Journal of Applied Psychology 2009, Vol. 94, No.

2, 445 464

2009 American Psychological Association 0021-9010/09/$12.00 DOI: 10.1037/a0013935

Testing and Extending the Group Engagement Model: Linkages Between Social Identity, Procedural Justice, Economic Outcomes, and Extrarole Behavior
Steven L. Blader and Tom R. Tyler
New York University
Two field studies tested and extended the group engagement model (Tyler & Blader, 2000, 2003) by examining the model with regard to employee extrarole behavior. Consistent with the group engagement models predictions, results of these studies indicate that the social identities employees form around their work groups and their organizations are strongly related to whether employees engage in extrarole behaviors. Moreover, the studies demonstrated that social identity explains the impact of other factors that have previously been linked to extrarole behavior. In particular, the findings indicate that social identity mediates the effect of procedural justice judgments and economic outcomes on supervisor ratings of extrarole behavior. Overall, these studies provide compelling indication that social identity is an important determinant of behavior within work organizations and provide strong support for the application of the group engagement model in organizational settings. Keywords: social identity, procedural justice, economic outcomes, extrarole behavior

The integration of social identity theory (Tajfel & Turner, 1979) into organizational research provides one of the most successful recent examples of how basic psychological research can enhance our understanding of people within organizational contexts. Since Ashforth and Maels (1989) seminal theorizing about the utility of social identity theory for organizational scholarship, applications of social identity theory to organizational phenomena have flourished. Indeed, the fundamental insights of social identity theory, which emphasize that group membership shapes peoples understanding of who they are, have been embraced by many organizational researchers. As a result, the theory has emerged as an important framework for analyzing the psychology of individuals in organizations (e.g., Haslam, 2004). Some theorists have even argued that social identity is essential to understanding organizational life and, moreover, makes organizational life possible (Haslam, Postmes, & Ellemers, 2003). One recent theoretical model that advances this perspective and places social identity center stage is the group engagement model (Tyler & Blader, 2000, 2003), which holds that social identity is a critical ingredient for understanding the psychological basis of peoples engagement with their groups, organizations, and societies. The insights derived from the group engagement model have already been integrated into much organizational research and theorizing (e.g., Boezeman & Ellemers, 2007, in press; Dukerich, Golden, & Shortell, 2002; Ellemers, de Gilder, & Haslam, 2004; Fuller et al., 2006; Haslam & Reicher, 2006; Olkkonen & Lipponen, 2006; Sparrowe, Soetjipto, & Kraimer, 2006). This degree of

integration attests to the widespread interest in applications of social identity to organizational phenomena. However, empirical tests and theoretical extensions of the model are still quite limited. Indeed, there has been only one comprehensive test of the model (Tyler & Blader, 2000), and the results of this work are particularly in need of replication, given their reliance on self-report measures of engagement. We therefore conducted the current research to further develop the group engagement model. We did so by examining the models utility for understanding employee extrarole behavior (i.e., the extent to which employees go beyond role requirements in carrying out their jobs), a widely studied behavior essential to organizational viability and success (Podsakoff, MacKenzie, Paine, & Bachrach, 2000). Specifically, we tested the group engagement models key argument that social identity plays an influential role in determining whether employees engage in extrarole behavior, insofar as social identity not only relates to extrarole behavior but helps to explain the impact of two factors (procedural justice judgments and economic outcome judgments) that have themselves been shown to have strong linkages with extrarole behavior. Support for this proposal would further validate the group engagement model and would provide broader support for the argument that social identity is important for the study of behavior in work settings, in part because it provides a critical pathway through which other workplace factors affect behavior.

The Group Engagement Model: Social Identity and Extrarole Behavior

A core principle of the group engagement model is that an individuals behavioral effort on behalf of a collective to which he or she belongs (which we will broadly refer to as their group) is influenced by the role the group plays in how the individual thinks and feels about themself (i.e., by the social identity they form
445

Steven L. Blader, Stern School of Business, New York University; Tom R. Tyler, Department of Psychology, New York University. Correspondence concerning this article should be addressed to Steven L. Blader, Stern School of Business, New York University, 40 West Fourth Street, Room 7-18, New York, NY 10012. E-mail: sblader@stern.nyu.edu

446

BLADER AND TYLER

around the group; Tajfel & Turner, 1979). The group engagement model explains that this effect occurs because group members with strong social identities vis-a ` -vis a group are intrinsically motivated to facilitate the viability and success of their group. Because their group is integrated with their self-concept (i.e., the group and the self become overlapping psychological entities), these individuals are inherently concerned with their groups welfare and are therefore likely to behave on behalf of their groups interests. For them, group success is tantamount to individual success (Ashforth & Mael, 1989; Dukerich et al., 2002; Kramer, Hanna, Su, & Wei, 2001), and they develop an inherent concern with meeting the needs of the group (Blader, van Knippenberg, & Sleebos, 2008) and advancing group goals. The reasons people may integrate the group into their self-concepts in the first place are varied and complex and can include the groups fulfillment of important psychological functions, such as fostering self-worth (Tajfel, 1978), making sense of people and situations (Tajfel, 1978), reducing uncertainty (Hogg & Abrams, 1993), and satisfying belongingness needs (Baumeister & Leary, 1995). The notion that social identity impacts behavioral effort is also supported by social identity theorys argument that group norms gain greater traction as behavioral guides when groups become a focal basis for self-definition. This increase in influence occurs because strong social identity prompts deindividuation, which in turn prompts internalization of group norms and characteristics. Greater behavioral effort follows from social identity to the extent that group norms prescribe such effort (Haslam & Ellemers, 2005; van Knippenberg, 2000; van Knippenberg & Ellemers, 2003). Moreover, social exchange processes may lead people to engage in greater behavioral effort on behalf of a group with which they identify, as a way of reciprocating (and maintaining) the groups fulfillment of their social-identity-related needs. Empirical research confirms that social identity can impact behavior on behalf of ones group. Social identity has been linked to in-group-favoring actions (Brewer, 1979; Tajfel, Billig, Bundy, & Flament, 1971), collective actions (Blader, 2007c; Kelly & Kelly, 1994), monetary donations (Mael & Ashforth, 1992), and loyalty (Abrams, Ando, & Hinkle, 1998; Ellemers, Spears, & Doosje, 1997; Zdaniuk & Levine, 2001). Experimental research on actions more conceptually similar to extrarole behavior has linked social identity to greater effort on behalf of ones group (Worchel, Rothgerber, Day, Hart, & Butemeyer, 1998) and to reduced social loafing (Karau & Williams, 1993). Similarly, in actual work settings, social identity has been linked to self-reports of behavior and motivation (Dukerich et al., 2002; Olkkonen & Lipponen, 2006; Tyler & Blader, 2000; van Knippenberg & van Schie, 2000) and to supervisor-rated work effort among a select sample of employees (Bartel, 2001). However, not all actions geared toward promoting ones ingroup are created equal (Ouwerkerk, Ellemers, & de Gilder, 1999), and thus findings for other group-oriented actions do not necessarily generalize to the issue of extrarole behavior. Indeed, extrarole behavior requires greater effort and has a wider variety of determinants (Organ & Ryan, 1995; Podsakoff et al., 2000) than do many of the group-serving behaviors examined in previous social identity research. In addition, because of the discretionary nature of extrarole behavior, social pressures to comply and engage in extrarole behavior may be weaker than they are for other groupserving behaviors. At the same time, due to its discretionary

nature, extrarole behavior may be especially open to influence by intrinsic motivators such as identity (Tyler & Blader, 2000). Further, of those studies listed above that examine behaviors most conceptually similar to extrarole behavior, several are experimental studies that may not replicate to real-world settings (e.g., Karau & Williams, 1993; Worchel et al., 1998), and the others are field studies that rely on self-report measures or specialized samples of employees (e.g., Bartel, 2001; Dukerich et al., 2002; Olkkonen & Lipponen, 2006; Tyler & Blader, 2000; van Knippenberg & van Schie, 2000). As such, it is difficult to draw definitive conclusions about social identitys impact on extrarole behavior, and stronger examination of the relationship between these constructs is needed. Examining this relationship can serve as a test of social identitys motivational consequences and can fill a gap in organizational identity research that has resulted from more rapid theory development than theory testing (Corley et al., 2006; Dukerich et al., 2002; Foreman & Whetten, 2002). In response, we tested the following hypothesis: Hypothesis 1: Employee social identity will be positively related to employee extrarole behavior.

The Group Engagement Model: Procedural Justice and Economic Outcomes

A major contribution of the group engagement model is that it not only emphasizes the importance of social identity but facilitates a more integrative understanding, because it proposes a causal framework that captures the relationship between multiple antecedents of engagement. That is, the group engagement model goes beyond previous work by explicitly locating social identity in relation to other factors that also influence the engagement of group members in their groups. The group engagement model accomplishes this by drawing on social psychological research and theorizing about the primary bases of peoples reactions to encounters with their groups and organizations. This work distinguishes between two primary bases of peoples reactions (Brockner & Wiesenfeld, 1996; Tyler & Blader, 2000; van den Bos & Lind, 2002). One basis of peoples reactions is the processes and treatment that people encounter in their group (i.e., their experience with how the group functions), which in turn impacts their experience as group members. Research has shown that fairness is a primary dimension that people use to evaluate the processes and treatment they encounter in their groups (Tyler, 2000; Tyler & Blader, 2000). This is reflected in their judgments of procedural fairness. The second basis of the reactions of people to encounters with their groups is the outcomesparticularly the economic outcomesthat they attain as a result of group membership. Research confirms the value of this framework and the importance of each of these bases of peoples reactions. For instance, the importance of procedural justice is supported by a rich set of findings indicating that employee evaluations of the fairness of decision-making processes and the quality of treatment (referred to by some researchers as interactional justice; cf. Blader & Tyler, 2003a, 2003b) have a strong influence on employee behavior (Blader & Tyler, 2005; Cohen-Charash & Spector, 2001; Colquitt, Conlon, Wesson, Porter, & Ng, 2001). This is particularly true for extrarole behaviors (Cohen-Charash & Spector, 2001; Colquitt et

TESTING AND EXTENDING THE GROUP ENGAGEMENT MODEL

447

al., 2001; LePine, Erez, & Johnson, 2002; Moorman, 1991; Organ & Ryan, 1995). Because they are discretionary, these behaviors provide greater opportunity for influence by factors such as procedural justice (Tyler & Blader, 2000). The group engagement model holds that this well-replicated impact of procedural justice on extrarole behavior (and, more generally, on employee behavior) is due in part to the role that procedural justice plays in determining whether or not employees link their social identity to their organization. This argument has its roots in relational models of procedural justice (Lind & Tyler, 1988; Tyler & Lind, 1992), one of several theoretical approaches that address why people react to procedural justice (Blader & Tyler, 2005). The relational models hold that the procedural justice experienced in a group setting provides a cue that people use to evaluate the nature of their relationship with their group. The group engagement model further develops this insight with the argument that people use their procedural justice judgments as a basis for evaluating whether or not a particular group is one in which they can safely invest their social identities (Tyler & Blader, 2002, 2003), because procedural justice communicates to them whether the group is likely to help them develop and maintain a satisfying, positive social identity. For this reason, procedural justice judgments are an antecedent of whether people socially identify with a particular group. Thus, the group engagement model holds that one function served by procedural justice information is to help people determine whether or not they will socially identify with their group. Therefore, the group engagement model integrates arguments that (a) procedural justice impacts behavior, (b) procedural justice impacts social identity, and (c) social identity impacts behavior, and reasons that social identity accounts for at least part of the reason that procedural justice impacts behavior. The model assigns social identity a role in explaining the effect of procedural justice on behavior because it regards social identity as central to peoples engagement in their groups. This assumption leads to our second hypothesis. Hypothesis 2: The relationship between procedural justice judgments and employee extrarole behavior will be partially mediated by employee social identity. The second basis of peoples reactions to encounters with their groups and organizations is their judgments about the economic outcomes that they receive as a result of group membership. This argument has a long and established history in the social sciences. The notion that employees are motivated by economic outcomes is central to several classic motivation theories (e.g., exchange and equity theories, Adams, 1965; Blau, 1964; Homans, 1961; Thibaut & Kelley, 1959), is a core assumption of economic models of behavior, and is prevalent among laymen, organizational practitioners, and academic researchers alike. It is widely accepted that issues such as pay and benefits, as well as other short- and long-term financial benefits that accrue from group membership, determine employees work behavior (e.g., Lawler, 1990; Milkovich & Newman, 1999; Rynes, Gerhart, & Parks, 2005). Importantly, economic outcomes have been linked specifically to the performance of extrarole behaviors. Research indicates that employees associate extrarole behaviors with favorable economic behaviors, even though these outcomes are discretionary and are not formalized in reward systems. As such, economic outcomes

can impact whether employees engage in these behaviors (Folger, 1993; Haworth & Levy, 2001; Hui, Lam, & Law, 2000; McAllister, Kamdar, Morrison, & Turban, 2007; Podsakoff et al., 2000). Such an impact is likely due to a sense among employees that managerial evaluations of their performanceand thus, the economic outcomes that result from performance evaluationsare influenced by whether or not they engage in extrarole behaviors (Haworth & Levy, 2001; Podsakoff et al., 2000). Notably, this judgment is likely reasonable and accurate (Allen & Rush, 1998; Hoffman, Blair, Meriac, & Woehr, 2007; Hui et al., 2000; MacKenzie, Podsakoff, & Fetter, 1991; Orr, Sackett, & Mercer, 1989). The prominence of economic outcomes in models of employee behavior reflects the assumption that employees are fundamentally resource-interested actors and that instrumentality considerations underlie the value they place on economic outcomes. However, considerable evidence demonstrates that people do not always act in the service of their instrumentality and resource concerns (see Gintis, Bowles, Boyd, & Fehr, 2005; Haslam & Ellemers, 2005; Kohn, 1999; Miller & Ratner, 1998; Pearce, 1987). This evidence suggests that alternate concerns may help explain employee reactions to economic outcomes. The influence of these alternate concerns is supported by findings that economic outcomes can carry important symbolic value for assessments of ones self-worth, relationships, and status (Furnham & Argyle, 1998; Jenkins, Mitra, Gupta, & Shaw, 1998; Mitchell & Mickel, 1999; Opsahl & Dunnette, 1966; Porter, Bigley, & Steers, 1996). Favorable economic outcomes communicate a positive message to employees about these issues, whereas unfavorable outcomes communicate a negative message. The group engagement model (Tyler & Blader, 2000, 2003) builds on this insight and holds that the symbolic value of economic outcomes can be relevant to social identity. In particular, the group engagement model states that economic outcomes can convey a message to group members that resembles the message conveyed by procedural justice, insofar as both messages inform people about the nature and quality of their group membership and, thus, whether or not they can safely invest their social identities in the group. As a result, the model postulates that one reason for the possible impact of economic outcomes on employee behavior is the social-identity-relevant symbolism (and social identity implications) of economic outcomes. Therefore, the current studies tested the following hypothesis: Hypothesis 3: The relationship between economic outcome evaluations and employee extrarole behavior will be partially mediated by employee social identity. In sum, Hypotheses 2 and 3 tested whether social identity is a psychological intermediary through which procedural justice and economic outcomes relate to employee extrarole behavior. Our focus on procedural justice and economic outcomes is rooted in the argument, from the social psychological literature, that these factors are primary bases for peoples reactions to encounters with groups and organizations. Examination of these factors in the current studies was further supported by findings that both procedural justice and economic outcomes are prominent predictors of extrarole behavior.

Conceptualizing Social Identity

The current studies conceptualize social identity directly as it is conceptualized in the social identity literature and by the group

448

BLADER AND TYLER

engagement model: as a multidimensional construct with various elements that work in coordination to produce the implications of group membership on the self-concept (Tajfel, 1978, 1982; see also, e.g., Bergami & Bagozzi, 2000; Haslam, 2004; Tyler & Blader, 2000; van Dick, Wagner, Stellmacher, & Christ, 2004). One component is the cognitive component or, in other words, an employees sense of oneness with their organization. This component captures the extent to which group membership is selfdefining and thus determines the implications of group membership for how people think of themselves. It also embodies the emphasis of self-categorization theory (Turner, 1985) on the extent to which people see themselves as group members. This component is the typical focus of organizational research that adopts a social identity framework (e.g., Ashforth & Mael, 1989; Dukerich et al., 2002; Dutton, Dukerich, & Harquail, 1994; Olkkonen & Lipponen, 2006). However, social identity theory emphasizes that there is a second component of social identity, the evaluative component. The evaluative component captures the value people attach to their group memberships, which in turn determines the significance of a group membership for how people think and feel about themselves. Although this element of social identity is not often examined in organizational research, it is central to classic social identity theorizing (Tajfel, 1978, 1982), where it is argued that the evaluative connotation of an identity is critical for understanding the identitys impact on, and significance to, the individual. This perspective has more recently been reflected in two arguments: that the impact of a social identity is determined by the extent to which it fosters positive self-definition and self-esteem (Ellemers, 1993; Haslam, Powell, & Turner, 2000; Hogg & Abrams, 1990; Mael & Ashforth, 1992; Tajfel & Turner, 1979) and that the evaluative aspects of group membership are critical for understanding the nature ofand reactions toidentity (Ashmore, Deaux, & McLaughlin-Volpe, 2004; Dutton et al., 1994; March & Simon, 1958). These arguments suggest that the omission of evaluative concerns from organizational research on social identity may represent an underspecification of the social identity construct, especially in work organizations in which permeable group boundaries, voluntary membership, and structural elements that accentuate intergroup and intragroup comparisons likely focus people on the evaluative implications of membership. The current studies assess the evaluative component of social identity via two judgments: pride and respect (Tyler & Blader, 2000, 2001; Tyler, Degoey, & Smith, 1996). Pride refers to evaluations of the standing of ones group (it is an intergroup evaluative judgment). Pride has its roots in early social identity theorizing (Tajfel, 1978, 1982; see also Hinkle & Brown, 1990; Smith & Tyler, 1997), which was primarily concerned with issues of intergroup relations. Pride directly addresses the evaluative connotation of group membership, because it is an assessment of the groups general worth and status. When group members regard their group as having high status, their social identity vis-a ` -vis the group is likely to positively contribute to their self-concept. However, as applications of social identity theory have begun to focus on intragroup phenomena, there has been an emerging recognition of the importance of respect. Respect refers to evaluations by people of their standing and acceptance within their group (it is an intragroup evaluative judgment). The core premise underlying the importance of respect is that someones standing within their

group shapes the social identity they form around the group as well as the implications of group membership on their self-concept. Indeed, group membership will have rather different social identity implications for low- versus high-status members of a given group; the thoughts and feelings of individuals about themselves that result from group membership (their social identity) will be more negative among low- versus high-status group members. Social psychological research confirms that peoples judgments about their intragroup standing (i.e., their respect judgments) influence how they relate to their groups (De Cremer, 2002; Jetten, Branscombe, & Spears, 2002; Noel, Wann, & Branscombe, 1995; Sleebos, Ellemers, & de Gilder, 2006; Smith, Tyler, Huo, Ortiz, & Lind, 1998) and that this influence is due to the relevance of respect for social identity (Simon & Sturmer, 2003; Smith & Tyler, 1997; Spears, Ellemers, Doosje, & Branscombe, 2006; Sturmer, Simon, & Loewy, 2008; Tyler & Blader, 2000, 2001). In other words, extensive research directly links respect to social identity and to group-oriented behavior and furthermore suggests that respect may be at least as important as intergroup status judgments (Crocker & Major, 1989; Davis, 1966; McFarland & Buehler, 1995). Such research reinforces arguments that respect is an issue related to the social self. Therefore, we assessed the cognitive and evaluative components of social identity in the current studies by measuring (a) cognitive representations of the connection between the group and the self (i.e., identification) and (b) assessment of pride and respect. Further, we examined this multidimensional conceptualization of social identity with regard to two separate social identity targets or foci: at the level of the work group in Study 1 and at the level of the overall organization in Study 2. This approach reflects the perspective that social identity theory and the group engagement model are concerned with the psychological experience of group membership. They focus on the part of the self-concept that captures the way the self relates to groups in general (i.e., they focus on collective identity; Brewer & Gardner, 1996). As such, their insights are not specific to a particular type of group or a particular level of nested subgroups. However, integration of these theories within organizational contexts has focused on social identity as it relates to membership in ones organization; only recently has there been an emerging recognition in the organizational literature of the importance of examining a broader set of social identity targets (Ashforth & Johnson, 2001; Blader, 2007c; Johnson, Morgeson, Ilgen, Meyer, & Lloyd, 2006; Olkkonen & Lipponen, 2006; Pratt & Foreman, 2000; van Dick et al., 2004; van Knippenberg & van Schie, 2000). As a result, we tested our hypotheses at the work group level (Study 1) and the organizational level (Study 2), i.e., we tested the group engagement model at both of these levels. We did so because extrarole behavior can benefitand express allegiance towardwork groups and organizations.

Current Research
The two field studies we present below tested three hypotheses based on the group engagement model. In them, we used the models multidimensional conceptualization of social identity and compared the utility of this approach with that of one that focuses solely on the cognitive component. The conceptual model that follows from these proposals is presented in Figure 1. Study 1 was

TESTING AND EXTENDING THE GROUP ENGAGEMENT MODEL

449

Procedural Justice

Hypothesis 2

Identification Social Identity Hypothesis 1 Extrarole Behavior

Pride

Respect

Hypothesis 3

Economic Outcomes

Future Evaluation Outcomes of Pay

Outcome Fairness

Incentives

Figure 1.

Conceptual model. Dotted line indicates predicted mediated effort.

conducted in a major financial services organization. The prominence of economic concerns in this company made it a particularly stringent testing ground for the influence of social identity on extrarole behavior. Study 2 was designed to provide a broader test and, thus, included respondents heterogeneous with regard to factors such as industry, job, income, and locale.

Study 1 Method
Participants and procedure. Five hundred and forty U.S. employees of a single division of an international financial services organization responded to a survey that included the measures described below (in all, 1,350 employees, representing the entire U.S. division, received the survey; thus, there was a 40% response rate). Employees received questionnaires via interoffice mail and were asked to complete them at work and then return them directly to the investigators using enclosed business-reply envelopes. The employees were assured confidentiality by both the investigators and the organizations management. Supervisor surveys were distributed to the 231 immediate supervisors of all 1,350 employees in the division. Of these supervisors, 64 returned assessments of the employees in their work group (a 28% supervisor response rate); in total, these 64 supervisors returned assessments of 415 individual employees (i.e., we received supervisor evaluations for 31% of our total sample). The overlap between these supervisor evaluations and our employee respondent pool led to a final employee supervisor matched sample of 112 employees.

Among this subsample of matched employeesupervisor respondents (n 112), 52% were female and 46% had pursued some postbachelors education. The mean tenure with the firm was 13.7 years, and the mean age was 44.47 years. The sample was only moderately heterogeneous with respect to race; 82% of respondents were Caucasian, 13% were Latino, 4% were African American, and 2% were Asian. It is significant that these demographics resembled those of the sample of employee respondents (N 540), in which 53% were female, 48% had pursued some postbachelors education, the mean tenure was 13.67 years, and the mean age was 42.9 years. Racial distribution was similar (the total sample was 74% Caucasian, 11% Latino, 5% African American, and 5% Asian). Further, the demographics among the full population of employees in the division were quite similar: 59% were female, the mean tenure was 13 years, and the mean age was 43 years. Race was distributed such that 63% were Caucasian, 17% were Latino, 11% were African American, and 8% were Asian. Reliable education data were unavailable for employees in the division overall. It is worth noting that the division in which we conducted Study 1 was structured into distinct work groups; divisions among these work groups were very salient. Work groups were a highly relevant social identity target for employees in this study. Our sample was widely distributed across work groups. Indeed, there were 64 different work groups represented in the data (33 work groups had only 1 respondent; 19 had 2 respondents; 10 had 3 respondents; 1 had 5 respondents; and 1 had 6 respondents).

450

BLADER AND TYLER

Measures. The items comprising each of the measures, presented in the Appendix, were adapted from previous research (Tyler & Blader, 2000, 2001, 2002, except where otherwise noted). All measures were included in the employee survey, except for the extrarole behavior measure, which was included on the supervisor survey. As indicated in Figure 1, social identity vis-a ` -vis the work group was measured with scales that assessed identification, pride, and respect and loaded them onto a common latent factor in the structural equation models presented below. Instructions preceding the procedural justice and various economic outcome measures oriented respondents to answer the questions with regard to the job they held within their work group. Economic outcome judgments were measured with scales that assessed evaluations of current pay, expectations of future economic outcomes, outcome fairness, and the perceived role of pay for performance (i.e., referred to as incentives below), which were loaded onto a common latent factor in the structural equation models.

Results
Table 1 presents the means, standard deviations, coefficient alphas, and interscale correlation matrix for all measures. All measures demonstrated strong reliability. Three control variables were included in all analyses: age, tenure, and gender. We conducted a series of structural equation models (via AMOS) to test the conceptual model presented in Figure 1 and, more specifically, to test the predicted positive relationship between employee social identity and extrarole behavior (Hypothesis 1), as well as the predicted mediating role of social identity in explaining the relationship between procedural justice (Hypothesis 2) and economic outcomes (Hypothesis 3) on extrarole behavior. The results of these structural equation models are presented in Table 2. One potential concern is that the sample represents a fairly small segment (n 112) of the total pool of respondents (N 540). Although response rates to the employee and supervisor surveys were independent from one another and were within traditional ranges for field research in organizational settings

(and, as such, there are no systematic methodological issues that would suggest the sample is biased), the overlap between them nevertheless yielded a relatively small sample of matched employee employer respondents. Thus, as an additional check, we examined the equivalence between the subsample of matched respondents and the remaining pool of employee respondents on the mean levels of all variables collected in the employee survey. The results indicate that there were no significant differences between the subsample of matched respondents and the broader employee respondent group (i.e., means on all variables except for age were not significantly different from one another between the two groups; all Fs 3.30, ps .05). Although the mean age of respondents in the subsample was somewhat higher than that of the other respondents (44.5 years vs. 42.9 years), this difference was relatively small (2 .009). Moreover, it is not clear how this age difference would impact interpretation of the results. We also compared supervisor ratings of extrarole behavior between our matched subsample and the remaining sample of employees for whom supervisor ratings were received; there was not a significant difference between these groups on these ratings, F(1, 413) 0.007, ns. Thus, there are methodological and empirical reasons to believe that our matched sample represented a nonbiased sample from the overall pool from which these respondents were drawn. Confirmatory factor analyses. Prior to testing our hypotheses and our proposed structural model, we examined the fit of our measurement model via a confirmatory factor analysis that included procedural justice, extrarole behavior, the three social identity measures (identification, pride, and respect) loaded onto a common latent factor (social identity), and the four economic outcome judgments (evaluations of pay, expectations of future economic outcomes, outcome fairness, and incentives) loaded onto a common latent factor (economic outcomes). To assess model fit in this analysis and throughout this article, we adopted the standards from the literature that reasonable fit is indicated by comparative fit index (CFI) and incremental fit index (IFI) values greater than .90 and root-mean-square error of approximation (RMSEA) values less than .10 (Diamantopou-

Table 1 Study 1: Means, Standard Deviations, Coefficient Alphas, and Correlations Among Scales
Scale 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. Procedural justice Identification Pride Respect Evaluations of current pay Future economic outcomes Outcome fairness Incentives Extrarole behavior Gender Tenure, years Age, years M (SD) 3.34 (1.10) 4.45 (0.85) 4.05 (1.09) 4.48 (1.19) 2.87 (1.13) 3.16 (1.01) 2.84 (1.19) 3.15 (1.15) 4.72 (0.98) 13.70 (8.56) 44.47 (8.75) 1 .96 .43 .36 .58 .50 .52 .69 .68 .26 .03 .00 .20 2 .85 .57 .50 .13 .33 .23 .26 .30 .08 .01 .13 3 4 5 6 7 8 9 10 11 12

.88 .48 .26 .41 .34 .39 .24 .01 .09 .09

.97 .43 .49 .43 .57 .39 .07 .08 .10

.95 .48 .68 .61 .21 .15 .05 .14

.70 .57 .57 .19 .00 .07 .07

.85 .74 .21 .05 .19 .17

.88 .27 .06 .06 .23

.94 .10 .05 .01

.03 .05

.44

Note. n 112. Higher numbers indicate greater endorsement of the construct (e.g., more positive fairness, stronger identification). Numbers on the diagonal represent coefficient alphas for the scale. All correlations .19 are statistically significant at p .05.

Table 2 Study 1: Results of Structural Equation Models

Hypothesis 2 Model 2 b M (SE) B M (SE) b M (SE) b M (SE) Model 3 Model 4 Model 5 b Hypothesis 3 Overall model Model 6 M (SE)

Hypothesis 1

Model 1

Variable

M (SE)

0.64 0.24 0.35 9% .36 44% 4.4 2 .98 .98 .10 1 1 .00 .95 .95 .12 0 0 11 4 2.6 5 .99 .99 .01 (.06) .66 21% 0.29 (.12) 10% (.08) .27

(.17)

.42 0.81 0.04 (.26) (.12) 0.82 0.00

.49 .05

(.29)

.46 .01 (.17) 22%

0.84 0.08 0.06

(.29) (.15) (.20) 23% 0.21 0.44 (.09) 44% 35.4 18 .95 .95 .09 .67 0.24 (.08) (.11) 49% 46 23 .95 .95 .09

.48 .09 .05

DV: Supervisor ratings of extrarole behavior Social identity Procedural justice Economic outcome judgments R2

19%

.41 .34

DV: Social identity Procedural justice Economic outcome judgments R2

TESTING AND EXTENDING THE GROUP ENGAGEMENT MODEL

2 df Fit indices CFI IFI RMSEAa

Note. n 112. All regression coefficients and R2 estimates are from analyses that include controls for the effects of age, gender, and tenure on extrarole behavior (though in no case did any of these controls have significant effects). DV dependent variable; CFI comparative fit index; IFI incremental fit index; RMSEA root-mean-square error of approximation. a RMSEA can be a poor estimator of fit for small samples, whereas IFI is relatively unaffected by sample size (Bollen, 1990; Curran et al., 2003). Therefore, RMSEA results should be interpreted with caution in this study. IFI, on the other hand, provides a more reliable estimate of fit for the current study. p .05. p .01. p .001. nonsignificant.

451

452

BLADER AND TYLER

los & Siguaw, 2000; Hu & Bentler, 1995; Schumacker & Lomax, 2004).1 This analysis confirmed that the proposed measurement model fit the data well, 2(23, N 112) 46.4, p .003, CFI .95, IFI .95, RMSEA .09. This model was better fitting than was an alternate single-factor model that loaded all the variables on a single factor, 2(4) 45.5, p .001.2 Moreover, the convergent validity of our social identity and economic outcome latent factors was supported by their factor loadings. The factor loadings of the social identity indicators were all statistically significant and of considerable magnitude (identification, .77, p .001; pride, .74, p .001; respect, .64, p .001), as were the factor loadings of the economic outcome indicators (evaluations of pay, .74, p .001; future economic outcomes, .65, p .001; outcome fairness, .89, p .001; incentives, .83, p .001). Is social identity related to extrarole behavior? Hypothesis 1 predicted that social identity would be significantly related to extrarole behavior. An analysis that tested this prediction (see Table 2, Model 1) indicates support for this hypothesis, as it shows a significant relationship between social identity and extrarole behavior ( .42, p .001, R2 19%). This result supports the prediction that social identity has important linkages to extrarole behavior. Does social identity mediate the relationship between procedural justice and extrarole behavior? Hypothesis 2 predicted that social identity would mediate the relationship between procedural justice and extrarole behavior. To test this hypothesis, we first examined whether or not procedural justice was significantly related to extrarole behavior (Baron & Kenny, 1986). That analysis (see Table 2, Model 2) indicates that procedural justice is indeed significantly related to extrarole behavior ( .27, p .01, R2 9%). This finding is consistent with previous research (Colquitt et al., 2001). Most critical to the current analysis, however, is whether or not work group social identity helps explain this relationship. The results support this hypothesis (see Table 2, Model 3). They show that the conditions for mediation (Baron & Kenny, 1986) are met in these data: procedural justice is related to social identity ( .66, p .001, R2 44%); social identity is related to extrarole behavior ( .49, p .01); and, most important, the prior relationship between procedural justice and extrarole behavior becomes nonsignificant ( .05, ns) when social identity is included in the analysis (Sobels z 2.76, p .01). These results suggest that work group social identity fully mediates the relationship between procedural justice and extrarole behavior. Does social identity mediate the effects of economic outcomes? Hypothesis 3 predicted that social identity would mediate the relationship between economic outcomes and extrarole behavior. To test this, we first examined whether or not economic outcome judgments were significantly related to extrarole behavior. The results (presented in Table 2, Model 4) indicate that they are ( .29, p .01, R2 10%), insofar as more positive evaluations of issues such as pay, future economic outcomes, outcome fairness, and incentives are associated with higher levels of extrarole behavior. Next, we tested Hypothesis 3 by examining whether social identity mediates the relationship between economic outcome judgments and extrarole behavior (see Table 2, Model 5). This

analysis revealed support for Hypothesis 3, as it shows that the conditions for mediation (Baron & Kenny, 1986) were met in these data: economic outcome judgments are related to social identity ( .67, p .001, R2 44%); social identity is related to extrarole behavior ( .46, p .01); and the prior relationship between economic outcome judgments and extrarole behavior becomes nonsignificant ( .01, ns) when social identity is included in the analysis (Sobels z 2.45, p .01). These results indicate that work group social identity fully mediates the relationship between economic outcomes and extrarole behavior. Overall model. We also explored the simultaneous relationships between social identity, procedural justice, economic outcome judgments, and extrarole behavior (see Table 2, Model 6). As might be expected on the basis of Models 3 and 5, the results indicate that social identity is related to extrarole behavior ( .48, p .01), whereas economic outcome judgments ( .05, ns) and procedural justice ( .09, ns) do not have significant associations with extrarole behavior when accounting for social identity (R2 23%) or each other. This analysis also revealed significant linkages between social identity and both procedural justice ( .41, p .01) and economic outcomes ( .34, p .05) (R2 49%). These linkages indicate that procedural justice and economic outcomes each uniquely contribute to work group social identity. Additional analyses. We conducted an additional set of analyses to compare the utility of our conceptualization of social identity, which includes both cognitive and evaluative components, with that of a conceptualization that focuses exclusively on the cognitive component (i.e., on identification). To do this, we repeated the analyses presented in Table 2 and described above but this time replaced the social identity latent factor with the single observed measure of identification. First, with regard to Hypothesis 1, although identification was related to extrarole behavior ( .29, p .001, R2 9.4%), this relationship was considerably weaker than was that of our multidimensional conceptualization of social identity (R2 19%; presented in Table 2, Model 1). Second, with regard to Hypothesis 2, although identification did mediate the relationship between procedural justice and extrarole behavior (i.e., in a model accounting for identification, .18, ns, Sobels z 2.10, p .05, for procedural justice), less variance in extrarole behavior was explained by this model, which included identification and proceRMSEA can be a poor estimator of fit for small samples, whereas IFI is relatively unaffected by sample size (Bollen, 1990; Curran, Bollen, Chen, Paxton, & Kirby, 2003). RMSEA results should be interpreted with caution in Study 1. IFI, on the other hand, provides a more reliable estimate of fit for the current study. 2 Furthermore, this model was a better specification of the measurement model than was a model that loaded pride and respect onto a common latent factor (evaluative component) that in turn was loaded, with identification, onto a common latent factor (social identity). This alternate model calculated a negative variance for the residual of the evaluative component latent factor; the result was an inadmissible solution that indicated misspecification and lack of fit for this proposed model (Chen, Bollen, Paxton, Curran, & Kirby, 2001). This result suggests that, in these data, identification, pride, and respect are best conceptualized as operating at the same conceptual level.
1

TESTING AND EXTENDING THE GROUP ENGAGEMENT MODEL

453

dural justice (R2 12%), than by the original model (R2 21%; see Table 2, Model 3). Third, with regard to Hypothesis 3, identification partially mediated the relationship between economic outcomes and extrarole behavior (i.e., in a model accounting for identification, .23, p .05, Sobels z 1.97, p .05, for economic outcomes), whereas our multidimensional conceptualization of social identity fully mediated the relationship between economic outcomes and extrarole behavior (see Table 2, Model 5). Once again, this alternate model explained less variance in extrarole behavior (R2 15%) than did its counterpart that used the multidimensional conceptualization of social identity (R2 22%; see Table 2, Model 5). These findings complement our argument that social identity is best conceptualized as containing cognitive and evaluative elements, insofar as the broader conceptualization is more strongly associated with extrarole behavior and is a more robust mediator of the relationship between extrarole behavior and both procedural justice and economic outcomes.

Although consistent with our predictions, the results of Study 1 require further examination. The sample was drawn from a single organization, and the effects for work group social identity are thus possibly an artifact of the specific organization or its context. It was therefore important to test the hypotheses on a sample that was more heterogeneous with regard to context, occupation, and industry. Moreover, given the complexity of some of the models examined (e.g., Model 6), the sample size in Study 1 was relatively small. Thus, it would be desirable to replicate these results on a larger sample. Finally, Study 1 examined the social identities that employees form around their work groups, but it is also critical to examine whether similar results emerge for the social identities employees form around their organizations. Study 2 directly addressed these issues by examining organizational social identity among a large (N 831) and heterogeneous sample of employees from throughout the United States.

Study 2 Method
Participants and procedure. Participants in Study 2 were a random sample drawn from a national panel of respondents designed to be representative of the entire U.S. population. Members of the panel, which is maintained by a private organization (Knowledge Networks, Inc.), were initially contacted for their participation via random digit dialing. An incentive of free Internet access via WebTV is provided to all those who join, and all surveys administered to the panel were conducted over WebTV (for more details on the panel, please contact the first author). Potential respondents for the current study were screened to ensure that they worked at least 20 hr a week, had a primary supervisor, and had worked at their current employer for at least 3 months. Respondents who met these criteria completed the survey in two parts, 1 week apart (the survey was not administered in a single time period because the length exceeded participants weekly time commitments to the panel). In total, there were 4,430 employee respondents in the sample. All respondents were asked if they would be willing to have their supervisor contacted about completing and returning a confidential questionnaire. Financial incentives were offered to employees and supervisors for each completed supervisor questionnaire received. In all, 2,020 (46%) employee respondents allowed their supervisor to be contacted, and 831 supervisors (41%) completed and returned the questionnaire they were sent. The data presented below are based on this subsample of 831 matched employee employer respondents. Employees in the subsample of matched employeesupervisor respondents came from a variety of organizations: 26% worked for small businesses, 19% worked for large companies in one location, 36% worked for large, multicity American companies, and 17% worked for multinational companies. Respondents were evenly split by gender (48% female); their average tenure at their current job was 4.49 years, and their mean age was 44 years. Of those sampled, 75% had at least some college experience (only 9% had earned a postbachelors degree). Measures. The items comprising each of the measures in this study are presented in the Appendix. Measures generally paralleled those in Study 1, with a few exceptions. In particular, in Study 2 we focused on social identity linked to the organization (rather

Discussion
The results of Study 1 support the three key hypotheses examined in the current research and, more generally, support application of the group engagement model to the issue of extrarole behavior. First, the social identities of employees vis-a ` -vis their work groups were significantly related to supervisor ratings of their extrarole behavior. This finding corroborates the value of social identity in organizational contexts and highlights the potential importance of examining social identity with the work group as the target. Second, beyond simply demonstrating that work group social identity matters, these findings go a step further to show how work group social identity relates to two key elements of the workplace that have been shown to impact extrarole behavior. Namely, the results indicate that work group social identity accounted for the relationship between extrarole behavior and employees procedural justice judgments as well as their economic outcome judgments. Evaluations of these workplace conditions are related to extrarole behavior because of the relevance of such evaluations to the social identities that employees form around their work group. It is noteworthy that the mediation analyses provided even stronger support for our hypotheses than was expected, insofar as they indicate full mediation of procedural justice and economic outcomes (rather than partial mediation, which had been predicted). A great deal of research attention has been devoted to understanding why people react to their procedural justice and economic outcome judgments. The current results contribute to that work and highlight that the social identity implications of procedural justice and economic outcomes should be a focus of those efforts. Overall, these results support the group engagement models argument that social identity is a psychological intermediary through which features of the workplace impact behavior. The results also show the potential value of adopting a multidimensional conceptualization of social identity, as this conceptualization was a more powerful predictor of extrarole behavior than was a conceptualization that focused solely on the cognitive component of social identity. In addition, it better accounted for the effects of procedural justice and economic outcomes on extrarole behavior.

454

BLADER AND TYLER

than the work group), and the social identity measures were adjusted accordingly. Also, instructions for the questionnaire oriented respondents to answer the questions in the survey with regard to their work organization (in Study 1, respondents were asked to focus on their current job in their work group). In addition, the procedural and distributive justice scales were expanded to capture assessments of the fairness of the culture of the overall organization (Naumann & Bennett, 2000), as in Study 1, as well as to index the perceived fairness directly experienced by respondents. The extrarole behavior measure completed by supervisors was expanded to inquire about a broader range of behaviors. Due to space constraints (in part as a result of expansion of these other scales), incentives were not measured in Study 2.

Results
Means, standard deviations, coefficient alphas, and the interscale correlation matrix for all measures are presented in Table 3. All scales demonstrated strong reliability. Table 3 also includes three control variables that were included in all analyses: age, tenure with the organization, and gender. Once again, we conducted a series of structural equation models (via AMOS) to test the conceptual model presented in Figure 1 and, in particular, to test the predicted interrelationships between social identity, procedural justice, economic outcome judgments, and extrarole behavior. The results of these structural equation models are presented in Table 4. As in Study 1, mean ratings on each of the variables collected from employees were compared between the subsample of matched employee employer respondents and the remaining pool of employee respondents (for whom no supervisor ratings were available). The results suggest that differences between the two groups were inconsequential. Due to the large sample sizes involved in this study, statistical significance does not provide a meaningful indicator of the equivalence of the groups, because even negligible differences will likely have statistical significance. However, investigation of the effect sizes of these differences indicates that there were no sizable differences in the means of any of the variables between the two respondent groups (all 2s .02). This result indicates that the subsample of matched employee supervisor respondents is likely equivalent to the larger pool of employees in our study. Confirmatory factor analyses. We conducted confirmatory factor analyses to examine the fit of the proposed measurement model. These analyses included procedural justice, extrarole behavior, the three social identity measures (identification, pride, and respect) loaded onto a common latent factor (social identity), and the three economic outcome judgments (evaluations of pay, expectations of future economic outcomes, and outcome fairness) loaded onto a common latent factor (economic outcomes). This analysis confirmed that the proposed measurement model fit the data well, 2(16, N 831) 118, p .001, CFI .94, IFI .94, RMSEA .087, and that it was better fitting than was a simpler, one-factor model, 2(4) 243.4, p .001.3 Moreover, the convergent validity of our social identity and economic outcome latent factors was again supported by their factor loadings. The factor loadings of the social identity indicators were all statistically significant and of considerable magnitude (identification, .78, p .001; pride, .75, p .001; respect, .60,

p .001), as were the factor loadings of the economic outcome indicators (evaluations of pay, .87, p .001; future economic outcomes, .75, p .001; outcome fairness, .78, p .001). Is social identity related to extrarole behavior? The analysis testing Hypothesis 1 (see Table 4, Model 1) indicates support for this hypothesis, as it shows a significant relationship between social identity and extrarole behavior ( .23, p .001, R2 6%). This result provides further support for our prediction that social identity has important linkages to employee engagement in extrarole behavior. Does social identity mediate the effects of procedural justice? The results presented in Table 4 (see Models 2 and 3) indicate support for Hypothesis 2, which predicted that social identity would mediate the relationship between procedural justice and extrarole behavior. They do so by showing that the conditions for mediation were met in this data, insofar as there was a significant association between procedural justice and extrarole behavior (Model 2, .17, p .001, R2 5%) and there was a significant association between procedural justice and social identity (Model 3, .79, p .001, R2 63%). In addition, the relationship between procedural justice and extrarole behavior becomes nonsignificant (Model 3, .01, ns) when accounting for the role of social identity, which was itself significantly related to extrarole behavior (Model 3, .14, p .01), (Sobels z 2.67, p .01). These findings support Hypothesis 2 by showing that social identity fully mediates the relationship between procedural justice and extrarole behavior. Does social identity mediate the effects of economic outcomes? The results presented in Table 4 (see Models 4 and 5) also indicate support for Hypothesis 3, which predicted that social identity would mediate the relationship between economic outcomes and extrarole behavior. They do so by showing that economic outcome judgments are related to extrarole behavior (Model 4, .11, p .01, R2 3%); that economic outcome judgments are related to social identity (Model 5, .78, p .001, R2 61%); and that the relationship between economic outcomes and extrarole behavior becomes nonsignificant (Model 5, .19, ns) when accounting for social identity, which was itself significantly related to extrarole behavior (Model 5, .36, p .01), (Sobels z 3.07, p .01). These findings support Hypothesis 3 by showing that the relationship between economic outcome judgments and extrarole behavior is fully mediated by social identity. Overall model. Once again, we conducted an analysis that simultaneously examined the effects of procedural justice, economic outcomes, and social identity. The results (see Table 4, Model 6) indicate that social identity ( .32, p .05), procedural justice ( .17, p .01), and economic outcome judgments
As in Study 1, we also considered a model in which pride and respect were loaded onto a common latent factor (evaluative component), which in turn was loaded, with identification, onto a common latent factor (social identity). However, once again this alternate model calculated a negative variance for the residual of the evaluative component latent factor. The result was an inadmissible solution that indicated misspecification and lack of fit for this proposed model (Chen et al., 2001). These results, in coordination with the identical results from Study 1, provide converging evidence that identification, pride, and respect may be best conceptualized as operating at the same conceptual level.
3

TESTING AND EXTENDING THE GROUP ENGAGEMENT MODEL

455

Table 3 Study 2: Means, Standard Deviations, Coefficient Alphas, and Correlations Among Scales
Scale 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. Procedural justice Identification Pride Respect Evaluations of current pay Future economic outcomes Outcome fairness Extrarole behavior Gender Tenure, years Age, years M (SD) 3.24 (0.99) 3.10 (0.91) 3.57 (0.95) 5.22 (1.46) 3.17 (0.95) 3.07 (0.98) 2.89 (0.95) 5.49 (1.16) 4.49 (0.70) 43.87 (10.36) 1 .95 .48 .56 .51 .66 .58 .83 .16 .02 .03 .11 2 .84 .58 .46 .38 .43 .44 .15 .11 .08 .13 3 4 5 6 7 8 9 10 11

.90 .44 .53 .51 .50 .07 .06 .08 .12

.94 .39 .44 .43 .23 .05 .01 .10

.89 .66 .71 .07 .07 .06 .09

.72 .60 .06 .06 .09 .01

.90 .12 .09 .03 .05

.91 .09 .05 .00

.03 .09

.19

Note. n 831. Higher numbers indicate greater endorsement of the construct (e.g., more positive fairness, stronger identification). Numbers on the diagonal represent coefficient alphas for the scale. All correlations 0.07 are statistically significant at p .05.

( .29, p .05) were all related to extrarole behavior. Therefore, there is evidence of a suppression situation (Tzelgov & Henik, 1991) in this data, which we address in the Discussion section below. Together, they explain 8% of the variance in extrarole behavior. This analysis also indicates that both procedural justice ( .31, p .001) and economic outcomes ( .56, p .001) have independent associations with social identity. Additional analyses. As in Study 1, we conducted an additional set of analyses to compare the utility of our multidimensional conceptualization of social identity with a conceptualization focused exclusively on the cognitive component (i.e., identification). Therefore, we repeated the analyses presented in Table 4 (and described above) but replaced the social identity latent factor with the observed measure of identification. The results again supported the utility of our conceptualization. More specifically, with regard to Hypothesis 1, although identification was related to extrarole behavior ( .14, p .001, R2 3%), this relationship was weaker than that found for our conceptualization of extrarole behavior (R2 6%; presented in Table 4, Model 1). With regard to Hypothesis 2, identification partially mediated the relationship between procedural justice and extrarole behavior (in a model accounting for identification, .13, p .001, Sobels z 2.36, p .05, for procedural justice). This finding contrasts with the full mediation described above (presented in Table 4, Model 3). Further, the overall amount of variance explained by the model that included only identification (R2 4.9%) was smaller than that in the model that used our conceptualization of identity (R2 6%; see Table 4, Model 3). Finally, with regard to Hypothesis 3, although identification fully mediated the relationship between economic outcomes and extrarole behavior (i.e., in a model accounting for identification, .05, ns, Sobels z 2.69, p .01, for economic outcomes), this alternate model explained considerably less variance in extrarole behavior than did its counterpart (presented in Table 4, Model 5), which was based on a broader conceptualization of social identity (R2s 3% vs. 6%). These empirical findings support the value of conceptualizing social identity as consisting of cognitive and evaluative components, because this broader conceptualization of social identity is more strongly related to extrarole behavior and is a stronger mediator of the effects of procedural justice and economic outcomes.

Discussion
Study 2 corroborates the results of Study 1 and supports the three hypotheses examined in this research. As such, it helps support several key propositions of the group engagement model by demonstrating the important linkages between social identity and employee extrarole behavior. The results of Study 2 indicate that the social identity that employees form around their work organization is strongly related to their extrarole behavior. Beyond mere demonstration that social identity matters, Study 2 indicates that social identity can account for the impact of other factors (i.e., procedural justice and economic outcome judgments) that are themselves related to extrarole behavior. It is noteworthy that in Study 2, as in Study 1, the results exceeded our predictions insofar as they indicate fullrather than merely partialmediation by social identity of the relationships between procedural justice and economic outcomes on extrarole behavior. In addition, the results again indicate the value of a multidimensional conceptualization of social identity, rather than a conceptualization focused exclusively on the cognitive component. An intriguing result in Study 2 was the indication of a suppression situation (Tzelgov & Henik, 1991) in the analysis presented in Model 6. In particular, there was evidence of reciprocal suppression (Conger, 1974) between procedural justice and economic outcomes; although both of these variables had a nonsignificant effect in Models 3 and 5 (respectively), both of their regression coefficients turned to significance in a model in which they were both included. More specifically, this result indicates an instance of relative suppression (Conger, 1974; Tzelgov & Henik, 1991), since there was no evidence of suppression in either of the threevariate cases (i.e., Models 3 and 5) but there was clear evidence of suppression in the four-variate case (Model 6), as the regression coefficients for procedural justice and social identity were of greater magnitude in this more complex analysis. This result implies that procedural justice and economic outcomes each suppressed criterion-irrelevant variance in the other, and when accounting for the effects of social identity this led the remaining variance in each predictor to be expressed as statistically significant regression coefficients. The substantive interpretation of these findings is intriguing. First, there appears to be a positive association between procedural

456

Table 4 Study 2: Results of Structural Equation Models

Hypothesis 2 Model 2 b M (SE) b M (SE) b M (SE) b M (SE) Model 3 Model 4 Model 5 b Hypothesis 3 Overall model Model 6 M (SE)

Hypothesis 1

Model 1

Variable

M (SE)

0.18 .20 0.15 5% 0.81 63% 0 0 1 1 .00

2

(.05) (.04) 0.11 6% (.04) 0.79 (.05) 3% .17

.23 0.27 0.02 (.10) (.09)

0.23 0.01

0.44 0.27

(.14)

.36 .19 (.16) 6%

0.38 0.20 0.45

(.10) (.10) (.16) 8% 0.31 0.89 (.06) 61% 0.78 0.72 (.08) (.11) 71%

0.32 0.17 0.29

DV: Supervisor ratings of extrarole behavior Social identity Procedural justice Economic outcome judgments R2

6%

BLADER AND TYLER 0.31 0.56 4.1 2 .99 .99 .04 74.8 12 .97 .97 .08 118 16 .94 .94 .09

DV: Social identity Procedural justice Economic outcome judgments 2 R 27.7 4 .98 .98 .08

2 df Fit indices CFI IFI RMSEA .99 .99 .05

5.7 2

Note. n 831. All regression coefficients and R estimates are from analyses that include controls for the effects of age, gender, and tenure. In all models, there was a significant, positive effect for gender, such that women were rated as engaging in more extrarole behavior. There were no significant effects for the other control variables. DV dependent variable; CFI comparative fit index; IFI incremental fit index; RMSEA root-mean-square error of approximation. p .05. p .01. p .001. nonsignificant.

TESTING AND EXTENDING THE GROUP ENGAGEMENT MODEL

457

justice and extrarole behavior after accounting for both social identity and economic outcomes. This finding is quite consistent with the consensus in the justice literature that procedural justice serves numerous functions and is valued for a variety of reasons (Blader & Tyler, 2005); it appears that these alternate reasons that also explain why people care about procedural justice may exert an impact on extrarole behavior once the social identity and outcome relevance of procedural justice are accounted for (i.e., once the identity and instrumental models of procedural justice are captured in the data). Second, these results indicate that there is a negative association between economic outcomes and extrarole behavior after accounting for both social identity and procedural justice. That is, once the identity- and justice-related significance of economic outcomes have been accounted for, economic outcomes may actually have a negative association with extrarole behavior. These results may be another instantiation of the classic finding that extrinsic rewards can have detrimental effects on behaviors that may be otherwise intrinsically motivated (Deci, Koestner, & Ryan, 1999). Additional work should investigate and substantiate these results, particularly as these results did not emerge in Study 1. Overall, however, the consistency in findings with Study 1 with regard to our focal hypotheses is striking, and this pattern considerably increases our confidence in the robust nature of these results. This consistency is reinforced by the fact that Study 2 was conducted on a large and extremely heterogeneous sample of employees. Respondents were drawn from throughout the United States and worked in a wide variety of occupations, industries, and organizations. Moreover, Study 2 focused on a different social identity target than did Study 1 (organizational social identity, rather than work group social identity), yet the overall pattern of results from Study 2 replicated that from Study 1. This finding confirms that extrarole behavior may be a form of behavioral engagement that is linked to the multiple social identity groups relevant to organizational life.

tionmay act in coordination to predict extrarole behavior. Further, they show these potent effects of social identity among employees working in a wide variety of organizations, occupations, and regions. More important, going beyond mere demonstration that social identity matters, the results reveal that social identity is especially critical for understanding extrarole behavior, insofar as it explains the impact of two organizational conditions that have received a great deal of attention for their relation to extrarole behavior: perceptions of the procedural justice experienced at work and evaluations of the economic outcomes received as a result of group membership. Determining that social identity provides the mechanism by which these organizational conditions relate to employee behavior provides critical insight into understanding how, when, and why efforts to shape the context of employee work experiences may affect employee behavior. It also contributes to the growing acknowledgment within psychology that the collective selfthe aspect of the self most closely linked to social identity plays a fundamental role in shaping behavior (Blader, 2007a; Hogg, 2001). The results for economic outcomes in these studies may seem somewhat surprising. Indeed, it is traditionally believed that pay, benefits, expectations of promotion (and, thus, hopes for greater pay), and other similar factors shape behavior because of their economic or instrumental value. However, studies on the utility of these economic, outcome-based approaches have raised questions about their actual impact (e.g., Jenkins et al., 1998; Kohn, 1993; Pearce, 1987). The current findings may provide insight into the equivocal effect of outcome concerns on behavior. In particular, they suggest that a primary way in which economic outcomes may affect behavior is through their impact on social identity, which helps explain earlier observations that social identity can trigger many of the same effects as economic incentives (Dukerich et al., 2002).

Practical Implications General Discussion

These studies make an important contribution to research on the group engagement model and to our understanding of the motivational implications of social identity in organizational contexts. They do so by showing that social identity has strong associations with extrarole behavior, and by demonstrating that social identity explains the impact of two key facets of the experiences people have with their work organizations (i.e., the procedural justice they experience and the economic outcomes they receive as a result of group membership). These findings are consistent with the group engagement models emphasis on social identity as the primary basis of peoples engagement in their groups and on the importance of locating social identity within the landscape of other factors that shape engagement. These studies further contribute to research on social identity in organizations by (a) highlighting the value of a multidimensional conceptualization of identity that is rooted in principles of social identity theory and (b) applying this conceptualization to two social identity targets (i.e., the work group and the organization). They advance previous work with regard to each of these issues and suggest that the cognitive and evaluative components of social identitywith regard to both the work group and the organizaThese findings suggest that the basis of employees relationships with their organizations is primarily linked to the role the organization plays in determining how employees think and feel about themselves. Many important practical implications follow from this insight. First and foremost, our findings highlight the motivational power of getting employees to develop social identities that are grounded in their work organizations. This is critical, as the current results show that identity is central to employee behavior. Furthermore, the results indicate two important levers that organizations can use to encourage the development of social identity. First, organizations can operate in ways that employees regard as procedurally fair by instituting fair decision-making processes and extending fair quality of treatment (Blader & Tyler, 2003a, 2003b). Second, they can provide economic outcomes that employees regard favorably and that encourage employees to link their social identities to their organization. Application of these insights, however, is not as straightforward as it may seem. Employee judgments of procedural justice are susceptible to a wide array of subjective influences (Blader, 2007b; Blader & Bobocel, 2005). It is not enough to implement fair procedures: Organizations must ensure that employees are well aware that such procedures are in operation. Another challenge is

458

BLADER AND TYLER

to provide economic outcomes in ways that encourage the development of a strong social identity. To do so, organizations must dispense such outcomes in ways that lead employees to regard their relationship with their organization positively. When economic outcomes fail to prompt these positive evaluations e.g., when they are dispensed in an environment marked by threats or guided by obligationsit becomes less likely that economic outcomes will have a positive impact on social identity or behavior. When employees must force their organizations to provide them with better economic outcomesfor instance, by looking for alternate jobs, getting outside offers, or taking collective actionsit is unlikely that the resultant economic outcomes will foster social identity or that they will shape extrarole behaviors. By highlighting the mechanism through which procedural justice and economic outcomes shape employee behavior, these studies provide critical insight as to how organizations should implement fair procedures and allot economic outcomes. Of course, the practical implications of these studies are influenced by the boundary conditions of our findings. Among employees at lower socioeconomic levels, for whom resources are critical, the impact of economic outcomes on behavior may be more direct. Similarly, social identity may be less likely to shape behavior among employees who have tenuous links to the organization (e.g., temporary and contract workers) and among workers who do not share values with the organization or who lack a shared demographic group membership. Still, the results of the current and related (Tyler & Blader, 2000) research clearly suggest the potential value, under many circumstances, of social identity for understanding employee attachment to organizations. Moreover, they suggest that social identity might work to stimulate higher levels of extrarole behavior.

Limitations and Future Research

There are a number of limitations to the current studies that are important to note. For instance, a note of caution is necessary regarding the causal directions advocated in our conceptual model. Although our research avoids the problem of common method variance by linking employee judgments to supervisor ratings of behavior, it cannot provide definitive evidence of the causal relations we have proposed. Consistent with the psychology and management literatures, we advocate the perspective that behavior flows from identity. However, it is possible that supervisor judgments that an employee is high performing might lead employees to identify with the group through a social exchange process. Although this is plausible, there is no clear theoretical basis for expecting that employees would reciprocate their supervisors views that they are high performers by integrating the group into the self (i.e., it is not clear why employees would use such a psychological currency in their exchange). Ultimately, there is need for alternate research methods that can help further elucidate causal relationships and provide more definitive support for the causal relations we propose. In addition, although the measures used in these studies were based on those used in prior studies (Tyler & Blader, 2000), it would be worthwhile to subject these scales to further psychometric evaluation and to determine if the current pattern of results replicate with other scales that are often used in the organizational literature to measure related constructs.

Various issues raised by this research require further investigation. First, it is important to investigate other organizational conditions that shape extrarole behavioraside from procedural justice and economic outcomesand to better understand the relationship between social identity and these other factors. Indeed, a myriad of factors has been linked to employee behavior, and the current contribution is only one small step toward developing a map of the antecedents of behavior and defining the relationships among these antecedents. Second, now that we have demonstrated these effects on extrarole behavior, it is important to consider the utility of our framework for other dependent variables. Doing so would not simply represent a test of the robustness of our results but would provide a means of unpacking some of the elements of our model and better understanding their function. For instance, our results confirm that the evaluative component of social identity makes an important contribution to prediction of extrarole behavior, perhaps because positive evaluations of ones group set in motion a motivational thrust that plays itself out in discretionary behaviors that help the group. Positive views of ones group and ones place within the group likely fuel a drive to see the group succeed, and this likelihood makes the evaluative component quite important for predicting extrarole behavior. However, the evaluative component may be less important for outcomes such as conformity to group norms, which likely result from deindividuation processes associated with identification. More generally, testing the robustness of our results on other dependent variables has the potential both to expand the scope of our model and to help us examine the functions served by various components of our model. Now that we have demonstrated the value of our conceptual framework, there is important work to be done in unpacking our broad conceptual constructs to better understand the contexts and dependent variables that make particular elements of our model matter most. Third, and related to our prior point, a deeper understanding of the psychological mechanism by which social identity impacts behavior is needed. Does social identity shape extrarole behavior because it prompts an intrinsic concern with the groups welfare, leads to conformity to group norms, invokes reciprocity processes, or leads people to seek acceptance via behavior that benefits the group? Uncovering the mechanism involved is theoretically important, but it also has practical significance because it facilitates an understanding of the circumstances under which social identity will matter most, the people for whom social identity is most important, and the behaviors that social identity is most likely to provoke. Fourth, future research should explore respect, which is an important element of our conceptualization of social identity. As we note in the Introduction, respect is a more recent innovation to the social identity literature and was made relevant by the shift from using social identity theory to understand intergroup relations toward using social identity to understand intragroup behavior. The current studies assessed respect by asking respondents about the respect they receive from group authorities, based on the premise from earlier work (Lind & Tyler, 1988; Tyler & Lind, 1992) that respect from a group authority indicates whether other group members respect the person (Smith et al., 1998, p. 472). This approach is supported by Tyler et al. (1996), who showed that, across four different social contexts (work, family, university,

TESTING AND EXTENDING THE GROUP ENGAGEMENT MODEL

459

and judicial system), perceptions of respect from authorities were linked to feelings that other group members respect the person because authorities express the values and norms of the group (Tyler et al., 1996, p. 914). However, it is important that research explore more direct approaches to assessing intragroup status (e.g., how respected are you in your group). One concern is that people may make these assessments on the basis of salient interpersonal ties, which may be more closely linked to their relational identities than to their social identities. Future work should explore this issue and more generally should explore optimal ways of assessing intragroup status. Finally, research is needed to understand the linkages between social identity and other variables that assess attachment to organizations, such as affective commitment (Meyer & Allen, 1997). As do many other identity researchers (Ashforth & Mael, 1989; Ashmore et al., 2004; Cameron, 2004; Ellemers, Kortekaas, & Ouwerkerk, 1999; Haslam, 2004; Ouwerkerk et al., 1999; Pratt, 2000), we feel that social identity is fundamentally different from affective commitment because of its reliance on the self-concept for understanding peoples attachment to organizations. Unlike affective commitment, social identity focuses on the basis of attachment to the group and is much more than a simple barometer of attachment. Thus, we view social identity as operating on a more fundamental level than affective commitment (Haslam, 2004) and regard affective commitment as a potential outcome of social identity (Tyler & Blader, 2000). Nevertheless, additional work is needed on the dynamics between social identity and affective commitment.

Conclusion
We hope that the current studies aid in highlighting the value of the group engagement model for organizational research. As does the group engagement model itself, these studies make strong theoretical contributions to research on procedural justice, on why money and related economic issues matter, and on how to foster higher levels of extrarole behavior. The current research impacts these diverse research areas by highlighting one key integrating concept: social identity. Perhaps the most important contribution of these studies is their indication that researchers and practitioners alike have much to gain by focusing on social identity, since social identity represents a fundamental issue that drives group life. And when the fundamental concerns that drive peoples behavior in groups are understood, groups will be more likely to operate in ways that both motivate and satisfy their members.

References
Abrams, D., Ando, K., & Hinkle, S. (1998). Psychological attachment to the group: Cross-cultural differences in organizational identification and subjective norms as predictors of workers turnover intentions. Personality and Social Psychology Bulletin, 24, 10271039. Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 2, pp. 267299). New York: Academic Press. Allen, T. D., & Rush, M. C. (1998). The effects of organizational citizenship behavior on performance judgments: A field study and a laboratory experiment. Journal of Applied Psychology, 83, 247260. Ashforth, B. E., & Johnson, S. A. (2001). Which hat to wear? The relative salience of multiple identities in organizational contexts. In M. Hogg &

D. Terry (Eds.), Social identity processes in organizational contexts (pp. 31 48). Philadelphia: Psychology Press. Ashforth, B. E., & Mael, F. (1989). Social identity theory and the organization. Academy of Management Review, 14, 20 39. Ashmore, R. D., Deaux, K., & McLaughlin-Volpe, T. (2004). An organizing framework for collective identity: Articulation and significance of multidimensionality. Psychological Bulletin, 130, 80 114. Baron, R. M., & Kenny, D. A. (1986). The moderatormediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 11731182. Bartel, C. A. (2001). Social comparisons in boundary-spanning work: Effects of community outreach on members organizational identity and identification. Administrative Science Quarterly, 46, 379 413. Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motive. Psychological Bulletin, 117, 497529. Bergami, M., & Bagozzi, R. P. (2000). Self-categorization, affective commitment and group self-esteem as aspects of social identity in the organization. British Journal of Social Psychology, 39, 555577. Blader, S. L. (2007a). Lets not forget the me in team: Investigating the interface of individual and collective identity. In C. Bartel, S. Blader, & A. Wrzesniewski (Eds.), Identity and the modern organization (pp. 61 84). Mahwah, NJ: Erlbaum. Blader, S. L. (2007b). What determines peoples fairness judgments? Identification and outcomes influence procedural justice evaluations under uncertainty. Journal of Experimental Social Psychology, 43, 986 994. Blader, S. L. (2007c). What leads organization members to collectivize? Injustice and identification as precursors of union certification. Organization Science, 18, 108 126. Blader, S. L., & Bobocel, R. (2005). Wanting is believing: Understanding psychological processes in organizational justice by examining perceptions of fairness. In S. Gilliland, D. Steiner, D. Skarlicki, & K. van den Bos (Eds.), Research in social issues in management (pp. 330). Greenwich, CT: Information Age. Blader, S. L., van Knippenberg, D., & Sleebos, E. (2008, August). Tit-fortat and give-and-take: Communal and exchange dynamics in work organizations. Paper presented at the annual meeting of the Academy of Management, Anaheim, CA. Blader, S. L., & Tyler, T. R. (2003a). Advancing the assessment of procedural justice: What constitutes fairness in work settings? Human Resource Management Review, 13, 107126. Blader, S. L., & Tyler, T. R. (2003b). A four-component model of procedural justice: Defining the meaning of a fair process. Personality and Social Psychology Bulletin, 29, 747758. Blader, S. L., & Tyler, T. R. (2005). How can theories of organizational justice explain the effects of fairness? In J. Greenberg & J. A. Colquitt (Eds.), Handbook of organizational justice (pp. 329 354). Mahwah, NJ: Erlbaum. Blau, P. (1964). Exchange and power in social life. New York: Wiley. Boezeman, E. J., & Ellemers, N. (2007). Volunteering for charity: Pride, respect, and the commitment of volunteer workers. Journal of Applied Psychology, 92, 771785. Boezeman, E. J. & Ellemers, N. (in press). Pride and respect in volunteers organizational commitment. European Journal of Social Psychology. Bollen, K. (1990). Overall fit in covariance structure models: Two types of sample size effects. Psychological Bulletin, 107, 256 259. Brewer, M. B. (1979). In-group bias in the minimal intergroup situation: A cognitivemotivational analysis. Psychological Bulletin, 86, 307324. Brewer, M. B., & Gardner, W. (1996). Who is this we? Levels of collective identity and self-representations. Journal of Personality and Social Psychology, 71, 8393. Brockner, J., & Wiesenfeld, B. M. (1996). An integrative framework for

460

BLADER AND TYLER Perceived external prestige and internal respect: New insights into the organizational identification process. Human Relations, 59, 815 846. Furnham, A., & Argyle, M. (1998). The psychology of money. Florence, KY: Taylor & Frances/Routledge. Gintis, H., Bowles, S., Boyd, R., & Fehr, E. (2005). Moral sentiments and material interests: The foundations of cooperation in economic life. Cambridge, MA: MIT Press. Haslam, S. A. (2004). Psychology in organizations: The social identity approach. London: Sage. Haslam, S. A., & Ellemers, N. (2005). Social identity in industrial and organizational psychology: Concepts, controversies and contributions. In G. P. Hodgkinson (Ed.), International review of industrial and organizational psychology (Vol. 20, pp. 39 118). Chichester, England: Wiley. Haslam, S. A., Postmes, T., & Ellemers, N. (2003). More than a metaphor: Organizational identity makes organizational life possible. British Journal of Management, 14, 357369. Haslam, S. A., Powell, C., & Turner, J. C. (2000). Social identity, selfcategorization, and work motivation: Rethinking the contribution of the group to positive and sustainable organizational outcomes. Applied Psychology: An International Review, 49, 319 339. Haslam, S. A., & Reicher, S. (2006). Stressing the group: Social identity and the unfolding dynamics of responses to stress. Journal of Applied Psychology, 91, 10371052. Haworth, C. L., & Levy, P. E. (2001). The importance of instrumentality beliefs in the prediction of organizational citizenship behaviors. Journal of Vocational Behavior, 59, 64 75. Hinkle, S. W., & Brown, R. J. (1990). Intergroup comparisons and social identity: Some links and lacunae. In D. Abrams & M. A. Hogg (Eds.), Social identity theory (pp. 48 70). New York: Springer-Verlag. Hoffman, B. J., Blair, C. A., Meriac, J. P., & Woehr, D. J. (2007). Expanding the criterion domain? A quantitative review of the OCB literature. Journal of Applied Psychology, 92, 555566. Hogg, M. A. (2001). Social identity and the sovereignty of the group: A psychology of belonging. In C. Sedikides & M. Brewer (Eds.), Individual self, relational self, collective self (pp. 125146). New York: Psychology Press. Hogg, M. A., & Abrams, D. (1990). Social motivation, self-esteem and social identity. In D. Abrams & M. A. Hogg (Eds.), Social identity theory: Constructive and critical advances (pp. 28 47). New York: Springer-Verlag. Hogg, M. A., & Abrams, D. (1993). Towards a single process uncertaintyreduction model of social motivation in groups. In M. A. Hogg & D. Abrams (Eds.), Group motivation: Social psychological perspectives (pp. 173190). New York: Harvester Wheatsleaf. Homans, G. (1961). Social behavior: Its elementary forms. New York: Harcourt, Brace, World. Hu, L., & Bentler, P. M. (1995). Evaluating model fit. In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues and applications (pp. 76 99). Thousand Oaks, CA: Sage. Hui, C., Lam, S. S., & Law, K. K. (2000). Instrumental values of organizational citizenship behavior for promotion: A field quasi-experiment. Journal of Applied Psychology, 85, 822 828. Jenkins, G. D., Mitra, A., Gupta, N., & Shaw, J. D. (1998). Are financial incentives related to performance? Journal of Applied Psychology, 83, 777787. Jetten, J., Branscombe, N. R., & Spears, R. (2002). On being peripheral: Effects of identity insecurity on personal and collective self-esteem. European Journal of Social Psychology, 32, 105123, Johnson, M., Morgeson, F., Ilgen, D., Meyer, C., & Lloyd, J. (2006). Multiple professional identities: Examining differences in identification across work-related targets. Journal of Applied Psychology, 91, 498 506. Karau, S. J., & Williams, K., D. (1993). Social loafing: A meta-analytic

explaining reactions to decisions: Interactive effects of outcomes and procedures. Psychological Bulletin, 120, 189 208. Cameron, J. E. (2004). A three-factor model of social identity. Self & Identity, 3, 239 262. Chen, F., Bollen, K. A., Paxton, P., Curran, P. J., & Kirby, J. B. (2001). Improper solutions in structural equation models. Sociological Methods & Research, 29, 468 508. Cohen-Charash, Y., & Spector, P. E. (2001). The role of justice in organizations: A meta-analysis. Organizational Behavior and Human Decision Processes, 86, 278 321. Colquitt, J. A., Conlon, D. E., Wesson, M. J., Porter, C., & Ng, K. Y. (2001). Justice at the millennium: A meta-analytic review of 25 years of organizational justice research. Journal of Applied Psychology, 86, 425 445. Conger, A. J. (1974). A revised definition for suppressor variables: A guide to their identification and interpretation. Educational and Psychological Measurement, 34, 35 46. Corley, K., Harquail, C. V., Pratt, M. G., Glynn, M. A., Fiol, C. M., & Hatch, M. J. (2006). Guiding organizational identity through aged adolescence. Journal of Management Inquiry, 15, 8599. Crocker, J., & Major, B. (1989). Social stigma and self-esteem: The self-protective properties of stigma. Psychological Review, 96, 608 630. Curran, P. J., Bollen, K. A., Chen, F., Paxton, P., & Kirby, J. B. (2003). Finite sampling properties of the point estimates and confidence intervals of the RMSEA. Sociological Methods & Research, 32, 208 252. Davis, J. A. (1966). The campus as a frog pond: An application of the theory of relative deprivation to career decisions of college men. American Journal of Sociology, 72, 1731. Deci, E., Koestner, R., & Ryan, R. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125, 627 668. De Cremer, D. (2002). Respect and cooperation in social dilemmas: The importance of feeling included. Personality and Social Psychology Bulletin, 28, 13351341. Diamantopoulos, A., & Siguaw, J. A. (2000). Introducing LISREL. London: Sage. Dukerich, J. M., Golden, B. R., & Shortell, S. M. (2002). Beauty is in the eye of the beholder: The impact of organizational identification, identity, and image on the cooperative behaviors of physicians. Administrative Science Quarterly, 47, 507533. Dutton, J. E., Dukerich, J. M., & Harquail, C. V. (1994). Organizational images and member identification. Administrative Science Quarterly, 39, 239 263. Ellemers, N. (1993). The influence of socio-structural variables on identity management strategies. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 4, pp. 2557). London: Wiley. Ellemers, N., de Gilder, D., & Haslam, A. (2004). Motivating individuals and groups at work: A social identity perspective on leadership and group performance. Academy of Management Review, 29, 459 478. Ellemers, N., Kortekaas, P., & Ouwerkerk, J. W. (1999). Selfcategorization, commitment to the group, and group self-esteem as related but distinct aspects of social identity. European Journal of Social Psychology, 29, 371389. Ellemers, N., Spears, R., & Doosje, B. (1997). Sticking together or falling apart: In-group identification as a psychological determinant of group commitment versus individual mobility. Journal of Personality and Social Psychology, 72, 617 626. Folger, R. (1993). Justice, motivation and performance beyond role requirements. Employee Responsibilities and Rights Journal, 6, 239 248. Foreman, P., & Whetten, D. (2002). Members identification with multiple-identity organizations. Organization Science, 13, 618 635. Fuller, J. B., Hester, K., Barnett, T., Frey, L., Relyea, C., & Beu, D. (2006).

TESTING AND EXTENDING THE GROUP ENGAGEMENT MODEL review and theoretical integration. Journal of Personality and Social Psychology, 65, 681706. Kelly, C., & Kelly, J. (1994). Who gets involved in collective action? Human Relations, 47, 63 88. Kohn, A. (1993). Why incentive plans cannot work. Harvard Business Review, 71, 54 63. Kohn, A. (1999). Punished by rewards. New York: Houghton Mifflin. Kramer, R. M., Hanna, B. A., Su, S., & Wei, J. (2001). Collective identity, collective trust and social capital: Linking group identification and group cooperation. In M. E. Turner (Ed.), Groups at work: Theory and research (pp. 173196). Mahwah, NJ: Erlbaum. Lawler, E. E., III (1990). Strategic pay. San Francisco: Jossey-Bass. LePine, J. A., Erez, A., & Johnson, D. E. (2002). The nature and dimensionality of organizational citizenship behavior: A critical review and meta-analysis. Journal of Applied Psychology, 87, 52 65. Lind, E. A., & Tyler, T. R. (1988). The social psychology of procedural justice. New York: Plenum Press. MacKenzie, S. B., Podsakoff, P. M., & Fetter, R. (1991). Organizational citizenship behavior and objective productivity as determinants of managerial evaluations of salespersons performance. Organizational Behavior and Human Decision Processes, 50, 123150. Mael, F., & Ashforth, B. E. (1992). Alumni and their alma mater: A partial test of the reformulated model of organizational identification. Journal of Organizational Behavior, 13, 103123. March, J., & Simon, H. A. (1958). Organizations. New York: Wiley. McAllister, D., Kamdar, D., Morrison, E. W., & Turban, D. (2007). Disentangling role perceptions: How perceived role breadth, discretion, instrumentality and efficacy relate to helping and taking charge. Journal of Applied Psychology, 92, 1200 1211. McFarland, C., & Buehler, R. (1995). Collective self-esteem as a moderator of the frog-pond effect in reactions to performance feedback. Journal of Personality and Social Psychology, 68, 10551070. Meyer, J. P., & Allen, N. J. (1997). Commitment in the workplace. Thousand Oaks, CA: Sage. Milkovich, G. T., & Newman, J. M. (1999). Compensation (6th ed.). Boston: Irwin/McGraw-Hill. Miller, D. T., & Ratner, R. K. (1998). The disparity between the actual and assumed power of self-interest. Journal of Personality and Social Psychology, 74, 53 62. Mitchell, T. R., & Mickel, A. E. (1999). The meaning of money: An individual-difference perspective. Academy of Management Review, 24, 568 579. Moorman, R. H. (1991). The relationship between organizational justice and organizational citizenship behaviors. Journal of Applied Psychology, 76, 845 855. Naumann, S., & Bennett, N. (2000). A case for procedural justice climate: Development and test of a multilevel model. Academy of Management Journal, 43, 881 889. Noel, J. G., Wann, D. L., & Branscombe, N. R. (1995). Peripheral ingroup membership status and public negativity toward outgroups. Journal of Personality and Social Psychology, 68, 127137. Olkkonen, M. E., & Lipponen, J. (2006). Relationships between organizational justice, identification with organization and work unit, and grouprelated outcomes. Organizational Behavior and Human Decision Processes, 100, 202215. Opsahl, R. L., & Dunnette, M. D. (1966). Role of financial compensation in industrial motivation. Psychological Bulletin, 66, 94 118. Organ, D. W., & Ryan, K. (1995). A meta-analytic review of attitudinal and dispositional predictors of organizational citizenship behavior. Personnel Psychology, 48, 775 802. Orr, J. M., Sackett, P. R., & Mercer, M. (1989). The role of prescribed and nonprescribed behaviors in estimating the dollar value of performance. Journal of Applied Psychology, 74, 34 40. Ouwerkerk, J. W., Ellemers, N., & de Gilder, D. (1999). Group commit-

461

ment and individual effort in experimental and organizational contexts. In N. Ellemers & R. Spears (Eds.), Social identity: Context, commitment, content (pp. 184 204). Oxford, England: Blackwell Science. Pearce, J. L. (1987). Why merit pay doesnt work: Implications from organizational theory. In D. B. Balkin & L. R. Gomez-Meija (Eds.), New perspectives on compensation (pp. 169 178). Englewood Cliffs, NJ: Prentice-Hall. Podsakoff, P. M., MacKenzie, S. B., Paine, J. B., & Bachrach, D. G. (2000). Organizational citizenship behaviors: A critical review of the theoretical and empirical literature and suggestions for future research. Journal of Management, 26, 513563. Porter, L. W., Bigley, G., & Steers, R. M. (1996). Motivation and work behavior. New York: McGraw-Hill. Pratt, M. G. (2000). The good, the bad, and the ambivalent: Managing identification among Amway distributors. Administrative Science Quarterly, 45, 456 493. Pratt, M. G., & Foreman, P. (2000). Classifying managerial responses to multiple organizational identities. Academy of Management Review, 25, 18 42. Rynes, S. L., Gerhart, B., & Parks, L. (2005). Personnel psychology: Performance evaluation and pay for performance. Annual Review of Psychology, 56, 571 600. Schumacker, R. E., & Lomax, R. G. (2004). A beginners guide to structural equation modeling (2nd ed.). London: Erlbaum. Simon, B., & Sturmer, S. (2003). Respect for group members: Intragroup determinants of collective identification and group-serving behavior. Personality and Social Psychology Bulletin, 29, 183193. Sleebos, E., Ellemers, N., & de Gilder, D. (2006). The paradox of the disrespected: Disrespected group members engagement in group-serving efforts. Journal of Experimental Social Psychology, 42, 413 427. Smith, H. J., & Tyler, T. R. (1997). Justice and power: When will justice concerns encourage the advantages to support policies which redistribute economic resources and the disadvantaged to willingly obey the law? European Journal of Social Psychology, 26, 171200. Smith, H. J., Tyler, T. R., Huo, Y. J., Ortiz, D. J., & Lind, E. A. (1998). The self-relevant implications of the group-value model: Group membership, self-worth, and treatment quality. Journal of Experimental Social Psychology, 34, 470 493. Sparrowe, R. T., Soetjipto, B. W., & Kraimer, M. L. (2006). Do leaders influence tactics relate to members helping behavior? It depends on the quality of the relationship. Academy of Management Journal, 49, 1194 1208. Spears, R., Ellemers, N., Doosje, B., & Branscombe, N. (2006). The individual within the group: Respect! In T. Postmes & J. Jetten (Eds.), Individuality and the group: Advances in social identity (pp. 175195). Thousand Oaks, CA: Sage. Sturmer, S., Simon, B., & Loewy, M. (2008). Intraorganizational respect and organizational participation: The mediating role of collective identity. Group Processes & Intergroup Relations, 11, 520. Tajfel, H. (1978). Social categorization, social identity and social comparison. In H. Tajfel (Ed.), Differentiation between social groups: Studies in the social psychology of intergroup relations (pp. 7798). London: Academic Press. Tajfel, H. (1982). Social psychology of intergroup relations. Annual Review of Psychology, 33, 139. Tajfel, H., Billig, M. G., Bundy, R. P., & Flament, C. (1971). Social categorization and intergroup behavior. European Journal of Social Psychology, 1, 149 178. Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33 47). Monterey, CA: Brooks. Thibaut, J., & Kelley, H. H. (1959). The social psychology of groups. New York: Wiley. Turner, J. C. (1985). Social categorization and the self-concept: A social

462

BLADER AND TYLER van den Bos, K., & Lind, E. A. (2002). Uncertainty management by means of fairness judgments. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 34, pp. 1 60). New York: Academic Press. van Dick, R., Wagner, U., Stellmacher, J., & Christ, O. (2004). The utility of a broader conceptualization of organizational identification: Which aspects really matter? Journal of Occupational and Organizational Psychology, 77, 171191. van Knippenberg, D. (2000). Work motivation and performance: A social identity perspective. Applied Psychology: An International Review, 49, 357371. van Knippenberg, D., & Ellemers, N. (2003). Social identity and group performance: Identification as the key to group-oriented effort. In S. A. Haslam, D. van Knippenberg, M. J. Platow, & N. Ellemers (Eds.), Social identity at work: Developing theory for organizational practice (pp. 29 42). Philadelphia: Psychology Press. van Knippenberg, D., & van Schie, E. C. M. (2000). Foci and correlates of organizational identification. Journal of Occupational and Organizational Psychology, 73, 137147. Worchel, S., Rothgerber, H., Day, E. A., Hart, D., & Butemeyer, J. (1998). Social identity and individual productivity within groups. British Journal of Social Psychology, 37, 389 413. Zdaniuk, B., & Levine, J. M. (2001). Group loyalty: Impact of members identification and contributions. Journal of Experimental Social Psychology, 37, 502509.

cognitive theory of group behavior. In E. J. Lawler (Ed.), Advances in group processes. Greenwich, CT: JAI Press. Tyler, T. R. (2000). Social justice: Outcome and procedure. International Journal of Psychology, 35, 117125. Tyler, T. R., & Blader, S. L. (2000). Cooperation in groups: Procedural justice, social identity, and behavioral engagement. Philadelphia: Psychology Press. Tyler, T. R., & Blader, S. L. (2001). Identity and cooperative behavior in groups. Group Processes & Intergroup Relations, 4, 207226. Tyler, T. R., & Blader, S. L. (2002). Autonomous vs. comparative status: Must we be better than others to feel good about ourselves? Organizational Behavior & Human Decision Processes, 89, 813 838. Tyler, T. R., & Blader, S. L. (2003). The group engagement model: Procedural justice, social identity, and cooperative behavior. Personality and Social Psychology Review, 7, 349 361. Tyler, T. R., Degoey, P., & Smith, H. J. (1996). Understanding why the justice of group procedures matters: A test of the psychological dynamics of the group-value model. Journal of Personality and Social Psychology, 70, 913930. Tyler, T. R., & Lind, E. A. (1992). A relational model of authority in groups. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 25, pp. 115191). New York: Academic Press. Tzelgov, J., & Henik, A. (1991). Suppression situations in psychological research: Definitions, implications, and applications. Psychological Bulletin, 109, 524 536.

Appendix Measures, Studies 1 and 2 Study 1

All items below are based on Tyler & Blader (2000, 2001), unless otherwise noted. strongly agree) with the following statements about their relationship with their work group: (a) When I talk about my work group, I usually say we rather than they; (b) I have a sense that I personally belong, (c) I feel like an important part of my work group; (d) I am close to other people in my work group; and (e) I feel like part of the family where I work. Pride. Pride in ones work group was measured with a fouritem measure (based on Tyler et al., 1996). Those items asked respondents to indicate the extent to which they agreed with the following statements about their particular work group (1 strongly disagree to 6 strongly agree): (a) I feel proud to be a part of my work group; (b) My work group is highly respected within the company; (c) My work group is one of the most desirable within ___; (d) I work in one of the best work groups in ___. Respect. Respect was measured with a six-item scale (based on Tyler et al., 1996). Items asked respondents to indicate the extent to which they agreed with the following statements (1 strongly disagree to 6 strongly agree): Would you agree or disagree that the manager of your work group (a) values you as a member of your work group, (b) respects your work-related ideas, (c) values what you contribute at work, (d) respects the work you do, (e) appreciates your unique contributions on the job, and (f) approves of how you do your job.

Procedural Justice
Five items assessed evaluations by respondents of the procedural justice they had experienced in their work group; these items were a direct measure of procedural fairness perceptions (Colquitt et al., 2001). Instructions preceding the items asked respondents to answer the following questions with respect to their current work group and the job they held within that work group at their organization: (a) How would you rate the overall fairness with which issues and decisions that come up at work are handled? (1 not fair at all to 6 very fair); (b) Overall, how fair would you say decisions and processes are where you work? (1 not fair at all to 6 very fair); (c) How often do you feel that decisions are made in fair ways at your job? (1 rarely to 6 very often); (d) Is there a general sense among employees that things are handled in fair ways where you work? (1 not at all to 6 definitely); (e) How much of an effort is made to be fair to employees when decisions are being made? (1 none to 6 a lot).

Social Identity
Identification. Identification with ones work group was measured with five items. Those items asked respondents to indicate the extent to which they agreed (1 strongly disagree to 6

Outcome Evaluations
As with the procedural justice scale, instructions preceding these measures asked respondents to answer the following questions

TESTING AND EXTENDING THE GROUP ENGAGEMENT MODEL

463

with respect to their current work group and the job they currently held within that work group at their organization: Evaluations of pay. Respondent evaluations of their pay were assessed with a four-item measure. Those items asked respondents to indicate the extent to which they agreed with the following statements (1 strongly disagree to 6 strongly agree): (a) Overall, I receive excellent pay at ___; (b) I am satisfied with my pay; (c) I am well compensated for the work I do; (d) Overall, I receive excellent financial benefits from working at ___. Expectations of future outcomes. Three items assessed respondents expectations of future outcomes. Those items asked respondents to indicate the extent to which they agreed with the following statements (1 strongly disagree to 6 strongly agree): (a) If I wish to, I can have a secure job well into the future; (b) I have good opportunities for promotion; (d) I think I will do well financially if I stay at my job. Outcome fairness. Three items assessed respondents evaluations of the outcome fairness (i.e., distributive justice) they had experienced. Those items asked respondents (a) How fairly are resources (e.g., salary, bonuses, etc.) allocated among employees where you work? (1 not at all fairly to 6 very fairly); (b) Overall, how fair is the salary you receive? (1 not fair at all to 6 very fair); and (c) Would you say that there is an emphasis where you work on distributing things fairly? (1 not at all to 6 definitely). Incentives. We conceptualized incentives as the link between pay and performance, or the perceived contingency between work performance and the benefits/rewards they received. Items asked respondents to indicate the degree to which they agreed with the following statements (1 strongly disagree to 6 strongly agree, except as noted): (a) The benefits I get are directly linked to the effort I put into my job; (b) If you do your job well, how much does that improve your pay and benefits? (1 very little to 6 a lot); (c) Rewards are distributed among employees according to the effort that they put into their jobs; (d) Employees are rewarded based on how well they do their jobs, not how well they maneuver through office politics; (e) Being good at ones job is more important to ones success than looking good; (f) Management is not impressed by employees who only try to make it seem like they are working hard.

pects of procedures (Blader & Tyler, 2003a). Participants indicated the extent to which they would agree or disagree with the following statements (1 strongly disagree to 5 strongly agree): (a) Decisions that affect me are usually made in fair ways at my company; (b) My company puts an effort into making decisions that affect me in fair ways; (c) Decisions are usually made in fair ways at my company; (d) My company puts an effort into making decisions in fair ways; (e) Most of the issues involving me are handled in fair ways where I work; (f) Overall, I am fairly treated where I work; (g) The procedures used to decide how much I am paid are fair; (h) Overall, people are treated fairly where I work; (h) The procedures used to make decisions about pay and promotions are fair; (i) There is a general sense among employees that things are handled in fair ways where I work.

Social Identity
Identification. Identification with ones work organization was measured with five items (based on Mael & Ashforth, 1992). Those items asked respondents to indicate the extent to which they agreed with the following statements (1 strongly disagree to 5 strongly agree): (a) Working at my company is important to the way that I think of myself as a person; (b) When someone praises the accomplishments of my company, it feels like a personal compliment to me; (c) When someone from outside criticizes my company, it feels like a personal insult; (d) The place I work says a lot about who I am as a person; and (e) I think I am similar to the people who work at my company. Pride. Pride in ones work organization was measured with a five-item measure (based on Tyler et al., 1996): (a) My company is one of the best companies in its field; (b) People are impressed when I tell them where I work; (c) My company is well respected in its field; (d) I think that where I work reflects well on me; (e) I am proud to tell others where I work. Respect. Respect was measured with a six-item scale (based on Tyler et al., 1996). We asked each respondent to assess (1 not at all to 7 a great deal) how much managers (a) respect the work you do, (b) respect your work-related ideas, (c) think highly of the quality of your work, (d) appreciate your unique contributions on the job, (e) think that you have valuable insights and ideas, (f) think it would be difficult to replace you.

Extrarole Behavior (Supervisor Rated)

Extrarole. Three items asked supervisors to specifically evaluate employees extrarole performance. In particular, supervisors were asked to think of things that this employee does voluntarily that are not explicitly part of his or her job description. How often does this employee (a) volunteer to do helpful things that are not required by the job description, (b) voluntarily assist you without being asked to do so, (c) do things that help without having to be asked to do so (1 never to 6 very often). Note. The company name appeared in lieu of the blank spaces (___) that appear in some of the items below; it has been removed to maintain the confidentiality of the organization.

Outcome Evaluations
Evaluations of pay. Respondent evaluations of pay were assessed with a five-item measure. Respondents answered the following questions (1 very unfavorable to 5 very favorable): (a) How favorable are the outcomes and resources you receive from the policies of your work organization? (b) How favorable are the benefits that you receive? (c) Compared to what you expect, how favorable is your compensation for the work you do? (d) How favorable is your pay? (e) How favorable are the resources that you are given to do your job? Expectations of future outcomes. Three items assessed respondents expectations of future outcomes: (a) How favorable will your long-term financial benefits be if you stay with your com-

Study 2 Procedural Justice

We used 10 items to measure perceptions of procedural justice with regard to both decision-making and treatment as-

464

BLADER AND TYLER

pany? (b) How favorable are your opportunities for promotion? (c) How favorable is the job security you have where you work? (1 very unfavorable to 5 very favorable). Outcome fairness. Eight items assessed respondents evaluations of the distributive justice they experienced at their work organization. Respondents indicated their agreement with the following items (1 strongly disagree to 5 strongly agree): (a) I am fairly paid at work; (b) I receive the opportunities I deserve; (c) The resources I receive are linked to how well I do my job; (d) My company places an emphasis on rewarding me fairly; (e) My company wants me to receive the pay and benefits I deserve; (f) In general, resources are fairly allocated among employees at my organization; (g) Employees receive the opportunities they deserve at my company; (h) The resources people receive are linked to how well they do their jobs.

Extrarole Behavior (Supervisor Rated)

Seven items asked supervisors to evaluate employees extrarole behavior. Supervisors were asked to rate on a scale from 1 (never) to 7 (always) how often employees do the following: (a) Volunteer to do things that are not required in order to help the organization; (b) Volunteer to help orient new employees; (c) Help others with work-related problems; (d) Volunteer to help others when they have heavy workloads; (e) Put an extra effort into doing their job well, beyond what is normally expected; (f) Share their knowledge with others even when they will not receive credit; (g) Work extra hours even when they will not receive credit for doing so. Received March 12, 2007 Revision received April 7, 2008 Accepted June 26, 2008