Leadership, supervisor-focused justice, and follower values: A comparison of three. leadership approaches in China

Size: px
Start display at page:

Download "Leadership, supervisor-focused justice, and follower values: A comparison of three. leadership approaches in China"

Transcription

1

2 Leadership, supervisor-focused justice, and follower values: A comparison of three leadership approaches in China Dissertation submitted to the Graduate School of the University of Cincinnati in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Management of the College of Business 2012 By Jie Li B.S., Sichuan University, China, 1996 M.B.A., Washington State University, 2003 Dissertation Committee: Chair: Dr. Suzanne Masterson Committee Members: Dr. Lawrence Gales; Dr. Wei Pan

3 ABSTRACT Although research has examined the relationship between employees justice perceptions of authorities and their subsequent attitudes and behaviors, inadequate efforts have been taken to integrate leadership theories and justice research. Using data from China, this dissertation simultaneously examines how supervisor-focused procedural, distributive, interpersonal, and informational justice perceptions affect the effectiveness of transformational leadership, transactional leadership, and paternalistic leadership. In addition, due to the profound socioeconomic changes in the past few decades in China, Chinese employees values have become more diverse and may affect the effectiveness of different leadership styles. Thus, this dissertation also examines how the Chinese cultural values of traditionality and modernity moderate the relationships between the three leadership approaches and supervisor-focused justice perceptions. The results of this study reveal that when the three leadership approaches are juxtaposed, they display differential total effect on in-role performance and organizational citizenship behaviors (OCB). Transactional leadership has a weak positive total effect on in-role performance while transformational leadership shows a strong positive total effect on OCB but not on in-role performance. However, paternalistic leadership does not display any significant total effect on the performance variables. This study also suggests that the three leadership approaches have differential impacts on the followers justice perceptions. Transformational leadership is the most influential antecedent of all four dimensions of supervisor-focused justice perceptions whereas paternalistic leadership only positively influences supervisor-focused informational justice. Transactional leadership has no significant effect on supervisor-focused justice perceptions when transformational leadership and paternalistic leadership are controlled ii

4 for. The findings also indicate that not all supervisor-focused justice dimensions are instrumental for the leadership approaches to affect the followers in-role and extra-role performance. Supervisor-focused procedural justice and supervisor-focused interpersonal justice mediate the effect of transformational leadership on the followers performance. Justice perceptions do not mediate transactional leadership and paternalistic leadership s effect on the followers performance when the three leadership styles are examined together. Furthermore, Chinese employees value of traditionality is found to moderate the relationship between transformational leadership and supervisor-focused informational justice and that between transformational leadership and supervisor-focused procedural justice. However, Chinese employees value of modernity does not moderate the leadership-justice relationships. By comprehensively examining multiple leadership approaches, multiple leader-focused justice perceptions and the followers values, this study is able to provide answers to some of the important leadership questions in the unique Chinese setting. The current study tests the effect of each of the three leadership approaches on the followers performance when the other two leadership approaches are controlled for. It also examines the efficacy of the different dimensions of supervisor-focused justice perceptions in mediating the effect of different leadership approaches on the followers performance. Furthermore, this study provides empirical evidence on how followers with different levels of certain values respond differently to transformational leadership. iii

5 Copyright by Jie Li 2012 iv

6 DEDICATION To my wife and my son, for the great sacrifices they have made while I was pursuing my doctorate. I cannot finish this journey without their unwavering love and support. To my unborn child who brings great joy to my family when I finish this dissertation. To my parents and my brother, for their everlasting care and encouragement. To all the teachers and mentors I have met, for their guidance and advice. v

7 ACKNOWLEDGEMENTS I would like to express my sincere gratitude to my dissertation committee for their support in the completion of this work. I thank my dissertation chair Dr. Suzanne Masterson for her insightful questions, thoughtful suggestions, endless patience, and invaluable guidance. I am very grateful for the confidence that Dr. Lawrence Gales has in me, and for his candid and concise critics on my work. Also, I greatly appreciate the encouragement and statistical support provided by Dr. Wei Pan. I want to thank Dr. Minli Yi at the Southwestern University of Finance and Economics in China. Dr. Yi has provided great support and advice to me in the past 13 years. Finally, many thanks to Dr. Elaine Hollensbe, Dr. Rajan Kamath, Dr. Ann Welsh, Dr. Ralph Katerberg, and Ms. Tricia Burger for their help and support. vi

8 TABLE OF CONTENTS ABSTRACT... II DEDICATION... V ACKNOWLEDGEMENTS... VI LIST OF TABLES... IX LIST OF FIGURES... X CHAPTER 1: INTRODUCTION... 1 AN OVERVIEW OF THE CURRENT LITERATURE... 2 POTENTIAL CONTRIBUTIONS... 6 STRUCTURE OF THE DISSERTATION... 7 CHAPTER 2: LITERATURE REVIEW... 8 TRANSFORMATIONAL AND TRANSACTIONAL LEADERSHIP... 8 Definition... 8 Measurement of transformational and transactional leadership How transformational and transactional leadership affect subordinates PATERNALISTIC LEADERSHIP Definition Measurement of paternalistic leadership How paternalistic leadership affect subordinates ORGANIZATIONAL JUSTICE Definition Measurement of organizational justice How organizational justice affect employees SUMMARY CHAPTER 3: THEORY AND HYPOTHESES THE DIRECT EFFECT OF LEADERSHIP ON OCB THE DIRECT EFFECT OF LEADERSHIP ON IN-ROLE PERFORMANCE THE DIRECT EFFECT OF LEADERSHIP ON SUPERVISOR-FOCUSED PROCEDURAL JUSTICE THE DIRECT EFFECT OF LEADERSHIP ON SUPERVISOR-FOCUSED DISTRIBUTIVE JUSTICE THE DIRECT EFFECT OF LEADERSHIP ON SUPERVISOR-FOCUSED INTERPERSONAL JUSTICE THE DIRECT EFFECT OF LEADERSHIP ON SUPERVISOR-FOCUSED INFORMATIONAL JUSTICE MEDIATION HYPOTHESES MODERATION HYPOTHESES CHAPTER 4: METHODS DESIGN vii

9 SAMPLE AND PROCEDURE Step 1: Subordinate Survey Step 2: Supervisor Survey MEASURES CONTROL VARIABLES ANALYSES Data Treatment Measurement Model Structural Model and Hypotheses Testing CHAPTER 5: RESULTS MEASUREMENT MODEL HYPOTHESES TESTING Mediation Moderation CHAPTER 6: DISCUSSION & CONCLUSION DISCUSSION FUTURE RESEARCH LIMITATIONS PRACTICAL IMPLICATION CONCLUSION viii

10 LIST OF TABLES Table Page 1. Subordinate Questionnaire Supervisor Questionnaire Results of Confirmatory Factor Analysis Parcels for Latent Constructs Descriptive Statistics, Correlations, and Reliabilities Comparison of Structural Equation Models of Mediation Bootstrapping Results of Total, Direct, and Total Indirect Effects Specific Indirect Effect of IV on DV through Proposed Mediators Test of Moderation ix

11 LIST OF FIGURES Figure Page 1. Theoretical Model Hypothesized 11-Factor Measurement Model Final Structural Model Traditionality Moderates the Transformational Leadership---Supervisor-focused Informational Justice Relationship Traditionality Moderates the Transformational Leadership---Supervisor-focused Procedural Justice Relationship. 78 x

12 CHAPTER 1: INTRODUCTION Globalized competition and cooperation and the increasing impact of China on the world economy call attention to the need to understand effective leadership in the contemporary Chinese business context. In the past few decades, China has gone through profound socioeconomic reforms and shifted toward a market-oriented economy. The coexistence of Chinese traditional cultural values and modern Western management philosophies and the business practices based on them respectively created an intriguingly rich and complex context. This dissertation explores how various leadership approaches function in Chinese organizations. It has been suggested that employees fairness judgments about leader behaviors affect followers performance (e.g., Masterson, Lewis, Goldman, & Taylor, 2000; Rupp & Cropanzano, 2002). This dissertation integrates leadership and justice by taking a relatively more comprehensive view on how justice perceptions associated with different leadership styles affect followers performance in Chinese organizations. Specifically, the four-factor model of organizational justice (Colquitt, 2001) and its role in leadership effectiveness is examined and empirically tested. In addition, given the particular context of this study, Chinese employees cultural values may also play a role in leader effectiveness. Thus, the dissertation also investigates how various leadership approaches affect Chinese employees who hold different cultural values. Through simultaneously examining transformational leadership, transactional leadership, and paternalistic leadership, this dissertation may also shed some light on the efficacy of each type of leadership in affecting followers performance in China. This chapter provides an overview of this dissertation. First, it briefly lays out the gaps in the extant literature on leadership in the Chinese context. Then, the potential contribution of this study is discussed. Finally, the structure of the dissertation is provided at the end of this chapter. 1

13 An Overview of the Current Literature Management scholars have taken some efforts in studying three leadership styles, i.e., transformational leadership, transactional leadership, and paternalistic leadership that are often used by Chinese business leaders. Several observations can be made regarding the current status of the literature. First, research attention has mainly focused on how each of these leadership styles independently affects followers. Several studies examined the mechanisms through which transformational leadership influences Chinese employees. For example, Wang and colleagues suggested that leader-member exchange mediated the relationship between transformational leadership and employees task performance and organizational citizenship behaviors (OCBs) (Wang, Law, Hackett, Wang, & Chen, 2005). Kirkman, Chen, Farh, Chen, and Lowe (2009) found that procedural justice mediated transformational leadership s effect on followers OCBs. In addition, using samples from banking and finance sectors in China and India, Walumbwa and colleagues suggested that collective efficacy mediated the relationship between transformational leadership and followers organizational commitment, job satisfaction, and withdrawal behaviors (Walumbwa, Wang, Lawler, & Shi, 2004). Although the most effective leaders were said to be both transformational and transactional (Bass, 1985), research on how transactional leadership influences Chinese employees has been lacking. Wei, Yuan, and Di (2010) found that transactional leadership was positively related to followers creative performance in teams with high empowerment climate, but negatively related to subordinates creative performance in lower empowerment climate. Such interactional effect was partially mediated by followers psychological empowerment perception. Given the economic development and the shift away from egalitarianism toward 2

14 equity-oriented outcome allocation in Chinese organizations (e.g., Chen, 1995), Chinese business leaders increasingly reward followers contingently on the basis of their work inputs and performance. Thus, transactional leadership is very likely a widely used leadership approach in Chinese organizations. Other research has focused on paternalism, a salient cultural value built on familism, that is prevalent in many parts of the world (Aycan, 2006). In China, paternalistic leadership derived from paternalism and is based on Chinese cultural roots in Confucian philosophy (Farh & Cheng, 2000). It is an important leadership style in Chinese societies. The components of the triad model of paternalistic leadership (i.e., authoritarianism, benevolence, and morality) proposed by Farh and Cheng (2000) were found to have significant impact on employees attitudes and behaviors even after the effect of three dimensions of transformational leadership (i.e., individualized consideration, high performance standard, and modeling) was controlled for in a Taiwanese sample (Cheng, Chou, Wu, Huang, & Farh, 2004). In a mainland Chinese sample, the effects of authoritarian leadership, benevolent leadership, and moral leadership were found to be mediated by fear of supervisor, gratitude and repayment to supervisor, and identification with supervisor respectively (Farh, Cheng, Chou, & Chu, 2006). However, how paternalistic leadership, as a holistic approach, influences followers and how it differs from transformational leadership as a whole, rather than some of its dimensions, have not been adequately researched in China. So far, only a limited number of studies have been conducted to examine transformational leadership, transactional leadership, and paternalistic leadership in China. Considering their coexistence in China, their importance to Chinese business leaders, and the ever increasing importance of China in the world economy, Chinese organizations provide an excellent context for investigating these leadership approaches. Although Chinese business 3

15 leaders can choose to engage in a mix of leadership behaviors that involve the three types of leadership to varying degrees, existing research on the three leadership approaches in China have largely treated them in isolation. To get a more realistic picture of the similarities and differences of the three types of leadership in China, it is necessary to simultaneously assess how they affect employees so that the impact of different leadership styles can be effectively compared. Therefore, this dissertation chooses to simultaneously focus on the three leadership styles in Chinese organizations. Second, more efforts are needed to examine how organizational justice theories inform leadership theories. Given the importance of fairness in leadership (cf. van Knippenberg & De Cremer, 2008; van Knippenberg, De Cremer, & van Knippenberg, 2007), there is a need to integrate leadership theories and organizational justice theories. Particularly, supervisor-focused justice is important as past research has shown that employees fairness evaluation of supervisors has significant impact on the followers attitudes and behaviors directed at both the leader- and the organization-levels (Liao & Rupp, 2005; Masterson et al., 2000; Rupp & Cropanzano, 2002). Despite the move toward studying supervisor-focused justice perceptions in organizational justice research, the extant research on transformational and transactional leadership mainly looked at organization-focused justice perceptions (e.g., Kirkman et al., 2009; Pillai, Scandura, & Williams, 1999; Pillai, Schriesheim, & Williams, 1999). Moreover, in the Chinese setting, only procedural justice has received researchers attention (Kirkman et al., 2009). Given the current strong emphasis on material wellbeing in the Chinese society, distributive justice is likely to assume a lot of weight when employees form fairness judgments (Kim & Leung, 2007). In addition, interpersonal justice and informational justice are also important antecedents of employees behaviors and attitudes (Colquitt, 2001; 4

16 Liao & Rupp, 2005). Clearly, all four aspects of justice have yet to be considered to improve our understanding of the relationship between leadership and follower performance. Thus, this dissertation examines the relationships between the three leadership styles and the four aspects of supervisor-focused justice comprehensively. Third, the dynamic changes in Chinese workplace need to be considered in order to provide us a realistic understanding of how leadership functions in Chinese organizations. Although the effects of transformational leadership, transactional leadership, and paternalistic leadership have been found to be generalizable across cultures (Aycan, 2006; Bass, 1997; Den Hartog, House, Hanges, Ruiz-Quintanilla, & Dorfman, 1999; Dorfman et al., 1997), the way these leadership approaches work may be bounded by employees cultural values. For example, employees power distance orientation was found to moderate the relationship between transformational leadership and procedural justice (Kirkman et al., 2009). In a similar vein, cultural values may also put constraints on how Chinese employees perceive transactional leadership and paternalistic leadership, thereby affecting the employees reactions to the leader. However, Chinese employees values have changed and become more diverse (Egri & Ralston, 2004; Ralston, Egri, Stewart, Terpstra, & Yu, 1999) during the profound and dynamic socioeconomic reform in the past few decades. As a result, different patterns of cultural value combinations may exist across individual employees who may react to various leadership approaches differently. Traditionality and modernity (Farh, Earley, & Lin, 1997) are cultural values that can effectively capture the differences between different groups of Chinese employees who coexist in China s modernization process. Researchers have yet to move beyond the commonly used Hofstede s cultural dimensions such as power distance (e.g., Kirkman et al., 2009) and collectivism (Walumbwa & Lawler, 2003) to capture the more recent value changes 5

17 in the Chinese workplace and investigate the boundary effect of employees cultural values on leadership effectiveness in China. To sum up, several major gaps in the extant literature on leadership in China need to be addressed. First, multiple leadership approaches such as transformational leadership, transactional leadership, and paternalistic leadership need to be examined simultaneously in Chinese organizations in order to holistically compare their effectiveness. Second, there is a need to integrate leadership and supervisor-focused justice theories. Third, all four aspects of supervisor-focused justice have yet to be considered concurrently in order to get a complete picture of how fairness affect leadership effectiveness. Finally, the cultural values of Chinese employees have changed in the recent few decades and consequently Chinese workplace has become more diverse. Researchers need to explore how employees cultural values may interact with various leadership styles and affect employees reactions to the leader. Potential Contributions This dissertation study contributes to the literature and practice in the following ways. First, this dissertation moves beyond the traditional transformational-transactional leadership paradigm by incorporating paternalistic leadership, a leadership approach that is of particular importance to organizations across the world including Chinese companies (Farh & Cheng, 2000; Pellegrini & Scandura, 2008). Given the different theoretical underpinnings of each leadership approach, only by juxtaposing multiple leadership styles in the same study can we fully understand their similarities and differences. This study has the potential to unravel how paternalistic leadership differs from transformational and transactional leadership in the Chinese setting. 6

18 Second, it extends the current literature by integrating multiple leadership theories and organizational justice theories. Several studies implied that the foci of employees justice perceptions have differential effects on employees performance, behaviors and attitudes, and that supervisor-focused justice perceptions are particularly relevant for us to understand leadership (Masterson et al., 2000; Rupp & Cropanzano, 2002). This dissertation comprehensively examines the roles that the various aspects of supervisor-focused justice perceptions play in affecting the effectiveness of multiple leadership styles. Third, this study extends the current literature by taking into consideration some dynamic characteristics of the Chinese workplace. It incorporates two cultural values that were developed indigenously in the Chinese setting (i.e., traditionality and modernity), thus developing an understanding of leadership more specifically suited to the contemporary Chinese business context. Through understanding how Chinese employees with different cultural values may perceive and react to different leadership approaches, business leaders can be more effective by adjusting their leadership styles when they interact with Chinese employees. Structure of the Dissertation The content of this dissertation is as follows. First, Chapter Two reviews the literature on leadership and organizational justice with an emphasis on cross-cultural research, particularly the studies done in the Chinese context, and a theoretical model is developed. Then, Chapter Three lays out specific hypotheses based on the theoretical model. Next, research methods are discussed in Chapter Four. In Chapter Five, the results are presented. Finally, the findings, implications for research and practice, and limitations are discussed in Chapter Six that also concludes this dissertation with some key contributions of it to the literature and practice. 7

19 CHAPTER 2: LITERATURE REVIEW This chapter reviews the literature on transformational/transactional leadership, paternalistic leadership, and organizational justice to provide a backdrop for the current study. First, I provide the definition for each of the above major constructs. Then, I discuss how each of the constructs is measured. Next, I review the processes through which the leadership behaviors and fairness perceptions affect employees attitudes and behaviors. Finally, I provide a brief summary of this section and develop the theoretical model. Transformational and Transactional Leadership Definition Building on Burns (1978) conceptualization of transactional and transforming leadership, Bass (1985) further elaborated and developed the theory of transformational and transactional leadership. According to Bass (1985), transformational leadership moves the followers beyond immediate self-interests through idealized influence (charisma), inspirational motivation, intellectual stimulation, or individualized consideration. It elevates the follower s level of maturity and ideals as well as concerns for achievement, self-actualization, and the well-being of others, the organization, and society. (Bass, 1999, p. 11) Among the transformational leadership factors (Bass & Avolio, 1993, pp ), idealized influence (charisma) involves leaders displaying strong convictions in their own beliefs and ideals, presenting their most important values, and emphasizing the purpose, commitment and ethical consequences of decisions. As a result, followers tend to identify with and trust them and see them as role models. Inspirational Motivation refers to leaders arousing followers emotions through the use of symbols and emotional arguments, articulating a compelling vision, challenging followers with high standards, 8

20 exhibiting optimism and enthusiasm, and providing meaning for what needs to be done. It may or may not overlap with idealized influence (charisma), depending on the extent to which followers identify with the leader. Intellectual stimulation involves leaders encouraging followers to question the status quo and to break with the past. The followers are encouraged by the leader to question the values, beliefs and expectations of their own, the leader, and the organization. The leader also promotes followers independent thinking and creative problem solving. Individualized consideration means that leaders treat followers differently but equitably on a one-to-one basis. Through delegating and coaching, leaders can raise the maturity of followers needs and make them more effective in meeting goals and challenges. In contrast, transactional leadership focuses on a cost-benefit, economic exchange to meet subordinates current material and psychic needs in return for contracted services rendered by the subordinate (Bass, 1985, p. 14). Among the transactional leadership factors (Bass & Avolio, 1993, pp ), contingent reward involves a positively reinforcing interaction such as the leader s providing appropriate rewards in exchange for the followers meeting objectives. Followers needs are linked to the leader s expectations and rewards. Management-by-exception represents another factor of transactional leadership whereby leaders use corrective criticism, negative feedback and negative contingent reinforcement only when things go wrong. Finally, laissez-faire refers to a situation in which leadership is absent. Such leaders avoid taking actions; delay decisions, feedback and rewards; and make no efforts to motivate followers or to recognize and satisfy their needs. Bass (1985) noted that transformational and transactional leadership are distinct but not mutually exclusive processes. Transactional leadership depends on the leader s power to reinforce subordinates in exchange for their successful completion of the tasks. However, 9

21 transformational leadership engages and commits the followers by adding their sense of selfworth, one of the strongest motivators, to the transactional exchange (Bass, 1997). Transformational leadership thus can induce extra efforts from employees and generate performance beyond expectations. In other words, transformational leadership was believed to augment transactional leadership in predicting employee outcome variables (Bass & Avolio, 1993). Evidence has shown that many effective leaders are both transformational and transactional (Bass, 1985). Transformational and transactional leadership is found to be applicable across many countries (Bass, 1997; Den Hartog et al., 1999) including China. Measurement of transformational and transactional leadership Bass (1985) proposed that transformational/transactional leadership consisted of seven factors, i.e. charisma (idealized influence), inspirational motivation, intellectual stimulation, individualized consideration, contingent reward, management by exception, and laissez-faire; and these factors are measured by the Multifactor Leadership Questionnaire (MLQ). Later, Bass and Avolio (1990) added another factor called active management by exception to the theory. In responding to criticisms of the earlier theory and measure (cf. Bass & Avolio, 1993), Bass and Avolio (1995) revised the transformational and transactional leadership measure and developed the MLQ (Form 5X). Together, transformational and transactional leadership are also called the Full Range Leadership Theory and it has been expanded to contain nine distinct factors (Bass & Avolio, 1997). In the nine-factor model, transformational leadership consists of five factors: Idealized Influence (attributed), Idealized Influence (behavior), Inspirational Motivation, Intellectual Stimulation, and Individualized Consideration. Transactional leadership subsumes three factors: Contingent Reward, Active Management-by-Exception, and Passive Management-by-Exception. 10

22 The last factor is the nontransactional Laissez-faire leadership. Antonakis, Avolio, and Sivasubramaniam (2003) examined the nine-factor model using large samples and concluded that the nine-factor model best represented the factor structure underlying the MLQ (Form 5X). The constructs under transformational leadership were often highly correlated indicating a higher-order construct of transformational leadership (Antonakis et al., 2003; Avolio, Bass, & Jung, 1999). Extant research indicates that the MLQ has been used in many countries including China (Bass, 1997). For example, Walumbwa and colleagues (Walumbwa et al., 2004; Wang & Walumbwa, 2007) found that the MLQ displayed measurement invariance across China, Kenya, Thailand, and India. Podsakoff, MacKenzie, Moorman, and Fetter (1990) proposed a different measure of transformational leadership, the Transformational Leadership Index. They identified six factors for transformational leadership. Specifically, Identifying and Articulating a Vision (leader identifying new opportunities for the unit/organization, and developing, articulating, and inspiring others with the vision), Providing an Appropriate Model (leader setting an example that is consistent with the values that the leader espouses), Fostering the Acceptance of Group Goals (leader promoting cooperation and getting followers to work together toward a common goal), High Performance Expectations (leader expressing expectations for excellence, quality, and/or high performance on the part of followers), Providing Individualized Support (leader respecting followers personal feelings and needs), and Intellectual Stimulation (leader challenging followers to re-examine some of their assumptions about their work and rethink how it can be performed). Strong evidence that supports the factor structure, reliability and validity of the Transformational Leadership Index has been found (e.g., Kirkman et al., 2009; MacKenzie, 11

23 Podsakoff, & Rich, 2001; Podsakoff, MacKenzie, & Bommer, 1996). Three factors, articulating a vision, providing an appropriate model, and fostering the acceptance of group goals were found to be highly intercorrelated and were called core transformational leader behavior (Podsakoff et al., 1990). Podsakoff, Todor, Grover, and Huber (1984) also developed a five-item measure of Contingent Reward as the operationalization of transactional leadership. This measure focuses on the extent to which a leader provides rewards in exchange for the followers work effort. It captures the exchange notions fundamental to transactional leader behavior and is consistent with Bass (1985) conceptualization of transactional leadership (Podsakoff et al., 1990, p. 113). The Transformational Leadership Index has been tested in Chinese business context. Wang and colleagues (Wang et al., 2005) collected data from Chinese employees and their supervisors and reported that an overall second-order factor of transformational leadership measured by Podsakoff et al. s (1990) scale fitted their data reasonably well. Kirkman et al. (2009) used a shortened version of the Transformational Leadership Index to survey Chinese employees and found that the multiple facets of transformational leadership formed an overall, higher-order factor. In addition, they found measurement equivalence for the Transformational Leadership Index between the U.S. and Chinese samples. How transformational and transactional leadership affect subordinates Transformational leadership has been found to have impact on followers attitudes, performance, and behaviors (e.g., Barling, Weber, & Kelloway, 1996; Dvir, Eden, Avolio, & Shamir, 2002; Wang et al., 2005). Although researchers are still trying to specify the mediating processes between leader transformational and transactional behavior and performance, a number of mechanisms through which leaders affect subordinates have been identified. Personal 12

24 identification, social identification, internalization, trust, value congruence, procedural justice, social exchange, emotions and moods, and instrumental compliance are among some of the identified mechanisms. Leader charisma was believed to be able to change the way followers perceive themselves (e.g., Shamir, House, & Arthur, 1993). Transformational leadership has a component of charisma (idealized influence and inspirational motivation). It may affect followers through the creation of personal identification with the leader and social identification with the work unit, and the different forms of identification could lead to different outcomes (Kark & Shamir, 2002). Personal identification could be defined as the process whereby an individual s belief about a person (a leader) becomes self-referential or self-defining (Kark & Shamir, 2002, p. 70). Pratt (1998) suggested that two modes of identification were associated with personal identification. First, evoking followers self-concepts in the recognition that they share similar values with the leader; and second, giving rise to followers desires to change their self-concepts so that their values and beliefs become more similar to those of the leader. Personal identification was found to mediate the relationship between transformational leadership and followers dependence on the leader (Kark, Shamir, & Chen, 2003). Social identification is defined as the process whereby an individual s belief about a group (or an organization) becomes self-referential or self-defining (Kark & Shamir, 2002, p. 70). Charismatic/transformational leaders could strengthen the connection between followers self-concepts and the group, and raise followers identification with the group (Bass & Avolio, 1994; Shamir, Zakay, Breinin, & Popper, 1998). As a result, followers became more willing to contribute to the group s mission and goals. Followers social identification was found to 13

25 mediate the relationship between transformational leadership and followers self-efficacy, organization-based self-esteem, and collective efficacy (Kark et al., 2003). Internalization is a process of influence through which the target person perceives the induced behavior as congruent with his value system and intrinsically rewarding (Kelman, 1958). Transformational leaders are likely to increase followers felt intrinsic motivation in the task by expressing high standards and expectations and setting challenging goals, encouraging followers independent thinking, empowering followers to increase their self-efficacy, delegating decision making to increase their autonomy and providing meaning to the task to increase its significance. Intrinsic motivation has been found to mediate the relationship between transformational leadership and sports performance (Charbonneau, Barling, & Kelloway, 2001). Trust has been identified as an important mediator between transformational leadership and followers attitudes and behaviors (e.g., Bennis & Nanus, 1985). Transformational leaders are likely to increase their perceived trustworthiness through several means. First, they display high self-confidence and strong conviction to their espoused ideals and vision, thus they are likely to be perceived as capable. Second, they give followers individualized consideration and coach followers, therefore they are likely to be seen as benevolent. Third, they tend to be perceived as having high integrity because they often honor agreements, make self-sacrifices and emphasize the most important values and ethical consequences of decisions (Bass, 1985). As a result, the perceived ability, benevolence and integrity enjoyed by transformational leaders are likely to generate trust in followers (Mayer, Davis, & Schoorman, 1995). Empirical evidence showed that trust was an important mediator between transformational leadership and followers attitudes and behaviors (Pillai, Schriesheim, et al., 1999; Podsakoff et al., 1990). 14

26 In addition, transformational leadership is able to commit followers to the important values and beliefs of the leader so that the followers move beyond their self-interest. In such a process, followers internalize the leader s values which become congruent with the followers own personal values. Value congruence has a positive influence on individual behaviors and attitudes (Meglino, Ravlin, & Adkins, 1989), and it is also found to be a mediator between transformational leadership and followers outcomes (Jung & Avolio, 2000). Procedural justice can be generally understood as the perceived fairness of the value allocation procedures. The group value model of procedural justice (Lind & Tyler, 1988) proposed that value allocation procedures could signify that the employees were valuable members of the group with good standing and status; therefore, those procedures were likely seen as fair. Because under transformational leadership, followers tend to go beyond self-interest and identify with the group and because they also receive individualized attention from the leader, followers tend to develop a strong sense of belongingness to the organization. This is consistent with the group value model of procedural justice. Kirkman et al. (2009) found that procedural justice mediates the relationship between transformational leadership and followers citizenship behaviors in China. In addition, Pillai and colleagues also found that procedural justice mediates the relationship between transformational leadership and employees attitudes and behaviors in the U.S. and other countries (Pillai, Scandura, et al., 1999; Pillai, Schriesheim, et al., 1999). Social exchange process is another mechanism through which transformational leaders impact their followers. Under transformational leaders, employees may receive individualized attention from the leader. As a result, they tend to reciprocate by supporting the leader s agenda and performing beyond expectations. Hence, transformational leaders can develop high quality 15

27 leader member exchange relationships with followers, through which they influence followers performance (e.g., Wang et al., 2005). Although the initial stage of LMX may be transactional, it can be transformational if the last stage is reached (Bass, 1999). Transformational leadership may also affect subordinates through positive emotional expressions and mood contagion. Transformational leaders often inspire followers through arousing their emotions. They use symbols and emotional arguments to articulate a compelling vision, display passion toward their ideals and beliefs, and exhibit high optimism and enthusiasm (Bass, 1985). Their positive emotional expression was able to transfer positive affect from them to the followers and consequently influenced followers ratings of leader effectiveness and attraction to the leader (Bono & Ilies, 2006). In contrast to transformational leadership, transactional leadership relies on a contractual or exchange process between leaders and followers. Transactional leaders clarify their expectations and use positive or negative reinforcement in exchange for followers performance (Bass, 1985). Essentially, the primary mechanism through which transactional leadership affect subordinates is using reward power to appeal to followers self-interest, thereby generating instrumental compliance from the followers (Yukl, 2010). Paternalistic Leadership Definition Paternalistic leadership derived from the notion of paternalism, but there is a lack of consensus on the definition of paternalism in the current literature (Aycan, 2006). Paternalism, as a cultural value prevalent in many parts of the world, is rooted in the indigenous traditional value of familism in Asian cultures such as China, Japan, Taiwan, Korea and India; however, in the 16

28 Western world such as the U.S. and European countries, paternalism originated from nineteenth century philanthropy, religious ideologies, and early industrialization (Aycan, 2006). Although researchers have not reached consensus on the definition of paternalistic leadership, they seem to agree that paternalistic leadership involves the coexistence of benevolence and authority (cf. Pellegrini & Scandura, 2008). In China, paternalism is rooted in the indigenous psychology of Confucianism and is congruent with Chinese collectivistic and high-power distance culture (Farh & Cheng, 2000). Developed in the Chinese cultural context, paternalistic leadership was defined by Farh and Cheng (2000, p. 91) as a style that combines strong discipline and authority with fatherly benevolence. Although the current literature demonstrates substantial disagreements regarding the extent to which benevolence should be displayed by a paternalistic leader, Pellegrini and Scandura s (2008) suggested that: In paternalistic leadership, the main focus is on employees welfare; a leader s care and protection are genuine, and employees show loyalty out of respect and appreciation for the leader s benevolence (Pellegrini & Scandura, 2008, p. 570). Overly emphasizing a leader s exploitation and control would turn this concept from paternalism toward authoritarianism (Aycan, 2006). In addition, Cheng and colleagues observed that the primary effects of paternalistic leadership come from benevolent and moral leadership (Cheng et al., 2004, p. 108) rather than authoritarian leadership. Therefore, the current research takes the view that benevolence and authority do coexist but the emphasis is on benevolence. Measurement of paternalistic leadership Similar to the definition of paternalistic leadership, the current literature does not have agreement on how to measure this construct (Pellegrini & Scandura, 2008). Multiple scales can be found in published research in this area. The various operationalizations of paternalism or 17

29 paternalistic leadership are perhaps a reflection of insufficient consensus on the conceptualization of this construct among researchers. Here I focus on two of the existing measures used by Cheng, Chou, and Farh (2000) and Pellegrini and Scandura (2006). In Cheng et al. s (2000) initial conceptualization, paternalistic leadership had three components: benevolence (holistic concern for subordinate s well-being), morality (demonstrating superior personal virtues), and authoritarianism (authority and control over subordinates). However, in later studies, Cheng and colleagues found that authoritarianism negatively correlated with the other two dimensions (e.g., Cheng et al., 2004). As a result, Farh, Cheng and colleagues suggested that paternalistic leadership was not very useful as an overall construct, and that the three scales for the three dimensions should be used separately (Farh et al., 2006). In agreement with Farh and colleagues (2006), Aycan (2006) also suggested that paternalistic leadership was not a monolithic construct and developed a 2 2 matrix depicting four different paternalistic leadership styles: benevolent paternalism, exploitative paternalism, authoritarian approach and authoritative approach. The current study intends to study paternalistic leadership as a whole and thus I turn to the 13-item measure used by Pellegrini and Scandura (2006). This measure was based on benevolent paternalism conceptualized by Aycan (2006). It is consistent with my view on paternalistic leadership in this dissertation. It is also consistent with Cheng et al. s (2004) observation that the primary effects of paternalistic leadership are not based on authoritarianism. In addition, it has been used in Turkey, India and North America and showed good reliability and validity (cf. Pellegrini & Scandura, 2008). Furthermore, this measure showed invariant factor structure across Indian and US samples (Pellegrini, Scandura, & Jayaraman, 2010). As a result, the present study uses this 13-item measure of paternalistic leadership. 18

30 How paternalistic leadership affect subordinates Although research has linked paternalistic leadership and some employee outcomes, such as higher employee flexibility, reduced cost, more control (Padavic & Earnest, 1994), lower turnover intentions (Cheng, Huang, & Chou, 2002), and organizational commitment (Farh et al., 2006), the mechanisms through which paternalistic leadership influences employees attitudes and behaviors have not been fully investigated. One possible mediating process between paternalistic leadership and subordinates outcomes might be social exchange. Farh and Cheng (2000) suggested that in the Chinese cultural context, because there is no compelling institutional force that requires the leader to act benevolently toward followers, benevolence on the part of the paternalistic leader generates indebtedness on the part of the employees, who will try to reciprocate in earnest. Empirical studies suggested that under benevolent leadership, followers could develop a genuine sense of gratitude and repayment to the leader (Cheng et al., 2004; Farh et al., 2006), which indicated a strong desire to reciprocate. In sum, the paternalistic leader takes care of the followers with fatherly benevolence in exchange for their felt indebtedness, personal loyalty, and obedience. Paternalistic leadership may also affect followers attitudes and behaviors through the mediating effect of identification with the leader. The benevolence and genuine care from the paternalistic leader are likely to engender a sense of being respected and valued by the leader on the part of the followers who are more likely to develop a sense of self-worth and self-esteem. As a result, followers are more likely to accept the leaders values and are more willing to imitate these values as if they were their own (Kark & Shamir, 2002). Therefore, followers are more likely to display positive attitudes and behaviors toward the leader. Empirical evidence showed that leader benevolence had a positive impact on Taiwanese followers identification with the 19

31 leader after three dimensions of transformational leadership (individualized consideration, high performance standard, and modeling) were controlled for (Cheng et al., 2004). Later, using a sample from mainland China, Farh et al. (2006) suggested that leader benevolence was positively associated with followers identification with the leader, which then led to compliance, satisfaction with the leader, and organizational commitment. Instrumental compliance may be another mediating process between paternalistic leadership and subordinates outcomes. Similar to the role that instrumental compliance plays for transactional leadership, paternalistic leadership can use authority to generate desirable attitudes, behaviors and performance on the part of the employees regardless how they truly feel. Empirically, leader benevolence and authoritarianism were found to be positively associated with followers compliance in China (Cheng et al., 2004; Farh et al., 2006). Furthermore, trust may also act as a mediating mechanism given the way paternalistic leadership is conceptualized in this study. Specifically, the benevolence displayed by the paternalistic leader may help form a close personal relationship with the followers, a relationship resembles those between family members. Such close personal relationship is likely to engender a high level of trust in the leader on the part of the followers. As a result, the followers may be willing to work hard toward the leader s goals. Justice perceptions are likely another mediator for the relationship between paternalistic leadership and employee outcome variables. The fatherly benevolence on the part of the leader signifies that the followers are valued members of the group with good standing and status. Consistent with the group value model (Lind & Tyler, 1988), followers tend to perceive higher procedural justice. The personal relationship between the leader and the follower and the individualized concern given by the leader may also show the leader s respect and propriety, 20

32 which often lead to perceived interpersonal justice (Bies & Moag, 1986; Colquitt, 2001). Given that paternalistic leadership has been found to have impact on followers attitudes, behaviors and performance, and that justice perceptions were also associated with employee outcomes (Colquitt, Conlon, Wesson, Porter, & Ng, 2001), organizational justice perceptions are likely one of the mediating mechanisms that go between paternalistic leadership and employee outcomes variables. However, the current literature has not adequately examined the mediating role of organizational justice in leadership research. Organizational Justice Definition In general, organizational justice can be understood as employees perceptions of fairness in organizations (Greenberg, 1987). Research in this area has focused primarily on three dimensions of organizational justice: distributive, procedural and interactional justice (cf. Greenberg & Colquitt, 2005). Distributive justice refers to the fairness of the allocation of outcomes in return for employee contributions (Adams, 1965). Individuals expect their outcomes to be commensurate with their inputs and consistent with the ratio of inputs and outcomes for selected comparison others (Adams, 1965). While equity-based allocation rules may be a natural extension of market-based economic systems, where economic value can often be associated with inputs and outcomes, need-based and equality-based distribution rules may be consistent with planned economies and socio-cultural systems that emphasize community welfare or reduction in social status differences (Deutsch, 1975). In addition to concerns about the distribution of outcomes, people also have concerns about the fairness of procedures used to allocate outcomes (Cropanzano & Greenberg, 1997). Procedural justice reflects the extent to which individuals perceive that they have voice 21

33 (influence) in procedures (Thibaut & Walker, 1975), and that procedures are consistent, unbiased, accurate, correctable, representative, and ethical (Leventhal, 1980). Procedural justice communicates an individual s status and standing in the group (Lind & Tyler, 1988) and it affects individuals satisfaction with and willingness to use procedures in the future, independent of outcome distributions (e.g., Tyler, 1994). Subsequent research expanded the two-factor (distributive/procedural) model by adding interactional justice, the extent to which social interactions are associated with perceptions of truthfulness and respectfulness of the information giver, justifications for actions taken, and propriety with which individuals are treated (Bies & Moag, 1986). Later, research investigated whether interactional justice is merely part of procedural justice or a separate construct (e.g., Bies & Shapiro, 1987; Tyler & Bies, 1990). Evidence showed that procedural justice was associated with individuals perceptions of and reactions toward the organization, while interactional justice was associated with perceptions of and reactions toward the supervisor (e.g., Masterson et al., 2000). Greenberg (1993) suggested that interactional justice itself might be multidimensional, with separate interpersonal and informational elements. Interpersonal justice involves respect and propriety elements, while informational justice involves justification and truthfulness rules (Bies & Moag, 1986; Shapiro, Buttner, & Barry, 1994). In addition to the recognition of various aspects of people s justice judgments, some researchers pointed out that employees consider different foci (e.g., agent vs. system) of their justice perceptions and tailor their behaviors accordingly (Colquitt et al., 2001). For example, Bies and Moag (1986) suggested that employees would focus more on their supervisor when evaluating interpersonal and informational justice but consider mainly the organization for 22

34 procedural justice judgments. From a social exchange perspective (Blau, 1964), employees form social exchange relationships with both the organization (Eisenberger, Huntington, Hutchison, & Sowa, 1986) and the leader (Graen & Scandura, 1987). The justice perceptions associated with each type of these social entities will affect how they reciprocate their obligations to them respectively. Organization-focused justice will lead to employees attitudes and behaviors directed at the organization whereas supervisor-focused justice perceptions engender employees reactions directed at the supervisor (Masterson et al., 2000) and also the organization (Rupp & Cropanzano, 2002). Later, Liao and Rupp (2005) found that supervisor-focused procedural and interpersonal justice climate had an impact on individual level attitudes and behaviors directed at the supervisor. Clearly, supervisor-focused justice is central in most organizations and the four aspects of the supervisor-focused justice may have significant implications on leadership. Measurement of organizational justice Along with the evolution of the conceptualization of the different dimensions of organizational justice, researchers have used a variety of measures for each dimension and the overall perception of organizational justice. Consistent with the four-factor conceptualization of organizational justice, Colquitt (2001) validated a four-factor measure that includes distributive justice, procedural justice, interpersonal justice and informational justice. The extant literature has not examined the four-factor justice model in leadership research, nor has it investigated the model in Chinese context. It is unclear how leader behaviors affect Chinese employees perceptions of the various aspects of justice and their reactions. How organizational justice affect employees A comprehensive review of the literature on how organizational justice affects employees is beyond the scope of this study. For more detailed review on this topic, please see Blader and 23

35 Tyler (2005), Cropanzano, Rupp, Mohler, and Schminke (2001), and Cropanzano, Byrne, Bobocel, and Rupp (2001). Instead, here I focus on the mechanisms through which justice perceptions affect employees attitudes and behaviors. In doing so, the commonalities between justice effect and leadership effect may be identified and integrated. Social exchange was found to be one of the major mediating processes between organizational justice and employee outcomes. For example, Masterson and colleagues found that leader-member-exchange mediated the relationship between interactional justice and supervisor-directed citizenship behaviors whereas perceived organizational support mediated that between procedural justice and organization-directed attitudes and behaviors (Masterson et al., 2000). In addition, Konovsky and Pugh (1994) suggested that trust as the basis for social exchange mediated the relationship between procedural justice and citizenship behaviors. In a multifoci social exchange model, Rupp and Cropanzano (2002) found that organization-focused and supervisor-focused social exchange mediated justice and employee performance. Social identification is another avenue for justice to affect employees outcomes. The relational model of authority proposed that inclusion in a group can provide a sense of self-worth and identity (Tyler & Lind, 1992). Fairness communicates one s status and value within a group (Tyler, Degoey, & Smith, 1996; Tyler, 1999). Therefore, when fairness perceptions signify that an employee is a valuable member of the group with good standing, he or she is likely to identify with the group to maintain a sense of self-worth. As a result, the employee is likely to be more willing to contribute to the group s goals and objectives. Essentially, justice perceptions can lead to employees attitudes, behaviors and performance (Colquitt et al., 2001) because fair treatment satisfies some important psychological needs (Cropanzano, Byrne, et al., 2001). The instrumental model suggested people have the need 24

36 to control their outcomes (Thibaut & Walker, 1975). Because justice satisfies people s need for control, they react to the extent that they perceive such need is met. People also have the need for belongingness. As has been pointed out earlier, justice implies one s inclusion in a social group, and his/her status and value in the group (Lind & Tyler, 1988; Tyler & Lind, 1992). Fairness indicates such important information about one s standing in the group. As a result, people will react to the different degrees of fairness that reflects the extent to which their need for belongingness is satisfied. Furthermore, people need to maintain meaningful existence. Folger (1998) argued that justice represented a moral virtue. People follow justice principles in keeping with morality in order to provide meaning to their lives (Becker, 1973). Therefore, people will react to justice to the degree that their need for meaningful existence is satisfied. Summary Several conclusions can be drawn from the above brief review of the literature. First, transformational leadership, transactional leadership, paternalistic leadership, and organizational justice seem to share some mechanisms in affecting employees attitudes and behaviors, such as social identification and social exchange. The extant literature on organizational justice has long recognized that the fairness of the treatment by the authorities has substantial impact on employees attitudes, behaviors, and performance (cf. Colquitt et al., 2001). However, to date, more effort is needed to integrate the fields of leadership and organizational justice (van Knippenberg & De Cremer, 2008; van Knippenberg et al., 2007). Second, so far, within the research related to transformational leadership, the focus appears to be the understanding how transformational and transactional leadership motivate and influence followers. Few studies have compared transformational and transactional leadership with other leadership approaches. In fact, frequently, transformational leadership is examined 25

37 alone. It is possible that a leader may need to engage in different leadership approaches in different situations in leading the same followers, yet the extant research is limited in informing us whether different leadership approaches have different patterns of impact on employees outcomes as well as which approach is more effective than others. Thus there is a need to compare and contrast in the same study multiple leadership styles that go beyond the transformational and transactional leadership paradigm. Third, although research that bridges leadership and organizational justice has been burgeoning recently (e.g., van Knippenberg & De Cremer, 2008), few studies have comprehensively investigated the four dimensions of organizational justice and their potential mediating role between leadership and employees outcomes. Pillai and colleagues (Pillai, Schriesheim, et al., 1999) showed that transformational leadership influenced only organizationfocused procedural justice whereas transactional leadership only impacted organization-focused distributive justice. Nonetheless, how interpersonal justice and informational justice can be influenced by different leadership approaches is still unknown. In addition, given the importance of leader fairness in affecting the followers, leader-focused organizational justice dimensions need to be examined in order to integrate leadership and organizational justice. Finally, there have been some cross-cultural studies on transformational, transactional, and paternalistic leadership in the Chinese context; specifically, employees cultural values such as collectivism and power distance (Hofstede, 1980) were used in examining cultures effect on leader effectiveness (e.g., Kirkman et al., 2009; Walumbwa & Lawler, 2003). However, although widely used, Hofstede s (1980) cultural dimensions generated considerable criticisms (e.g., McSweeny, 2002). Additionally, China has undergone significant socio-economic and political changes within a few decades and Chinese employees values have become more diverse 26

38 (Ralston, Holt, Terpstra, & Yu, 1997; Xu & Wang, 1991). Hofstede s dimensions may not adequately capture such dynamic changes in Chinese employees values, considering the fact that they were originally conceptualized to describe cultural differences at the country level (Hofstede, 1980). Therefore, researchers may need to investigate the effect of a different set of cultural value variables that can capture the diverse values held by Chinese employees in the workplace. In doing so, a more current and refined understanding about how different leadership affect Chinese employees can be developed. Based on the above review and discussion, a theoretical model can be developed (Figure 1). To address the aforementioned gaps in the extant literature, one focus of this model is the mediating role of the four aspects of supervisor-focused organizational justice in the relationships between transformational, transactional, and paternalistic leadership and followers in-role and extra-role performance. The other focus of this model is the moderating effect of two indigenously developed Chinese cultural values, traditionality and modernity (cf. Farh et al., 1997) on the relationship between leadership and supervisor-focused organizational justice. The next chapter will specify the hypothesized relationships depicted in the model. 27

39 Transformational Leadership Transactional Leadership Paternalistic Leadership Traditionality H7a & H9a H10 H12 H3a Supervisor- Focused PJ H3b H8& H9b H14 H11 H4a H4b H5a H5b Supervisor- Focused DJ H7b & H9c Supervisor- Focused IPJ OCB In-Role Performance H6a H7c & H9d H13 H15 H6b Supervisor- Focused IFJ Modernity Figure 1 Theoretical Model Note. Solid arrows depict positive relationships; dashed arrows represent negative relationships. H1a H2c H2a H2b H1b

40 CHAPTER 3: THEORY AND HYPOTHESES This chapter lays out hypotheses with regard to the relationships in the theoretical model (Figure 1). There are two main foci in this dissertation research. One is how supervisor-focused justice perceptions mediate the relationships between different leadership styles and employee attitudes and behaviors. The other is how Chinese employees cultural values can moderate the relationship between the different leadership approaches and supervisor-focused organizational justice. These two foci are associated with four sets of hypotheses: H1 and H2 establish the direct effect of leadership on followers performance, while H3 through H6 are about the relationships between leadership and justice perceptions. The mediating role of justice perceptions between leadership approaches and employee outcomes are specified in H7 through H9, and how the cultural values of traditionality and modernity moderate the effect of leadership approaches on justice perceptions are laid out in H10 through H16. The Direct Effect of Leadership on OCB Transformational leadership should have a positive direct effect on OCB. Transformational leadership is able to move subordinates beyond their self-interest to focus on what is good for the organization (Bass, 1985). A transformational leader often emphasizes the collective identity and group membership, inspires followers to buy into the desirable vision that the leader espouses, instills enthusiasm in employees, and provides meaning for what followers do in the organization (Bass & Avolio, 1994; Shamir et al., 1998). As a result, followers of transformational leaders tend to care more about the organizations wellbeing and they tend to be more willing to engage in extra-role behaviors such as OCB for the success and survival of the organization. Empirically, a positive link has been found between transformational leadership 29

41 and OCB in Chinese samples (Kirkman et al., 2009; Wang et al., 2005). Therefore, I expect that transformational leadership will be positively associated with OCB in China. Transactional leadership is unlikely to have a direct impact on OCB. The core notion of transactional leadership conceptualized as contingent reward was based on the leaders clarifying their expectations and providing rewards contingently in exchange for followers performance (Bass, 1985; Podsakoff et al., 1984). Such contractual exchange between the leader and the followers is unlikely to affect followers extra-role behaviors such as OCB because OCBs are discretionary behaviors that are often not explicitly stipulated by the authority (Organ, 1988) and thus less likely to be rewarded. Therefore, transactional leadership is unlikely to have direct influence on the followers OCBs. Paternalistic leadership s impact on OCB can be multifaceted. On the one hand, a paternalistic leader tries to create a family-like environment in the work place. He/she often knows about the followers personal problems and family life, consults followers on job matters, and gives advice to the followers on different matters as if he/she were an elder family member (Pellegrini & Scandura, 2006). In creating such a family ambiance, the leader s benevolence toward followers resembles that of an elder family member (father/mother, elder brother/sister); thus followers are more likely to move beyond their in-role behaviors and engage in extra-role activities to benefit the leader and the family-like larger group, similar to a child s effort to benefit the parents, elder siblings or the whole family. However, the paternalistic leader can be authoritarian. For example, he/she may believe that he/she is the only one who knows what is best for his/her followers and make decisions on behalf of them without asking for their approval (Pellegrini & Scandura, 2006). Such behaviors may discourage followers initiative and proactivity (Aycan et al., 2000). As a result, followers are less likely to engage in extra-role 30

42 behavior such as OCB. Overall, because the concept of paternalistic leadership used in the current study emphasizes the benevolence aspect of the construct, I predict that paternalistic leadership will have a positive but weak direct relationship with OCB. Hypothesis 1a: Transformational leadership has positive direct effect on followers OCB. Hypothesis 1b: Paternalistic leadership has positive direct effect on followers OCB. The Direct Effect of Leadership on In-role Performance In both Bass s (1985) and Podsakoff et al. s (1990) conceptualization, transactional leadership clarifies expectations toward followers performance and provides rewards to followers contingently on the level of their performance. Followers will be motivated to meet performance expectations and fulfill their end of the contract in order to be rewarded accordingly (Bass, 1985). A strong empirical support for the relationship between leaders contingent reward and followers in-role performance has been found (cf. Podsakoff, Bommer, Podsakoff, & MacKenzie, 2006). However, transformational leadership inspires followers with attractive vision, expresses optimism and high expectations for excellence and performance on the part of followers. It should be able to move followers beyond their normal level of performance (Bass, 1985). A positive relationship between transformational leadership and followers in-role performance has been found in both lab (Howell & Frost, 1989) and field (Bass, 1985) settings. Thus both transformational and transactional leadership are expected to have positive direct effect on followers in-role performance. Paternalistic leadership is very likely positively related to followers in-role performance. Paternalistic leaders take a personal interest in their followers lives and display genuine concern in protecting their followers personal wellbeing (Pellegrini & Scandura, 2006). In return, followers are likely to work hard to meet the performance expectations from the paternalistic 31

43 leader as a way to show their loyalty and deference toward the leader (Aycan, Kanungo, & Sinha, 1999). In addition, paternalistic leaders can put authoritarian demand on followers performance. Thus paternalistic leadership will likely have a positive direct relationship with followers in-role performance. To date, there have been few studies that simultaneously investigate the efficacy of different leadership approaches in eliciting followers in-role performance. Although transformational and transactional leadership have been found to have positive impact on followers performance, few empirical studies have been done on the direct effect of paternalistic leadership on followers in-role and extra-role performance (Pellegrini & Scandura, 2008). By juxtaposing the three different leadership styles, the current study has the potential to shed light on the direct effect of each of the leadership approaches on followers performance when controlling for the direct effects of the other two. Based on the aforementioned rationale, the following hypotheses can be expected. Hypothesis 2a: Transformational leadership has positive direct effect on followers in-role performance. Hypothesis 2b: Transactional leadership has positive direct effect on followers in-role performance. Hypothesis 2c: Paternalistic leadership has positive direct effect on followers in-role performance. The Direct Effect of Leadership on Supervisor-Focused Procedural Justice Transformational leaders have been found to be able to enhance followers perceptions of organization-focused procedural justice (Kirkman et al., 2009; Pillai, Scandura, et al., 1999; Pillai, Schriesheim, et al., 1999). Since the supervisor is often viewed as the agent that represents 32

44 the organization, the general procedural justice perceptions are likely to seep into followers judgments about procedural justice associated with the supervisor. More importantly, transformational leaders often encourage followers to question the status quo, give followers individualized consideration and delegate decision making to the followers (Bass, 1985; Podsakoff et al., 1990). Followers will likely perceive more control in important work related decisions. Also, when followers are allowed to question the status quo and their assumptions regarding their work, and when their concerns are taken into consideration by the leader, they tend to perceive certain attributes that usually are associated with procedural justice such as correctability, bias suppression, accuracy, etc. (Leventhal, 1980) As a result, transformational leaders are more likely perceived as procedurally fair by the followers. In business organizations where productivity and performance are often emphasized, equity rule for allocation is naturally aligned with leaders behaviors in allocating outcomes, which are consistent with contingent reward. In line with previous research (e.g., Podsakoff et al., 1990), contingent reward is the focal transactional leadership behavior of this study. Because contingent reward is mainly concerned about making sure that followers outcome is contingent on their performance, its implication on supervisor-focused procedural justice is unclear in the form it is operationalized. Consistent with Pillai et al. (1999), transactional leadership is unlikely to be associated with supervisor-focused procedural justice. As Pellegrini and Scandura noted, In paternalistic leadership, the main focus is on employees welfare; a leader s care and protection are genuine, and employees show loyalty out of respect and appreciation for the leader s benevolence (Pellegrini & Scandura, 2008, p. 570). Paternalistic leaders often assume the role of parents and consider it an obligation to protect 33

45 followers, who in turn, reciprocate the care and protection from the paternalistic leader with loyalty, deference and compliance (Aycan, 2006). The relationship between paternalistic leadership and supervisor-focused procedural justice is complex. On the one hand, a paternalistic leader strives to create a family environment in the workplace, in which followers may feel that they have more freedom to voice their opinions on work-related matters. And they also consult followers on job matters, which affords followers more control (Thibaut & Walker, 1975). Thus followers will perceive procedural justice. In addition, the fact that followers are often treated as family members by paternalistic leaders may communicate that the followers are valued members of the group with good standing. According to the group value model (Lind & Tyler, 1988), followers will perceive procedural justice. To the extent that control and group value are seen as afforded by the paternalistic leader, followers will perceived the leader as procedurally fair. However, on the other hand, paternalistic leaders can be authoritarian. A paternalistic leader often believes that he/she is the only one who knows what is best for the followers and makes decisions on behalf of the followers without asking for their approval (Pellegrini & Scandura, 2006). In doing so, paternalistic leaders may deprive followers of voice and control in work-related matters. Such leader behaviors in turn will likely reduce followers procedural justice perceptions. Nonetheless, the notion of paternalistic leadership used in the current study emphasizes the leaders benevolence more than their authoritarianism (Pellegrini & Scandura, 2008). Additionally, given the high power distance nature of the Chinese culture (Hofstede, 1980), followers are more used to paternalistic leaders being authoritarian (Brockner et al., 2001). Therefore, on balance, paternalistic leadership is likely to be positively associated with supervisor-focused procedural justice. 34

46 Hypothesis 3a: Transformational leadership has positive direct effect on supervisor-focused procedural justice Hypothesis 3b: Paternalistic leadership has positive direct effect on supervisor-focused procedural justice. The Direct Effect of Leadership on Supervisor-focused Distributive Justice Among multiple allocation rules, equity is the basis upon which distributive justice is evaluated (cf. Colquitt, 2001). It is appropriate for this study to use equity instead of other allocation rules because, this study investigates leadership and justice in Chinese business organizations where productivity and performance are typically emphasized. Equity is the prevalent allocation rule in such type of organizations (e.g., Deutsch, 1975). Transformational leadership is unlikely to be associated with supervisor-focused distributive justice. In the current formulation of transformational leadership, the focus is on the leaders providing a vision and role model, advocating group goals, expressing high performance expectations, providing individualized support and intellectually stimulating the followers. The implications of such leader behaviors on equity-based allocation are unclear. Transactional leaders use contractual processes in which they clarify their expectations, appeal to followers self-interest, and use reinforcement in exchange for followers performance (Bass, 1985). In order to generate performance that meets expectations, transactional leaders make rewards contingent on followers work outcomes. Such notion of contingent reward is aligned with the equity principle of distributive justice (Adams, 1965). The relationship between transactional leadership and distributive justice was empirically supported (Pillai, Schriesheim, et al., 1999). Thus, to the extent that transactional leaders enact contingent reward in accordance with equity principle, followers will perceive supervisor-focused distributive justice. 35

47 Paternalistic leadership is expected to be negatively associated with supervisor-focused distributive justice. Paternalistic leaders genuinely care about subordinates welfare and the relationships between them and subordinates tend to be personalized because they often consider personal factors (Pellegrini & Scandura, 2008; Redding, 1990) in their decision making and are often involved in many aspects of subordinates personal lives. In the family-like ambiance created by paternalistic leaders, interpersonal harmony becomes more important in the leaders decision making. As a result, allocation may be more based on need or equality rather than equity, which tend to facilitate individual productivity (Deutsch, 1975; Leung & Park, 1986). In addition, equity rules have not yet been well defined nor well institutionalized in allocation practices in Chinese organizations (e.g., Ding & Warner, 2001; Leung & Kwong, 2003). Thus, paternalistic leadership is likely negatively associated supervisor-focused distributive justice. Hypothesis 4a: Transactional leadership has positive direct effect on supervisor-focused distributive justice. Hypothesis 4b: Paternalistic leadership has negative direct effect on supervisor-focused distributive justice. The Direct Effect of Leadership on Supervisor-focused Interpersonal Justice Transformational leaders tend to foster high quality leader-member exchange (Wang et al., 2005). They often provide individualized support to followers through respecting followers personal feelings and needs (Podsakoff et al., 1990). Such leader behaviors are consistent with interpersonal justice principles that emphasize treating employees with respect and dignity (Bies & Moag, 1986; Colquitt, 2001). So, transformational leadership will be positively related to supervisor-focused interpersonal justice (Cho & Dansereau, 2010). 36

48 Transactional leadership is mainly concerned with outcome allocation that is contingent on followers performance (Bass, 1985; Podsakoff et al., 1990). It uses reinforcement in exchange for followers work outcomes. Whether the transactional leader treats followers with respect and dignity is not specified or implied from the definition and operationalization of transactional leadership. Thus it is unlikely related to supervisor-focused interpersonal justice. The relationship between paternalistic leadership and supervisor-focused interpersonal justice can be complex. On the one hand, paternalistic leaders care about followers wellbeing and treat them like family members. Their concern toward followers may be displayed with respect, similar to the respect a parent displays toward a child. Therefore the followers are likely to perceive interpersonal justice. On the other hand, paternalistic leaders benevolence can be delivered in an authoritarian way and they don t have to show respect for the followers (cf. Pellegrini & Scandura, 2008). Therefore, a negative relationship between paternalistic leadership and supervisor-focused interpersonal justice can be expected. Overall, because the notion of paternalistic leadership used in this study emphasize leaders benevolence rather than their authoritarianism (Pellegrini & Scandura, 2008), paternalistic leadership should be positively associated with supervisor-focused interpersonal justice. Hypothesis 5a: Transformational leadership has positive direct effect on supervisor-focused interpersonal justice. Hypothesis 5b: Paternalistic leadership has positive direct effect on supervisor-focused interpersonal justice. The Direct Effect of Leadership on Supervisor-focused Informational Justice Transformational leaders tend to articulate important values and emphasize the purpose, commitment and ethical consequences of decisions (Bass, 1985; Bass & Steidlmeier, 1999). 37

49 Such behaviors are aligned with the notion of informational justice that involves providing adequate justification for allocation decisions and communicating candidly with employees (Bies & Moag, 1986; Colquitt, 2001). Thus, followers are likely to perceive supervisor-focused informational justice under transformational leaders. Transactional leadership uses reinforcement in exchange for followers work outcomes and it was often operationalized as contingent reward (e.g., Podsakoff et al., 1990), and it was unclear whether transactional leaders provide adequate justification and explanation about their allocation decisions. Thus, the relationship between transactional leadership and supervisorfocused informational justice is not hypothesized in this study. Paternalistic leadership is likely to be negatively associated with supervisor-focused informational justice. Paternalistic leaders often treat followers as children and make decisions without consulting them. They tend to feel less obligation in providing adequate justifications and explanations. Since it is assumed that paternalistic leaders will look out for subordinates best interest, candid communication between paternalistic leaders and followers would seem unnecessary. These characteristics of paternalistic leadership are not in alignment with the principles of informational justice (Colquitt, 2001). Therefore, when judged against the relevant rules, paternalistic leaders will likely be perceived by followers as low on informational justice. Hypothesis 6a: Transformational leadership has positive direct effect on supervisor-focused informational justice. Hypothesis 6b: Paternalistic leadership has negative direct effect on supervisor-focused informational justice. 38

50 Mediation Hypotheses Next, I provide the rationale for the mediating role played by the four aspects of supervisor-focused justice perceptions (Figure 1) between the three leadership approaches and the dependent variables. Transformational leadership s indirect effect on OCB and in-role performance may be explained through supervisor-focused procedural, interpersonal and informational justice. This is because transformational leadership behaviors will positively affect followers evaluations of the three aspects of justice associated with the supervisor, which in turn lead to higher quality of social exchange between the leader and the followers (Masterson et al., 2000; Rupp & Cropanzano, 2002), and higher trust in the leader (Konovsky & Pugh, 1994; Pillai, Schriesheim, et al., 1999; Podsakoff et al., 1990). As a result, followers are willing to display higher level of in-role and extra-role performance as a way to reciprocate. According to Moorman and Byrne (2005), (a) fair treatment could be considered a benefit to be reciprocated; and (b) justice perceptions can offer employees evidence that it is appropriate to be in a social exchange relationship with the supervisor or organization. Empirically, various aspects of organizational justice perceptions were found to be positively associated with employees OCB and in-role performance (cf. Colquitt et al., 2001). Therefore, the following relationships can be expected. Hypothesis 7a-1: Transformational leadership has indirect effect on followers OCB through supervisor-focused procedural justice. Hypothesis 7a-2: Transformational leadership has indirect effect on followers in-role performance through supervisor-focused procedural justice. Hypothesis 7b-1: Transformational leadership has indirect effect on followers OCB through supervisor-focused interpersonal justice. 39

51 Hypothesis 7b-2: Transformational leadership has indirect effect on followers in-role performance through supervisor-focused interpersonal justice. Hypothesis 7c-1: Transformational leadership has indirect effect on followers OCB through supervisor-focused informational justice. Hypothesis 7c-2: Transformational leadership has indirect effect on followers in-role performance through supervisor-focused informational justice. The indirect effect of transactional leadership on the outcome variables can be explained via supervisor-focused distributive justice. Leaders contingent reward behaviors are aligned with equity principle, which is the rule used in evaluating supervisor-focused distributive justice (Colquitt, 2001). In contrast to procedural justice, distributive justice is associated with economic exchange and transactional contracts (Konovsky & Pugh, 1994). Followers fairness perceptions about the outcomes allocated by the leader can facilitate the economic exchange between the leader and the followers. As a result, the followers will likely work harder to deliver higher inrole performance in exchange for higher rewards. However, according to Moorman and Byrne (2005), it is social exchange rather than economic exchange that facilitates employees OCBs. Empirically, distributive justice was not found to be related to OCBs (e.g., Konovsky & Pugh, 1994; Pillai, Schriesheim, et al., 1999). Thus, supervisor-focused distributive justice is unlikely to mediate the relationship between transactional leadership and OCB. Hypothesis 8: Transactional leadership has indirect effect on followers in-role performance through supervisor-focused distributive justice. The indirect relationship between paternalistic leadership and followers OCB and in-role performance is multifaceted. On the one hand, paternalistic leadership is likely to enhance followers perception of supervisor-focused procedural and interpersonal justice. On the other 40

52 hand, paternalistic leadership may inhibit followers perceptions of supervisor-focused distributive and informational justice. As mentioned earlier, fairness perceptions facilitate trust building and high quality social exchange between the followers and the organization (Moorman & Byrne, 2005), which in turn lead to followers reciprocation with higher in-role performance and OCB. Therefore, the indirect effect of paternalistic leadership on outcome variables via supervisor-focused procedural and interpersonal justice will be positive whereas that through supervisor-focused distributive and informational justice will be negative. Hypothesis 9a-1: Paternalistic leadership has indirect effect on followers OCB through supervisor-focused procedural justice. Hypothesis 9a-2: Paternalistic leadership has indirect effect on followers in-role performance through supervisor-focused procedural justice. Hypothesis 9b: Paternalistic leadership has indirect effect on followers in-role performance through supervisor-focused distributive justice. Hypothesis 9c-1: Paternalistic leadership has indirect effect on followers OCB through supervisor-focused interpersonal justice. Hypothesis 9c-2: Paternalistic leadership has indirect effect on followers in-role performance through supervisor-focused interpersonal justice. Hypothesis 9d-1: Paternalistic leadership has indirect effect on followers OCB through supervisor-focused informational justice. Hypothesis 9d-2: Paternalistic leadership has indirect effect on followers in-role performance through supervisor-focused informational justice. 41

53 Moderation Hypotheses In this section, the potential boundary effect of Chinese cultural values is explicitly considered. Specifically, I hypothesize that Chinese employees cultural values of traditionality and modernity will moderate the relationships between leadership behaviors and supervisorfocused justice perceptions. Confucianism is one important root of Chinese culture which has a strong emphasis on maintaining the five cardinal hierarchical relationships (wu lun) in work and family (Hofstede, 1980). The cultural value of traditionality, a construct developed in Chinese context, focuses on expressive ties among people and values such as respect for authority, filial piety, ancestor worship, male-domination, fatalism and a general sense of powerlessness (Farh et al., 1997, p. 424). It was operationalized with an emphasis on respect for authority, which is consistent with Confucianism. Modernity, another cultural value indigenous to Chinese societies, on the other hand, reflects the orientation toward egalitarianism and open-mindedness, optimism and assertiveness, affective hedonism, sexual equality, and self-reliance (Farh et al., 1997, p. 424), and it was operationalized with a focus on egalitarianism and open-mindedness. The individual followers cultural values of traditionality and modernity are chosen in this study for several reasons. First, traditionality and modernity involve differential views about authority and hierarchy. Since hierarchy is also embedded in the dyadic leader-follower relationship, traditionality and modernity may have differential impact on leadership effectiveness. Second, traditionality and modernity are not the two ends of a unidimensional construct, and it is possible for an individual to be both traditional and modern at the same time. This is particularly interesting because China has undergone fast profound transition from the traditional hierarchical tightly controlled society to a much more modernized society with more 42

54 autonomy afforded to entrepreneurs and individuals (Walder, 2000). During this transition, both traditional Chinese values such as respect for authority and modern Western values such as egalitarianism and open-mindedness may coexist in Chinese employees and have important implications for leadership practices. For example, younger generations of Chinese employees have more exposure to modern Western business, educational and entertainment material through Internet, TV and other media than older generations of Chinese employees. They may have developed modern values such as egalitarianism and open-mindedness. However, at the same time, all Chinese employees are embedded in Chinese traditional culture that emphasizes hierarchical authority. As a result, some Chinese employees may possess strong values of traditionality and modernity while other Chinese employees values on one or both of them could be weak. Thus investigating how these values may affect Chinese employees fairness perceptions about the leader will give us more refined understanding of the effectiveness of various leadership approaches. Finally, consistent with the level of analysis of this study, traditionality and modernity as cultural values have been validated at individual level of analysis. Other cultural values such as Hofstede s (1980) cultural dimensions, although have been measured at individual level (e.g., Brockner et al., 2001), were originally developed for societal level of analysis. Chinese employees with strong traditionality values are inclined to show higher level of respect for the leader and see their relationship with the leader as covenantal. Traditional Chinese employees tend to expect the leader to be more authoritarian and less participative. Because transformational leaders often delegate and empower followers, traditional Chinese subordinates are likely to perceive transformational leadership actions, although fair, as out of the expected realm of typical leader behaviors. However, because traditional Chinese employees have high 43

55 respect for authority, they tend to be more accepting for leaders behaviors and perceive a wider range of leader behaviors as fair. Therefore, although the leaders may engage in more or less transformational leadership behaviors, traditional Chinese employees will perceive similar level of supervisor-focused procedural, interpersonal and informational justice. On the other hand, Chinese employees with low traditionality have less respect for the authority and they will evaluate the leader with less leniency than traditional Chinese. When the leader engages in less transformational behaviors, less traditional Chinese employees will perceive a significantly lower level of supervisor-focused procedural, interpersonal and informational justice than traditional Chinese employees because of the lower respect they have for authority. As a result, transformational leadership will have a stronger effect on perceived supervisor-focused procedural, interpersonal and informational justice among followers who are low rather than high on traditionality. This is consistent with Kirkman et al. s (2009) finding that individual power distance orientation negatively moderated the relationship between group level transformational leadership and individual level procedural justice perceptions. Hypothesis 10a: Traditionality moderates the relationship between transformational leadership and supervisor-focused procedural justice. The relationship is stronger among Chinese employees with low rather than high traditionality. Hypothesis 10b: Traditionality moderates the relationship between transformational leadership and supervisor-focused interpersonal justice. The relationship is stronger among Chinese employees with low rather than high traditionality. Hypothesis 10c: Traditionality moderates the relationship between transformational leadership and supervisor-focused informational justice. The relationship is stronger among Chinese employees with low rather than high traditionality. 44

56 In a similar vein, because traditional Chinese tend to view a wider range of leader behaviors as fair, although the leader may allocate rewards contingently or non-contingently on followers performance, traditional Chinese employees will perceive similar level of supervisorfocused distributive justice. However, less traditional Chinese employees will have less tolerance for the leader s non-contingent reward behavior and perceive much less fairness than traditional Chinese employees. Hypothesis 11: Traditionality moderates the relationship between transactional leadership and supervisor-focused distributive justice. The relationship is stronger among Chinese employees with low rather than high traditionality. The same logic for Hypotheses 10 and 11 applies when the interaction between paternalistic leadership and traditionality is concerned. Although paternalistic leadership behaviors may deviate from justice principles, the traditional Chinese employees will give the leader more leeway and thus their perception of supervisor-focused procedural, distributive, interpersonal and informational justice may not fluctuate dramatically. However, the same violation of justice principles will engender a much more negative perception in less traditional Chinese employees because of the low reverence they have toward authority. Hypothesis 12a: Traditionality moderates the relationship between paternalistic leadership and supervisor-focused procedural justice. The relationship is stronger among Chinese employees with low rather than high traditionality. Hypothesis 12b: Traditionality moderates the relationship between paternalistic leadership and supervisor-focused distributive justice. The relationship is stronger among Chinese employees with low rather than high traditionality. 45

57 Hypothesis 12c: Traditionality moderates the relationship between paternalistic leadership and supervisor-focused interpersonal justice. The relationship is stronger among Chinese employees with low rather than high traditionality. Hypothesis 12d: Traditionality moderates the relationship between paternalistic leadership and supervisor-focused informational justice. The relationship is stronger among Chinese employees with low rather than high traditionality. Chinese employees may also hold the cultural value of modernity considering the modernization in China in the past few decades. Those employees who have a strong value of modernity will expect equal treatment between the powerful and powerless, appreciate opportunities for voice, and embrace independence and open-mindedness. Chinese employees who have a low value of modernity tend to see the relationship between them and the authority as covenantal and based on societal expectations of roles (Farh et al., 1997). Therefore, whether or not the leaders behavior is fair is less of a question for less modern Chinese employees. As a result, their perceptions of the supervisor-focused procedural, distributive, interpersonal, and informational justice may not vary to a great extent when transformational, transactional, and paternalistic leadership behaviors deviate from justice principles. On the other hand, Chinese employees high on modernity will likely see the fairness of the leader s behavior as an indicator of whether they are afforded with the respect and status they deserve as individuals in the group. As a result, the fairness of the leader s behaviors is of more importance for modern Chinese employees than it is for less modern Chinese employees. Therefore, modern Chinese employees will react with much less positive perceptions of supervisor-focused procedural, interpersonal, and informational justice when the leader s behaviors move away from justice principles. Thus, 46

58 the relationships between leadership behaviors and followers justice perceptions will be more pronounced among Chinese employees high rather than low on modernity. Hypothesis 13a: Modernity moderates the relationship between transformational leadership and supervisor-focused procedural justice. The relationship is stronger among Chinese employees with high rather than low modernity. Hypothesis 13b: Modernity moderates the relationship between transformational leadership and supervisor-focused interpersonal justice. The relationship is stronger among Chinese employees with high rather than low modernity. Hypothesis 13c: Modernity moderates the relationship between transformational leadership and supervisor-focused informational justice. The relationship is stronger among Chinese employees with high rather than low modernity. Hypothesis 14: Modernity moderates the relationship between transactional leadership and supervisor-focused distributive justice. The relationship is stronger among Chinese employees with high rather than low modernity. Hypothesis 15a: Modernity moderates the relationship between paternalistic leadership and supervisor-focused procedural justice. The relationship is stronger among Chinese employees with high rather than low modernity. Hypothesis 15b: Modernity moderates the relationship between paternalistic leadership and supervisor-focused distributive justice. The relationship is stronger among Chinese employees with high rather than low modernity. Hypothesis 15c: Modernity moderates the relationship between paternalistic leadership and supervisor-focused interpersonal justice. The relationship is stronger among Chinese employees with high rather than low modernity. 47

59 Hypothesis 15d: Modernity moderates the relationship between paternalistic leadership and supervisor-focused informational justice. The relationship is stronger among Chinese employees with high rather than low modernity. Although I propose that traditionality and modernity have opposing moderating effect on the relationship between the three leadership approaches and the supervisor-focused justice perceptions, because traditionality and modernity are not the two ends of the same continuum, it is meaningful to investigate how Chinese employees with different combinations of the two cultural values react to the three different leadership approaches. From the above moderation analysis, it seems that high traditionality and low modernity will have an inhibiting effect on followers perceptions of justice whereas low traditionality and high modernity will amplify leadership s influence on follower s justice perceptions. Therefore, I predict that, among Chinese employees who are high on modernity but low on traditionality, the relationships between leadership behaviors and followers justice perceptions will be the strongest. On the other hand, among Chinese employees who are low on modernity but high on traditionality, the same relationships will be the weakest. For those Chinese employees who are high on both, the effect of high traditionality may mitigate the effect of high modernity on the relationships between leadership behaviors and followers justice perceptions. Similarly, among the Chinese employees who are low on both values, the effect of low modernity may mitigate the effect of low traditionality on these relationships. Thus the relationships between the focal leadership behaviors and the followers justice perceptions will be of medium strengths among the Chinese employees who are high on both traditionality and modernity or low on both values. Hypothesis 16: The strengths of the relationships between leadership and justice are the strongest among Chinese employees who are high on modernity but low on traditionality. 48

60 The strengths of the relationships between leadership and justice are the weakest among Chinese employees who are low on modernity but high on traditionality. Among those who are high on both or low on both, the strengths of the relationships between leadership and justice are in the middle. 49

61 CHAPTER 4: METHODS In this chapter, the methods employed in this study are laid out to address the research questions and test hypotheses. Design This dissertation research collected cross-sectional data through questionnaires distributed to Chinese part-time MBA and EMBA students and their supervisors across more than 25 industries and business organizations. Given that past research has found that transformational and transactional leadership effectiveness are contingent on the type of the organization (Lowe, Kroeck, & Sivasubramaniam, 1996), a cross-sectional design provides confidence for external validity especially when the data is collected from various types of companies and industries in China. Although all participants in this study are students, they all have sufficient work experience and their responses are based on their observations in real work setting rather than artificially created leadership experience. In order to reduce common source bias, this study collected dependent variable data from leaders while independent variable, mediator and moderator data were collected via their subordinates self-report responses. Sample and Procedure Step 1: Subordinate Survey The targeted subordinate subjects were part time MBA students and executive MBA students enrolled at a major public university located in a major metropolitan area in Southwest China. All MBA/EMBA student participants must be 18 or older and must have worked under the current supervisor for at least four months in order to be eligible for participation. The subordinate (MBA/EMBA student) participants were recruited in a classroom setting. However, they could choose to complete the Subordinate Questionnaire at locations of their choice after 50

62 class. Then, they returned the completed Subordinate Questionnaire in a sealed envelope to a collection box located in a corner of the classroom. The return envelope was provided by the investigator. By completing and returning the Subordinate Questionnaire, the MBA/EMBA students explicitly indicated their consent for participating in this research and for their responses to be used in this study. All participation was voluntary. In total, 700 subordinate questionnaires were distributed and 414 were returned, resulting in a response rate of 59%. Among the 414 supervisors who received the supervisor questionnaire, 366 of them filled out the survey and returned them, resulting in a response rate of 88%. After 17 unusable questionnaires were excluded from the sample due to large number of unanswered questions, the sample consists of matched questionnaires filled out by 349 supervisorsubordinate dyads. Fifty percent of the subordinate participants are female. The percentage of the subordinates in different age groups are: between 18 and 30 (59.5%); between 31 and 35 (19%); between 36 and 40 (11.7%); between 41 and 45 (6.1%); between 46 and 50 (3.5%); between 51 and 55 (.3%). Sixty-one percent of the subordinates are regular employees; first line managers take up 15.8%; mid-level managers 14%; and 6.9% of the subordinates are executives. Most of the subordinates (92%) have received at least a four-year undergraduate college education and 39.5% of them work in publicly traded companies. Sixty-five percent of the subordinates work in companies in which state ownership is dominant (above 51%), 30% in organizations dominated by private ownership and 5% are from firms dominated by foreign ownership. The average job tenure and average time working with the supervisor are 3.1 years and 2.4 years respectively. 51

63 Step 2: Supervisor Survey The subordinate participants passed a sealed envelope to their direct supervisors that contained an invitation letter, consent information for supervisor, Supervisor Questionnaire, and pre-addressed, stamped envelope. Supervisors who met the following two criteria were eligible to participate in this study. First, the supervisor participant had to be 18 or older and have supervised the subordinate for at least 4 months. Second, the subordinate should have completed the Subordinate Questionnaire and returned it in Step 1. Eligible supervisor participants read the invitation and supervisor consent information. Participating supervisor sent the completed survey back directly to the investigator via regular mail in the provided pre-addressed envelope. In order to match the responses for each subordinate-supervisor dyad, each of the two surveys was marked with a pre-assigned, 5-digit identification number located in the upper right corner of the survey. However, no other identifying information was collected. There was no master list for the purpose of matching the questionnaires to individuals, and the investigator had no way of knowing who participated in this study. Thus complete anonymity was ensured. Thirty-two percent of the supervisor participants are female. The percentage of the subordinates in different age groups are: 30 and younger (18.8%); between 31 and 35 (24.9%); between 36 and 40 (27.9%); between 41 and 45 (17.6%); between 46 and 50 (7.0%); between 51 and 55 (3.2%); between 56 and 60 (.3%) and above 60 (.3%). The rank of the supervisors varies from low level employees (3%), first line supervisors (32%), mid-level managers (39%), and executives (26%). Seventy-six percent of the supervisors have received either undergraduate or graduate college degrees. The average job tenure is 5 years. It should be noted that subordinate participants included both part time MBA students who tended to work at lower levels within their organizations and EMBA students who were 52

64 more likely to have higher level positions. This implies that their supervisors also occupied different levels of positions in the organization. There are several reasons why it is appropriate to include leaders at different organizational levels. First, transformational and transactional leadership could be used by leaders at different levels (Bass, 1999). The same is true for paternalistic leadership. Therefore, in order to reflect such reality and ensure the generalizability of findings across organizational levels, supervisors of both MBA and EMBA students should be included. Second, the effectiveness of transformational and transactional leadership varies across leaders rank in organizations (Judge & Piccolo, 2004). This source of variation can be controlled for by including supervisors at different position levels. However, to use student type, (i.e., MBA vs. EMBA) as a control variable may not be the optimal way to control for the variation due to leaders organizational rank. This is because, in the EMBA program of the Chinese university, a significant number of EMBA students do not necessarily occupy executive positions. Many of them can be mid-level managers. Therefore, level of position may be a better control variable than student type. Measures Subordinate Questionnaire contained measures of nine constructs. All items included in this survey were followed by a five-point Likert scale anchored by 1= strongly disagree and 5= strongly agree. The English version of the questionnaire was developed first. Then the questionnaire was translated into Chinese using translation-back-translation techniques to ensure the meaning of the Chinese version was equivalent to that of the English version (Brislin, 1980). There are two questionnaires in this study, one for the subordinate (Table 1) and the other for the supervisor (Table 2). Table 1 Subordinate Questionnaire Transformational Leadership (Podsakoff et al., 1990) 53

65 Articulating a vision 1. My supervisor has a clear understanding of where we are going. 2. My supervisor paints an interesting picture of the future for our group. 3. My supervisor is always seeking new opportunities for the organization. 4. My supervisor inspires others with his/her plans for the future. 5. My supervisor is able to get others committed to his/her dream. Providing an appropriate model 6. My supervisor leads by doing, rather than simply by telling. 7. My supervisor provides a good model for me to follow. 8. My supervisor leads by example. Fostering the acceptance of group goals 9. My supervisor fosters collaboration among work groups. 10. My supervisor encourages employees to be team players. 11. My supervisor gets the group to work together for the same goal. 12. My supervisor develops a team attitude and spirit among employees. High performance expectations 13. My supervisor shows us that he/she expects a lot from us. 14. My supervisor insists on only the best performance. 15. My supervisor will not settle for second best. Providing individualized support 16. My supervisor acts without considering my feelings. (R) 17. My supervisor shows respect for my personal feelings. 18. My supervisor behaves in a manner thoughtful of my personal needs. 19. My supervisor treats me without considering my personal feelings. (R) Intellectual stimulation 20. My supervisor challenges me to think about old problems in new ways. 21. My supervisor asks questions that prompt me to think. 22. My supervisor has stimulated me to rethink the way I do things. 23. My supervisor has ideas that have challenged me to reexamine some basic assumptions about my work. Transactional Leadership (Podsakoff et al., 1990) 1. My supervisor always gives me positive feedback when I perform well. 2. My supervisor gives me special recognition when my work is very good. 3. My supervisor commends me when I do a better than average job. 4. My supervisor personally compliments me when I do outstanding work. 5. My supervisor frequently does not acknowledge my good performance. (R) Paternalistic Leadership (Pellegrini & Scandura, 2006) 1. My supervisor is interested in every aspect of his/her employees lives. 2. My supervisor creates a family environment in the workplace. 3. My supervisor consults his/her employees on job matters. 4. My supervisor is like an elder family member (father/mother, elder brother/sister) for his/her employees. 5. My supervisor gives advice to his/her employees on different matters as if he/she were an elder family member. 6. My supervisor makes decisions on behalf of his/her employees without asking for their approval. 7. My supervisor knows each of his/her employees intimately (e.g., personal problems, family life, etc.). 8. My supervisor exhibits emotional reactions in his/her relations with the employees; doesn t refrain from showing emotions such as joy, grief, anger. 9. My supervisor participates in his/her employees special days (e.g., wedding, funerals, etc.). 54

66 10. My supervisor tries his/her best to find a way for the company to help his/her employees whenever they need help on issues outside work (e.g., setting up home, paying for children s tuition). 11. My supervisor expects his/her employees to be devoted and loyal, in return for the attention and concern he/she shows them. 12. My supervisor gives his/her employees a chance to develop themselves when they display low performance. 13. My supervisor believes he/she is the only one who knows what is best for his/her employees. Supervisor-Focused Procedural Justice (adapted from Colquitt, 2001) 1. I have been able to express my views and feelings during the procedures used by my supervisor to determine my outcomes. 2. I have been able to voice my opinion about the procedures used by my supervisor to arrive at my outcomes. 3. Have the procedures that your supervisor used to allocate your outcomes been applied consistently? 4. Have the procedures used by your supervisor to allocate your outcomes been applied without bias? 5. Have the procedures that your supervisor used to your outcomes been based on accurate information? 6. Would you be able to appeal the outcomes arrived at by these procedures that your supervisor used? 7. Have the procedures that your supervisor used to allocate your outcomes upheld ethical and moral standards among your coworkers? Supervisor-focused Distributive Justice (adapted from Colquitt, 2001) 1. Do your outcomes allocated by your supervisor reflect the effort you have put into your work? 2. Are your outcomes allocated by your supervisor appropriate for the work you have completed? 3. Do your outcomes allocated by your supervisor reflect what you have contributed to your organization? 4. Based on your performance, are your outcomes allocated by your supervisor justified? Supervisor-Focused Interpersonal Justice (adapted from Colquitt, 2001) 1. My supervisor who was responsible for the procedures used to allocate my outcomes treated me in a polite manner. 2. My supervisor who was responsible for the procedures used to allocate my outcomes treated me with dignity. 3. My supervisor who was responsible for the procedures used to allocate outcomes treated me with respect. 4. My supervisor who was responsible for the procedures used to allocate my outcomes refrains from making improper remarks or comments. Supervisor-Focused Informational Justice (adapted from Colquitt, 2001) 1. My supervisor who was responsible for the procedures used to allocate my outcomes has been candid in his/her communication about the allocation with me. 2. My supervisor who was responsible for the procedures used to allocate my outcomes explained the allocation thoroughly. 3. My supervisor who was responsible for the procedures used to allocate my outcomes gave reasonable explanations about the allocation. 4. My supervisor who was responsible for the procedures used to allocate my outcomes communicated details about the allocation in a timely manner. 5. My supervisor who was responsible for the procedures used to allocate my outcomes seemed to tailor his/her communication about the allocation to my specific needs? Traditionality (Farh et al., 1997) 1. The chief government official is like the head of a house-hold. The citizen should obey his decisions on all state matters. 55

67 2. The best way to avoid mistakes is to follow the instructions of senior persons. 3. Before marriage, a woman should subordinate herself to her father. After marriage, to her husband. 4. When people are in dispute, they should ask the most senior person to decide who is right. 5. Those who are respected by parents should be respected by their children. Modernity (Farh et al., 1997) 1. If the chief government official makes a mistake, the citizen may criticize him openly. 2. If the teacher makes a mistake, the student can argue with him by reason. 3. People who seek political reforms should have the right to make a speech in public places. 4. To pursue advanced study or better employment opportunity, it is okay for someone to leave his home and family. 5. The sexual scenes in a TV program should not be censored if they are an integral part of the script. The Supervisor Questionnaire contained measures of two constructs. All measurement items in this survey were followed by a five-point Likert scale anchored by 1= strongly disagree and 5= strongly agree. The English version of the questionnaire was developed first. Then the questionnaire was translated into Chinese using translation-back-translation techniques to ensure the equivalency of meaning (Brislin, 1980). Table 2 Supervisor Questionnaire Organizational Citizenship Behavior (Farh et al., 1997) Identification with the company 1. This subordinate is willing to stand up to protect the reputation of the company. 2. This subordinate is eager to tell outsiders good news about the company and clarify their misunderstandings. 3. This subordinate makes constructive suggestions that can improve the operation of the company. 4. This subordinate actively attends company meetings. Altruism toward colleagues 5. This subordinate is willing to assist new colleagues to adjust to the work environment. 6. This subordinate is willing to help colleagues solve work-related problems. 7. This subordinate is willing to cover work assignments for colleagues when needed. 8. This subordinate is willing to coordinate and communicate with colleagues. Conscientiousness 9. This subordinate complies with company rules and procedures even when nobody watches and no evidence can be traced. 10. This subordinate takes one s job seriously and rarely makes mistakes. 11. This subordinate does not mind taking on new or challenging assignments. 12. This subordinate tries hard to self-study to increase the quality of work outputs. 13. This subordinate often arrives early and starts to work immediately. Interpersonal harmony 14. This subordinate uses illicit tactics to seek personal influence and gain with harmful effect on interpersonal harmony in the organization (R). 15. This subordinate uses position power to pursue selfish personal gain (R). 56

68 16. This subordinate takes credits, avoids blames, and fights fiercely for personal gain (R). 17. This subordinate often speaks ill of the supervisor or colleagues behind their backs (R) Protecting company resources 18. This subordinate conducts personal business on company time (e.g., trading stocks, shopping, going to barber shops). (R) 19. This subordinate uses company resources to do personal business (company phones, copy machines, computers and cars). (R) 20. This subordinate views sick leave as benefit and makes excuse for taking sick leave. (R) In-Role Performance (Rupp & Cropanzano, 2002) 1. This subordinate adequately completes assigned duties. 2. This subordinate fulfills responsibilities specified in job description. 3. This subordinate performs tasks that are expected of him/her. 4. This subordinate meets formal performance requirements of the job. 5. This subordinate engages in activities that will directly affect his/her performance evaluation. 6. This subordinate neglects aspects of the job he/she is obligated to perform (R). 7. This subordinate fails to perform essential duties (R). Control Variables Consistent with past research on transformational and transactional leadership, several demographic variables were considered as potential control variables. Those regarding the subordinates are: age, length of time working with the supervisor, level of organizational position (employee, low-level manager, mid-level manager, executive-level manager), and education (high school and under, associate college degree, bachelor s degree, master s degree, and Ph.D.). The variables collected from the supervisors include: age, gender, job tenure (in years), level of organizational position (employee, low-level manager, mid-level manager, executive-level manager), education (high school and under, associate college degree, bachelor s degree, master s degree, and Ph.D.), state ownership of the firm (percentage of stock shares owned by the government), and private ownership of the firm (percentage of stock shares owned by private persons). 57

69 Analyses Data analyses were conducted with structural equation modeling (SEM) using AMOS (Arbuckle & Wothke, 1999). The SEM approach was chosen for at least two reasons. First, SEM can account for measurement error and thus it estimates the relationships between latent variables more accurately. Second, because past research has documented relatively high correlations between some of the leadership variables included this research, such as the correlation between transformational and transactional leadership (Judge & Piccolo, 2004), a regression approach may not be optimal for data analysis because it assumes that the predictor variables are independent from each other. In SEM, by allowing the independent variables to covary, the high correlations among them can be statistically controlled for. Thus the relationships estimated in the model should be more valid. Several model fit indices were chosen in the SEM analyses, including Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), Root Mean Square Error of Approximation (RMSEA), and χ 2 /df value. CFI and TLI values of 0.90 or higher and RMSEA values between.08 and.10 indicate acceptable fit; RMSEA values between.05 and.08 indicate good model fit, and RMSEA values lower than.05 indicate excellent fit (Browne & Cudeck, 1993; Hu & Bentler, 1999). Data Treatment Attention was given to the amount and pattern of missing data because incomplete data can seriously bias the conclusions drawn from this empirical study. There are no clear guidelines on how to handle missing data. Kline (1998) suggested that missing data should be less than 10% of the data. In the current study, because the missing data are 7.6%, mean substitution was used 58

70 for imputing missing data (Donner, 1982). A normality check shows that all continuous variables are normally distributed. Measurement Model Following Anderson and Gerbing (1988), a two-step approach was employed in data analyses. First, confirmatory factor analysis (CFA) was conducted to analyze construct validity. Because transformational leadership and OCB are used as holistic constructs, second order CFA was used to test whether the first order factors of transformational leadership and OCB indeed form coherent second order factors. After the unidimensionality was established for each construct, a CFA was performed with all the 11 latent constructs used in this study. Considering the exceptionally large number of indicators (98) associated with the 11 constructs, parceling (Bagozzi & Edwards, 1998) was used to reduce the number of indicators in the 11-factor CFA model. Alternative CFA models were also tested to examine discriminant validity of the constructs. First, a 4-factor model was investigated. In this model, all the leadership parcels were forced to load on one latent construct; all the justice parcels were forced to load on one latent factor; all the parcels for the outcome variables were forced to load on one latent construct; and all the parcels for the two cultural value constructs were force to load on one latent factor. Second, in a 2-factor model all the constructs rated by the subordinates were merged into one factor and all the constructs rated by the supervisors were treated as driven by one latent factor. Finally, a one-factor model in which all the constructs were merged into one latent factor was tested. 59

71 Structural Model and Hypotheses Testing Two steps were taken to test the structural model and the hypotheses. In step 1, the mediation model was tested. In step 2, the moderation effects of the two cultural values were examined. When testing for mediation, especially when multiple mediators are present, recent publications have recommended a bootstrapping approach (Hayes, 2009; Preacher & Hayes, 2008). The bootstrapping approach involves repeatedly sampling from the data set with replacement and estimating the indirect effect in each bootstrap sample. After thousands of times of estimation of the indirect effect in the bootstrapping samples, confidence intervals can be derived based on the empirical approximation of the distribution of the indirect effect. According to Preacher and Hayes (2008), this approach emphasizes the direction and size of the indirect effect instead of the statistical significance of the paths between the independent variables, mediators, and dependent variables. In addition, the bootstrapping approach is able to tease apart the specific indirect effect of each mediator while controlling for the other mediators. Considering the fact that the supervisor-focused justice factors were significantly correlated with each other as reported in past research (e.g., Liao & Rupp, 2005), it is theoretically meaningful to examine each of the four justice mediators unique ability to mediate the indirect effect of leadership on the followers performance in the presence of the other three justice mediators. Therefore, the bootstrapping approach (Preacher & Hayes, 2008) was used to test the proposed theoretical mediation model (Figure 1). In Step 2, the test for moderation, because the mediated effect of the leadership on employee performance will differ across employee groups with different values, a median split approach was used to create groups of employees with different levels of traditionality and 60

72 modernity; and multi-group analysis was conducted to test moderation effect across employee groups (Byrne, 2001; Hayes, 2009). A factor score was generated respectively for the constructs of traditionality and modernity after their construct validity was examined. Factor scores higher than the median were coded as 1 and those lower than the median were coded as 0 for both traditionality and modernity. Along traditionality, the subsample with high traditionality has a sample size of 175 and the one with low traditionality has 174 cases. Similarly, the subsample size with high modernity is 175 and that associated with low modernity is 174. In testing how different levels of traditionality moderate the leadership-justice relationships (H10-H12), first, the mediation model as the result of Step 1 was tested simultaneously across the two groups (high traditionality & low traditionality) with all the paths freely estimated. Then, the paths between leadership and the mediators were constrained to be equal and the model was once again tested across the two groups. Chi-square difference was used in determining whether the paths were invariant across the two groups. The same approach was used in testing whether modernity moderates the paths between leadership and the justice mediators (H13-H15). To test H16, the main sample was split into four subsamples of nearly equal size. The subsample with all the cases that have both traditionality and modernity coded as 1 represents the group of Chinese employees who are high on both traditionality and modernity. The other three subsamples represent Chinese employees who are high on one of the two cultural values and low on both respectively. Then the mediation model as the result of Step 1 was tested using multi-group analysis. 61

73 CHAPTER 5: RESULTS Measurement Model Confirmatory Factor Analysis was used to examine construct validity (Table 3). The 23- item transformational leadership measure showed a poor fit (χ 2 /df = 5.21, CFI=.88, TLI=.86, RMSEA=.11). The factor loadings of two items under the factor Providing Individualized Support were very low (.37 and.33, standardized). After they were excluded from the analysis, model fit improved significantly (χ 2 /df = 3.02, CFI=.95, TLI=.94, RMSEA=.08). Then I proceeded to examine the 21-item transformational leadership as a holistic concept. A secondorder CFA model was not admissible and the factor Providing Individualized Support had a very low factor loading (.14) on transformational leadership as the second-order construct. Therefore, the last two items under this factor were removed and second-order CFA was conducted on the remaining 19-item (5-factor) construct. Results show that the 5-factor transformational leadership concept does form an overall second-order construct (χ 2 /df = 3.29, CFI=.95, TLI=.94, RMSEA=.08). The reliability (Cronbach s α) of the 19-item scale was.97. Similar CFA analyses were conducted on organizational citizenship behavior. The 5- factor OCB model showed a reasonable fit (χ 2 /df = 2.82, CFI=.92, TLI=.91, RMSEA=.07). Second-order CFA, however, showed a poor fit (χ 2 /df =3.84, CFI=.87, TLI=.86, RMSEA=.09). Modification Indices suggested that the residuals of Interpersonal Harmony and Protecting Company Resources were highly correlated. Considering the fact that all the items under these two factors were negatively worded, it is very likely that the wording pattern may have caused the correlation between the residuals. After the residuals were allowed to correlate, the secondorder CFA model showed a reasonable fit (χ 2 /df = 2.86, CFI=.92, TLI=.91, RMSEA=.07) 62

74 indicating that the 5-factor OCB constructs indeed form a coherent overall construct. Internal consistency of the 20-item OCB scale was.92. Table 3 Results of Confirmatory Factor Analyses Models χ 2 /df CFI TLI RMSEA α Transformational Leadership 2 nd Order Organizational Citizenship Behavior 2 nd Order Transactional Leadership Paternalistic Leadership Supervisor-focused Organizational Justice In-role Performance Traditionality & Modernity Three leadership approaches together (3-factor) N/A Three leadership approaches together (1-factor) N/A 11-Factor Measurement Model N/A 4-Factor Alternative Measurement Model N/A 2-Factor Alternative Measurement Model N/A 1-Factor Alternative Measurement Model N/A The 5-item transactional leadership measure showed a good fit (χ 2 /df = 3.27, CFI=.99, TLI=.98, RMSEA=.08). However, the reversed coded item had a very low factor loading (.25). When this item was excluded from the model, the CFA showed reasonable fit (χ 2 /df = 4.91, CFI=.99, TLI=.98, RMSEA=.10) and all factor loadings were above.80. The reliability of the 4-item scale was.93. Next, CFA was conducted on the one-factor paternalistic leadership scale. The 13-item measure did not fit the data well (χ 2 /df = 5.47, CFI=.88, TLI=.85, RMSEA=.11). The factor loading of one item (My manager makes decisions on behalf of his/her employees without asking for their approval) was non-significant (.04, p =.47). The factor loading on another item (My manager believes he/she is the only one who knows what is best for his/her employees.) was.35, lower than the commonly used threshold of.40. The model fit improved significantly to an acceptable level (χ 2 /df = 4.70, CFI=.93, TLI=.91, RMSEA=.10). The reliability of the 11-item scale was

75 The 4-factor CFA model of organizational justice fit the data reasonably well (χ 2 /df = 3.00, CFI=.94, TLI=.93, RMSEA=.08) and the 20-item scale showed a reliability of.95. The reliabilities for Procedural Justice, Distributive Justice, Interpersonal Justice and Informational Justice are.86,.92,.91, and.91 respectively. In-role performance did not fit the data well (χ 2 /df = 7.81, CFI=.89, TLI=.83, RMSEA=.14). One item s (This subordinate engages in activities that will directly affect his/her performance evaluation.) factor loading was very low (.17; p<.01). The remaining 6-item measure still did not fit well with the data (χ 2 /df = 9.90, CFI=.90, TLI=.84, RMSEA=.16). The Modification Indices suggested that the error terms of the two negatively worded items should be correlated. It is likely that the wording pattern may have caused higher correlation between the error terms. After these two errors were allowed to correlate, model fit was significantly improved (χ 2 /df = 1.90, CFI=.99, TLI=.98, RMSEA=.05). The reliability of the 6-item scale was.84. The CFA of the two cultural values of traditionality and modernity produced a model that fit the data reasonably well (χ 2 /df = 3.03, CFI=.92, TLI=.89, RMSEA=.08). Cronbach s α for each of the 5-item scales of Traditionality and Modernity was.73. In order to examine whether the three leadership constructs are distinct from each other, the validity of the three leadership scales was tested simultaneously in one CFA model. Transformational leadership (5-factor model tested earlier) was treated as a second-order factor while transactional leadership and paternalistic leadership were treated as first order factors in the model. The resulting CFA model showed a good fit with the data (χ 2 /df = 2.55, CFI=.93, TLI=.92, RMSEA=.07). When all the indicators in the three leadership scales were forced on 64

76 one single latent construct, results showed that this single-factor leadership CFA model did not fit the data (χ 2 /df = 6.11, CFI=.75, TLI=.74, RMSEA=.12). In order to examine the discriminant validity of the constructs of interest, they need to be tested in a series of CFA models. Given the large number of indicators involved in these constructs, the limited sample size, and unidimensionality established for each construct, parcels (Bagozzi & Edwards, 1998) were used to reduce the number of indicators for each construct. Each of the five dimensions of transformational leadership was used as an indicator. OCB was also represented by five parcels, each representing one of the five dimensions of this construct. The four items of transactional leadership were aggregated into two parcels. For each of the four justice constructs, two parcels were formed. Traditionality and modernity each have two parcels. The 6 items of in-role performance were aggregated into three parcels. Table 4 shows how the parcels were formed for each latent construct. Table 4 Parcels for Latent Constructs Parcel 1 Parcel 2 Parcel 3 Parcel 4 Parcel 5 TFL Items 1-5 Items 6-8 Items 9-12 Items Items TSL Items 1-2 Items 3-4 PTL Items 1, 3, 9, 10 Items 4, 5, 11, 12 Items 2, 7, 8 PJ Items 1, 2, 6 Items 3, 4, 5, 7 DJ Items 1-2 Items 3-4 IPJ Items 1-2 Items 3-4 IFJ Items 1-2 Items 3-5 PFMS Items 1-2 Items 3-4 Items 6-7 OCB Items 1-4 Items 5-8 Items 9-13 Items Items TRAD Items 1-2 Items 3-5 MOD Items 1, 3, 5 Items 2, 4 Note. TFL= transformational leadership; TSL= transactional leadership; PTL= paternalistic leadership; PJ= supervisor-focused procedural justice; DJ= supervisor-focused distributive justice; IPJ= supervisor-focused interpersonal justice; IFJ= supervisor-focused informational justice; PFMS= in-role performance; OCB=organizational citizenship behaviors; TRAD= traditionality; MOD= modernity. Next, using the parcels specified above, the 11-factor measurement model was tested in one CFA model (Figure 2). Results showed that it fit the data reasonably well (χ 2 /df = 2.62, CFI=.93, TLI=.91, RMSEA=.07). Then, a few alternative measurement models were also 65

77 tested. First, a 4-factor model was examined. In this model, all the leadership parcels were forced to load on one latent construct; all the justice parcels were forced to load on one latent factor; all the parcels for the outcome variables were forced to load on one latent construct; and all the parcels for the two cultural value constructs were force to load on one construct. This four-factor model did not fit the data (χ 2 /df = 4.55, CFI=.82, TLI=.80, RMSEA=.10). Second, a 2-factor Figure 2 Hypothesized 11-Factor Model model, in which all the constructs rated by the subordinates were merged into one factor and all the constructs rated by the supervisors were treated as driven by one latent factor, fit the data poorly (χ 2 /df = 6.33, CFI=.72, TLI=.70, RMSEA=.12). Finally, a one-factor model in which all 66