Structuring Reality through the Faultlines Lens: The Effects of Structure, Fairness, and Status Conflict on the Activated Faultlines-performance Relationship

We investigate how activated team faultlines represent an informal sensemaking structure through which teammates interpret their social reality. Constructed from inter-subgroup comparisons, activated faultlines likely result in status perceptions that are ambiguous or illegitimate. Thus, activated faultlines threaten the justice climate within the team, which drives status conflict, impairing team performance. We explore the effects of team structure clarity in providing certainty or legitimacy around status and structure, ameliorating the negative effect of activated faultlines on team justice climate. We tested our model using a multi-source (three sources), multi-wave cross-lagged design (four waves) on a sample of 271 employees and 41 leaders in 41 teams. We found that the negative relationship between activated faultlines and team performance was mediated by the team justice climate—status conflict causal chain. We also found that team structure clarity reduced activated faultlines negative effect on team justice climate. The results highlight the value of using team faultlines, the social identity approach, and justice theories to understand how diverse teams interpret their social reality that influences their performance. Furthermore, our research provides practical guidance to managers in building clear team structures that minimize the harmful effects of activated faultlines on justice perceptions and team performance. Academy of Management Journal

We investigate how activated team faultlines represent an informal sensemaking structure through which teammates interpret their social reality. Constructed from inter-subgroup comparisons, activated faultlines likely result in status perceptions that are ambiguous or illegitimate. Thus, activated faultlines threaten the justice climate within the team, which drives status conflict, impairing team performance. We explore the effects of team structure clarity in providing certainty or legitimacy around status and structure, ameliorating the negative effect of activated faultlines on team justice climate. We tested our model using a multi-source (three sources), multi-wave cross-lagged design (four waves) on a sample of 271 employees and 41 leaders in 41 teams. We found that the negative relationship between activated faultlines and team performance was mediated by the team justice climate-status conflict causal chain. We also found that team structure clarity reduced activated faultlines negative effect on team justice climate. The results highlight the value of using team faultlines, the social identity approach, and justice theories to understand how diverse teams interpret their social reality that influences their performance. Furthermore, our research provides practical guidance to managers in building clear team structures that minimize the harmful effects of activated faultlines on justice perceptions and team performance.
conflict; this view of team faultlines helps reconcile some of the extant disparate findings in the field.
The current state of research synthesized in a quantitative review (Thatcher & Patel, 2012), shows that team faultlines create fractures within teams that result in conflict, inhibit team information elaboration and reduce key outcomes, such as decision quality, accuracy and performance. Although the majority of studies associate negative outcomes with dormant faultlines, there are inconsistencies in the findings; some studies show that dormant faultlines can result in positive group processes and performance outcomes under certain contexts (Bezrukova, Jehn, Zanutto, & Thatcher, 2009;Cooper, Patel, & Thatcher, 2013;Lau & Murnighan, 2005;Meyer, Shemla, & Schermuly, 2011;Xie, Wang, & Qi, 2015). We argue that such inconsistencies may be due to at least two issues present in existing research, which hinder a complete understanding of team faultlines and their effects on teams.
First, faultline researchers assume that dormant faultlines are perceived and thus will have similar (albeit less strong) effects to activated faultlines (defined as dormant faultlines that are also perceived, Lau & Murnighan, 1998), as borne out by Thatcher & Patel (2012).
However, from a conceptual perspective, dormant faultlines may exist that do not become activated. Hence, it is important to investigate the true impact that activated faultlines have on organizational team processes and outcomes. Relatedly, in studies on both dormant and perceived faultlines there is an assumption that all attributes underlying the faultlines are equally important in determining their strength. However, recent research has shown this assumption to be erroneous (e.g., Choi & Sy, 2010, Chung et al., 2015Jehn & Bezrukova, 2010). It is necessary to consider what attributes team members perceive as relevant within their working context (i.e., which attributes are relevant for faultline activation within their team) to accurately assess how activated faultlines affect team processes and performance. Accordingly, our work extends current thinking in the study of faultlines by explicitly and simultaneously examining both the dormant and perceptual conceptualizations of the faultlines construct.
Second, faultline scholars have mainly focused on conceptualizing faultlines as a compositional feature of the team. Nevertheless, such a perspective restricts the view that teams shape the content and meaning of their diversity. We know from recent research that the emergent processes of salience are critical to explaining the effects of diversity on team process and outcomes (Joshi & Neely, 2018). Building on the social identity approach (incorporating self-categorization and social identity theories; Haslam, 2001, Chattopadhyay, Tluchowska, & George, 2004a, we conceptualize activated faultlines as an informal sensemaking structure through which team members interpret their social reality. Employees composing teams interpret their social reality and develop perceptions based on demographic similarities/differences as well as status similarities/differences across subgroups (Chattopadhyay, Finn, & Ashkanasy, 2010). In the absence of a legitimating organizational mechanism (e.g., a formal structure), such perceptions drive intergroup comparisons yielding an atmosphere of competition (Sherif, 1966;Correll & Park, 2005) that may be associated with perceptions of inequality and unfairness between faultline-based subgroups (Mannix, 1993;Sachdev & Bourhis, 1991). To explain these relationships, we incorporate the aspects of status and legitimacy that are relevant to the social identity approach (Chattopadhyay et al., 2004a;Tajfel & Turner, 1986), rather than incorporating distal status and legitimacy theories 1 (Magee & Galinsky, 2008;Suddaby, Bitektine & Haak, 2017). In doing so, we maintain conceptual coherence as the faultlines literature is strongly rooted in the social identity approach (Thatcher & Patel, 2012).
Overall, our study makes several contributions. First, rather than conceptualizing activated faultlines solely as a more fine-grained diversity measure that emphasizes differences in demographic attributes, we view activated faultlines as a sensemaking structure that conveys information about reciprocal influence, status and power (Brown, Lawrence, & Robinson, 2005) and is used by team members to order and understand their social environment. This conceptualization moves us away from viewing team members as passive actors guided by composition structures and moves us toward accepting that individuals use team attributes to create a reality that guides team interactions.
As a second contribution, we explain how faultlines as a sensemaking structure trigger team members to be conscious of issues associated with subgroup fairness and inter-subgroup status. Perceptions of unfairness and inter-subgroup differences lead teams to experience status conflict (i.e., the attempt to defend or elevate one's own [subgroup's] relative status - Bendersky & Hays, 2012). We show how organizations can legitimize inter-subgroup differences and promote a sense of fairness among employees by using formal structures that communicate clear roles (Chattopadhyay et al., 2004a). We do this by developing a model linking activated faultlines, team justice climate (the shared perceptions of the extent to which team members treat each other fairly -Cropanzano, Li, & James, 2007), status conflict, team structure clarity, and team performance. Using the social identity approach, we integrate fairness perceptions and status conflict into the faultlines literature contributing to a more nuanced view of how activated faultlines impact organizational teams (Bendersky & Hays, 2012;Chattopadhyay, George & Shulman, 2008;Jost & Banaji, 1994); this also directly addresses Thatcher & Patel's (2012) call for research to examine the link between faultlines and status.
Third, we provide evidence that leaders can reduce the negative impact of activated faultlines on fairness perceptions by introducing a clear structure that legitimates and/or clarifies inter-subgroup status differences, contributing to effective faultlines management.
Our fourth contribution is the development of a new faultlines measure that incorporates both the dormant and perceived facets of faultlines. This new measure complements existing measures and stimulates new empirical opportunities for faultline researchers. Our hypotheses are tested using longitudinal data over four time periods on a sample of 271 subordinates and 41 supervisors composing 41 teams from a healthcare organization in Spain.

Theoretical Background and Hypotheses
Previous research has shown that faultline-based alignments create fractures within teams that inhibit team performance through group processes (such as increased team conflict or reduced elaboration of task relevant information; Rico, Sánchez-Manzanares, Antino, & Lau, 2012), and that such effects are stronger when faultlines are activated (Thatcher & Patel, 2012). These empirical results are mainly explained by the social identity approach (Lau & Murnighan, 1998;Kunze & Bruch, 2010, Thatcher, Jehn, & Zanutto, 2003, which focuses on how attribute salience results in a categorization process yielding both in-group and outgroup perceptions (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987;Shemla, Meyer, Greer, & Jehn, 2016). More specifically, categorization processes can be explained through the saliency of diversity dimensions, based on comparative and normative fit (i.e., the degree to which the diversity dimensions are related to actual differences between members and how meaningful they are for the group members within their group context, respectively) and cognitive accessibility (i.e., how easily members perceive the differences, and how quickly they come to mind). When these three components are present, categorization induces subgroup formation and inter-subgroup bias processes that disrupt inter-subgroup relations (Turner et al., 1987;van Knippenberg et al., 2004). Based on these theories, Lau and Murnighan (1998) proposed that dormant faultlines become activated when team members perceive faultline-based subgroups. In their original theorization, the negative effects of faultlines on team processes and outcomes are associated with their activation (Lau & Murnighan, 1998).
Notwithstanding the original theorization, the majority of research on faultlines has investigated dormant faultlines (Meyer et al., 2014;Thatcher & Patel, 2012); unfortunately, this research is unable to conclusively state what attributes contribute to a team's activated faultline (Minichilli, Corbetta, & MacMillan, 2010). The results attributed to dormant faultlines may or may not be caused by unmeasured faultline activation. Despite the importance of investigating activated faultlines in organizations, and some recent efforts in studying perceived diversity in the field (Mayo, van Knippenberg, Guilen, & Firfiray, 2016), there is a dearth of research on this topic due to the difficulty of obtaining such data (for a review, see Meyer et al., 2014). Accordingly, and consistent with the original conceptualization of faultlines, we submit that simultaneously considering both the dormant and the perceived aspects of faultlines is essential to wholly understanding their effects on teams.
Apart from the operationalization of activated faultlines, what is missing in the current conversation is how the meanings associated with subgroup configurations caused by activated faultlines influence perceptions of fairness and inter-subgroup status. Faultlines represent a social configuration of team members that simultaneously considers multiple diversity dimensions, where the same team members may be aligned on one dimension (sharing a specific attribute that results in a subgroup) and crossed on another dimension (sharing a specific attribute with members of other subgroups; Lau & Murnighan, 1998).
Thus, it is more complex for team members to interpret their reality using a faultlines lens (i.e., considering several dimensions at the same time) than focusing on a single diversity dimensions separately (e.g., only gender or only age). For example, in a team composed of long-tenured female engineers and administrative assistants, and newcomer male engineers and administrative assistants, team members differ on the attributes of tenure, gender, and job function. Some of the attributes are associated with high status (e.g., long-tenured, male, the engineer job function) and some of the attributes are associated with low status (e.g., newcomer, female, the administrative assistant job function) (Howell, Harrison, Burris & Detert, 2015;Phillips, Duguid, Thomas-Hunt, & Uparna, 2013). If the team members perceive their reality only on job function (single diversity dimension) then it is very clear that the engineers would represent a high-status subgroup and the assistants would represent a low-status subgroup. However, if team members perceive their reality using a faultline lens, things are more complicated as the faultline creates subgroups of team members who contain both high and low status dimensions (for example, long-tenured female engineers have both high-status and low-status attributes simultaneously). Subgroups of members containing both high-and low-status attributes may perceive unfairness because of their gender, age, or functional background. In this scenario, potential activated faultlines produce an inherently ambiguous inter-subgroup status structure; this sensemaking view of faultlines has not yet been accounted for in extant faultlines research.
In accordance with the above rationale, we argue that activated faultlines are an informal sensemaking structure that team members use to interpret the social reality derived from the interactions between different subgroups. These interpretations create mental representations of reciprocal influence, inter-subgroup status, and power that are based on perceptions of interpersonal differences (Brown, Lawrence, & Robinson, 2005;Fiol, O'Conner, & Aguinis, 2001;Harrison, Price, Gavin, & Florey, 2002). Due to the complex nature of faultlines these mental representations are often ambiguous with regard to the status hierarchy. Furthermore, with the drive to create a positive sense of self through gaining status, faultlines may also result in perceptions of illegitimate inter-subgroup status distribution, as explained by the social identity approach (Hogg & Terry, 2000;Chattopadhyay et al., 2004a, Chattopadhyay, George, & Lawrence, 2004b. In short, the social identity approach offers theoretical grounding to explain how faultlines provide a social sensemaking structure, but the inter-subgroup status underpinnings of such a structure are likely to be perceived as ambiguous and/or illegitimate. Thus, advancing extant theory, we predict that activated faultlines, as an informal sensemaking structure, drives unfairness perceptions and conflicts over status. We explain these relationships in the next subsections.

The Relationship between Faultlines, Status Conflict, and Team Performance
Since its inception, there have been conceptual arguments (Lau & Murnighan, 1998) and empirical evidence supporting the relationship between dormant faultlines and relationship conflict ["interpersonal incompatibilities among group members, which typically includes tension, animosity, and annoyance among members within a group"; Jehn, 1995, p. 258] and/or task conflict ["disagreements among group members about the content of the tasks being performed, including differences in viewpoints, ideas, and opinions"; Jehn, 1995, p. 258] (e.g., Choi & Sy, 2010;Thatcher et al., 2003;Molleman, 2005). Furthermore, Jehn & Bezrukova (2010) found that perceived faultlines in student groups led to an increase in relationship conflict. Although studying the relationship between dormant or perceived faultlines and task and relationship conflict has been fruitful, we posit that activated faultline teams are especially susceptible to conflict associated with status (attempts to defend or elevate one's own [subgroup's] relative status -Bendersky & Hays, 2012). Our investigation of status conflict as a specific mechanism related to status as a social resource and as an identity threat, explains how faultlines influence performance, independent of the interpersonal aspects associated with relationship-and task-related conflict (Bendersky & Hays, 2012). To back our assertion, we first characterize status conflict within the intergroup dynamics that the social identity approach demarcates, and then explain why it is likely to mediate the relationship between activated faultlines and team performance.
According to the social identity approach, individuals and (sub)groups proactively pursue a positive social identity (as a fundamental human motive, Tajfel & Turner, 1986;Chattopadhyay et al., 2004b). A basic strategy to achieve a positive social identity is through social competition, embodied as status conflict. Besides social competition, subgroups can aim for a positive social identity through other strategies, such as social mobility or social creativity (Chattopadhyay et al., 2004a); however, these last two strategies are unlikely to work in activated faultline teams as we explain next.
Through social mobility team members dissociate from a low status subgroup to gain membership into a higher status subgroup, such as by assimilating high-status subgroup norms (Chattopadhyay et al., 2004a). Social mobility is unlikely to be successful in activated faultline teams as it concurrently requires that the high-status subgroup members accept the low status subgroup members as equals, and the low status members acquire the norms and values of the high-status subgroup. Additionally, this strategy calls for permeable boundaries between social categories; this is unlikely in faultline-based teams when activated faultlines have resulted in subgroup identification and consequent inter(sub)group biased interactions (Carton & Cummings, 2012).
Through a social creativity strategy, low-status team members move up in the status hierarchy by generating alternative ingroup-outgroup comparisons on dimensions that are more positive for their social identity (Tajfel & Turner, 1986), and are not directly related to status in the current context (e.g., athletic or artistic abilities, Oyserman & Harrison, 1998). In faultline-based teams social creativity is unlikely to work because this strategy is a collective response that requires coordination within the entire low-status subgroup. Additionally, when team members differ simultaneously on several dimensions, it complicates efforts for an entire subgroup to create a completely new, unrelated identity.
Thus, although social creativity and social mobility are common strategies for individual employees pursuing status change within organizations (Chattopadhyay et al., 2010), we consider social competition to be the most viable strategy for team members wishing to alter their status within an activated faultline-based team. Importantly, social competition takes into account the desires by both those with perceived low status (e.g., to improve their status) and those with perceived high status (e.g., to defend their status).
Accordingly, activated faultline teams will engage in social competition to enhance positive identities as manifested through status conflict (Bendersky & Hays, 2012;Jehn, 1995).
Four features describe status conflict: (1) it is motivated by instrumental interests to increase or defend one's status associated with a positive identity; this means that individuals and subgroups attempt to legitimate their own group over others independently of their interpersonal relations; (2) it involves a coalition of actors; (3) it is zero-sum whereby if one party wins status, another party loses status; and (4) it is reflected by denigrating or aggrandizing behaviors (Bendersky & Hays, 2012). The uniqueness of status conflict (compared to other types of conflict) is related to the view of status not as a static emergent state, but rather as a negotiable and dynamic social resource. The identity-based structure caused by activated faultlines invokes the first two features of status conflict, as explained by social identity theories. Note that we are not necessarily concerned with the content of the status (e.g., race, gender) or the specific ascription of status to a particular subgroup (e.g., high status, low status); rather, our arguments are based on the idea that the presence of subgroups derived from an informal sensemaking structure (activated faultlines) results in differential status perceptions between subgroups. Thus, regardless of whether a subgroup is in a dominant, submissive, or equal position relative to other subgroups, the saliency of subgroups and perceived threats regarding job resources and status push subgroups to manipulate the social construction of status relations fueling status conflict (Chattopadhyay et al., 2008;Porath, Overbeck, & Pearson, 2008;Zhou, 2005). Consequently, we expect a positive relationship between teams with strong activated faultlines and status conflict.
The latter two features of status conflict reflect outcomes associated with status conflict in the form of behaviors and status distribution within a team. Status conflict, unlike relationship or task conflict, is defined by its zero-sum nature (Bendersky & Hays, 2012).
Meta-analyses of relationship conflict have generally shown that relationship conflict results in a negative net loss for everyone within a team (de Wit, Greer, & Jehn, 2012), with some exceptions (e.g., Jung & Lee, 2015). Likewise, task conflicts are generally seen as detrimental to teams, although some studies have found positive effects (e.g. Jehn, 1995;Hollenbeck et al., 1995). Because status conflict is inherently zero-sum, the negative interactions between subgroups in the form of aggrandizing and denigrating behaviors reinforce subgroup salience, which has negative consequences for teams. For example, research has found that coalitions formed with the aim of increasing or defending status differences creates an environment of reduced communication that impairs team performance (Tost, Gino, & Larrick, 2013). For this reason, we consider status conflict as a unique and specific type of conflict that operates above and beyond other types of conflict and results from perceptions associated with ambiguity or the lack of status legitimacy in teams.
When activated faultline teams experience status conflict, they will focus at least a portion of subgroup actions on maintaining or enhancing status positions rather than focusing on the team task; as a result, these teams are likely to make suboptimal decisions regarding performance (Bendersky & Hays, 2012;Loch, Huberman, & Stout, 2000;Groysberg, Polzer, & Elfenbein, 2011). In addition to poor use of time management, activated faultline teams experiencing status conflict are likely to have low levels of performance for two other reasons. First, members of a subgroup experiencing status conflict are likely to feel less connected to members of other subgroups resulting in differentiated information seeking and strong judgment biases (Vescio, Snyder, & Butz, 2003;Ruscher & Fiske, 1990). Second, according to Worchel, Rothgerber, Day, Hart, & Butemeyer (1998), status conflict can result in increased attention to subgroup goals and tasks, benefitting subgroup performance at the expense of overall team performance. Consequently, we expect that activated faultline teams will experience status conflict, impairing team performance.
Hypothesis 1a: The negative relationship between activated faultlines and team performance will be mediated by status conflict.
In the preceding rationale, we argued that activated faultline teams will experience status conflict leading to low levels of performance; but it is also important to explore a mechanism that explains why activated faultlines cause status conflict. As the intergroup relations literature posits, intergroup reactions and status conflicts are influenced by team members' views about the inter-subgroup status distribution within the team and the legitimacy of their subgroup's status (Tajfel 1974;Tajfel & Turner, 1979;Doosje, Spears, & Ellemers, 2002;van Dijk & van Engen, 2013;Christie & Barling, 2010). In the following section, we elaborate on a mechanism that explains the positive relationship between activated faultlines and status conflict: team justice climate (i.e., a team-level cognition expressing shared fairness perceptions of treatment by leaders or other team members; - Whitman, Caleo, Carpenter, Horner, & Bernerth, 2012).

The mediational role of team justice climate (TJC) in the activated faultlines and status conflict relationship
In activated faultline-based teams, individuals across subgroups may differ on several dimensions that may include both high and low status characteristics, generating a sense of status ambiguity. Such ambiguity complicates employee's efforts to use salient dimensions (e.g., demographic characteristics) to classify and systematize their work context (Chattopadhyay et al. 2004). In fact, in ambiguous status situations, the desire for team members to build a positive identity through categorization processes (e.g. classifying themselves into a valued ingroup - Turner, 1987) is difficult because of the possibility of contradictory stereotypes (Franke, Keinz, & Klausberger, 2013). Hence, the ambiguity caused by activated faultline-based teams will likely cause team members to experience uncertainty.
In parallel, when there is a clearly-aligned faultline demarcating high status and low status subgroups, the informal sensemaking structure represented by the faultline may drive perceptions of status illegitimacy caused by stereotype-based information (Hogg & Terry, 2000;Jost & Banaji, 1994). Status attributions ascribed by society or the organizational environment (e.g., being a female in a male dominated industry) can result in non-ambiguous, but illegitimate status perceptions. For example, in a male dominated industry, men (a dominant category) may be viewed as competent and committed, while women (a subordinate category) may be seen as incompetent and not committed (Fiske, Cuddy, & Glick, 2007). These societal attributions influence perceived structural relations between subgroups thereby justifying discrimination between subgroups (Fiske, Cuddy, Glick, & Xu, 2002;Glick & Fiske, 1999). Thus, activated faultlines induce uncertainty either through the ambiguity that mixed attributions create, and/or through the perceived illegitimate status attributions dictated by the social context where teams are embedded.
As a consequence of these non-mutually exclusive ambiguous or illegitimate perceptions, members of activated faultline teams will likely experience unfairness perceptions (low TJC). Based on the social identity approach (Chattopadhyay et al., 2004a), the legitimacy of a specific social stratification relies on two assumptions: that the status linked to a specific category reflects the true status in an organizational context (distributive justice, Greenberg 1987); and that the distribution process is perceived to be fair (procedural justice, Lind & Tyler, 1988). Because the status structure of activated faultline teams lends itself to behaviors aimed at increasing or defending one's subgroup status (Chattopadhyay, 1999;Chattopadhyay, et al., 2004a), members of activated faultline teams will experience low levels of TJC (Lamertz, 2002;Lind, Kray, & Thompson, 1998). In sum, the informal sensemaking structure derived from activated faultlines allows for an interpretation of the inter-subgroup status hierarchy that is ambiguous or illegitimate, driving low levels of fairness perceptions (Chattopadhyay et al., 2004a). In this situation, according to the social identity approach, team members will change their behaviors and attitudes to enhance a positive social identity through social competition behaviors (e.g., status conflict), as a way to reduce uncertainty and/or unfairness perceptions (Hogg & Mullin, 1999;Major et al., 2002). Consequently, status conflict in activated faultline teams results from a desire to enhance a positive identity, and/or to create a more legitimate intersubgroup status configuration (Hogg & Terry, 2000;Naumann & Bennett, 2000;Chattopadhyay et al., 2004a;Porath et al., 2008). Thus, we submit that a team's sense of unfairness, as manifest by a low TJC causes status conflict in activated faultline teams, which impairs team performance. This results in the following hypotheses: Hypothesis 1b: Team justice climate will mediate the relationship between activated faultlines and status conflict. Hypothesis 1c: The negative relationship between activated faultlines and team performance will be mediated by team justice climate and status conflict, such that activated faultlines will reduce team justice climate and therefore promote status conflict, which in turn will impair team performance.

The moderating role of team structure clarity in the relationship between team faultlines and team justice climate
Despite the presence of ambiguity and illegitimacy in teams with activated faultlines, organizations and leaders can take action to ameliorate the negative effects of activated team faultlines on team justice climate. The use of stereotypes and social comparison mechanisms provide an inter-subgroup status configuration that results in perceptions of unfairness and ultimately, behaviors associated with status conflict. However, organizations can provide a clear team structure to create conditions that legitimate inter-subgroup status differences, reduce ambiguity and attenuate the negative effects of activated faultlines on TJC. Team structure clarity (the lucidity of the structure) is the extent to which a team is organized through an elaborated division of vertical and horizontal labor, and has clear procedures for According to the social identity approach, the extent to which the status of a particular category is legitimated can be explained by the norms of organizations and other social structures, such as the broader society (Bargh, 1999; van Knippenberg & Ellemers, 1993).
For example, in a team composed of scientists and managers, acknowledgement of the scientist's status by the managers and respectful acknowledgement of the management's status by the scientists will result in status legitimacy and a sense of fairness (Zitek & Tiedens, 2012). Under these circumstances, status differences are unlikely to result in stereotypical perceptions (Hornsey, 2008) or a desire for status change. Accordingly, the negative impact of activated faultlines on fairness perceptions (i.e., TJC) can be reduced by creating a context where team members clearly perceive inter-subgroup status differences.
Managers can create such a context by providing teams with clear team structures that reduce uncertainty, increase legitimacy around status, and provide common schemas for expectations (Fiol, et al., 2001). In fact, when teammates know their roles, procedures and authority relations (i.e., status hierarchy), their interactions become predictable, such that work-related information exchange becomes more effective (Anderson & West, 1998;Baron, Jennings, & Dobbin, 1988;Degoey, 2000).
To better characterize how team structure clarity impacts the activated faultlines-TJC relationship, we consider Bunderson and Boumgarden's (2010) main indicators of team structure: hierarchy, formalization and specialization. Specifically, clarifying formal responsibility positions (i.e., hierarchy) can ensure that different pieces of information are shared and acknowledged during task-related interactions, setting the stage for justice-related information sharing (Larson, Foster-Fishman, & Franz, 1998;Lind, 1995). In this regard, clear identification of expert roles (by clarifying who possesses information and where particular types of information reside within the team) facilitates information sharing  (Stasser, Steward, & Wittenbaum, 1995). In addition, clarity around roles, procedures and priorities (i.e., formalization), makes member relations and interactions predictable and eliminates uncertainty by grounding common perceptions (Edmondson, 1999;Sitkin & Roth 1993). Finally, specialization occurs when team members engage in dyadic exchanges, sharing and seeking out information related to their own and their teammates' capabilities and responsibilities within the team (i.e. role identification behaviors, Kozlowski, Gully, Nason, & Smith, 1999). Through specialization, team members develop a deep understanding of their own role, and create reciprocal role expectations that in turn are conducive to a shared cognitive structure regarding other's roles and expectations (the basis for the construction of a justice climate). Thus, team structure clarity creates the conditions that allow members in activated faultline teams to reduce their focus on demographic categorizations and refocus their efforts on shared understandings of interaction and hence, develop a common view of fairness, as evidenced by high levels of TJC. In this regard, perceived inter-subgroup status legitimacy reduces potential threats caused by differences among subgroups (Chattopadhyay et al., 2004a;Tyler, 2006) and when inter-subgroup status distribution in a team is clear and legitimated through organizational actions, team members are more likely to accept the status differences (Chattopadhyay et al., 2004a). For these reasons, we submit that team structure clarity will reduce the negative impact of activated faultlines on TJC.
In contrast, unclear team structures perpetuate ambiguity around role relationships, as well as uncertainty and illegitimacy perceptions around the inter-subgroup status hierarchy, created by activated faultlines that jeopardizes fairness perceptions (Fiol, et al., 2001). In other words, without a clear formal structure, activated faultlines provide the baseline informal sensemaking structure guiding team member interactions centered on subgroups and perceived status. Under such circumstances, common interpretations of fairness will be

Organizational Context
The data were collected from a Spanish healthcare organization delivering social healthcare services, such as psychological services and social rehabilitation for socially marginalized or brain-damaged people, and those suffering from intellectual incapacitation.
Accordingly, the organization has a range of diverse employees, who are equipped with a variety of technical skills and expertise. The organization has a team-based structure, where teams consist of social workers, sociologists, psychologists, educators and in some cases social technicians (workers who perform many of the same functions as social workers, but do not have the formal education or qualifications to hold the title of social worker). The main task of these teams is to monitor and provide daily support for the service beneficiaries who have a high risk of social exclusion. These teams establish social rehabilitation programs for the beneficiaries and, when required, intervene in emergencies (i.e., specific psychiatric treatment) by contacting the Spanish public healthcare system. As one example of a social rehabilitation program, beneficiaries were trained to collect used oil from families in private households, transform the used oil into soap, and return the soap to the families who provided the used oil. To facilitate this program, the team members discussed and agreed on action protocols with respect to the expected impact of this activity on the beneficiaries. The three main goals linked with this particular program were: stigma reduction associated with the beneficiaries from the community; development of a new set of skills (soap-production); and an increase in the beneficiaries' social network.
This organization is ideal for examining the relationships proposed in our model for three main reasons. First, the organization operates under a team-based structure whereby each team is led by a supervisor and has its own set of beneficiaries and performance assessments. Second, each team is an intact unit such that an individual belongs to only one team and the team members see themselves as being part of a distinct team. Effective completion of tasks requires a high level of interdependence and there is mutual responsibility for the team outcomes. And third, team supervisors are given latitude with respect to team organization and functioning, such that some teams function as decentralized units and other teams operate within formal structures.

Sample
Our study required collection of data via multisource survey instruments at four different time periods at six-week intervals. During the first wave of data collection we distributed questionnaires to 512 team members and received 411 complete questionnaires (first response rate = 80.27%); additionally, we surveyed team supervisors (response rate of 89.39%, 59 out of 66 surveys distributed). During the second data collection, we sent surveys to the 411 members who completed the first survey and received 367 completed surveys (second response rate = 71.67%). A response rate of 60.35% was achieved during the third data collection stage, based on our receipt of 309 completed surveys from the 367 surveys solicited (those that completed the first two surveys). The fourth and final data collection wave focused on surveying team supervisors and yielded a response rate of 77.27% (51 surveys returned out of 66 surveys distributed). Teams with lower than 60% within-team response rate, which is established as the minimum requirement for meaningful aggregation of data to the team level (Timmerman. 2005), or with no matched upper-level manager data were excluded. Thus, the final study sample included 271 employees in 41 teams. 2 Of the 271 2 In a posthoc power analysis, with N = 41, α = .050, and f 2 = .315 (estimated through the average R 2 Adjusted ), we obtained a power of 1 -β = .868, which is consistent with the requirements for mediation analysis (Pan, Liu, Miao & Yuan, 2018). This analysis was calculated using the G*Power 3.1.9.2 software (Faul, Erdfelder, Lang, & Buchner, 2007). employees, 68% were female, the average age was 35.55 years (SD = 7.54) and mean organizational tenure was 4.94 years (SD = 3.33). Team sizes ranged from 4 to 9 people.
Overall, 25 teams and 241 employees were excluded from the final data analysis; there were no statistically significant differences between respondents and non-respondents with respect to age, gender, organizational tenure, and educational background.

Procedure
Archival demographic data, including information on age, gender, organizational tenure, and educational background was collected for the entire population of employees from the Human Resources Department. Before starting data collection, one of the authors met with team supervisors to explain the data collection procedure (e.g., to describe the process to ensure respondent confidentiality), to obtain buy-in, and to answer any questions.
Surveys were completed and returned to a secure mailbox located at one of the author's universities. We anonymized the sample by providing each potential participant (the entire population of employees) with a code that was included on the surveys to enable matching of participant data across the different surveys. Our first wave of data collection (Time 1) was designed to elicit team member's responses to survey items around faultline perceptions, team structure clarity, and some of the control variables (intra-team conflict and task complexity  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 attributes (Thatcher et al., 2003;Meyer et al., 2014), extant evidence suggests that faultline effects are more pronounced when they are perceived (Jehn & Bezrukova, 2010;Rico et al., 2012). Consistent with our theoretical approach, our activated faultlines measurement integrates both perspectives, weighting a dormant faultline's strength with the perception of activation; this means that attributes that are perceived to be the most influential in creating an activated faultline-based subgroup are weighted most heavily. Calculating our measure of activated faultlines involves five steps. In Step 1, we selected the demographic characteristics that were relevant for calculating activated faultlines (i.e., educational background, gender, organizational tenure, and age) in our sample. Our selection was based on two criteria: a) extant empirical evidence showing that team members categorize themselves based on these attributes (Tsui, Egan, & O'Reilly, 1992); and b) the relevance of these attributes for the teams included in our sample (Thatcher & Patel, 2012), assessed through a round of interviews with several team members and supervisors not included in our sample. For example, these interviewees informed us that organizational tenure is a very important dimension in the company, as team members often view the opinions of long-tenured employees to be more valuable than short-tenured employees.

In
Step 2 we estimated faultline strength using Shaw's (2004) procedure, which offers high measurement quality for teams with up to 10 members (Meyer et al., 2014). Following Shaw's (2004) procedure, we converted the continuous variables into categorical ones 3 , and then created four separate dormant faultline scores whereby each attribute was used as the primary attribute to ascertain the extent of alignment. For example, if we begin with the attribute of age, we determine the strength of the dormant faultline based on age and then measure the extent to which the other attributes of gender, educational background, and organizational tenure align with the age faultline. To obtain the overall dormant faultline 3 We used quartiles based on the mean values in the sample (i.e., 4 categories for each continuous variable). To overcome such limitations and to capture a team's activated faultline, we introduce a perceptual element in the next step ( Step 3). Following Jehn and Bezrukova (2010) In Step 4, we multiplied the dormant faultline strength score derived for each attribute in Step 2 by the extent to which a faultline along the same attribute was perceived to exist as described in Step 3. For example, the dormant faultline strength score for age was multiplied by the perception that an activated faultline was perceived to exist based on age. This allowed us to capture the extent to which each possible demographic attribute contributes to an overall activated faultline. Therefore, our final activated faultline measure (Step 5)  validity of our activated faultline measure).
As described earlier, many of our variables were measured via a survey instrument.
Unless otherwise noted, all survey items are measured on a 5-point Likert scale where the anchors range from 1 (strongly disagree) to 5 (strongly agree).
Team Justice Climate. We adapted the intraunit justice climate scale developed by Li & Cropanzano (2009) so that it captures perceptions of fairness within the team.
According to Li and Cropanzano (2009), the construct includes three dimensions of justice climate that are applicable to the unit (i.e., the team): distributive justice, procedural justice, and interactional justice. We measured distributive team justice climate using five items reflecting the extent to which team members contribute equitably to the team effort (e.g., " The recognition my teammates have received for their performance is appropriate considering the quality of the work they have completed"). Procedural team justice climate was also measured with a five-item scale reflecting members' evaluations of the procedures used within the teams. These items are similar to Colquitt's (2001) items (e.g., My teammates are able to express their views and feelings about the way decisions are made in the team).
We adapted Donovan, Drasgow, and Munson's (1998) four-item scale to measure interactional team justice climate (e.g., The way my teammates make decisions is applied consistently). The inclusion of all fourteen items into our measure presents an acceptable reliability score (α= .84), and team aggregation was warranted (ICC1 = .23; ICC2 =.58).

Team Structure Clarity. We adapted Bunderson and Boumgarden's (2010) 5-item
scale to reflect team structure clarity from the supervisor's point of view (e.g., In my team individual roles are very clear and teammates don't stray from them). The scale showed an acceptable reliability (α= .87). We also asked team members to rate team structure clarity from their own perspective, using the same 5 items (ICC1 = .28; ICC2 = .66). As additional evidence for validity, the team member and the supervisor scores are highly related (r = .65; p <.01); we decided to use the supervisors' ratings to differentiate the sources of information.
Team Performance. We combined the criteria proposed by Ancona and Caldwell (1992) and van der Vegt and Bunderson (2005) to measure team performance: efficiency, quality of innovations, productivity, adherence to schedules, adherence to budgets, and overall achievement. Team supervisors responded to one item for each criterion (e.g., The team accomplishes the task smoothly and efficiently). The response anchors for these items ranged from 1, "far below average," to 5, "far above average." Each supervisor was asked to compare the performance of his or her team with the performance of teams that performed similar tasks. The scale reliability was acceptable (α= .84).

Control variables.
Several variables were controlled for in our research, due to their potential to affect our results. We controlled for task complexity, since it has been shown to be a key variable in understanding the impact of diversity in complex teams (Wegge, Roth, Neubach, Schmidt, & Kanfer, 2008 Harrison & Klein, 2007). Following the procedure suggested by Jehn et al (1999), we averaged the heterogeneity variables to calculate the overall team heterogeneity control variable. We also controlled for team size, central among the compositional variables that have been shown to influence team processes and outcomes (Ancona & Caldwell, 1992); team size was obtained by the human resources department of the company.
Finally, to ensure that we captured the mediation effect of status conflict rather than other types of conflict as hypothesized in H1, we controlled for intra-team conflict. In doing so, we used Jehn's (1995) eight-item scale that showed acceptable reliability (α= .80).

Analytical Approach
Before testing our hypotheses, we assessed the measurement model by running several Confirmatory Factor Analyses (CFA). Given the nature of our data (categorical) we ran CFA using Robust Maximum likelihood (MLM in the Mplus Software) using the polychoric correlation matrix, a method that is demonstrated to be robust under our circumstances (Finney & Distefano, 2006). Our results showed that the three-factor model (team justice climate, team structure clarity, status conflict) presented a reasonable fit to our data (CFI .92. NFI .90. RMSEA .07), while the single-factor model (CFI .72. NFI .85. (Hu & Bentler, 1999).

Hypotheses Tests
To test our hypotheses, we conducted several bootstrapping analyses (5000 resamples) following Hayes's (2013) recommendation employing his Process macro.

DISCUSSION Theoretical implications
Building on the social identity approach, our model offers new insights on how faultlines can disrupt team performance and provides information on strategies to counteract such negative effects. Deviating from the predominant view of team faultlines, we describe how activated faultlines represent an informal sensemaking structure that team members adopt to interpret and provide meaning about their team social environment and intersubgroup status differences. To date, most of the faultlines research has centered on studying the direct effect of objective (dormant) configurations of diversity on team processes and outcomes. We show the importance of understanding how activated faultlines result in unfairness perceptions that lead teams to engage in conflict over status, instead of focusing on their common task; we show how this is especially detrimental for team performance in the absence of clear and legitimizing information (i.e., team structure clarity). By conceptualizing faultlines in a way that incorporates inter-subgroup status perceptions, we begin to understand why activated faultlines can be so damaging for teams. Not only do demographic similarities and differences divide team members, but the belief that some team members may be unfairly reaping the benefits associated with ambiguous or illegitimate status distinctions helps to explain the disruptive nature of activated faultlines. Accordingly, our findings complement and extend previous theory on the wellgrounded team faultlines-team conflict-team performance causal chain (Thatcher & Patel, 2012). Despite being a key aspect of organizational working environments (Gould, 2002), status conflict has not been investigated in extant faultlines literature (Thatcher & Patel, 2012), and has not been investigated as related to fairness perceptions within a team. In this regard, we introduce a couple of unique twists to the status conflict literature by suggesting that: first, in the absence of other formal structures (Tiedens, Unzueta, & Young, 2007), activated faultlines play a prominent role in establishing an ambiguous and/or illegitimate inter-subgroup status hierarchy built from team member differences; and second, although (inter-subgroup) status hierarchies are negotiable and contested (Bendersky & Hays, 2012), conflict about status can emerge through perceptions of unfairness caused by perceived rather than actual inter-subgroup status differences stemming from activated faultlines.
Our findings are also highly relevant for team justice theory and research. First, by focusing on team-level justice perceptions and moving beyond fairness perceptions associated with organizational authorities, we strengthen recent developments in justice theory (e.g., Whitman et al., 2012). Second, by revealing how activated faultlines drive justice perceptions within the team, we extend justice theories to contexts that provide informal sensemaking structures (i.e., activated faultlines), triggering social identity based mechanisms that can disrupt the construction of high levels of TJC. Finally, based on Spell, Bezrukova, Haar, and Spell's (2011) findings that faultlines exacerbate the relationship between organizational distributive injustice perceptions and relational and task conflict perceptions. Our study suggests that activated faultlines can actually be the impetus for perceptions of low levels of fairness. Thus, our work complements research on individuallevel views of justice perceptions and advocates the importance of looking at justice from a team-level perspective.  (Lind, 1995), such as those built around activated faultlines. Further, our pattern of results aligns with extant research suggesting that team structure augments agreement about the tasks to be done and team member relationships (Mathieu, Heffner, Goodwing, Salas, & Cannon-Bowers, 2000). This is especially relevant for activated faultline-teams, where teammates may be informally structured around identity-based subgroups rather than identified with the team as a whole (Carton & Cummings, 2012;Thatcher & Patel, 2012).
In sum, and consistent with social identity theory (Chattopahyay et al., 2004a;Hogg & Terry, 2000) a clear team structure provides an alternative and legitimate interpretative framework for activated faultline-teams. The clear team structure precludes team members from relying on the ambiguous and/or illegitimate sensemaking structure provided by activated faultlines to interpret their social reality. Our findings are important for researchers interested in the relationship between structure-based theories and social identity theories.
Whereas previous research shows that structuring teams around task roles prevents inter(sub)group bias and facilitates information elaboration and high inclusive identification (van Ginkel & van Knippenberg, 2008, Rico et al., 2012, our results provide evidence that team structure clarity can also reduce the negative effects of inter(sub)group bias after subgroup categorizations have already occurred. The value of this finding suggests that  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 categorization effects, and the theories that explain these effects, should incorporate relevant situational factors. Additionally and consistent with social identity theories (Hogg & Terry, 2000), our findings suggest that identity-based mechanisms are the "default" option that individuals use in ordering the social world. This finding also explains why in the absence of other information, activated faultlines may be a natural, yet disruptive, sensemaking structure for teams.

Measurement implications
In addition to the theoretical implications described above, our activated faultlines Jehn & Bezrukova (2010) investigated both dormant and activated faultlines in student groups where group composition was manipulated, but they kept the two aspects of the faultlines construct separate, without studying their joint impact on performance and processes. Accordingly, being able to measure activated faultlines in organizational settings represents a new contribution to the faultlines literature.
By reconciling the two streams of faultlines studies -streams that utilize measures of dormant faultlines and streams that measure perceptions of faultlines or manipulate faultlines so they are activated, we offer a promising measurement approach to those interested in gauging the effects of faultlines on teams in a variety of settings. We believe it is promising for two main reasons: first, it provides researchers with the ability to assess the importance of differentially-weighted faultlines in teams; and second, it simultaneously captures the objective structure and perceived categorization salience of different diversity attributes. Our measure enables researchers to address recent calls in the teamwork research to attend to the dynamic evolution of team constructs (Collins, Gibson, Quigley, & Parker, 2016). For example, if a company introduces specific policies around gender equality and integration, this would not change the underlying dormant gender faultline of a specific team, but it may change the internal "weighting", so that the resulting activated faultline may look different from the original activated faultline. Through our measurement approach, researchers can trace changes in the faultline configuration of a given team over time. This opens the door to exploring the dynamic evolution of team faultlines, a neglected area in the faultlines literature (Thatcher & Patel, 2012), but which represents a new frontier for future research. For instance, it is now possible to determine whether different attributes contribute to faultline activation (as a function of specific contextual elements) or whether an activated faultline crystallizes over time.

Practical implications
Our study also has important practical implications that deserve to be discussed. From a functional approach to leadership (Fleishman el al., 1992), our activated faultline measure allows organizations and managers to know both when faultlines have the potential to form (dormant), and to gauge from time to time the degree to which team members perceive that  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 faultlines have become activated. This information can aid managers in knowing when the threat from faultlines is real, rather than having managers worry about potential unlimited attribute combinations fostering faultlines that are not activated.
In addition to the diagnostic recommendations, our results show that in a team with activated faultlines, low levels of TJC and status conflict are both related to reductions in team performance. Combining our results regarding status conflict with extant leadership research (Cho, Overbeck, & Carnevale, 2011), there are certain elements that leaders should consider. First, leaders may clarify and adjust both structural and task characteristics (i.e., interdependence/autonomy, Rico, Molleman, Sanchez-Manzanares & van der Vegt, 2007) to provide teams with a legitimate sensemaking structure that increases fairness perceptions and diminishes status conflict and its subsequent impairment on team performance. More specifically, leaders can assign internal roles and responsibilities according to the task, responding to the necessity for team-members to have a clear understanding of the social order and status hierarchy. Second, team leaders need to be aware of team member's status expectations, especially in diverse teams where members may feel that their specific knowledge is more valuable than other types of knowledge. Status expectations may also exist because of individual differences, where people with high self-esteem and self-efficacy tend to expect high status (Ali, McWhirter, & Chronister, 2005). In order to alter the nature of the underlying biases of an informal sensemaking structure caused by activated faultlines, managers could employ environmental stimulation strategies, emphasizing or de-emphasizing some of the faultline attributes to change the internal perceptual configuration of identitybased subgroups, with the goal of reducing perceptions of unfairness. For example, managers who fear a strong activated age-based faultline may be able to defuse the activation by adopting age-inclusive strategies (such as diversity training or showing the value of age diversity) to reduce the salience and potential split around an age faultline (Homan, Van Knippenberg, Van Kleef & De Dreu, 2007).
Beyond these general recommendations and building on our results, a managerial strategy that provides teams with a clear structure (i.e., clarity around roles, procedures and priorities) will make team member interactions more predictable and less uncertain.
Introducing a clear structure should help teams that are struggling with ambiguity or illegitimacy perceptions as a result of activated faultlines by reducing perceptions of unfairness. Further, based on research by Antino, Rico, Sanchez-Manzanares & Lau (2013), team leaders can strengthen team knowledge regarding who knows what in the team, to reinforce both role clarity as well as a sense of justice. In this regard and from a job design perspective, it is important to note that clarifying roles makes it easier for team members to understand how best to work together.

Limitations and future research
There are some opportunities for future research that build off the limitations of our study. To begin, like most survey-based studies, our measures (although we employed three sources of information) are perceptual in nature, except those based on demographic information that are used as part of our activated faultline measure. In this regard, further research will benefit by including non-obtrusive process measures (i.e., behaviorally-coded indicators of status conflict) and objective performance indicators. In this way, and in addition to subjective indicators, we can obtain a more comprehensive and robust view of activated faultlines on team performance.
Further, our study (as in most organizational field research) incorporates a research design that does not allow manipulation of independent variables. Thus, we were unable to run the study under the necessary controls that causality assertions require. Accordingly, although we opted for a cross-lagged and multisource approach, we cannot ensure causality in our results. Further studies should replicate our findings with research designs that enable testing of causal relationships. Also, our results come from a specific organizational and  (Merriam et al., 2001).
Although our cross-lagged design offers temporal separation in the measurement of different variables, we did not completely capture the temporal dynamics associated with our activated faultlines measure. In order to rectify this situation, faultlines researchers could employ longitudinal research designs to study how activated faultlines change in response to team contexts or external stimuli. For example, future researchers could study whether there is a pattern with respect to which diversity dimensions are more relevant to the activated faultline over time (e.g., do most teams experience activated faultlines based mainly on Additionally, our results provide new avenues for future research addressing the relationship between activated faultlines and TJC. Specifically, the negative relation between activated faultlines and TJC offers new insights for social information processing theory (Salancik & Pfeffer, 1978), since the categorization mechanisms (potentially related to the faultline-based splits) could corrupt team members' work-related discussions and interpretation of work-related events. In a similar manner, attraction-selection-attrition research might benefit from our results as a way of understanding subgroup asymmetries of common climate perceptions (Schneider, 1987). Although we found that activated faultlines led to low levels of TJC, teams with many activated faultlines represent an ideal context for investigating how subgroupings may impair the emergence of a collective climate over time (Harrison et al., 2002), as well as a collective shared representation of the team structure.
In our theorizing, we focused on exploring the relationship between activated faultlines and constructs associated with fairness, legitimacy and inter-subgroup status.
However, there are other potential constructs that might be impacted by the informal sensemaking structure caused by activated faultlines. Given the renewed attention to the importance of psychological safety (Anderson & West, 1998& Edmonson, 1999 for team performance in highly innovative technological environments (Duhigg, 2016), researchers could study the impact of activated faultlines on the creation of a shared sense of safety within a team. As an informal sensemaking structure, activated faultlines may lead team members to trust out-subgroup members less than in-subgroup members, reducing safety climate perceptions, and harming team learning and performance (Edmonson, 1999).
Furthermore, in line with our results showing the benefits that a clear team structure provides to faultline-based teams, future research could study whether specific leadership styles help to provide such clarity, for example studying the impact of directive versus empowering leadership styles (Lorinkova, Pearsall, & Sims, 2013).

Measurement Computation Example
To better illustrate the functioning of our activated faultlines measure, in Table A1 below we provide a comparison of four teams, where differences based on age, gender and ethnicity are used to calculate the activated faultline score. Using the measurement steps described in the manuscript, we first compute the faultline strength score using Shaw's (2004) measure based on age (e.g., we determine where the split exists on age and then determine the extent to which the attributes of gender and ethnicity align with that split; seen in Column 1 of table A1). More specifically, we employ the following calculation: where IA= internal alignment within the subset of group members sharing the same age category and CGAI (cross group internal alignment) = the extent to which group members belonging to another subgroup (by falling in a different age category) are similar on other attributes such as gender, or ethnicity (Meyer et al., 2014).
Similarly, Columns 2 and 3 show the faultline strength scores based on gender and ethnicity. Columns 4, 5, and 6 reflect the faultline perceptions score based on age, gender, and ethnicity, respectively. Column 7 provides our activated faultline score using the following calculation: where FLS = Faultline strength based on Shaw (2004) and PS = perception of split based on Bezrukova et al., (2010). We use Shaw's (2004) approach as it allows us to estimate faultlines strength separately for each attribute (based on internal and cross group alignment).
In this example, Teams 1 and 2 share the same team demographic composition (same heterogeneity in terms of age, gender and ethnicity). Regardless of the faultline measure used (various measurement approaches are described below in this Appendix) these teams will receive the same dormant faultlines score, as the objective diversity within the team is the same. However, our measurement approach allows us to discriminate between these two teams, by considering the extent to which the perceptions of subgroups formed by the demographic attributes differ.
Similarly, Team 1 and Team 3 share the same faultline perceptions across these three demographic characteristics. Thus, if we only measured the perceptual faultlines, these two teams would look identical; however, by incorporating the dormant (objective) faultline strength score, we show how these two teams have different activated faultline scores, and likely experience faultlines differently. In sum, our measure allows us to capture both the objective team configuration (through the faultline strength estimation) and the perception of faultline activation, allowing for differentiation between teams with the same objective structure but different perceptions of activation.

Additional Evidence for Validity of our Activated Faultline Measure
Predictive validity. Predictive validity refers to the extent to which a measure is able to predict related constructs (DeVellis, 1991). In an unpublished presentation, the faultlines measure proposed in this paper was a better predictor of elaboration of task relevant information in a different sample, compared to Shaw's (2004) measure. In a sample of multidisciplinary innovative teams working in an IT firm, the relationship between Shaw's Page 50 of 54 Academy of Management Journal   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 faultline measure and elaboration of task relevant information (measured with the four-item scale developed by Kearney & Gebert, 2009) was r = -.222, p < .05. In addition, the relationship between perceptual faultlines and elaboration of task relevant information was r = -.185, p < .05. Using the measure described in our manuscript, the relationship between activated faultlines and elaboration of task-relevant information was r = -.336, p < .05.

Content validity.
Content validity refers to the extent to which a given measure reflects a content domain (DeVellis, 1991). Our activated faultlines measure has content validity because it is consistent with Lau and Murnighan's (1998) original theorization that dormant (or objective) faultlines are activated when the subgroups' alignments are salient.
Our measure includes both the dormant and perceptual facets of the faultline construct that result in its activation. Dormant faultlines (operationalized as faultline strength using Shaw's (2004) measure) inform us about the extent to which demographic attributes are aligned within subgroups and deviate between subgroups. Perceptions of subgroup existence (Jehn & Bezrukova, 2010) are valuable for capturing the salience of a faultline split. By combining both the dormant and perceptual facets of the faultline in a single measure, we ensure that the strength of the dormant split (within vs. between alignment of attributes) is weighted according to its salience (the perception of a faultline based on the attributes) to provide a valid measure as originally conceptualized.

Brief Description of Other Faultlines Measures
As described in our manuscript, there are no faultline measures that incorporate both a dormant and perceptual element; thus, it is difficult to establish construct validity as no other measures provide a direct comparison to our activated faultline measure. However, we feel it is valuable to provide a brief review of the current approaches to measuring objective faultlines (as reflected in Meyer et al., 2014) and explain why we used Shaw's (2004) measure to calculate the dormant (objective) portion of the activated faultline construct rather than many of the other popular and useful existing faultline measures.
Thatcher's fau (Thatcher, Jehn, & Zanutto, 2003) is a variance-based approach that detects the two-subgroup configuration delivering the largest ratio of between-group variance over the total group variance of attributes. Further developments of this measure (Bezrukova et al., 2009;Zanutto et al., 2011) incorporate a measure of the distance between subgroups.
While these approaches are useful when researchers want to focus on a specific subgroup configuration (the one that maximizes the variance difference), they are not able to take into account all possible subgroup splits (which is necessary for the subsequent weighting with the perceived measure) nor do they take into account the perceptual measure.
Gibson & Vermeulen's (2003) approach considers the overlap of all potential dyads to estimate the subgroup strength. Because of its focus on dyads, this approach is not suitable for estimating the strength of a specific subgroup configuration based on the alignment of attributes necessary for our activated faultlines measure. Similar rationale applies to Carton & Cummings' (2013) subgroup measure which is useful when the focus is only on subgroup splits, without considering the perceptual configuration.
Factional faultlines (Li & Hambrick, 2005) considers the extent to which attributes align with the focal attribute of interest. This measure is useful when researchers focus on a specific attribute, but it does not provide information regarding specific subgroup configurations and does not allow researchers to integrate the perceptual facet of the faultline.
In sum, although the above faultline measures are valuable for understanding how dormant faultlines are configured when researchers do not have information regarding the extent to which faultlines are perceived, they are not suitable for use in measuring activated faultlines that include both dormant and perceptual facets.

Limitations and further developments
Although our faultlines measure presents an advancement with respect to faultline measurement, there are limitations that should be considered. First, as required when using Shaw's (2004) dormant faultlines measure, our measure also requires converting continuous variables into categorical variables. Future researchers may be able to determine how to combine perceptual faultlines measures with faultlines strength indicators estimated through continuous variables. Second, although our proposed faultlines measure provides content validity as it is consistent with the original faultlines construct description (Lau & Murnighan, 1998), we are unable to clearly partition the variance between the dormant and activated aspects of the faultlines, which might be useful in some cases. Finally, our faultlines measure proposes adding all faultlines attribute alignments (product of objective and perceptual estimations of faultlines); however further developments of the measure may consider alternative ways of integrating different faultlines attributes.  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60