A multilevel review of stressor research in teams

J Organ Behav. 2020;41:185–209. Summary The contemporary work environment is characterized by an ongoing trend to embed employees in teams because of their expected abilities for handling complex tasks and integrating diverse sets of knowledge and skills. However, leveraging this potential is endangered by stimuli within and outside of teams that take a toll on cohesion and teamwork among team members. Understanding the role that stressors and demands play in the work‐related functioning of teams and their members is therefore an increasingly important challenge in the organizational behavior literature. Whereas research on stressors and demands has primarily focused on the individual level, we expand the research scope by considering these phenomena to be multilevel. We perform an interdisciplinary review of the literature on these stimuli in teams and show how related research, such as that on destructive leadership, may benefit from a more balanced account and integration of frameworks on stressors. Our multilevel review is informative for the literature on stressors and demands at the individual and team levels, as it offers an important conceptual grounding for how and why various stimuli in this social environment differentially influence both the collective entity and its individual team members.

Studies accounting for higher level stressors have revealed the complex nature of team stressors, identifying some team stressors as beneficial for team performance and others as detrimental (Drach-Zahavy & Freund, 2007;Gardner, 2012;Pearsall et al., 2009). Studies on cross-level effects have shown that among other factors, team members' exhaustion and engagement depend on the stressfulness of their team's climate (i.e., ranging from distressed to eustressed; Kozusznik et al., 2015). Most of these findings are in line with findings at the individual level showing that stressors differentially affect outcomes such as performance, commitment, and engagement (for reviews at the individual level, please refer to Bakker, Demerouti, & Sanz-Vergel, 2014;LePine, Podsakoff, & LePine, 2005;Podsakoff, LePine, & LePine, 2007).
Notably, however, only a few studies conducted in actual team contexts have explicitly drawn upon and theoretically extended stressor frameworks that were originally conceptualized at the individual level (e.g., Ellis & Pearsall, 2011;Pearsall et al., 2009). This is surprising, given that the need for multilevel approaches in occupational stress research was articulated long ago . Consequently, although team research is gradually expanding the knowledge of how teams are affected by specific stimuli (e.g., time or performance pressures), the development of explicit stressor research in teams is potentially stalled. This is particularly critical because the bifurcation of literature leading to two separate streams (i.e., micro and macro) impedes the advancement of knowledge (House, Rousseau, & Thomas-Hunt, 1995), thereby creating unnecessary pluralism (Goldspink & Kay, 2004). A more integrative approach to the study of stressors and teams is hence needed to effectively use any potential synergies currently lying dormant.
By extending the informative literature reviews of workplace stress (Ganster & Rosen, 2013) and stressors (Cooper et al., 2001) at the individual level, we provide a unified overview of the empirical research addressing stressors and demands in teams to allow such synergies to surface. Our contribution is thus threefold. First, as team research on stressors appears to be unstructured in terms of a rather random selection of stressors, outcomes, and potential buffering mechanisms, we align and structure previous findings to put scholars who are interested in the same specific phenomena within those categories (e.g., stressors intrinsic to a team's job or originating from relationships at work) on the same page. This approach will connect their research ideas to findings from potentially related stressors. Second, although research on work stressors has become increasingly invested in the study of collectives, much of the discussion is implicitly applied as analogous to the individual level. However, especially when stressors affect individuals embedded in collective structures such as teams, interactions and team-internal processes are pivotal to the effects of the stressors. To avoid an overly restricted and static understanding of stressors and demands at work, we systematically review the literature across multiple levels of analysis and discuss the multilevel nature of stressors (Bliese, 1998;. Third, as scholars from related fields of team research arguably investigate stressor-like phenomena without explicitly drawing on this stream of research, we raise their awareness of a more balanced account of team-stressor research. We exemplify this by the literature on destructive leadership behaviors (e.g., Mawritz, Dust, & Resick, 2014;Tepper, 2007). Applying a team-stressor lens to these behaviors will give rise to substantial progress that would otherwise be overlooked.
Taken together, we contribute to the existing literature by explicitly focusing on stressors in collective entities (e.g., Maruping et al., 2015;Rodríguez-Escudero, Carbonell, & Munuera-Aleman, 2010;Savelsbergh, Gevers, Van der Heijden, & Poell, 2012), and we advance related streams of team research that center on stressful stimuli jeopardizing the proper functioning of teams and their members.

| Theoretical approach and structuring
Because teams and work groups are characterized as collectives operating in settings that link individuals to one another (Ilgen, Hollenbeck, Johnson, & Jundt, 2005;McGrath, Arrow, & Berdahl, 2000), we integrate studies on both forms of interdependent collective work in our systematic literature review. Some authors consider their studies to relate to teams exclusively, but blurring the lines between the two terms may provoke others to classify these studies as research on work groups, or vice versa. In keeping with previous reviews (e.g., Cohen & Bailey, 1997;Costa, Fulmer, & Anderson, 2018;Thatcher & Patel, 2012), we use the terms "team" and "group" interchangeably for simplicity's sake, although we more frequently use the term team.
As Kerr and Tindale (2004, p. 624) noted, "the distinction is a rather artificial one that reflects more about subdisciplinary territoriality than about fundamental differences in focus or objectives." If a distinction is to be made, then team studies typically, but not exclusively, tend to be of an applied nature, with data collected in real work teams, whereas group studies tend to be of an experimental nature in that they investigate groups in laboratory settings (Kerr & Tindale, 2004;McGrath et al., 2000).
Because research on stressors in the organizational context of teams appears to be rather unstructured, the approach to organizing our interdisciplinary literature review is twofold. On the one hand, we classify the extant empirical research on stressors and their effects on and within teams into three broader categories reflecting the hierarchical levels considered. First, there are studies investigating individual-level stressors and their effects on team members (e.g., Jex & Thomas, 2003). These studies apply a single-level approach by observing individuals embedded in teams and how individual-level stressors affect their work. Second, by applying this single-level approach to the higher level, that is, the team level, some studies consider the team-specific and team-external influences that affect collective entities as a whole (e.g., Ellis, 2006;Maruping et al., 2015). This vein of research has started to emerge due to the increasing relevance of teams in organizational practice. Third, studies perform multilevel investigations of stressors within teams. Articles simultaneously accounting for individual-level and team-level models (i.e., homologous multilevel models; Klein & Kozlowski, 2000; e.g., Griffith, 1997;Sacramento et al., 2013) must be differentiated from those investigating actual cross-level effects of either higher level stressors or buffering resources on outcomes pertaining to individual team members (e.g., Savelsbergh et al., 2012;Triana, Porter, DeGrassi, & Bergman, 2013).
On the other hand, within each of these broad categories structured around the hierarchical levels, we cluster the reviewed studies according to the categories of the stressors they examine. We follow the categorization of workplace stressors developed by Cooper and Marshall (1976) in which stressors affecting employees in their work typically fall into one of the following six categories: (a) stressors intrinsic to a job (e.g., workload, time pressure); (b) stressors due to relationships at work (e.g., interpersonal animosity, poor leadership); (c) stressors due to roles in the organization (e.g., role ambiguity, role conflict); (d) stressors related to career development (e.g., job insecurity, thwarted ambition); (e) stressors due to organizational structure and climate (e.g., politics, lack of participation); and (f) stressors at the work-home interface (e.g., family problems, financial difficulties).
Drawing upon this well-established categorization facilitates the integration of previous reviews of stressor research at the individual level (e.g., Cooper et al., 2001) with the review of the research on teams presented in our article.

| Literature search
Before demonstrating how this article's multilevel approach fits into and contributes to the existing research on stressors in teams, we delve into an interdisciplinary review of the previous work carried out in the three broader categories outlined above. In line with best practices (e.g., Hodgkinson & Ford, 2014;Newman, Ucbasaran, Zhu, & Hirst, 2014;Short, 2009), we first systematically searched for reasonable combinations of relevant key terms relating to contexts (e.g., "team," "group," and "collective") and stressors (e.g., "stressor," "demand," and "stimuli" and specific stressors such as "time pressure," "workload," and "ambiguity") in Web of Science, EBSCO (i.e., Business Source Complete, PsycARTICLES, and SocINDEX), and Google Scholar. We then screened the results of our search by reading the abstracts and consulting the references of the remaining research to identify further research that was undetected by our initial search strings. We identified 90 empirical journal articles, one dissertation, and one book chapter for our multilevel literature review, with the most recent study having been published in June 2019 (i.e., Shen, Chang, Cheng, & Kim, 2019). As illustrated in Figure 1, the body of research on stressors in teams has not only persistently grown over the last quarter century but changed its perspective since the turn of the millennium, with an increasing proportion of the research now considering multiple hierarchical levels (i.e., the individual "and" team levels) rather than only a single level (i.e., the individual "or" team level).
Given the nature of the topic, our systematic literature review is highly interdisciplinary. Figure 2 provides an illustrative distribution of the 90 reviewed journal articles across disciplines. Whereas 14.4% of the reviewed articles were published in business and management journals (e.g., Academy of Management Journal and Administrative Science Quarterly), 41.1% were published in psychology journals (e.g., Group Dynamics: Theory, Research, and Practice and Journal of Occupational Health Psychology). Roughly one-third of them (30.0%) were published in organizational behavior journals (e.g., Journal of The relative evolution of research on stressors in teams by hierarchical level FIGURE 2 The number of reviewed journal articles across disciplines Applied Psychology and Journal of Organizational Behavior), that is, in journals that are interdisciplinary in nature and thus ranked by the Social Sciences Citation Index within the fields of both management and psychology. The remaining articles (14.4%) were published in journals from other disciplines, such as sociology, hospitality, and ergonomics.

| Stressors intrinsic to a job
Research on individual-level stressors and their effects on team members has identified different causes of stressors intrinsic to team members' jobs. Workload and time pressure are certainly the most prominent stressors from this category. For example, team decisionmaking and responsibility were shown to increase and team-based job rotation to decrease both of these stressors (Cruz & Pil, 2011).
These stressors, in turn, tend to translate into heightened experiences of stress (Cruz & Pil, 2011), anxiety (de Jonge, van Breukelen, Landeweerd, & Nijhuis, 1999, and acute job strain (Gevers, van Erven, de Jonge, Maas, & de Jong, 2010). In team-based settings, this "strainbased process" is critical because strains-particularly from the emotional domain-tend to impair teamwork behaviors (Gevers et al., 2010). These findings are in line with those from the military context showing that individual-level stressors (such as overload) negatively affect team members' group perceptions, job satisfaction, and wellbeing (Jex & Thomas, 2003). Worsening matters, pressures intrinsic to the work of teams seem to disrupt team members' perceptions of team cohesiveness and to increase their competitive mindset (Klein, 1996), which potentially leads to a dysfunctional upward spiral in which stress and competition (or disadvantaging behaviors) mutually reinforce one another. As a pleasant side effect, however, team members in highly competitive teams were shown to experience less pressure from other stressors (Sonnentag, Brodbeck, Heinbokel, & Stolte, 1994), arguably because the existence of permanent competition constitutes a chronic stressor that makes team members more used to (and robust against) stressful events.
Although these studies consistently show the detrimental effects of stressors intrinsic to team members' tasks at hand on their wellbeing, empirical evidence also suggests that the decision-making process in teams becomes more decentralized under excessive stressors.
When experiencing stressors in the military context, for example, team leaders and members become more open and willing to accept input from one another, thereby decentralizing authority within the team (Driskell & Salas, 1991). Likewise, in the student context, Brown and Miller (2000) found less centralized communication between psychology students when their teams were working on highly complex and taxing tasks, pointing to a "behavior-based process" among team members that resulted from their individual experience of stressors. In addition to individual-level stressors that are potentially beneficial for team members, buffering resources were also shown to help mitigate the potential detrimental effects that other, more negative stressors may have. For example, in team-based care, team members' job autonomy was revealed to buffer the negative effect that jobintrinsic stressors had on the work motivation of team members (de Jonge et al., 1999). Other studies with designs that account for team-level buffers assessed via team members' individual perceptions (and therefore are not conceptualized as cross-level studies, such as those reviewed below) have shown that the strain imposed on members of army companies (i.e., a team-based military unit) by jobintrinsic stressors is effectively buffered when they have strong selfefficacy beliefs . Likewise, the sheer perception of support from others is shown to be functional in this regard (Bliese & Castro, 2000). Evidence from student teams suggests that when such support is not available, team identification can compensate for this deficit by buffering the strain-related effects of individual-level stressors (Jimmieson, McKimmie, Hannam, & Gallagher, 2010). This identification seems to have its most beneficial stressor-buffering effect on satisfaction for prototypical team leaders (Cicero, Pierro, & van Knippenberg, 2007), which shows that the complex interrelationships and dependencies between team leaders and members are critical to the study of stressors and demands in teams.

| Stressors due to relationships at work
Research on stressors resulting from team members' relationships at work has predominantly focused on the role of social identification given the unique social structure of teams and work groups. Extant individual-level investigations of the work relationships of team members have hence emphasized that identity-based processes play a pivotal role not only in stressors' effects on performance at the team level (as shown below) but also in team members' individual experience of stress (Haslam & Reicher, 2006). In this regard, members of teams with a strong collective sense of team identification favor the "we" over the "I" (Ashforth & Mael, 1989;Duffy, Scott, Shaw, Tepper, & Aquino, 2012). Reflecting the idiom that "when 'I' is replaced by 'we,' even 'illness' becomes 'wellness,'" teams and their members can profit from a strong shared sense of identification. More specifically, such a shared identity should not only motivate team members to provide more support for their own in-group (Haslam & Reicher, 2006) but also trigger feelings of team cohesiveness. In the student context, this sense of a shared social identity, which makes a team cohesive, was shown to translate into less stress, confusion, and mood disturbances among the members of a varsity team (Henderson, Bourgeois, LeUnes, & Meyers, 1998) and to compensate for the neuroendocrine stress reactions of students collectively working on an experimental task (Häusser, Kattenstroth, van Dick, & Mojzisch, 2012). Because identification with a particular team is easier for team members to develop when they are not required to work simultaneously in more than one team, it is unsurprising that Pluut, Flestea, and Curşeu (2014) Team members' time pressure and workload, which translate into stress, seem to be triggered by team decision-making and responsibility, but to be decreased by team-based job rotation.
-… conceptualized as individual-level predictors: Bliese and Castro (2000) 1,538 U.S. Army soldiers of 53 companies (U.S.) Overload tends to induce strain, and this effect is shown to be efficiently buffered by role clarity only in high-supportive teams.
- Brown and Miller (2000) 216 psychology students in 48 teams (U.S.) Communication seems to be more centralized in teams working on tasks of low complexity (irrespective of time pressure).
- Driskell and Salas (1991) 78 U.S. Navy students in 39 teams (U. S.) When being stressed, team members seem to become more receptive to information provided by others. Acute job demands tend to impede effective teamwork behavior only when they result in acute job strain. This detrimental effect is shown to be particularly true for strain from the emotional domain. Team members with strong (vs. weak) self-efficacy seem to be less (vs. more) psychologically and physically strained by long work hours and work overload, and they appear to be more satisfied with their job when working on tasks of high significance. Overload, interpersonal conflict, and work-family conflict tend to impair team members' job satisfaction, well-being, and group perceptions. Highly team-identified members seem to be less negatively affected by role ambiguity in their job satisfaction. Strong team identification seems to buffer against the harms of role ambiguity for their psychological well-being only when team members lack co-worker support.
- Klein (1996) 1,676 employees from 6 plants (U.S.) Work pressure tends to disrupt perceptions of team-cohesive behaviors and to trigger perceptions of intrateam competition.
- Sonnentag et al. (1994) 180 software professionals from 29 teams (Germany and Switzerland) The experience of stressors like overload seem to decrease team members' identification with their team and increase their perceived pressure. While team members are even less identified when cognitive requirements are high, they tend to be more pressured when cognitive or learning requirements are high or when there is low competition.
-… conceptualized as individual-level moderators: Cicero et al. (2007) 329 employees from 3 organizations (Italy) Team leaders' prototypicality is shown to relate more strongly to their members' job satisfaction when team members experience more stress and identify more strongly with their team.
-Stressors due to relationships at work … … conceptualized as individual-level criteria: Haslam and Reicher (2006) 15 adult men in 2 teams (U.K.) A shared identity not only appears to increase the provision of more social support, but also to allow for effectively resisting the adverse effects of situational stressors like inequality.  Team members perceiving strong team cohesion tend to experience less stress (e.g., depression, confusion, mood disturbance).
-… conceptualized as individual-level mediators: (Continues) found that teamwork demands in the organizational context strain team members more who possess multiple team memberships.
Although the social relationships within a team can serve as a valuable resource for its members when a shared identity is able to arise, work relationships also bear the risk of creating conditions under which individual team members are harmed. Specifically, Haslam, Jetten, O'Brien, and Jacobs (2004) show that team members experience more stress when the bearer of a stressful message is from their in-group (i.e., their own team) than from an out-group, which reflects existing research showing that discouragement from in-group members rather than out-group members provokes downward performance spirals (Rees et al., 2013). Moreover, perceptions of team-internal competition in student teams are known to function as a stressor that leads to disadvantaging behaviors toward fellow team members, more stress, and less calmness and satisfaction (Boos, Franiel, & Belz, 2015). Notably, social stressors associated with cultural differences between members of binational teams have been shown to be more stressful than actual work-related stressors (Pines & Zaidman, 2014). Thus, the social context in which team members work obviously influences the severity of the effects of a stressful event.

| Stressors at the work-home interface
Our review shows that most research on social stressors in teams considers team-internal stressors resulting from relationships at work.
However, individual-level stressor research reveals that the social con-

| Stressors intrinsic to a job
Although studies that investigate the effects of team stressors exclusively at the team level have grown in popularity over the last two While taskwork demands (i.e., task load) and both teamwork demands (i.e., team process load and conflict) increase team members' job strain, only the teamwork demands appear to be enhanced by multiple team membership.
-… conceptualized as individual-level predictors: Boos et al. (2015) 200 students in 20 teams (Germany) Intrateam competition seem to provoke disadvantaging behaviors, more stress, and less calmness and satisfaction.
- Haslam et al. (2004) 40 psychology students in team-based structures (U.K.) For team members, stressful messages from in-group members appear to be more stressful than that provided by members of an out-group. For challenging messages, in contrast, more stress is created when it is delivered by out-rather than in-group members.

Pines and Zaidman
Social stressors associated with cultural differences between team members tend to be significantly more stressful than the stressors associated with working with members of the same culture. -

Stressors at the work-home interface … … conceptualized as individual-level predictors:
ten Brummelhuis et al.
495 team members of 95 teams (the Netherlands) Team members experiencing private conflicts with their partners at home show less helping behaviors toward their fellow team members at work.
decades, research on the antecedents of stressors in teams is still underrepresented. Beyond research showing that the provision of information is helpful in avoiding overload and ambiguity in teams experiencing organizational change (Rafferty & Jimmieson, 2010), only two studies to date have investigated the antecedents of team-level stressors by examining the role of team composition in particular.
Whereas Hudson and Shen (2018) focused on actual staffing within teams and showed that understaffing influences teams' quantitative and qualitative workload, Keller (2001) examined team composition from a team-diversity perspective. His research revealed that crossfunctional teams are more stressed than their homogeneous counterparts and that this stress taxes their performance by impairing team cohesiveness, thus pointing to an "identity-based process" by which team stressors affect team performance.
Empirical evidence of the consequences of workload and time pressure experienced by an entire team is mixed. On the one hand, team workload tends to undermine the performance and effectiveness of teams (Brown, 2011). This is partly because workload appears to inhibit teamwork (Entin & Serfaty, 1999) (Ellis, 2006) and to impair the accuracy of such mental models and the allocation of information within teams (Ellis & Pearsall, 2011). From a knowledge perspective, such performance decrements under pressure were shown to be transmitted by team members' tendency "to overly rely on general expertise while discounting domain-specific expertise" (Gardner, 2012, p. 1), which points to an "information-based process" that enables team stressors to affect team performance.
On the other hand, empirical findings suggest that quantitative stressors via team commitment and qualitative stressors via organic structuring positively influence team effectiveness (Drach-Zahavy & Freund, 2007), which can be explained by experimental findings from the student context. Student teams tend to develop relatively pronounced cohesion when they must collectively function under fear (Morris et al., 1976), and moreover, the structuring of teams into non-specialized rather than specialized team members appears to counteract the performance decrements associated with elevated degrees of workload (Urban, Bowers, Monday, & Morgan, 1995).
Because appointing leaders in positive physical environments has been shown to enhance team members' attitudes toward their team (Worchel & Shackelford, 1991), student teams appear to make use of their leaders' structuring behaviors that are typically triggered by more complex tasks (Marta, Leritz, & Mumford, 2005). This is in line with findings from the organizational context showing that the team orientation of team leaders' supervision tends to buffer the negative effects of work pressures on team cohesion (Klein, 1996). Finally, Urban, Weaver, Bowers, and Rhodenizer (1996) show that work overload caused by a lack of resources does not necessarily degrade the performance of student teams, although it triggers more complaints about the availability of team resources. This finding resonates with the organizational phenomenon of team members potentially becoming self-declared victims of their restrictions (Hoegl, Gibbert, & Mazursky, 2008).
Considering these equivocal findings, it is unsurprising that in the years following the establishment of the challenge-hindrance stressor framework at the individual level (LePine et al., 2005;Podsakoff et al., 2007), team studies on the dual nature of team stressors began to accumulate in organizational behavior research. Whereas challengerelated team stressors were found to improve the performance and transactive memory of student teams, hindrance-related team stressors tend to impair both (Pearsall et al., 2009). These differential performance effects of team stressors were identified as being transmitted through the coping strategies used by student teams (Pearsall et al., 2009)  Many investigations of time pressure in teams have not only applied experimental manipulations of time pressure in teams of students but also drawn on the attentional focus model (Karau & Kelly, 1992). Providing further support for an information-based process of team stressors, this model suggests that members of time-pressured teams focus more on tasks and their completion, thereby filtering out other information that they perceive as less important to accomplishing a task successfully. This is partly the case because although time pressure results in elevated communication quantity and quality (Pfaff, 2012), it tends to increase normative and decrease informational influences in teams (Kelly, Jackson, & Hutson-Comeaux, 1997). However, whereas only one further study (Kelly & Loving, 2004) found support for such a filtering-out effect in student teams, others failed to provide empirical evidence that time pressure directly affects teams' information seeking (e.g., Durham, Locke, Poon, & McLeod, 2000;Kelly & Karau, 1999;Parks & Cowlin, 1995). The process of filtering information-which would explain why teams under time pressure often show lower decision-making quality and decreased accuracy-hence appears to be more complex. It is thus unsurprising that the effect of time pressure on student teams' decision-making has been revealed to be contingent on the initial decision preference of team members (e.g., strong and shared with others; Kelly & Karau, 1999).
Likewise, in the organizational context, two more recent studies examined the more complex effects of time pressure. This research revealed time pressure to be beneficial for team performance at moderate levels (i.e., having an inverted U-shaped relationship) because learning orientation and knowledge sourcing (Khedhaouria, Montani, & Thurik, 2017) as well as specific team processes (i.e., team transition, action, and interpersonal processes; Maruping et al., 2015) are efficiently activated at such levels of time pressure. The latter indirect effect was found to vary with the level of team temporal leadership (i.e., the structuring, coordination, and management of task pacing in teamwork; Maruping et al., 2015).
Whereas traditional stressor frameworks conceptualize stressors as the causes (i.e., predictors) of the stress process, studies conceptualizing stressors intrinsic to a team' job as moderators are equally informative to our understanding of stressors in teams. Similar to the direct effects of team workload discussed above, this job characteristic appears to have opposite contingency effects. Whereas work overload seems to inhibit the positive relationships that team resources have with both team resilience and team performance (Meneghel, Martinez, & Salanova, 2016), it may constitute a condition under which task conflicts benefit team performance (Bang & Park, 2015).
Similarly, although work overload tends to limit the positive effects that team improvisation has on team performance (Magni & Maruping, 2013), it may strengthen the positive effect of team reflexivity on team innovation (Schippers, West, & Dawson, 2015). Although workload does not seem to alter the degree to which familiarity among team members is beneficial for a team's success in the military context (Espevik, Johnsen, & Eid, 2011), it hampers the positive effect of a team's adaptation and coordination training on its performance (Serfaty, Entin, & Johnston, 1998 Thompson, & Pigeau, 2000) but also the positive potential of team cohesion for the performance of student teams (Zaccaro, Gualtieri, & Minionis, 1995). Thus, in contrast to the research that conceptualizes stressors as predictors, team-level studies conceptualizing job characteristics such as work overload (Magni & Maruping, 2013;Schippers et al., 2015) and time pressure (McCann et al., 2000;Zaccaro et al., 1995) as moderators rather than predictors suggest that individual-level stressor frameworks are still insufficiently accounted for at-and extended to-the team level.

| Stressors due to organizational structure and climate
The way in which organizations structure both their teams and the organizational context constitutes an important stressor affecting organizations' efficiency. Initial research has pointed to the adaptive responses of leaders who must cope with understaffed teams, showing that headcount understaffing tends to trigger more initiating structure behaviors of team leaders, whereas expertise understaffing results in more consideration behaviors (Shen et al., 2019). In terms of the workplace climate affecting a team's proper functioning, conformity pressure in student teams was shown to stifle team creativity when team members were creative and to boost it when they lacked such creative talent (Goncalo & Duguid, 2012

| Stressors at the work-home interface
Finally, initial research on team-level stressors associated with the work-home interface has revealed that family demands on teams decrease teamwork and team performance (through reduced task work), but their detrimental effect on teamwork appears to be effectively buffered by the support of both coworkers and supervisors (ten Brummelhuis, Oosterwaal, & Bakker, 2012). Table 2 presents a detailed summary of the team-level studies reviewed above.

| Homologous multilevel investigations of stressors and demands
Replicating lower level stressor effects at the higher level of analysis is the main goal of research applying homologous multilevel models. A further distinction is hence needed for multilevel investigations of stressors within teams because articles simultaneously accounting for individual-and team-level models (Klein & Kozlowski, 2000) must be differentiated from those evaluating cross-level effects of teamlevel predictors on individual-level criteria.  While time pressure and threat may reduce both mental model accuracy and information allocation in teams, such stressors are shown to enhance tension among team members. In cross-trained teams, the former effects appear to be less negative and the latter effect to be less positive.

Experimentally manipulated
Entin and Serfaty (1999) 6 five-person teams of naval officers (U.S.) Whereas workload tends to inhibit the performance of and teamwork within teams, it seems not to affect the associations between team training and both these outcomes.

Experimentally manipulated
Fritzsche et al. (2014) 56 car assembly teams with 623 team members (Germany) Physical workload tends to increase the absenteeism duration of teams.

Key-informant design
Gardner (2012) 72 audit and consulting teams (U.S.) Although performance pressure appears to generally increase team performance, such pressured teams are also more likely to engage in performance-detracting behaviors as they tend to rely on general rather than domain-specific expertise.
Key-informant design Karau and Kelly (1992) 36 three-person teams of psychology students (U.S.) Time-pressured teams tend to focus proportionally more on direct task activities than on non-task activities, whereas it is the other way round for teams experiencing little time pressure.
Experimentally manipulated Kelly et al. (1997) 164 three-person teams of psychology students (U.S.) Time pressure in teams tends to increase their normative influence and to decrease their informational influence.
Experimentally manipulated Kelly and Karau (1999) 71 three-person teams of psychology students (U.S.) Time pressure is shown to trigger teams to work at a faster rate and to focus more on the completion of tasks.
Experimentally manipulated Kelly and Loving (2004) 80 three-person teams of students (U. S.) Although time pressure seems not to affect the recall of information discussed in teams (i.e., refuting an encoding process), time-Experimentally manipulated (Continues) Teams having to work with more complex task descriptions tend to produce plans of both higher quality and originality, partly because task complexity appears to trigger structuring behaviors of team leaders.
Experimentally manipulated Maruping et al. (2015) 111 teams with 1,115 team members from one software firm (U.S.) The effect of time pressure on team processes, which promote overall team performance, seems to be inverted U-shaped such that a certain degree of time pressure is needed to realize optimal performance.
Referent-shift consensus model Parks and Cowlin (1995) 81 four-person teams of psychology students (U.S.) Although time pressure tends to inhibit the repetition of information, it is shown to not bear on the introduction of facts. Time pressure is shown to increase the quantity of and the efficiency in the communication between team members.
Experimentally manipulated Urban et al. (1995) 24 five-person teams of students (U.S.) Overload tends to impair the performance of hierarchically structured teams (i.e., team members are specialized, or hold information and capabilities that are unique to some extent), whereas structuring teams non-hierarchically (i.e., team members are non-specialized, or share common information and capabilities) tends to buffer against this performance decrement.
Experimentally manipulated Urban et al. (1996) 36 five-person teams of students (U.S.) In contrast to resource demands, time pressure is shown to impair team performance.
Experimentally manipulated Worchel and Shackelford (1991) 41 teams with 263 psychology students (U.S.) The structuring of teams (i.e., appointing a leader and identifying discussion rules) tends to impair (vs. facilitate) team members' perceptions of team functioning, interpersonal attraction, and their desire to remain in the team when their physical environment in terms of noisiness and crowdedness is negative (vs. positive).

Experimentally manipulated … conceptualized as team-level moderators:
Bang and Park (2015) 153 teams with 5,579 team members of a semiconductor company (Korea) The effect of task conflict on team performance is shown to be positive (vs. negative) when demanding task characteristics in teams are high (vs. low).

Direct consensus model
Cannon-Bowers et al. under high (vs. low) promotion focus or low (vs. high) prevention focus (Sacramento et al., 2013). In both of these cited studies, the authors replicated their models at the team level and found that team self-efficacy decreases team job demands, which in turn have opposite effects on team creativity at different levels of promotion focus (i.e., positive when promotion focus is high and negative when it is low). The positive effect of team reflexivity on team innovation is shown to be stronger either for high (than for low) levels of workload or for low (than for high) quality work environments. Although workload seem to hamper the positive effect that a team adaptation and coordination training has on team performance, teams receiving such a training still perform better under high levels of workload than non-trained teams under low levels of workload.

Key-informant design
Experimentally manipulated Volpe et al. (1996) 40 two-person teams of psychology students (U.S.) Whereas cross-training of teams is shown to be beneficial for their effective teamwork, communication, and performance, their workload-although degrading teamwork and communicationappears not to interact with cross-training.

S.)
Under high (vs. low) levels of time pressure, highly cohesive teams are shown to outperform (vs. perform equally well compared to) less cohesive teams.

Stressors due to roles in the organization … … conceptualized as team-level predictors:
Hu and Liden (2011)  For role ambiguity and role conflict, team job satisfaction is shown to be key for transmitting their linear and non-linear effects on new product performance as measured by market success, adherence to budget and schedule, and product quality.

Key-informant design
Stressors due to organizational structure and climate … … conceptualized as team-level predictors: Akgün et al. (2007) 96 R&D teams with 192 marketing and engineering managers (U.S.) Team anxiety seems to foster desirable outcomes like idea generation and the development and commercialization of a product only under high levels of management support.

S.)
Conformity pressure may help increasing team creativity only in teams that lack creative talent. Headcount understaffing tends to trigger more initiating structure behaviors of team leaders, whereas expertise understaffing result in more consideration behaviors.

Referent-shift consensus model
Stressors at the work-home interface … … conceptualized as team-level predictors: ten Brummelhuis et al.
61 teams with 520 team members from various industries (the Netherlands) Family demands seem to impair team performance through reduced taskwork and this negative effect is shown to be attenuated by support from both supervisors and organizations. However, family demands may even foster teamwork when coworker and supervisor support are high.

Additive model
Similarly, team members' job demands were revealed to positively predict both their emotional exhaustion (Bakker, Van Emmerik, & Van Riet, 2008) and burnout (Consiglio et al., 2013). Both studies replicated these findings for team job demands and their association with team emotional exhaustion and team burnout. Finally, Griffith (1997) showed that work overload in military teams decreases their cohesion (conceptualized as a team's task and emotional peer support), which in turn results in perceptions of inferior performance at both the individual and team levels. This overall negative performance effect of work overload is buffered at both levels by low disintegration (i.e., the dissolution of a team) and only at the individual level by strong team task support and well-being (i.e., less strain).

| Stressors intrinsic to a job
Although team-internal processes may induce stressors that affect teams and their members, such as overload and time pressure, extant research has predominantly focused on team-environmental influences leading to the emergence of stressors intrinsic to team members' jobs. Research has revealed that the leaders of student teamsrather than their members-experience heavier workloads due to their elevated coordination demands (Guastello, Correro, & Marra, 2018), and surprisingly, reflexivity interventions for organizational teams have also been found to be ineffective in reducing individuals' overload (Chen, Bamberger, Song, & Vashdi, 2018). Moreover, time pressure has been found to predominantly arise in the wake of external or environmental factors rather than internal disturbances, and therefore, teams encountering crisis events due to time pressure and threats most effectively cope with these factors through teamexternal activities (Choi, Sung, & Kim, 2010). Interestingly, experimental research has suggested that members of student teams are more likely to attribute charismatic leadership qualities to their team leaders during stressful crises (Halverson, Murphy, & Riggio, 2004).
At this point, it is worth mentioning that research is conclusive in that team-level stressors-although relevant for team members' functioning-show weaker associations with outcomes pertaining to individual team members than do respective individual-level stressors.
This might be one reason why cross-level research on team stressors has been slow to accumulate. Nonetheless, it has been found that quantitative team overload negatively affects both team-and individual-level performance via team learning behaviors (Savelsbergh et al., 2012) and that stressed teams generally perform worse because their team members typically experience a loss of the team perspective, which is required to perform well (Driskell, Salas, & Johnston, 1999). In contrast, team members' experiences of work-to-family conflict are primarily attributable to their own quantitative and emotional job demands rather than to those of their team (van Emmerik & Peeters, 2009). A similar picture emerges for team members' organizational commitment, which is predominantly hampered by their individual (rather than team) work pace (Clausen & Borg, 2010), and for their psychosomatic symptoms, which are essentially caused by experiences of individual time pressure and physical demands rather than by team understaffing (Busch, Deci, & Laackmann, 2013). The latter findings resonate with more recent research showing that team members' individual job demands trigger psychosomatic symptoms and presenteeism and decrease their subjective general health, mental health, and work ability (Schulz, Zacher, & Lippke, 2017). Evidence that effects are less likely to be identified across levels also appears when potential buffering resources are conceptualized at different levels. Whereas team members' individual perceptions of support appear to somewhat compensate for the detrimental effects that individual-level time pressure has on their well-being and identification, collective perceptions of cohesion do not buffer these lower level effects (Griffith, 2002 Bliese and Britt (2001) show that the social environment of military team members buffers the detrimental impact of work stressors on their morale and depression, research in nursing also shows that team members can uphold their motivation despite team job demands when the contextual environment provides them with a great deal of autonomy (de Jonge et al., 1999).
In contrast to these functional buffering mechanisms, some resources interplay to aggravate the demands intrinsic to team members' jobs. Although a proactive personality tends to buffer the positive effect of emotional job demands on team members' intention to quit, high team potency inverts this effect by aggravating highly proactive team members' intention to quit (Loi, Liu, Lam, & Xu, 2016).
Likewise, team cohesiveness and social support may equally trigger dysfunctional buffering mechanisms, as they have been shown to increase the likelihood of job demands and exhaustion crossing over within highly cohesive and supportive teams (Westman et al., 2011).
The latter effect appears to be particularly critical either when the team performs badly or when team members lack psychological capital. With respect to social support, one might assume that team members who must carry a relatively high share of their team's workload should receive more support. However, it has been shown that after receiving negative feedback, members of student teams with relatively high workloads benefit from significantly less helping behavior from their teammates when they are racially dissimilar than when they are racially similar (Triana et al., 2013). As the number of cross-level studies on team stressors increases, research may come closer to verifying whether the burden is shared and the grief divided in teams under stress or such situations may reveal the shady sides of interpersonal collaboration (as this latter evidence suggests).

| Stressors due to relationships at work
Although this stressor has been insufficiently addressed in the existing literature, studies have begun to examine how the team itself can function as a stressor for its members. With respect to intrateam conflict, Hon and Chan (2013) show team task conflict to be positively related to team members' job satisfaction and performance as it triggers challenge-related stress; additionally, they show team relationship conflict to be negatively associated with such outcomes in individuals as this form of conflict triggers feelings of hindrance-related stress.
Relationship conflicts in teams may equally spill over into the nonwork domain (i.e., causing strain-based work-life conflict), which is attenuated by team members' optimism and resilience (Martinez-Corts, Demerouti, Bakker, & Boz, 2015). In line with these findings, it has been shown that relationship conflict in teams taxes team members' health (i.e., burnout) and performance (Leon-Perez, Antino, & Leon-Rubio, 2016), with the latter association being buffered by a conflict management climate. Although still underrepresented, these findings serve as the initial evidence of the potential contagion effects of stressors. This evidence resonates with research showing that team burnout tends to cross over to individual team members (Bakker, van Emmerik, & Euwema, 2006) and that team leaders' distress is at risk of crossing over into team members' lives because distressed leaders tend to demonstrate more abusive behaviors (Li, Wang, Yang, & Liu, 2016).

| Stressors due to organizational structure and climate
Cross-level research on stressors due to organizational structure and climate has exclusively considered a team's climate. Haber (2016) shows that team membership becomes stressful only in climates characterized by conflict and distrust. Likewise, a climate of empowerment within teams is beneficial, as it further attenuates the already negative effect that distributive justice has on team members' stress and ultimately reduces their turnover intentions (Choi, Moon, Nae, & Ko, 2013). This might partly be the case because individuals working in distressed team climates seem to be significantly more exhausted than those in eustressed team climates, and they tend to show significantly less vigor and dedication at work than members of teams with a more balanced team climate (Kozusznik et al., 2015). Similarly, in a highly ambiguous team climate, team members typically engage in less extra-role activity, as their affective commitment is impaired (Mañas et al., 2018). Table 3 lists the reviewed multilevel articles applying either homologous multilevel or cross-level models.

| FUTURE RESEARCH AGENDA
In summary, Figure 3 offers a multilevel framework of the extant literature following the methodological logic of our review above. The framework suggests that individual-level stressors influence team performance-relevant outcomes in team members via two mediating pathways that are either strain based or behavior based. For the effects of team-level stressors on actual team performance, such mediating pathways are either identity based or information based.
Identity-based processes at the team level, in turn, influence the lower level effects of stressors either directly by moderating lower level effects of stressors or indirectly by triggering individual-level moderators such as team members' feelings of being part of and their perceptions of in-and out-groups. This framework is also helpful because it identifies four important research gaps (accentuated with circled grey numbers) left by the extant research in this field. We discuss below how research can further advance the field by investigating (1) stressors related to career development, (2) dysfunctional antecedents of stressors grounded in the formation of subgroups and the contagion of stress among team members, (3) the conceptual relationship of stressors across levels, and (4) phenomena of related research streams more consistently from a team-stressor lens.

| Advancing research through new types of stressors in teams
As reviewed above, in the context of teams, stressor research has predominantly addressed aspects intrinsic to the job or originating from workplace relationships. However, in contrast to the original stressor research that exclusively considered individuals (Cooper et al., 2001), no study in the context of teams has thus far looked at the stressors related to career development. This is unfortunate given that there are important stressors within this category potentially affecting teams and their members.
Conceptualized at the individual level, such stressors can include individual team members experiencing job insecurity (e.g., facing the prospect or threat of being removed from the team) or career stagnation (e.g., seeing a protégé being advantaged or treated with favor by the team leader). In terms of insecurity, a growing body of literature is becoming invested in temporary team membership given that membership dynamics have become pervasive in organizational practice (e.g., Hirst, 2009;Mathieu, Tannenbaum, Donsbach, & Alliger, 2014).
As the temporary nature of fluid teams makes it difficult for their members to familiarize themselves with each other (Edmondson & Nembhard, 2009), such teams are often challenged in fostering teamwork (Bushe & Chu, 2011). Therefore, knowledge about team members' behaviors toward this type of team and the strain caused by this demanding setting is needed to allow organizations to more successfully benefit from temporary team memberships in the future. In terms of career stagnation, Dasborough et al. (2009, p. 574) suggest that when team members "perceive unwarranted favoritism by their leader towards some members, individuals not so treated are likely to experience negative emotions." Whereas the literature on individuals has shown that threats to their career advancement or promotion drive their dissatisfaction and strain (Cooper et al., 2001), insights on how actual favoritism or team members' perceptions thereof affects teamwork behaviors or information sharing advance not only teamstressor research but also the scarce but growing body of empirical On both the team and individual levels, self-efficacy appears to decrease job demands like monotony and time pressure, which, in turn, increase burnout in teams.

Additive model
… conceptualized as individual-and team-level predictors: Bakker et al. (2008) 71 teams with 176 team members from a temporary employment agency (the Netherlands) On both the team and individual levels, job demands like work pressure tend to result in emotional exhaustion, which, however, does not translate into performance on both levels.
Additive model Griffith (1997) 112 companies with 9,013 U.S. Army soldiers (U.S.) While work overload is shown to increase individual performance, it does not directly affect team performance. On the team level, such a job stressor appears to predominantly take a toll on team performance via impairing the task cohesion of teams.
Additive model Sacramento et al. (2013) 41 R&D teams with 123 team members and 36 team leaders (Portugal) On both the team and individual levels, the effect of job demands on creativity tend to be positive (vs. negative) for high (vs. low) levels of promotion focus. Moreover, on the individual level, this effect seems to be positive (vs. negative) when prevention focus is low (vs. high).

Cross-levelinvestigations
Stressors intrinsic to a job … … conceptualized as individual-level criteria: Guastello et al. (2018) 44 teams with 348 psychology students (U.S.) Team leaders are shown to experience greater workload and coordination demands than non-leaders.
-… conceptualized as team-level criteria: Choi et al. (2010) 30 teams with 108 team members from various industries (Korea) The effectiveness of managing team crisis (e.g., time pressure and threat) appears to depend strongly on the external activities of teams.
Referent-shift consensus model … conceptualized as individual-level mediators: Chen et al. (2018) 73 teams with 469 team members of an electronics company (China) Neither are changes in qualitative overload shown to predict any of the three burnout dimensions (i.e., inefficacy, emotional exhaustion, and cynicism), nor a team reflexivity intervention to affect changes in team members' qualitative role overload.
-… conceptualized as individual-and team-level predictors: Bakker et al. (2006) 85 teams with 2,229 police officers (the Netherlands) In particular, the emotional and expectation demands of team members tend to trigger their individual burnout (i.e., emotional exhaustion, cynicism, and reduced professional efficacy). Moreover, burnout in teams appears to cross over from the team to the individual team members.
Direct consensus model Busch et al. (2013) 33 teams with 265 team members from various industries (Germany) Whereas team understaffing is shown to lead to dysfunctional coping on the team level, it does not explain variance in psychosomatic symptoms over and above individual-level job stressors like time pressure and physical demands.

Direct consensus model
Clausen and Borg (2010) 301 eldercare teams with 6,299 team members (Denmark) Team members' affective organizational commitment seems to be decreased by a team's work pace and increased by its emotional demands. These effects seem to hold only as long as individual job demands are not considered, Additive model (Continues) Only individual-level qualitative and quantitative demands tend to trigger emotional exhaustion and job-related anxiety. In contrast, work motivation is shown to be decreased by such demands on both levels, but only the lower-level effect to be efficiently buffered by autonomy.
Additive model Savelsbergh et al. (2012)  While team members' overload appears to be detrimental to their subjective health, this relationship is shown to be less negative in highly cohesive teams. Work stressors like a lack of privacy tend to decrease morale and this effect is shown to be less negative when team consensus is strong. When team potency is low, emotional job demands appear to enhance (vs. decrease) team members' intention to quit if their proactive personality is low (vs. high). When team potency is high, both little and highly proactive team members tend to show an increase in intention to quit as emotional demands increase.
- Schulz et al. (2017) 621 teams with 6,449 team members working in a health insurance organization (Germany) While team members' individual job demands tend to trigger psychosomatic complaints and presenteeism, they are shown to decrease their subjective general health, mental health, and work ability. After having received negative feedback, such recipients who had a disproportionately heavy share of their team's workload are shown to experience significantly less helping behaviors from their teammates when they are racially distant rather than racially similar. - research on the role of unwarranted favoritism in the quality of leader-member relationships in teams (e.g., Hsiung & Bolino, 2018;Omilion-Hodges & Baker, 2013). It should be mentioned that such stressors (i.e., insecurity and career stagnation) could equally affect the entire team. As such, the security of teams is threatened when they face the risk of their projects being terminated (e.g., Shepherd, Patzelt, Williams, & Warnecke, 2014), whereas some teams can stagnate when they receive fewer resources than comparable teams within the same organization (e.g., Hoegl et al., 2008). Both these realities of team-based work have implications in terms of how teams and their members function and should therefore be featured in stressor research.

| The role of team diversity and faultlines
In addition to the lack of research on career development stressors in teams, our integrative framework presented in Figure 3 uncovers  Driskell et al. (1999) 32 three-person teams of U.S. Navy students (U.S.) In interdependent teams, stressors like task load and time pressure seem to decrease team performance by narrowing team members' team perspective.
Experimentally manipulated Halverson et al. (2004) 55 three-person teams of arts students (U.S.) When team members are stressed (vs. not stressed), they tend to perceive their leaders to be more (vs. less) charismatic.

Experimentally manipulated
Stressors due to relationships at work … … conceptualized as team-level predictors: Hon and Chan (2013) 50 teams with 265 team members working in the hotel industry (China) Whereas challenge stressors tend to mediate the positive relationship of team task conflict with both job performance and satisfaction, hindrance stressors are shown to mediate the negative relationship of team relationship conflict with both job performance and satisfaction.
- -Perez et al. (2016) 55 teams with 798 team members from a vehicle safety and emission inspection firm (Spain)

Leon
Team relationship conflict seems to increase team members' burnout. Moreover, while team process conflict is shown to decrease the quality of service, this negative relationship is shown to be buffered by a strong conflict management climate.
Referent-shift consensus model Li et al. (2016) 86 teams with 351 team members and 86 leaders from various industries (China) Team leaders' psychological distress appears to increase distress in their team members via an increase in abusive supervision. This cross-over effect is shown to be even strengthened when either team performance or team members' psychological capital is low.

Key-informant design
Stressors due to organizational structure and climate … … conceptualized as individual-level predictors: Choi et al. (2013) 90 teams with 4,432 team members (South Korea) Team members appear to be less stressed and thus to show less turnover intentions with increasing levels of distributive justice. This effect is shown to be even more negative in strong empowerment climates. -… conceptualized as team-level predictors: Kozusznik et al. (2015) 78 teams with 535 team members working in social service firms (Spain) In distressed team climates, team members seem to be significantly more exhausted compared to those in eustressed team climates. Team members in balanced (i.e., neither distressed nor eustressed) team climates appear to show the strongest engagement as measured by vigor, dedication, and absorption.

| From bad apples to bad barrels?
Beyond looking at the (diverse) composition of teams as a source of stressors, our review should also encourage scholars to further study how teams themselves become stressors for their individual members.
This resonates with initial attempts in team-conflict literature, for example, to understand how conflict can become contagious (Jehn, Rispens, Jonsen, & Greer, 2013). In terms of emotional contagion theory (e.g., Barsade, 2002;Sy, Côté, & Saavedra, 2005), it is relevant not only to investigate how team members collectively make sense of stressors affecting their team but also, and perhaps more importantly, to understand how other team members are affected by an individual member experiencing stronger stressors than his or her colleagues.
This might be particularly relevant for teams operating in extreme environments in which the consequences of failure are more dramatic (Driskell, Salas, & Driskell, 2018).
In regard to time pressure, for example, it is well known that the degree to which individuals perceive this stressor (partly) depends on their personality (e.g., neurotic individuals tend to overestimate time pressure; Freedman & Edwards, 1988). Thus, if a team consists of members with dissimilar personalities, it is plausible that the experience of this specific stressor may differ among its members, as well.
If such perceptions diverge, then longitudinal research on the adaptive processes within teams is important to improve our understanding of how stressors in teams unfold. On the one hand, the other team members may help this outlier better cope with the situation, thus decreasing his or her experience of this specific stressor. Such an adaptation (i.e., the convergence of perceptions of stressors) will likely be functional in that it ultimately helps lower the amount of time pressure experienced within the entire team, which is recommended in teams failing at establishing the contingencies that may allow time pressure to benefit teams (such as temporal leadership, as reviewed above). On the other hand, there is also the risk of a more dysfunctional adaptation occurring if the perceptions of a team member who feels highly time pressured were to infuse the perceptions of his or her fellow team members. Such a dysfunctional adaptive process may metaphorically transform bad apples into bad barrels, thus endangering the proper functioning and performance of the entire team. Consequently, further research that connects to the scant knowledge on the cross-over effects of stressors within teams (e.g., Bakker et al., 2006;Li et al., 2016;Westman et al., 2011) is not only needed but also promising. This is particularly true because membership in multiple teams is becoming increasingly relevant in organizational practice (Mortensen & Gardner, 2017), which may lead to a contagion effect of stressors even between different teams that share the same "bad apple" (i.e., a team member who feels an exaggerated amount of pressure).

| Conceptualizing and measuring stressors in teams
More generally, our literature review emphasizes that collecting crosslevel rather than single-level data within teams is worthwhile (Klein & Kozlowski, 2000). Although the same stressors potentially affect team members when they work in teams, their perceptions and responses sometimes differ, leading to a non-independence bias that contributes to the non-equivalence of the results at the individual and team levels (Bliese, 1998;. Thus, collecting only single-level data from individuals within teams makes it more difficult to detect actual team-level effects . Cross-level data collection efforts are therefore promising because they are theoretically and empirically relevant for the multilevel phenomena of specific stressors Farh & Chen, 2014;Korsgaard, Soyoung Jeong, Mahony, & Pitariu, 2008). Tables 2 and 3 research on team stressors more frequently uses either direct consensus or referent-shift consensus approaches to construct team stressors from individual-level data. In our review, 11 of the studies applied a direct consensus approach (e.g., Hudson & Shen, 2018;van Emmerik & Peeters, 2009). When collecting data using this approach, the operationalization of the stressor is referred to team members themselves and is then aggregated to the team level, typically by using the construct's arithmetic mean across all team members once sufficient agreement within the team is reached (i.e., values for r WG exceed. 70; Klein & Kozlowski, 2000). Finally, 13 articles conceptualized team stressors by using the referent-shift consensus model (e.g., Maruping et al., 2015;Sacramento et al., 2013), which differs from the direct consensus approach only by referring the stressor items to the entire team. Thus, although the operationalization of team stressors during the data collection stage differs between these approaches, the data analyses are identical.

As listed in
Consequently, perhaps the most critical (albeit promising) challenge for future research on stressors in teams to overcome is the question of how to operationalize team stressors for quantitative field studies, which requires understanding the convergence among individuals that allows such higher level phenomena to emerge (Fulmer & Ostroff, 2016). This is important because when studying team-level stressors in the organizational context, applying a key-informant design (which seven of the reviewed field studies did, e.g., Drach- Once such data are aggregated, they can be used not only to examine how the respective team-level constructs differ in the degree of agreement (i.e., the consensus) and the effect size but also how the direct consensus measurement, which is in essence an aggregated operationalization of an individual-level stressor, works differently (or similarly) at the lower level. van Mierlo et al. (2009) offer a framework on the distinction between and the baseline psychometric qualities of composed group constructs that might be informative in this respect. Empirically complementing and qualifying previous cross-level studies on team stressors using their approach is therefore a promising avenue of future research. Specifically, investigating whether the differential team-performance effects of challenge-and hindrancerelated team stressors also hold true for outcomes at the lower level of analysis by using data from both composition models may contribute significantly to the developing body of cross-level studies on team stressors (e.g., Consiglio et al., 2013;Kozusznik et al., 2015;Savelsbergh et al., 2012).

| Reconsidering phenomena of related literature streams
Beyond informing extant research on stressors within teams, the reviewed findings should demonstrate that other streams of team research have already been devoted to stressor-like phenomena without consciously making use of what is known from the research presented above. The research on destructive leadership behaviors, such as abusive supervision (Duffy, Ganster, & Pagon, 2002) and supervisor undermining (Tepper, 2007), proves to be a worthwhile exemplar to consider in this regard. These behaviors are associated with hostile climates that are "characterized by consistent acrimonious, antagonistic, and suspicious feelings among coworkers" (Mawritz et al., 2014, p. 737). Along with traditional stressor frameworks, abusive supervision is unequivocally a hindrance-related stressor for employees who are confronted with supervisors applying such destructive leadership behaviors (e.g., Liu, Liao, & Loi, 2012;Nandkeolyar, Shaffer, Li, Ekkirala, & Bagger, 2014;Restubog, Scott, & Zagenczyk, 2011). Depending on the victim(s) targeted and the structural context, abusive supervision can occur at both the individual and team levels (Farh & Chen, 2014 (Chi & Liang, 2013), conscientiousness (Nandkeolyar et al., 2014), and psychological capital (Li et al., 2016) ameliorate such detrimental effects.
At the team level, team member support (Hobman, Restubog, Bordia, & Tang, 2009) and organizational support (Kim, Kim, & Yun, 2015) have been shown to attenuate the undesirable effects of abusive supervision. Therefore, expanding the concept of stressors to the team level is informative for studies on destructive leadership (such as abusive supervision) once such leadership behaviors are understood and theorized as stressors. Although some studies have begun to examine abusive supervision in light of traditional stressor theories (e.g., Chi & Liang, 2013;Hobman et al., 2009), further integrating such theories is helpful for the theoretical development in both areas.
To conclude, our multilevel perspective allows organizational behavior scholars to more easily embed their research into a compelling framework, thereby avoiding fragmented theorizing (Sparrowe & Mayer, 2011). Applying this integrative approach to the theories of other domains may generate promising insights. This is particularly important because such related theories sometimes unknowingly share obscure links to concepts established in different streams of research (Mayer & Sparrowe, 2013), such as those on stressors and demands. Bridging the gaps and embracing the similarities between seemingly disparate fields of research are hence helpful for allowing scholars to draw upon a more elaborate rationale regarding how team stressors could shape the behaviors and performance of teams and their members beyond what is already known. Our review is a first step toward creating such a more comprehensive understanding of the functioning of teams and their members despite (or because of) certain stressors.