Team Psychological Safety and Conflict Trajectories’ Effect on Individual’s Team Identification and Satisfaction

Abstract

Using a multilevel field study with data collected over a 9-month period, we tested how team psychological safety interacts with levels of team relationship conflict to influence an individual’s team identification and satisfaction with their team. We propose that team psychological safety measured early in a team’s time together influences what team members can expect to experience in subsequent team interactions. We use identification and team conflict theory to theorize that through sense-making processes, team members evaluate early experiences with their team relative to initial levels of team psychological safety, which then influences their levels of team identification. When team members experience high levels of team psychological safety initially followed by an increasing trajectory of relationship conflict within the team over time, we predicted and found that individual’s team identification decreased, resulting in lower satisfaction with their team. The theoretical and practical implications for aligning early perceptions of team’s psychological safety with patterns of perceived relationship conflict and its effect on team identification and satisfaction with the team are discussed.

Keywords

identification social identity theory team psychological safety trajectory team conflict satisfaction with team

Despite acknowledgment that identification is shaped through multiple interactions between individuals and groups they belong to (Ashforth, 1998; Ashforth, Harrison, & Corley, 2008; Pratt, 1998), there has been relatively little attention focused on how identification with one’s team is formed. We know that individuals engage in a sense-making process whereby they enact an identity and reflect on what occurs as a result of that identity; these reflections then affect subsequent levels of identification (Ashforth et al., 2008). Here we consider how initial perceptions of the teams’ psychological safety affect an individual’s subsequent levels of team identification and satisfaction. We reasoned that initial perceptions about the team provide cues to team members about whether their needs will be fulfilled and satisfied by their team. Team psychological safety in particular is a key team factor to consider because it represents collective perceptions about how comfortable team members expect to feel in interacting with team members and sharing their perspectives (Edmonson, 1999). Then, individual team members accrue experiences with the team that either confirm or disconfirm those initial expectations, affecting subsequent levels of identification and satisfaction with the team. We investigate team relationship conflict because it characterizes an important aspect of the interpersonal interactions that take place within a team (Jehn, 1995).

Antecedents to team identification include outgroup salience, low intragroup competition, distinctiveness, prestige, and positive interpersonal factors within a team (Ashforth & Mael, 1989). Yet, prior research has not examined how the fulfillment of expectations for a team’s level of psychological safety affects subsequent levels of team identification. Prior research has shown that the prestige of a group was positively related to team identification because being members of a prestigious group enhances self-esteem (Dukerich, Golden, & Shortell, 2002; Dutton, Dukerich, & Harquail, 1994; Pratt, 1998; Smidts, Pruyn, & van Riel, 2001). Interpersonal factors that contribute to identification with a team such as diversity, goal congruence, and goal/task interdependence affect whether members identify with their team based on how similar members perceive themselves to be to each other (Kearney & Gebert, 2009; Kearney, Gebert, & Voelpel, 2009; Lee, Farh, & Chen, 2011; Polzer, Milton, & Swann, 2002; Van der Vegt, Van de Vliert, & Oosterhof, 2003; Van Dick, Van Knippenberg, Hägele, Guillaume, & Brodbeck, 2008). We also know that higher levels of team identification positively predict higher levels of team learning and performance (Van der Vegt & Bunderson, 2005) and weaken the adverse effects of workgroup faultlines on performance, coalition formation, and conflict (Bezrukova, Jehn, Zanutto, & Thatcher, 2009; Jehn & Bezrukova, 2010). However, these antecedents have been evaluated as static antecedents to team identification rather than as team interactions that accrue over time. Individuals engage in sense-making throughout a team’s time together (Ashforth et al., 2008) resulting in team identification. A key contribution of the current study involves our examination of how team members interpret the level of relationship conflict in their group that emerges over time relative to the team’s early expectations for team psychological safety.

Our study contributes to existing literature in three ways. First, our study expands our understanding of antecedents to identification (Ashforth, 1998; Ashforth et al., 2008; Ashforth & Mael, 1989; Pratt, 1998) to include team factors such as team psychological safety (Edmonson, 1999). Perceptions about the psychological safety of a team can also influence the likelihood of an individual’s team identification beyond the interpersonal factors discussed earlier, which are focused on the makeup of the team or specific task-related factors.

Second, we respond to calls to examine how team identification forms (Ashforth, 1998; Ashforth et al., 2008; Pratt, 1998). We do so by assessing the interplay between initial perceptions of team psychological safety and the trajectory of team relationship conflict (increasing, decreasing, or consistent) over time on subsequent levels of individual’s team identification. Relationship conflict in particular is relevant to our theorizing because initial perceptions of team psychological safety provide team members with an indication of the type of interpersonal interactions they should expect among team members including potential levels of conflict in those interactions.

Third, we evaluate the level of consistency between initial perceptions of team psychological safety and the subsequent relationship conflict trajectory, and its effect on individual’s team identification and satisfaction with the team. It is not only initial perceptions about the team or the trajectory of relationship conflict interactions among team members over time that affect team identification and satisfaction, but the extent to which these two factors are consistent. See Figure 1 for our theoretical model.

Figure 1.

Structural equation modeling results.^a

Team Identification

Team identification refers to a sense of oneness with a group or a perceived overlap in the attributes of a group and themselves (Ashforth, 2001; Dutton et al., 1994). With higher levels of team identification, members are more likely to behave in a manner that aligns with their team’s efforts (Ashforth, 2001), such as being more cooperative in support of the team’s values, norms, and interests (Ashforth & Mael, 1989; Dukerich et al., 2002; Tajfel & Turner, 1986). Supporting this position, physicians’ level of identification with their organization predicted more cooperative behaviors with other physicians and toward their overall health care organization (Dukerich et al., 2002). Higher levels of team identification have also been shown to attenuate the negative effects of diversity (e.g., demographic, functional) on team outcomes (Bezrukova et al., 2009; Jehn & Bezrukova, 2010) and reduce social undermining (Duffy, Scott, Shaw, Tepper, & Aquino, 2012).

Identification with a particular team can be understood by first examining why individuals identify with those teams and then how this identification evolves over time (Ashforth et al., 2008). According to social identity theory, one of the core reasons why individuals identify with a group is to fulfill needs for uncertainty reduction and self-enhancement (Hogg & Terry, 2000). Individuals seek identities (oftentimes linked to desired groups) that provide them with a sense of their place in a group, organization, or community (reducing uncertainty), while enhancing their levels of self-esteem. Individuals are also motivated to identify with groups because they contribute to their feelings of competence (self-efficacy; Ashforth, 2001; Brickson, 2013).

How identification is strengthened or diminished is related to the sense-breaking, sense-giving, and sense-making processes that occur between individuals and their team membership (Ashforth et al., 2008; Weick, 1995). First, through sense-breaking, teams can challenge individuals’ sense of self such that individuals perceive a void that needs to be filled. A team and its members can then guide individuals through sense-giving processes, which provide informational cues about what it means to identify with their team. Individuals also undergo sense-making processes as they observe and reflect upon how the environment and/or others in their team respond to the way they think, feel, and/or behave with respect to the team’s identity (Ashforth et al., 2008). These observations also provide individuals with information about whether identifying with their team will help fulfill their psychological motives (Brickson, 2013). The more consistent (inconsistent) the interactions and sense-making are with what individuals expect from a team, the more (less) identified individuals become with their team. (Ashforth, 2001).

Theoretical Model and Hypotheses

We use identification and conflict theory to examine how initial levels of team psychological safety interact with the trajectory of team relationship conflict to affect individual team members’ team identification and satisfaction. We propose that higher perceived psychological safety at the beginning of team formation is expected to reduce team members’ uncertainty about working with their team, allowing team members to feel comfortable sharing information with one another. If team members judge the team as safe, they will feel less uncertainty about how team members will respond to their opinions, behaviors, risk-taking, and mistakes, resulting in higher levels of team identification and in turn satisfaction.

We also know that as teams engage in accomplishing their work, interactions with team members frequently result in conflict. This may occur because members seek to maintain their individuality or do not have the means to effectively address escalations in conflict (Tuckman, 1965). The specific types of conflict that team members experience are related to relationships among team members, tasks the team is undertaking, and processes the teams use to achieve their goals (Jehn, 1995).

In Tuckman’s (1965) classic work on groups, he described how they move beyond conflicts that derail the group’s development to a group where members accept each other and turn their attention to maintaining group harmony. Subsequent team research shows that overcoming conflict was important to the development and effectiveness of teams (Tekleab, Quigley, & Tesluk, 2009). We argue that conflict trajectories cue individual members about how their group engages to complete tasks after formation, which will then be compared to how the team expected the group to engage based on initial perceptions of team psychological safety. To the extent that early expectations are inconsistent with the team’s actual experiences with team relationship conflict, lower levels of team identification should result over time. Specifically, we expect lower levels of identification to occur when a team has high initial expectations for the team’s level of psychological safety, but then experience an increasing trajectory of relationship conflict. We also expect that more inconsistency (e.g., higher positive initial expectations and greater unexpected conflict) will result in lower satisfaction with one’s team due to lower team identification.

Team Conflict Trajectories

As teams interact, its members make sense of their experiences including the type and degree of conflict they experience. Through this sense-making process, members try to understand how the conflict occurred, as well as the incompatibilities members feel they have with each other in terms of expectations, tasks, goals, processes, and so on (Jehn, 1995). Authors taking an information-processing view of conflict argue that there are initial benefits of conflict, but that these benefits will eventually diminish resulting in hampering team performance (De Dreu & Weingart, 2003). Therefore, team conflict should be minimized.

Team conflict represents the presence of disagreement among team members. Relationship conflict refers to disagreements that arise based on interpersonal differences among team members, such as differences related to member experience, personality, and values (Jehn, 1995). We examined how the trajectory of team relationship conflict measured over three periods of time (Chen, 2005; Chen, Ployhart, Thomas, Anderson, & Bliese, 2011) relates to individual’s team identification and satisfaction. The pattern of team relationship conflict experienced is a signal to team members about how their team is progressing in terms of its formation. Where there is a rising level of relationship conflict, we would expect team members to be less identified with their team because interpersonal factors such as their relationships with each other can affect levels of team identification (Ashforth & Mael, 1989). Team members who are different from each other are also less likely to be identified with their team unless there are factors that help to mitigate these differences, such as leadership (Kearney & Gebert, 2009) or goal and task interdependence (Van der Vegt et al., 2003). When a team experiences an increasing trajectory of relationship conflict, signaling persistent incompatibilities and differences among team members, we would expect those experiences to result in lower levels of an individual’s team identification. Here, we were particularly interested in the relationship conflict trajectory because as we explain in the next section, relationship conflict is most likely to interact with perceived psychological safety among team members. Therefore, we propose the following:

Hypothesis 1: The team relationship conflict trajectory that occurs over time will negatively predict individual’s team identification.

Initial Team Psychological Safety

Psychological safety describes the “shared belief that the team is safe for interpersonal risk taking” (Edmonson, 1999, p. 354). Perceptions of team psychological safety derive from members feeling trust in and respect for each other, resulting in a feeling that members can speak up without risk of punishment or embarrassment. Team psychological safety is distinct from team relationship conflict because team psychological safety refers to team members’ perceptions about the team environment or climate, whereas team relationship conflict represents the observations of disagreements among team members. Higher levels of team psychological safety have been associated with team learning because members feel safe sharing different perspectives, seeking feedback, and discussing mistakes (Edmonson, 1999). Team psychological safety has also been found to encourage individual voice, again resulting in a sharing of each member’s opinions (Detert & Burris, 2007). However, in these two studies, psychological safety was measured after the teams had formed, and was therefore not examined as a perception at the beginning of the teams’ formation.

In our study, we specifically investigate initial team psychological safety, which refers to team members’ perceptions of team psychological safety at the beginning of a team’s time together before team members have spent much time together and begun work on projects. During this period of time, team members develop perceptions about whether the team will have an environment/climate that allows for mistakes and risk-taking based on impressions they develop early in the team’s time together. This measure of initial team psychological safety represents the expectations team members have for how their team will work together.

One of the motivations for developing identification with a group is based on self-categorization theory (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987) and relates to uncertainty reduction (Hogg, 2000; Hogg & Terry, 2000). Individuals are motivated to categorize themselves as members of groups because this reduces feelings of uncertainty about their sense of self. Within a team, uncertainty can result from disagreements about beliefs, attitudes, or behaviors (Hogg, 2000). We suggest that when individuals feel a high level of team psychological safety, they will have more confidence in how other team members will respond to their questions, ideas, and disagreements, thus reducing feelings of uncertainty and risk associated with their team (Hogg, 2007). With higher levels of team psychological safety, when disagreements and conflict occur, team members should be more confident they will not be punished for sharing contrarian opinions, which should result in lower levels of uncertainty, and higher team identification (Hogg, 2000; Hogg & Terry, 2000). Team identification would result from uncertainty reduction because taking on a team’s membership provides clarity about how one should think, feel, and behave (Hogg & Terry, 2000). Therefore, we propose the following:

Hypothesis 2: Initial team psychological safety will positively predict individual’s team identification.

Initial team psychological safety provides sense-giving cues to team members that suggest whether their psychological motives will be met by identifying with their team. However, team identification forms as a result of team members enacting their roles and observing interactions within their team (Ashforth et al., 2008) with some of those interactions being characterized by conflict. It is important here to reiterate that perceptions of initial team psychological safety represent team member perceptions early in a team’s time together or during formation, and are developed based on relatively few interactions with team members. Therefore, these perceptions do not represent the actual environment or climate within the team but instead, expectations for how team members think the team will interact in their upcoming time working together.

Increasing levels of team relationship conflict are likely to contradict team members’ expectations for a psychologically safe environment in which they are comfortable offering ideas and opinions. We expect this to be the case as a psychologically safe environment derives primarily from interpersonal relationships and is tied to the trust and respect felt among team members (Edmonson, 1999). With higher levels of relationship conflict, team members would experience more clashes of opinion that could result in a lack of respect for each other (Jehn, Greer, Levine, & Szulanski, 2008), contradicting their initially high expectations for team psychological safety. We do not focus on task or process conflict in the current study because they do not as clearly confirm or disconfirm initial perceptions of psychological safety. A disagreement about a work problem or process could signal to team members that the team environment is safe for sharing opinions. However, disagreements related to personalities and values more directly threaten feelings of respect and trust within the team so increasing levels of relationship conflict would be expected to contradict high initials perceptions for team psychological safety.

Where a team initially expected a psychologically safe environment, an increasing trajectory of team relationship conflict should contradict those initial expectations, resulting in lower levels of individual’s team identification. As we discussed earlier, team identification is developed as a result of sense-breaking, sense-giving, and sense-making processes. When team members join a new team, they may experience a sense-breaking process in which team members consider whether to incorporate the team into their identities. Team members then experience a sense-giving process. Initial perceptions of team psychological safety contribute to the sense-giving process by providing clues to team members about what being a member of the team will be like in terms of risk-taking and making mistakes. Then, team members actually engage in experiences within the team and undergo a sense-making process. Here, team members may compare the level of relationship conflict they experience with what they expected from the team based on their initial perceptions of team psychological safety. Based on identification theory, team identification is facilitated to the extent that interactions and sense-making are consistent with what individuals expected from the team (Ashforth, 2001).

For example, a team may judge the level of safety at team formation as being high, only to discover through an increasing trajectory of relationship conflict that the team was not a safe place to share divergent ideas. This inconsistency between experiences and expectations would be expected to result in lower levels of identification. However, if the team thought that team psychological safety was relatively low to begin with, we would expect that an increasing trajectory of team relationship conflict would have less of an effect on subsequent levels of individual’s team identification. Our position is also generally supported by research showing that when newcomers’ experiences did not align with what they expected prior to entering an organization, newcomers experienced more negative attitudes toward the organization (Schaubroeck, Shaw, Duffy, & Mitra, 2008; Wanous, Poland, Premack, & Davis, 1992). Therefore, we propose there will be an interactive effect between team psychological safety at formation, and the perceived trajectory of team relationship conflict on individual’s team identification measured at the end of the team’s life cycle as follows:

Hypothesis 3: There will be an interactive effect of team relationship conflict trajectory with initial team psychological safety in predicting an individual’s team identification, such that there will be a negative relationship between an increasing team relationship conflict trajectory and team identification when initial team psychological safety is high, and this relationship will be attenuated when initial team psychological safety is low.

Satisfaction With Team

Our focus on the consistency between initial team psychological safety and the relationship conflict trajectory over a team’s time together is critical because this consistency ultimately affects the effectiveness of the team. Key outcomes that indicate team effectiveness include performance, the extent to which team members’ needs are met, and the desire of team members to stay with their teams (Kozlowski & Ilgen, 2006; Luciano, Mathieu, & Ruddy, 2014; Mathieu, Maynard, Rapp, & Gilson, 2008; Shaw et al., 2011). Separate from a team’s actual performance, it is essential to consider team members’ feelings toward the team and team viability when evaluating team effectiveness (Mathieu et al., 2008). Hackman (1987) emphasized the importance of team member need fulfillment, and satisfaction with the team indicates whether team members’ needs are being fulfilled (Phillips, 2001). If team members’ needs are not met and they are dissatisfied, the viability of the team could be threatened (Balkundi & Harrison, 2006; Mathieu et al., 2008). Satisfaction has been evaluated as a measure of team effectiveness in numerous studies on teams (e.g., Janz, Colquitt, & Noe, 1997; Kirkman & Rosen, 1999; Tesluk & Mathieu, 1999). We assess satisfaction from the individual team members’ perspective in line with our focus on examining each member’s level of team identification.

We argued that the consistency between initial levels of team psychological safety and relationship conflict trajectories would influence team members’ beliefs about whether team identification will result in need fulfillment. If the trajectory is increasing after the team had initially positive expectations for safe interactions, lower levels of individual’s team identification should result. We propose this inconsistency would also affect members’ satisfaction with their team, and that team identification is one mechanism for explaining why we would expect a negative effect on satisfaction.

Satisfaction with the team captures team members’ affect toward the team including interest in working with the team in the future (Phillips, 2001; Van der Vegt, Emans, & Vliert, 2001). When team members do not feel their needs are met, we expect them to experience lower levels of satisfaction with their team based on a dissonance argument, which has been used to explain why lower levels of job satisfaction result (Hollon & Chesser, 1976; Viswesvaran, Deshpande, & Joseph, 1998). Specifically, when dissonance results between initial expectations for psychological safety within a team and subsequent team interactions, members will experience lower levels of satisfaction with their team. We suggest the reason for this negative effect on individual’s team satisfaction results from higher levels of dissonance that negatively affect individual’s team identification (Ashforth, 2001; Ashforth et al., 2008; Riketta, 2005). In addition, feeling close to one’s team has been found to positively relate to satisfaction (Tekleab et al., 2009; Van Dick et al., 2004; Van Dick, van Knippenberg, Kerschreiter, Hertel, & Wieseke, 2008). Therefore, we propose the following:

Hypothesis 4: Individual’s team identification at the end of the team’s time together will mediate the negative interactive effect of team relationship conflict trajectory and initial team psychological safety on satisfaction with one’s team at the end of the team’s time together.

Method

Participants and Study Design

Participants in this 9-month study were 556 Master of Business Administration (MBA) and Master of Health Administration (MHA) students from a large Northwestern university. Seventy-four students were dropped from our study because they did not have complete data due to either dropping out of the program or not answering one of the surveys over the duration of the 9-month study. The remaining sample included 482 graduate students participating in 96 teams. The size of the teams ranged from two to seven team members with an average of five team members.¹ From the two different academic programs, 89 students (20 teams) were from an MBA program and 393 students (76 teams) were from three different MHA programs. Students within the MBA program were from one cohort and students from the MHA program represented five different cohorts, graduating from the MHA program across five consecutive years.

Teams were formed at the start of the students’ first year in the program and measures were assessed three times throughout the students’ first year in their respective 2-year graduate degree programs. Time 1 was several weeks into the first year of the program during which teams were being formed, and Times 2 and 3 were chosen to coincide with key transition points in their curriculum at approximately 3 to 4 month intervals. Time 3 was near the end of the team’s time together. All students worked with teams on various assignments, for example, case analyses, presentations, and a consulting project in all of their classes during their first year. Students’ membership in teams remained constant throughout this study for the entire period of time.

Measures

Team conflict

Team conflict was measured at Time 1, Time 2, and Time 3 with Jehn and Mannix’ (2001) nine items, which included three items representing each of the three types of conflict—relationship, task, and process. All items were measured on a 7-point Likert-type scale ranging from 1 = none/not at all to 7 = a lot and aggregated to the team level. An example item for relationship conflict is “how much relationship tension is there in your work group?” “How much conflict of ideas is in your group?” is an example item for task conflict and “how much conflict is there in your group about task responsibilities?” is an example item for process conflict. The scores on the Cronbach alphas for relationship, task, and process conflict scales, respectively, were as follows: Time 1: .90, .87, and .87, Time 2: .90, .87, and .90, and Time 3: .93, .89, and .90. To justify aggregating individual-level ratings of team conflict to the team level, we calculated $r_{w g}$ , ICC(1), and ICC(2) (Bartko, 1976; Shrout & Fleiss, 1979). These checks yielded acceptable values. According to LeBreton and Senter (2008), a value for ICC(1) above .25 is considered a “large” effect and a value between .70 and .85 for ICC(2) is sufficient to justify aggregation. See Table 1 for a summary of the agreement statistics for each conflict scale over three time periods.

Table 1.

Summary of Agreement Statistics for Aggregating Team Conflict.

Variable	R _WG	ICC(1)	ICC(2)	F statistic
Relationship conflict (T1)	.74	.44	.80	F(95, 386) = 4.96, p < .001
Relationship conflict (T2)	.75	.41	.78	F(95, 386) = 4.55, p < .001
Relationship conflict (T3)	.68	.45	.80	F(95, 386) = 5.05, p < .001
Task conflict (T1)	.68	.26	.63	F(95, 386) = 2.74, p < .001
Task conflict (T2)	.73	.38	.75	F(95, 386) = 4.06, p < .001
Task conflict (T3)	.68	.35	.73	F(95, 386) = 3.64, p < .001
Process conflict (T1)	.76	.19	.53	F(95, 386) = 2.14, p < .001
Process conflict (T2)	.72	.18	.53	F(95, 386) = 2.13, p < .001
Process conflict (T3)	.66	.25	.63	F(95, 386) = 2.70, p < .001

Note. ICC = intraclass correlation.

Initial team psychological safety

Initial perceptions of teams psychological safety were measured at Time 1 with Edmonson’s (1999) seven items and aggregated to the team level. An example item is “if you make a mistake on this team, it is often held against you (reverse-coded).” All items were measured on a 5-point Likert-type scale ranging from 1 = strongly disagree to 5 = strongly agree. We also included team psychological safety at Time 3 as a control variable. The scores on the Cronbach alpha at Times 1 and 3 were .75 and .82, respectively. The checks to justify aggregating individual-level ratings of team psychological safety to the team level showed acceptable values for Time 1 $r_{w g (j)}$ = .87, ICC(1) = .41, ICC(2) = .78; F(95, 384) = 4.47, p < .001 and Time 3 $r_{w g (j)}$ = .85, ICC(1) = .43, ICC(2) = .79; F(95, 380) = 4.81, p < .001.

Team identification

Individual’s team identification was measured at Time 3 with Van der Vegt and Bunderson’s (2005) four items on a 7-point Likert-type scale ranging from 1 = strongly disagree to 7 = strongly agree and analyzed at the individual level. An example item is “I do not feel a strong sense of belonging to my team (reverse-coded).” We also included team identification at Time 1 as a control variable. The scores on the Cronbach alpha at Times 1 and 3 were .72 and .79, respectively.

Satisfaction with team

Individual’s satisfaction with the team was measured at Time 3 with four of Phillips’ (2001) items and analyzed at the individual level. Phillips (2001) originally used eight items to measure satisfaction with the leader. A shortened scale was needed to capture satisfaction with the team over the 5 years of data collection. Items that best measured satisfaction with the team were selected and adapted to the team as the referent. All items were measured on a 5-point Likert-type scale ranging from 1 = strongly disagree to 5 = strongly agree. An example item is “overall, I am satisfied with my team.” The score on the Cronbach alpha was .84.

Analyses

Bayes-estimated slopes

We measured the trajectory of team relationship conflict at the team level using Bayes slope estimates (Chen et al., 2011). The slope represents the extent to which team conflict increases, decreases, or stays consistent. Consistent with Chen and colleagues (2011), we calculated the Bayes slope estimate using hierarchical linear models (Bliese & Ployhart, 2002) in HLM7 (Raudenbush, Bryk, Cheong, Congdon, & du Toit, 2011). A Bayes slope estimate was generated by conducting a separate two-level hierarchical linear model for each team where the aggregated relationship conflict score for each team was regressed on the three time periods coded as 0, 1, and 2. The empirical Bayes-estimated slope represents the change in team relationship conflict. Positive values of the slope indicate an increasing trajectory of team relationship conflict and negative values represent a decreasing trajectory. The range of the team relationship conflict trajectory in this study was from −1.50 to 1.64. The empirical Bayes-estimated slope was a Level 2 variable, representing differences in team relationship conflict trajectory between teams.

Multilevel analyses

We employed single-indicator multilevel structural equation modeling (ML-SEM; Leslie, Manchester, Park, & Mehng, 2012; Liu, Zhang, & Wang, 2012; Preacher, Zyphur, & Zhang, 2010; Zhang, LePine, Buckman, & Wei, 2014) using Mplus version 7.31 (Muthén & Muthén, 2004) to test our model. Multilevel methodologies are superior to aggregation models because they partition variance due to differences in individuals (within-level) from variance based on differences in groups (between-level; Heck, 2001). ML-SEM is preferable to other multilevel modeling techniques such as hierarchical linear modeling because it allows for the simultaneous evaluation of multiple relationships (Snijders, 1999). As recommended by Enders and Tofighi (2007), the Level 2 variables were grand mean centered.

Results

Confirmatory factor analysis (CFA)

We conducted CFA to demonstrate the distinction between relationship, task, and process conflict using Mplus version 7.31. First, to determine the appropriate estimator to use, we evaluated the skewness and kurtosis at the item level. According to Muthén and Kaplan (1985), skewness and kurtosis in the range of −1 to +1 is acceptable. Five of the 18 team conflict items exceeded +1 for skewness and three of the 18 team conflict items exceeded +1 for kurtosis. The range of skewness for the items that fell within the acceptable range was .039 to .996. The range of kurtosis for the items that fell within the acceptable range was −.989 to .666. Given there is some evidence of nonnormality at the univariate level, we proceeded to calculate Mardia’s coefficient of multivariate kurtosis to test for multivariate normality and it was significant (p > .001) suggesting the data are not multivariate normal. Therefore, we utilized the MLM estimator in Mplus 7.31 for the CFAs which is a maximum likelihood estimator robust for testing nonnormal data.

We conducted CFAs for team conflict at Times 1, 2, and 3. See Table 2 for a summary of our results. Although there have been studies challenging whether specific cut-off indices should be adhered to (e.g., Greiff & Heene, 2017), general guidelines presented by Hu and Bentler (1999) indicate that acceptable fit indices include values close to or lower than .08 for standardized root mean square residual (SRMR) and close to or higher than .95 for comparative fit index (CFI). For root mean square error approximation (RMSEA), Browne and Cudeck (1993) state that values of .08 or lower are reasonable. First, to support a three-factor structure for the three types of conflict, we loaded each set of three items onto relationship, task, and process conflict. Then we compared the three-factor structure to a two-factor structure in which we loaded the items for the two types of conflict with the highest correlation—relationship and process—onto one factor, and the items for task conflict onto a second factor. We then tested a single-factor structure in which all nine items loaded onto one conflict factor. We found consistent results for Times 1, 2, and 3. Overall, the fit indices for the three-factor structures at Times 1 and 3 met the general guidelines for cut-offs indicated by Hu and Bentler (1999) and Browne and Cudeck (1993). Collectively, the results of our CFAs provide confidence that the three types of conflict are distinct constructs.

Table 2.

Confirmatory Factor Analyses Results for Relationship, Task, and Process Conflict.

	χ²(N = 241)	df	Δχ²	Δdf	RMSEA (90% CI)	SRMR	CFI
Time 1
Three-factor structure	46.23**	24	—	—	.04 [.02, .06]	.03	.99
Two-factor structure	292.52***	26	+246.29***	2	.15 [.13, .16]	.07	.88
One-factor structure	485.61***	27	+193.09***	1	.19 [.17, .20]	.08	.79
Time 2
Three-factor structure	72.02***	24	—	—	.06 [.05, .08]	.03	.98
Two-factor structure	447.30***	26	+375.28***	2	.17 [.16, .19]	.07	.85
One-factor structure	625.94***	27	+178.64***	1	.22 [.20, .23]	.09	.75
Time 3
Three-factor structure	43.91**	24	—	—	.04 [.02, .06]	.02	.99
Two-factor structure	244.55***	26	+200.64***	2	.13 [.12, .15]	.05	.92
One-factor structure	423.36***	27	+178.81***	1	.18 [.16, .19]	.07	.85

Note. MLM estimator used due to evidence of multivariate nonnormality. Three-factor structure = three items each for relationship conflict, task conflict, and process conflict items load onto their respective factors; two-factor structure = relationship and process conflict items loaded onto one factor and task conflict items loaded onto a second factor; one-factor structure = all items loaded onto one factor. RMSEA = root mean square error approximation; CI = confidence interval; SRMR = standardized root mean square residual; CFI = comparative fit index; MLM = maximum likelihood estimator.

**p < .01. ***p < .001.

To examine the distinction between our mediator and dependent variables, we conducted an average variance extracted (AVE) test (Fornell & Larcker, 1981). According to Fornell and Larcker (1981), there are two requirements when using an AVE test to demonstrate that two constructs are distinct. First, the AVE for each construct must be greater than .5 which shows the variance captured by these constructs is greater than the variance due to measurement error. Second, the AVE for each of the constructs must be greater than the square of the correlation between the two constructs. The AVE for team identification (.65) and satisfaction with team (.58) were both above .5 and each are greater than the square of the correlation between the two constructs (.61 × .61 = .37), supporting that team identification and satisfaction with the team were distinct.

Multilevel analyses

To justify using multilevel methods, we calculated the intraclass correlation (ICC; Stapleton, 2006) to examine whether there is sufficient variance due to group level effects. The ICC represents the proportion of between-group variance to total variance (between-group plus within-group variance; Muthén, 1997). Sufficient between-group variance justifies using multilevel methods because variance exists due to being in groups. Our ICC for satisfaction with the team was sufficiently large, justifying the use of multilevel methods. The ICC suggests that 45% of the variance in satisfaction with the team was due to group-level effects (p < .001). A nonzero ICC suggests there is variance in the data due to group-level effects and multilevel methods are appropriate (Nezlek, 2008).

Hypotheses tests

Table 3 summarizes the means, standard deviations, and intercorrelations. We used Mplus version 7.31 to test our hypotheses. See Figure 1 for the summary of the structural equation modeling results. In the model and in all hypotheses, we control for relationship, task, and process conflict at Time 1, as well as type of academic program. Existing literature has shown relatively high correlations among the different types of team conflict (De Dreu & Weingart, 2003). Based on our theoretical arguments, initial ratings of team psychological safety were expected to signal team members that it is okay to speak up with their opinions without worrying about negative reactions from team members. Conflict that is associated with team member relationships would be expected to create an environment that felt less safe. Therefore, we control for all three types of conflict at Time 1 to examine the relationship conflict trajectory in our model. We control for team relationship conflict at Time 1 in particular because we expected that teams would differ in their initial levels of team relationship conflict. Our interest is in understanding the effect of team relationship conflict trajectory on team identification after taking into consideration initial levels of team relationship conflict. Past empirical studies that have examined trajectories have utilized a control variable (Hausknecht, Sturman, & Roberson, 2011; Liu, Mitchell, Lee, Holtom, & Hinkin, 2012).

Table 3.

Means, Standard Deviations, Intercorrelations.

Measure	M	SD	1	2	3	4	5	6	7	8	9	10	11	12
1. Day MHA program^a	.68	.47	—
2. Executive MHA program^b	.59	.59	−.58***	—
3. MBA program^c	.82	.39	−.33***	−.40***	—
4. Team relationship conflict (Time 1)	2.55	.99	−.04	.11*	−.03	(.90)
5. Team task conflict (Time 1)	3.29	.82	.04	.18***	−.30***	.71***	(.87)
6. Team process conflict (Time 1)	2.31	.65	−.05	.10*	−.03	.77**	.65***	(.87)
7. Team identification (Time 1)	4.96	1.32	−.20***	−.23***	.55***	−.18***	−.33***	−.17***	(.72)
8. Team relationship conflict trajectory (Times 1-3)	.18	.53	.09	−.04	−.03	−.38***	−.11*	−.20***	−.02	—
9. Initial team psychological safety (Time 1)	3.84	.48	−.22***	−.41***	.82***	−.43***	−.59***	−.40***	.62***	.08	(.75)
10. Team psychological safety (Time 3)	3.80	.57	−.13**	−.43***	.66***	−.31***	−.56***	−.35***	.52***	−.39***	.79***	(.82)
11. Team identification (Time 3)	5.13	1.45	−.18***	−.28***	.56***	−.15**	−.34***	−.17***	.62***	−.13**	.59***	.63***	(.79)
12. Satisfaction with team (Time 3)	3.96	.84	−.07	−.22***	.28***	−.34***	−.39***	−.31***	.39***	−.26***	.42***	.60***	.61***	(.84)

Note. n = 96 at the team level; n = 482 at the individual level. Only for this correlation table, the Level 2 variables (team relationship conflict, team task conflict, team process conflict, team relationship conflict trajectory, initial team psychological safety, team psychological safety) were calculated as team-level means and then assigned back to the individuals. Cronbach’s alphas are in italics in parentheses on the diagonal (except for the Day MHA, Executive MHA, MBA program controls, and relationship conflict trajectory).

Day MHA program coded as: 0 = Day MHA; 1 = Executive MHA, MHA Information Systems, or MBA.

Executive MHA program coded as: 0 = Executive MHA; 1 = Day MHA, MHA Information Systems, or MBA.

MBA program coded as: 0 = MBA; 1 = Day MHA, Executive MHA, or MHA Information Systems.

p < .05. **p < .01. ***p < .001.

We also tested alternate models with task and process conflict trajectories in place of relationship conflict trajectory. There were no significant interactions with team psychological safety (Time 1) and task conflict trajectory (b = −.24, p = .64) or process conflict trajectory (b = −.55, p = .22) on individual’s team identification (Time 3). We also controlled for team identification at Time 1 and team psychological safety at Time 3 to isolate the effects of the team relationship conflict trajectory (Time 1-Time 3) and initial team psychological safety (Time 1) on team identification at Time 3. Finally, we controlled for the type of academic program (MBA, day MHA, MHA information systems, and executive MHA) for participants in this study.

Hypothesis 1 stated that team relationship conflict trajectory from Time 1 through Time 3 will have a negative effect on individual’s team identification at Time 3. We find support for this negative relationship (b = −.35, p = .001). Hypothesis 2 states that initial team psychological safety ratings at Time 1 will have a positive effect on team identification at Time 3. We find support for the positive relationship between team psychological safety and team identification (b = 1.95, p < .001) providing support for Hypothesis 2. Hypothesis 3 tests the interactive effect of initial team psychological safety ratings at Time 1 and team relationship conflict trajectory from Time 1 through Time 3 on team identification at Time 3. We found a marginally significant interactive effect (b = −.48, p ≤ .05; see Figure 2). Simple slopes analyses support this interaction, revealing a significant negative effect of an increasing relationship conflict trajectory (Times 1-3) on team identification (Time 3) when team psychological safety is initially high (simple slope = −.59, SE =.08, t = −6.96, p < .001), but not when it is low (simple slope = −.11, SE =.10, t = −1.07, p = .29). This pattern indicates that at high levels of initial team psychological safety, an increasing trajectory of relationship conflict had a negative effect on team identification at Time 3. However, when there were low levels of initial team psychological safety, this relationship was absent, providing support for Hypothesis 3.

Figure 2.

Interaction of initial team psychological safety and team relationship conflict trajectory on team identification.

Hypothesis 4 proposed that individual’s team identification at Time 3 mediated the negative interactive effect of initial team psychological safety at Time 1, and team relationship conflict trajectory from Time 1 through Time 3 on satisfaction with the team at Time 3. To support this mediated moderation hypothesis, we also need to demonstrate a significant effect of our mediator on our dependent variable (Liu, Zhang, & Wang, 2012). This relationship was supported in that a higher level of individual’s team identification at Time 3 was associated with higher levels of satisfaction with the team at Time 3 (b = .33, p < .001). We used the Monte Carlo method for assessing mediation (Selig & Preacher, 2008) by constructing bias-corrected 95% confidence intervals (CIs) for the indirect effects by generating 20,000 random resamples. If the CI does not contain zero, the indirect effect is supported. Hypothesis 4 was supported by a significant negative indirect effect of the interaction between initial team psychological safety and team relationship conflict trajectory on satisfaction with team through individual’s team identification (95% CI = [−.26, −.08]). Therefore, the interaction between initial team psychological safety and the increasing trajectory of team relationship conflict had a negative effect on satisfaction with team because of its effect on team identification.

General Discussion

Our study provided support for our primary theoretical contribution in that lower levels of individual’s team identification resulted when the trajectory of team relationship conflict over time was inconsistent with initially high ratings of team psychological safety. An increasing trajectory of relationship conflict was related to lower levels of individual’s team identification (Hypothesis 1), and higher initial perceptions of team psychological safety were positively related to team identification (Hypothesis 2). However, the effect on team identification depended on the interaction between initial team psychological safety and team relationship conflict trajectory. If team members initially perceived higher levels of team psychological safety (Time 1), then increasing levels of team relationship conflict (Times 1-3) resulted in lower levels of individual’s team identification later in their teams’ life cycle (Time 3; Hypothesis 3). However, if team members initially perceived lower levels of team psychological safety, there was no effect of team relationship conflict trajectory on team identification. This inconsistency ultimately had an effect on satisfaction with the team. The interaction between initial team psychological safety and team relationship conflict trajectory had a negative effect on satisfaction with the team, mediated by team identification (Hypothesis 4).

Theoretical Implications and Contributions

We contribute to prior team and organizational research in three important ways. First, we contribute to the identification literature (Ashforth, 1998; Ashforth et al., 2008; Ashforth & Mael, 1989) by expanding the types of antecedents to identification that have been examined to include initial perceptions of team psychological safety (Edmonson, 1999). Existing studies have demonstrated salience, levels of intragroup competition, and prestige and/or distinctiveness as antecedents to team identification (Dukerich et al., 2002; Dutton et al., 1994; Riketta, 2005; Smidts et al., 2001), and we expand this to include a team factor. Future research should now investigate how other team factors beyond team psychological safety can influence how team members’ personal, emotional, or work-related needs are met through the fostering of individual’s identification with a team, particularly during initial team formation.

Second, we investigated how individual’s team identification formed as a result of interactions within a team over a 9-month time span, and were able to show that trajectories of team relationship conflict versus just examining initial perceptions of conflict predicted subsequent levels of team identification.

The development of identification has been repeatedly depicted as an unfolding process (Ashforth, 1998; Ashforth et al., 2008) based on qualitative research studies that have shown how higher levels of identification resulted from successful sense-breaking and sense-making practices (Pratt, 2000). However, to our knowledge, there has been no prior empirical research testing how team identification is developed as a result of team experiences over time. We take the first step by focusing on the effects of team relationship conflict and comparing whether it was increasing, decreasing, or remaining the same, as is true in other research domains such as with turnover (Chen et al., 2011; Liu, Mitchell, et al., 2012). Further, we demonstrate the value of examining the trajectory of relationship conflict in predicting individual’s team identification. Beyond initial levels of team conflict, the pattern or trajectory of team conflict has an impact on individual’s team identification later in a team’s time together.

Third, we demonstrate the importance of considering the consistency in initial expectations and subsequent experiences. We found that it was not only one’s initial perceptions of team psychological safety during team formation that affected subsequent levels of individual’s team identification, but also how the trajectory of team relationship conflict interacted with those initial perceptions to affect team identification. If team members experienced an increasing trajectory of relationship conflict after initially perceiving higher levels of team psychological safety, those team members subsequently experienced lower levels of identification, and also rated their satisfaction with the team lower nearly 9 months later. Considering whether initial expectations are met by subsequent team interactions and how this consistency affects team identification is also new to the team identification literature (Ashforth, 1998; Ashforth et al., 2008; Ashforth & Mael, 1989; Pratt, 1998).

We suggest that our research provides a glimpse into a more dynamic view of how members in teams form and interact, particularly with respect to how their team identification is affected by the trajectory of relationship conflict teams experience over time. This is in line with Ashforth (1998) who described how identification is formed through individuals, teams, and organizations shaping each other and comingling.

Practical Implications and Future Directions

Our study highlights the importance of the dynamics that are established early in a team’s formation and how those dynamics may influence the formation of individual’s team identification. Based on our findings, it would make sense to introduce training and developmental interventions that foster realistic and hopefully higher levels of team psychological safety during the team formation process. These higher feelings of team psychological safety could then reduce feelings of uncertainty regarding team member interactions. This type of developmental intervention could help to build increasingly higher levels of team identification, as they serve to reinforce each other over time.

These developmental interventions might also focus on how to manage relationship conflict when it arises and build a team environment that is more transparent and safe for team members to exchange views. Initial evidence to support this focus was provided by Tekleab and colleagues (2009), who concluded based on their longitudinal study of teams that conflict management strategies moderated the relationship between team relationship conflict and team cohesion when there was higher cohesion and conflict was present. Future research might also explore designing interventions that involve team members gaining a better understanding of one another. In addition, mechanisms such as conflict strategies and creating contracts that clarify norms and expectations could set expectations that help address conflict, rather than only relying upon the formation of expectations based on first or early impressions. By developing a more realistic view of how the team can operate, we expect team members will interact and experience conflict in a manner that is more consistent with initial expectations, fostering greater team identification.

Our study sets the foundation for future research to expand upon examining teams’ initial perceptions by exploring other individual and team psychological states. Perceptions of team efficacy during team formation (Bandura, 1997) could fulfill members’ needs for self-efficacy (Ashforth, 2001; Erez & Earley, 1993) and thereby foster higher levels of team identification and positive team outcomes (Hogg & Terry, 2000). At the team level, team affect could be studied to determine whether it (when positive) facilitates team identification by fulfilling members’ need for belongingness (Leonardelli, Pickett, & Brewer, 2010).

A deeper understanding of the processes underlying the influence of initial perceptions on team identification can be gained by specifying the psychological motive(s) that is (are) fulfilled. For example, existing research has identified the moderating effect of diversity on team identification (Kearney & Gebert, 2009; Kearney et al., 2009). However, mediators for the relationship between diversity and team identification can also be examined in that a requisite level of diversity should fulfill an individual’s need for distinctiveness. Future research could also examine the types of individual and contextual factors that influence the impression a team forms initially that then impacts their perceived team psychological safety, team identification, and satisfaction.

Another potential area to investigate is to identify factors that may influence whether teams can manage their relationship conflict trajectory effectively to minimize negative effects. The existing empirical studies on team conflict demonstrate mixed findings on whether conflict has positive or negative effects (De Wit, Greer, & Jehn, 2012). Therefore, future research can identify moderators beyond team psychological safety that may influence whether an increasing relationship conflict trajectory has a positive or negative effect on team effectiveness. For example, a transformational leader who takes time to understand individual team members and relationships among team members may be able to manage relationship conflict more effectively (Bass & Bass, 2008; Sivasubramaniam, Jung, Avolio, & Murry, 2002).

For each of the areas above, mixing both qualitative and quantitative methods may serve to move this research forward in ways that either method alone could not accomplish (Creswell, 2014). For example, it would be quite beneficial to both interview and observe teams interacting during formation to assess how uncertain members are in interacting with each other, as well as what may be the early roots for explaining conflict trajectories. It would be interesting to observe the events during formation that help shape initial perceptions of feeling safe to fully engage with other team members. Also, interviewing team members during this early formation about their feelings of psychological safety and interactions that may have contributed to these feelings would be a useful addition to future research methods.

Finally, we recommend that future research on conflict trajectories replicate our study. Replicating interactive effects is always a gold standard for empirical research and would further bolster our findings and interpretation of our results.

Limitations

There are limitations to our research design and methods that are worth noting. First, our study involved graduate students, which may affect our ability to generalize our findings to other team contexts. However, our sample included master’s-level students with considerable work experience pursuing degrees in health care administration or business, and participants were engaged in real team activities such as consulting projects and simulations that are relevant to teamwork in organizations. Both graduate programs from which the teams came from were designed to provide students with practical experience and challenges working in teams. Nevertheless, future research should determine whether our findings generalize to other organizational settings where teams may also stay together longer.

Second, all of the measures used over the three time periods were self-report, which introduced the potential for common method/source effects or bias. However, each of these measures were shown to represent distinct constructs/measures and were collected over a considerable period of time. Specifically, the independent variable in our model representing team psychological safety was measured at Time 1, while the trajectory of perceived conflict (a Bayes-estimated slope) was examined over three points in time over a 9-month period. Our mediator and dependent variables were measured at Time 3, thus reducing the risk of common method/source bias associated with our independent variables (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). We also provided support for the discriminant validity of our measures using confirmatory factor analyses and an AVE test. Future research should minimize the effects of common method/source bias by measuring dependent variables after the mediator data are collected and if possible, using different sources.

It also would have been useful to focus on measuring how team members perceived the gaps between what they initially expected in terms of psychological safety and what they experienced at each point in time. Finally, there are certainly other constructs that were not measured such as team leadership, how teams set goals for their work, and feedback mechanisms that could enhance our understanding of each of the theoretical links tested in our model.

Although it would be beneficial to examine how initial perceptions and subsequent experiences within the team influence team performance, our study focused on perceptions, motivations, and attitudes of team members. Therefore, our outcomes of interest were how individuals’ team identification predicted their satisfaction with the team. The ultimate effect on how the team performs, based on each member’s identification, should be investigated in future research. Our field study was also not an experiment, where we could test cause and effect. One could experiment with different strategies for influencing levels of team psychological safety and subsequent interventions to help resolve team conflict to determine the effects on team member identification and satisfaction. Finally, the timing used to collect our measures may not necessarily have coincided with key team developmental points. Future research could examine how the timing of data collection impacts the trajectories of conflict observed here.

Conclusion

The formation of identification within teams occurs over time and initial perceptions such as team psychological safety appear to play an important role in team members’ identification and satisfaction. We found that initial perceptions of team psychological safety fostered individual’s team identification, but that inconsistencies with subsequent trends in perceived relationship conflict can actually be detrimental to an individual’s team identification and satisfaction with team. The sense-giving and sense-making processes individuals experience within teams occurs over time to affect team identification and should be high priorities for future research.

Footnotes

Acknowledgements

We would like to thank Crystal Farh, Cristiano Guarana, Fong Keng-Highberger, Dong Liu, Elizabeth Umphress, and Zhen Zhang for helpful feedback on this research.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Associate Editor: M. Travis Maynard

Notes

Author Biographies

Hana Huang Johnson is an assistant professor of Management and Human Resources in the College of Business and Economics, University of Idaho. She received a PhD in organizational behavior from the University of Washington Foster School of Business, an MBA in marketing from the Anderson School at University of California at Los Angeles, and a BBA in Accounting from the University of Washington. Prior to pursuing her PhD, she worked in brand management for the Starbucks Coffee Company and Procter and Gamble.

Bruce J. Avolio, PhD, is the Mark Pigott Chair in Business Strategic Leadership in the Foster School of Business at the University of Washington. Dr. Avolio serves as the executive director of the Center for Leadership & Strategic Thinking in the Foster School and can be reached at bavolio@uw.edu.

References

Ashforth

B. E.

(1998). Becoming: How does the process of identification unfold? In Whetten

D. A.

Godfrey

P. C.

(Eds.), Identity in organizations: Building theory through conversations (pp. 213-222). Thousand Oaks, CA: Sage. doi:10.4135/9781452231495

Ashforth

B. E.

(2001). Role transitions in organizational life: An identity-based perspective. Mahwah, NJ: Lawrence Erlbaum. doi:10.2307/3094833

Ashforth

B. E.

Harrison

S. H.

Corley

K. G.

(2008). Identification in organizations: An examination of four fundamental questions. Journal of Management, 34, 325-374. doi:10.1177/0149206308316059

Ashforth

B. E.

Mael

(1989). Social identity theory and the organization. Academy of Management Review, 14, 20-39. doi:10.5465/amr.1989.4278999

Balkundi

Harrison

D. A.

(2006). Ties, leaders, and time in teams: Strong inference about network structure’s effects on team viability and performance. Academy of Management Journal, 49, 49-68. doi:10.5465/AMJ.2006.20785500

Bandura

(1997). Self-efficacy: The exercise of control. New York, NY: W.H. Freeman.

Bartko

J. J.

(1976). On various intraclass correlation reliability coefficients. Psychological Bulletin, 83, 762-765. doi:10.1037/0033-2909.83.5.762

Bass

B. M.

Bass

(2008). The handbook of leadership. New York, NY: Free Press.

Bezrukova

Jehn

K. A.

Zanutto

E. L.

Thatcher

S. M. B.

(2009). Do workgroup faultlines help or hurt? A moderated model of faultlines, team identification, and group performance. Organization Science, 20, 35-50. doi:10.1287/orsc.1080.0379

10.

Bliese

P. D.

Ployhart

R. E.

(2002). Growth modeling using random coefficient models: Model building, testing, and illustrations. Organizational Research Methods, 5, 362-387. doi:10.1177/109442802237116

11.

Brickson

S. L.

(2013). Athletes, best friends, and social activists: An integrative model accounting for the role of identity in organizational identification. Organization Science, 24, 226-245. doi:10.1287/orsc.1110.0730

12.

Browne

M. W.

Cudeck

(1993). Alternative ways of assessing model fit. SAGE Focus Editions, 154, 136-136.

13.

Chen

G. L.

(2005). Newcomer adaptation in teams: Multilevel antecedents and outcomes. Academy of Management Journal, 48, 101-116. doi:10.5465/amj.2005.15993147

14.

Chen

G. L.

Ployhart

R. E.

Thomas

H. C.

Anderson

Bliese

P. D.

(2011). The power of momentum: A new model of dynamic relationships between job satisfaction change and turnover intentions. Academy of Management Journal, 54, 159-181. doi:10.5465/amj.2011.59215089

15.

Creswell

(2014). Research design: Qualitative, quantitative, and mixed methods approaches. Thousand Oaks, CA: Sage.

16.

De Dreu

C. K.

Weingart

L. R

. (2003). Task versus relationship conflict, team performance, and team member satisfaction: A meta-analysis. Journal of Applied Psychology, 88, 741-749. doi:10.1037/0021-9010.88.4.741

17.

Detert

J. R.

Burris

E. R.

(2007). Leadership behavior and employee voice: Is the door really open? Academy of Management Journal, 50, 869-884. doi:10.5465/amj.2007.26279183

18.

De Wit

F. R.

Greer

L. L.

Jehn

K. A

. (2012). The paradox of intragroup conflict: A meta-analysis. Journal of Applied Psychology, 97, 360-390. doi:10.1037/a0024844

19.

Duffy

M. K.

Scott

K. L.

Shaw

J. D.

Tepper

B. J.

Aquino

(2012). A social context model of envy and social undermining. Academy of Management Journal, 55, 643-666. doi:10.5465/amj.2009.0804

20.

Dukerich

J. M.

Golden

B. R.

Shortell

S. M.

(2002). Beauty is in the eye of the beholder: The impact of organizational identification identity and image on the cooperative behaviors of physicians. Administrative Science Quarterly, 47, 507-533. doi:10.2307/3094849

21.

Dutton

J. E.

Dukerich

J. M.

Harquail

C. V.

(1994). Organizational images and member identification. Administrative Science Quarterly, 39, 239-263. doi:10.2307/2393235

22.

Edmonson

(1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44, 350-383. doi:10.2307/2666999

23.

Enders

C. K.

Tofighi

(2007). Centering predictor variables in cross-sectional multilevel models: A new look at an old issue. Psychological Methods, 12, 121-138. doi:10.1037/1082-989X.12.2.121

24.

Erez

Earley

P. C.

(1993). Culture, self-identity, and work. New York, NY: Oxford University Press. doi:10.1093/acprof: oso/9780195075809.001.0001

25.

Fornell

Larcker

D. F.

(1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39-50. doi:10.2307/3151312

26.

Greiff

Heene

(2017). Why psychological assessment needs to start worrying about model fit. European Journal of Psychological Assessment, 33, 313-317. doi:10.1027/1015-5759/a000450

27.

Hackman

J. R.

(1987). The design of work teams. In Lorsch

(Ed.), Handbook of organizational behavior (pp. 315-342). New York, NY: Prentice Hall.

28.

Hausknecht

J. P.

Sturman

M. C.

Roberson

Q. M.

(2011). Justice as a dynamic construct: Effects of individual trajectories on distal work outcomes. Journal of Applied Psychology, 96, 872-880. doi:10.1037/a0022991

29.

Heck

R. H.

(2001). Multilevel modeling with SEM. In Marcoulides

G. A.

Schumacker

R. E.

(Eds.), New developments and techniques in structural equation modeling (pp. 89-127). Mahwah, NJ: Lawrence Erlbaum.

30.

Hogg

M. A.

(2000). Subjective uncertainty reduction through self-categorization: A motivational theory of social identity processes. European Review of Social Psychology, 11, 223-256. doi:10.1080/14792772043000040

31.

Hogg

M. A.

(2007). Uncertainty–identity theory. In Mark

P. Z.

(Ed.), Advances in experimental social psychology (Vol. 39, pp. 69-126). Cambridge, MA: Academic Press. doi:10.1016/s0065-2601(06)39002-8

32.

Hogg

M. A.

Terry

D. J.

(2000). Social identity and self-categorization processes in organizational contexts. Academy of Management Review, 25, 121-140. doi:10.5465/amr.2000.2791606

33.

Hollon

C. J.

Chesser

R. J.

(1976). The relationship of personal influence dissonance to job tension, satisfaction and involvement. Academy of Management Journal, 19, 308-314. doi:10.2307/255780

34.

L.-t.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1-55. doi:10.1080/10705519909540118

35.

Janz

B. D.

Colquitt

J. A.

Noe

R. A.

(1997). Knowledge worker team effectiveness: The role of autonomy, interdependence, team development, and contextual support variables. Personnel Psychology, 50, 877-904. doi:10.1111/j.1744-6570.1997.tb01486.x

36.

Jehn

K. A.

(1995). A multimethod examination of the benefits and detriments of intragroup conflict. Administrative Science Quarterly, 40, 256-282. doi:10.2307/2393638

37.

Jehn

K. A.

Bezrukova

(2010). The faultline activation process and the effects of activated faultlines on coalition formation, conflict, and group outcomes. Organizational Behavior and Human Decision Processes, 112, 24-42. doi:10.1016/j.obhdp.2009.11.008

38.

Jehn

K. A.

Greer

Levine

Szulanski

(2008). The effects of conflict types, dimensions, and emergent states on group outcomes. Group Decision and Negotiation, 17, 465-495. doi:10.1007/s10726-008-9107-0

39.

Jehn

K. A.

Mannix

E. A.

(2001). The dynamic nature of conflict: A longitudinal study of intragroup conflict and group performance. Academy of Management Journal, 44, 238-251. doi:10.2307/3069453

40.

Kearney

Gebert

(2009). Managing diversity and enhancing team outcomes: The promise of transformational leadership. Journal of Applied Psychology, 94, 77-89. doi:10.1037/a0013077

41.

Kearney

Gebert

Voelpel

S. C.

(2009). When and how diversity benefits teams: The importance of team members’ need for cognition. Academy of Management Journal, 52, 581-598. doi:10.5465/amj.2009.41331431

42.

Kirkman

Rosen

(1999). Beyond self-management: Antecedents and consequences of team empowerment. Academy of Management Journal, 42, 58-74. doi:10.2307/256874

43.

Kozlowski

S. W. J.

Ilgen

D. R.

(2006). Enhancing the effectiveness of work groups and teams. Psychological Science in the Public Interest, 7, 77-124. doi:10.1111/j.1529-1006.2006.00030.x

44.

LeBreton

J. M.

Senter

J. L.

(2008). Answers to 20 questions about interrater reliability and interrater agreement. Organizational Research Methods, 11, 815-852. doi:10.1177/1094428106296642

45.

Lee

Farh

J. L.

Chen

Z. J.

(2011). Promoting group potency in project teams: The importance of group identification. Journal of Organizational Behavior, 32, 1147-1162. doi:10.1002/job.741

46.

Leonardelli

G. J.

Pickett

C. L.

Brewer

M. B.

(2010). Optimal distinctiveness theory: A framework for social identity, social cognition, and intergroup relations. Advances in Experimental Social Psychology, 43, 63-113. doi:10.1016/s0065-2601(10)43002-6

47.

Leslie

Manchester

Park

T. Y.

Mehng

S. A.

(2012). Flexible work practices: A source of career premiums or penalties? Academy of Management Journal, 55, 1407-1428. doi:10.5465/amj.2010.0651

48.

Liu

Mitchell

T. R.

Lee

T. W.

Holtom

B. C.

Hinkin

T. R.

(2012). When employees are out of step with coworkers: How job satisfaction trajectory and dispersion influence individual- and unit-level voluntary turnover. Academy of Management Journal, 55, 1360-1380. doi:10.5465/amj.2010.0920

49.

Liu

Zhang

Wang

(2012). Mono-level and multilevel mediated moderation and moderated mediation: Theorization and test. In Chen

Tsui

Farh

(Eds.), Management research methods (in Chinese) (pp. 545-479). Beijing, China: Peking University Press.

50.

Luciano

M. M.

Mathieu

J. E.

Ruddy

T. M.

(2014). Leading multiple teams: Average and relative external leadership influences on team empowerment and effectiveness. Journal of Applied Psychology, 99, 322-331. doi:10.1037/a0035025

51.

Mathieu

Maynard

M. T.

Rapp

Gilson

(2008). Team effectiveness 1997-2007: A review of recent advancements and a glimpse into the future. Journal of Management, 34, 410-476. doi:10.1177/0149206308316061

52.

Muthén

B. O.

(1997). Latent variable modeling of longitudinal and multilevel data. Sociological Methodology, 27, 453-480. doi:10.1111/1467-9531.271034

53.

Muthén

B. O.

Kaplan

(1985). A comparison of some methodologies for the factor analysis of non-normal Likert variables. British Journal of Mathematical and Statistical Psychology, 38, 171-189. doi:10.1111/j.2044-8317.1985.tb00832.x

54.

Muthén

B. O.

Muthén

L. K.

(2004). Mplus user’s guide. Los Angeles, CA: Author.

55.

Nezlek

J. B.

(2008). An introduction to multilevel modeling for social and personality psychology. Social & Personality Psychology Compass, 2, 842-860. doi:10.1111/j.1751-9004.2007.00059.x

56.

Phillips

J. M.

(2001). The role of decision influence and team performance in member self-efficacy, withdrawal, satisfaction with the leader, and willingness to return. Organizational Behavior and Human Decision Processes, 84, 122-147. doi:10.1006/obhd.2000.2922

57.

Podsakoff

P. M.

MacKenzie

S. B.

Lee

J. Y.

Podsakoff

N. P.

(2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879-903. doi:10.1037/0021-9010.88.5.879

58.

Polzer

J. T.

Milton

L. P.

Swann

W. B.

Jr. (2002). Capitalizing on diversity: Interpersonal congruence in small work groups. Administrative Science Quarterly, 47, 296-324. doi:10.2307/3094807

59.

Pratt

M. G.

(1998). To be or not to be? Central questions in organizational identification. In Whetten

D. A.

Godfrey

P. C.

(Eds.), Identity in organizations: Building theory through conversations (pp. 171-207). Thousand Oaks, CA: Sage. doi:10.4135/9781452231495.n6

60.

Pratt

M. G.

(2000). The good, the bad, and the ambivalent: Managing identification among Amway distributors. Administrative Science Quarterly, 45, 456-493. doi:10.2307/2667106

61.

Preacher

K. J.

Zyphur

M. J.

Zhang

(2010). A general multilevel SEM framework for assessing multilevel mediation. Psychological Methods, 15, 209-233. doi:10.1037/a0020141

62.

Raudenbush

Bryk

Cheong

Congdon

du Toit

(2011). HLM 7. Lincolnwood, IL: Scientific Software International.

63.

Riketta

(2005). Organizational identification: A meta-analysis. Journal of Vocational Behavior, 66, 358-384. doi:10.1016/j.jvb.2004.05.005

64.

Schaubroeck

Shaw

J. D.

Duffy

M. K.

Mitra

(2008). An under-met and over-met expectations model of employee reactions to merit raises. Journal of Applied Psychology, 93, 424-434. doi:10.1037/0021-9010.93.2.424

65.

Selig

J. P.

Preacher

K. J.

(2008). Monte Carlo method for assessing mediation: An interactive tool for creating confidence intervals for indirect effects [Computer software]. Available from http://quantpsy.org/

66.

Shaw

J. D.

Zhu

Duffy

M. K.

Scott

K. L.

Shih

H. A.

Susanto

(2011). A contingency model of conflict and team effectiveness. Journal of Applied Psychology, 96, 391-400. doi:10.1037/a0021340

67.

Shrout

Fleiss

J. L.

(1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86, 420-428. doi:10.1037/0033-2909.86.2.420

68.

Sivasubramaniam

Jung

D. I.

Avolio

B. J.

Murry

W. D.

(2002). A longitudinal model of the effects of team leadership and group potency on group performance. Group & Organization Management, 27, 66-96. doi:10.1177/1059601102027001005

69.

Smidts

Pruyn

A. T. H.

van Riel

C. B. M.

(2001). The impact of employee communication and perceived external prestige on organizational identification. Academy of Management Journal, 44, 1051-1062. doi:10.2307/3069448

70.

Snijders

T. A. B.

(1999). Multilevel analysis: An introduction to basic and advanced multilevel modeling. London, UK: Sage.

71.

Stapleton

L. M.

(2006). An assessment of practical solutions for structural equation modeling with complex sample data. Structural Equation Modeling: A Multidisciplinary Journal, 13, 28-58. doi:10.1207/s15328007sem1301_2

72.

Tajfel

Turner

J. C.

(1986). Social identity theory and intergroup behavior (2nd ed.). Chicago, IL: Nelson-Hall Publishers.

73.

Tekleab

A. G.

Quigley

N. R.

Tesluk

P. E.

(2009). A longitudinal study of team conflict, conflict management, cohesion and team effectiveness. Group & Organization Management, 34, 170-205. doi:10.1177/1059601108331218

74.

Tesluk

P. E.

Mathieu

J. E.

(1999). Overcoming roadblocks to effectiveness: Incorporating management of performance barriers into models of work group effectiveness. Journal of Applied Psychology, 84, 200-217. doi:10.1037/0021-9010.84.2.200

75.

Tuckman

B. W.

(1965). Developmental sequence in small groups. Psychological Bulletin, 63, 384-399. doi:10.1037/h0022100

76.

Turner

J. C.

Hogg

Oakes

P. J.

Reicher

D. D.

Wetherell

M. S.

(1987). Rediscovering the social group: Self-categorization theory. Oxford, UK: Basil Blackwell. doi:10.2307/2073157

77.

Van der Vegt

G. S.

Bunderson

J. S

. (2005). Learning and performance in multidisciplinary teams: The importance of collective team identification. Academy of Management Journal, 48, 532-547. doi:10.5465/amj.2005.17407918

78.

Van der Vegt

G. S.

Emans

B. J. M.

Vliert

. (2001). Patterns of interdependence in work teams: A two-level investigation of the relations with job and team satisfaction. Personnel Psychology, 54, 51-69. doi:10.1111/j.1744-6570.2001.tb00085.x

79.

Van der Vegt

G. S.

Van de Vliert

Oosterhof

. (2003). Informational dissimilarity and organizational citizenship behavior: The role of intrateam interdependence and team identification. Academy of Management Journal, 46, 715-727. doi:10.2307/30040663

80.

Van Dick

Christ

Stellmacher

Wagner

Ahlswede

Grubba

. . . Tissington

P. A

. (2004). Should I stay or should I go? Explaining turnover intentions with organizational identification and job satisfaction. British Journal of Management, 15, 351-360. doi:10.1111/j.1467-8551.2004.00424.x

81.

Van Dick

Van Knippenberg

Hägele

Guillaume

Brodbeck

. (2008). Group diversity and group identification: The moderating role of diversity beliefs. Human Relations, 61, 1463-1492. doi:10.1177/0018726708095711

82.

Van Dick

van Knippenberg

Kerschreiter

Hertel

Wieseke

. (2008). Interactive effects of work group and organizational identification on job satisfaction and extra-role behavior. Journal of Vocational Behavior, 72, 388-399. doi:10.1016/j.jvb.2007.11.009

83.

Viswesvaran

Deshpande

Joseph

(1998). Job satisfaction as a function of top management support for ethical behavior: A study of Indian managers. Journal of Business Ethics, 17, 365-371.

84.

Wanous

J. P.

Poland

T. D.

Premack

S. L.

Davis

K. S.

(1992). The effects of met expectations on newcomer attitudes and behaviors: A review and meta-analysis. Journal of Applied Psychology, 77, 288-297. doi:10.1037/0021-9010.77.3.288

85.

Weick

K. E.

(1995). Sensemaking in organizations. Thousand Oaks, CA: Sage.

86.

Zhang

LePine

J. A.

Buckman

B. R.

Wei

(2014). It’s not fair . . . or is it? The role of justice and leadership in explaining work stressor-job performance relationships. Academy of Management Journal, 57, 675-697. doi:10.5465/amj.2011.1110