Exploring the Dynamic Team Cohesion–Performance and Coordination–Performance Relationships of Newly Formed Teams

Abstract

Emergent states, like cohesion, and behavioral processes, like coordination, are considered critical for team performance, yet little is known regarding their relative importance over time. Testing several hypotheses and exploring multiple research questions, this study used a laboratory design to better understand the evolution of cohesion–performance and coordination–performance relationships of newly formed teams. Forty-two teams of three completed 10 task episodes of an interdependent team task over the course of two and a half hours. Results reveal that cohesion and performance have a reciprocal relationship such that cohesion predicts subsequent performance, which then predicts subsequent cohesion. Moreover, coordination predicts subsequent performance, but performance does not predict subsequent coordination. Both the cohesion–performance and coordination–performance relationships weaken as the team works together, indicating that other states and processes predict performance at later stages of team development. Meanwhile, the relative importance of cohesion compared with coordination strengthens with increased team interactions.

Keywords

team cohesion team coordination team performance longitudinal data analysis dominance analysis

As the world evolves, so does the structure of jobs and organizations. Through this restructuring process, teams—whether they are action teams, project teams, virtual teams, or short- or long-term teams—have become the fundamental building blocks of organizations (Campion et al., 1993; Kozlowski & Ilgen, 2006). Personal and organizational successes and failures now hinge on the functioning and performance¹ of the teams at the helm. In some cases, team performance not just influences profit margins, but also has serious implications for societal well-being. Not surprisingly, the increased reliance on teams in organizations spurred a surge of interest in understanding and predicting team performance. More than six decades of research on small group and team performance exists (see Kozlowski & Ilgen, 2006; Mathieu et al., 2008; Salas et al., 2004, for reviews) and has identified team emergent states and team behavioral processes as proximal predictors of team success (Marks et al., 2001).

Despite a wealth of research examining the link between emergent states and behavioral processes with team performance that has resulted in a multitude of important findings, there are many untested assumptions regarding how these relationships function over time. Research on team phenomena, like many constructs in the organizational sciences, suffers from an overreliance on self-report, cross-sectional research designs (Cronin et al., 2011; Kozlowski, 2015). Unfortunately, cross-sectional findings frequently do not generalize to temporal relationships (Molenaar, 2004), which has been demonstrated many times across a wide range of phenomena (e.g., Block, 1995; Kozlowski & Chao, 2012; Vancouver et al., 2001). One untested implicit assumption in the team performance literature is that effect size estimates are stable across time. Meta-analytic estimates based on primarily cross-sectional data are not qualified depending on when the measurement of emergent states, behavioral processes, or team performance occurred during the task or life cycle of the team (e.g., Beal et al., 2003; Castano et al., 2013; LePine et al., 2008), implicitly equating time. Likewise, any meta-analytic variability in the predictor–outcome relationships is typically attributed entirely to between-study differences, ignoring potential within-study moderators such as measurement timing. Even longitudinal studies of team performance (e.g., Mathieu et al., 2015; Tekleab et al., 2009) often implicitly treat the strength of the relationships as stable across time by estimating a single parameter across all measurement points.

One implication of implicitly assuming stable effect sizes in the relationships between emergent states and behavioral processes with team performance is that the time scale for when and how strongly these constructs predict performance in newly formed teams is currently unknown. Similarly, how these relationships change as the team begins to work together and evolve through the various phases of team development (e.g., Kozlowski et al., 1999) is not well understood. Emergent states, in particular, form through the interactions among team members (Kozlowski & Klein, 2000) leading to their mental representations changing through increased interactions (Carter et al., 2015). Understanding how team emergent states and behavioral processes impact performance initially and across multiple task episodes is critical for improving their utility to enhance team performance outcomes.

Through the six-plus decades of research on team performance, a multitude of critical emergent states and behavioral processes have been identified. An exhaustive examination of all of them is infeasible and well beyond the scope of this study. Rather, we focus on one emergent state, team cohesion, and one behavioral process, team coordination, to examine how these phenomena affect team performance over time in newly formed teams. Team cohesion was chosen because it is one of the most commonly studied team states (LePine et al., 2008), being consistently demonstrated as vital for team success through multiple meta-analyses (e.g., Beal et al., 2003; Castano et al., 2013; Gully et al., 1995). Likewise, coordination was chosen because it has routinely been empirically and theoretically considered critical for positive team performance (e.g., Bowers et al., 1992; LePine et al., 2008; Summers et al., 2012).

Therefore, the purpose of this study is to better understand the initial evolution of cohesion–performance and coordination–performance relationships with regard to the temporal stability of these phenomena with the ultimate goal of improving how cohesion and coordination can be used to better the outcomes of teams. As most cohesion and coordination research utilizes less than three measurement points (Castano et al., 2013), the time scale of change in the constructs and their relationships is unknown. However, theory suggests that the early phases of team development are formative (Kozlowski et al., 1999). As such, we utilize a laboratory design to ensure we study teams from inception (Epstein, 1999) and capture the microdynamics of development (Dormann & Griffin, 2015) so that any fluctuations in the relationships can be identified. We first test a set of basic theoretically derived hypotheses related to the reciprocal relationships between cohesion and performance as well as coordination and performance summarized across 10 task episodes, primarily replicating prior work. Then, to better understand the dynamics of the relationships, we examine research questions related to whether the strength of the reciprocal relationships changes across episodes and, using general dominance analysis (Budescu, 1993), whether the relative importance of cohesion versus coordination for predicting performance changes across task episodes. Finally, we discuss the theoretical and applied implications of these findings for improving the functioning of teams and suggest avenues for future research.

Importance of Cohesion and Coordination for Team Performance

Cohesion

Team cohesion is a compositional emergent state (Marks et al., 2001) that is conceptualized as multidimensional (Festinger, 1950) and composed of social and task (action) components, as well as horizontal and vertical components (e.g., Beal et al., 2003). There are a multitude of cohesion definitions in the literature. Some representative examples include the following: “team members’ attraction and commitment to their team, team members, and the team’s task” (LePine et al., 2008, p. 290); “the resultant of all forces acting on the members to remain in the group” (Festinger, 1950, p. 274); “a dynamic process that is reflected in the tendency for a group to stick together and remain united in the pursuit of its goals and objectives” (Carron, 1982, p. 124); “a cognition about the group that exists in the minds of individual group members” (Carron & Spink, 1995, p. 91); and “the tendency for a group to stick together and remain united in the pursuit of its instrumental objectives” (Tekleab et al., 2009, p. 174). Based on these definitions, this study defines cohesion as “team members’ shared commitment or attraction to their task/goal and to one another.”

As one of the most commonly examined team states (LePine et al., 2008), the concept of cohesion is repeatedly identified as critical for understanding teamwork and team effectiveness. Not surprisingly, cohesion has been the focus of a plethora of studies, resulting in a number of important findings. For example, cohesion positively relates to team effectiveness outcomes, including satisfaction, performance, and viability at single (e.g., Beal et al., 2003; Evans & Dion, 1991; Mullen & Cooper, 1994) or multiple (typically two or three) time points (Castano et al., 2013; Mathieu et al., 2015; Tekleab et al., 2009).

Coordination

Team coordination is a team behavioral process (Marks et al., 2001) that represents efforts toward synchronizing behavior (Brannick et al., 1993; Fleishman & Zaccaro, 1992; Zalesny et al., 1995). A currently accepted definition of coordination that is adopted by the research team is “orchestrating the sequence and timing of independent actions” (Marks et al., 2001, p. 363). As teams are increasingly recognized as complex systems (Arrow et al., 2000), coordination is increasingly considered a vital process for effective performance (e.g., Summers et al., 2012). Teams that are better coordinated perform actions in a more efficient manner, allowing for greater levels of performance (e.g., Salas et al., 1992; Tannenbaum et al., 1992; Zalesny et al., 1995). Not surprisingly, meta-analytic evidence identifies coordination as a central mechanism by which teams combine their effort to achieve higher performance (e.g., LePine et al., 2008).

Reciprocal Relationships

Teams higher in cohesion have better intrateam relationships (Barrick et al., 1998; Tjosvold et al., 1986; Widmeyer et al., 1985), are more committed to being part of a team (Somech et al., 2009), emphasize team-level goals (Barrick et al., 1998; Carron et al., 1985), and share a more unified vision for how to complete the task (Campion et al., 1993; Watson & Michaelsen, 1988). As such, teams with higher cohesion should work together better and perform more effectively than teams lower in cohesion. As mentioned previously, this logically consistent relationship has been supported by multiple meta-analyses (e.g., Beal et al., 2003; Castano et al., 2013; Gully et al., 1995). Although most of this research is cross-sectional, there is no strong reason to believe that the directionality of the effects should change when examined at multiple time points assuming that the structure of task interdependence (Day et al., 2004; Steiner, 1972; Van De Ven et al., 1976) remains constant. Not surprisingly, when the cohesion–performance relationship is studied over two or three time points (e.g., Mathieu et al., 2015; Tekleab et al., 2009), the same positive relationship is revealed. As such, it is expected that teams exhibiting higher levels of cohesion during a given task episode will perform better during that episode.

H1: Within newly formed teams, those higher in cohesion will have higher levels of performance than teams lower in cohesion.

Teams with higher levels of coordination better synchronize their actions as they work interdependently (Brannick et al., 1993; Fleishman & Zaccaro, 1992; Zalesny et al., 1995) compared with teams with lower levels of coordination. As such, teams higher in coordination function more efficiently and are able to reach higher levels of performance, an effect supported meta-analytically (e.g., LePine et al., 2008). Like cohesion, there is no strong rationale for why this relationship should be altered when examining teams performing multiple task episodes of a given task. Therefore, it is expected that teams who are better coordinated during a given performance episode will perform better during that episode compared with teams worse at coordinating.

H2: Within newly formed teams, those higher in coordination will have higher levels of performance than teams lower in coordination.

Team cohesion can be influenced by a variety of factors. Of particular importance to this study is that feedback from prior performance is thought to influence subsequent levels of team cohesion (Mathieu et al., 2015). Teams deemed to be performing well likely receive an affective boost through recognition of their exemplary performance, increasing feelings of liking and unification among the team members (Casey-Campbell & Martens, 2009). Alternatively, teams performing poorly likely exhibit negative affective reactions to feedback, potentially resulting in internal conflict among members. Therefore, it is anticipated that teams higher in performance will receive a subsequent increase in team cohesion compared with teams lower in performance.

H3: Within newly formed teams, those with higher levels of performance during a given task episode will exhibit higher levels of cohesion during the subsequent episode compared with teams with lower levels of performance.

Like cohesion, coordination within teams is malleable. Feedback allows positive actions to be recognized and reinforced, whereas it emphasizes negative actions that should be avoided. Not surprisingly, training interventions that use feedback significantly improve the coordinated efforts of team members (e.g., Salas et al., 2008; Serfaty et al., 1998). Therefore, it is likely that teams performing well will have their positive actions reinforced, allowing them to build upon their success and improve their coordinated behavior.

H4: Within newly formed teams, those with higher levels of performance during a given task episode will exhibit higher levels of coordination during the subsequent episode compared with teams with lower levels of performance.

Combining H1 through H4 inherently indicates expected reciprocal relationships across task episodes between both cohesion and performance and coordination and performance (e.g., Mathieu et al., 2015). Within a given performance episode, it is expected that cohesion (H1) and coordination (H2) will predict performance. Performance from a given episode is then expected to predict subsequent cohesion (H3) and coordination (H4). Figure 1 summarizes the hypothesized relationships.

Figure 1.

Summary of hypothesized relationships.

Exploring Relationship Dynamics

Although it is expected that there is an overall, positive relationship between both cohesion and coordination with performance across task episodes, the strength of those relationships may vary depending on when the measurement occurs during the life cycle of the team. Relationships between predictors and performance have previously been found to change over time (e.g., Ackerman, 1989; Murphy, 1989), although theories of team development would suggest that the internal processes the team engages in differ depending on what phase of development the team is currently experiencing (e.g., Dierdorff et al., 2011; Kozlowski et al., 1999; Tuckman, 1965). As such, when examining the initial development of teams from their inception, it may be possible to observe temporal variability in the cohesion–performance or coordination–performance relationships. Unfortunately, limited theoretical rationale or empirical evidence exists to inform the directionality or magnitude of any potential dynamic changes. Therefore, we propose the following research question:

RQ1: Within newly formed teams, how do the cohesion–performance and coordination–performance relationships change across task episodes?

Similarly, the relative importance of cohesion and coordination for predicting performance may not be stable across task episodes. For action teams like those simulated in this study, coordination dynamics directly relate to the performance outcomes of interest (Marks et al., 2001). As such, coordination is likely a more important predictor of performance than cohesion. That said, as team capabilities develop (Kozlowski et al., 1999; McGrath, 1991; Tuckman, 1965), cohesion perceptions become more meaningful in the minds of team members (Carter et al., 2015). That is, it takes time for accurate and shared understanding of the team to form in the minds of the constituent team members. Therefore, cohesion likely becomes a more meaningful predictor with increased interactions, indicating that the relative importance of cohesion compared with coordination may increase as time goes on. However, like exploring the dynamics of the overall cohesion–performance and coordination–performance relationships, examining relative importance between cohesion and coordination and how that relative importance changes across task episodes is novel. Limited theory or empirical evidence exists to make a strong argument for exact directionality and magnitude of effects. As such, we propose the following research question:

RQ2: Within newly formed teams, how does the relative importance of cohesion and coordination for predicting performance change across task episodes?

Method

Participants

The sample comprised 42 three-person teams (n = 126) of undergraduate students participating for research credit at a large Midwestern university. The mean age for the sample was 19.71 years (SD = 1.76), the majority were female (69%), and their self-reported grade point average (GPA) was 3.1 out of 4.0 (SD = 0.53). As this study included 10 task episode points per team, the total sample size for testing the four hypotheses and the first research question was 420. For evaluating RQ2, the sample size was 42, the number of teams observed during each task episode.

Experimental Task

Participants engaged in a team-based, interactive resource foraging simulation designed by the research team in which each team member was tasked with finding and collecting unique resources. Participants were presented with a series of different scenarios that required them to strategize a plan to navigate the synthetic world, coordinate the collection of resources, and return to home base under time pressure, while still allowing for variance in the degree to which team members worked as a team. The task was ideal for accomplishing the goals of this study because it captures move-by-move and trial-by-trial objective behavior (process) and distinct performance data for individuals and teams, allowing team members’ actual behaviors to be assessed throughout each scenario. In addition, this task was chosen because of the high degree of interdependence required, which is critical for evaluating the cohesion–performance relationship (e.g., Beal et al., 2003). Team members sat at networked computer stations randomly distributed across the room, could not see one another, and communicated solely through headsets via Skype.

Scenario design and rules

Each team consisted of three members who each occupied a distinct role on the team. Information was distributed across the team such that some information was shared by all members, whereas other information was unique to a particular role (e.g., Stasser & Titus, 1985). More specifically, the entire team knew the location and type (i.e., penetrable or impenetrable) of geographic barriers in the world, but each member only had information relevant to the location and amount of their own role’s resources and, thus, their own individual performance. The team collectively had to share information and cooperate to plan an effective strategy to navigate barriers and collect as many resources as possible, while balancing individual and team goals. Each scenario presented teams with a unique challenge of comparable difficulty to complete their goals.

A cursor represented the team’s location in the grid. To move, team members had to agree on a direction, and then all team members had to coordinate their actions for the team to move. Specifically, members had to hit the corresponding direction arrow key (up, down, left, right). The team (cursor) would move in the direction chosen by the majority of the team members (i.e., at least two out of the three team members), and the team would not move until all three team members locked in a direction. Team members were not able to see which direction other members selected and thus had to communicate to move effectively. If a team approached an impenetrable barrier, the team had to navigate around it. If a team approached a penetrable barrier, they had the option of moving through the barrier, at the cost of slowing the team down. All moves were independent of one another meaning that the direction of prior moves had no bearing on subsequent actions the team took.

Once a team landed on a resource square, the team had to decide whether to collect the resource before moving on. If the team decided to collect the resource, each team member needed to press the spacebar to lock in that decision. Similar to the rules for moving, the resource would not be collected until all team members locked in. Each resource collected added one point to the team member’s score who owned that resource. The task was built such that the overall team score was computed as a simple composite score of each team member’s individual scores (Steiner, 1972). At all times, the team members could view the individual scores of all members as well as the total team score. For each scenario, participants were given the goal to maximize both their individual and the team’s scores.

Scenarios were designed to permit comparisons across task episodes, while still providing novelty in surface-level experiences and behavioral strategies (e.g., how to navigate the space). All scenario worlds were equivalent in size (20 × 20 grid), with the home base located in the bottom left grid square. Each scenario was created to ensure that all team members had the same number of individual resources (40) and total resources (120) available. The specific resource locations and densities (number of resources per location) varied across scenarios. In addition, the number, type, and location of barriers in the world varied across scenarios. Teams had 7 min to navigate each scenario.

Experimental procedure

Initial consent and all control variables were assessed via an online survey prior to the experiment. Participants then reported to the lab to participate in a 2.5-hr experiment session. Immediately upon arrival, participants were seated at an individual cubicle, randomly assigned to three-person teams (one or two teams were tested in a session), and informed consent was obtained. The experimenter then guided participants through a brief PowerPoint© training on operating the task. Prior to and during training, team members did not have the ability to communicate with one another. After the completion of training, all teams engaged in 10 7-min trials of the task separated by brief self-report assessments. Participants were then debriefed and dismissed.

Measures

Cognitive ability

Self-reported GPA was gathered as a proxy for general cognitive ability. GPA was used because academic learning and declarative knowledge is consistently highly related to cognitive ability (r values range from .3 to .7; Chen et al., 2000; Colquitt et al., 2000; Gully & Chen, 2010). GPA was averaged across members within a team to form a team-level cognitive ability proxy (Chan, 1998), which served as a Level 2 control variable for subsequent analyses.²

Cohesion

Self-reported cohesion was measured following each trial using a six-item measure adapted from Kozlowski et al. (2010) which has shown good reliability and convergent/discriminant validity evidence (Powers, 2012; Powers & Kozlowski, 2011). Participants were prompted to rate each item on a five-point scale (1 = strongly disagree to 5 = strongly agree) based on their experiences during the previous trial. Sample items included “Our team was unified in its task focus” (task cohesion) and “Our team members got along well with each other” (social cohesion). Coefficient alphas were computed for the scale at each of the 10 task episodes, with an average value of α = .94. Due to the strong correlation between task and social cohesion (r = .87, p < .001), a single scale score was computed to represent the overall cohesion felt by each member of the team. Self-reported cohesion scores were aggregated to the team level for each task episode (average r_WG = .96; James, 1982; James et al., 1993).

Coordination

As coordination represents a team behavioral process (Marks et al., 2001), it was measured using objective behaviors recorded by the task software. In particular, three team-level behaviors were assessed. The first was the percent of the total decisions (behaviors/actions) within a trial in which the team was in unanimous agreement (e.g., all pressed the same arrow key or spacebar), labeled team unity. The second was the average difference in team members’ scores (e.g., top scoring team member’s score − lowest scoring team member’s score) within a team on a given trial, labeled team priority. The final coordination behavior was the elapsed time in seconds between when the first person on a team makes a decision and when the last person on the team makes a decision for a given move (averaged across all moves in a trial), labeled team synchrony. Taken together, these behaviors capture the extent to which the team is unified in its actions, focused on team-level outcomes, and is temporally synchronized. The three indicators were on different scales so they were each standardized prior to any analyses being conducted. As behaviors were utilized to capture multiple elements of coordination, exploratory factor analysis³ (EFA) was used to determine the weighting of each behavior to create a composite coordination score. The results from the EFA are displayed in Table 1 and the final equation used to create the composite is as follows:

\begin{array}{l} {Coordination}_{i j} = 0.585 \times Team {Unity}_{i j} + 0.443 \times Team {Priority}_{i j} \\ - 0.314 \times Team {Synchrony}_{i j}, \end{array}

(1)

where i is the team and j is the time.

Table 1.

Rotated Factor Loadings for Cohesion and Coordination Indicators.

Indicator	Factor 1—cohesion	Factor 2—coordination
Cohesion—Item 1	0.892	0.230
Cohesion—Item 2	0.971	0.168
Cohesion—Item 3	0.979	0.118
Cohesion—Item 4	0.983	0.090
Cohesion—Item 5	0.975	0.105
Cohesion—Item 6	0.813	0.271
Coordination—team unity	−0.141	0.862
Coordination—team priority	0.275	0.733
Coordination—team synchrony	−0.331	−0.549

Note. Team synchrony was a reaction time variable and, thus, is scored in the opposite direction of the other indicators. Bolded values represent the indicators retained for each factor.

Team performance

Performance was assessed in this study using objective data collected directly from the task. It was operationalized as the total number of points (score; one resource gathered = one point) acquired during a trial. There was no unique team-level resource, rather each individual role acquired relevant resources and the team score (resources) equaled the sum of all individual scores.

Data Analysis

The data collected in this study were multilevel in nature. Team-level variables of cohesion, coordination, and performance all varied within teams across task episodes as well as between teams. As shown on the diagonal of Table 2, all time-varying constructs demonstrated a significant effect of clustering (i.e., ICC(1)) over time (Bliese, 2000), necessitating the use of random coefficient modeling (RCM; Raudenbush & Bryk, 2002; Singer & Willett, 2003; Snijders & Bosker, 1999). Prior to conducting longitudinal RCM analyses, it is recommended to check all criterion variables for the presence of time-dependent error terms (Braun et al., 2013; Kuljanin et al., 2011). As such, the augmented Dickey–Fuller test was administered to each team’s longitudinal data and it was determined that a majority of the teams did not exhibit time dependencies in the error term, indicating that longitudinal RCM was appropriate to use. All RCM analyses were conducted in the statistical software, R (R Core Team, 2017), using the lmer function with full maximum likelihood estimation in the lme4 package because it allows chi-square difference tests between models with different fixed- and random-effects structures (Finch et al., 2014).

Table 2.

Overall Team-Level Correlations Across Time Among Level 1 Time-Varying Variables.

Variable	M	SD	1	2	3	4
1. Cohesion	4.31	0.39	(.57)
2. Coordination	0.78	0.43	.19*	(.22)
3. Performance_{T − 1}	37.60	17.24	.26*	.41*	(.34)
4. Performance	38.30	17.10	.23*	.60*	.83*	(.34)

Note. The subscript T − 1 on performance indicates a lagged version of the variable. All other variables were related within the same time point.

p < .01.

Results

Hypothesis Testing

The four primary hypotheses in this study, depicted in Figure 1, examined whether cohesion (H1) and coordination (H2) significantly predicted performance within a given task episode and whether performance significantly predicted subsequent levels of cohesion (H3) and coordination (H4). Table 3 presents the nested models used to test the first two hypotheses. As observed in Model 7, both cohesion (b = 3.89, p < .05) and coordination (b = 12.69, p < .01) significantly predicted performance across task episodes, supporting H1 and H2. These results demonstrate that teams that felt more cohesive during a trial (i.e., performance episode) and were better at coordinating their behaviors during the trial performed better. Interestingly, the relationship between cohesion and performance was consistent across all teams because the insertion of cohesion as a random effect resulted in the model failing to reach convergence. Alternatively, the strength of the relationship between coordination and performance did vary depending on team membership (∆R² = .04), indicating the presence of potential moderators of this relationship.

Table 3.

Unstandardized RCM Results for Team Performance.

Parameters	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
Fixed effects
Intercept	38.30** (1.67)	20.16** (1.64)	−12.21 (13.37)	−26.51 (14.79)	−28.98 (14.72)	−24.33* (11.51)	−23.40* (10.98)	14.14 (8.43)
Level 1
Trial		3.30** (0.021)	3.30** (0.021)	3.26** (0.21)	3.25** (0.22)	2.66** (0.21)	2.50** (0.26)	2.80** (0.29)
Cohesion				3.65* (1.81)	3.70* (1.82)	4.30** (1.66)	3.89* (1.61)	6.99** (2.16)
Coordination						10.15** (1.31)	12.69** (1.87)	19.45** (1.60)
Trial × Cohesion								−0.98* (0.45)
Trial × Coordination								−3.09** (0.35)
Level 2
Cog. ability			10.53** (4.32)	10.13** (4.21)	10.68* (4.14)	6.98* (3.09)	7.05* (2.88)	4.27 (2.70)
Variance components
Level 2
Intercept	98.42	73.50	62.92	58.56	28.99	11.58	15.13	14.21
Trial		0.86	0.86	0.83	0.85	0.60	1.43	2.45
Cohesion					0.28
Coordination							61.80
Level 1
Residual	193.27	85.63	85.63	85.37	85.29	81.82	72.84	64.45
Fit indices
AIC	3,484.7	3,202.0	3,198.5	3,196.5		3,151.1	3,132.4	3,092.6
BIC	3,496.9	3,226.3	3,226.8	3,228.8		3,187.5	3,180.9	3,137.1
−2*log likelihood	3,478.8	3,190	3,184.6	3,180.4		3,133.2	3,108.4	3,070.6
Variance explained between		0.25	0.36	0.40		0.88	0.85	0.86
Variance explained within		0.56	0.56	0.56		0.58	0.62	0.67
Overall pseudo R²	.34	.71	.70	.70		.69	.73	.77
∆χ²		228.69**	5.53*	4.02*		47.35**	24.77**	62.51***
∆df		3	1	1		1	3	2

Note. Values represent unstandardized coefficients with standard errors in parentheses. Model 5 failed to properly converge and, thus, output is not interpreted. In Model 8, all Level 1 variables were centered prior to creating the interaction terms. Coordination was not modeled as a random effect in Model 8 due to failed convergence; as such, the χ² test uses Model 6 as a baseline. Covariance values among random effects for Models 5 and 7 were not tabled for parsimony as they were not interpreted. Overall pseudo R² was calculated using the r.squaredGLMM() function in the MuMIn package in R. RCM = random coefficient modeling; AIC = Akaike information criterion; BIC = Bayesian information criterion.

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

Table 4 presents the RCM results used to evaluate H3 and H4. Models 9 and 10 for cohesion demonstrate that performance significantly predicted subsequent levels of cohesion (b = 0.01, p < .01). That is, teams that performed better on a given trial reported feeling higher levels of cohesion on the subsequent trial, supporting H3. Once again, there was no evidence of between-team variability in the performance–cohesion relationship as the insertion of performance as a random effect failed to improve model fit (∆R² = .00). Performance did not significantly predict subsequent coordination, such that teams that performed better in a given trial were not necessarily more likely to coordinate better in the subsequent trial (b = 0.00, p > .05). However, the insertion of performance as a random effect did improve overall model fit (∆R² = .04), once again indicating between-team variability in the performance–coordination relationship. Therefore, H4 was not supported, but the results indicate the presence of potential moderators of this relationship. Figure 2 summarizes the results for the hypotheses.

Table 4.

Unstandardized RCM Results for Cohesion and Coordination.

Parameters	Cohesion			Coordination
Parameters	Model 9	Model 10	Model 11	Model 9	Model 10	Model 11
Fixed effects
Intercept	4.16** (0.06)	4.16** (0.06)	4.40** (0.06)	0.61** (0.07)	0.62** (0.08)	0.71** (0.06)
Level 1
Trial	−0.01 (0.01)	−0.01 (0.01)	−0.01 (0.01)	0.04** (0.01)	0.04** (0.01)	0.03** (0.01)
Performance_{T − 1}	0.01** (0.00)	0.01** (0.00)	0.01** (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)
Trial × Performance_{T − 1}			−0.001* (0.00)			−0.00 (0.00)
Variance components
Level 2
Intercept	0.11	0.12	0.11	0.13	0.19	0.08
Trial	0.00	0.00	0.00	0.00	0.00	0.00
Performance_{T − 1}		0.00			0.00	0.00
Level 1
Residual	0.05	0.05	0.05	0.09	0.09	0.09
Fit indices
AIC	128.9	134.4	126.4	281.8	274.9	276.6
BIC	156.4	173.6	157.8	309.3	314.2	319.9
−2*log likelihood	114.9	114.2	110.4	267.8	254.9	245.6
Variance explained within	0.29	0.29	0.29	0.40	0.40	0.40
Overall pseudo R²	.65	.65	.66	.41	.45	.45
∆χ²		0.65	4.53*		12.93**	0.25
∆df		3	1		3	1

Note. Values represent unstandardized coefficients with standard errors in parentheses. Due to only including Level 1 variables, only variance explained within was calculated. The baseline variance explained within was 0.07 for cohesion and 0.15 for coordination. Covariance values among random effects for Model 10 were not tabled for parsimony as they were not interpreted. RCM = random coefficient modeling; AIC = Akaike information criterion; BIC = Bayesian information criterion.

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

Figure 2.

Asymmetric reciprocal relationships between performance, cohesion, and coordination.

Exploring the Research Questions

In addition to testing the hypotheses, several analyses were conducted to answer the research questions that explore the temporal relationships between cohesion, coordination, and performance. Specifically, we tested whether the hypothesized relationships were stable across trials as well as examined the relative importance of cohesion and coordination for predicting performance and evaluated whether the relative importance findings were consistent across task episodes. To determine whether the longitudinal relationships were stable, an interaction term was created between the trial variable and each focal predictor, shown in Tables 3 (Model 8) and 4 (Model 11). Results demonstrate that the predictive utility of cohesion (b = −0.98, p < .05) and coordination (b = −3.09, p < .01) for explaining variance in team performance decreases over time (Table 3, Model 8). This suggests that, early in a team’s life cycle, performance is more heavily influenced by the coordinated behaviors and cohesive feelings of team members. However, as the team develops (e.g., Kozlowski et al., 1999; McGrath, 1991; Tuckman, 1965), other factors play a larger role in influencing team performance. Not surprisingly, the reciprocal relationship between performance on subsequent cohesion followed a similar pattern. Model 11 for cohesion (Table 4) shows that, over time, performance is less impactful for predicting subsequent cohesion (b = −0.001, p < .05), indicating an overall weakening of the cohesion–performance relationship throughout the team life cycle. Given that performance did not significantly predict subsequent coordination, it is not surprising that this relationship did not significantly change across trials (b = −0.00, p > .05).

In addition to exploring the overall longitudinal relationships between cohesion, coordination, and performance across task episodes, the relationships were tested in each trial separately to better understand the relative impact of cohesion and coordination on performance. Overall, the predictors combined to explain a significant amount of variance in performance, on average explaining 42% of the variance (average overall R² = .42, p < .01). The relative impact (i.e., importance) of each predictor was calculated using general dominance analysis (Braun et al., 2019; Budescu, 1993). General dominance weights represent the average unique variance explained across all possible submodels including a given predictor. In the case with only two predictors, each predictor contributes unique variance explained in a model including only a single predictor and a model with both predictors. In the model with two predictors, unique variance explained is calculated using the change in R² between a model with and without the focal predictor. Table 5 presents the standardized general dominance weights (by dividing each weight by the model R² so the weights sum to 1) of cohesion and coordination for team performance, indicating their relative importance.

Table 5.

Standardized General Dominance Weights for Predicting Team Performance.

Trial	Cohesion	Coordination	Overall R²
1	0.09	0.91	.76
2	0.01	0.99	.67
3	0.11	0.89	.57
4	0.11	0.89	.47
5	0.10	0.90	.41
6	0.25	0.75	.45
7	0.18	0.82	.05
8	0.11	0.89	.32
9	0.01	0.99	.38
10	0.31	0.69	.12
M	0.13	0.87	.42

Note. All dominance weights were standardized by dividing them by the final model R² to show relative importance.

Two general patterns emerged. First, coordination was a more important predictor of performance in all cases, typically explaining much more unique variance than cohesion. The second pattern was that the relative importance of cohesion for predicting team performance increased across episodes (correlation with trial is .44), whereas the relative impact of coordination decreased across episodes. This demonstrates that cohesion becomes more important above and beyond coordination for predicting performance as the team works together longer.

It can be seen from Table 5 that although there is certainly variability in the relative predictive utility of cohesion, during the first half of the trials, cohesion explained only about 8% of the unique variance in performance. Alternatively, in the latter half of the trials, the average relative incremental prediction of cohesion more than doubled and explained a little more than 17% of the variance explained in performance. Consistent with the argument by Carter et al. (2015) that emergent states take time to form and become meaningful in the minds of team members, early cohesion ratings provided less relative predictive utility over and above coordination. However, over time, members’ shared perceptions of team cohesion became relatively more predictive of team performance.

It is important to clarify the distinction between the two sets of exploratory analyses because the trends may seem contradictory. The initial set of analyses (Table 3, Model 8; Table 4, Model 11) demonstrates that the absolute amount of variance explained in team performance by cohesion and coordination decreased across trials. That is, both cohesion and coordination were weaker predictors of performance during later task episodes compared with earlier episodes. The second set of exploratory analyses (Table 5) demonstrates that the relative importance of cohesion compared with coordination increases across trials. That is, cohesion adds more predictive utility above and beyond coordination throughout the life of the team, regardless of the absolute amount of variance explained in team performance.

Discussion

The primary goal of this study was to evaluate implicit temporal assumptions to better understand the evolution of team cohesion–performance and coordination–performance relationships within newly formed teams. Not surprisingly, results demonstrated that both cohesion and coordination uniquely predicted team performance consistently across multiple task episodes. Performance then predicted subsequent levels of cohesion, indicating the presence of a reciprocal relationship in the form of a positive feedback loop. Higher cohesion resulted in higher levels of performance, which resulted in subsequent higher cohesion. These results are consistent with the recent findings by Mathieu et al. (2015) and highlight the important role that performance feedback has in shaping the cohesion perceptions of team members (e.g., Casey-Campbell & Martens, 2009). Alternatively, performance did not have the expected positive relationship with subsequent coordination. It could be that because performance feedback is generally helpful and informative regardless of the level of performance (e.g., Bell & Kozlowski, 2002), coordination simply improved across trials as a result of feedback for most, if not all, teams. The significant fixed effect of trial on coordination (b = 0.04, p < .01) supports this by indicating improved coordination with additional experience and feedback opportunities.

Interestingly, these results demonstrate that both cohesion and coordination are important for team performance outcomes almost immediately upon inception. Although this might be expected for coordination because it encompasses actual task-relevant behavior, it is more surprising for cohesion. Despite emergence needing time and interactions to occur (Kozlowski & Klein, 2000) and cohesion becoming more stable and meaningful in the minds of individuals over time (Carter et al., 2015), cohesion was a significant predictor across all of the task episodes and explained 5% to 7% of the total variance in performance throughout the first hour of the team interacting. This suggests that fostering cohesion immediately is important, placing an emphasis on building cohesive teams through selection (Acton et al., 2019) and reinforcing it through leadership. More broadly, these results demonstrate the general importance of both team emergent states and behavioral processes for influencing team outcomes in newly formed teams.

A significant amount of variability existed in both the coordination–performance and performance–subsequent coordination relationships (demonstrated by improved model fit with random effects). As such, there are likely moderators of both relationships indicating that some teams demonstrate the hypothesized reciprocal relationship, whereas others do not. For example, it could be that teams with higher motivation would see a stronger reciprocal relationship between coordination and performance compared with teams lower in motivation. Given the well-known importance of motivation for performance (e.g., Campbell et al., 1993), well-coordinated teams with high levels of motivation should conduct their actions with urgency and frequently capitalize on being temporally synchronized and efficient, achieving higher levels of performance. Likewise, higher levels of motivation would make teams more likely to attend to and incorporate lessons learned from feedback, allowing for improvements in subsequent coordination. Alternatively, even well-coordinated teams without motivation are likely to perform poorly due to low effort and teams lacking the motivation to attend to or internalize information from feedback would not see any potential benefits. It is important for future research to explore processes, such as motivation, that could moderate the strength of the relationship between coordination and performance to improve training interventions and other attempts to bolster performance through improved coordination.

In addition to simply examining the reciprocal relationships between cohesion and coordination with performance, we also wanted to evaluate temporal assumptions through a series of exploratory analyses. We first tested whether the reciprocal relationships were stable across task episodes. Results indicated that all significant relationships weakened as teams worked together. That is, the degree to which cohesion and coordination predicted performance lessened across the life cycle of the teams. Likewise, the degree to which performance influenced subsequent cohesion also weakened. These findings indicate that, initially, cohesion and coordination play a critical role in how the team performs; however, as teams develop and work together, performance becomes less dependent on these constructs. It could be that, through repeatedly working together, teams’ level of cohesion and coordination eventually reach a steady state or equilibrium (von Bertalanffy, 1950a, 1950b), meaning that the relative rank ordering of teams becomes somewhat constant on these constructs. As such, variability in performance would not be explainable by these steady variables.

This weakening of the cohesion–performance and coordination–performance relationships during the life of the team is important for numerous reasons. First, these results present an alternative explanation to those currently explored through meta-analysis regarding variable effect sizes in the cohesion–performance and coordination–performance relationships. When measurement occurs during the life or task cycle of the team could be having a significant impact on the relative magnitude of these relationships. Therefore, it is critical to consider when measurement should/does occur in studies and report on this information so that current inferences are properly interpreted and qualified and future meta-analytic efforts can adequately evaluate this potential explanatory mechanism. Second, it appears that cohesion and coordination are most critical to focus on early during the developmental process of teams. This is consistent with many theories of team development (e.g., Kozlowski et al., 1999; McGrath, 1991; Tuckman, 1965), which highlight the relative volatility of team members’ roles and perceptions during early phases. Interventions aimed at improving coordination dynamics or cohesion levels of teams are likely less effective and less important at later phases in the teams’ cycle once equilibrium is reached. Finally, these results demonstrate the need to identify additional critical predictors of performance for different phases of team development. Current theory and empirical evidence regarding the proximal predictors of team performance, emergent states, and team processes (Marks et al., 2001) implicitly assume that the relevant constructs are stable across time. To the extent to which this is not the case necessitates a better theory of time (Mitchell & James, 2001) regarding which states and processes are most relevant and important during each phase of the task or life cycle of the team (e.g., Acton et al., 2019).

The final important contribution of this article is examining the relative importance (Budescu, 1993) of cohesion and coordination for predicting performance across multiple trials. Although, overall, cohesion and coordination decreased in their predictive utility of performance, mirrored in the overall R² estimates from within task episode models (Table 5), the relative importance of these predictors demonstrated unique and interesting trajectories. First, coordination was always more important than cohesion across all task episodes, indicating a consistently stronger association with performance. Perhaps, this is not surprising given that actual task behavior was used as a measure for coordination. Teams exhibiting higher levels of coordination were actually engaging in more efficient task completion behaviors so it is only logical that coordination would be a stronger predictor of performance than team member affective reactions measured via self-report survey at the conclusion of the trial. That said, as the team worked together, cohesion became relatively more important, whereas coordination became relatively less important. This finding is consistent with the argument by Carter et al. (2015) that affective states take time to develop in the mind of participants and become meaningful variables. These findings imply that, if possible, the best means of understanding and ultimately improving team performance is by leveraging coordination behaviors. Changing the actual actions of team members will likely result in dramatic changes in performance outcomes. That said, it is important to consider and foster cohesion for teams that will be working together for longer periods of time because as teams work together and cohesion emerges it becomes a relatively more important and useful predictor of performance.

Limitations and Future Directions

Although the results of this study are interesting and promising, like all studies it has some limitations. The participants in this study performed over a relatively short time frame in a lab setting. This design has definite benefits for understanding the development of temporal processes in which the time horizon for meaningful change is unknown (Dormann & Griffin, 2015). Similarly, for swift action teams, which are becoming increasingly critical to many organizations (Wildman et al., 2012), these processes likely closely match those experienced in organizations. Complexity theory (Anderson, 1999) and multilevel theory (Kozlowski & Klein, 2000) both suggest that early interactions are foundational to the formation of perceptions and routines and these initial impressions are likely to have long-lasting effects (e.g., Kenny, 1994). However, to ensure that these findings generalize to teams in organizations that operate over longer time horizons, further research needs to be conducted utilizing event sampling and behavioral observations of field samples, ideally with newly formed teams. In addition, the task that the participants in this study performed was highly interdependent (Van De Ven et al., 1976), which accentuates the effects of cohesion (Beal et al., 2003). It is important to evaluate these findings under various task conditions to determine any boundary conditions of the results.

Despite these limitations, the findings from this study encourage additional work in a number of important areas. First, the identification of within-study variability in the coordination–performance temporal reciprocal relationship demonstrates the need for additional theory and empirical investigations to identify moderators or boundary conditions of this relationship. Similarly, the difference in the cohesion–performance and coordination–performance relationships over time highlights the importance of creating and testing theory regarding the intersection between team development and commonly applied input–process–output (i.e., input–process–output [IPO; see McGrath, 1964] and input–mediator–output–input [IMOI; see Ilgen et al., 2005]) models of team performance. This study implies that critical inputs or processes (mediators) may differ depending on the developmental phase of the team so it is important to evaluate how our commonly studied relationships work at different points in time. The relative importance analyses conducted demonstrate the need to understand how different mechanisms uniquely contribute to team success and how their relative importance may change over time. Theory and testing surrounding what processes to target for improving team performance is critical as the organizational sciences continue to strive for bettering the functioning and outcomes of teams. Finally, this study focused on team cohesion and team coordination as predictors due to their well-established relationships with team performance. However, there are many additional emergent states (e.g., efficacy, conflict) and processes (e.g., communication) that are relevant to team performance outcomes. As such, it is important that future studies examine these additional states and processes within newly formed teams to understand how their relationships with performance evolve during the life or task cycle of the team.

Conclusion

Emergent states such as cohesion and behavioral team processes such as coordination are considered proximal predictors of team performance (Marks et al., 2001) and, thus, represent critical leverage points for improving the success of teams. Despite their importance, little is known regarding the temporal cohesion–performance and coordination–performance relationships, especially in newly formed teams. Cohesion and performance demonstrate a positive reciprocal relationship across task episodes, whereas coordination significantly affects performance but does not demonstrate a reciprocal relationship. In addition, across trials, the strength of these relationships varies and the relative importance of cohesion and coordination changes. Therefore, it is important that organizational scientists develop and test temporally specified theory regarding the importance of team phenomena for predicting team outcomes so that targeted interventions may be created to increase performance.

Footnotes

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors gratefully acknowledge the National Aeronautics and Space Administration (NASA; NNX09AK47G, S.W.J.K., principal investigator; R.P.D., coinvestigator) for support that, in part, assisted the composition of this manuscript. Any opinions, findings, and conclusions or recommendations expressed are those of the authors and do not necessarily reflect the views of NASA.

ORCID iD

Michael T. Braun

Notes

Author Biographies

Michael T. Braun received his PhD in organizational psychology from Michigan State University and is currently an assistant professor of Organizational Behavior and Human Resource Management at DePaul University (USA). His expertise is in the areas of team knowledge emergence and decision making, emergent leadership, team cohesion, and modeling multilevel dynamics.

Steve W. J. Kozlowski, PhD, is a professor of Industrial Organizational Psychology at the University of South Florida (USA). He is a recognized authority in the areas of multilevel theory; team leadership and team effectiveness; and learning, training, and adaptation.

Tara A. (Rench) Brown is a senior scientist and the instructional strategy and support capability lead at APTIMA (USA). She has expertise in unobtrusive measurement, learning, and training. She received her PhD and MA in industrial-organizational psychology from Michigan State University and her BS in psychology from Wright State University.

Richard P. DeShon earned his PhD from the University of Akron and is currently a professor of Organizational Psychology at Michigan State University (USA). His research addresses individual and collective performance, human motivation, dynamics, measurement theory, and machine learning.

References

Ackerman

P. L.

(1989). Within-task intercorrelations of skilled performance: Implications for predicting individual differences? A comment on Henry & Hulin, 1987. Journal of Applied Psychology, 74, 360–364. https://doi.org/10.1037/0021-9010.74.2.360

Acton

B. P.

Braun

M. T.

Foti

R. J.

(2019). Built for unity: Assessing the impact of team composition on team cohesion trajectories. Journal of Business and Psychology, 1–16. https://doi.org/10.1007/s10869-019-09654-7

Anderson

(1999). Perspective: Complexity theory and organization science. Organization Science, 10, 216–232. https://doi.org/10.1287/orsc.10.3.216

Arrow

McGrath

J. E.

Berdahl

J. L.

(2000). Small groups as complex systems. SAGE.

Barrick

M. R.

Stewart

G. L.

Neubert

M. J.

Mount

M. K.

(1998). Relating member ability and personality to work-team processes and team effectiveness. Journal of Applied Psychology, 83, 377–391. https://doi.org/10.1037/0021-9010.83.3.377

Beal

D. J.

Cohen

R. R.

Burke

M. J.

McLendon

C. L.

(2003). Cohesion and performance in groups: A meta-analytic clarification of construct relations. Journal of Applied Psychology, 88, 989–1004. https://doi.org/10.1037/0021-9010.88.6.989

Bell

B. S.

Kozlowski

S. W. J.

(2002). Adaptive guidance: Enhancing self-regulation, knowledge, and performance in technology-based training. Personnel Psychology, 55, 267–306. https://doi.org/10.1111/j.1744-6570.2002.tb00111.x

Bliese

P. D.

(2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and analysis. In Klein

K. J.

Kozlowski

S. W. J.

(Eds.), Multilevel theory, research, and methods in organizations (pp. 349–381). Jossey-Bass.

Block

(1995). A contrarian view of the fie-factor approach to personality description. Psychological Bulletin, 117, 187–215. https://doi.org/10.1037/0033-2909.117.2.187

10.

Bowers

Salas

Prince

Brannick

(1992). Games team play: A method for investigating team coordination and performance. Behavior Research Methods, Instruments, & Computers, 24, 503–506. https://doi.org/10.3758/BF03203594

11.

Brannick

M. T.

Roach

R. M.

Salas

(1993). Understanding team performance: A multimethod study. Human Performance, 6, 287–308. https://doi.org/10.1207/s15327043hup0604_1

12.

Braun

M. T.

Converse

P. D.

Oswald

F. L.

(2019). Estimating the accuracy of dominance analysis as a metric of relative importance: The joint impact of sampling error variance and unreliability. Journal of Applied Psychology, 104, 593–602. https://doi.org/10.1037/apl0000361

13.

Braun

M. T.

Kuljanin

DeShon

R. P.

(2013). Spurious relationships in the analysis of longitudinal data in organizational research. Organizational Research Methods, 16, 302–330. https://doi.org/10.1177/1094428112469668

14.

Brown

T. A.

(2006). Confirmatory factor analysis for applied research. Guilford Press.

15.

Budescu

D. V.

(1993). Dominance analysis: A new approach to problem of relative importance of predictors in multiple regression. Psychological Bulletin, 114, 542–551. https://doi.org/10.1037/0033-2909.114.3.542

16.

Campbell

J. P.

McCloy

R. A.

Oppler

S. H.

Sager

C. E.

(1993). A theory of performance. In Schmitt

Borman

(Eds.), Personnel selection in organizations (pp. 35–70). Jossey-Bass.

17.

Campion

M. A.

Medsker

G. J.

Higgs

A. C.

(1993). Relations between work group characteristics and effectiveness: Implications for designing effective work groups. Personnel Psychology, 46, 823–850. https://doi.org/10.1111/j.1744-6570.1993.tb01571.x

18.

Carron

A. V.

(1982). Cohesiveness in sport groups: Interpretation and considerations. Journal of Sport Psychology, 4, 123–138. https://doi.org/10.1123/jsp.4.2.123

19.

Carron

A. V.

Spink

K. S.

(1995). The group size-cohesion relationship in minimal groups. Small Group Research, 26, 86–105. https://doi.org/10.1177/1046496495261005

20.

Carron

A. V.

Widmeyer

W. N.

Brawley

L. R.

(1985). The development of an instrument to assess cohesion in sport teams: The Group Environment Questionnaire. Journal of Sport Psychology, 7, 244–266. https://doi.org/10.1123/jsp.7.3.244

21.

Carter

N. T.

Carter

D. R.

DeChurch

L. A.

(2015). Implications of observability for the theory and measurement of emergent team phenomena. Journal of Management, 44, 1398–1425. https://doi.org/10.1177/0149206315609402

22.

Casey-Campbell

Martens

M. L.

(2009). Sticking it all together: A critical assessment of the group cohesion–performance literature. International Journal of Management Reviews, 11, 223–246. https://doi.org/10.1111/j.1468-2370.2008.00239.x

23.

Castano

Watts

Tekleab

A. G.

(2013). A reexamination of the cohesion-performance relationship meta-analyses: A comprehensive approach. Group Dynamics: Theory, Research, and Practice, 17, 207–231. https://doi.org/10.1037/a0034142

24.

Chan

(1998). Functional relations among constructs in the same content domain at different levels of analysis: A typology of composition models. Journal of Applied Psychology, 83, 234–246. https://doi.org/10.1037/0021-9010.83.2.234

25.

Chen

Gully

S. M.

Whiteman

J. A.

Kilcullen

B. N.

(2000). Examination of relationships among trait-like individual differences, state-like individual differences, and learning performance. Journal of Applied Psychology, 85, 835–847. https://doi.org/10.1037/0021-9010.85.6.835

26.

Colquitt

J. A.

LePine

J. A.

Noe

R. A.

(2000). Toward an integrative theory of training motivation. Journal of Applied Psychology, 85, 678–707. https://doi.org/10.1037/0021-9010.85.5.678

27.

Cronin

M. A.

Weingart

L. R.

Todorova

(2011). Dynamics in groups: Are we there yet? The Academy of Management Annals, 5, 571–612. https://doi.org/10.1080/19416520.2011.590297

28.

Day

E. A.

Arthur

W. J.

Miyashiro

Edwards

B. D.

Tubré

T. C.

Tubré

A. H.

(2004). Criterion-related validity of statistical operationalizations of group general cognitive ability as a function of task type: Comparing the mean, maximum, and minimum. Journal of Applied Social Psychology, 34(7), 1521–1549. https://doi.org/10.1111/j.1559-1816.2004.tb02017.x

29.

Dierdorff

E. C.

Bell

S. T.

Belohlav

J. A.

(2011). The power of “we”: Effects of psychological collectivism on team performance over time. Journal of Applied Psychology, 96, 247–262. https://doi.org/10.1037/a0020929

30.

Dormann

Griffin

M. A.

(2015). Optimal time lags in panel studies. Psychological Methods, 20(4), 489–505. https://doi.org/10.1037/met0000041

31.

Epstein

J. M.

(1999). Agent-based computational models and generative social science. Complexity, 4, 41–60. https://doi.org/10.1002/(SICI)1099-0526(199905/06)4:5<41::AID-CPLX9>3.0.CO;2-F

32.

Evans

C. R.

Dion

K. L.

(1991). Group cohesion and performance: A meta-analysis. Small Group Research, 22, 175–186. https://doi.org/10.1177/1046496491222002

33.

Festinger

(1950). Informal social communication. Psychological Review, 57, 271–282.

34.

Finch

W. H.

Bolin

J. E.

Kelley

(2014). Multilevel modeling using R. CRC Press.

35.

Fleishman

E. A.

Zaccaro

S. J.

(1992). Toward a taxonomy of team performance functions. In Swezey

R. W.

Salas

(Eds.), Teams: Their training and performance (pp. 31–56). Ablex.

36.

Gully

S. M.

Chen

(2010). Individual differences, attribute-treatment interactions, and training outcomes. In Kozlowski

S. W. J.

Salas

(Eds.), SIOP organizational frontiers series. Learning, training, and development in organizations (pp. 3–64). Jossey-Bass.

37.

Gully

S. M.

Devine

D. J.

Whitney

D. J.

(1995). A meta-analysis of cohesion and performance: Effects of level of analysis and task interdependence. Small Group Research, 26, 497–520. https://doi.org/10.1177/1046496495264003

38.

Ilgen

D. R.

Hollenbeck

J. R.

Johnson

Jundt

(2005). Teams in organizations: From input-process-output models to IMOI models. Annual Review of Psychology, 56, 517–543. https://doi.org/10.1146/annurev.psych.56.091103.070250

39.

James

L. R.

(1982). Aggregation bias in the estimates of perceptual agreement. Journal of Applied Psychology, 67, 219–229. https://doi.org/10.1037/0021-9010.67.2.219

40.

James

L. R.

Demaree

R. G.

Wolf

(1993). r_wg: An assessment of within-group interrater agreement. Journal of Applied Psychology, 78, 306–309. https://doi.org/10.1037/0021-9010.78.2.306

41.

Kenny

D. A.

(1994). Interpersonal perception: A social relations analysis. Guilford Press.

42.

Kozlowski

S. W. J.

(2015). Advancing research on team process dynamics: Theoretical, methodological, and measurement considerations. Organizational Psychology Review, 5, 270–299. https://doi.org/10.1177/2041386614533586

43.

Kozlowski

S. W. J.

Chao

G. T.

(2012). The dynamics of emergence: Cognition and cohesion in work teams. Managerial and Decision Economics, 33, 335–354. https://doi.org/10.1002/mde.2552

44.

Kozlowski

S. W. J.

Gully

S. M.

Nason

E. R.

Smith

E. M.

(1999). Developing adaptive teams: A theory of compilation and performance across levels and time. In Ilgen

D. R.

Pulakos

E. D.

(Eds.), The changing nature of work performance: Implications (pp. 240–292). Jossey-Bass.

45.

Kozlowski

S. W. J.

Ilgen

D. R.

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

46.

Kozlowski

S. W. J.

Klein

K. J.

(2000). A multilevel approach to theory and research in organizations: Contextual, temporal, and emergent processes. In Klein

K. J.

Kozlowski

S. W. J.

(Eds.), Multilevel theory, research and methods in organizations: Foundations, extensions, and new directions (pp. 3–90). Jossey-Bass.

47.

Kozlowski

S. W. J.

Ployhart

R. P.

Lim

B. C

. (2010, April). The developmental role of team leaders. In Resick

Doty

(Chairs), Current perspectives on leadership in collective work arrangements [Symposium]. 25th Annual Conference of the Society for Industrial and Organizational Psychology, Atlanta, GA, United States.

48.

Kuljanin

Braun

M. T.

DeShon

R. P.

(2011). A cautionary note on applying growth models to longitudinal data. Psychological Methods, 16, 249–264. https://doi.org/10.1037/a0023348

49.

LePine

J. A.

Piccolo

R. F.

Jackson

C. L.

Mathieu

J. E.

Saul

J. R.

(2008). A meta-analysis of teamwork processes: Tests of a multidimensional model and relationships with team effectiveness criteria. Personnel Psychology, 61, 273–307. https://doi.org/10.1111/j.1744-6570.2008.00114.x

50.

Marks

M. A.

Mathieu

J. E.

Zaccaro

S. J.

(2001). A temporally based framework and taxonomy of team processes. Academy of Management Review, 26, 356–376. https://doi.org/10.2307/259182

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. https://doi.org/10.1177/0149206308316061

52.

Mathieu

J. E.

Kukenberger

M. R.

D’Innocenzo

Reilly

(2015). Modeling reciprocal team cohesion–performance relationships, as impacted by shared leadership and members’ competence. Journal of Applied Psychology, 100, 713–734. https://doi.org/10.1037/a0038898

53.

McGrath

J. E.

(1964). Social psychology: A brief introduction. Holt, Rinehart and Winston.

54.

McGrath

J. E.

(1991). Time, interaction, and performance (TIP): A theory of groups. Small Group Research, 22, 147–174. https://doi.org/10.1177/1046496491222001

55.

Mitchell

T. R.

James

L. R.

(2001). Building better theory: Time and the specification of when things happen. Academy of Management Review, 26, 530–547. https://doi.org/10.2307/3560240

56.

Molenaar

P. C. M.

(2004). A manifesto on psychology as idiographic science: Bringing the person back into scientific psychology, this time forever. Measurement, 2, 201–218. https://doi.org/10.1207/s15366359mea0204_1

57.

Mullen

Cooper

(1994). The relation between group cohesiveness and performance: An integration. Psychological Bulletin, 115, 210–227. https://doi.org/10.1037/0033-2909.115.2.210

58.

Murphy

K. R.

(1989). Is the relationship between cognitive ability and job performance stable over time? Human Performance, 2, 183–200. https://doi.org/10.1207/s15327043hup0203_3

59.

Powers

C. L.

(2012, April). What does cohesion capture? An empirical and conceptual analysis [Poster presentation]. 27th Annual Conference of the Society for Industrial and Organizational Psychology, San Diego, CA, United States.

60.

Powers

C. L.

Kozlowski

S. W. J.

(2011, July). Team cohesion and conflict: Where’s the connection? Factor analysis and nomological network for critical team processes [Paper presentation]. 6th Annual INGRoup Conference, Minneapolis, MN, United States.

61.

Raudenbush

S. W.

Bryk

A. S.

(2002). Hierarchical linear models: Applications and data analysis methods. SAGE.

62.

R Core Team. (2017). R: A language and environment for statistical computing. R Foundation for Statistical Computing.

63.

Salas

DiazGranados

Klein

Burke

C. S.

Stagl

K. C.

Goodwin

G. F.

Halpin

S. M.

(2008). Does team training improve team performance? A meta-analysis. Human Factors, 50, 903–933. https://doi.org/10.1518/001872008X375009

64.

Salas

Dickinson

T. L.

Converse

S. A.

Tannenbaum

S. I.

(1992). Toward an understanding of team performance and training. In Swezey

R. W.

Salas

(Eds.), Teams: Their training and performance (pp. 3–29). Ablex.

65.

Salas

Stagl

K. C.

Burke

C. S.

(2004). 25 years of team effectiveness in organizations: Research themes and emerging needs. In Cooper

C. L.

Robertson

I. T.

(Eds.), International review of industrial and organizational psychology (Vol. 19, pp. 47–91). John Wiley.

66.

Serfaty

Entin

E. E.

Johnston

J. H.

(1998). Team coordination training. In Cannon-Bowers

J. A.

Salas

(Eds.), Making decisions under stress: Implications for individual and team training (pp. 221–245). American Psychological Association.

67.

Singer

J. D.

Willett

J. B.

(2003). Applied longitudinal data analysis modeling change and event occurrence. Oxford University Press.

68.

Snijders

Bosker

. (1999). Multilevel analysis. an introduction to basic and advanced multilevel modeling. Thousand Oaks, CA: SAGE.

69.

Somech

Desivilya

H. S.

Lidogoster

(2009). Team conflict management and team effectiveness: The effects of task interdependence and team identification. Journal of Organizational Behavior, 30, 359–378. https://doi.org/10.1002/job.537

70.

Stasser

Titus

(1985). Pooling of unshared information in group decision making: Biased information sampling during discussion. Journal of Personality and Social Psychology, 48, 1467–1478. https://doi.org/10.1037/0022-3514.48.6.1467

71.

Steiner

I. D.

(1972). Group process and productivity. Academic Press.

72.

Summers

J. K.

Humphrey

S. E.

Ferris

G. R.

(2012). Team member chance, flux in coordination, and performance: Effects of strategic core roles, information transfer, and cognitive ability. Academy of Management Journal, 55, 314–338. https://dx-doi-org.web.bisu.edu.cn/10.5465/amj.2010.0175

73.

Tannenbaum

S. C.

Beard

R. L.

Salas

(1992). Team building and its influence on team effectiveness: An examination of conceptual and empirical developments. In Kelley

(Ed.), Issues, theory, and research in industrial/organizational psychology (pp. 117–153). Elsevier.

74.

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. https://doi.org/10.1177/1059601108331218

75.

Tjosvold

Wedley

W. C.

Field

R. H. G.

(1986). Constructive controversy, the Vroom Yetton model, and managerial decision making. Journal of Occupational Behavior, 7, 125–138. https://doi.org/10.1002/job.4030070205

76.

Tuckman

B. W.

(1965). Developmental sequence in small groups. Psychological Bulletin, 63, 384–389. https://dx-doi-org.web.bisu.edu.cn/10.1037/h0022100

77.

Vancouver

J. B.

Thompson

C. M.

Williams

A. A.

(2001). The changing signs in the relationships among self-efficacy, personal goals, and performance. Journal of Applied Psychology, 86, 605–620. https://doi.org/10.1037/0021-9010.86.4.605

78.

Van De Ven

A. H.

Delbecq

A. L.

Koenig

(1976). Determinants of coordination modes within organizations. American Sociological Review, 41, 322–338. https://doi.org/10.2307/2094477

79.

von Bertalanffy

. (1950a). An outline of general systems theory. The British Journal for the Philosophy of Science, 1, 134–165. https://doi.org/10.1093/bjps/I.2.134

80.

von Bertalanffy

. (1950b). The theory of open systems in physics and biology. Science, 111, 23–29. https://doi.org/10.1126/science.111.2872.23

81.

Watson

W. E.

Michaelsen

L. K.

(1988). Group interaction behaviors that affect performance on an intellective task. Group and Organization Studies, 13, 495–516. https://doi.org/10.1177/105960118801300406

82.

Widmeyer

W. N.

Brawley

L. R.

Carron

A. V.

(1985). The measurement of cohesion in sport teams: The Group Environment Questionnaire. Sports Dynamics.

83.

Wildman

J. L.

Shuffler

M. L.

Lazzara

E. H.

Fiore

S. M.

Burke

C. S.

Salas

Garven

(2012). Trust development in swift starting action teams: A multilevel framework. Group & Organization Management, 37(2), 137–170. https://doi.org/10.1177/1059601111434202

84.

Zalesny

M. D.

Salas

Prince

(1995). Conceptual and measurement issues in coordination: Implications for team behavior and performance. Research in Personnel and Human Resources Management, 13, 81–115. https://doi.org/10.1037/1089-2699.7.3.179