Journal of Applied Psychology 2006, Vol. 91, No. 2, 298 –310

Copyright 2006 by the American Psychological Association 0021-9010/06/$12.00 DOI: 10.1037/0021-9010.91.2.298

A Longitudinal Study of the Moderating Role of Extraversion: Leader–Member Exchange, Performance, and Turnover During New Executive Development Talya N. Bauer and Berrin Erdogan

Robert C. Liden and Sandy J. Wayne

Portland State University

University of Illinois at Chicago

Identifying factors that help or hinder new executives in “getting up to speed” quickly and remaining with an organization is vital to maximizing the effectiveness of executive development. The current study extends past research by examining extraversion as a moderator of relationships between leader–member exchange (LMX) and performance, turnover intentions, and actual turnover for an executive sample. The sample consisted of 116 new executives who were surveyed prior to starting their employment and at 3 months postentry. A total of 67 senior executives rated these new executives in terms of overall performance at 6 months postentry. Turnover data were gathered from company records 31⁄2 years later. Hierarchical regression results showed that LMX was not related to performance or turnover intentions for those high in extraversion; but for individuals low in extraversion, there was a relation between LMX, performance, and turnover intentions. Furthermore, survival analyses showed that LMX was only related to turnover-hazard rate for individuals low in extraversion. Keywords: leader–member exchange, extraversion, executives, job performance, turnover

ence decisions regarding a variety of important factors. But coupled with these seemingly positive aspects of executive jobs is the pressure put on executives to perform well and to do so quickly (McCall et al., 1988). Identifying factors such as personality characteristics and insider support that may help or hinder new executives in “getting up to speed” is vital to maximizing the effectiveness of new executives and retaining them (McCall et al., 1988). The literature on new employee socialization may be useful in understanding how executives adjust to their new roles. Research on socialization shows that establishing effective working relationships with leaders early on is a key way that new employees can successfully integrate into their organizations (e.g., Bauer & Green, 1998; Kammeyer-Mueller & Wanberg, 2003; Major, Kozlowski, Chao, & Gardner, 1995). Similarly, the executive development literature has established the importance of developing relationships (Mainiero, 1994), finding mentors (Lyness & Thompson, 2000; McCall et al., 1988), and receiving top management support (McCauley, Ruderman, Ohlott, & Morrow, 1994) for executive success. Therefore, examining relationship development with new executives and their superiors may increase understanding of new executive success in terms of enhancing performance and avoiding organizational withdrawal. One theory that examines the relationship quality between superiors and subordinates is leader–member exchange (LMX). A distinctive feature of this approach to leadership is its focus on the dyadic level of leadership. Rather than thinking about a leader as having one set style, LMX theory proposes that leaders form unique relationships with each of their subordinates so that highLMX employees receive greater growth opportunities (Graen & Scandura, 1987) and higher levels of support (Kraimer, Wayne, & Jaworski, 2001). Theorists posit that social networks are important to organizational members (Brass, Butterfield, & Skaggs, 1998).

Executives are among the highest paid workers in the United States, and they occupy high-level jobs within organizations, with over 3 million of them working at the turn of the century (Bureau of Labor Statistics, U.S. Department of Labor, 2000). Organizations spend considerable time, money, and effort to transform executives into productive assets. Research has shown that the cost of losing executive talent is high. For example, Overton (2001) estimated that the direct and indirect costs of replacing an executive amount to more than that executive’s annual salary, and they have been estimated to be $500,000 per lost executive (McCall, Lombardo, & Morrison, 1988). In addition to replacement costs, losing top talent can also result in the loss of competitive industry secrets. This is especially true for Fortune 500 organizations, which tend to hire employees from competing Fortune 500 firms (Williamson & Cable, 2003). Executives hold powerful and prestigious positions and are key organizational members who have the ability to shape and influ-

Talya N. Bauer and Berrin Erdogan, Department of Management, School of Business, Portland State University; Robert C. Liden and Sandy J. Wayne, Department of Managerial Studies, University of Illinois at Chicago. An earlier version of this article was presented at the annual meeting of the Academy of Management, New Orleans, Louisiana, August 2004. This research was funded by a research grant from the Society of Human Resource Management Foundation and by a grant from the University of Illinois’s Center for Human Resource Management. We thank Murray Barrick, Carl Maertz, and Donald Truxillo for their helpful comments on earlier versions of this article. Correspondence concerning this article should be addressed to Talya N. Bauer, Department of Management, School of Business, Portland State University, P.O. Box 751, Portland, OR 97207-0751. E-mail: [email protected] 298

LMX-TO-PERFORMANCE AND -TURNOVER RELATIONSHIPS

More specific to LMX, Sparrowe and Liden (1997) contended and empirically verified (Sparrowe & Liden, 2005) that high-LMX employees are more readily integrated into the leader’s personal network. These researchers discovered that being incorporated into the leader’s set of trusted contacts was related to focal members being perceived by others as influential players in the organization. Paralleling these results, Seibert, Kraimer, and Liden (2001) found that establishing a social network (including contacts in different functional areas) was related to the procurement of resources and information and, ultimately, to career success. Given that executives tend to be charged with making high-profile decisions that require access to a wide variety of information as well as being charged with the responsibility of executing tasks that demand access to resources, establishing oneself as an influential member of a far-reaching social network may be especially critical for an individual holding an executive position. Thus, we expect the relationship that new executives form with their superiors to be a salient factor in the executive development process. To date, the executive development literature has not incorporated the findings of LMX theory. Similarly, the LMX literature has typically not examined employees who hold high-level positions. Therefore, our main objective in this study was to integrate these literatures by exploring the role of LMX in the new executive development process as measured by executive performance and withdrawal. Furthermore, we contend that the importance of highquality LMXs will be contingent on the personality of new executives. We propose that introverted executives—individuals who are less inclined to seek social attention (Ashton, Lee, & Paunonen, 2002)—will benefit more from forming high-quality relationships than will their more extraverted counterparts. Our study makes three contributions to the literature. First, we contribute to the executive development literature by examining how performance and withdrawal are influenced by the interaction between executive personality and the unique relationships that form between new executives and their immediate superiors. This is especially important because the executive development literature has primarily focused on gender as a key individual difference (e.g., Lyness & Thompson, 2000; McCauley et al., 1994), but it has not examined how executives with different personalities adjust to their new roles. Second, we make a contribution to the LMX literature by introducing a potential moderator of LMX-to-outcome relations. Several LMX researchers have expressed concern that LMX studies have adopted a universalistic perspective that has resulted in a dearth of knowledge on the circumstances that make LMX more or less important for executive success (e.g., Erdogan & Liden, 2002; Schriesheim, Castro, & Yammarino, 2000; Vecchio, Griffeth, & Hom, 1986). Finally, our third contribution is reexamining the LMX-toactual-turnover relationship while using methodological advancements in the turnover literature. Although a number of studies have revealed nonsignificant relationships between LMX and turnover, researchers have called for additional studies that examine moderators (Gerstner & Day, 1997), are conducted with high-level employees (Vecchio, 1985), and use longitudinal designs (Singer & Willett, 1991) to further test the relationship between LMX and actual turnover. We address these calls for further examination by using survival analysis, examining a new moderator (extraversion), and studying a high-level executive sample across time.

299

Hypotheses Executive Performance One of the main goals for new executives is to adjust to their new jobs so as to perform well. LMX is associated with better role adjustment in terms of performance—research has consistently found that LMX and performance are related across a wide range of jobs (Gerstner & Day, 1997). For example, in a cross-sectional study, Masterson, Lewis, Goldman, and Taylor (2000) found that in their sample of university employees with an average tenure of 7 years, performance was related to LMX. Similarly, Wayne, Shore, and Liden (1997) studied employees with at least 5 years of work experience and found a positive relationship between LMX and performance. Settoon, Bennett, and Liden (1996) discovered that LMX and performance were related in their sample of hospital employees with approximately 6 years of tenure. For their sample of expatriates, Kraimer et al. (2001) found a similar relationship. In one of the few studies of new employees that included performance as an outcome, Bauer and Green (1996) studied new college graduates and found that LMX and performance were related at 2 months on the job as well as at 9 months on the job. LMX researchers have posited several mechanisms relating LMX to performance (Liden, Sparrowe, & Wayne, 1997). For example, new employees need information to do their jobs and to gain clarity for role expectations (Bauer, Morrison, & Callister, 1998). Indeed, the literature shows a strong relationship between LMX and role clarity, perhaps because individuals in high-LMX relationships interact more frequently with their leaders (Kramer, 1995). Further, high-LMX members often enjoy more challenging assignments, sponsorship, and greater access to information relevant to the job (Graen & Scandura, 1987). Although the existing literature shows a relationship between LMX and job performance, to date no study has examined this relationship in an executive sample, in which “who you know” can be as important as “what you know” (Burke, 1984). Having a strong LMX relationship should help with these dimensions of executive performance, which are often ambiguous and high profile. Research has shown that developing relationships is key to advancing to executive positions (Mainiero, 1994). Lyness and Thompson (2000) found that mentoring was related to executive success even after they controlled for tenure, gender, age, education, and human capital. Further, interviews with executives have shown that their superiors are key in helping them to understand management values and navigate the politics of the organization (McCall et al., 1988). McCauley et al. (1994) found that major areas of learning for executives experiencing job transitions included having higher responsibility and visibility and dealing with nonauthority relationships to get their jobs done. In line with findings from the LMX and executive development literatures, forming high-LMX relationships should help new executives deal with job transitions, garner greater responsibility, learn insider information, and overcome obstacles. All of these factors should enhance new executive job performance. Therefore, we first examined whether the previously established relationship held true with this sample of new, high-level executives. Hypothesis 1: LMX is positively related to new executives’ job performance as rated by their supervisors.

BAUER, ERDOGAN, LIDEN, AND WAYNE

300 Executive Turnover Intentions

Only a handful of studies have examined LMX and turnover intentions. Despite the small number of studies of this relationship, meta-analytic results have found an overall negative relationship between LMX and turnover intentions (Gerstner & Day, 1997). The socialization literature has also examined interpersonal interaction and turnover intentions. For example, relationship building was found to influence turnover intentions across three waves of data collection (Kammeyer-Mueller & Wanberg, 2003). Kammeyer-Mueller and Wanberg’s (2003) study provided preliminary evidence for the salience of leaders as key socializing agents, but no previous studies examined new executive employees. Executives who are able to form high-LMX relationships should become more attached to their workplace because they feel more supported (Kacmar, Carlson, & Brymer, 1999; Liden, Wayne, & Sparrowe, 2000; Schriesheim et al., 2000), receive more feedback from their superiors (Kramer, 1995), and belong to a network at work (Sparrowe & Liden, 2005). On-the-job embeddedness refers to employees being situated or connected in a social web at work that creates a disinclination to consider leaving (Lee, Mitchell, Sablynski, Burton, & Holtom, 2004). Job embeddedness increases the cost of leaving the organization for the person in question. If executives do not want to lose valued relationships with their superiors, they are more psychologically attached to the organization and should have lower turnover intentions (Maertz & Campion, 1988). Hypothesis 2: LMX is negatively related to new executive turnover intentions.

Executive Turnover Although it is clear that LMX is negatively related to turnover intentions, research on LMX and actual turnover has tended to show mixed results, with results from a meta-analysis indicating that no relationship exists (Gerstner & Day, 1997). However, we propose that the same mechanisms that link turnover intentions with LMX will operate for actual turnover as well. We revisited the relation between LMX and actual turnover for several reasons. First, relatively few studies of LMX have examined turnover, and none of these (to our knowledge) have been conducted with a sample of executives. Indeed, we expect that factors that relate to executive turnover will be related, at least partially, to the ability of new executives to learn from their leaders (e.g., McCall et al., 1988; McCauley et al., 1994). Although Ferris (1985) replicated the findings of Graen, Liden, and Hoel (1982) for LMX relating to actual turnover, Vecchio (1985) was unable to replicate these results in his sample of bank tellers. He noted, however, that “the leader–member exchange approach has potentially greater predictive utility for turnover among high-level employees than among low-level employees” (p. 483). Consistent with this speculation, the sample in the Graen et al. (1982) research, which found a strong negative correlation between LMX and turnover, consisted of highly educated computer and systems analysts. Second, there are characteristics associated with executives that provide support for the relation between LMX and turnover. In a study of the predictors of executive career success, Judge, Cable,

Boudreau, and Bretz (1995) found that higher level executives were more “promotion motivated” than lower level executives, despite the more limited prospective promotion opportunities given their high level in the organization. Furthermore, Judge et al. (1995) found that job tenure negatively predicted number of promotions. These findings suggest that high-level executives tend to be very ambitious and desirous of promotion. As a result, such executives may be less likely to tolerate a low-LMX relationship with their superior, because this person typically has a significant impact on promotion decisions. Furthermore, it seems that executives may have many alternative job opportunities given the “war” for executive talent (Michaels, Handfield-Jones, & Axelrod, 2001). Thus, the LMX relationship between new executives and their immediate superiors may be especially salient with respect to turnover. Additionally, Gerstner and Day (1997) concluded that the relationship between LMX and actual turnover is worthy of further examination, and Vecchio et al. (1986) indicated that the search for this relationship should not be abandoned. Finally, research since Gerstner and Day’s meta-analysis has found some support for the importance of leadership. For example, Kammeyer-Mueller and Wanberg (2003) compared different socializing influences and found that across seven different organizations, leader influence was the only socializing agent that was related to actual turnover. If new executives have low-LMX relationships, they may not fit into the organization, and they are likely to have an unfavorable work situation that will motivate job-search behaviors and cause them to leave more rapidly. High LMX may create a sense of patience not to “rush out the door” and an obligation to one’s manager that may lengthen tenure. Thus, in both cases, LMX would be related to duration of stay as well as turnover. In terms of the methods used when studying actual turnover, Singer and Willett (1991) noted that turnover research should be longitudinal, allow sufficient time for turnover to take place, use survival analysis to analyze predictors of turnover, and capture turnover in terms of days stayed rather than simply as a dichotomous stayed–left variable. Survival analysis uses information regarding when turnover occurred in addition to whether it occurred. Therefore, it provides a richer understanding of the process by which employees leave an organization. For example, over a 3-year period, all study subjects could have left an organization. However, from the organization’s perspective (as well as theoretically), there is a difference between an executive who leaves after staying for 3 months and someone who leaves after staying for 21⁄2 years. Even if both individuals would have turned over by the end of the study, it makes theoretical sense to use information regarding how long individuals stayed. In fact, Morita, Lee, and Mowday (1993) demonstrated that survival analysis and logistic regression may give contradictory findings and that the use of survival analysis provides a more accurate method of estimating relations between turnover and other variables. All of the published studies of LMX and turnover have used logistic regression procedures (e.g., Ferris, 1985; Graen et al., 1982; Vecchio, 1985; Vecchio et al., 1986). Logistic regression is appropriate for examining predictors of turnover. However, it discards information regarding how long individuals stayed before leaving. Using this information may reveal that even though LMX may not be related to whether individuals leave or stay, those with high LMX may stay for longer periods of time before they leave.

LMX-TO-PERFORMANCE AND -TURNOVER RELATIONSHIPS

One goal of this study was to build on past research and examine the LMX-to-actual-turnover relationship, following the recommendations of Singer and Willett (1991). Hypothesis 3: LMX is negatively related to new executive actual turnover.

The Moderating Effect of Extraversion To this point, we have posited main effects for LMX relating to performance and withdrawal. However, LMX researchers have consistently called for the examination of moderators—in particular, individual-difference moderators— of the LMX-to-performance and LMX-to-turnover relationships (e.g., Erdogan & Liden, 2002; Gerstner & Day, 1997; Vecchio et al., 1986). Our search of the literature for potential moderators identified a key personality variable: extraversion. Extraversion has surfaced as an individualdifference variable that is stable and exerts an influence on a wide range of employee behaviors. Extraversion is perhaps the dominant factor in the five-factor model (FFM) of personality (e.g., Costa & McCrae, 1992; Digman, 1990; Judge, Bono, Ilies, & Gerhardt, 2002), and of the five factors, it is the one most theoretically and empirically related to the constructs of interest in the current investigation. Most major personality inventories include some form of extraversion (Watson & Clark, 1997). Research has shown that extraversion is a valid predictor across criterion types for managers (including foremen through top-level executives). And in their meta-analysis on the FFM, Barrick and Mount (1991) concluded that being extraverted was helpful for people in managerial jobs but less important for people with other occupations, such as secretaries, accountants, production workers, engineers, and architects. They also found that extraversion was related to higher training success. Of particular importance, given that our sample consisted of executives learning new jobs, are Barrick and Mount’s findings indicating that extraversion is a key personality characteristic for managers. In fact, Judge et al. (2002) found meta-analytic evidence that extraversion was the most consistent correlate of leadership across study settings and leadership criteria. LMX is a theory of social interaction. Therefore, its effects should be contingent on the extent to which employees interact effectively with others. Personality theory and research have shown that people high on extraversion are characterized as seeking out interaction opportunities with others, generally liking other people, being gregarious and talkative (Costa & McCrae, 1992), and being high in reward sensitivity. Depue and Collins (1999) define reward sensitivity as the tendency to experience “an incentive motivational state that facilitates and guides approach behavior to a goal” (p. 495). Research has identified reward sensitivity— which included ascendance (social dominance, enjoyment of leadership roles, assertiveness, and goal accomplishment)—as a higher order factor underlying extraversion (e.g., Lucas, Diener, Grob, Suh, & Shao, 2000). Subsequently, Ashton et al. (2002) found that a central feature of extraversion is the drive to garner social attention—which they defined as the tendency for people to behave in ways that attract and hold social attention—and that extraverted individuals enjoy engaging in these behaviors. Introverts, conversely, tend to be reserved and to prefer their own company or that of habitual companions. In terms of job characteristics, extraverts seek out novel, complex, challenging, and

301

varied tasks as well as intense stimuli, whereas introverts tend to prefer predictable paths in their work (Costa & McCrae, 1992). Successful executives have reported developmental experiences associated with their success, such as networking, having developmental job assignments, and encountering novel situations (Lyness & Thompson, 2000). It is interesting to note that the benefits of high-LMX relationships and the personality characteristics associated with being extraverted are strikingly similar in terms of the extent to which they influence success as an executive. Because extraverts are characterized as seeking situations in which they garner social attention and interaction, they are likely to do well in executive positions, without the necessity of forming high-quality relations with their leader. Extraversion is a trait that fits well with the demands of an executive position, and therefore, new executives high on extraversion will seek information, engage in social networking (Bozionelos, 2003), and behave in ways that should lead to higher levels of adjustment in their new positions. For example, in their study of new employees, Wanberg and Kammeyer-Mueller (2000) found that extraversion was positively related to feedback seeking and interactions with organizational insiders. High-LMX relationships are characterized by frequent interaction (e.g., Bauer & Green, 1996; Kramer, 1995), and extraverts desire interaction. High-LMX relationships are characterized by challenging assignments and a wide range of job responsibilities (e.g., Liden et al., 2000), and extraverts enjoy job challenge and novel work. In essence, the preferences and behavioral tendencies that characterize extraverts parallel the qualities of both typical high-quality LMX relationships and executive success (Lyness & Thompson, 2000). Therefore, for extraverts, LMX and performance should be weakly related, because extraverts should be able to perform well at executive jobs regardless of the quality of their LMX relationships. Introverts are in a seemingly precarious situation in which their personalities do not match the needs of new employees to seek information to help them adjust to their new roles. Without a high-LMX relationship, introverts may find it difficult to access needed information and resources and to navigate complex social networks. Therefore, it is possible that for introverts, LMX relationships play a critical role in their adjustment. Without the benefits derived from establishing high-LMX relationships, introverts’ tendency to be reserved and reclusive may make it difficult for them to perform at the level of their extraverted counterparts. We propose that for an introvert, performance will be related to LMX such that the higher their LMX, the higher their performance. Those introverts who are able to successfully establish and maintain high-LMX relationships should be able to accrue benefits similar to those enjoyed by their extraverted coworkers. Hypothesis 4: Extraversion moderates the relationship between LMX and job performance such that there is a positive relationship for those low on extraversion (introverts), whereas the relationship is not significant for individuals high on extraversion (extraverts). Similarly, for introverts, we propose that LMX will be crucial for their integration into the organization such that the higher their LMX, the lower their turnover intentions and actual turnover. If introverts working as executives are unable to establish high-LMX

BAUER, ERDOGAN, LIDEN, AND WAYNE

302

relationships and subsequently are unable to become successful in terms of social integration and/or job performance, they should be less embedded in the organization and more likely to want to leave it (Lee et al., 2004). However, although a leader can be instrumental in supporting and motivating employees (Eisenberger, Stinglhamber, Vandenberghe, Sucharski, & Rhoades, 2002), subordinates are dependent on such a leader only to the extent that alternate forms of support, guidance, and feedback are lacking. Extraverts are more inclined to seek out these alternative sources of support, because they tend to be more independent and are more likely to engage in proactive socialization behaviors like relationship building and feedback seeking than are introverts (Crant, 1995). Related to their tendency to be proactive, extraverts seek out information, opportunities for social interaction, and new activities (Bozionelos, 2003; Judge, Martocchio, & Thoresen, 1997)—all behaviors that have been linked to socialization success (Jablin, 2000). Although introverts can garnish information and develop contacts with key individuals in the organization through a strong relationship with a leader (Sparrowe & Liden, 2005), we contend that extraverts can do so without relying on their leader. This argument is consistent with substitutes for leadership theory (Kerr & Jermier, 1978), which suggests that subordinate independence from others, a characteristic of extraverts, serves as a substitute for leadership. Thus, we propose that for introverts, LMX is crucial for their successful attainment of work-related goals such that the higher their LMX, the lower their turnover intentions and actual turnover. If introverts working as executives are unable to establish highLMX relationships, they will be less likely to become effectively socialized into the organization and, thus, will be more likely to leave the organization. Because extraverts are able to secure the resources and support needed to become successfully socialized into the organization without the help of the immediate leader, for extraverts, LMX, turnover intentions, and actual turnover should be unrelated. Hypothesis 5: Extraversion moderates the relationship between LMX and turnover intentions such that there is a negative relationship for those low on extraversion (introverts), whereas the relationship is not significant for individuals high on extraversion (extraverts). Hypothesis 6: Extraversion moderates the relationship between LMX and actual turnover such that there is a negative relationship for those low on extraversion (introverts), whereas the relationship is not significant for individuals high on extraversion (extraverts).

time frame for understanding subsequent performance and turnover (Bauer et al., 1998). The timing of the data collections included the following: prior to entry (Time 1) and 3 months postentry (Time 2) for the new executives, a survey completed by senior executives about the new executives at 6 months postentry (Time 3), and data obtained from organizational records 31⁄2 years later (Time 4). The organization sent the Time 1 preentry survey via postal mail to all new hires along with their official job offer. Transferees completed the same survey online after receiving the link from the research team. This survey assessed the extraversion level of new executives, and participants returned their completed surveys (a) directly to the researchers via U.S. mail in self-addressed, postage-paid envelopes or (b) online (for transferees). The Time 2 and Time 3 surveys were Web-based, and the research team forwarded the links for these surveys directly to new executives at 3 months following their start dates and to the supervising senior executives at 6 months. The Time 2 employee survey assessed the relationship quality between new executives and their senior executives and turnover intentions from the new executives’ perspective, and the Time 3 superior survey assessed performance of new executives from the senior executives’ perspective. Time 4 turnover data were gathered from company records 31⁄2 years later. A total of 168 new executives completed the preentry survey (response rate ⫽ 73%). One hundred and forty-six new executives completed the 3-month survey (response rate ⫽ 63%). One hundred and eleven senior executives participated in the 6-month senior executive survey (response rate ⫽ 48%). The number of employees who responded at both preentry and at 3 months was 116, which was the sample size for Hypotheses 2, 3, 5, and 6 (overall response rate ⫽ 50%). The number of dyads (new executives paired with supervising senior executives) for which we had complete data across the first three time periods was 67 (with 58 senior executives rating only 1 new executive, 3 senior executives rating 2 new executives, and 1 senior executive rating 3 new executives)—a response rate of 29%. This was the sample size available to test Hypotheses 1 and 4. Sixty-six percent of the new executives were new hires, whereas 34% were employees transferring within the company. However, as noted below, whether the executive was a new hire or transferee did not affect the results. According to company records, the average age of new executives was 41 years (SD ⫽ 7). Fifty-nine percent of the respondents were male, and 41% were female. Seventy-seven percent of the sample was identified as Caucasian, 11.5% as Asian/Pacific Islander, 5.3% as Hispanic, 1.8% as African American, and 0.9% as Puerto Rican. The respondents had high levels of education: 37% of the sample held a bachelor’s degree, 33% had a master’s degree, and 30% held a doctoral degree. The job titles held by the new executives included Director and Associate Director (n ⫽ 75), Senior Manager (n ⫽ 24), and Vice President (n ⫽ 11) . Senior executives who rated new executives were mostly male (75%) and Caucasian (75%), with 8.8% of the superior sample identified as Asian/Pacific Islander, 4.4% as Hispanic, and 1.5% as African American. The average age of senior executives was 45 years (SD ⫽ 6.54). In terms of education level, 22% of the managers held a bachelor’s degree, 30% had a master’s degree, and 48% held a doctoral degree. The job titles held by the senior executives included President (n ⫽ 5), Vice President (n ⫽ 32), and Director (n ⫽ 40).

Method Participants and Procedure Our sample consisted of new employees and transferees holding executive positions in a Fortune 500 pharmaceutical organization. New employees (N ⫽ 231; new hires and transferees to executive-level positions) and their respective superiors were invited to participate in this study. Data were collected across four points in time via three surveys and organizational records. Research on new employee socialization has shown that it takes employees several months before they begin to feel adjusted to both the social and the task aspects of their job, suggesting that entry is a key

Measures Control variables. We identified several potentially relevant control variables. For example, gender and age have been shown to relate to executive relationship formation (Lyness & Thompson, 2000). Research has shown that transferees and new hires may differ (e.g., Kramer, Callister, & Turban, 1995). Therefore, age, gender, and new executive status (i.e., new hire or transferee) were all gathered from company records. New hires were coded as 1, and transferees were coded as 0. Gender was coded 1 for men and 0 for women. In addition, research has shown that the amount of interaction between subordinates and superiors is sometimes

LMX-TO-PERFORMANCE AND -TURNOVER RELATIONSHIPS related to LMX (Schiemann, 1977). Therefore, we asked new executives to report on the number of hours per week (on average) that they interacted with their superior at Time 2. Extraversion. At preentry, we measured extraversion using 8 adjectives developed by Saucier (1994) as part of his measure of the FFM of personality. New executives reported the extent to which these adjectives were accurate in describing their personality using a 9-point scale ranging from 1 (extremely inaccurate) to 9 (extremely accurate). Example adjectives include bold, energetic, and talkative (␣ ⫽ .82). LMX. Three months after the new executives started their new positions (Time 2), we measured LMX from their perspective, using the 12-item LMX–Multidimensional (LMX-MDM) scale developed by Liden and Maslyn (1998). Although the LMX-MDM can be used as a multidimensional scale, Liden and Maslyn found support for a higher order factor, allowing for a composite of all items to be used as a measure of global LMX. New executives reported their level of agreement using a 7-point scale ranging from 1 (strongly disagree) to 7 (strongly agree). An example item is “My supervisor would come to my defense if I were ‘attacked’ by others.” We averaged all items to create the LMX score (␣ ⫽ .90). Job performance. At 6 months postentry (Time 3), we measured job performance from the superior’s perspective using four items developed by Welbourne, Johnson, and Erez (1998). Managers used a 5-point scale ranging from 1 (needs much improvement) to 5 (excellent) to rate new executive job performance (␣ ⫽ .91). Sample items included ratings of new executive “quality of work output” and “quantity of work output.” Turnover intentions. Three months (Time 2) after the new executives started their new positions, they reported turnover intentions using a 5-item scale (Wayne et al., 1997). Employees reported their level of agreement using a 7-point scale ranging from 1 (strongly disagree) 7 (strongly agree). An example item is “I am seriously thinking of quitting my job” (␣ ⫽ .87). Turnover. Three and a half years following organizational entry (Time 4), information regarding new executive turnover was gathered from company records. Of the 168 executives who participated in this study at Time 1, 68 had left the organization by Time 4 (40%). In addition to turnover, we gathered the dates of separation to examine the length of stay in the organization. Methods relating to the measurement and analysis of turnover have garnered considerable attention (e.g., Dickter, Roznowski, & Harrison, 1996; Huselid & Day, 1991; Morita et al., 1993; Singer & Willett, 1991). The general consensus is that ordinary least squares regression is inappropriate when dealing with binary dependent variables (e.g., Huselid & Day, 1991; Morita et al., 1993). As Morita et al. (1993) noted, examining when an event such as turnover happens as well as whether it happens is more informative than what logistic and probit regression can test. Therefore, researchers have recommended using survival analysis when examining actual turnover. Survival analysis examines both when and if turnover occurs. Singer and Willett (1991) offered guidelines for survival analyses. Of these, we followed several suggestions, such as gathering longitudinal data, having a specific “start” time (such as entry), gathering data over a meaningful period of time so that the event has a chance to occur, comparing respondents and nonrespondents to ensure that no major differences exist, reporting the survivor function, examining interactions, and examining the shape of the hazard profile. We calculated days stayed as the number of days between the day each new executive completed a Time 2 survey and the day they quit or Time 4. We did not operationalize the number of days stayed starting from their date of entry because this would have involved using LMX (a Time 2 variable) to predict survival before Time 2. To address a potential concern about using a Time 2 variable (LMX) to predict turnover throughout a participant’s tenure (between Time 1 and Time 4), we operationalized “days stayed” as the difference between Time 2 and Time 4. Because the difference between Time 1 and Time 2 was equal to 90 days for all participants, we feel that the late-entry bias was less of a problem. Unlike past research (e.g., Lyness & Judiesch, 2001) that made corrections using

303

SAS, in which employees entered the dataset before Time 0 at varying time periods, in the present case, Time 2 was equidistant from the start date for all participants. Those who eventually left the organization had been with the organization for an average of 643 days (SD ⫽ 357). The average number of days remaining for the entire sample (both leavers and those who remained) was 1,017 (SD ⫽ 403). The earliest time anyone in our sample left was 20 days after completing the Time 2 survey. Replication sample. In addition to data from this executive sample, we wanted to gather additional data to address the concern that our interaction results might not generalize to other populations. Therefore, we gathered data from 102 working adults at a university located in the northwestern United States using a snowball sampling technique. Jobs of the participants were diverse and included financial analyst, test engineer, and salesperson. Participants were an average of 28 years old (SD ⫽ 6.83), had an average of 10 years of work experience (SD ⫽ 6.04), and 60% of the respondents were male. All of the measures collected were parallel to the executive sample to examine Hypothesis 5 (LMX; extraversion; turnover intentions; and the control variables of gender, age, and part-time vs. full-time work status). Because manager ratings and actual turnover were not available for this sample, we did not test hypotheses involving performance or actual turnover (Hypotheses 4 and 6).

Results Means, standard deviations, and intercorrelations are presented in Table 1. Before testing our hypotheses, we examined the relationships between respondents and nonrespondents for the main sample as well as between the variables of interest and several demographic control variables to make sure that significant differences did not affect our results. On the basis of t-test results, new executives who responded at both Time 1 and Time 2 did not differ from Time 2 nonrespondents in terms of age, gender, new executive status, hours of interaction, or extraversion. Further, new executives who had superiors complete a Time 3 performance survey did not differ from those who did not in terms of age, gender, new executive status, hours of interaction, extraversion, LMX, or turnover intentions. And finally, there were no differences among those who participated in our study and those who did not in terms of number of days remaining on the job or actual turnover rate. To examine the potential influence of control variables, we examined the correlation matrix. As shown in Table 1, age and gender were correlated with extraversion such that older executives were less extraverted than younger ones and that women were more extraverted than men. In addition, gender was positively related to LMX quality, with men reporting stronger relationships, and it was negatively correlated with turnover intentions, with men reporting lower intentions than women. Finally, number of hours of interaction was positively correlated with LMX. Therefore, we controlled for age, gender, and hours of interaction in all analyses, but we did not control for employee status, because it was unrelated to our variables of interest. To test Hypotheses 1, 2, 4, and 5, we performed two separate moderated hierarchical regression analyses. In the first step, we regressed job performance and turnover intentions on employee age, gender, and hours of interaction as well as extraversion after it was centered. In the second step, we entered LMX after it was centered. The coefficient estimate at this step provided tests of Hypotheses 1 and 2. Then, in the third step, we entered the product

BAUER, ERDOGAN, LIDEN, AND WAYNE

304

Table 1 Means, Standard Deviations, and Intercorrelations Among Study Variables Variable 1. 2. 3. 4. 5. 6. 7. 8. 9.

Age (years; T1) Gender (T1) Employee status (T1) Hours of interaction (T2) Extraversion (T1) LMX (T2) Turnover intentions (T2) Job performance (T3) Days stayed (T4)

M

SD

1

2

3

4

5

6

7

8

9

40.97 0.59 0.66 4.56 6.67 5.84 1.71 4.37 1,017.18

7.03 0.49 0.47 4.93 1.13 0.77 0.87 0.74 403.18

— .06 ⫺.16 ⫺.15 ⫺.18* ⫺.12 ⫺.07 ⫺.09 .03

— ⫺.03 ⫺.03 ⫺.19* .18* ⫺.31** ⫺.07 .06

— ⫺.02 .04 .15 ⫺.11 ⫺.02 ⫺.08

— ⫺.01 .18* .09 .11 ⫺.05

(.82) .16 ⫺.04 .25* .04

(.90) ⫺.37** .33** .13

(.87) ⫺.02 ⫺.20*

(.91) .25*

—

Note. Ns ⫽ 67–120. Age, gender, employee status, and days stayed in the organization were all gathered from company records. Gender was coded 1 for male and 0 for female. Employee status was coded 1 for new hire and 0 for transferee. Coefficient alpha reliability estimates are presented in parentheses along the diagonal. T1 ⫽ Time 1 (preentry; rated by new executive); T2 ⫽ Time 2 (3 months postentry; rated by new executive); LMX ⫽ leader–member exchange; T3 ⫽ Time 3 (6 months postentry; rated by superior of new executive); T4 ⫽ Time 4 (31⁄2 years postentry). * p ⬍ .05. ** p ⬍ .01.

of the centered LMX and extraversion to the equation (Aiken & West, 1991), providing tests of Hypotheses 4 and 5. We tested Hypotheses 3 and 6 via survival analysis. Because there were several individuals who remained with the organization at the conclusion of the study, the turnover data were right censored (meaning that a majority of the sample did not leave the organization during our data collection). Therefore, following Singer and Willett (1991) we used survival analysis via Cox’s proportional hazard model. The dependent variable in these analyses was the number of days before turnover occurred. The results of the hierarchical regression analyses are presented in Tables 2 and 3. Our analyses suggested that LMX was significantly related to both job performance (␤ ⫽ .31, p ⬍ .05) and turnover intentions (␤ ⫽ ⫺.35, p ⬍ .01), providing support for Hypotheses 1 and 2. LMX explained 10% of the variance in job

performance and 11% of the variance in turnover intentions after age, gender, hours of interaction, and extraversion were controlled for. To test Hypothesis 3, we first plotted the sample survivor profile, as recommended by Singer and Willett (1991). Specifically, we examined the probability of remaining with the organization (the survival function) for individuals high and low in LMX. Figure 1 depicts the survival functions for low LMX (below mean) and high LMX (above mean) levels. These curves suggest that the survival rate was higher for individuals high in LMX. Survival curves provide preliminary information about the differences between groups, but because our interest was in examining the more complex relationship of the interaction between LMX and extraversion, this plot does not provide an actual test of our hypotheses. Therefore, we used the results from the Cox propor-

Table 2 Hierarchical Regression Analysis With Job Performance as the Dependent Variable

Table 3 Hierarchical Regression Analysis With Turnover Intentions as the Dependent Variable

Job performance

Turnover intentions

Step 2 ␤ (Hypothesis 1)

Step 3 ␤ (Hypothesis 4)

Variable

Step 1 ␤

⫺.03 ⫺.01 .06 .23

.00 ⫺.03 .03 .21 .31*

.07 1.17 4, 62 .07 .01 1.17 4, 62

.10 7.00* 1, 61 .17 .10 2.42* 5, 61

.00 ⫺.03 .06 .20 .25* ⫺.32** .10 8.26** 1, 60 .27 .19 3.64** 6, 60

Age Gender Hours of interaction Extraversion LMX LMX ⫻ Extraversion ⌬R2 ⌬F df R2 Adjusted R2 Overall F df

Variable

Step 1 ␤

Age Gender Hours of interaction Extraversion LMX LMX ⫻ Extraversion ⌬R2 ⌬F df R2 Adjusted R2 Overall F df

Note. n ⫽ 67. Standardized regression coefficients are reported. Extraversion and LMX (leader–member exchange) are centered variables. Confidence intervals for the hypothesized and supported relationships did not include 0. Gender was coded 1 for male and 0 for female. * p ⬍ .05. ** p ⬍ .01.

Step 2 ␤ (Hypothesis 2)

Step 3 ␤ (Hypothesis 5)

⫺.06 ⫺.32** .07 ⫺.12

⫺.09 ⫺.25* .13 ⫺.05 ⫺.35**

.12 3.79** 4, 111 .12 .08 3.79** 4, 111

.11 16.13** 1, 110 .23 .19 6.67** 5, 110

⫺.10 ⫺.22** .12 ⫺.04 ⫺.31** .23** .05 7.72** 1, 109 .28 .24 7.18** 6, 109

Note. n ⫽ 116. Standardized regression coefficients are reported. Extraversion and LMX (leader–member exchange) are centered variables. Confidence intervals for the hypothesized and supported relationships did not include 0. Gender was coded 1 for male and 0 for female. * p ⬍ .05. ** p ⬍ .01.

LMX-TO-PERFORMANCE AND -TURNOVER RELATIONSHIPS

305

Figure 1. Sample survivals function at high and low levels of leader–member exchange (LMX). The upper line represents survival probability for executives with high LMX; the lower line represents survival probability for executives with a low LMX.

tional hazard model, which is able to test more complex models (Singer & Willett, 1991). The results from the second step of the Cox regression suggested that the coefficient for LMX was significant (␤ ⫽ ⫺.44, Wald ⫽ 4.44, p ⬍ .05) and that the coefficient was in the expected direction. In addition, the change in chi-square when LMX was added to the equation was significant, ⌬␹2(1) ⫽ 4.14, p ⬍ .05. LMX was negatively related to the hazard ratio such that a 1 standard deviation increase in LMX was associated with a 36% decrease in the odds of leaving. These results provide support for Hypothesis 3. The hierarchical regression results presented in Tables 2 and 3 provided preliminary support for Hypotheses 4 and 5 in that the interaction terms of LMX and extraversion were significantly related to performance and turnover intentions. To examine the pattern of relationships, we plotted the significant interactions following the procedure described by Cohen and Cohen (1983). Figure 2 depicts the relation between LMX and performance at high and low values of extraversion (1 standard deviation above and below the mean). Post hoc analyses (Aiken & West, 1991) demonstrated that, as expected, there was a positive relationship between LMX and job performance only for individuals low in extraversion (introverts; ␤ ⫽ .58, p ⬍ .01), but the relationship was not significant for individuals high in extraversion (␤ ⫽ ⫺.01, p ⬎ .05). The LMX ⫻ Extraversion interaction explained 10% of the variance in job performance, and the entire model accounted for 19% of the variance. These results provide support for Hypothesis 4.

Figure 3 depicts the nature of the relationship between LMX and turnover intentions at high and low levels of extraversion. Supporting Hypothesis 5, there was a negative relationship between LMX and turnover intentions for individuals low in extraversion (introverts; ␤ ⫽ ⫺.53, p ⬍ .01), but no relationship for individuals high in extraversion (␤ ⫽ ⫺.03, p ⬎ .05). The LMX ⫻ Extraversion interaction accounted for 5% of the variance in turnover intentions, and the entire model accounted for 24% of the variance. Proportional hazard model results provided support for Hypothesis 6. As presented in Table 4, the LMX ⫻ Extraversion interaction was significantly related to the hazard rate (B ⫽ 0.47, Wald ⫽ 7.24, p ⬍ .01). Following Aiken and West’s (1991) technique, we found that for individuals high in extraversion, there was no relationship between LMX and the hazard rate (B ⫽ 0.06, Wald ⫽ 0.04, p ⬎ .05). However, for individuals low in extraversion, LMX was negatively related to the hazard rate (B ⫽ ⫺1.02, Wald ⫽ 11.92, p ⬍ .01), meaning that LMX was negatively related to the probability of turnover. We also calculated the D statistic, which is similar to the R2 used in regressions, following the formula reported in Sheridan (1992). The LMX ⫻ Extraversion interaction explained 7% of the variance in hazard rates, and the overall model explained 11% of the variance. These results provide support for Hypothesis 6. We also tested the proportional hazards assumption inherent in Cox regression (Singer & Willett, 1991). We plotted hazard functions for different levels of all predictors and found the hazard plots to be parallel in all cases. Furthermore, we entered the

BAUER, ERDOGAN, LIDEN, AND WAYNE

306

Figure 2. The relationship between leader–member exchange (LMX) and job performance at high and low levels of extraversion.

interaction term of our variables of interest (LMX, Extraversion) with time into the equation and found them to be nonsignificant. These analyses indicated that the proportionality assumption was not violated.

Replication of Results We tested Hypothesis 5 using the replication sample of working adults. We ran a hierarchical moderated regression in which we regressed turnover intentions on the control variables, LMX, and extraversion in the first step and on the interaction term in the second step. The Extraversion ⫻ LMX interaction was significantly related to turnover intentions (␤ ⫽ .19, p ⬍ .05). The Extraversion ⫻ LMX interaction explained 3% of the variance in turnover intentions. The shape of the interaction was the same as the one presented in Figure 3. The relationship between LMX and turnover intentions was stronger for introverts than for extraverts. These findings are consistent with the findings from the executive sample. Thus, we established that the results for Hypothesis 5 were replicated in our general adult working population.

Discussion At the outset of this article, we argued that organizations that are able to maximize the effectiveness of their new executives have a

competitive advantage over those that do not. Yet few studies have examined the socialization of senior-level executives. We proposed that extraversion moderates the relation between LMX quality, executive performance, and withdrawal. Strong support was found for all of the hypothesized moderating relationships, as evidenced by ⌬R2s for the interaction terms ranging between 5% and 10% above the main effects and control variables. We found support for the expectation that extraversion would moderate the relationship between LMX, turnover intentions, and turnover. Our findings also support the idea that performance is related to LMX but that this relationship must take into account the personality of the employee being rated such that new executives low on extraversion need the support of organizational insiders, such as their immediate superiors, to help them succeed and remain with the organization. At the multivariate level, the performance of extraverts appeared unrelated to LMX, with extraverts performing at the same level regardless of their LMX relationships. Conversely, for introverts, having a high-LMX relationship was critical to high performance. Those individuals who were introverted and who were unable to establish a high-LMX relationship had lower ratings of performance. Our results indicated that high LMX and high extraversion provided similar advantages to new executives, such that only for

Figure 3. The relationship between leader–member exchange (LMX) and turnover intentions at high and low levels of extraversion.

LMX-TO-PERFORMANCE AND -TURNOVER RELATIONSHIPS

307

Table 4 Results of the Proportional Hazards Model Using Cox Regression Step 1

Step 2

Independent variable

B

SE

Wald

OR

Age Gender Hours of interaction Extraversion LMX LMX ⫻ Extraversion ⌬␹2 ␹2 D

⫺0.01 ⫺0.19 0.01 ⫺0.04

0.02 0.31 0.03 0.14

0.20 0.36 0.22 0.09

0.99 0.83 1.01 0.96

B ⫺0.01 ⫺0.05 0.03 ⫺0.01 ⫺0.44

0.86 0.87 .00

Step 3

SE

Wald

OR

B

0.02 0.32 0.03 0.14 0.21

0.26 0.02 0.73 0.00 4.44*

0.99 0.95 1.03 0.99 0.64

⫺0.02 0.07 0.03 0.12 ⫺0.49 0.47

4.14* 5.30 .04

SE 0.02 0.33 0.03 0.15 0.22 0.17 6.86** 13.32* .11

Wald

OR

0.54 0.04 1.20 0.58 5.23* 7.24**

0.98 1.07 1.03 1.12 0.61 1.61

Note. n ⫽ 116 for Hypothesis 6. Extraversion and leader–member exchange (LMX) are centered variables. Confidence intervals for the hypothesized relationships did not include 1. The degree of freedom corresponding to each chi-square value in this analysis is equal to the number of variables in the step. Gender was coded 1 for male and 0 for female. OR ⫽ odds ratio. * p ⬍ .05. ** p ⬍ .01.

relatively introverted executives was LMX significantly related to performance, turnover intentions, and turnover. The jobs of executive-level employees require social interaction, networking, and dealing with novel situations. Extraverts may manage these situations via their more attention-seeking personalities, whereas introverts seem to need the assistance of high-LMX relationships. Thus, for an introverted executive, a high-LMX relationship seems essential for success, but extraverts’ ability to seek social interaction, resources, and support suggests that extraversion may serve as a substitute for leadership (Kerr & Jermier, 1978). Even though we hypothesized that extraversion would act as a moderator of LMX, it is also possible to interpret the results in a way in which LMX is the moderator. In other words, our results could be interpreted as indicating that high LMX acts as a substitute for extraversion. Our results indicate that there would be a difference in performance and turnover between introverts and extraverts for only those with low LMX. This is an encouraging finding given that personality is a fairly stable construct that is not likely to change over short periods of time. Even though extraverts seem to be at an advantage as new executives, introverts can deal with their limitations by seeking and forming high-LMX relationships with their immediate managers. Taken together, it seems apparent that personality matters to new employees, at least at the executive level. It is interesting to note that this sample had a relatively high level of extraversion (the average was 6.67 on a 9-point scale). Future research should be conducted to see whether these relationships hold in other samples of employees who hold lower level positions and/or have more variance in terms of personality, because research has shown that different behaviors for success are needed for those above and below middle management (e.g., McCauley et al., 1994).

Implications In terms of turnover intentions and turnover, it is apparent that those who are introverted and do not have high-quality LMX relationships are the most vulnerable executives. This has important implications for organizations, because they invest large sums of money in new hires, especially at the executive level, and a loss

of a high-level employee can also cause potential problems for competitiveness. This study was conducted in the pharmaceutical industry, in which industry secrets are closely held, and information is an important commodity. This underscores the need to work with new executives, especially those who are more introverted, to enhance retention. Future research that further investigates these relationships as well as other ways to avoid turnover (e.g., Barrick & Zimmerman, 2005) is encouraged. Extraverts were rated more highly in terms of performance at the bivariate level. This implies that organizations may want to consider including some measures of personality in their selection batteries. This is consistent with Barrick and Mount’s (1991) findings showing that personality is related to job outcomes for managers. Conversely, the negative effects of introversion on task performance can be ameliorated by building effective relationships with managers. Organizations can maximize the effectiveness of their human resources by reaching out to more introverted employees by means of mentoring or even providing LMX training for superiors (as described in Scandura & Graen, 1984) as a way to smooth the transition for introverted employees. It is also possible that the observed lower performance for introverts with low-LMX relationships reflects an observation bias on the part of the senior executives who were rating performance. For example, the true performance of introverts, who are less likely to seek out interactions, may go unnoticed. Conversely, it may be that extraverts are better able to communicate and clarify their ideas with their superiors even given a relatively low-quality exchange. However, it is also possible that for executives, the ability to communicate with others is a key job-performance dimension and that, thus, the ratings may reflect a true difference in performance levels that is based on personality. Therefore, it would be helpful to extend the present study with future research that includes objective performance ratings and/or peer performance ratings. Another set of interesting findings revolved around the demographics of new executive age and gender, which were unrelated to superior ratings of performance. However, men (59% of the new executive employee sample) reported having higher-quality LMX

308

BAUER, ERDOGAN, LIDEN, AND WAYNE

relationships than did women, whereas women reported having higher turnover intentions. This pattern of findings is consistent with previous research demonstrating a subtle difference in how men and women perceive the environment at higher levels of an organization (Lyness & Thompson, 2000). We would characterize this Fortune 500 organization as a traditional, hierarchical organization. The superiors who participated in this sample were primarily male (75% of the sample), and previous research on executives has found that women and younger employees may have a harder time forming strong relationships (Lyness & Thompson, 2000) and receiving developmental job experiences (Ohlott, Ruderman, & McCauley, 1994). This may, in part, relate to the “glass ceiling” that has been reported to exist in Fortune 500 organizations (Davies-Netzley, 1998; Davis & Thomas, 2004; Lyness & Thompson, 1997). Future research is encouraged that further examines the role that gender and age play in the formation of strong LMX relationships for new executives to ensure a level playing field and welcoming environment.

Potential Limitations Although 168 new executives and 111 senior executives participated, the final sample used for this study was 116 new executives (a 50% response rate), with 67 for the performance analyses (a 29% response rate). Although these response rates are consistent with previous longitudinal, new employee studies (Bauer et al., 1998), a potential concern exists that the lack of power somehow compromised our findings. The usual concern with low power is one of inability to detect true differences (Cohen & Cohen, 1983). The concern for low power is even more evident in studies of interaction, in which problems such as small sample size, unreliability in predictor variables, and range restriction in predictor variables have been cited as reasons that interactions often go undetected (Aguinis & Stone-Romero, 1997). These concerns are lessened in the present study by the strong support garnered for all of the interaction hypotheses. Interaction effects are very difficult to detect and typically account for 1%–3% of the variance (Aiken & West, 1991). The interactions we examined accounted for 5%–10% of the variance in our outcomes. Therefore, the moderation effects detected seem strong in comparison with those described as typical by Aiken and West (1991), and our ability to detect them with a modest sample size may be indicative of the magnitude of the effect sizes in the overall population. In fact, for the R2 found here for the interaction for job performance, our power was .84 (Borenstein et al., 2001). Another potential concern regarding the sample size is generalizability. However, our response and nonresponse analyses revealed no differences between respondents and nonrespondents at Time 2 or Time 3. Further, no differences were found between participants and nonparticipants in terms of actual days on the job or in their actual turnover rates. This implies that our sample was not overly skewed in terms of the study variables. Finally, we were able to replicate the interaction finding for turnover intentions with a cross-section of lower level jobs. This leads us to be less concerned about generalizability. Nonetheless, future research that extends the generalizability of this study is encouraged. In addition, correlations between two measures may be inflated if both are obtained from the same person at the same point in time using the same data-collection technique. We designed this study

to minimize this potential limitation. Our data collection was across three time periods over several months and then again after 31⁄2 years, which should have mitigated common-stimulus cues and consistent responding. Also, data were collected from three sources: superiors, new executives, and organizational records. This eliminates the concern of common-method bias for those measures. Finally, we ran tests conservatively by controlling for any additional variables that were theoretically related and that covaried with predictors and criteria (in this case, age, gender, and interaction hours). Therefore, we have confidence that the observed relationships are not simply attributable to common-method variance. A final limitation is our inability to make causal inferences, a shortcoming that plagues nonexperimental research. It is difficult to address causality, even in a longitudinal study, because we cannot know how respondent cognitions influence behaviors occurring between data collections. For example, when performance is measured at a later time than LMX, leaders may be thinking about the performance of subordinates prior to that measurement, and this could influence future interactions. Liden, Wayne, and Stilwell’s (1993) finding that LMX relationships appear to form within the first 2 weeks postentry, largely on the basis of perceived similarity and liking and not performance, indicates an LMX-toperformance causal direction. In fact, researchers (e.g., Bauer & Green, 1996) have argued that the relationship between LMX and performance is probably reciprocal. It would be helpful to have future research specifically address this issue of causality. With respect to causality between LMX and turnover, it is important to note that none of the executives in our sample left their jobs before 3 months. To the extent that Liden et al.’s (1993) finding that LMX relationships formed in 2 weeks generalizes to other organizations, it is unlikely that turnover intentions would determine LMX quality. We contend that newcomers who have exerted the effort associated with searching for and transitioning into a new position would not have decided within the first 2 weeks that they wanted to leave the organization. Therefore, although we acknowledge that it is possible for turnover intentions to drive LMX, we feel that an LMX-to-turnover causal direction is much more likely. As noted earlier, executives are key organizational members who have the ability to shape and influence the decisions made regarding a variety of important factors. We have identified key factors that may help or hinder new executives in “getting up to speed” quickly and remaining with an organization. This study illustrates the importance of personality as a moderator of the LMX-to-outcome relationship for new executives. We found that high LMX is helpful in achieving new executive adjustment, but this is only so for introverted executives. Extraverted new executives are likely to be successful in their new positions, but the same benefits can be achieved among introverts who succeed in forming high-LMX relationships with their superiors. Our results demonstrate the importance of paying attention to both personality and relationships for executive development.

References Aguinis, H., & Stone-Romero, E. F. (1997). Methodological artifacts in moderated multiple regression and their effects on statistical power. Journal of Applied Psychology, 82, 192–206.

LMX-TO-PERFORMANCE AND -TURNOVER RELATIONSHIPS Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage. Ashton, M. C., Lee, K., & Paunonen, S. V. (2002). What is the central feature of extraversion? Social attention versus reward sensitivity. Journal of Personality and Social Psychology, 83, 245–252. Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimensions and job performance: A meta-analysis. Personnel Psychology, 44, 1–26. Barrick, M. R., & Zimmerman, R. D. (2005). Reducing voluntary, avoidable turnover through selection. Journal of Applied Psychology, 90, 159 –166. Bauer, T. N., & Green, S. G. (1996). Development of leader–member exchange: A longitudinal test. Academy of Management Journal, 39, 1538 –1567. Bauer, T. N., & Green, S. G. (1998). Testing the combined effects of newcomer information seeking and manager behavior on socialization. Journal of Applied Psychology, 83, 72– 83. Bauer, T. N., Morrison, E. W., & Callister, R. R. (1998). Organizational socialization: A review and directions for future research. Research in Personnel and Human Resources Management, 16, 149 –214. Borenstein, M. T., Rothstein, H., Cohen, J., Schoenfeld, D., Berlin, J., & Lakatos, E. (2001). Power and precision: A computer program for statistical power analysis and confidence intervals (Version 2) [Computer software and manual]. Mahwah, NJ: Erlbaum. Bozionelos, N. (2003). Intra-organizational network resources: Relation to career success and personality. International Journal of Organizational Analysis, 11, 41– 66. Brass, D. J., Butterfield, K. D., & Skaggs, B. C. (1998). Relationships and unethical behavior: A social network perspective. Academy of Management Review, 23, 14 –31. Bureau of Labor Statistics, U.S. Department of Labor. (2000). Top executives. Retrieved from http://stats.bls.gov/oco/ocos012.htm Burke, R. J. (1984). Relationships in and around organizations: It’s both who you know and what you know that counts. Psychological Reports, 55, 299 –307. Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Costa, P. T., Jr., & McCrae, R. R. (1992). Revised NEO Personality Inventory (NEO-PI–R) and NEO Five-Factor Inventory (NEO-FFI) professional manual. Odessa, FL: Psychological Assessment Resources. Crant, J. M. (1995). The Proactive Personality Scale and objective performance among real estate agents. Journal of Applied Psychology, 80, 532–537. Davies-Netzley, S. A. (1998). Women above the glass ceiling: Perceptions on corporate mobility and strategies for success. Gender and Society, 12, 339 –355. Davis, J., & Thomas. K. M. (2004, August). The roles of candidate race and gender on promotions and developmental opportunity: Is there a double whammy? Paper presented at the annual meeting of the Academy of Management Conference, New Orleans, LA. Depue, R. A., & Collins, P. F. (1999). Neurobiology of the structure of personality: Dopamine facilitation of incentive motivation and extraversion. Behavior and Brain Sciences, 22, 491–569. Dickter, D. N., Roznowski, M., & Harrison, D. A. (1996). Temporal tempering: An event history analysis of the process of voluntary turnover. Journal of Applied Psychology, 81, 705–716. Digman, J. M. (1990). Personality structure: Emergence of the five-factor model. Annual Review of Psychology, 41, 417– 440. Eisenberger, R., Stinglhamber, F., Vandenberghe, C., Sucharski, I. L., & Rhoades, L. (2002). Perceived supervisor support: Contributions to perceived organizational support and employee retention. Journal of Applied Psychology, 87, 565–573. Erdogan, B., & Liden, R. C. (2002). Social exchanges in the workplace: A review of recent developments and future research directions in leader–

309

member exchange theory. In L. L. Neider & C. A. Schriesheim (Eds.), Leadership (pp. 65–114). Greenwich, CT: Information Age. Ferris, G. R. (1985). Role of leadership in the employee withdrawal process: A constructive replication. Journal of Applied Psychology, 70, 777–781. Gerstner, C. R., & Day, D. V. (1997). Meta-analytic review of leader– member exchange theory: Correlates and construct issues. Journal of Applied Psychology, 82, 827– 844. Graen, G. B., Liden, R. C., & Hoel, W. (1982). Role of leadership in the employee withdrawal process. Journal of Applied Psychology, 67, 868 – 872. Graen, G. B., & Scandura, T. A. (1987). Toward a psychology of dyadic organizing. Research on Organizational Behavior, 9, 175–208. Huselid, M. A., & Day, N. E. (1991). Organizational commitment, job involvement, and turnover: A substantive and methodological analysis. Journal of Applied Psychology, 76, 380 –391. Jablin, F. M. (2000). Organizational entry, assimilation, and disengagement/exit. In F. M. Jablin & L. L. Putnam (Eds.), The new handbook of organizational communication (pp. 732– 818). Thousand Oaks, CA: Sage. Judge, T. A., Bono, J. E., Ilies, R., & Gerhardt, M. W. (2002). Personality and leadership: A qualitative and quantitative review. Journal of Applied Psychology, 87, 765–780. Judge, T. A., Cable, D. M., Boudreau, J. W., & Bretz, R. D. (1995). An empirical investigation of the predictors of executive career success. Personnel Psychology, 48, 485–519. Judge, T. A., Martocchio, J. J., & Thoresen, C. J. (1997). Five-factor model of personality and employee absence. Journal of Applied Psychology, 82, 745–755. Kacmar, K. M., Carlson, D. S., & Brymer, R. A. (1999). Antecedents and consequences of organizational commitment: A comparison of two scales. Educational and Psychological Measurement, 59, 976 –994. Kammeyer-Mueller, J. D., & Wanberg, C. R. (2003). Unwrapping the organizational entry process: Disentangling multiple antecedents and their pathways. Journal of Applied Psychology, 88, 779 –794. Kerr, S., & Jermier, J. M. (1978). Substitutes for leadership: Their meaning and measurement. Organizational Behavior and Human Performance, 22, 375– 403. Kraimer, M. L., Wayne, S. J., & Jaworski, R. A. (2001). Sources of support and expatriate performance: The mediating role of expatriate adjustment. Personnel Psychology, 54, 71–99. Kramer, M. W. (1995). A longitudinal study of superior–subordinate communication during job transfers. Human Communication Research, 22, 39 – 64. Kramer, M. W., Callister, R. R., & Turban, D. B. (1995). Informationgiving and information-receiving during job transitions. Western Journal of Communication, 59, 151–170. Lee, T. W., Mitchell, T. R., Sablynski, C. J., Burton, J. P., & Holtom, B. C. (2004). The effects of job embeddedness on organizational citizenship, job performance, volitional absences, and voluntary turnover. Academy of Management Journal, 47, 711–722. Liden, R. C., & Maslyn, J. M. (1998). Multidimensionality of leader– member exchange: An empirical assessment through scale development. Journal of Management, 24, 43–72. Liden, R. C., Sparrowe, R. T., & Wayne, S. J. (1997). Leader–member exchange theory: The past and potential for the future. Research in Personnel and Human Resource Management, 15, 47–119. Liden, R. C., Wayne, S. J., & Sparrowe, R. T. (2000). An examination of the mediating role of psychological empowerment on the relations between the job, interpersonal relationships, and work outcomes. Journal of Applied Psychology, 85, 407– 416. Liden, R. C., Wayne, S. J., & Stilwell, D. (1993). A longitudinal study on the early development of leader–member exchange. Journal of Applied Psychology, 78, 662– 674.

310

BAUER, ERDOGAN, LIDEN, AND WAYNE

Lucas, R. E., Diener, E., Grob, A., Suh, E. M., & Shao, L. (2000). Cross-cultural evidence for the fundamental features of extraversion. Journal of Personality and Social Psychology, 79, 452– 468. Lyness, K. S., & Judiesch, M. K. (2001). Are female managers quitters? The relationships of gender, promotions, and family leaves of absence to voluntary turnover. Journal of Applied Psychology, 86, 1167–1178. Lyness, K. S., & Thompson, D. E. (1997). Above the glass ceiling? A comparison of matched samples of female and male executives. Journal of Applied Psychology, 82, 359 –375. Lyness, K. S., & Thompson, D. E. (2000). Climbing the corporate ladder: Do female and male executives follow the same route? Journal of Applied Psychology, 85, 86 –101. Maertz, C. P., & Campion, M. A. (1998). 25 years of voluntary turnover research: A review and critique. In C. L. Cooper & I. T. Robertson (Eds.), International review of industrial and organizational psychology (Vol. 13, pp. 49 – 81). New York: Wiley. Mainiero, L. A. (1994). Getting anointed for advancement: The case of executive women. Academy of Management Executive, 8, 53– 67. Major, D. A., Kozlowski, S. W. J., Chao, G. T., & Gardner, P. D. (1995). A longitudinal investigation of newcomer expectations, early socialization outcomes, and the moderating effect of role development factors. Journal of Applied Psychology, 80, 418 – 431. Masterson, S. S., Lewis, K., Goldman, B. M., & Taylor, M. S. (2000). Integrating justice and social exchange: The differing effects of fair procedures and treatment on work relationships. Academy of Management Journal, 43, 738 –748. McCall, M. W., Lombardo, M. M., & Morrison, A. M. (1988). The lessons of experience: How successful executives develop on the job. New York: Free Press. McCauley, C. D., Ruderman, M. N., Ohlott, P. J., & Morrow, J. E. (1994). Assessing the developmental components of managerial jobs. Journal of Applied Psychology, 79, 544 –560. Michaels, E., Handfield-Jones, H., & Axelrod, B. (2001). The war for talent. Boston: Harvard Business School Press. Morita, J. G., Lee, T. W., & Mowday, R. T. (1993). The regression-analog to survival analysis: A selected application to turnover research. Academy of Management Journal, 36, 1430 –1464. Ohlott, P. J., Ruderman, M. N., & McCauley, C. D. (1994). Gender differences in managers’ developmental job experiences. Academy of Management Journal, 37, 46 – 67. Overton, B. (2001). Using peer group performance in executive pay. Executive Compensation, 27(1), 1– 8. Saucier, G. (1994). Mini-Markers: A brief version of Goldberg’s unipolar Big-Five markers. Journal of Personality Assessment, 63, 506 –516. Scandura, T. A., & Graen, G. B. (1984). Moderating effects of initial leader–member exchange status on the effects of a leadership intervention. Journal of Applied Psychology, 69, 428 – 436. Schiemann, W. A. (1977). The nature and prediction of organizational communication: A review of the literature and an empirical investiga-

tion. Unpublished doctoral dissertation, University of Illinois at Urbana– Champaign. Schriesheim, C. A., Castro, S. L., & Yammarino, F. J. (2000). Investigating contingencies: An examination of the impact of span of supervision and upward controllingness on leader–member exchange using traditional and multivariate within- and between-entities analysis. Journal of Applied Psychology, 85, 659 – 677. Seibert, S. E., Kraimer, M. L., & Liden, R. C. (2001). A social capital theory of career success. Academy of Management Journal, 44, 219 – 237. Settoon, R. P., Bennett, N., & Liden, R. C. (1996). Social exchange in organizations: The differential effects of perceived organizational support, leader–member exchange, and employee reciprocity. Journal of Applied Psychology, 81, 219 –227. Sheridan, J. E. (1992). Organizational culture and employee retention. Academy of Management Journal, 35, 1036 –1056. Singer, J. D., & Willett, J. B. (1991). Modeling the days of our lives: Using survival analysis when designing and analyzing longitudinal studies of duration and the timing of events. Psychological Bulletin, 110, 268 –290. Sparrowe, R. T., & Liden, R. C. (1997). Process and structure in leader– member exchange. Academy of Management Review, 22, 522–552. Sparrowe, R. T., & Liden, R. C. (2005). Two routes to influence: Integrating leader–member exchange and network perspectives. Administrative Science Quarterly, 50, 505–535. Vecchio, R. P. (1985). Predicting employee turnover from leader–member exchange: A failure to replicate. Academy of Management Journal, 28, 478 – 485. Vecchio, R. P., Griffeth, R. W., & Hom, P. W. (1986). The predictive utility of the vertical dyad linkage approach. Journal of Social Psychology, 126, 617– 625. Wanberg, C. R., & Kammeyer-Mueller, J. D. (2000). Predictors and outcomes of proactivity in the socialization process. Journal of Applied Psychology, 85, 373–385. Watson, D., & Clark, L. A. (1997). Extraversion and its positive emotional core. In R. Hogan, J. Johnson, & S. Briggs (Eds.), Handbook of personality psychology (pp. 767–793). San Diego, CA: Academic Press. Wayne, S. J., Shore, L. M., & Liden, R. C. (1997). Perceived organizational support and leader–member exchange: A social exchange perspective. Academy of Management Journal, 40, 82–111. Welbourne, T. M., Johnson, D. E., & Erez, A. (1998). The Role-Based Performance Scale: Validity analysis of a theory-based measure. Academy of Management Journal, 41, 540 –555. Williamson, I. O., & Cable, D. M. (2003). Organizational hiring patterns, interfirm network ties, and interorganizational imitation. Academy of Management Journal, 46, 349 –358.

Received July 23, 2004 Revision received March 21, 2005 Accepted April 6, 2005 䡲