Nationality Heterogeneity and Interpersonal Relationships at (PDF)

CHRISTIAN TRÖSTER Nationality Heterogeneity and Interpersonal Relationships at Work CHRISTIAN TRÖSTER - Nationality Heterogeneity and Interpersonal Relationships at Work ERIM PhD Series Research in Management Erasmus Research Institute of Management - E R I M 233 E R I M Design & layout: B&T Ontwerp en advies (www.b-en-t.nl) Print: Haveka (www.haveka.nl) NATIONALITY HETEROGENEITY AND INTERPERSONAL RELATIONSHIPS AT WORK In this dissertation I test three new approaches to extend the ‘classical’ model of work - place diversity. The ‘classical’ model of workplace diversity assumes that diversity affects work outcomes via the mediating effects of social networks. I hypothesize that this model fruitfully can be extended by 1) considering that diversity forms a context in which employees act, 2) testing alternative predictors of network formation and employee behavior (i.e., employee voice), and 3) integrating diversity and social network perspectives in a contingency model. Three empirical studies support these hypotheses. In the first study, I show that the association between leadership and employee voice is stronger for nationality dissimilar employees. The second study finds that employee voice affects the strength of friendship relations but that this effect is contingent on employees’ past position in the social network. Finally, the third study demonstrates that group performance is maximized at moderate levels of task network centralization but lowest at high and low levels of centralization but that this relation is moderated by nationality diversity. Nationality diverse teams required more centralization to achieve high performance than homoge - neous teams. Finally, I discuss the implications of these findings for research on diversity and social networks. ERIM The Erasmus Research Institute of Management (ERIM) is the Research School (Onder - zoek school) in the field of management of the Erasmus University Rotterdam. The founding participants of ERIM are Rotterdam School of Management (RSM), and the Erasmus School of Econo mics (ESE). ERIM was founded in 1999 and is officially accre dited by the Royal Netherlands Academy of Arts and Sciences (KNAW). The research under taken by ERIM is focused on the management of the firm in its environment, its intra- and interfirm relations, and its busi ness processes in their interdependent connections. The objective of ERIM is to carry out first rate research in manage ment, and to offer an ad vanced doctoral pro gramme in Research in Management. Within ERIM, over three hundred senior researchers and PhD candidates are active in the different research pro - grammes. From a variety of acade mic backgrounds and expertises, the ERIM commu nity is united in striving for excellence and working at the fore front of creating new business knowledge. Erasmus Research Institute of Management - ERIM Rotterdam School of Management (RSM) Erasmus School of Economics (ESE) P.O. Box 1738, 3000 DR Rotterdam The Netherlands Tel. +31 10 408 11 82 Fax +31 10 408 96 40 E-mail [email protected] Internet www.erim.eur.nl B&T11161_ERIM_omslag Tröster_14mrt11  Nationality Heterogeneity and Interpersonal Relationships at Work   Nationality Heterogeneity and Interpersonal Relationships at Work Nationaliteiten heterogeniteit en interpersoonlijke relaties op de werkplek Thesis to obtain the degree of Doctor from the Erasmus University Rotterdam by command of the rector magnificus Prof.dr. H.G. Schmidt and in accordance with the decision of the Doctorate Board The public defence shall be held on Friday, 13th May 2011 at .9.30hrs by Christian Tröster Essen, Germany  Doctoral Committee Promoter: Prof.dr.D.L. van Knippenberg Other members: Dr. D. van Dierendonck Dr. G. Jakobs-Belschak Dr. A. Mehra Erasmus Research Institute of Management – ERIM The joint research institute of the Rotterdam School of Management (RSM) and the Erasmus School of Economics (ESE) at the Erasmus University Rotterdam Internet: http://www.erim.eur.nl ERIM Electronic Series Portal: http://hdl.handle.net/1765/1 ERIM PhD Series in Research in Management, 233 ERIM reference number: EPS-2011-233-ORG © 2011, Christian Tröster Design: B&T Ontwerp en advies www.b-en-t.nl This publication (cover and interior) is printed by haveka.nl on recycled paper, Revive®. The ink used is produced from renewable resources and alcohol free fountain solution. Certifications for the paper and the printing production process: Recycle, EU Flower, FSC, ISO14001. More info: http://www.haveka.nl/greening All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system, without permission in writing from the author.  1 Table of Contents SUMMARY .................................................................................................................................................. 2 CHAPTER 1: INTRODUCTION ........................................................................................................................ 5 CHAPTER 2: LEADERSHIP, NATIONALITY, DISSIMILARITY, AND VOICE IN MANAGEMENT TEAMS ................. 13 THEORY AND HYPOTHESES ...................................................................................................................................... 14 METHODS ........................................................................................................................................................... 19 DISCUSSION ......................................................................................................................................................... 33 CONCLUSION ....................................................................................................................................................... 38 CHAPTER 3: EMPLOYEE VOICE AND SOCIAL NETWORKS ............................................................................. 39 THEORETICAL BACKGROUND AND HYPOTHESES.......................................................................................................... 40 METHODS ........................................................................................................................................................... 45 DISCUSSION ......................................................................................................................................................... 55 CONCLUSION ....................................................................................................................................................... 59 CHAPTER 4: TASK NETWORK CENTRALIZATION, NATIONALITY DIVERSITY, AND TEAM PERFORMANCE ........ 61 THEORETICAL BACKGROUND AND HYPOTHESES.......................................................................................................... 62 METHODS ........................................................................................................................................................... 69 DISCUSSION ......................................................................................................................................................... 76 CONCLUSION ....................................................................................................................................................... 80 CHAPTER 5: GENERAL DISCUSSION ............................................................................................................ 83 REFERENCES ............................................................................................................................................. 90 ABOUT THE AUTHOR ............................................................................................................................... 113  2 SUMMARY In this dissertation I argue and test three new extensions of what I refer to as the ‘classical’ model of workplace diversity. This model assumes that diversity affects work outcomes via the mediating effect on social networks. I show that this model can be fruitfully extended via 1) considering diversity a contextual variable rather than a main effect 2) considering alternative predictors of network formation and in particular employee behavior (i.e. employee voice), and 3) integrating the diversity perspective and the social network perspective in a contingency model. Three empirical studies support these extensions. In chapter two I investigated the leader directed voice behavior (i.e. challenging the status quo with constructive ideas for change) of managers who were dissimilar to their coworkers in terms of nationality. We hypothesized that employee nationality dissimilarity will strengthen the relationships between leader openness and leader nationality similarity with supervisor-directed voice. Results showed that leader openness and nationality similarity to the leader were more strongly related to supervisor-directed voice of nationality dissimilar managers than for more nationality similar managers. Nationality dissimilar managers were more likely to speak up to their leaders when they had the same nationality as their leaders and when leaders were open for ideas. We also explored two psychological mechanisms (i.e. psychological safety and affective commitment). Affective commitment mediated the interaction of leader openness and nationality dissimilarity but psychological safety the interaction of leader similarity and nationality dissimilarity. The second empirical study (Chapter 3) tested the argument that the relationship between a focal employee’s voice behaviors on a coworker’s felt friendship strength for this employee is contingent on the structural holes (i.e. missing links in a network) that disconnect the focal employee’s friends from each other. The hypothesis was supported in a longitudinal study in the health care sector. We found that employee voice was related to decreases in friendship strength if employees had highly redundant networks (few structural holes) and increased friendship strength if redundancy was low (many structural holes). Finally, the fourth chapter tested the argument that the influence of team task  3 network structure on team performance is contingent on the nationality diversity of team members. A study of five-person teams revealed that the relationship between network centralization (in the task/workflow network) and team performance was curvilinear. Performance was higher at moderate levels of centralization than at low and high levels of centralization. This relationship was moderated by nationality diversity. Nationality diverse teams required more centralization to achieve high performance than homogeneous teams.  4  5 CHAPTER 1: INTRODUCTION Understanding the antecedents of attitudes and behaviors of individuals and teams is crucial if companies want to effectively manage their employees’ efforts and increase organizational effectiveness. One perspective on team and individual outcomes that has gained a lot of attention is organizational diversity (Pfeffer, 1982, 1983). Diversity can be conceptualized as the distribution of differences among team members with regard to some attribute (Harrison & Klein, 2007). The central premise of the diversity perspective is that diversity in organizational members’ backgrounds leverages organizational performance because it increases the available pool of experiences and tacit knowledge in organizations. Team members from diverse backgrounds are supposed to challenge each other’s taken-for-granted assumptions, stimulate debate, introduce a greater range of solutions to problems, and engender more careful considerations of information and take better decisions (van Knippenberg, De Dreu, & Homan, 1994; van Knippenberg, & Schippers, 2007). In the following I will discuss the importance of nationality diversity, introduce the various extensions of this dissertation to what I refer to as the ‘classical’ model of diversity, and provide a summary of the empirical studies in this book. Nationality diversity Of the many attributes that could be considered, I focused in this book on nationality. Not only do individuals from different countries bring different ideas and ways of thinking to a team, they also bring with them different beliefs about how best to organize for task performance (e.g., Hofstede, 2001; Kirkman & Shapiro, 1997; Laurent, 1986). Nationality is relevant for work teams because it is a super-ordinate determinant of one’s identity that is engrained from birth (Earley & Mosakowski, 2000; Gibson & Gibbs, 2006; Hofstede, 2001) and the generic effects of diversity will be most prominent when team members hold different nationalities (Stahl, Maznevski, Voigt, & Jonsen, 2010). Individuals from different countries bring different ideas and ways of thinking to a team (e.g., Kirkman & Shapiro, 1997; Laurent, 1986) and how to best prioritize and interpret the team’s problems  6 (Earley, 2006). From a sense making perspective, variety within a team is an important prerequisite for registering the variety in a team’s environment (Weick, 1979). Thus, nationality diversity can be a perquisite for team members’ efforts to uncover their differences. Nationality diversity is particularly relevant in companies who execute global operations in a competitive international business (Jackson & Alvarez, 1992) because their nationality diverse employees form a pool of experiences and knowledge with respect to companies’ international business environments (Jackson, Joshi, & Erhardt, 2003). Thus, a focus on nationality diversity seems promising because of its salience in organizations and the increasing diversity of the work force in times of globalization. Diversity as context Despite almost five decades of research no consistent main effects of diversity on team performance have been established (Williams & O’Reilly, 1998; van Knippenberg & Schippers, 2007). In fact, diversity has rarely had positive effects and often was unrelated to or negatively impacted work outcomes. In this dissertation I argue that we can gain new insights into the effects of diversity at the workplace by extending the ‘classical’ model of diversity. In its most basic form this model assumes that diversity has main effects on work outcomes through its effect on social networks (Pfeffer, 1982, 1983). Diversity affects social networks because people tend to associate with similar others (i.e. homophily) and these social networks, in turn, influence performance because they facilitate or constrain the flow of resources and are the prisms through which identities are formed. I argue that a first limitation of this perspective is that it neglects possible contingency affects of diversity. Indeed, researchers have pointed out the promise of departing from the ‘classical’ main effect model of diversity and shift the focus on contingencies of diversity (van Knippenberg, DeDreu, Homan, 2004; Kearney & Gebert, 2009). In this study we will investigate diversity as context of teams (Study 1 and Study 3). Context is where other teams, departments and organizations live (Hackman, 1999). Context provides the purpose, resources, social cues, norms, and meanings that shape behavior in teams (Jackson, Joshi, & Erhard, 2003). Whereas there is some research that shows that the effect of diversity differs with contexts (e.g. organizational culture) (see Jackson, Joshi, & Erhard, 2003 for a  7 review) there is little research that has considered that diversity itself forms a context that affects the salience of social cues (i.e. leadership in Study 1) and a team’s need for coordination (i.e. task network centralization in Study 3). Hence, a first contribution of this dissertation is to investigate the role of diversity as a contextual variable. This way I hope to facilitate a fluid conversation from main effect models of diversity to diversity as contextual factor. The diversity-social networks relationship revisited Whereas the diversity perspective focuses on individual employee attributes a complementary but different perspective on individual and team outcomes in organizations has focused on the interpersonal relationships between people at work and in particular the structure of their relationships. This social network perspective argues that relationships are important because they are the pipes through which resources flow and the prisms through which perceptions are formed (Podolny, 2001). Importantly, however, the diversity literature (Reagans, Zuckerman, & McEvily, 2004; Williams and O’Reilly, 1998 for reviews) has tended to assume that these alternative approaches are interchangeable and that priority should be given to the diversity perspective (Lawrence, 1997). Combining the network perspective and the diversity perspective was motivated by the argument that the effect of demographic composition on performance is mediated by intervening processes such as network-based dynamics. Since demographic variables are easier to “access and reliably measure” and networks are influenced by demography researchers focused on demographics rather than measuring networks (Pfeffer, 1982: 351; cf. Tsui and Gutek, 1999). But this focus on demographics has created a ‘black box’ that has limited our understanding of why diversity has all kinds of effects. In fact, there are good reasons to doubt that diversity can be that easily linked to social networks. First, Lawrence (1997) has questioned whether organizational networks are characterized by a significant amount of homophily. She reviewed literature that indicates that the observed levels of homophily are generally weak because a variety of contextual factors can reduce the degree of homophily inside an organization or team. For example, past experience in interacting with members of different social categories (Westphal & Milton, 2000) or  8 organization culture can fosters the creation of non-homphilious relationships (Chatman et al., 1998; Ely and Thomas, 2001; Polzer, Milton, and Swann, 2002). And generally, homophily varies with the degree to which group members identify with their social category (Mehra, Kilduff, & Brass, 1998). As the level of identification declines so does the tendency to form homophilious relationships. Thus, whereas there is strong support for the homophily argument there is reason to believe that its effects on networks are contingent on a range of other factors and that the strength of this effects might at times be weak. In addition to Lawrence’s (1997) critique Reagans, Zuckerman, and McEvily (2004) formulated and tested a second objection. They argued that diversity affects networks in opposing ways. First, diversity might decrease a group’s cohesion as team members find it more attractive to interact with similar rather than dissimilar others. In this view, demographic diversity reduces internal coordination, which hinders a team’s ability to succeed. On the other hand, diverse group members should have contacts into different social circles of people with whom they share some significant (demographic) characteristics. These connections potentially reach beyond the boundaries of the team or department and generate “information benefits” for the team because they connect team members with groups that are likely to have non-redundant information and resources (Burt 1992, Granovetter, 1973). In their study Reagans, Zuckerman, and McEvily (2004) found support for both hypotheses. Diverse teams were less cohesive, which decreased performance but they also had a broader range of external connections, which increased performance. According to these findings it seems difficult to exclusively focus on diversity and neglect social networks because diversity had ambiguous effects on social networks and hence performance A third shortcoming of the diversity-network relationships is that diversity describes the creation of homopilious relationships. According to the diversity perspective employees form relations by choosing from the menu of attractive interaction partners offered in a given context (McPherson, Popielarz, & Drobnic, 1992). In fact, a popular perspective on diversity assumes that people will form ties with those with whom they entered the organization (i.e. tenure, Pfeffer, 1982) but does not argue  9 how these associations might change over time. But even in groups without compositional changes one can observe changes in relationships that thus cannot be explained solely by the invariable characteristics of group members (Moody, McFarland, Bender-DeMoll, 2002). The diversity perspective has prospered during a time when research has gravitated to the view that social networks are relatively stable. But this apparent stability does not account for change and adjustment in the ties underlying the network (e.g., Moody, 2002; Kossinets & Watts, 2006, Kilduff & Krackhardt, 2008). For example, a recent study showed that while some people had stable relationships others “dance between friends” throughout time (Moody, McFarland, Bender-DeMoll, 2005: 1208). Thus, diversity might have limited capabilities to explain network change because it assumes that actors “merely choose from the menu of associations presented by the social structure surrounding them” (McPherson, Popielarz, & Drobnic, 1992: 168) but begs the question how “intentional, creative human action serves in part to constitute those very social networks that so powerfully constrain actors in turn” (Emirbayer & Goodwin, 1994: 1413). Thus, despite the strong evidence for homophily researchers might profit from investigating other, non-stable actor attributes and consider the active role that people can play in shaping their networks. Overall, the previous discussion concludes that the diversity-network linkage might be weak, ambiguous, and neglects the importance of human agency. Since networks are supposed to be the more proximate causes for work outcomes the next logical step in the study of intra-organizational processes should be to explore alternative explanations for the creation of social networks. In this dissertation I will highlight the importance of past networks and human agency (i.e. employee voice) for changes in relationship strength. Thus, a second contribution of this dissertation will be to tap into new antecedents of social networks and enhance our understanding of individual attributes that shape networks beyond the effects of homophily. Integrating social networks and diversity perspective: A contingency approach The literature on the antecedents of team performance has, at least until recently, tended to ignore the possible effects of the structure of a team’s task network. But it is equally true that the  10 literature on team networks has tended to ignore the possibility that the effects of networks on teams may be contingent on the individual attributes of its members. Perhaps this mutual neglect stems from the fact that these perspectives are rooted in the very different intellectual lineages of psychology and structural sociology. In fact, considering the role of individual attributes has for a while been antithetical in the study of social structure (Mayhew, 1980). In this dissertation I argue that integrating these perspectives can advance our understanding of both. In particular, I argue that networks are essentially about the structural opportunities for the mobilization and flow of resources (Lin, 2005) but this alone says little about the quality of resources in the network. Our study suggests that the effectiveness of different network properties depends on the variety of the resources that flow through the network (cf. Hansen, 1999) and diversity is one of the team dimensions that have repeatedly been associated with resource variety (van Knippenberg & Schippers, 2007). Thus, a third contribution of this study is to investigate the contingent effect of networks and diversity on team performance. Conclusion In a nutshell, the literature on diversity has for many years proposed that diversity affects social networks, which, in turn, affect team performance. This dissertation taps into the different components of this model and extends it in three important ways. First, rather than assuming main effects of diversity I offer a test of its relevance as a contextual factor (Chapter 2 and Chapter 4). Second, given the abundance of research on the importance of homophily for the creation of networks I propose and test new antecedents of change in networks over time (Chapther 3). And third, I show how the diversity-network-team performance perspective can be re-framed into a contingency model (Chapter 4). This dissertation provides theoretical arguments and tests of an attempt to reframe the currently dominant model of diversity in multiple and new ways. Overview of this dissertation In addition to the current introductory chapter, this dissertation consists of three empirical chapters and a final chapter where I present the findings and general conclusions of the present