Out of Sight, Out of Sync: Understanding Conflict in Distributed Teams

The bulk of our understanding of teams is based on traditional teams in which all members are collocated and communicate face to face. However, geographically distributed teams, whose members are not collocated and must often communicate via technology, are growing in prevalence. Studies from the field are beginning to suggest that geographically distributed teams operate differently and experience different outcomes than traditional teams. For example, empirical studies suggest that distributed teams experience high levels of conflict. These empirical studies offer rich and valuable descriptions of this conflict, but they do not systematically identify the mechanisms by which conflict is engendered in distributed teams. In this paper, we develop a theory-based explanation of how geographical distribution provokes team-level conflict. We do so by considering the two characteristics that distinguish distributed teams from traditional ones: Namely, we examine how being distant from one’s team members and relying on technology to mediate communication and collaborative work impacts team members. Our analysis identifies antecedents to conflict that are unique to distributed teams. We predict that conflict of all types (task, affective, and process) will be detrimental to the performance of distributed teams, a result that is contrary to much research on traditional teams. We also investigate conflict as a dynamic process to determine how teams might mitigate these negative impacts over time. (Distributed Work; Distributed Teams; Virtual Teams; Conflict) In response to a variety of factors that characterize the modern economy—including the global expansion of the marketplace and the businesses that serve it, the rise in mergers and acquisitions, and heightened competitive pressures to reduce the time to develop products— organizations increasingly are assembling teams whose members are drawn from sites far and near. Geographically distributed teams face a number of unique challenges, including being coached from a distance, coping with the cost and stress of frequent travel, and dealing with repeated delays (Armstrong and Cole 2002). Many scholars and practitioners have noted and expressed concern about one such challenge facing these teams: the prevalence and severity of conflict. Justifying their concern, reports from the field indicate that conflict is disruptive to performance in distributed teams. Field studies further indicate that geographically distributed teams may experience conflict as a result of two factors: The distance that separates team members and their reliance on technology to communicate and work with one another. Distance and technology mediation have gone unexplored in existing models of conflict and performance in teams because their authors, for the most part, assumed that team members were collocated and communicating face to face. As a result, whether these two factors spur new antecedents of conflict is not known, nor is it clear how conflict in distributed teams might be reduced. In this paper, we consider the possibility that distance and technology mediation give rise to conflict in distributed teams. We also examine how conflict might manifest itself over time as members of distributed teams learn how to work and communicate across distances and use technology more effectively. Geographically distributed teams, whose members reside in different cities, countries, or continents, share a number of properties commonly associated with traditionally conceived teams. Namely, they are groups of individuals that work together interdependently to accomplish a task, constitute distinct social entities, and jointly manage their team boundaries (Cohen and Bailey 1997, Hackman 1987). Recent studies demonstrate the kinds of problems that arise uniquely in the case of distributed teams and that render questionable the comprehensiveness of past models of group conflict and performance. For example, Armstrong and Cole (2002) reported that conflicts in geographically distributed teams went unidentified and unaddressed longer than conflicts in collocated teams. Beyond such empirical evidence, however, there is no comprehensive theory-driven prediction and explanation 1047-7039/03/1406/0615 1526-5455 electronic ISSN Organization Science © 2003 INFORMS Vol. 14, No. 6, November–December 2003, pp. 615–632 PAMELA J. HINDS AND DIANE E. BAILEY Understanding Conflict in Distributed Teams for conflict in distributed teams. It is not known whether conflict in distributed teams is triggered in a manner similar to that for traditional teams, nor is it clear whether the impact of conflict on performance is the same as in traditional teams. We investigate these issues by reviewing findings from research on distance and technology mediation and blending those findings with research on conflict in traditional teams. We also consider evidence from the growing number of empirical studies of distributed teams, which provide support for the propositions that we build inductively. Our theoretical analysis is intended to help establish a roadmap for future empirical work on distributed teams. Our analysis reveals that geographical distribution will have a significant impact on each type of group conflict proposed in recent organizational studies: task, affective, and process. Task conflict refers to disagreements focused on work content. Affective conflict (sometimes referred to as relationship or emotional conflict) refers to team disagreements that are characterized by anger or hostility among group members. Process conflict refers to disagreements over the team’s approach to the task, its methods, and its group processes. Affective conflict has been differentiated from task conflict (Eisenhardt et al. 1997, Pelled 1996, Pelled and Adler 1994) and from process conflict (Jehn 1997) partly in an effort to explain contradictory findings regarding the impact of conflict on team performance. Our analysis not only identifies a number of new antecedents for each type of conflict in distributed teams, it also reveals that the impact of group conflict will in some cases be different for distributed teams than for traditional, collocated teams. For example, task conflict has been found to be beneficial for performance on many traditional teams, but we contend that it will not be so for their distributed counterparts. Although we predict worse outcomes for distributed teams, we acknowledge that they have certain advantages over collocated teams. Distributed teams enable firms to take advantage of expertise around the globe, to continue work around the clock, and to create closer relationships with far-flung customers. We argue that these benefits will be diminished by the conflict engendered by distance and technology mediation, but we acknowledge that distributed teams may, at times, be the only viable option for achieving organizational goals. We therefore extend our analysis to consider conflict as a dynamic rather than a static process to illustrate how teams might mitigate the negative effects of distribution over time. Although we contend that the negative effects cannot be fully overcome, the preventative measures we identify may facilitate the performance of distributed teams when the realities of business dictate their use. In our analysis, we specify that distributed teams differ from traditional teams in only two respects: members separated by distance and forced to rely on technologies to mediate their communication and collaborative work. At first glance, this approach may appear to rule out other traits that might distinguish a distributed team from a collocated one. We contend that all other traits that may be associated with geographical distribution derive from distance or technology mediation, and we consider them in our analysis of these two factors. For example, some, but not all, distributed teams may experience incongruent temporal rhythms because members work in different time zones, but different time zones occur as a result of distance. Likewise, distributed teams may have members from different cultures. Beyond the impact of distance, distributed teams are no more or less likely than collocated teams to have members working at different times or to have a culturally diverse membership. Moreover, the effects of technology mediation are distinct from those of distance. Although distributed team members must rely on technologies because of their distance, technology mediation has impacts even for teams that are collocated, and that at times choose to rely heavily on technology rather than meet face to face (Mortensen and Hinds 2001). Among the many traits that might distinguish geographically distributed teams from collocated ones, separation by distance and heavy reliance on mediating technologies are the only two factors that hold true for all distributed teams. In the balance of the paper, we use the term “group” to refer to ad hoc collections of individuals brought together solely for the purpose of research, typically for a short period, and with limited shared past or anticipated future (e.g., as in experimental studies). Although groups of this nature may not closely resemble organizational teams, studies employing them have closely examined the effects of technology mediation and have strong, consistent findings that are extremely relevant to the case of distributed teams. We reserve the term “team” for groups in organizations. We often abbreviate “geographically distributed teams” to simply “distributed teams,” but in doing so do not intend to include other forms of distributed teams (e.g., organizationally distributed teams) whose experiences with conflict we do not consider in this paper. Finally, we treat as synonymous the terms “traditional teams” and “collocated teams” even though what is new about distributed teams is their increasing prevalence, not their existence. Collocated teams represent how scholars have traditionally conceived of teams. 616 Organization Science/Vol. 14, No. 6, November–December 2003 PAMELA J. HINDS AND DIANE E. BAILEY Understanding Conflict in Distributed Teams Distance and Conflict In this section, we build on research demonstrating that distance has a detrimental impact on team members’ shared context, familiarity, and friendship, all factors that can heighten conflict in teams. We further argue that distance is likely to bring with it increased heterogeneity, particularly cultural differences, that will reduce the similarity of team members. We describe how the negative impact of distance on shared context, familiarity, friendship, and homogeneity will precipitate conflict for distributed teams. Shared Context Because they are distant from each other, members of distributed teams may have difficulty establishing a shared context. Different contexts may derive from and be revealed in different work and geographic environments, different technologies, and different cultures. Occupying different physical contexts makes it more difficult to make and interpret references to objects of interest (Schober 1998) and to co-orient in a particular context. For example, in a study of the use of new machines in a factory, Tyre and von Hippel (1997) observed that engineers and operators had trouble resolving equipment problems over the phone because the engineers needed to “see for themselves” the technology in context. In this way, distance fosters different perspectives on and information about the work in which distributed team members are engaged. In the absence of a shared context, team members will have difficulty developing mutual understanding (Fussell and Krauss 1992, Clark and Brennan 1991). When team members have different understandings of the task, task conflict is likely to result (Jehn et al. 1997). Moreover, when team members’ understanding of the issues differs, conflict is difficult to resolve (Brehmer 1976). Team members who lack a sense of a shared context as a result of distance also are likely to adhere to different norms. Offices, meetings rooms, cafeterias, and so forth are associated with behavioral norms and mental schemas that affect team members’ behaviors and expectations of one another (Kiesler and Cummings 2002). Team members who do not share the same social setting may have different perceptions about what behaviors are appropriate, consequently holding one another to different standards. Karnoe (1995), for example, observed that Danish and American workers used different paradigms for understanding problems and solutions. He attributed these differences to disparities in local routines and behavioral norms. Armstrong and Cole (2002) also observed that site-specific cultures and expectations acted as significant sources of misunderstandings and conflict between distant sites. In short, dissimilar paradigms, norms, and behavioral expectations are likely to result in more task and affective conflict. In addition to disrupting shared understanding and the development of common behavioral norms, occupying different contexts may make it difficult for groups to establish a shared temporal rhythm, or to become “entrained.” Ancona and Chong (1996) argued that groups establish a rhythm that serves as a powerful coordination mechanism. Isomorphic processes at the group level may create similarity in temporal cycles as team members signal to each other the pace and timing of activities (Ancona and Chong 1996). In distributed teams, signals among distant team members may be difficult to observe and interpret. Grinter et al. (1999) found that members of distributed software development teams, regardless of the way they structured their work, were “constantly surprised” and confused about the activities of their distant colleagues. In the absence of triggers available to traditional teams, distributed teams have been observed to use face-to-face meetings and other interactions to establish the rhythm of the team (Maznevski and Chudoba 2000). Such difficulties in developing a shared temporal rhythm may make coordination in distant teams more fraught with conflict as team members continually find their expectations of others unmet and their work processes “out of sync.” We posit that distant team members will experience more incongruent temporal rhythms, which in turn will engender unfavorable attributions (Cramton 2002) and process conflict as confusion arises about who is doing (or has done) what and when. Familiarity Whereas shared context either exists or does not, familiarity can build over time when people are continually copresent. Mere exposure to others powerfully affects peoples’ feelings about one another (Zajonc 1968). As compared to collocated teams, distributed team members tend to receive less passive information about their distant colleagues. Armstrong and Cole (2002) reported that distance blocked casual visual observation, consequently inhibiting learning across sites. Similarly, proximity increases the amount of informal interaction that can occur among team members. When people are collocated, the number of casual encounters, unplanned conversations, and multipurpose interactions increases (Kraut et al. 2002). Such opportunities promote familiarity as team members learn about the personalities, concerns, and work processes of others. Familiarity, in turn, is associated with reduced conflict. Deutsch (1969) reported that lack of familiarity increased conflict Organization Science/Vol. 14, No. 6, November–December 2003 617 PAMELA J. HINDS AND DIANE E. BAILEY Understanding Conflict in Distributed Teams about roles and responsibilities. Similarly, Goodman and Leyden (1991) found that not being familiar with the work habits of other team members increased coordination problems in the team. These studies suggest that process conflict will be greater in distributed teams because their members have fewer opportunities to become familiar with one another. Friendship Proximity also is associated with friendship. Festinger et al. (1950) found that graduate students and their families who were randomly assigned to housing near one another were more likely to become friends. Friendship is easier to establish when people casually encounter one another and interact spontaneously. Grinter et al. (1999) observed that distant team members had difficulty building rapport and developing long-term relationships without meeting face to face. Surprisingly, when team members are friends, conflict, particularly affective conflict, is likely to be more prevalent. In an experiment conducted with business students, Shah and Jehn (1993) found that friend groups experienced more emotional conflict than groups of strangers working on decision-making tasks. Murnighan and Conlon (1991) also reported higher levels of conflict in string quartets in which members were friends, but friend groups were better able to manage conflict successfully. Because bonds of friendship are built on trust, expressing affective conflict may be perceived as safer and more readily accepted. These studies suggest that friendship increases affective conflict in teams, but that these teams also are better able to harness the conflict to improve task performance. In sum, research on friendship suggests that distributed teams will experience less friendship and, thus, less affective conflict. Homogeneity Finally, distance is likely to reduce homogeneity among team members by increasing demographic heterogeneity, particularly ethnic or racial heterogeneity. In a comparison of collocated and distributed product development teams, Mortensen and Hinds (2001) reported that distributed teams were somewhat more culturally diverse than collocated teams. Such diversity has been shown to increase task and affective conflict (O’Reilly et al. 1997, Pelled 1996) because diversity prompts different perspectives on, and approaches to, work and fuels different attitudes, beliefs, and expectations. In their review, Williams and O’Reilly (1998, p. 115) conclude that over 40 years of research has found that “diverse groups are more likely to be less integrated, have less communication, and have more conflict.” We thus expect distant teams to be more heterogeneous and to experience more task and affective conflict. Overall Impact of Distance In sum, the social and psychological effects of distance are likely to lead to more task and process conflict due to challenges resulting from different perspectives, inconsistent norms, incongruent temporal rhythms, reduced familiarity, and demographic heterogeneity. The effect of distance on affective conflict, however, is less immediately apparent. On one hand, distance should lead to more affective conflict as team members adhere to inconsistent norms and attempt to work through demographic differences. On the other hand, distance may lessen affective conflict because team members do not have a basis of friendship that would enable them to express affective conflict openly. Although one might expect these opposing forces to result in little ultimate impact on affective conflict, we argue that distance will heighten affective conflict. Our reasoning lies in the relationship between task and affective conflict. Task conflict can lead to increased affective conflict, especially in teams with low trust, perhaps because low trust leads to more faulty attributions regarding the source of the disagreement (Simons and Peterson 2000). When trust is missing, team members are more likely to question others’ intentions and make attributions that do not adequately account for situational factors. In distributed teams, trust can be fragile and often fractures rapidly (Jarvenpaa and Leidner 1999). Thus, although distance may not directly lead to affective conflict, increased task conflict will result in more affective conflict for distributed teams. This argument is consistent with Cramton’s (2001) observations that reduced information about team members’ actions on distributed teams will lead to more harsh attributions about their intentions. Overall, we propose that distance will engender conflict of all types for distributed teams. Proposition 1. As a result of the different perspectives and norms and reduced homogeneity that it occasions, distance engenders task conflict in teams. Proposition 2. As a result of the different norms and temporal rhythms and reduced homogeneity that it occasions, distance engenders affective conflict in teams, despite the ameliorating effect of reduced friendship. Proposition 3. As a result of the different temporal rhythms and reduced familiarity that it occasions, distance engenders process conflict in teams. Proposition 4. As a result of the reduced trust that it occasions, distance engenders a strong positive relationship between task conflict and affective conflict. 618 Organization Science/Vol. 14, No. 6, November–December 2003 PAMELA J. HINDS AND DIANE E. BAILEY Understanding Conflict in Distributed Teams Technology Mediation and Conflict When distributed team members wish to communicate with one another or work together, they typically employ technology. With limited opportunities for talking face to face, they may hold discussions by phone or e-mail. Unable to carry a document to a colleague’s cubicle, they may resort to electronic options, such as posting it on a Web page or storing it in an intranet-accessible directory. In this manner, technology mediates both communication and collaborative work for distributed teams. Our reading of prior research suggests that the effects of technology mediation can be categorized according to their impact on relational outcomes, information transfer, and coordination. Although significant interest has been paid to the relational outcomes of technology mediation for distributed teams, we conclude that issues of information transfer and coordination may have an equal, if not greater, bearing on group conflict. The bulk of prior research on technology mediation examines use of communication technologies such as computer-supported meeting systems, audioand videoconferencing, and e-mail. Unfortunately, relatively little work considers the use of information technologies such as shared electronic workspaces or version control software that mediate collaborative work absent direct interpersonal communication. Such technologies are equally important for a distributed team’s functioning because team members spend only a portion of their time meeting or otherwise communicating with one another. A few technologies support both communication and information sharing and have been studied (see Mark et al. 1999, Olson and Teasley 1996, Orlikowski 1992). As we assess the impact of technology mediation on group conflict, we consider research on the use of communication technologies and, to the extent possible, information technologies that support collaborative work. Relational Outcomes For several decades, researchers have been concerned about the effect that interacting via technology has on group member relations, including cohesiveness, competition, group behaviors and attitudes, and group identity. One of the earliest theories explaining the relational effects of technology mediation is social presence theory, which argues that interpersonal and group processes are negatively affected when people interact over media that reduce their feeling of “being there” with their communication partners (Short et al. 1976). Social presence theory predicts that this reduced social presence will impair interpersonal relations. Later work claims that mediation via technology reduces social cues and, because of their absence, alters the nature of group communication and group processes. With the potential for a reduction in critical contextual cues (e.g., status and gender), the social context may become less visible, causing people to display more disinhibited behaviors, to become less aware that they are engaged in social interaction, and to tend toward less consensus (Sproull and Kiesler 1991, Siegel et al. 1986). As a result of depersonalized interactions, groups communicating via technology are less likely to exchange relational information than face-to-face groups (Siegel et al. 1986). Other studies show that mediated groups are less cohesive than face-to-face groups (Straus and McGrath 1994), have lower group identity (Bouas and Arrow 1996), and exhibit more competitive behavior (Purdy et al. 2000). In short, mediated communication appears to negatively impact the ability of teams to build and maintain strong interpersonal relationships. Differences among technologies in their ability to support communication have been explained by the media richness theory of Daft and Lengel (1984), which holds that media vary in the richness of the information transmitted and that richer media are more effective at reducing ambiguity and facilitating shared meaning (Daft et al. 1987). Media richness theory suggests that distributed team members might lessen the negative effects of mediation by choosing richer media. The premises of social presence, reduced cues, and media richness theories are challenged by studies that suggest that time can remedy the relational problems that ensue from technology mediation. Many studies have found that mediated groups work more slowly than faceto-face groups (Walther and Burgoon 1992, Weisband 1992). In a meta-analysis of such studies, Walther et al. (1994) found that when task time limits were imposed on these more slowly paced groups, the groups exhibited lower socioemotional communication. When limits were expanded, relational outcomes often improved. There are caveats, however, to the benefits of time. Time does not universally improve socioemotional communication because not all groups are equally willing or able to develop relational closeness (Walther 1994). Also, we suspect that as team membership changes over time, a team’s relational closeness must be reestablished as reconfigured teams learn anew how to communicate effectively via technology. Social presence, reduced cues, and media richness theories also come under fire from scholars who decry the technological determinism these theories represent. Markus (1994a) noted two alternatives to technological determinism, both of which turn attention to technology users, their intentions, and the social context of technology use rather than the material characteristics of the Organization Science/Vol. 14, No. 6, November–December 2003 619 PAMELA J. HINDS AND DIANE E. BAILEY Understanding Conflict in Distributed Teams technologies. The first alternative is the “rational-actor” perspective (Markus and Robey 1988, Kling 1980), which contends that individuals make choices about when and how to employ technologies. Research taking this perspective suggests that at times individuals desire and pursue the distance afforded by technology. They may choose, for example, to employ technology rather than talk face to face when interpersonal relationships are strained (Markus 1994a). A second alternative to technological determinism is the emergent-process view (Markus and Robey 1988, Pfeffer 1982), which holds that the effects of technology use are emergent and thus unpredictable. Given this, individuals may inadvertently worsen the impact of mediation by assuming they can prevent possible ill effects of technology use. Recent extensions of Giddens’ (1984) structuration theory also challenge the technological deterministic perspective by suggesting that people will engage in social practices that produce a particular structure of technology use, which may or may not be consistent with its intended use (Orlikowski 2000, DeSanctis and Poole 1994, Barley 1986). Structuration theory implies that distributed teams may alter their use of existing technologies in ways that will better serve the team. For example, there is evidence that teams find ways to share extensive and detailed information over e-mail systems (Hinds and Kiesler 1995, Lee 1994) and adapt technologies to improve their ability to coordinate better (Kraut et al. 1998). No matter which theoretical perspective one takes, the anticipated effect of technology mediation on group conflict appears to be negative. The technological deterministic perspective suggests that distributed teams will experience greater opportunities for affective conflict as a result of technology mediation. Feelings of not “being there” with one’s communication partners stand to prevent distributed team members from sharing relational information that help teams to develop trust. Fewer inhibited behaviors and a lower tendency for consensus may prompt affective conflict as team members neglect to censor their comments and fail to accommodate their team members’ preferences. The rational-actor perspective highlights the potential for agency among distributed team members, but it fails to significantly lessen the negative implications of mediation because distributed team members, primarily lacking opportunities for face-to-face discourse, can choose only among technologies when wishing to communicate with one another. The rational-actor perspective provides hope, however, that for short periods and for specific purposes, choices that individuals make might deflect the onset of affective conflict. For example, Simons and Peterson (2000) report that task conflict is more likely to lead to affective conflict when voices are raised, suggesting that distributed team members might opt for technologies like e-mail rather than the telephone for discussions in which task conflict is expected. This is consistent with empirical work suggesting that asynchronous communication mitigated negative interpretations of competitive behaviors, perhaps because these behaviors were obscured by the technology (MontoyaWeiss et al. 2001). The emergent-process perspective, which allows for unintended and unanticipated effects of technology use, provides little reason to suspect that technology mediation will not occasion conflict in distributed teams. Lee (1994), for example, illustrated that e-mail use is socially embedded and that e-mail users add meaning to the messages that they receive. However, the series of messages Lee examined also highlighted the potential for conflict, as Lee noted when describing the “politically sensitive and managerially troublesome meaning” (p. 153) occasioned by the initial e-mail. Finally, with regard to structuration theory, we concur with the conclusion of Kraut et al. (1998), that although users may modify technologies to suit their needs, technologies possess certain material limits that cannot be overcome. As evidence, Markus (1994b) showed that after the introduction of a new e-mail system, users felt that their interpersonal relationships were weakened even though they regularly used the telephone to “keep in touch.” Although structuration theory implies that teams can limit the impact of technology on relational outcomes, the material properties of technologies, combined with the challenges of distance, render unlikely the possibility that distributed teams will be able to modify the technology adequately and consistently enough to match face-to-face communication. In short, technology mediation engenders negative relational effects that we contend will precipitate affective conflict. These effects, including reduced cohesion and group identity, increased competitiveness, reduced consensus, and less sharing of relational information, contribute to lower trust, familiarity, and a sense of belonging on the team, and ultimately, we argue, induce affective conflict. Because the effects we have considered here concern group relations, we expect that their impact is primarily on affective conflict, although task and process conflict may be indirectly affected. Information Transfer Technology mediation also impacts information transfer among team members. Several problems related to information sharing and seeking emerge from the literature, including uneven distribution of information, unevenly 620 Organization Science/Vol. 14, No. 6, November–December 2003 PAMELA J. HINDS AND DIANE E. BAILEY Understanding Conflict in Distributed Teams weighted information, and information that resists transmission. Each problem portends negative effects with respect to conflict. Uneven distribution can occur in at least two ways: Team members may be purposely or accidentally excluded from communications, or members may not reveal information that they uniquely hold. Work by Cramton (2001) highlights problems that can arise from exclusion. In her study of distributed student teams, not all members of the teams were copied on e-mails sent by one team member or another. As a result of this limited distribution, some team members worked with incomplete information while their colleagues assumed information had been universally shared. Conflict arose in the confusion that ensued. To complicate matters, team members attributed disagreements and miscommunications to individual-level factors rather than to the technology or the situation, a practice that further fueled interpersonal friction. The technologies upon which distributed teams rely vary in the degree to which they promote inclusion and prevent exclusion. Audioand videoconferencing, as well as Web-based meeting systems, support inclusion by allowing distant members to “attend” meetings, but they fail to prevent exclusion because they cannot guarantee that all members will be notified of the meeting. Additionally, Web-based systems like NetMeeting make it difficult for participants to keep track of who is remotely included in a meeting (Mark et al. 1999). Cramton’s (2001) study reveals that technologies like e-mail, despite having features that support inclusion, may be more apt to facilitate exclusion of various team members, either through sender intent (acting as a rational actor) or mistake (in line with the emergent-process view of mediation). Her work points out the limitations of a purely technological determinist perspective by showing that even advanced features cannot ensure the uniform sharing of information. Uneven distribution of information also results when team members fail to share uniquely held information. Information exchange is less complete and more biased in mediated groups as compared with face-to-face ones, and mediated groups are less likely to uncover information uniquely held by one member (Hollingshead 1996). Group members who rely on communication technologies find it more difficult to cue one another for the information they need and to interpret the cues being conveyed by other group members (Hollingshead 1996), which may be why unique information is not shared. The problem of incomplete information sharing, especially of uniquely held information, may be particularly detrimental to distributed teams because organizations often rely on such teams as a means of assembling expertise from a variety of distant locations. Technology mediation also can impact the weight that various team members place on different pieces of information. Cramton (2001) noted that despite whatever importance a sender intended to attach to various topics within a single e-mail, team members often assigned different salience to the topics. As a result, some topics never received the attention the sender desired for them and at times were entirely overlooked, which led to frustration and misunderstanding. Finally, some information is not readily transmitted via technology. Certainly some types of information, particularly those that can be digitized, can now be more easily transmitted via technologies such as file-transfer protocols and electronic workspaces, thus increasing the amount of this information that can be shared with distant teammates. However, many types of data continue to resist transmission via technology, with negative implications for conflict in distributed teams. This is particularly true in the case of contextual information that leads to shared awareness, such as who is in the office, what they are doing, what problems they are confronting, and the moment-to-moment social dynamics of the workplace. Weisband (2002) reported that although some mediated groups using a Web-conferencing system and e-mail were successful in conveying contextual information, many were not. Differences also exist in the degree to which technologies facilitate the exchange of contextual information (Olson and Olson 2000, Clark and Brennan 1991). As compared with the phone, which transmits little contextual information beyond background noises, newly developing awareness technologies that display information such as time zones, holidays, and current availability by type of medium for team members around the world are far better at aiding shared awareness (Atkins et al. 2002). Even with these technologies, considerable contextual information, such as unplanned actions that occur away from one’s desk, remains untransmitted. Research on technology mediation’s information transfer effects suggests that distributed teams will be prone to conflict. Uneven distribution of information implies that team members will work and communicate on the basis of different information. As a result, they will be unlikely to recognize or resolve differences in perspectives, which have been shown to increase task conflict in groups with weak relationships (Brehmer 1976). Because technology impedes the ability of distributed teams to collect contextual information, these different perspectives may become entrenched, increasing the possibility of task conflict. Process conflict also Organization Science/Vol. 14, No. 6, November–December 2003 621 PAMELA J. HINDS AND DIANE E. BAILEY Understanding Conflict in Distributed Teams is likely to arise when some members are excluded from messages. The progression of the team’s work, seemingly purposeful and rational for members who receive all messages, may appear disjoint to those members who do not. Members who lack complete information may call into question the team’s methods and trajectory. Finally, affective conflict, already engendered by mediation’s relational effects, also may derive from information transfer effects. Uneven distribution of information, which can lead to frustration and misplaced blame, increases the potential for affective conflict. Coordination Coordination constitutes a third aspect of communication and collaborative work that is impaired by technology mediation. Purdy et al. (2000) reported that student groups working face to face collaborated more than distributed groups working over video, telephone, or chat. Moreover, collaborative efforts were less likely to be perceived in the mediated conditions. Field studies of distributed teams further suggest that technologies designed to increase shared awareness may exacerbate coordination problems. For example, an awareness technology that allowed team members to see what documents other team members had read resulted in team members opting not to read what others had in an effort to avoid duplication (Espinosa et al. 2000). Consequently, the team was limited in its ability to discuss materials. Technology mediation also may induce time lags and sequencing problems that further hamper coordination. Cramton (2001) found that distributed teams using online chat to hold discussions with members around the globe experienced time lags that delayed some members’ comments, rendering them “out of sync” with the larger conversation. With the sequence of responses disrupted, members may think their comments have been ignored, prompting frustration and irritation. Cramton’s (2001) study also highlighted problems with asynchronous communication. Disparities in when messages were sent, received, and responded to resulted in individuals working with different information at different times. The coordination problems imposed by technology mediation may precipitate conflict of all types. Process conflict is likely when the use of technologies renders some team members “out of sync” and makes it difficult to coordinate use of shared resources. Incompatibilities in work processes may give rise to disagreements about how work should get done. Coordination problems also may occasion task conflict, as can be expected when individuals have not examined the same materials as their colleagues or when they work on the basis of different information. Finally, the inability to coordinate work may lead to frustration and misconceptions, which in turn feed the potential for affective conflict. Overall Impact of Technology Mediation In short, notwithstanding the different affordances of technologies and the ability of users to make choices about how and when to use them, technology mediation has implications for conflict in distributed teams. Teams using the most advanced technologies experience difficulties, as do teams that have a host of technologies available to them. Just because one technology is found to be superior to another does not mean it is absent negative outcomes; research to date reveals that problems arise from nearly every available technology. The coordination problems we mention, for example, were uncovered with respect to computer conferencing and support systems, awareness technologies, e-mail, and shared workspaces. Moreover, advanced technologies are accompanied by their own litany of usability problems that compound the difficulties of mediation (see Fish et al. 1993, Gaver et al. 1992). Success of groupware tools, for instance, can be dependent on achieving a “critical mass” of users and on users’ willingness to enter data into the system (Atkins et al. 2002). NetMeeting allows distant members to “attend” meetings, but Mark et al. (1999) found that many members could not participate because they were late in implementing the technology or had no one at their own site to consult with about the technology. Although technological advances may lessen a particular impact of mediation, it seems unlikely that they will ameliorate all such impacts and rather likely that they will occasion new ones. On these grounds, we propose that technology mediation engenders conflict of all types in distributed teams. Proposition 5. As a result of the uneven information and difficulties in coordination that it occasions and its inability to transmit certain information, technology mediation engenders task conflict in teams. Proposition 6. As a result of the negative relational effects, uneven information, unevenly weighted information, and difficulties in coordination that it occasions, technology mediation engenders affective conflict