases with its similarity to the category’s ideal. Therefore, in some cases, categories may be organized around ideal prototypes rather than typical prototypes. For example, a follower might judge a leader based on an ideal notion (ILT) when evaluating whether a leader is worthy of influence (Barsalou, 1985). Similarly, the closer the stimulus is to a category’s ideal, or the more category exemplars the stimulus resembles, the more likely the observer will classify the stimulus as a member of that category (Kenney, Schwartz-Kenney, & Blascovich, 1996). In this example, the category is a leader who is worthy of influence.

It has also been argued that there is a possible feedback loop between employees’ categorization of a supervisor as congruent to the implicit leadership profile and their perceptions of the supervisor’s behavior (Lord & Maher, 1993). Leadership categorization processes can have an influence on perceptions of actual leader behavior, but they are also likely to be affected by a person’s general day-to-day experiences with a manager. It might be through their exposure to their actual leader’s behaviors that employees engage in a categorization process and make active comparison between their actual supervisor and the implicit profile stored in memory. Furthermore, perceivers may then rely on existing categorizations and evaluations of congruence to further simplify the processes required to recognize leadership in others. Once someone is categorized as close or distant to a perceiver’s ILTs, the relevant leadership schema generates further assumptions about the person’s behavior, affecting perceptions of that behavior. Additionally, once people attach a label to an object, person, or event, this process of categorization guides how they interpret much of the subsequent information they encounter concerning that object, person, or event. Over time, this additional information tends to strengthen the original categorization (Lord & Maher, 1993).

Measure of ILTs and Effects on Validity

To date, there is no single and widely accepted measure of implicit leadership theories. Several researchers have developed independent lists of traits to measure ILTs, but such lists are rarely replicated in other research, and they often use different clusters of traits. However, attributes such as intelligent, honest, dynamic, and motivated seem to exist in all the lists elicited, and a distinction between positive (prototypic) and negative (anitprototypic) traits has been made by most researchers (Epitropaki & Martin, 2004; Offermann, Kennedy, & Wirtz, 1994). In addition, ILT trait lists can be fairly long, as in Lord et al.’s (1984) scale of 59 items, the Schein Descriptive Index’s (SDI; Schein, 1973) 92 item scale, and Offermann et al.’s (1994) scale of 41 items (Epitropaki & Martin, 2004). Other scales include the Campbell Leadership Indicator (CLI; Campbell, 1991), Kenney et al.’s (1996) Leaders Described as Worthy of Influence, and House et al.’s (1999) Culturally Endorsed Implicit Leadership Theories (CILTs).

It is important to note here that leadership instruments as a whole have used both traits and behaviors to measure the construct. Specifically, the trait approach emphasizes attributes of leaders such as personality, motives, values, and skills, whereas the behavior approach emphasizes the importance of what leaders actually do (e.g., activities, responsibilities, functions, etc.; Yukl, 2002.). However, behavioral theories are often considered a stronger approach in leadership literature, and thus will be the approach taken in the present study. Specifically, a meta-analysis examining the Ohio State leadership behaviors of Consideration and Initiating Structure dimensions illustrated important support for the validity of Initiating Structure and Consideration in leadership research (Judge, Piccolo, & Ilies, 2004).

Generalizability of ILTs

A variety of individual difference variables have been considered as potentially being related to ILTs. The present study primarily examines personality as individual differences in relation to ILTs. If individual differences do not affect ILTs, then the ILTs would be generalizable across different groups of people. Such generalizability has been examined in relation to some groups of people, including gender, type of role or job, and culture.

However, the generalizability of ILTs is supported by contradictory evidence. Generalizability findings have mainly focused on the areas of consistency across gender (Deal & Stevenson, 1998; Offermann et al., 1994), consistency from students to professional samples (Offermann et al., 1994), and consistencies across cultures (Bryman, 1987; House et al., 1999). All three areas are generally supported across the generalizability studies, as findings support claims that ILTs remain unaffected by individual and contextual differences, and the degree to which employees resort to categorical thinking and use ILTs as a map to evaluate their actual supervisor’s behavior does not vary as a function of context or individual differences (Epitropaki & Martin, 2005).

However, other findings argue that some variations do exist. For example, even though men and women have similar perceptions of prototypic leaders, men rated traits such as “aggressive, competitive, and feelings not easily hurt” higher than women did, and women rated traits such as “being aware of others’ feelings, helpful, and self-confident” higher than men did (Deal & Stevenson, 1998). In other words, one group’s ILT is not necessarily the same as another group’s.

In addition, there is much discrepancy in ILT ratings in the cross-cultural literature on leadership. For instance, Gerstner (1994) compared leadership prototypes of “a business leader” across several countries using an attribute-rating task. Results indicated that there are reliable differences in the prototypical leadership perceptions of members from the various countries sampled. Such findings were built upon other literatures asserting that perception is not solely an innate, physiological function of the cognitive process, but is also a subjective process reflecting the self, including cultural background (Markus & Kitayama, 1991). Likewise, it was also previously found that characteristic traits of a leader in one culture may be very different from prototypical traits in another culture (Shaw, 1990). Additionally, Hofstede (1976) has proposed that the more similar two individuals are in terms of nationality, the more likely they are to perceive their social environment similarly. Overall, therefore, the generalizability claims of ILTs appear inconclusive, but pertinent information may be lacking. As such, further investigation should be undertaken.

Additionally, other hypotheses by Eptriopaki and Martin (2005) examined the degree to which people use ILTs as a benchmark to make sense of their supervisor’s behavior, and suggested that it might change as a function of context. For instance, a limited capacity model has been applied to leadership perceptions and leader behavior ratings (Epitrokpaki & Martin, 2005; Lord & Maher, 1990). This model acknowledges people’s limited memory capacity and their reliance on “general cognitive simplification mechanisms” (such as ILTs), particularly in conditions of high cognitive load. Furthermore, it was also proposed that employees in jobs of high demand will rely more on their ILTs to rate the quality of the leader-member exchange (LMX) relationship they develop with their manager (Epitropaki & Martin, 2005). It was further proposed that employees in exchanges of short duration would rely more on their ILTs to rate the quality of the relationship they develop with their supervisors than those in exchanges of long durations. Such variations have not been examined further, and there is still a clear need for future examination of all demographical variables. When considering the formation of ILTs, the present proposal will hold cultural differences constant, as only U.S. residents will serve as participants. Ethnicity and gender differences will be controlled.

As described above, the definition and illustration of ILTs often refers to past experiences playing a role in the formation of such theories. However, studies have neglected to examine directly how one’s past leaders have affected the formation process. The first hypothesis in the present study attempts to explain how the role of past experiences affects the ILT formation process by examining ratings of subordinates’ most and least effective leaders. Therefore, the following is hypothesized about the effects of experience on predicting implicit leadership theories:

Hypothesis 1: Subordinates’ past experiences with various types of leaders will predict their implicit leadership theories. ILTs will be more strongly related to descriptions of one’s most effective leader than to descriptions of one’s least effective leader.

The Formation of ILTs Expanded: The Effects of Personality

The present study also examines how ILTs could be formed by personality. Historically, results of investigations relating personality traits to leadership have been inconsistent and often disappointing. Most reviews of the literature have concluded that the trait approach has fallen out of favor among leadership researchers. The original source of skepticism with the trait approach is often attributed to Stogdill’s (1948) influential review. Although Stogdill did find some consistent relations, he concluded, “The findings suggest that leadership is not a matter of passive status or of the mere possession of some combination of traits” (Stogdill, 1948, p. 66). As Bass (1990) noted, after Stogdill’s (1948) review, “situation-specific analyses took over, in fact, dominating the field” of leadership theory and research (p. 59). Many attempts have examined the personality traits in effective leaders (Hogan, 1994; Judge & Bono, 2000; Tett & Burnett, 2003), but the definition of what comprises an effective leader may be partly responsible for theory abandonment. Perhaps this is because there are unique differences among individuals’ personalities, it seems likely that individuals’ implicit leadership theories, or their expectations of an effective leader, will be differ from person to person, and these differences could be related to personality of the person whose ILT is examined.

The ILTs approach reflects a resurgence of interest in leadership traits, but the emphasis is now placed on the perceptual processes underlying the conceptualization of leadership. When defining ideal leadership, it is important to remember that one’s personality may affect what behaviors he or she thinks are most effective for leading others (Cucina, Vasilopoulos, & Sehgal, 2005).

A similar case has already been argued for the role that supervisor’s implicit theories play on performance appraisals (Borman, 1987; Cronbach, 1955; Uggerslev & Sulsky, 2008). For instance, it has been found that raters often use their own distinctive dimensions to evaluate the performance of their employees (Borman, 1987). Such evidence has led to the implementation of frame-of-reference (FOR) training, or a training program that helps raters hold a common understanding of how to rate an organization’s performance dimensions. FOR training has been shown to help increase the accuracy of performance ratings, thereby countering the effects of supervisors’ implicit theories of performance. Therefore, if it has been found that supervisors have implicit theories based on their own personal performance ideology, it only seems logical that subordinates could too have their own personal theories on the performance of leaders.

Social Perceptions based on Similarity

Some of the ILT literature has suggested that implicit theories can provide stability to dyadic relationships when they fulfill a similarity paradigm (Epitropaki & Martin, 2001; Turban et al., 1990). For instance, the “similarity-attraction” paradigm (Byrne, 1971) states that similarity between individuals with regard to personal attributes or other characteristics is linearly related to interpersonal attraction, such that similarity gives rise to attraction while dissimilarity engenders repulsion. In the context of outcomes, it has been argued that interpersonal attraction fosters supervisor-subordinate compatibility and in turn, performance. For instance, the resulting interpersonal connection encourages accurate perception of supervisor performance expectations and, consequently, improved subordinate performance (Deluga, 1998).

In addition, there is evidence that perceptions of similarity, both in a general sense (Wexley & Pulakos, 1983) and with regard to specific attributes such as demographic characteristics (Epitropaki & Martin, 2001), attitudes (Phillips & Bedeian, 1994), values (Ashkanasy & O’Connor, 1997), competence (Kim & Organ, 1982) and personality traits (Bauer & Green, 1996; Keller, 1999), have been associated with perceptions of leader-member exchange quality. Moreover, it has been found that subordinates who regard themselves as being similar to their supervisors communicate more with them, and are rated as higher performers than those who do not (Turban et al., 1990). This may describe the previously mentioned gender differences found in terms of individual ratings on effective leader attributes. Other relationships, like mentorships, have also evidenced success based on attraction due to the similarity of race and gender, and perceived similarity of beliefs and attitudes (Turban, Dougherty, & Lee, 2001). Overall, perceived similarity in a dyad (such as a supervisor-subordinate pair) is often related to favorable or pleasurable outcomes.

Social Perceptions and Self-Serving Biases

Individuals often believe that what they do, or how they contribute at work, is important. For instance, literature on positive self-illusions suggests that individuals may prefer leaders similar to the self, because such individuals tend to hold unrealistic, positive illusions of the self (Taylor & Brown, 1988). Assuming that becoming a leader and leadership are construed as socially desirable, individuals may keep unrealistic expectations of assuming a leadership position and project their own traits onto idealized leadership images. In other words, as individuals tend to view themselves in an overly positive manner, believing that they themselves could be leaders, they may be most satisfied and most likely to want to continue working with leaders who are similar to them (Taylor & Brown, 1988). They might judge such leaders to be more effective, which is similar to ILT perceptions.

The social cognition literature also suggests that individuals engage in self-serving biases when describing attributes that are necessary for performance (Cucina, Vasilopoulos, & Sehgal, 2005). It has been suggested that individuals tend to indicate that their own personality traits are those that are necessary for successful performance. Research in this domain initially (Lewicki, 1983) examined self-serving biases when looking at prototypes of social categories. It was evidenced that if an individual has a favorable impression of a target (e.g., a leader), then he or she will be more likely to assume that the target has the same characteristics as one’s self. In addition, she will ascribe her own strengths and positive attributes to the target.

While there is currently no available literature on self-serving biases and implicit leadership theories, several studies have examined how self-serving biases affect ratings of leadership orientation (Dunning, Perie, & Story, 1991; McElwee, Dunning, Tan, & Hollman, 2001). In these studies, participants tended to rate leaders as goal- or people-oriented depending on which orientation the participant possessed. Interestingly, such self-serving biases were only present when rating leadership performance, as opposed to leadership creativity (Dunning, Perie, & Story, 1991). In addition, liking the target leader was not found to mediate or moderate the relationship between orientation of the leader and orientation of the participant (McElwee, Dunning, Tan, and Hollman, 2001).

Perhaps the trait activation model (Tett & Burnett, 2003) can assist in explaining how people develop the previously discussed self-biased belief systems, and consequently, their expectations. For instance, according to the trait activation model, a person will rate specific behaviors high on importance for successful performance, because those specific behaviors (and their underlying personality trait) are what make him or her effective. Because of such beliefs, those behaviors become valued by the individual. For example, a conscientious person will likely engage in organizational behaviors and work in a methodical fashion, because that is what he or she thinks is necessary in order to perform successfully on a job.

When a person’s traits are valued by one’s organization (or perhaps one’s leader), the person will probably experience more satisfaction with the job (Tett & Burnett, 2003). In return, an individual may expect a leader to have similar traits as the self if these traits are seen as valuable and related to performance. Applying this rationale, it is hypothesized that a self-serving belief system will affect the formation of followers’ implicit leadership theories in terms of trait-based principles. Such a hypothesis lies on the foundation that personality is the underlying driving force behind how individuals behave, or why certain types of behaviors are valued. In other words, it is speculated that a subordinate will rate how an effective leader should perform (i.e., ILT), based on implicit assumptions about the behaviors that make the subordinate most effective. It is proposed that followers will hold implicit leadership theories based on a self-serving bias: that is, a bias that successful people are like themselves. Thus, followers will expect an effective leader to exhibit behaviors congruent with the followers’ own personalities.

Hypothesis 2: Subordinates’ personalities will affect the formation of their implicit leadership theories. As such, subordinates will rate leaders who exhibit behaviors that correspond to their own personality traits as more effective than leaders whose behaviors do not correspond to the subordinate’s personality traits.

Specifically:

Subordinate conscientiousness is positively associated with rating a business leader high on initiating structure.

Subordinate agreeableness is positively associated with rating a business leader high on consideration.

Subordinate extraversion is positively associated with rating a business leader high on inspiring commitment.

Subordinate agreeableness is positively associated with rating a business leader high on participative style.

The rationale for this hypothesis rests on inferences about the relationships between specific personality traits and specific (leader) behaviors. Regarding Hypothesis 2a, the correspondence between scoring high on conscientiousness and rating a business leader high on initiating structure stems from the constructs’ definitions. For instance, because those who score high on conscientiousness are said to have high levels of thoughtfulness, with good impulse control and goal-directed behaviors (Goldberg, 1992), it seems feasible to assume that such persons would assign high ILT ratings for leader dimensions like initiating structure. Initiating structure refers to leader behaviors such as encouraging the use of uniform procedures and maintaining definite standards of performance (Stogdill, 1963), leader behaviors that parallel the definition of conscientiousness.

Similarly, because those who score high on agreeableness are said to have high levels of trust, altruism, kindness, affection, and other prosocial behaviors, it is predicted that such persons would have high ILT ratings for the leader dimension of consideration. Consideration refers to leader behaviors such as treating all group members as one’s equal and looking out for the personal welfare of group members, behaviors that seem to parallel the definition of agreeableness. Similarly, it is also predicted that those who score high on agreeableness will have high ILT ratings for the leader dimension participative style. Leaders with a participative style consult with their subordinates when facing problems and ask them for suggestions concerning how ideas should be implemented (Northhouse, 2004).

Lastly, because those who score high on extraversion are said to have high levels of excitability, sociability, assertiveness, emotional expressiveness, and optimism, it is predicted that such persons would assign high ILT ratings for leader dimensions like inspiring commitment. Inspiring commitment refers to the leader behaviors of proposing change with great enthusiasm and exerting optimism when describing a difficult activity or mission for a work unit (Yukl, 1998). These behaviors seem to parallel the definition of extraversion.

Leadership Self-Image

Self-schemas are “sets of cognitive structures that provide for individual expertise in particular domains of social behavior” (Markus, Smith, & Moreland, 1985). As “schematicism” increases in a conceptual domain, such as leadership, so does sensitivity to the behavior of others in that domain. Thinking of oneself as a leader may make one’s own characteristics more salient for judging leaders.

As a result, it will be important to note whether or not individuals see themselves as leaders, because it may affect if they expect a leader to be like them. That is, if the people think of themselves as leaders, then the Hypotheses 2 are more likely to be true. If, on the other hand, the person has no self-image as a leader and does not ever want to be a leader, the self-serving hypothesis might be less true. It would not be very self-serving to see another person being effective at something (e.g., leadership) as similar to oneself if leadership plays no part in one’s self-image. Therefore, the relationships in Hypotheses 2 should be moderated by one’s self-image regarding leadership.

Hypothesis 3: Subordinate leadership self-image will moderate the relationship between subordinate personality and formation of implicit leadership theories. The relationship between personality and the formation of implicit leadership theories will be stronger when subordinates have a leadership self-image.

A proposed model that is consistent with the preceding hypotheses can be viewed in Figure 1. By examining how ILTs may be formed, especially if one’s past experience plays a role, important implications could be explored. For instance, Eptiropaki and Martin (2005) noted the potential role ILTs could play in the socialization of newcomers. As such, ILTs can potentially have “a significant impact on the development of interpersonal relationships during the organizational socialization process” (Eptiropaki & Martin, 2005, p. 673). Subsequently, current training programs of leadership could seriously benefit by including some form of ILTs training to increase managers’ awareness of their subordinate ILTs.

Figure 1. The effects of personality and experience on predicting implicit leadership theories.

Experience with Behaviors of Multiple Leaders

Personality of Subordinates

Implicit Leadership Theory

Leadership Self-Image

Demographic Considerations

Previously mentioned generalization issues will be considered when conducting the study. Gender effects on ILTs are of interest in past research, as are age and experience. Younger and less experienced employees might have different conceptions of ideal leadership from employees with more experience and more exposure to organizational leaders (Brown & Lord, 2001). Age, education level, occupational title, and years employed will therefore be examined and controlled if necessary. In addition, ethnicity will be recorded, but the sample will be limited to United States residents to avoid previously found cross-cultural differences.

Participative Leadership

Even though there are often individual differences among people, maybe working individuals just generally prefer a participative leader over a certain “type” of leadership that is based on their personalities or past experiences. For instance, practicing participative leadership offers a variety of potential benefits, including likelihood to increase the quality of decision-making (Scully, Kirkpatrick & Locke, 1995), to contribute to the quality of employees’ work life (Somech, 2002), and to increase employees’ motivation (Locke and Latham, 1990), commitment (Armenakis, Harris, & Mossholder, 1993), and satisfaction (Smylie, Lazarus, & Brownlee-Conyers, 1996), all of which suggest that subordinates react well to participation leadership.

Participative leadership is based on the process of joint decision-making by two or more parties in which the decisions have future effects on those making them. The amount of participation by any individual is the amount of influence he or she has on the decisions and plans agreed upon (Vroom, 1959). Specifically, participative leadership involves the efforts of a supervisor to encourage and facilitate participation by subordinates when making decisions that could have been made by the manager alone. Participative leadership can take many forms, such as revising a tentative decision after receiving protests, asking for suggestions before making a decision, or allowing others to make a decision subject to a supervisor’s final authorization (Yukl, 2002). Field and House (1990) examined the validity of the Vroom-Yetton decision-making model of leadership, and found it was validated for managers, but not for subordinates. Although they were not explicitly studying ILTs, their results suggest that because the model was developed and validated on managers’ perceptions of leader styles in relation to their perceptions of effectiveness, it is biased towards managers’ implicit leadership theories. Subordinates in their study, however, perceived effectiveness simply when they perceived the leader to have a more participative decision making style. This may indicate that typical subordinates have a common element to their own ILTs, which is that participative leadership styles are effective.

The idea that people value different attitudes when they are in different positions has long held empirical support (Lieberman, 1956). In other words, a person’s attitude will be influenced by his or her role, stemming from the fact that a new role requires one to take on a new and different set of tasks and a new knowledge domain. Lieberman demonstrated how people’s attitudes change when they shift roles by making workers become foremen and by demoting foremen back into the worker role. Interestingly, when workers were made foremen, their attitudes became more “pro-management and anti-union” (p. 400). However, when they were demoted back down to the worker role, their attitudes shifted back to “pro-union and anti-management” (p. 400). Nevertheless, attitudes clearly changed in response to the position switch, supporting the idea that those holding different positions may have different implicit theories when rating ideal behavior.

Based on such information, the following hypothesis is proposed:

Hypothesis 4: Participants who have only held subordinate positions will rate participative ILT behaviors higher than participants who have also held supervisory positions.

Overall, the present study seeks to investigate how employed individuals form Implicit Leadership Theories. Past literature has neglected to examine directly how experiences, or working with multiple leaders, have affected the formation process. Therefore, this study assesses the extent to which subordinates’ experiences with multiple leaders affect the formation of ILTs. As a competing hypothesis, subordinates’ personalities will also be assessed to examine the extent to which subordinates form ILTs congruent with their own personality. Lastly, based on past research, it is also hypothesized that subordinates might share a common element when assessing ILTs, such that they might rate participative leadership behaviors higher than any other leadership behaviors.

Methods

Participants and Procedure

All participants will be recruited using the online subject pool the StudyResponse Project (studyresponse.org), which facilitates online research for behavioral, social, and organizational science researchers by distributing email participation requests to adult research participants. Participants who work for a large variety of organizations, are invited to complete a particular survey when they meet the eligibility criteria of the study. Motivation to participate is prompted with a five-dollar incentive. Several published studies have used data collected from this source (e.g., Harris, Anseel, & Lievens, 2008; Piccolo & Colquitt, 2006) The StudyResponse project is hosted by the School of Information Studies at Syracuse University and has received institutional review board approval.

Before selection, all participants will be pre-screened by StudyResponse to determine that they are currently employed, are currently working for some kind of supervisor (i.e., a leader), and have worked for at least three supervisors (including their present supervisor) at some point during their employment history. Such pre-screening procedures are necessary to ensure that all participants are currently in a subordinate employment position and have had experiences with multiple supervisors or leaders. In addition to employment status, pre-screening will also ensure that all participants are U.S. residents, in order to avoid cross-cultural differences in leadership beliefs.

Upon pre-screening, all eligible participants will be notified of the study and offered an incentive if it is completed. One week after initial notification, participants who have not yet completed the study will be sent a reminder. If SurveyResponse participants choose to complete the study, they will be sent all materials via the internet. At this time, all participants will complete a personality inventory, a questionnaire about their work situation and leaders, and a demographic survey. The leadership questionnaire will be completed twice at Time 1, once regarding their most effective leader’s behaviors and once regarding their least effective leader’s behaviors. The same participants who completed all parts of the survey will be sent a second incentive to complete the second half of the study two weeks after completion of the first. At Time 2, all participants will complete a leadership questionnaire similar to the one from Time 1, except that they will be asked to rate their own idea of how they expect a business leader to behave (i.e., their Implicit Leadership Theory). Data will be gathered at two time periods, such that independent and dependent variables will be collected separately, due to findings that note temporal separation of predictor and criterion during the data gathering process may serve to remedy the effects of common method variance (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003).

Measures

Personality. At Time 1, three personality dimensions (agreeableness, conscientiousness, and extraversion) will be assessed for all participants. An adapted version of Goldberg’s (1992) measure of the Big Five Factor Structure will be used to assess the three dimensions. The three personality dimensions will be measured with ten items each. Sample items of the conscientiousness scale are “I like order” and “I am precise in my work.” Sample items of the agreeableness scale are “I sympathize with others’ feelings” and “I make people feel at ease.” Sample items of the extraversion scale are “I feel comfortable around people” and “I don’t mind being the center of attention.” The items will be answered on a 5-point Likert scale ranging from “Strong disagree” to “Strongly agree.” Goldberg (1992) reported a reliability of .79 for conscientiousness, .82 for agreeableness, and .87 for extraversion.

Additionally, in order to assess Hypothesis 3, three extra items will be included to asses how strongly participants consider themselves leaders. The items are as follows: “I consider myself a leader,” “I have the potential to be a leader,” and “I would be a good leader.”

Past Experience with Leadership. At Time 1, all participants will also rate four dimensions of leadership behaviors (initiating structure, consideration, inspiring commitment, and participative style) twice, once for a participant’s most effective leader and once for a participant’s least effective leader. For each leader, the dimensions will be rated using a 5-point Likert scale ranging from “Strong disagree” to “Strongly agree” based on the extent that each leadership behavior was exhibited.

The initiating structure and consideration items were adapted from the 1963 version of the Leader Behavior Description Questionnaire (LBDQ) of the Ohio State Leadership Studies. There are ten items for each dimension, with a reliability of .83 for initiating structure and a reliability of .92 for consideration (Stogdill, 1963). The four-item inspiring commitment scale was adapted from Yukl’s (1998) Managerial Practices Survey. The participative style items were adapted from the Path-Goal Leadership Questionnaire cited in Northouse (2004). There are five items representing the participative style dimension.

The specific leadership dimensions and personality traits were chosen in order to align the leadership behaviors with corresponding personality traits. Sample items of the initiating structure scale are “Assigns group members to particular tasks” and “Maintains definite standards of performance.” Sample items of the consideration scale are “Is friendly and approachable” and “Treats all group members as his/her equals.” A sample item of the inspiring commitment scale is “Describes a proposed change with great enthusiasm and conviction.” A sample item of the participative style scale is “Consults with members when facing a problem.” These items and their instructions can be seen in Appendices B and C.

Implicit Leadership Theories. Because there is to date no single, widely accepted measure of implicit leadership theories, the present study took the approach of using well-known and established measures of a variety of leader behaviors and adapting them to measure ILTs. At Time 2, adaptations of the same scales used to assess past experience with leadership will be administered again to assess the implicit leadership theories of all participants. Participants will rate the same four dimensions of leadership (initiating structure, consideration, inspiring commitment, and participative style), but this time will rate them on the extent that each behavior is essential for what they consider a business leader. All ratings will be made on a five-point Likert scale. These items and their instructions can be seen in Appendix D.

Demographics. StudyResponse asks all participants demographic questions when they register to participate in the subject pool. The following demographics will be recorded and sent to the researcher for each participant: gender, age, education level, ethnicity, job title, job position, years at present workplace, and years in the workforce. These items can be seen in Appendix E.

Analysis

Correlations and multiple regressions will be used to test the hypotheses. Hypothesis 1 states that subordinates’ experiences with various types of leaders will predict their implicit leadership theories. It is predicted that ILTs will be more strongly related to descriptions of one’s most effective leader than to descriptions of one’s least effective leader. Pearson correlation coefficients will be calculated in order to determine the strength of the relationship between ratings of one’s most effective leader and ratings of ILTs and between ratings of one’s least effective leader and ratings of ILTs. The ratings of one’s most effective leader should be more strongly correlated with ILTs than are the ratings of one’s least effective leader. Therefore, the significance of the difference between these correlations will be calculated. In addition, both ratings of one’s least and most effective leaders will be entered into a regression equation as predictors of ILTs, and the betas will be inspected.

Hypothesis 2 states that subordinates’ personalities will affect the formation of their implicit leadership theories and that subordinates will think leaders who exhibit behaviors that correspond to their own personality traits are more effective than leaders whose behaviors do not correspond to the subordinate’s personality traits.

More specifically:

Subordinate conscientiousness is positively associated with rating a business leader high on initiating structure.

Subordinate agreeableness is positively associated with rating a business leader high on consideration.

Subordinate extraversion is positively associated with rating a business leader high on inspiring commitment.

Subordinate agreeableness is positively associated with rating a business leader high on participative style.

A Pearson correlation coefficient will be calculated in order to determine the significance and strength of the relationship between each personality trait and ratings of each corresponding ILT behavior in the hypothesis. An underlying assumption of the hypothesis is that the personality trait named will more strongly predict one specific ILT behavior than will the other personality traits. Therefore, the significance of the difference of the correlations of personality characteristics with ILT behaviors will be calculated. In addition, each ILT behavior will be regressed on all three personality traits, and the betas will be inspected.

Hypothesis 3 states that subordinates who have a self-image as a leader will be more likely to engage in a self-serving bias when forming implicit leadership theories. A moderated regression analysis regarding each pair of subordinates’ personality traits and leaders’ behaviors with self-image as the moderator will be calculated to determine if leadership self-image significantly moderates the relationship between subordinate personality and ILT ratings. The expected direction of the moderated relationship is presented in Figure 2.

Operationally, engaging in a self-serving bias is tested in Hypotheses 2a, 2b, 2c, and 2d. As an example analysis, in order to test Hypothesis 3 in relation to the variables in Hypothesis 2a, subordinate conscientiousness and leadership self-image will be entered at the first step of a multiple regression with the ILT dimension of initiating structure as the criterion variable. In the second step, the interaction between conscientiousness and leadership self-image will be entered as a predictor. The R-square change between the first and second step and the beta for the interaction term will be examined to test Hypothesis 3.

Figure 2. Expected moderation effects for leadership self-image on the relationship between engaging in a self-serving bias and ILT ratings.

Lastly, Hypothesis 4 states that participants who have only held subordinate positions will rate participative ILT behaviors higher than participants who have also held supervisory positions. In order to test Hypothesis 4, the ILT rating differences between the participants who have only held subordinate positions and the participants who have also held supervisory positions will be compared in a 1 x 4 within-subjects analysis of variance, with the expected outcome being higher mean scores on participative behaviors for participants who have only held subordinate positions.

In addition, consistent with the model in Figure 1, multiple regressions will be used in order to assess the overall prediction of implicit leadership theory (ILTs) behaviors from past experiences (i.e., best leaders in the past) and engaging in self-serving biases (i.e., personality). The multiple regression analysis will allow the simultaneous testing and modeling of multiple independent variables, control variables, and the moderator variable.