The Evolution and Maturation of Team in Organizations

Meinecke et al. State Space Grids and Teams across different SSGs is particularly useful if their underlying Team interactions can be chaotic and teamwork may move dimensions are the same. in spurts rather than flow evenly toward team outcomes. This is particularly true for teamwork in the face of trends Entropy is a measure of predictability and describes the toward increasing team fluidity and temporary organizing (i.e., level of organization of the system. In GridWare entropy can quick changes in team composition), distributed teamwork be calculated based on cell visits (i.e., visited entropy), cell (i.e., members collaborating from a distance and interacting transitions (i.e., transitional entropy), and duration (i.e., duration and coordinating their actions in intervals), and multiple team entropy). To clarify, consider the following sequence of coded memberships (i.e., employees finding themselves in different behavior ABABABAB with A and B being discrete codes, such roles across different teams). In light of such developments, as a joint occurrence of idea expression and low positivity. teams are discussed as “dynamic hubs of participants” rather This particular sequence is much easier to recreate than the than clearly bounded structures (Mortensen and Haas, 2018). We following sequence, ACBFDAAB, which seems rather random. expect that the interactions that ensue in these dynamic hubs are For computing entropy, a conditional probability is calculated even less likely to follow linear rules than in traditional teams, and for each cell. For example, the probability of visiting cell A is SSGs can account for this possibility. calculated by dividing the number of visits in cell A by the total number of visits. These individual probabilities are then summed The second strength of SSGs constitutes visualizing team up for the entire grid based on the formula by Shannon and interaction patterns and making complex team dynamics more Weaver (1949). Lower entropy values indicate a highly organized accessible. This can be tremendously helpful especially for pattern, whereas high entropy denotes unpredictability. The exploratory research stages, for example when there is little exact formula and implementation in GridWare is described in or no prior empirical research on team dynamics and team Hollenstein (2013; see also Dishion et al., 2004). In our example, interactions in a particular team setting. As discussed by visit entropy was 1.89. The interpretation of this measure should Granic and Hollenstein (2003), SSGs can summarize complex be based on the respective study and the structure of the SSG. For interactional data in an intuitively appealing manner (Granic example, a comparison across different teams who have worked and Hollenstein, 2003; Pennings and Mainhard, 2016). Whereas on a similar task and whose interaction were analyzed with the the theoretical underpinnings of dynamic systems theory may same coding system would likely yield interesting insights. seem daunting, the visualization of such system dynamics via SSGs helps team researchers grasp the characteristics of the team Of note, the SSG technique offers a range of measures and, as an interacting system from a holistic perspective. Visualizing although tempting, these measures should not be used blindly in the complexity of team interactions may be particularly subsequent analyses. Instead, the choice of a specific SSG setting helpful for understanding team contexts that involve frequent and accompanying measure in GridWare or Interact software changes or “upheaval” and that require teams to develop should be guided by theoretical considerations. swift trust and rapid collaboration (i.e., quickly settling into new routines). This includes action teams (e.g., first response BENEFITS AND IMPLICATIONS FOR teams) as well as agile teams (e.g., software development TEAM SCIENCE teams), where behavioral interaction patterns emerge quickly and where teams are often characterized by fluidity and low Team interactions are dynamic and can be rather messy (e.g., stability in team boundaries (Mortensen and Haas, 2018). Cronin et al., 2011). Adopting a fine-grained behavioral approach In those contexts, the adoption of dynamic systems theory to investigate team interactions typically generates large amounts for team science will be particularly fruitful, and SSGs as a of data that can be difficult to make sense of (e.g., Kozlowski visualization tool can help position and guide the scholarly et al., 2015). The SSG technique can address this challenge and thought process in this regard. innovate the study of team evolution and maturation processes. In the following, we first describe the strengths of the SSG When utilizing SSGs as a visualization tool, it is important approach before we outline how this technique complements to decide how to best arrange the different categories along the existing analysis strategies. two axes of the grid. Rearranging the categories may be very helpful for “reading” the interaction more intuitively but should Strengths of the SSG Approach align with the theoretical underpinnings of the respective study. Moreover, the use of SSGs as a visualization tool for complex The strengths of the SSG approach to innovate team science team interaction dynamics also incorporates a movie function broadly fall into three areas. First and foremost, the conceptual that allows the inspection of a team trajectory evolving over time approach underlying SSGs can innovate team science by (see Hollenstein, 2013). Team researchers can either explore the applying non-linear dynamic systems theory and changing the cumulative trajectory of an overall observed team interaction, or epistemology of teams (for a detailed discussion, see Ramos- they can select specific time windows for shorter trajectories (e.g., Villagrasa et al., 2018). The opportunity afforded by SSGs of for highlighting particularly eventful or critical episodes within a embracing the notion of teams as complex and dynamic systems longer stream of team interaction). While this analysis remains and moving away from the typical linear thinking that has qualitative, it can facilitate more dynamic theorizing about the predominated team research (cf. Ramos-Villagrasa et al., 2018) evolution and maturation of team processes. Furthermore, the is particularly fruitful for advancing our understanding of the visualization of complex team dynamics via SSGs may generate evolution and maturation of teamwork and team processes. innovative research hypotheses to be tested in further analyses. Frontiers in Psychology | www.frontiersin.org 560 April 2019 | Volume 10 | Article 863

Meinecke et al. State Space Grids and Teams The third strength of the SSG approach concerns novel p. 108), “[SSG] are an important tool but often it takes many opportunities for empirical research and hypothesis testing based tools to complete the understanding of the phenomenon at on the quantitative measures for complex interaction patterns hand.” We have identified two techniques that, in our opinion, derived by SSG software. SSGs provide a wide array of different are useful complements to the analysis of SSGs, specifically measures that can be compared to traditional measures or added recurrence quantification analysis (e.g., Eckmann et al., 1987; to existing models. Measures cannot only be obtained in a Webber and Zbilut, 2005; Knight et al., 2016) and sequence cumulative fashion, as in our example above, but also for smaller analysis (e.g., Bakeman and Quera, 2011; Herndon and Lewis, time slices within a larger data set. For example, we could request 2015; Klonek et al., 2016). In the following, we briefly compare the number of events per cell for every 5 min of an observed the main similarities and differences between the SSG technique team meeting interaction and thus obtain information about on the one hand and recurrence quantification analysis and the dominant speaker (or any other measure of interest) for sequence analysis on the other hand, respectively. Readers each temporal slice of interest. Such an approach opens up new interested in an overview of additional methods for pattern possibilities for investigating how team processes evolve at a recognition in team process data are referred to Poole (2018) or quicker pace and within much smaller time frames than typically Ramos-Villagrasa et al. (2018). investigated in temporal team process research, and departs from larger-scale temporal frames for conceptualizing team emergence As described earlier, SSGs are a tool for visualizing and (e.g., Kozlowski, 2015). quantifying the trajectories of categorical time-series data such as coded team interactions. Turning to team interactions during Relying on the SSG technique to quantify team interaction organizational meetings as an example, researchers may ask dynamics may be especially useful in the context of infrequent questions such as: Does team behavior A typically coincide with or rare team interaction behaviors. When applying a quantitative team behavior B? Do certain behavioral pairings occur more behavioral observation approach, team researchers may feel often than others? Is the interaction evenly distributed across inclined to neglect such behaviors given their low base rate, or the state space (i.e., flexible patterns) or “boxed” into specific choose to combine them with other behaviors in order to obtain corners (i.e., rigid patterns)? Is each team unique in terms more frequent categories (see Lehmann-Willenbrock and Allen, of exhibiting qualitatively different patterns (e.g., distinctive 2018, for a more detailed discussion of decisions to be made when trajectories resulting in idiosyncratic attractors) or can we coding team interactions). The SSG technique is sensitive to such identify similarities in interaction patterns across different teams? low frequency behaviors, which are sometimes highly informative (e.g., when a rare behavior only occurs in successful but not in Another non-linear approach based on the visualization of unsuccessful teams). time-series data is recurrence quantification analysis (Eckmann et al., 1987; Webber and Zbilut, 2005). The visualizations at As a guiding reminder, team researchers looking to apply the heart of this approach are called recurrence plots (Marwan SSGs to study team interaction dynamics need to be aware et al., 2007; Marwan, 2011). In its most classical application, a and make informed decisions about how their approach to recurrence plot spans two dimensions, but shows the same time coding the observed data will affect the results regarding system series on both axes (e.g., ABACABC, with A, B, and C denoting dynamics that can be obtained using the SSG technique. Of discrete behavioral codes). In contrast to a SSG visualization, note, this does not necessarily mean that SSGs are applied to the recurrence plot does not show specific values along the evaluate entire theories, but rather refers to making conceptually two axes, and the plot does not become denser with time as sound decisions about the operationalization of relevant team more and more events are entered. Instead, the recurrence plot constructs at the behavioral event level. Decisions about how shows when a specific value in the time series repeats itself relevant team interaction phenomena can adequately be captured (e.g., the code “A” reoccurs at positions 3 and 5) and the plot in terms of observable behavioral units should be guided by itself gets larger when the time series is longer. Whenever there conceptual arguments (cf. Lehmann-Willenbrock and Allen, is a repetition in the time series, these recurrence points are 2018), which also applies to decisions about SSGs. In other words, marked black in the recurrence plot (Marwan, 2011). The basic when choosing SSGs to quantify interaction dynamics, team idea underlying the use of recurrence plots is that researchers researchers need to be mindful when conceptualizing the state can recognize repetitive sequences in the time series with the space to ensure that those phenomena or variables of interest naked eye, which resembles the basic notion of SSGs. Similarly, that will later fall onto the two dimensions of the grid will be recurrence quantification analysis offers various measures that assessed at the same time. Moreover, especially when measures can be obtained from the visualizations such as the percentage of duration are of interest to a researcher, clear unitizing rules are of recurrence (Webber and Zbilut, 2005). imperative (i.e., deciding when each behavioral unit within the temporal team interaction stream starts and ends). Since recurrence quantification analysis typically focuses on the repetitive properties of a dynamic system within itself, this Complementary Analyses method may seem less intuitive to team researchers at first glance (but for previous applications in team science, see Ramos- The SSG technique shares common ground with some other Villagrasa et al., 2012; Knight et al., 2016). Moreover, recurrence analytical strategies that aim to distil higher-level emergent quantification analysis focuses exclusively on the structure of patterns from lower-level interaction among individual elements. a system’s dynamics; implications regarding the content of the Thus, we do not want to position SSGs as the new “holy grail” system dynamics are limited. Results of this type of analysis of team research. To put it in the words of Hollenstein (2013, need to be interpreted within a precisely elaborated theoretical Frontiers in Psychology | www.frontiersin.org 571 April 2019 | Volume 10 | Article 863

Meinecke et al. State Space Grids and Teams context. Consequently, recurrence quantification analysis is provides a statistical check for the sequential relationships found less suitable for exploratory research stages. Sample research in the coded data. Although this is certainly also possible questions when applying recurrence quantification to coded with quantifications derived from SSGs, the SSG technique team meeting interactions could include: does the team show in and of itself is much more descriptive in nature. In fact, structural recurrence in interaction data or are their interaction this was one of the main reasons for the development of patterns chaotic? Are repetitions in behavior more apparent at the SSGs (Hollenstein, 2013). Sequence analysis is more rigid beginning or end of the meeting? Are there breakpoints during in comparison to SSGs because it requires the researcher the meeting after which the interaction is more/less structured? to make specific assumptions about the expected patterns How complex are the detected recurrence structures? of behaviors. In addition, behavioral contingencies at higher lags are increasingly difficult to model because they require A benefit of recurrence quantification analysis concerns larger amounts of data (Quera, 2018). Yet, “often, meaningful its ability to process continuously sampled signals (e.g., responses in interpersonal interactions are not immediate” physiological data). When working with continuous measures, (Hollenstein, 2013, p. 109). researcher need to specify a recurrence threshold (i.e., specifying when an event is marked as recurrent), which illustrates that A more recent sequential analysis technique that addresses the method is mathematically more demanding than an analysis some of these caveats is time-window sequential analysis based on SSGs as it includes finding optimal parameters (Yoder and Tapp, 2004; Bakeman and Quera, 2011). Group (Marwan, 2011). In sum, we would argue that the SSG technique researchers can use this technique to test whether a certain is to some extent more accessible for team researchers than response occurs within a pre-defined time window such as recurrence quantification analysis, even though the two methods a 5 s time-window (i.e., a behavior is contingent if we see build on similar ideas—both conceptually and methodologically. a response within 5 s; Bakeman and Quera, 2011). From We are not aware of any studies that use a combination of both a conceptual point of view, this approach can solve some techniques, but we certainly consider this promising (see also of the difficulties associated with specifying meaningful Hollenstein, 2013). time lags. However, its practical implementation is more difficult, since time-window sequential analysis is not Another methodological approach to the study of team integrated in common observational software such as Interact dynamics is to focus on and identify “sub-sequences” in coded (Quera, 2018). team interactions (Poole, 2018). Approaches in this tradition explore more immediate temporal contingencies among coded Likewise, team researchers rarely turn to sequence analysis events and can be subsumed under the umbrella term sequence for exploring co-occurrences in parallel coded strings of analysis (Quera, 2018). Notably, sequence analysis is not one events, although there are procedures that allow this (Quera, particular technique but rather “a toolbox of techniques” 2018). As a result, sequence analysis is often used in (Bakeman and Quera, 2011, p. 134). Over the years, different a simplified form (Herndon and Lewis, 2015). To recall, and increasingly advanced procedures for sequence analysis have with SSGs the combination of at least two variables or been developed (Quera, 2018). dimensions is of interest. As such, the two analysis strategies could by combined by using the observed co-occurrences The types of research questions that can be explored with revealed with the aid of SSGs as a basis for a subsequent sequence analysis include the following: does behavior A trigger sequence analysis. In return, SSGs could be used to visualize or inhibit behavior B, C, or D? Which behaviors A, B, or the results obtained from sequence analysis and make the C increase the likelihood for behavior D? Which behaviors findings more tangible. A, B, or C can inhibit behavior D? Most frequently in team research, studies using sequence analysis explore the extent to Finally, despite its many advantages and application which team members reciprocate verbally (i.e., does behavior possibilities, sequence analysis is not particularly sensitive to low A trigger more of the same). For example, previous research frequency behaviors (for a detailed discussion of the limitations has explored whether complaining leads to further complaining of the sequence analysis approach, see also Chiu and Khoo, during organizational team meetings (Kauffeld and Meyers, 2005). Common practice is therefore to collapse fine-grained 2009). Other research has utilized sequence analysis to test categories into larger macro codes and/or to pool the data across whether monitoring behaviors trigger different responses in groups in order to base the analysis on a larger number of higher- vs. lower-performing anesthesia teams (Kolbe et al., codes (e.g., Klonek et al., 2016). However, this approach regards 2014). For such research questions, the researcher needs to groups as largely homogeneous, which has been criticized as specify a specific time lag. Time lags refer to the number a simplistic reductionist view on teams and team processes of steps that separate a particular behavior from a criterion (Hewes and Poole, 2012). event. Lag1 refers to a coded event directly following the previous one (e.g., does code B immediately follow code A); In sum, the SSG technique has much to offer for team lag2 refers to second-order transitions when a coded event science. To date, SSGs have mainly been used for studying is followed by the next but one coded event, and so forth interactions in dyadic settings, outside the realm of team science (Bakeman and Quera, 2011). Lag sequential analysis can then test (e.g., Pennings et al., 2014; Guo et al., 2017). We hope that whether a certain sequence of events is statistically meaningful team researchers will begin to embrace the SSG technique by comparing the observed transition frequencies to those for enabling novel insights into the complex interactional expected by change. In contrast to SSGs, sequence analysis dynamics at the core of team functioning and performance (e.g., Ramos-Villagrasa et al., 2018). Frontiers in Psychology | www.frontiersin.org 582 April 2019 | Volume 10 | Article 863

Meinecke et al. State Space Grids and Teams APPLICATION EXAMPLE AND TUTORIAL overview and serve the purpose of applying SSGs as an analytical and/or visualization tool. To make the application of SSGs more tangible to team research and development, we will now present an example based on In the fifth step, once all these decisions have been taken, real team data. We provide step-by-step suggestions for using behavioral process data (video/audio recordings or live coding) the technique and hope to highlight the various opportunities can be gathered and coded. It is worth ensuring high-quality data that SSGs offer. through appropriate training of coders and establishing inter- rater reliability. Depending on the sample population, questions A Step-by-Step Overview around data storage and privacy policies should be clarified before data collection and coding. As we have pointed out above, researchers should not begin considering SSGs in the final stages of an investigation. Rather, In the sixth step, once the coding is completed and the decision to employ SSGs should be made early in order to visualizations are available for each team, the SSGs can be be able to account for the requirements of this technique. In interpreted and appropriate measures for describing both the Table 1 we have summarized the key steps for using SSGs in team content and structure of the trajectories can be calculated. These research and development. measures can be easily exported and used for further analysis in other statistic software programs. The first step involves defining the research aim and identifying the theoretical foundations for capturing team Finally, beyond research purposes, the coded data may be used phenomena at the behavioral event level and specifying for team development as detailed below. The visualizations, even temporally sensitive interaction dynamics in the study context. more so than the measures, can serve as a basis for feedback. The two chosen variables should be meaningfully related and their interaction should be grounded in theory. Most likely, the The Data Set nodes or data points (i.e., the observed behavioral units) will not be randomly scattered across the state space but organized Data for this application example were sampled from a recently into clusters. It is advisable to find theoretical support for gathered data set that has not been published to date. The data grouping the expected patterns of nodes into meaningful clusters. set comprises videotapes of the first (T1) and the final (T2) Hence, theory-based considerations should drive how a SSG team meeting of a 6-week long student project at a large Dutch is structured, and how this relates to the overarching team university. The project resembled the work of organizational phenomenon that is studied. This step will ensure an early consultants and required the teams to develop a managerial integration of the SSG technique as a methodological tool into strategy for an organizational change project. The study was the concept of the study. approved by the Economics and Business Ethics Committee at the University of Amsterdam. Participation in the study was The second step entails defining the variables of interest. Since voluntary, and all participants provided their written informed the variables need to fulfill specific norms to be used for SSG consent. From this pool we selected two five-person teams with analyses, it is imperative to account for such norms early on as roughly equal meeting durations on the basis of their productivity well. In particular, it is important that the chosen dimensions (high vs. low). On average, these four team meetings lasted for underlying the SSG can be observed and coded in a sequential 55.14 min (SD = 4.08). As a proxy for productivity, we took fashion (i.e., moment-to-moment). Likewise, the dimensions the rate of solutions mentioned per hour. The productive team should be constructed in a way that they allow for mutually produced 19.45 solutions per hour at T1 and 21.15 solutions per exclusive and exhaustive coding. It is therefore important to hour at T2. The unproductive team produced 6.94 solutions per choose two variables that have similar granularity. hour at T1 and 9.66 solutions per hour at T2. As shown in Table 2, the productive team consistently scored higher on positive team Closely related, the third step includes that both variables characteristics like reflexivity, cohesion, and meeting satisfaction need to be unitized identically. For instance, if one variable was and lower on team conflict measures. measured every 2 min (e.g., mood), the second variable (e.g., number of solutions mentioned) has to provide a data point for Formatting the Data every 2 min as well. Hence, this aspect is important to consider at the research design stage, when making decisions regarding the We coded the observed team meeting interaction using operationalization of variables. The chosen software may pose the act4teams coding scheme (e.g., Kauffeld and Lehmann- additional requirements. For instance, the smallest time scale Willenbrock, 2012; Kauffeld et al., 2018) and Interact software GridWare processes are seconds. Missing data should be avoided (Mangold, 2017). Act4teams is a mutually exclusive and as this interrupts the interaction flow and thus the trajectory. exhaustive coding scheme for measuring problem-solving dynamics that occur in groups and teams. Using the In the fourth step, an appropriate coding scheme can be act4teams coding scheme, a behavioral code is assigned chosen or developed. Available fine-grained coding schemes to each verbal thought unit, which is typically a single may be adjusted and summarized into broader categories to sentence. In order to reduce complexity, we collapsed the fit the purpose at hand. Note that each dimension (variable) 43 fine-grained act4teams codes into six broader aspects of may be coded with a different scheme (e.g., verbal and non- interaction. These covered elements of interactions that were verbal interaction). Although it is not a theoretical requirement, knowledge-oriented, problem-focused, structural, action- for practical reasons a smaller number of coding categories, oriented, relational, and counterproductive. To ensure that for example six to eight on each dimension, will yield a better the coding was exhaustive, we included an additional filler Frontiers in Psychology | www.frontiersin.org 593 April 2019 | Volume 10 | Article 863

Meinecke et al. State Space Grids and Teams TABLE 1 | Basic steps for applying SSGs in team research. Considerations Basic steps (1) Define the (research) aim - Clarify how the context and purpose of the study is linked to the dynamic systems perspective. (2) Define phenomena and variables of interest - Describe the theoretical fundaments for temporally sensitive interaction dynamics. (3) Select unitizing rule (e.g., turn of talk) - Identify the underlying dimensions of the state space. (4) Choose existing coding scheme(s) or develop a new one - Decide how the state space is constructed and define the variables of interest. (5) Gather interaction data and code the data - Variables must be observable simultaneously. (6) Visualize and quantify data in regards of the research question - The variables should be mutually exclusive and exhaustive. - Units for the variables observed should be measured at the same time intervals. (7) Provide feedback to the team - Preferably, time units should not be smaller than 1 s. - Chose or develop one or several coding schemes that fit the research question. - A smaller number of categories will yield a better overview. - Record data such that the variables of interest can be measured effectively. - Train coders and establish inter-rater reliability. - Create a SSG for each team using Interact (Mangold, 2017) or GridWare (Lamey et al., 2004) software. - Interpret the SSGs and derive adequate measures from the visualizations. - Several types of analyses can be conducted on the measures the software offers. - Chose a format that communicates the contents of the analysis which are relevant feedback for the target recipients. TABLE 2 | Aggregated scores on team characteristics for each team at T1 and T2. Team characteristic T1 T2 Unproductive Productive Unproductive Productive Reflexivitya 3.35 4.35 2.85 4.05 Meeting satisfactionb 4.13 4.87 4.10 4.50 Social cohesionc 3.20 3.80 2.83 4.00 Task cohesionc 4.47 4.87 4.27 4.67 Intragroup conflict (relationship)d 1.45 1.05 2.00 1.05 Intragroup conflict (task)d 2.10 1.55 2.70 1.50 Answers were provided on a Likert scale ranging from 1 = very low to 5 = very high. aSchippers et al. (2007), bRogelberg et al. (2010), cCarless and De Paola (2000), dJehn (1995). code labeled “other behavior.” An overview of the simplified The plots in Figures 2, 3 depict the interaction trajectory coding scheme including sample statements for each code for the first 5 min, for the first 20 min, and for the entire is shown in Table 3. With each coded statement, we also meeting, respectively. recorded who the speaker was. Thus, our data format meets the requirements for SSGs explained in section In the following we will discuss the grids and the quantitative “Visualizing Patterns of Dynamic Interactions.” The coding measures with regard to the two teams in a more generalized leads to a multivariate time series of sequentially coded way and point out benefits for both team research and team categorical data. development where relevant. Again, we used the SSG application in Interact software Visual Inspection for visualization and GridWare software to further analyze the coded team data. Each cell in the grid represents a distinct Figure 2 shows the developing SSG for the two teams interactive state defined by the mutual occurrence of a specific during their initial meeting. At first inspection of the entire speaker (x-axis) and the corresponding verbal behavior (y-axis). meetings, we can observe clear differences between them. To visualize how the interaction unfolds over the time of a Starting with the columns (i.e., speakers), we can see an meeting, we created three plots per meeting for each of the two interesting difference concerning the length and distribution teams (see Figures 2, 3). This is possible through a function of speaker turns. First, there is a clearer pattern of cells that integrated in both software applications, i.e., a time slider allows are visited more often than others in the productive team us to choose specific time ranges of interest within the recorded compared to the unproductive team. Second, more circles time. The SSG then builds up gradually. The SSG measures in the productive team are larger which indicates longer can also be calculated for each of the individual time intervals. lasting contributions. Third, the distribution of circles across columns (speakers) in general and that of large circles in Frontiers in Psychology | www.frontiersin.org 1540 April 2019 | Volume 10 | Article 863

Meinecke et al. State Space Grids and Teams TABLE 3 | Behavioral categories, descriptions, and sample statements. Examples Behavioral category Description “Well, I format it like this . . .”, “The guidelines are on blackboard.”, We should ask Marisa about that.” Knowledge-oriented Sharing organizational knowledge, referring to experts, Problem solving and asking questions about opinions, content, or “We have not yet clarified the concept.”, “We have to experience. narrow our focus.”, “We should stick to the marking guidelines.” Identifying, describing, and analyzing problems and “So in sum, we have to start with this point.”, “Let me solutions. write that down.”, “This is a key aspect.”, “We still have 15 min.” Structural Structuring the conversation by clarifying, summarizing Action-oriented content as well as structuring the procedure in terms of “I am curious about the results.”, “That will bring us goals and priorities, time management and task ahead.”, “Okay, I will research that.”, “I will do that next distribution. week then.” “If that’s okay with you, Jim.”, “Yes, exactly.”, “Hmm, Showing interest in change and new ideas as well as yes.”, “I have understood that.” taking responsibility and planning concrete steps. “If everyone did it my way . . .”, “We will wait and see.”, Relational Positive socio-emotional behavior such as humor, “What if that ends up nowhere?”, “Mark should have Counterproductive involving and supporting other team members as well prepared that.” Other as appreciating their contributions. Behavior which disrupts the productivity of the team such as complaining, denying responsibility or side conversations and self-promotion. Behavior which does not fit in any of the previous categories (e.g., pauses, incomplete or incomprehensible sentences). FIGURE 2 | State space grids (SSGs) representing verbal team interactions for a productive and an unproductive team at three time points for the first meeting. The (top) three panels show the SSGs of the productive team. The (bottom) three panels show the SSGs of the unproductive team. A, B, C, D, and E label each of the five team members per team. The size of the circles denotes the duration of each event. KnowEx, knowledge exchange; ProbSolve, problem solving; Struct, structuring; TakeAction, taking initiative; Relat, relational; CMB, counterproductive meeting behavior; Other, verbal behaviors that do not fit any of the six functional categories. particular reveals that in the productive team speakers do not Turning to the rows and looking at the functional interaction seem to have an equal share in the amount and length of categories, more differences arise. In the productive team, their contributions. Some (speakers D and E) dominate the the distribution of circles in the rows shows that some are interaction and others (speaker A) are rather quiet. In the visited more frequently than others. For instance, cells on unproductive team the differences between speakers are more the structural level (e.g., clarifying, prioritizing, and time difficult to characterize. It seems that the conversational floor is management statements) are visited more often than cells on more equally shared. the action-oriented level (e.g., interest in change and action Frontiers in Psychology | www.frontiersin.org 1551 April 2019 | Volume 10 | Article 863

Meinecke et al. State Space Grids and Teams FIGURE 3 | State space grids (SSGs) representing verbal team interactions for a productive and an unproductive team at three time points for the last meeting. These are the same teams as in Figure 2. The (top) three panels show the SSGs of the productive team. The bottom three panels show the SSGs of the unproductive team. A, B, C, D, and E label each of the five team members per team. The size of the circles denotes the duration of each event. KnowEx, knowledge exchange; ProbSolve, problem solving; Struct, structuring; TakeAction, taking initiative; Relat, relational; CMB, counterproductive meeting behavior; Other, verbal behaviors that do not fit any of the six functional categories. planning). Again, the unproductive team lacks such a clear Figure 3 represents the equivalent interaction trajectories trend. Finally, in the productive team we see a dark horizontal for the final meeting. The patterns for each team look rather shade across the relational level. The shade indicates intensive different compared to the patterns for the first meeting. For interaction within that level, that is relational contributions instance, observing the final grid for the productive team, it is are often followed by other relational contributions. These less easy to identify a dominant speaker, members seem more observations are relatively rough but they provide an overview of equally involved in the interaction compared to the first meeting. the interaction and thus an accessible form of feedback that can Especially team member A who was very quiet at T1 is now fully be insightful for team leaders and team members themselves (e.g., integrated in the interaction at T2. Circles in the top three rows Who dominates the conversation? Who tends to structure the are larger than in the bottom rows. Thus, knowledge-oriented, meeting? Who takes action? What contributions occur at what structural, and problem-solving contributions take up more time point during the meeting?). Before turning to the quantification than other types of contributions in the productive team. The of these observations, we will briefly examine the plots that unproductive team shows two dark horizontal shadows, one on represent earlier interaction stages within the same meetings. the top row suggesting an intensive exchange of knowledge- After 5 min, in both teams one individual seems to dominate the oriented contributions, and one on the relational level indicating interaction: in the productive team, member E makes a number of strong positive socio-emotional exchange. contributions and a particularly lengthy knowledge-oriented one. This active role seems to remain stable across the meeting. In the Quantitative Inspection unproductive team, after 5 min, member D has a similar role with a prominent problem-solving contribution. D, however, does For many of these observations we can obtain quantitative not remain dominant throughout the meeting. Further, in this measures. These help to analyze the content and structure of the first grid the productive team shows more relational interaction interaction within and across grids. In practical terms, it means compared to the unproductive team. This pattern intensifies that we could establish dominant speakers, dominant interaction throughout the meeting. The unproductive team, however, shows categories or characterize speakers with regard to their types pronounced interaction on the knowledge-oriented level after of interactive contributions. In addition, we can quantify if the 5 min that increases over time. To conclude, the two teams show interaction was rigid or flexible such that structural patterns specific and different trends from the beginning, and these may in the trajectory can be identified. For example, if we want to explain higher or lower productivity. Such conclusions highlight know who of the speakers dominated the interaction we can the potential of identifying dysfunctional processes early on look for the number of events that we find within that speaker’s during the meeting to be able to correct them guiding the team column or we might look at the proportion of the total time into more productive dynamics. taken up by the events of that speaker. Taking the example of the productive team at T1 (Figure 2) makes clear how critical it is to Frontiers in Psychology | www.frontiersin.org 1562 April 2019 | Volume 10 | Article 863

Meinecke et al. State Space Grids and Teams determine these measures beforehand and rooting this decision diversity training may be aimed at enhancing the willingness to in theoretical grounds: considering the number of events per cooperate in diverse teams (e.g., affective changes), increasing speaker yields member C as the dominant individual (226 events) knowledge regarding the potential benefits and pitfalls of while we can record much less events for speaker D (141 events) diversity for teamwork (cognitive changes), providing the skills and speaker E (153 events) which we had identified as dominant to more effectively utilize the heterogeneity of ideas and through our visual inspection. Considering the proportion of perspectives present in diverse teams (skill-based changes), the total time per speaker results in a different conclusion: the leading to measurable performance improvements (e.g., Homan contributions of the three speakers are rather similar, although et al., 2015). In contrast to team training, team development speaker E slightly dominates the conversational floor (C = 23.5%, (e.g., team coaching or developmental assignments) tends to D = 23.6%, and E = 27.3%). Overall, the standard deviation be broader in scope and has a longer-time perspective. The for these percentages was 9.23. Looking at the unproductive skills to be acquired also typically go beyond those required team, the standard deviation for the proportion of the total time for effectively accomplishing current tasks, jobs, and/or roles per speaker was 5.27. This supports our preliminary conclusion (Aguinis and Kraiger, 2009). Yet, boundaries between training about a more even distribution of speaker contributions in and development are fluid, and both show considerable overlap the unproductive team at T1. Still, interesting differences exist. in the principles followed to ensure effectiveness. Therefore, Specifically, speaker E’s contributions composed 23.2% of the unless specified otherwise, we use both terms interchangeably and overall conversation whereas speaker C only contributed 7.1%. assume that both formats can benefit from SSGs in similar ways. Turning to measures of structure, findings reveal that all teams Training and development strategies typically follow several rather exhibit flexible interaction. The teams explored large parts principles to ensure effectiveness (e.g., Salas and Cannon-Bowers, of the grids with an average cell range of 38.75. Likewise, and 2001). These entail presenting concepts and information relevant because all team members did contribute to the discussion, the to the participant; showcasing the knowledge, skills, and abilities values for dispersion ranged between 0.97 and 0.98. Values for (KSAs) to be learnt; allowing for practicing the KSAs; and visit entropy were in the range of 3.22–3.49. Taken together, these supplying participants with feedback during practicing and values indicate a highly variable interaction style and show that on improvements made over time. We believe that the SSG interaction is rather difficult to predict. Contrary to other studies technique is particularly useful to support the feedback element with SSGs (e.g., van Dijk et al., 2017), our coded team data was of effective training and development. not boxed into a specific corner of the SSG. This is not necessarily characteristic for team interaction patterns in general but is, in The SSGs allow for detailed and visually appealing feedback part, due to how we defined the dimensions in our particular based on actual behavior. This feedback can support teams in example with speakers on one axis and coded talk on the other. diagnosing the state they are in terms of team processes (e.g., knowledge sharing and utilization), in reflecting on emergent BENEFITS AND IMPLICATIONS FOR states (e.g., relational conflict), and in improving on important TEAM TRAINING AND DEVELOPMENT team processes. For instance, teams could receive feedback on their status quo as well as how their status quo has changed over We would like to conclude this article with suggestions and the course of a training or developmental activity. Scholars have ideas for the practical application of SGGs. Of note, these argued that feedback tools with a higher temporal resolution are suggestions require future empirical work to evaluate their especially suitable for providing developmental feedback (e.g., actual utility for team training and development. Yet, overall, Rosen and Dietz, 2017). An important advantage of SSGs is we foresee multiple benefits of the application of SSGs in the that they allow teams and those involved in team training and context of team training and development, facilitating team development (e.g., leaders, trainers, and coaches) to gain an easily maturation and evolution over time. First of all, getting teams accessible overview of micro-level team interaction data that to consider their team as a system of interactions, rather otherwise would be perceived as messy and difficult to grasp. than a collection of people, may inspire novel understanding The software’s “movie function,” as described earlier, may further and insights regarding interdependencies and team dynamics. support such practicing and feedback over time, as it adds further However, such a perspective can be quite complex and requires visual stimulation to other established forms of presentation a holistic picture of the team interaction space. Visualizing this (Myer et al., 2013). In addition, as SSGs can be administered holistic picture via SSGs and presenting the behavioral feedback repeatedly, (lack of) improvements could be detected, allowing to the team can likely serve as a development trigger in this regard teams to redirect or strengthen efforts if needed. (cf. Lehmann-Willenbrock and Kauffeld, 2010). In the following, we point out specific ways in which SSGs might be used for As SSGs are based on actual behavior, using this technique effective delivery and transfer of training and development, along for feedback purposes might help circumvent validity and with recommendations for differing team contexts. fairness issues. Such issues may arise when feedback is based on attributions or interpretations of behaviors, or of attitudes Training is considered effective when it produces changes and underlying traits (e.g., by means of a rating scale completed in cognitive, affective, and/or skill-based outcomes (Salas and by one’s supervisor or team members, or by means of a Cannon-Bowers, 2001), and leads to transfer of learning to supervisor’s forced ranking of members in a team). Furthermore, the work context (Blume et al., 2010). For instance, a team feedback on relatively stable dimensions (e.g., intellectual ability) does not offer guidance regarding how to improve one’s behavior. Comprehensible feedback based on actual behavior, Frontiers in Psychology | www.frontiersin.org 1573 April 2019 | Volume 10 | Article 863

Meinecke et al. State Space Grids and Teams however, increases the likelihood that feedback leads to improved seek to improve their phase-specific behavior over time (e.g., performance (e.g., Bandura, 1986; Kluger and DeNisi, 1996; by increasing reflexivity in transition phases and improving on Roter et al., 2004). coordination in action phases). In this case, using SSGs repeatedly across multiple performance cycles may prove most conducive to Besides their role in feedback, SSGs may be used to continuous learning. demonstrate the KSAs to be learnt during training and development, and facilitate subsequent practicing. For example, Finally, certain types of teams may particularly benefit from for more standardized procedures, teams may watch a video- using SSGs as a feedback and development tool. As our based example of both an ineffective and effective team application example shows, there are visible differences in the interaction. This demonstration could be accompanied by SSGs interaction patterns not only between teams but also across reflecting the respective patterns of observed interactions in the different stages in the team’s life cycle (e.g., as determined by effective and ineffective example. The trainer or coach could the duration of a project). Identifying characteristic patterns for then discuss concrete steps to bring the ineffectively interacting team processes and emergent states embedded in certain stages team closer to the effectively interacting team. Alternatively, of a project could help evaluate team processes in a standardized team members could identify ways to approximate the effectively way. This could be especially interesting in and applicable to interacting team’s profile. Yet, “it is important to remember that the context of SCRUM teams. While their project phases are all teams are not equal” (Salas et al., 2017, p. 21). Especially in relatively short and contents may vary according to project, the complex situations, the results of a SSG analysis of a successful general procedures employed in SCRUM teams follow similar team should not necessarily serve as a model for other teams patterns across projects (Schwaber, 1997; Rising and Janoff, (i.e., “one size fits all”). In such cases, it is particularly important 2000). Furthermore, teams undergoing intense training (e.g., that the trainer or coach stimulates reflection, so that the team in the form of simulations) before entering the performance members themselves can decide which elements can serve as a stage such a crisis or emergency teams, aviation or astronautic model for their own teamwork. Building a shared understanding crews, or firefighter and special force units may be particularly of successful team interaction patterns is key to make sure that all attuned to benefit from the fine-grained, behavior-based feedback team members equally benefit from team training with SSGs. This opportunities of the SSG technique. Systematically studying SSGs brings us to our next point, i.e., using SSG for team development. obtained during training and development in these team contexts may afford the opportunity to extract knowledge on more generic Compared to team training, team development may entail a patterns of effective behavior across types of teams. longer and less formalized process, allowing for more profound and longer-lasting maturation and evolution processes in teams. ETHICS STATEMENT Less emphasis is given on how a team compares to other teams (e.g., by comparing the team’s current SSG with the average SSG The study was approved by the Economics and Business Ethics in the department, organization, or branch). Rather, development Committee at the University of Amsterdam. Participation in the is concerned with the team’s growth over time (e.g., Aguinis study was voluntary, and all participants provided their written and Kraiger, 2009). We expect SSGs to be helpful in stimulating informed consent. this growth, as the technique allows for observing the same aspects of a team’s interaction at different points in time. These AUTHOR CONTRIBUTIONS points in time may demarcate different “life stages” such as at team formation and in the middle and end of a project (cf. AM developed the original idea for the manuscript, took the Tuckman, 1965; Gersick, 1988) or phases in a team’s performance lead in writing, and performed the analyses. CH contributed cycle (e.g., action versus transition phases; Marks et al., 2001). to writing the manuscript and aided in data analysis and Depending on the exact purpose, it might be useful to employ interpretation. NL-W and CB collected the data, critically revised the same or different state spaces at different points in time. To the manuscript for intellectual content, and contributed to observe development on a given behavioral pattern, using the writing the manuscript. All authors approved the manuscript same state space is likely to be most suitable. To understand to be published. whether teams appropriately deal with the unique demands that differing stages or phases impose, using phase- or stage- specific state spaces might be more insightful. Teams might also REFERENCES Bakeman, R., and Quera, V. (2011). Sequential Analysis and Observational Methods for the Behavioral Sciences. New York, NY: Cambridge University Press. Aguinis, H., and Kraiger, K. (2009). Benefits of training and development for individuals and teams, organizations, and society. Ann. Rev. Psychol. 60, 451– Bandura, A. (1986). Social Foundations of Thought and Action: A Social Cognitive 474. doi: 10.1146/annurev.psych.60.110707.163505 Theory. Englewood Cliffs, NJ: Prentice-Hall. Alligood, K. T., Sauer, T. D., and Yorke, J. A. (1996). Chaos: An Introduction to Blume, B. D., Ford, J. K., Baldwin, T. T., and Huang, J. L. (2010). Transfer Dynamical Systems. New York, NY: Springer. of training: a meta-analytic review. J. Manag. 36, 1065–1105. doi: 10.1177/ 0149206309352880 Arrow, H., McGrath, J. E., and Berdahl, J. L. (2000). Small Groups as Complex Systems: Formation, Coordination, Development, and Adaptation. Thousand Butler, E. A., Hollenstein, T., Shoham, V., and Rohrbaugh, N. (2014). A dynamic Oaks, CA: Sage. state-space analysis of interpersonal emotion regulation in couples who smoke. J. Soc. Pers. Relationsh. 31, 907–927. doi: 10.1177/0265407513508732 Frontiers in Psychology | www.frontiersin.org 1584 April 2019 | Volume 10 | Article 863

Meinecke et al. State Space Grids and Teams Carless, S. A., and De Paola, C. (2000). The measurement of cohesion in work Kauffeld, S., and Lehmann-Willenbrock, N. (2012). Meetings teams. Small Group Res. 31, 71–88. doi: 10.1177/104649640003100104 matter: effects of team meetings on team and organizational success. Small Group Res. 43, 130–158. doi: 10.1177/10464964114 Chiu, M. M., and Khoo, L. (2005). A new method for analyzing sequential 29599 processes. Small Group Res. 36, 600–631. doi: 10.1177/1046496405279309 Kauffeld, S., Lehmann-Willenbrock, N., and Meinecke, A. L. (2018). “The Couto, A. B., Barbosa, E., Silva, S., Bento, T., Teixeira, A. S., Salgado, J., et al. Advanced Interaction Analysis for Teams (act4teams) coding scheme,” in (2016). Client’s immersed and distanced speech and therapist’s interventions in The Cambridge Handbook of Group Interaction Analysis, eds E. Brauner, M. emotion-focused therapy for depression: an intensive analysis of a case study. Boos, and M. Kolbe (New York, NY: Cambridge University Press), 422–431. Res. Psychother. 19, 136–149. doi: 10.4081/ripppo.2016.205 doi: 10.1017/9781316286302.022 Cronin, M. A., Weingart, L. R., and Todorova, G. (2011). Dynamics in groups: Kauffeld, S., and Meyers, R. A. (2009). Compliant and solution-oriented circles: are we there yet? Acad. Manag. Ann. 5, 571–612. doi: 10.1080/19416520.2011. interaction patterns in work group discussions. Eur. J. Work Organ. Psychol. 18, 590297 267–294. doi: 10.1080/13594320701693209 Dishion, T. J., Nelson, S. E., Winter, C. E., and Bullock, B. M. (2004). Adolescent Klonek, F. E., Quera, V., Burba, M., and Kauffeld, S. (2016). Group interactions and friendship as a dynamic system: entropy and deviance in the etiology and time: using sequential analysis to study group dynamics in project meetings. course of male antisocial behavior. J. Abnorm. Child Psychol. 32, 651–663. Group Dyn. 20, 209–222. doi: 10.1037/gdn0000052 doi: 10.1023/B:JACP.0000047213.31812.21 Kluger, A. N., and DeNisi, A. (1996). The effects of feedback interventions Eckmann, J.-P., Oliffson Kamphorst, S., and Ruelle, D. (1987). Recurrence plots of on performance: a historical review, a meta-analysis, and a dynamical systems. Europhys. Lett. 4, 973–977. doi: 10.1209/0295-5075/4/9/004 preliminary feedback intervention theory. Psychol. Bull. 119, 254–284. doi: 10.1037/0033-2909.119.2.254 Erickson, K., Côté, J., Hollenstein, T., and Deakin, J. (2011). Examining coach athlete interactions using state space grids: an observational analysis in Knight, A. P., Kennedy, D. M., and McComb, S. A. (2016). Using recurrence competitive youth sport. Psychol. Sport Exerc. 12, 645–654. doi: 10.1016/j. analysis to examine group dynamics. Group Dyn. 20, 223–241. doi: 10.1037/ psychsport.2011.06.006 gdn0000046 Gersick, C. J. G. (1988). Time and transition in work teams: toward a new model of Kolbe, M., Grote, G., Waller, M. J., Wacker, J., Grande, B., Burtscher, M. J., group development. Acad. Manag. J. 31, 9–41. doi: 10.2307/256496 et al. (2014). Monitoring and talking to the room: autochthonous coordination patterns in team interaction and performance. J. Appl. Psychol. 99, 1254–1267. Granic, I. (2005). Timing is everything: developmental psychopathology from a doi: 10.1037/a0037877 dynamic systems perspective. Dev. Rev. 25, 386–407. doi: 10.1016/j.dr.2005.10. 005 Koster, E. H. W., Fang, L., Marchetti, I., Ebner-Priemer, U., Kirsch, P., Huffzinger, S., et al. (2015). Examining the relation between affect and rumination in Granic, I., and Hollenstein, T. (2003). Dynamic systems methods for models of remitted depressed individuals: a dynamic systems analysis. Clin. Psychol. Sci. developmental psychopathology. Dev. Psychopathol. 15, 641–669. doi: 10.1017/ 3, 619–627. doi: 10.1177/2167702615578129 s0954579403000324 Kozlowski, S. W. J., Chao, G. T., Chang, C.-H., and Fernandez, R. (2015). “Team Granic, I., and Lamey, A. V. (2002). Combining dynamic systems and dynamics: using “big data” to advance the science of team effectiveness”, in Big multivariate analyses to compare the mother-child interactions of externalizing Data at Work: The Data Science Revolution and Organizational Psychology, eds subtypes. J. Abnorm.l Child Psychol. 30, 265–283. doi: 10.1023/A:101510691 S. Tonidandel, E. King, and J. Cortina (New York, NY: Routledge Academic), 3866 272–309. Guastello, S. J., and Liebovitch, L. S. (2009). “Introduction to nonlinear dynamics Kozlowski, S. W. J. (2015). Advancing research on team process dynamics: and complexity,” in Chaos and Complexity in Psychology: The Theory of theoretical, methodological, and measurement considerations. Organ. Psychol. Nonlinear Dynamic Systems, eds S. J. Guastello, M. Koopmans, and D. Rev. 5, 270–299. doi: 10.1177/2041386614533586 Pincus (New York, NY: Cambridge University Press), 1–40. doi: 10.1017/ cbo9781139058544.002 Kozlowski, S. W. J., and Ilgen, D. R. (2006). Enhancing the effectiveness of work groups and teams. Psychol. Sci. Publ.c Inter. 7, 77–124. doi: 10.1111/j.1529-1006. Guo, Y., Garfin, D. R., Ly, A., and Goldberg, W. A. (2017). Emotion coregulation 2006.00030.x in mother-child dyads: a dynamic systems analysis of children with and without autism spectrum disorder. J. Abnorm. Child Psychol. 45, 1369–1383. doi: 10. Lamey, A., Hollenstein, T., Lewis, M. D., and Granic, I. (2004). GridWare (Version 1007/s10802-016-0234-9 1.1) [Computer software]. Available at: http://statespacegrids.org (accessed August 28, 2018). Ha, T., and Granger, D. A. (2016). Family relations, stress, and vulnerability: biobehavioral implications for prevention and practice. Fam. Relat. 65, 9–23. Lehmann-Willenbrock, N., and Allen, J. A. (2018). Modeling temporal interaction doi: 10.1111/fare.12173 dynamics in organizational settings. J. Bus. Psychol. 33, 325–344. doi: 10.1007/ s10869-017-9506-9 Herndon, B., and Lewis, K. (2015). Applying sequence methods to the study of team temporal dynamics. Organ. Psychol. Rev. 5, 318–332. doi: 10.1177/ Lehmann-Willenbrock, N., and Kauffeld, S. (2010). “The downside of 2041386614538276 communication: Complaining cycles in group discussions,” in The Handbook For Working With Difficult Groups: How They Are Difficult, Why They Hewes, D. E., and Poole, M. S. (2012). “The analysis of group interaction processes,” Are Difficult, What You Can Do, ed. S. Schuman (San Francisco, CA: in Research Methods for Studying Groups and Teams: A Guide to Approaches, Jossey-Bass/Wiley), 33–54. Tools, and Technologies, eds A. B. Hollingshead and M. S. Poole (New York, NY: Routledge), 358–385. Lewis, M. D. (2000). The promise of dynamic systems approaches for an integrated account of human development. Child Dev. 71, 36–43. doi: 10.1111/1467-8624. Hollenstein, T. (2007). State space grids: analyzing dynamics across development. 00116 Int. J. Behav. Dev. 31, 384–396. doi: 10.1177/0165025407077765 Lewis, M. D., and Granic, I. (2000). Emotion, Development, And Self-Organization: Hollenstein, T. (2011). Twenty years of dynamic systems approaches to Dynamic Systems Approaches To Emotional Development. New York, NY: development: significant contributions, challenges, and future directions. Child Cambridge University Press. Dev. Persp. 5, 256–259. doi: 10.1111/j.1750-8606.2011.00210.x Lewis, M. D., Lamey, A. V., and Douglas, L. (1999). A new dynamic systems Hollenstein, T. (2013). State Space Grids: Depicting Dynamics Across Development. method for the analysis of early socioemotional development. Dev. Sci. 2, New York, NY: Springer. 457–475. doi: 10.1111/1467-7687.00090 Homan, A. C., Buengeler, C., Eckhoff, R. A., van Ginkel, W. P., and Voelpel, Mangold, A. (2017). INTERACT (Lab Suite Version 2017) [Computer software]. S. C. (2015). The interplay of diversity training and diversity beliefs on Mangold International GmbH. www.mangold-international.com (accessed team creativity in nationality diverse teams. J. Appl. Psychol. 100, 1456–1467. March 1, 2018). doi: 10.1037/apl0000013 Marks, M. A., Mathieu, J. E., and Zaccaro, S. J. (2001). A temporally based Jehn, K. A. (1995). A multimethod examination of the benefits and detriments of framework and taxonomy of team processes. Acad. Manag. Rev. 26, 356–376. intragroup conflict. Administ. Sci. Quart. 40, 256–282. doi: 10.2307/2393638 doi: 10.2307/259182 Johnson, S. L., and Nowak, A. (2002). Dynamical patterns in bipolar depression. Pers. Soc. Psychol. Rev. 6, 380–387. doi: 10.1207/S15327957PSPR0604_12 Frontiers in Psychology | www.frontiersin.org 1595 April 2019 | Volume 10 | Article 863

Meinecke et al. State Space Grids and Teams Marrone, J. A. (2010). Team boundary spanning: a multilevel review of past Salas, E., and Cannon-Bowers, J. A. (2001). The science of training: a decade of research and proposals for the future. J. Manag. 36, 911–940. doi: 10.1177/ progress. Ann. Rev. Psychol. 52, 471–499. doi: 10.1146/annurev.psych.52.1.471 0149206309353945 Salas, E., Reyes, D. L., and Woods, A. L. (2017). “The assessment of Marwan, N. (2011). How to avoid potential pitfalls in recurrence plot based team performance: Observations and needs,” in Innovative Assessment of data analysis. Int. J. Bifurc. Chaos 21, 1003–1017. doi: 10.1142/S02181274110 Collaboration. Methodology of Educational Measurement and Assessment, eds 29008 A. A. von Davier, M. Zhu, and P. C. Kyllonen (Cham: Springer), 21–36. doi: 10.1007/978-3-319-33261-1_2 Marwan, N., Romano, M. C., Thiel, M., and Kurths, J. (2007). Recurrence plots for the analysis of complex systems. Phys. Rep. 438, 237–329. doi: 10.1016/j. Schippers, M. C., Den Hartog, D. N., and Koopman, P. L. (2007). Reflexivity physrep.2006.11.001 in teams: a measure and correlates. Appl. Psychol. Int. Rev. 56, 189–211. doi: 10.1111/j.1464-0597.2006.00250.x McGrath, J. E. (1964). Social Psychology: A Brief Introduction. New York, NY: Holt, Rinehart, & Winston. Schwaber, K. (1997). “Scrum development process,” in Proceedings of the Business Object Design and Implementation. OOPSLA ’95 Workshop. eds J. Sutherland, D. McGrath, J. E., Arrow, H., and Berdahl, J. L. (2000). The study of groups: Patel, C. Casanave, J. Miller, and G. Hollowell (London, UK: Springer), 117–134. past, present, and future. Pers. Soc. Psychol. Rev. 4, 95–105. doi: 10.1207/ doi: 10.1007/978-1-4471-0947-1_11 S15327957PSPR0401_8 Sesemann, E. M., Kruse, J., Gardner, B. C., Broadbent, C. L., and Spencer, Mortensen, M., and Haas, M. R. (2018). Rethinking teams: from bounded T. A. (2017). Observed attachment and self-report affect within romantic membership to dynamic participation. Organ. Sci. 29, 341–355. doi: 10.1287/ relationships. J. Couple Relationsh.p Ther. 16, 102–121. doi: 10.1080/15332691. orsc.2017.1198 2016.1238794 Myer, G. D., Stroube, B. W., DiCesare, C. A., Brent, J. L., Ford, K. R., Heidt, Shannon, C. E., and Weaver, W. (1949). The Mathematical Theory of R. S., et al. (2013). Augmented feedback supports skill transfer and reduces Communication. Urbana, IL: University of Illinois Press. high-risk injury landing mechanics: a double-blind, randomized controlled laboratory study. Am. J. Sports Med. 41, 669–677. doi: 10.1177/036354651247 Thelen, E., and Smith, L. B. (1998). “Dynamic systems theories,” in Handbook of 2977 Child Psychology, eds W. Damon, R. M. Lerner, and N. Eisenberg (New York, NY: Wiley), 563–634. Pennings, H. J., and Mainhard, T. (2016). “Analyzing teacher–student interactions with state space grids,” in Complex Dynamical Systems in Tomicic, A., Martínez, C., Pérez, J. C., Hollenstein, T., Angulo, S., Gerstmann, Education: Concepts, Methods and Applications, eds M. Koopmans and A., et al. (2015). Discourse-voice regulatory strategies in the psychotherapeutic D. Stamovlasis (Basel: Springer), 233–271. doi: 10.1007/978-3-319-275 interaction: a state-space dynamics analysis. Front. Psychol. 6:378. doi: 10.3389/ 77-2_12 fpsyg.2015.00378 Pennings, H. J. M., van Tartwijk, J., Wubbels, T., Claessens, L. C. A., van der Want, Tuckman, B. W. (1965). Developmental sequence in small groups. Psychol. Bull. A. C., and Brekelmans, M. (2014). Real-time teacher–student interactions: a 63, 384–399. doi: 10.1037/h0022100 dynamic systems approach. Teach. Teach. Educ. 37, 183–193. doi: 10.1016/j.tate. 2013.07.016 Turnnidge, J., Cote, J., Hollenstein, T., and Deakin, J. (2014). A direct observation of the dynamic content and structure of coach-athlete interactions in a model Poole, M. S. (2018). “Coding and analyzing multiple levels,” in The Cambridge sport program. J. Appl. Sport Psychol. 26, 225–240. doi: 10.1080/10413200.2013. Handbook of Group Interaction Analysis, eds E. Brauner, M. Boos, and M. 821637 Kolbe (New York, NY: Cambridge University Press), 370–386. doi: 10.1017/ 9781316286302.019 van Dijk, R., Dekovic´, M., Bunte, T. L., Schoemaker, K., Zondervan-Zwijnenburg, M., Espy, K. A., et al. (2017). Mother-child interactions and externalizing Quera, V. (2018). “Analysis of interaction sequences,” in The Cambridge behavior problems in preschoolers over time: inhibitory control as a mediator. Handbook of Group Interaction Analysis, eds E. Brauner, M. Boos, and M. J. Abnorm. Child Psychol. 45, 1503–1517. doi: 10.1007/s10802-016-0258-1 Kolbe (New York, NY: Cambridge University Press), 295–322. doi: 10.1017/ 9781316286302.016 Waller, M. J., Okhuysen, G. A., and Saghafian, M. (2016). Conceptualizing emergent states: a strategy to advance the study of group dynamics. Acad. Ramos-Villagrasa, P. J., Marques-Quinteiro, P., Navarro, J., and Rico, R. (2018). Manag. Ann. 10, 561–598. doi: 10.1080/19416520.2016.1120958 Teams as complex adaptive systems: reviewing 17 years of research. Small Group Res. 49, 135–176. doi: 10.1177/1046496417713849 Webber, C. L., and Zbilut, J. P. (2005). “Recurrence quantification analysis of nonlinear dynamical systems,” in Tutorials in Contemporary Nonlinear Methods Ramos-Villagrasa, P. J., Navarro, J., and García-Izquierdo, A. L. (2012). Chaotic for the Behavioral Sciences, eds M. A. Riley and G. C. Van Orden (Arlington, VA: dynamics and team effectiveness: evidence from professional basketball. National Science Foundation), 26–94. Eur. J. Work Organ. Psychol. 21, 778–802. doi: 10.1080/1359432X.2012.66 9525 Yoder, P. J., and Tapp, J. (2004). Empirical guidance for time-window sequential analysis of single cases. J. Behav. Educ. 13, 227–246. doi: 10.1023/B:JOBE. Rising, L., and Janoff, N. S. (2000). The scrum software development process for 0000044733.03220.a9 small teams. IEEE softw. 17, 26–32. doi: 10.1109/52.854065 Conflict of Interest Statement: The authors declare that the research was Rogelberg, S. G., Allen, J. A., Shanock, L., Scott, C., and Shuffler, M. (2010). conducted in the absence of any commercial or financial relationships that could Employee satisfaction with meetings: a contemporary facet of job satisfaction. be construed as a potential conflict of interest. Hum. Res. Manag. 49, 149–172. doi: 10.1002/hrm.20339 Copyright © 2019 Meinecke, Hemshorn de Sanchez, Lehmann-Willenbrock and Rosen, M. A., and Dietz, A. S. (2017). “Team performance measurement,” in The Buengeler. This is an open-access article distributed under the terms of the Creative Wiley Blackwell Handbook of the Psychology of Team Working and Collaborative Commons Attribution License (CC BY). The use, distribution or reproduction in Processes, eds E. Salas, R. Rico, and J. Passmore (Hoboken, NJ: Wiley-Blackwell), other forums is permitted, provided the original author(s) and the copyright owner(s) 481–502. are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted Roter, D. L., Larson, S., Shinitzky, H., Chernoff, R., Serwint, J. R., Adamao, which does not comply with these terms. G., et al. (2004). Use of an innovative video feedback technique to enhance communication skills training. Med. Educ. 38, 145–157. doi: 10.1046/j.1365- 2923.2004.01754.x Frontiers in Psychology | www.frontiersin.org 1606 April 2019 | Volume 10 | Article 863

ORIGINAL RESEARCH published: 03 May 2019 doi: 10.3389/fpsyg.2019.00992 The Emergence of Group Potency and Its Implications for Team Effectiveness Hayden J. R. Woodley1*†, Matthew J. W. McLarnon2† and Thomas A. O’Neill3 1 Faculty of Business, University of Prince Edward Island, Charlottetown, PE, Canada, 2 Department of Psychology, Oakland University, Rochester, MI, United States, 3 Department of Psychology, University of Calgary, Calgary, AB, Canada Edited by: Much of the previous research on the emergence of team-level constructs has Eduardo Salas, overlooked their inherently dynamic nature by relying on static, cross-sectional Rice University, United States approaches. Although theoretical arguments regarding emergent states have underscored the importance of considering time, minimal work has examined the Reviewed by: dynamics of emergent states. In the present research, we address this limitation by Ishani Aggarwal, investigating the dynamic nature of group potency, a crucial emergent state, over Brazilian School of Public time. Theory around the “better-than-average” effect (i.e., an individual’s tendency to and Business Administration, Brazil think he/she is better than the average person) suggests that individuals may have elevated expectations of their group’s early potency, but may decrease over time as Gary Pheiffer, team members interact gain a more realistic perspective of their group’s potential. In University of Hertfordshire, addition, as members gain experience with each other, they will develop a shared understanding of their team’s attributes. The current study used latent growth and United Kingdom consensus emergence modeling to examine how potency changes over time, and its relation with team effectiveness. Further, in accordance with the input-process-output *Correspondence: framework, we investigated how group potency mediated the relations between team- Hayden J. R. Woodley level compositions of conscientiousness and extraversion and team effectiveness. We collected data at three time points throughout an engineering design course from 337 [email protected] first-year engineering students that comprised 77 project teams. Results indicated that group potency decreased over time in a linear trend, and that group consensus †These authors have contributed increased over time. We also found that teams’ initial potency was a significant predictor equally to this work as first authors of team effectiveness, but that change in potency was not related to team effectiveness. Finally, we found that the indirect effect linking conscientiousness to effectiveness, Specialty section: through initial potency, was supported. Overall, the current study offers a unique This article was submitted to understanding of the emergence of group potency, and facilitate a number theoretical and practical implications, which are discussed. Organizational Psychology, a section of the journal Keywords: group potency, emergence, team effectiveness, conscientiousness, extraversion Frontiers in Psychology INTRODUCTION Received: 01 November 2018 Accepted: 15 April 2019 According to the input-process-outcome (IPO) framework (McGrath, 1964) and related models Published: 03 May 2019 (e.g., the input-mediator-output-input [IMOI] model; Ilgen et al., 2005), emergent states are integral to understanding the effectiveness of teams. In this light, extensive research has been Citation: conducted in effort to improve our understanding of how emergent states influence team Woodley HJR, McLarnon MJW and O’Neill TA (2019) The Emergence of Group Potency and Its Implications for Team Effectiveness. Front. Psychol. 10:992. doi: 10.3389/fpsyg.2019.00992 Frontiers in Psychology | www.frontiersin.org 611 May 2019 | Volume 10 | Article 992

Woodley et al. The Emergence of Group Potency effectiveness (Kozlowski and Ilgen, 2006). Marks et al. (2001) Schaubroeck et al., 2016, among others). Although it has been defined emergent states as, “constructs that characterize described in different forms previously (see Stajkovic et al., properties of the team that are typically dynamic in nature 2009), we adhere to its conventional definition as a team’s and vary as a function of team context, inputs, processes, generalized confidence in its ability to perform across a variety and outcomes” (p. 357). Examples of emergent states include of situations (see Guzzo et al., 1993). Potency differs from collective efficacy, group potency, and cohesion. Overall, meta- efficacy, in that “efficacy represents a shared, task-specific analyses have found that the previously mentioned emergent expectation that the team can accomplish its goals, whereas states are positively related to team effectiveness (e.g., Gully potency is a more generalized sense of competence” (Kozlowski, et al., 2002; Beal et al., 2003; Stajkovic et al., 2009, respectively). 2018, p. 208). To date, two meta-analyses have investigated Although these findings have been influential in building the relations between group potency and team performance our understanding of team effectiveness, little research has (Gully et al., 2002; Stajkovic et al., 2009), with both reporting that investigated the temporal, dynamic aspects of emergent states group potency is positively related to team performance, ρ = 0.35 (Kozlowski et al., 2016; Waller et al., 2016). Ilgen et al. (2005) and 0.29, respectively. argued that time plays an important role in understanding the emergence of states in teams, and without more direct insight Nevertheless, these meta-analyses are based on research that into the temporal nature of emergent team processes, theoretical has used static, cross-sectional approaches (Marks et al., 2001), advancements, and practical recommendations will be limited which unfortunately may not adequately address the inherently (see also Collins et al., 2016; Salas et al., 2017a,b). To address this dynamic nature of group potency. As such, the dynamic issue, the current investigation sought to examine: (1) how group aspects of group potency, which we expand on subsequently, potency, a critical emergent state, changes over time, (2) the have been relatively ignored by past research (Kozlowski and relation between the dynamics of potency and team effectiveness, Ilgen, 2006; cf. Collins and Parker, 2010; Collins et al., 2016; and (3) the mediating effect the dynamics of potency have on Salas et al., 2017a,b). There are two potential reasons for this: (1) the relation between inputs (i.e., team-level personality) and gathering longitudinal data with teams can be difficult because team effectiveness. team membership and/or project assignments may change over time (see McClurg et al., 2017), and (2) the analytical approaches In this research, data were gathered from student engineering for investigating emergence and growth had not developed project teams over multiple time points during an academic until recently (see Collins et al., 2016; Lang et al., 2018). course. We then used latent growth and consensus emergence In this research, we address these methodological challenges modeling to examine the dynamic nature and emergent and present a novel investigation into the dynamics of group properties of group potency. Throughout, we use the term potency over time. dynamic to reflect the separate factors of the initial starting point of teams’ potency, the rate of change in potency over EMERGENCE time, and also the emergence of the construct (see Ployhart and Vandenberg, 2010; Wang et al., 2016). Further, we The concept of emergence in multilevel phenomena (e.g., investigated the role of team-level input variables (i.e., team- teams) has been the focus of recent theoretical discussions (see level conscientiousness and extraversion) as predictors of the Kozlowski et al., 2013; Waller et al., 2016; Grossman et al., 2017). dynamicity of group potency. Additionally, we examined whether Here, we establish a theoretical model for the emergence and the dynamics of group potency mediated the relations for both dynamics involved with group potency. Kozlowski and Klein conscientiousness and extraversion on team effectiveness. (2000) defined an emergent state as a characteristic of a team that “is amplified by their interactions, and manifested as a In the following sections, we utilize conservation of resources higher-level, collective phenomenon” (p. 55). An emergent state, (COR) theory to discuss the importance of group potency therefore, is a dynamic construct, which theoretically changes as a team-level resource that influences team effectiveness, in or emerges over time (Kozlowski and Ilgen, 2006). We adopt accordance within the broad IPO and IMOI frameworks. In this as the basis for our investigation because it makes an addition, we invoke COR to support our theoretical rationale important distinction that other definitions do not address (e.g., for how potency changes over time, and how this change Marks et al., 2001). In Kozlowski and Klein (2000) definition, predicts team effectiveness. Then, we theorize that specific emergence is not a singular attribute; rather there are two personality traits (i.e., conscientiousness and extraversion) are distinct underlying processes that develop as a result of group both antecedents (i.e., inputs) and resources that contribute to interactions: (1) amplification, and (2) consensus. Amplification the process of group potency dynamics and the prediction of refers to the growth aspect, or in broader terms, reflects the team effectiveness. notion of changing levels over time, of a construct. Consensus refers to the emergence of a collective phenomenon from the GROUP POTENCY shared perceptions of individual members. Broadly speaking, the literature on emergent states has ignored the dynamic Group potency is one of the most frequently investigated nature of both amplification and consensus (Cronin et al., emergent states and team processes associated with effective 2011; Kozlowski et al., 2016). In particular, the vast majority of teamwork (LePine et al., 2008), and recent research suggest previous research has used cross-sectional data, which is poorly this trend is going to continue (e.g., O’Neill et al., 2016; Frontiers in Psychology | www.frontiersin.org 622 May 2019 | Volume 10 | Article 992

Woodley et al. The Emergence of Group Potency suited to examining the role time plays in both amplification Hypothesis 1: Perceptions of group potency will decrease and consensus processes (Cronin et al., 2011; Roe et al., 2012; over time. Vantilborgh et al., 2018). Emergent states should demonstrate changes in level and consensus over time, and result from team To be clear, we suggest that the downward trend of group interactions and collective experiences that lead to increasingly potency would be approximated well by a linear trajectory (see shared perceptions and consensus between individual members Ployhart and Vandenberg, 2010). Rather than a series of discrete (Kozlowski and Klein, 2000; Marks et al., 2001; Kozlowski et al., step-wise drops, or patterns of punctuated change, we anticipate 2013; Kozlowski, 2018). an incremental series of changes over time. Particularly, as teams meet on a set schedule during their lifecycle (i.e., three Group Potency Levels Across Time times a week during course and laboratory sessions) interacting with each other may lead to gradual changes in perceptions For group potency – and other emergent states – to develop, of group potency. Thus, rather than sudden, dramatic changes team members need time and a reason to interact and develop (i.e., discontinuous, non-linear change) in perceptions of group an understanding of “who they are” as a group (Marks et al., potency, teams will demonstrate a consistent, linear, downward 2001; Kozlowski, 2018). This suggests that potentially, at first, pattern over time. teams would be less confident in their ability to perform because they do not have enough experience with each other to Group Potency Over Time and Implications for Team develop a shared understanding of their collective ability. Then, Effectiveness conceivably, as team members interact over time they will gain To improve our understanding of the dynamic nature of group insight into each member’s work habits and abilities, leading potency, it is crucial to investigate its criterion-related validity to increases in collective confidence. This perspective, however, and examine how group potency relates to team effectiveness. rests on the assumption that team members enter teams without Meta-analytic research at both the individual- (e.g., Stajkovic any pre-existing expectations. It seems more likely that team and Luthans, 1998) and team-level (Gully et al., 2002; Stajkovic members enter their teams with high expectations, optimism, and et al., 2009) suggests strong, positive relations with performance. confidence, especially without evidence to suggest otherwise. In However, these results, as previously mentioned, are based on support of the latter, Allen and O’Neill (2015b) theorized that static research methods and do not take into consideration the early agreement they found among team members on ratings changes over time. of emergent states (e.g., group potency) might be attributed to an early positivity bias. They reasoned that this bias may lead to Within a time-limited project, group potency may function inflated perceptions of potency early in teams’ lifecycle, indicating as a team-level resource that takes time to coalesce through a strong need to consider the role of time in investigating team consensus, but can be drawn upon by the team to influence processes. Unfortunately, limited research has been conducted effectiveness and the achievement of team tasks and goals. on the dynamic nature of group potency. One study, however, According to the COR theory, resources play an important by Lester et al. (2002) measured group potency at two time role in understanding behavioral outcomes (e.g., performance; points, and using differences scores found that group potency Halbesleben and Bowler, 2007). Halbesleben et al. (2014) defined decreased over time. Although difference scores have several resources as “anything perceived by the individual to help attain methodological shortcomings (see Edwards, 2001, for a review), his or her goal” (p. 1338). Although defined at the individual level, this finding is not overly surprising. In fact, research on the this definition could easily be translated to the team context by “better-than-average” effect (e.g., Svenson, 1981) – a common defining a team resource as anything perceived by the members social comparison bias – would suggest that team members’ initial that can help the team attain its goal(s). This definition allows expectations of their team’s collective general ability might be group potency to be considered a team-level resource that can inflated. The better-than-average effect has also been found to be used to optimally influence team effectiveness (see Guzzo be stronger when the comparison target is ambiguous (Alicke et al., 1993; Gully et al., 2002; Stajkovic et al., 2009). In this et al., 1995), as in a newly formed team might be, and is positively light, there are two key components of COR to consider: (1) related to over-confidence in one’s individual ability (Larrick initial resource losses lead to future resource losses, and (2) a et al., 2007). It may therefore stand to reason that confidence in greater amount of a resource can reduce the vulnerability to one’s team may occur early in a team’s lifecycle. Yet, as members resource losses (Hobfoll, 2001, 2011), as in a buffering effect. may rate their team artificially high early on in their tenure Concerning initial resource loss, Hobfoll et al. (2018) argued that (Lester et al., 2002), scores will tend to decrease over time as resource loss begets stress, which leads to further resource loss. members interact with each other and face ongoing challenges In support of this theorizing, research by Demerouti et al. (2004) with the task that may reduce their potency resources that demonstrated that resource loss (due to work pressure) leads to are available for subsequent performance episodes. Continuing increased stress (i.e., work-life role conflict) in individuals, which interactions and experience with the task may facilitate more then leads to further resource loss (i.e., exhaustion). Demerouti realistic perceptions of how the team can reasonably be expected et al. (2004) referred to this phenomenon as a “loss spiral,” which to perform (i.e., a demonstrating a decreasing trend over has also been reported by De Cuyper et al. (2012) and Whitman time), in conjunction with increasing consensus across members. et al. (2014). Consistent with these findings, we anticipate that Together, this underscores the emergent and dynamic nature of teams that are unable to conserve their potency resources over potency. Based on this theorizing, the following is hypothesized: time will lose further resources over time, and experience worse Frontiers in Psychology | www.frontiersin.org 633 May 2019 | Volume 10 | Article 992

Woodley et al. The Emergence of Group Potency team effectiveness. Concerning the buffering effect, Hobfoll et al. used this methodology to provide an assessment of group potency (2018) argued that individuals who have more resources are less emergence over time. likely to lose resources and are more likely to gain resources. For example, Hakanen et al. (2008) found that individuals with As a collective phenomenon, group potency fits into Chan greater job resources were more engaged in their work, which (1998) referent-shift consensus model. Group potency, therefore, led to increased innovativeness in their work group. Chen et al. requires consensus amongst group members to demonstrate (2009) also found that by boosting individuals’ resources through the collective or shared aspect of the construct. Commensurate training, they were more likely to adapt to changing work with Kozlowski et al. (2013) theorizing on emergent processes, contexts and were less likely to experience resource loss (i.e., group members need time to interact with each other and exhaustion). We therefore propose that teams that start with engage with the task to develop a shared understanding of the higher potency (i.e., initially have more potency resources than team-level phenomenon. Initially, group members’ perceptions other teams) will perform better than teams that have lower initial of their potency will be based on minimal information potency. Together, we therefore hypothesize the following: as they have had limited time interacting. As a result, initial ratings of group potency will be more indicative of Hypothesis 2: Changes in group potency (i.e., the downward individual members’ perceptions rather than shared perceptions. trend described by Hypothesis 1) will be negatively related to It can therefore be theorized that agreement between group team effectiveness. members will increase over time. Accordingly, we forward the Hypothesis 3: Initial group potency will be positively related to following hypothesis: team effectiveness. Hypothesis 4: Consensus on group potency will increase Group Potency Consensus Over Time over time. Emergent states, as previously defined, describe the development ANTECEDENTS OF GROUP POTENCY’S of a collective phenomenon from the sharedness of individual DYNAMIC NATURE members’ perceptions of a team-level attribute. Emergent states therefore exist as constructs at the collective level According to the IPO framework, inputs play an important role (e.g., team, group, unit, and organization), underscoring in the development of team processes. Inputs are conditions their theoretical foundations based on differing composition or characteristics of team members that exist prior to the frameworks. Detailed considerations of composition models is team interacting and performing together, including – but not available elsewhere (e.g., Chan, 1998; Kozlowski and Klein, 2000); limited to – personality, and other dispositional characteristics. however, we note here that research on emergent states (e.g., Inputs can therefore be considered as antecedents to emergent group potency) requires that a level of consensus (i.e., agreement states, such as group potency. We selected conscientiousness or sharedness), which is based on a theoretically appropriate and extraversion as two input variables (i.e., resources) that composition model, be demonstrated. Emergent state research will contribute to group potency (i.e., a resource gain). Our has generally relied on rwg, intraclass correlations (ICCs), and rationale for selecting conscientiousness and extraversion is other agreement statistics (see LeBreton and Senter, 2008, for a two fold. First, meta-analytic research by Ng and Feldman review) as indices of consensus. Kozlowski et al. (2013) noted (2014), structured around COR theory, demonstrated that both that although these statistical approaches for assessing agreement conscientiousness, and extraversion contribute to resource gains have been used in both cross-sectional and longitudinal research (e.g., salary attainment). Second, meta-analytic research by Bell to demonstrate emergence, their use has predominantly been (2007) found that team-level conscientiousness and extraversion restricted to static interpretations (even when averaged across were positively related to team effectiveness (ρ = 0.14 and time in longitudinal research), and therefore ignores the temporal ρ = 0.10, respectively). Although the latter supports the direct aspect of emergence. More specifically, in both cross-sectional relation between our selected inputs and team effectiveness, there and longitudinal data, these consensus statistics have been is a dearth of research investigating the full IPO framework used to demonstrate that emergence has taken place, but and the implied indirect effects of how the inherently dynamic only provide a snapshot of sharedness, thereby ignoring the nature of team processes and resources (e.g., group potency) dynamicity of the emergence process. For example, in cross- transmit the effects of input resources to outputs. LePine et al. sectional research, after demonstrating some level of consensus, (2011) described the issues involved with this piecemeal approach researchers are left to assume a team-level phenomenon has of only assessing the input-output, or process-output relations, emerged, without actually assessing the pattern of change in for example, rather than a more theoretically aligned model of consensus that may more accurately represent the emergence input → process → output. Further, LePine et al. (2011) noted process (O’Neill and Allen, 2012; Allen and O’Neill, 2015a). that more advanced research designs and analyses should be Although this is informative from a descriptive standpoint, forwarded to improve understanding of the complete framework interpreting isolated ICC estimates may not provide a strict test of (see also Pitariu and Ployhart, 2010). Finally, Mathieu et al. whether emergence has occurred. To address this issue, Lang et al. (2014) pointed out that team personality composition might (2018) introduced the consensus emergence model, which allows not just be relevant for static teamwork variables but also their researchers to examine change in consensus over time, a key change over time. component of the emergence process. The current investigation Frontiers in Psychology | www.frontiersin.org 644 May 2019 | Volume 10 | Article 992

Woodley et al. The Emergence of Group Potency In the current research, we investigated the full IPO 2007), as it may facilitate positive interpersonal interactions framework by incorporating team-level conscientiousness and between team members (Barry and Stewart, 1997). Further, extraversion as inputs (i.e., antecedent resources), initial levels extraverts tend to have higher confidence in their ability to and change in group potency as process variables (i.e., team work in a self-managed group (Thoms et al., 1996), suggesting process resources), and team effectiveness as an output. Together, a positive relation between team-level extraversion and group indirect relations are described with group potency’s dynamics potency. Finally, extraversion involves facets related to energy, mediating the relations between team-level personality and activity, and excitement seeking (Hastings and O’Neill, 2009), all team effectiveness. of which would encourage strong willingness to engage in the work and exploration required for team success. Conscientiousness Similar to team-level conscientiousness, team-level Individuals with high conscientiousness are characterized extraversion can be considered a resource that is brought by being hardworking and achievement-oriented (Goldberg, to the team by its individual members and functions as an 1990). Further, conscientious individuals tend to be confident input for team processes (i.e., group potency). Thus, considering (Chen et al., 2004; Ebstrup et al., 2011), and likely behave team-level extraversion as a team resource, it may lead to in a manner that is conducive to operating in a team increased initial group potency and help teams preserve their environment (e.g., O’Neill and Allen, 2011). Even further, as group potency over time. This will permit teams to conserve and noted, Bell’s (2007) meta-analysis found that team-level mean maintain their potency resources during its lifecycle, potentially conscientiousness was positively related to team performance. leading to increased team effectiveness. Based on this theorizing, Thus, past research has illustrated positive relations between the following is hypothesized: team-level conscientiousness and both group potency and team effectiveness. Hypothesis 6a: The initial level of group potency will mediate the relation between extraversion and team effectiveness. We again draw upon COR theory, and apply a resource- Hypothesis 6b: The rate of change of group potency based perspective to propose how team-level conscientiousness will mediate the relation between extraversion and relates to the dynamics of group potency and team effectiveness. team effectiveness. Another key proposition of COR is that initial resources can combine to positively influence the achievement of desired MATERIALS AND METHODS outcomes, and can help produce gains in resources, or alternatively, can provide additional resources to help maintain Participants and Procedure resources levels that may otherwise become depleted over time. Hobfoll (2011) argued that resources should be considered as This study was reviewed and approved by Western University’s “caravans,” in which the combined functioning of resources Non-Medical Research Ethics Board and participants provided best facilitates achieving desired outcomes (e.g., meeting goals, written informed consent prior to participating. Participants were coping with stress). Based on the importance of team-level 337 first-year engineering students. The majority of participants conscientiousness, we argue that team-level conscientiousness (81%) were male, and ranged in age from 16 to 33 years (M = 18.5, can function as a team “input” resource that can lead to gains SD = 1.9). Participants were randomly assigned to one of 77 in (i.e., higher) initial group potency. For instance, groups that project teams, which consisted of either four (62% of teams) or see themselves as more collectively hard working will likely see five (38%) members. Each team had two small design projects themselves as having higher initial confidence in their ability to (taking place over 2 months each) and one large design project achieve the team’s goals, because they know they will persist even (taking place over 4 months) to complete over the course of when the task difficulty increases. In addition, increased team- an academic year. For the large design project, students were level conscientiousness may provide another resource to the team required to create a prototype of a device that individuals with to protect against loss of potency resources over time. Thus, a disability could use to improve their well-being. teams with higher levels of conscientiousness will be able to better conserve their potency resources over time. This, in turn, will lead Survey data were collected at five different time points to increased team effectiveness. Thus: throughout the academic year. Conscientiousness and extraversion data was collected on the first day of class Hypothesis 5a: The initial level of group potency will before students were assigned into their project teams (i.e., Time mediate the relation between conscientiousness and 1). Group potency data was collected at three subsequent time team effectiveness. points: 2 months (Time 2), 5 months (Time 3), and 8 months Hypothesis 5b: The rate of change of group potency (Time 4) after the start of the semester. Grades on the large will mediate the relation between conscientiousness and design project were collected at the conclusion of the semester team effectiveness. (i.e., Time 5) and serve as our measure of team effectiveness. Extraversion Measures Highly extraverted individuals tend to be talkative and sociable Conscientiousness (Goldberg, 1990). Research on team-level extraversion has Conscientiousness was measured with ten items from the generally revealed positive relations with team performance (Bell, International Personality Item Pool (IPIP; Goldberg et al., 2006; Frontiers in Psychology | www.frontiersin.org 655 May 2019 | Volume 10 | Article 992

Woodley et al. The Emergence of Group Potency α = 0.81). The IPIP items correlate highly with Costa and structure and function of a scale (McLarnon and Carswell, 2013). McCrae’s (1992) NEO-PI-R. There were five positively worded The configural invariance model assesses whether the same and five negatively worded items. A sample item is “I am pattern of factor loadings holds over time. For determining always prepared.” Participants responded to these items on a configural invariance, we – in part – assumed support because all five-point Likert-type agreement scale (1, strongly disagree; 5, seven potency items, which measure a single factor, were assessed strongly agree). at each time point. In addition, we also considered indicators of model-data fit rendered by the comparative fit index (CFI) Extraversion and root mean square error of approximation (RMSEA). CFI Extraversion was also measured with ten items from the IPIP values > 0.95 and RMSEA values <0.08 can be taken as evidence (Goldberg et al., 2006; α = 0.86) that correlate highly with for acceptable model fit (e.g., Hu and Bentler, 1999). Building the NEO-PI-R. There were five positively worded and five on the configural invariance model, metric invariance then negatively worded items. A sample item is “I feel comfortable constrains respective factor loadings to equality, scalar invariance around people.” Participants responded to these items on a places additional equality constraints on respective intercepts, five-point Likert-type agreement scale (1, strongly disagree; 5, and strict invariance places equality constraints on respective strongly agree). item residuals. To assess plausibility of each of these sets of invariance constraints, the χ2 test can be used because each Group Potency set of constraints imposed represent a nested model. However, Group potency was measured with seven items from Guzzo et al. as χ2 may be overly sensitive to sample size, changes in the (1993), which measure a team’s confidence in their general ability CFI of less than 0.010 and/or changes in the RMSEA of less than to be effective. A sample item is “No task is too tough for this 0.015 can support invariance in each step (Chen, 2007). In each team.” Participants responded to these items on a five-point longitudinal invariance analysis, autocorrelated residuals were Likert-type agreement scale (1, strongly disagree; 5, strongly specified between respective items (Little, 2013). agree). Sosik et al. (1997) found that these group potency items have strong internal consistency with a Cronbach’s α ranging Our invariance analyses used individual-level data in order to from 0.87 to 0.98 across three time points. achieve a balance between sample size and model complexity. However, to account for the nested nature of our data Team Effectiveness (i.e., individuals within teams), we used robust maximum Associated with the large design project, teams submitted a likelihood estimation, implemented as Mplus’ MLR estimator, comprehensive written report that was typically about 100 in conjunction with the TYPE = COMPLEX specification to pages in length. The report contained a variety of detailed furnish model fit indices and standard errors that were robust to information pertaining to the project including, design sketches, non-independence (Muthén and Muthén, 2012; McNeish et al., mathematical models, and implications for practice. Team 2017). Given the use of the MLR estimator, χ2 nested model reports were rated based on their overall quality by experienced comparisons were facilitated through Satorra and Bentler’s (2001) course instructors, who were blind to this study’s objectives, and scaled χ2 statistic. grades were assigned to the team as a whole (i.e., no unique grades were assigned to individual members). Each rater rated a unique An additional wrinkle in estimating the longitudinal subset of the reports (see O’Neill et al., 2018). invariance models concerns the correct specification of the longitudinal null model (Little, 2013), which is used in the Analytical Procedure derivation of the CFI. If the null model is incorrect, the CFIs Using Mplus 7.4 (Muthén and Muthén, 2012, 2015) throughout used to judge invariance may also be biased and may result in for our focal analyses, we implemented a sequential model erroneous inferences. As discussed by Widaman and Thompson testing procedure to conduct (1) longitudinal measurement (2003), the correct longitudinal null model should specify zero invariance analyses, (2) latent growth modeling, and (3) covariances between any indicators (as in the typical null model), consensus emergence modeling (Lang et al., 2018). The full model but equal variances and equal means for respective indicators assessed is illustrated in Figure 1. Examinations of change over across time points. As such, our use of the CFI was based on the time requires measurement invariance to ensure that a measure corrected longitudinal null model. functions and means the same thing over time, and to facilitate meaningful longitudinal inferences (Ployhart and Vandenberg, Then, using latent growth modeling (Chan, 2002), and the 2010). Longitudinal measurement invariance assesses the stability aggregated potency scores, we examined the dynamics involved of a scale’s measurement model over time, and without with group potency. First, we estimated an unconditional model this support misleading interpretations may result, akin to to estimate the mean and variability around the latent intercept comparing apples to oranges over time (Chen and West, 2008). and slope of group potency. The latent growth model was Demonstrating invariance requires several analytical steps, which specified in a typical fashion with the factor loadings for the include: (a) configural invariance, (b) metric invariance, (c) scalar latent intercepts all fixed at 1.00, and the factor loadings for the invariance, and (d) strict invariance. Ployhart and Vandenberg latent slope were fixed at zero, 1.00, and 2.00, for each of the (2010) noted that configural, metric, and scalar invariance are measures (i.e., Time 2, 3, and 4; see above), respectively. The sufficient for longitudinal invariance, yet strict invariance was parameterization for the slope follows from equal time spacing also investigated as it can provide additional insight into the between Times 2 and 3, and Times 3 and 4, as both reflected 3-month time lags. We then incorporated team effectiveness, as a simultaneous outcome of both the latent intercept and Frontiers in Psychology | www.frontiersin.org 666 May 2019 | Volume 10 | Article 992

Woodley et al. The Emergence of Group Potency FIGURE 1 | Focal analytical model. Numeric factor loadings for LGM presented. Direct effects, c’paths, and shown in dashed lines. Indirect effects, comprising respective a and b paths and associated aj × bj effects, and shown in solid lines. slope, and the personality predictors to assess the indirect effects. demonstrated adequate fit, CFI = 0.95 and RMSEA = 0.06. Using bias-corrected bootstrapping, with 10,000 samples, indirect Adding equality constraints on the factor loadings resulted in effects were deemed significant if their 95% confidence intervals (CIs) excluded zero. Notably, the personality predictors used the CFI = –0.001 and RMSEA = –0.002, supporting metric mean-aggregation of scores from each individual member and invariance. This suggests that the potency measure retains a as mean-aggregated personality is not a shared-unit property of similar meaning across occasions. The scalar invariance model a team (Kozlowski and Klein, 2000) justifying aggregation (via resulted in a CFI = –0.003 and RMSEA < 0.0004 versus the ICCs, etc) is therefore not required (e.g., O’Neill and Allen, 2011). metric invariance model. This lends support to scalar invariance, which suggests that the potency measure functions similarly Finally, we used Lang et al.’s (2018) multilevel procedure to over time. As a final stage in the invariance analyses, additional examine consensus emergence of group potency. This allowed us equality constraints were placed on respective item residuals to to assess emergence of the group-level potency construct from assess strict invariance. This model resulted in CFI = 0.003 and the sharedness, or more specifically the increasing degree of sharedness, of individual members’ ratings over time. RMSEA = –0.004, supporting strict invariance, and suggests that each item had equivalent reliability over time. Together, RESULTS these invariance analyses suggest equivalence of group potency over time, facilitating our focal latent growth models. Table 1 presents the team-level correlation matrix, the intraclass correlation estimates [ICC(1) and ICC(2)] for group potency Given the ICCs provided support for aggregating group at each time point, and Cronbach’s alpha internal consistency potency to the team-level, we averaged individual members’ estimates. Notably, the ICC estimates increased slightly over group potency scores within each team, and used the aggregated time, indicating a growing proportion of variance in group scores to estimate our latent growth model. The unconditional potency that could be attributed to the team-level rather growth model demonstrated adequate fit to the data, χ2(1) = 0.73, than the individual-level. This suggests increasing consensus in p = 0.39, CFI = 1.00, RMSEA = 0.00. With respect to Hypothesis perceptions of team potency over time and stronger emergence. 1, the mean of the latent slope was of central interest, which We revisit this pattern to more formally substantiate the was estimated as -0.07, p < 0.05. This supported Hypothesis emergence of group potency and provide a test of Hypothesis 4. 1, suggesting that group potency decreased over time (by 0.07 units at each time point). The estimate of the latent intercept Table 2 presents the results of the longitudinal measurement was 4.06, p < 0.01, and the variances for the latent intercept invariance analyses. The configural invariance model and slope were 0.20, p < 0.01, and 0.04, p < 0.05, respectively. The correlation between the latent intercept and slope was -0.14, p = 0.59. Interestingly, freeing the slope’s factor loading for the Frontiers in Psychology | www.frontiersin.org 677 May 2019 | Volume 10 | Article 992

Woodley et al. The Emergence of Group Potency TABLE 1 | Team-level Descriptives and Intercorrelations. M SD ICC(1) ICC(2) 1 2 3 4 56 1. Conscientiousness 3.66 0.26 – – (0.73) (0.77) (0.90) (0.93) (0.94) – 2. Extraversion 3.49 0.26 – – –0.00 0.22 0.41∗ ∗ 0.62∗ ∗ 0.32∗ ∗ 3. Potency, Time 1 4.06 0.37 0.27 0.60 0.13 –0.06 0.30∗ ∗ 0.35∗ ∗ 4. Potency, Time 2 4.06 0.52 0.35 0.65 0.24∗ 0.00 0.14 5. Potency, Time 3 3.97 0.62 0.37 0.67 0.30∗ ∗ –0.16 6. Team effectiveness 82.28 11.01 – – 0.14 n = 77. ICCs not applicable to conscientiousness, extraversion, or team effectiveness measures. Individual-level Cronbach’s α estimates given in parentheses on diagonal; Team Effectiveness was a single score (grade), therefore reliability could not be estimated. ∗∗p < 0.01, ∗p < 0.05. TABLE 2 | Longitudinal measurement invariance analyses. χ2 χ2c df #fp CFI RMSEA χ2 χ2 df CFI RMSEA Configural 355.46∗ 1.22 165 87 0.95 0.06 – – – – Metric 370.09∗ 1.21 177 75 0.95 0.06 13.61 12 –0.001 –0.002 Scalar 392.72∗ 1.20 189 63 0.95 0.06 21.12∗ 12 –0.003 +0.001 Strict 394.32∗ 1.27 203 49 0.95 0.06 13.75 14 0.003 –0.004 χ2c, scaling correction factor for χ2; df, degrees of freedom; #fp, number of parameters estimated; CFI, comparative fit index (calculated using corrected longitudinal null model; Widaman and Thompson, 2003); RMSEA, root mean square error of approximation; χ, Satorra-Bentler scaled χ2 difference statistic (Satorra and Bentler, 2001); χ2 df, degrees of freedom for Satorra-Bentler χ2; CFI, RMSEA, change in CFI and RMSEA estimates, respectively, from less restricted to more restricted models (i.e., change in CFI from configural invariance model to metric invariance model). ∗p < 0.05. second group potency measure, as in a latent basis model (Grimm significant increases in the support for emergence over the three et al., 2013) did not suggest an improvement in fit. Specifically, measurement occasions. Thus, Hypothesis 4 was supported. χ2(1) = 0.71, p = 0.39, and both the Akaike Information Criteria Finally, we incorporated the conscientiousness and and Bayesian Information Criteria were higher in the latent basis extraversion team-level predictors into the latent growth model than the latent growth model. Thus, based on parsimony, model. This also resulted acceptable model-data fit: χ2 (4) = 3.02, we proceed with the linear latent growth model. Notably, even in p = 0.55, CFI = 1.00, RMSEA = 0.00. Neither of the indirect the latent basis model, the trend did not deviate significantly from effects involving the latent slope had 95% CIs that excluded a linear trajectory, thus lending further credibility to Hypothesis zero: the conscientiousness → latent slope → team effectiveness 1, and the underlying linear, downward pattern of change indirect effect was -0.01, 95% CI = –3.11–2.65, and the in group potency. Next, incorporating team effectiveness as a extraversion → latent slope → team effectiveness indirect effect simultaneous outcome of the latent intercept and slope factors was 0.08, 95% CI = –2.36–4.55. The indirect effect involving also resulted in adequate model-data fit: χ2 (2) = 1.22, p = 0.54, extraversion → latent intercept → team effectiveness was also CFI = 1.00, RMSEA = 0.00. Specifying regressions between both not significant, -0.73, 95% CI = –9.34–4.65. However, the indirect intercept and slope factors and effectiveness revealed that the effect of conscientiousness → latent intercept → effectiveness regression of effectiveness on the latent slope was b = 0.07, was significant, 5.57, 95% CI = 0.59–23.58. Thus, there was no p = 0.99, but that for the latent intercept it was b = 10.23, evidence for the mediating role for change in potency, but instead p < 0.01. Thus, there was no influence of change in potency on the latent intercept transmitted the effect of conscientiousness team effectiveness, but the starting point of teams’ potency was on team effectiveness. In sum, Hypothesis 5a was supported, but positively related to effectiveness. Accordingly, Hypothesis 2 was Hypotheses 5b, 6a, and 6b did not receive support. not supported, whereas Hypothesis 3 was supported. DISCUSSION To more formally assess the emergence of the group potency construct, we used Lang et al.’s (2018) consensus emergence There are four intriguing findings from the current investigation model. This model uses longitudinal changes in the individual- that contribute to both the group potency and the multilevel level residual variances as evidence of emerging consensus. emergence literatures. First, the latent growth model revealed Specifically, decreasing residual variances can be taken as a significant negative slope for group potency. Group potency indicative of increasing consensus emergence, and therefore levels therefore decreased over time, on average across teams. reflects more agreement about a team-level phenomenon. Previous research by Lester et al. (2002) also found a decrease Indeed, in our model the estimated change in residual variance in group potency over time; however, that study had only two was δ = –0.11, p < 0.05. This suggests significantly less time points and a much shorter time span in comparison to the individual-level variance and comparably greater sharedness at current investigation (i.e., 9 weeks vs. 6 months, respectively). We the team-level over time. In other words, this negative coefficient supports the proposition that group potency demonstrated Frontiers in Psychology | www.frontiersin.org 688 May 2019 | Volume 10 | Article 992

Woodley et al. The Emergence of Group Potency theorized that individuals would generally tend to start with high inference of the actual process of consensus emergence, which expectations of how their team would perform (Svenson, 1981). is temporally defined. Using Lang et al.’s (2018) methodology, As well, due to the “better-than-average” effect when teams we were able to utilize an analytical approach that is sensitive to first get together they may experience a “honeymoon period” emergence’s inherently temporal nature and provide an empirical where they have unrealistic positive expectations of how they estimate of group potency’s emergence. Commensurate with will do as a group (Forsyth, 2018). Over time, it is probable Allen and O’Neill (2015b), we found support for early emergence, that the honeymoon dissolves as team members spend more with Time 1 ICCs meeting acceptable levels of agreement time interacting, debating, dealing with internal conflicts, and (LeBreton and Senter, 2008). Nevertheless, our findings also other challenges associated with teamwork and the team task suggest that agreement still increased over longer durations as (O’Neill and McLarnon, 2018; O’Neill et al., 2018). In this study, team members interact and get a better understanding of “who we drew upon COR theory to argue that these challenges they are” as a collective. negatively affect team resources (e.g., group potency), resulting in a decrease in magnitude over time. Interestingly, similar results Although the findings of decreasing group potency levels have been found in other research domains. For example, in and increasing consensus on group potency may seem in examining changes in organizational commitment, an integral opposition, these are independent phenomena. Conceivably, workplace resource, Lance et al. (2000) and Bentein and Meyer consensus could emerge over any level of a construct, which (2004) found that organizational newcomers experienced loss could be static or dynamic in nature. Future research may of this resource over time as they interacted with their new be able to leverage Lang et al.’s (2018) framework and settings. Our results, and those from the domain of organizational incorporate predictors of emergence, such as relationship commitment, therefore support the argument that resources and process conflict (O’Neill et al., 2018), psychological can be depleted over time as individuals interact with their safety (Edmondson, 1999), intrateam communication, and peer environment, whether the environmental context is a workplace feedback (Donia et al., 2018), among others. or a team. This suggests that early team experiences (i.e., socialization) are important for establishing strong, initial group The fourth important finding reflects the application of the potency resources. IPO framework to test key COR principles. More specifically, two input resources – conscientiousness and extraversion – were This paved the way for the second intriguing finding from included as antecedents of group potency’s dynamic nature. We this study: in the latent growth model, teams’ initial group found that the relation between conscientiousness and team potency predicted overall team effectiveness. This implies that, effectiveness was mediated by initial group potency. Contrary although group potency takes time to emerge (which we discuss to our expectations, no effect was found for extraversion, or subsequently), early interactions might play an important role for the link between conscientiousness and team effectiveness, in setting a team up for future success. Although teams may as mediated by the rate of change in potency level. These have elevated potency ratings during a honeymoon period, they findings suggest that teams that comprise individuals with are still able to effectively leverage their potency resources, higher levels of conscientiousness are more likely to get such that it helps explain teams’ effectiveness later on during off to a “good start,” and utilize their collective personality project completion (i.e., 6 months later). This finding supports composition as a resource to develop higher levels of initial Kozlowski et al. (2013) argument that it is important to assess group potency (another resource), thereby leading to greater emergent states as early in a team’s lifecycle as possible. Even team effectiveness. though group potency resources may decrease over time, early potency, and the intrateam resources it provides, may have a role Practical Implications in determining future strategizing, planning, and cooperation, which helps to set the stage for the future goal and task Stemming from these results, an important practical implication accomplishment. Thus, despite the decreasing trend experienced is that early team interactions need to be managed effectively to by teams over time, what appears to be an important component enable a strong starting point for teams’ group potency. With of a team’s effectiveness is each team’s perception of potency early an emphasis on early group potency, rather than the change on in their respective lifecycle. in potency over time, teams may be able to leverage initial potency as a critical team resource and more effectively navigate The third intriguing contribution that this research provides hurdles encountered during project completion. Nevertheless, is that we documented an increase in consensus on group future research may want to also consider how the potentially potency within teams. Thus, members gained an increasingly negative effects of overconfidence (Goncalo et al., 2010) can shared perception of their group’s potency over time. This is be mitigated with early team experiences such as developing an important aspect of what Kozlowski et al. (2013) described a team charter, engaging in informal socialization, and other generally as exemplifying the multilevel emergence process: as activities that may assist in developing a healthy level of team members interact they will develop a stronger, shared early group potency. understanding of the team’s emergent properties (e.g., group potency). Historically, “sharedness” or consensus has only been A second practical implication is that interteam investigated using cross-sectional analyses, and inferred via ICC differences in personality composition play an important estimates, with “high values” taken to support the occurrence role in developing early group potency. We found that of emergence. This approach, however, does not facilitate an teams that had members with higher conscientiousness were more likely to develop group potency early on, leading to increased team effectiveness. Drawing from Frontiers in Psychology | www.frontiersin.org 699 May 2019 | Volume 10 | Article 992

Woodley et al. The Emergence of Group Potency an integration of COR theory and the IPO framework, with a larger number of more heterogeneous teams conscientiousness is an important resource that sets the engaged in alternative projects that take place over longer stage for teams’ early potency, which reflects another (or shorter) time periods and lifecycles. Such research critical team resource that, in turn, influences effectiveness. endeavors may highlight alternative forms of group Therefore, teams can utilize the resources made available by potency change over time (i.e., non-linear, discontinuous). their aggregated level of conscientiousness to establish and However, we would still likely anticipate consensus to develop group potency allowing them to be more effective. emerge and solidify over time, though it may taper off Thus, it is important to consider personality traits, like during longer lifecycles. Though we have substantiated conscientiousness, when selecting members for a team (see a linear, downward trend in group potency over time, Allen and West, 2005; Morgeson et al., 2005; O’Neill and Allen, it may also be interesting to examine whether distinct 2011; Allen and O’Neill, 2015a). types of teams occupy differential trajectories of group potency dynamics. Specifically, leveraging growth mixture Limitations modeling, future researchers could examine nuanced trajectories of potency that may be illustrated by distinct One of the limitations of the current study is the use of a types of teams (Muthén, 2001; McLarnon and O’Neill, 2018; student sample that, on average, was relatively young (18.5 years O’Neill et al., 2018). old). As well, the participants were predominantly male. It is therefore somewhat difficult to generalize the current findings Additionally, future research could be dedicated toward to more heterogeneous work environments. Furthermore, our whether similar emergent states (e.g., collective efficacy) results may only apply to time- and project-limited teams. Teams may exhibit differential patterns of change over time. For that are tasked with multiple performance cycles may experience instance, collective efficacy, as previously mentioned, is an a different form of change in potency over time, during the emergent state that represents a group’s confidence in their completion of their projects. Future research will be needed to ability to perform a specific task, rather than the general assess the form and function of potency in alternative types of ability to perform that is measured by group potency. teamwork, which may also facilitate insight into Marks et al.’s Thus, as teams engaged in a specific task (e.g., a product (2001) recommendation to investigate multiphasic perspectives development initiative), they could experience increasing on team processes. collective efficacy as they gain task-specific knowledge, and expertise through practice – similar to how training A second limitation concerns the ability to apply these can increase self-efficacy (Blume et al., 2010) – while also results to the dynamic nature that is exemplified by other experiencing decreasing group potency as they recognize emergent states. Specifically, the dynamics of potency may vary how challenging it can be to effectively function as a team, in from the growth inherent with other emergent states (e.g., general. Nonetheless, we believe the current research provides cohesion). Although potency may emerge after a relatively substantial value to the literature, and our methodological short duration, and then decline over time, cohesion (i.e., a approach may assist future studies, which we eagerly motivational force that drives teams to stay together) may take await so as to equip the literature with a comprehensive longer to emerge as teams take time to decide whether they understanding of form, function, predictors, and implications want to stay together. Thus, future research should be conducted of group potency. using similar research methods and analytical procedures (i.e., latent growth modeling, paired with Lang et al.’s (2018) CONCLUSION consensus emergence model) to investigate the dynamics of other emergent states. The current investigation improves our understanding of the dynamic aspect of group potency. Results demonstrated A third limitation is that the measures marking the that potency decreased over time, which we attributed beginning of the potency growth trajectories were collected to a honeymoon period associated with a team’s early 2 months into teams’ lifecycle. This timeframe was selected interactions. Further, teams tended to agree more on their because members had limited time to interact over the team’s potency over time, suggesting that it takes time for first 2 months, but would have still been representative of the group potency construct to emerge. Even further, early teams’ “honeymoon” levels of potency, as they had yet to group potency predicted team effectiveness, however, the receive any substantial feedback on their team effectiveness. change in group potency did not. This suggests that early The results of this study, however, demonstrate that interactions play an important role in establishing group potency had already begun to emerge by the beginning potency, which may emerge relatively quickly, and may set of the trajectory. Future research should measure and the tone for future success. Finally, initial group potency examine potency even earlier on in a team’s inception (see mediated the relation between team-level conscientiousness Kozlowski et al., 2013). and team effectiveness, suggesting that conscientiousness plays an important role in influencing the dynamics of Directions for Future Research group potency, which subsequently leads to increased team effectiveness. Although this research presents several unique and valuable contributions to the literature, there are a number of crucial questions future research should investigate. First and foremost is cross-validation of these findings Frontiers in Psychology | www.frontiersin.org 1700 May 2019 | Volume 10 | Article 992

Woodley et al. The Emergence of Group Potency ETHICS STATEMENT data collection, and provided the comments throughout the manuscript. This study was carried out following the protocol approved by the university’s Non-Medical Research Ethics Board, which in FUNDING accordance with the Declaration of Helsinki, all subjects provided written informed consent prior to participating. This work was supported by a grant from the Social Sciences and Humanities Research Council of Canada. AUTHOR CONTRIBUTIONS HW developed the study idea, coordinated the data ACKNOWLEDGMENTS management, and wrote sections of the manuscript. MM analyzed the data and wrote sections of the manuscript. The authors would like to thank Dr. Natalie Allen for her TO’N developed the study materials, coordinated the comments and feedback on a previous version of this manuscript. REFERENCES Collins, C. G., Gibson, C. B., Quigley, N. R., and Parker, S. K. (2016). Unpacking team dynamics with growth modeling: an approach to test, refine, and integrate Alicke, M. D., Klotz, M. L., Breitenbecher, D. L., Yurak, T. J., and Vredenburg, theory. Organ. Psychol. Rev. 6, 63–91. doi: 10.1177/2041386614561249 D. S. (1995). Personal contact, individuation, and the better-than-average effect. J. Personal. Soc. Psychol. 68, 804–825. doi: 10.1037//0022-3514.68.5.804 Collins, C. G., and Parker, S. K. (2010). Team capability beliefs over time: distinguishing between team potency, team outcome efficacy, and team Allen, N. J., and O’Neill, T. A. (2015a). “Team composition and performance: process efficacy. J. Occup. Organ. Psychol. 83, 1003–1023. doi: 10.1348/ considering the project team challenge,” in The Psychology and Management 096317909X484271 of Project Teams: An Interdisciplinary Perspective, eds F. Chiocchio, E. K. Kelloway, and B. Hobbs (Oxford: Oxford University Press), 301–328. Costa, P. T., and McCrae, R. R. (1992). Normal personality assessment in clinical practice: the NEO personality inventory. Psychol. Assess. 4, 5–13. doi: 10.1037/ Allen, N. J., and O’Neill, T. A. (2015b). The trajectory of emergence of /1040-3590.4.1.5 shared group-level constructs. Small Group Res. 46, 352–390. doi: 10.1177/ 1046496415584973 Cronin, M. A., Weingart, L. R., and Todorova, G. (2011). Dynamics in groups: are we there yet? Acad. Manage. Ann. 5, 571–612. doi: 10.1080/19416520.2011. Allen, N. J., and West, M. A. (2005). “Selection for teams,” in Handbook of Personnel 590297 Selection, eds A. Evers, N. Anderson, and O. Voskuijl (Oxford: Blackwell Publishing), 476–494. De Cuyper, N., Mäkikangas, A., Kinnunen, U., Mauno, S., and De Witte, H. (2012). Cross-lagged associations between perceived external employability, Barry, B., and Stewart, G. L. (1997). Composition, process, and performance in job insecurity, and exhaustion: testing gain and loss spirals according to the self-managed groups: the role of personality. J. Appl. Psychol. 82, 62–78. doi: conservation of resources theory. J. Organ. Behav. 33, 770–788. doi: 10.1002/ 10.1037/0021-9010.82.1.62 job.1800 Beal, D. J., Cohen, R. R., Burke, M. J., and McLendon, C. L. (2003). Cohesion Demerouti, E., Bakker, A. B., and Bulters, A. J. (2004). The loss spiral of work and performance in groups: a meta-analytic clarification of construct relations. pressure, work–home interference and exhaustion: reciprocal relations in a J. Appl. Psychol. 88, 989–1004. doi: 10.1037/0021-9010.88.6.989 three-wave study. J. Vocat. Behav. 64, 131–149. doi: 10.1016/S0001-8791(03) 00030-7 Bell, S. T. (2007). Deep-level composition variables as predictors of team performance: a meta-analysis. J. Appl. Psychol. 92, 595–615. doi: 10.1037/0021- Donia, M. B. L., O’Neill, T. A., and Brutus, S. (2018). The longitudinal effects of peer 9010.92.3.595 feedback in the development and transfer of student teamwork skills. Learn. Ind. Diff. 61, 87–98. doi: 10.1016/j.lindif.2017.11.012 Bentein, K., and Meyer, J. P. (2004). doi: 10.1037/0021-9010.92.3.595 Blume, B. D., Ford, J. K., Baldwin, T. T., and Huang, J. L. (2010). Transfer Ebstrup, J. F., Eplov, L. F., Pisinger, C., and Jørgensen, T. (2011). Association between the five factor personality traits and perceived stress: is the of training: a meta-analytic review. J. Manage. 36, 1065–1105. doi: 10.1177/ effect mediated by general self-efficacy? Anxiety Stress Coping 24, 407–419. 0149206309352880 doi: 10.1080/10615806.2010.540012 Chan, D. (1998). Functional relations among constructs in the same content domain at different levels of analysis: a typology of composition models. J. Appl. Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Psychol. 83, 234–246. doi: 10.1037//0021-9010.83.2.234 Admin. Sci. Quart. 44, 350–383. doi: 10.2307/2666999 Chan, D. (2002). “Latent growth modeling,” in Measuring and Analyzing Behavior in Organizations: Advances in Measurement and Data Analysis, eds F. Drasgow Edwards, J. R. (2001). Ten difference score myths. Organ. Res. Methods 4, 265–287. and N. Schmitt (San Francisco, CA: Jossey-Bass), 302–349. doi: 10.1177/109442810143005 Chen, F. F. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Struct. Equat. Model. 14, 464–504. doi: 10.1080/107055107013 Forsyth, D. R. (2018). Group Dynamics. Boston, MA: Cengage Learning. doi: 10. 01834 1177/109442810143005 Chen, F. F., and West, S. G. (2008). Measuring individualism and collectivism: the importance of considering differential components, reference groups, and Goldberg, L. R. (1990). An alternative “description of personality”: the big-five measurement invariance. J. Res. Personal. 42, 259–294. doi: 10.1016/j.jrp.2007. factor structure. J. Personal. Soc. Psychol. 59, 1216–1229. doi: 10.1037/0022- 05.006 3514.59.6.1216 Chen, G., Gully, S. M., and Eden, D. (2004). General self-efficacy and self-esteem: toward theoretical and empirical distinction between Goldberg, L. R., Johnson, J. A., Eber, H. W., Hogan, R., Ashton, M. C., Cloninger, correlated self-evaluations. J. Organ. Behav. 25, 375–395. doi: 10.1002/ C. R., et al. (2006). The international personality item pool and the future of job.251 public-domain personality measures. J. Res. Personal. 40, 84–96. doi: 10.1016/j. Chen, S., Westman, M., and Eden, D. (2009). Impact of enhanced resources on jrp.2005.08.007 anticipatory stress and adjustment to new information technology: a field- experimental test of conservation of resources theory. J. Occup. Health Psychol. Goncalo, J. A., Polman, E., and Maslach, C. (2010). Can confidence come 14:219. doi: 10.1037/a0015282 too soon? Collective efficacy, conflict and group performance over time. Organ. Behav. Hum. Decis. Process. 113, 13–24. doi: 10.1016/j.obhdp.2010. 05.001 Grimm, K. J., Steele, J. S., Ram, N., and Nesselroade, J. R. (2013). Exploratory latent growth models in the structural equation modeling framework. Struct. Equat. Model. 20, 568–591. doi: 10.1080/10705511.2013.824775 Frontiers in Psychology | www.frontiersin.org 1711 May 2019 | Volume 10 | Article 992

Woodley et al. The Emergence of Group Potency Grossman, R., Friedman, S. B., and Kalra, S. (2017). “Teamwork processes and LeBreton, J. M., and Senter, J. L. (2008). Answers to 20 questions about emergent states,” in The Wiley Blackwell Handbook of the Psychology of Team interrater reliability and interrater agreement. Organ. Res. Methods 11, 815–852. Working and Collaborative Processes, eds E. Salas, R. Rico, and J. Passmore doi: 10.1177/1094428106296642 (Sussex: John Wiley & Sons), 245–269. LePine, J. A., Buckman, B. R., Crawford, E. R., and Methot, J. R. (2011). A review Gully, S. M., Incalcaterra, K. A., Joshi, A., and Beaubien, J. M. (2002). A meta- of research on personality in teams: accounting for pathways spanning levels analysis of team-efficacy, potency, and performance: interdependence and of theory and analysis. Hum. Resour. Manage. Rev. 21, 311–330. doi: 10.1016/j. level of analysis as moderators of observed relationships. J. Appl. Psychol. 87, hrmr.2010.10.004 819–832. doi: 10.1037/0021-9010.87.5.819 LePine, J. A., Piccolo, R. F., Jackson, C. L., Mathieu, J. E., and Saul, J. R. (2008). Guzzo, R. A., Yost, P. R., Campbell, R. J., and Shea, G. P. (1993). Potency in groups: A meta-analysis of teamwork processes: tests of a multidimensional model articulating a construct. Br. J. Soc. Psychol. 32, 87–106. doi: 10.1111/j.2044-8309. and relationships with team effectiveness criteria. Pers. Psychol. 61, 273–307. 1993.tb00987.x doi: 10.1111/j.1744-6570.2008.00114.x Hakanen, J. J., Perhoniemi, R., and Toppinen-Tanner, S. (2008). Positive gain Lester, S. W., Meglino, B. M., and Korsgaard, M. A. (2002). The antecedents and spirals at work: from job resources to work engagement, personal initiative and consequences of group potency: a longitudinal investigation of newly formed work-unit innovativeness. J. Vocat. Behav. 73, 78–91. doi: 10.1016/j.jvb.2008.0 work groups. Acad. Manage. J. 45, 352–368. doi: 10.5465/3069351 1.003 Little, T. D. (2013). Longitudinal Structural Equation Modeling. New York, NY: Halbesleben, J. R., and Bowler, W. M. (2007). Emotional exhaustion and job Guilford Press. doi: 10.5465/3069351 performance: the mediating role of motivation. J. Appl. Psychol. 92, 93–106. doi: 10.1037/0021-9010.92.1.93 Marks, M. A., Mathieu, J. E., and Zaccaro, S. J. (2001). A temporally based framework and taxonomy of team processes. Acad. Manage. Rev. 26, 356–376. Halbesleben, J. R., Neveu, J. P., Paustian-Underdahl, S. C., and Westman, doi: 10.5465/amr.2001.4845785 M. (2014). Getting to the “COR” understanding the role of resources in conservation of resources theory. J. Manage. 40, 1334–1364. doi: 10.1177/ Mathieu, J. E., Tannenbaum, S. I., Donsbach, J. S., and Alliger, G. M. (2014). 0149206314527130 A review and integration of team composition models: moving toward a dynamic and temporal framework. J. Manage. 40, 130–160. doi: 10.1177/ Hastings, S. E., and O’Neill, T. A. (2009). Predicting workplace deviance 0149206313503014 using broad and narrow personality traits. Personal. Ind. Diff. 47, 289–293. doi: 10.1016/j.paid.2009.03.015 McClurg, C. E., Chen, J. L., Petruzzelli, A., and Thayer, A. L. (2017). “Challenges and new directions in examining team cohesion over time,” in Research Hobfoll, S. E. (2001). The influence of culture, community, and the nested-self in on Managing Groups and Teams, Vol. 18, eds E. Salas, W. Vessey, and L. the stress process: advancing conservation of resources theory. Appl. Psychol. Landon (Bingley: Emerald Publishing Limited), 261–286. doi: 10.1108/s1534- 50, 337–421. doi: 10.1111/1464-0597.00062 085620160000018011 Hobfoll, S. E. (2011). Conservation of resource caravans and engaged settings. McGrath, J. E. (1964). Social Psychology: A Brief Introduction. New York, NY: Holt, J. Occup. Organ. Psychol. 84, 116–122. doi: 10.1111/j.2044-8325.2010.02016 Rinehart and Winston. doi: 10.1108/s1534-085620160000018011 Hobfoll, S. E., Halbesleben, J., Neveu, J. P., and Westman, M. (2018). Conservation McLarnon, M. J. W., and Carswell, J. J. (2013). The personality differentiation by of resources in the organizational context: the reality of resources and intelligence hypothesis: a measurement invariance investigation. Personal. Ind. their consequences. Annu. Rev. Organ. Psychol. Organ. Behav. 5, 103–128. Diff. 54, 557–561. doi: 10.1016/j.paid.2012.10.029 doi: 10.1146/annurev-orgpsych-032117-104640 McLarnon, M. J. W., and O’Neill, T. A. (2018). Extensions of auxiliary variable Hu, L.-t., and Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance approaches for the investigation of mediation, moderation, and conditional structure analysis: conventional criteria versus new alternatives. Struct. Equ. effects in mixture models. Organ. Res. Methods 21, 955–982. doi: 10.1177/ Modeling 6, 1–55. doi: 10.1080/10705519909540118 1094428118770731 Ilgen, D., Hollenbeck, J., Johnson, M., and Jundt, D. (2005). Teams in McNeish, D., Stapleton, L. M., and Silverman, R. D. (2017). On the unnecessary organizations: from input-process-output models to IMOI models. ubiquity of hierarchical linear modeling. Psychol. Methods 22, 114–140. Annu. Rev. Psychol. 56, 517–543. doi: 10.1146/annurev.psych.56.091103. doi: 10.1037/met0000078 070250 Morgeson, F. P., Reider, M. H., and Campion, M. A. (2005). Selecting individuals Kozlowski, S. W. (2018). Enhancing the effectiveness of work groups and teams: a in team settings: the importance of social skills, personality characteristics, and reflection. Perspect. Psychol. Sci. 13, 205–212. doi: 10.1177/1745691617697078 team work knowledge. Pers. Psychol. 58, 583–611. doi: 10.1111/j.1744-6570. 2005.655.x Kozlowski, S. W., Chao, G. T., Grand, J. A., Braun, M. T., and Kuljanin, G. (2013). Advancing multilevel research design: capturing the dynamics of emergence. Muthén, B. O. (2001). “Second-generation structural equation modeling with a Organ. Res. Methods 16, 581–615. doi: 10.1177/1094428113493119 combination of categorical and continuous latent variables: New opportunities for latent class/latent growth modeling,” in New Methods for the Analysis of Kozlowski, S. W., and Ilgen, D. R. (2006). Enhancing the effectiveness of work Change, eds L. M. Collins and A. Sayer (Washington, D.C: APA), 291–322. groups and teams. Psychol. Sci. Public Interest 7, 77–124. doi: 10.1111/j.1529- 1006.2006.00030.x Muthén, L. K., and Muthén, B. O. (2012). Mplus User’s Guide, 7th Edn. Los Angeles, CA: Mutheìn & Mutheìn. Kozlowski, S. W., Mak, S., and Chao, G. T. (2016). Team-centric leadership: an integrative review. Annu. Rev. Organ. Psychol. Organ. Behav. 3, 21–54. Muthén, L. K., and Muthén, B. O. (2015). Mplus 7.4. [Computer Program]. Los doi: 10.1146/annurev-orgpsych-041015-062429 Angeles, CA: Muthén & Muthén. Kozlowski, S. W. J., and Klein, K. J. (2000). “A multilevel approach to theory Ng, T. W., and Feldman, D. C. (2014). A conservation of resources perspective and research in organizations: contextual, temporal, and emergent processes,” on career hurdles and salary attainment. J. Vocat. Behav. 85, 156–168. in Multilevel Theory, Research, and Methods in Organizations: Foundations, doi: 10.1016/j.jvb.2014.05.008 Extensions, and New Directions, eds K. J. Klein and S. W. J. Kozlowski (San Francisco, CA: Jossey-Bass), 3–90. O’Neill, T. A., and Allen, N. A. (2012). Team meeting attitudes: conceptualization and investigation of a new construct. Small Groups Res. 43, 186–210. Lance, C. E., Vandenberg, R. J., and Self, R. M. (2000). Latent doi: 10.1177/1046496411426485 growth models of individual change: the case of newcomer adjustment. Organ. Behav. Hum. Decis. Process. 83, 107–140. O’Neill, T. A., and Allen, N. J. (2011). Personality and the prediction of team doi: 10.1006/obhd.2000.2904 performance. Eur. J. Personal. 25, 31–42. doi: 10.1002/per.769 Lang, J. W. B., Bliese, P. D., and de Voogt, A. (2018). Modeling consensus O’Neill, T. A., and McLarnon, M. J. W. (2018). Optimizing team conflict dynamics emergence in groups using longitudinal multilevel methods. Pers. Psychol. 71, for high performance teamwork. Hum. Resour. Manage. Rev. 28, 378–394. 255–281. doi: 10.1111/peps.12260 doi: 10.1016/j.hrmr.2017.06.002 Larrick, R. P., Burson, K. A., and Soll, J. B. (2007). Social comparison and O’Neill, T. A., McLarnon, M. J. W., Hoffart, G. C., Woodley, H. J., and Allen, N. J. confidence: when thinking you’re better than average predicts overconfidence (2018). The structure and function of team conflict state profiles. J. Manage. 44, (and when it does not). Organ. Behav. Hum. Decis. Process. 102, 76–94. 811–836. doi: 10.1177/0149206315581662 doi: 10.1016/j.obhdp.2006.10.002 O’Neill, T. A., McLarnon, M. J. W., Xiu, L., and Law, S. J. (2016). Core self-evaluations, perceptions of group potency, and job performance: the Frontiers in Psychology | www.frontiersin.org 1722 May 2019 | Volume 10 | Article 992

Woodley et al. The Emergence of Group Potency moderating role of individualism and collectivism cultural profiles. J. Occup. Svenson, O. (1981). Are we all less risky and more skillful than our fellow drivers? Organ. Psychol. 89, 447–473. doi: 10.1111/joop.12135 Acta Psychol. 47, 143–148. doi: 10.1016/0001-6918(81)90005-6 Pitariu, A. H., and Ployhart, R. E. (2010). Explaining change: theorizing and testing dynamic mediated longitudinal relationships. J. Manage. 36, 405–429. Thoms, P., Moore, K. S., and Scott, K. S. (1996). The relationship between doi: 10.1177/0149206308331096 self-efficacy for participating in self-managed work groups and the big five Ployhart, R. E., and Vandenberg, R. J. (2010). Longitudinal research: the personality dimensions. J. Organ. Behav. 17, 349–362. doi: 10.1002/(SICI)1099- theory, design, and analysis of change. J. Manage. 36, 94–120. doi: 10.1177/ 1379(199607)17:4<349::AID-JOB756<3.0.CO;2-3 0149206309352110 Roe, R. A., Gockel, C., and Meyer, B. (2012). Time and change in teams: where Vantilborgh, T., Hofmans, J., and Judge, T. A. (2018). The time has come to we are and where we are moving. Eur. J. Work Organ. Psychol. 21, 629–656. study dynamics at work. J. Organ. Behav. 39, 1045–1049. doi: 10.1002/job. doi: 10.1080/1359432x.2012.729821 2327 Salas, E., Rico, R., and Passmore, J., (eds) (2017a). The Wiley Blackwell Handbook of the Psychology of Team Working and Collaborative Processes. West Sussex: John Waller, M. J., Okhuysen, G. A., and Saghafian, M. (2016). Conceptualizing Wiley & Sons. doi: 10.1016/j.hrmr.2006.03.004 emergent states: a strategy to advance the study of group dynamics. Acad. Salas, E., Vessey, W. B., and Landon, L. B. (eds) (2017b). Team Dynamics Over Manage. Ann. 10, 561–598. doi: 10.1080/19416520.2016.1120958 Time: Research on Managing Groups and Teams, Vol. 18. Bingley: Emerald Publishing Limited. doi: 10.1016/j.hrmr.2006.03.004 Wang, M., Beal, D. J., Chan, D., Newman, D. A., Vancouver, J. B., and Vandenberg, Satorra, A., and Bentler, P. M. (2001). A scaled difference chi-square test statistic R. J. (2016). Longitudinal research: a panel discussion on conceptual issues, for moment structure analysis. Psychometrika 66, 507–514. doi: 10.1007/ research design, and statistical techniques. Work Aging Retirement 3, 1–24. bf02296192 doi: 10.1093/workar/waw033 Schaubroeck, J. M., Peng, A. C., and Hannah, S. T. (2016). The role of peer respect in linking abusive supervision to follower outcomes: dual moderation of group Whitman, M. V., Halbesleben, J. R., and Holmes, O. IV (2014). Abusive supervision potency. J. Appl. Psychol. 101, 267–278. doi: 10.1037/apl0000050 and feedback avoidance: the mediating role of emotional exhaustion. J. Organ. Sosik, J. J., Avolio, B. J., and Kahai, S. S. (1997). Effects of leadership style and Behav. 35, 38–53. doi: 10.1002/job.1852 anonymity on group potency and effectiveness in a group decision support system environment. J. Appl. Psychol. 82, 89–103. doi: 10.1037/0021-9010.8 Widaman, K. F., and Thompson, J. S. (2003). On specifying the null model for 2.1.89 incremental fit indices in structural equation modeling. Psychol. Methods 8, Stajkovic, A. D., Lee, D., and Nyberg, A. J. (2009). Collective efficacy, group 16–37. doi: 10.1037/1082-989x.8.1.16 potency, and group performance: meta-analyses of their relationships, and test of a mediation model. J. Appl. Psychol. 94, 814–828. doi: 10.1037/a0015659 Conflict of Interest Statement: The authors declare that the research was Stajkovic, A. D., and Luthans, F. (1998). Self-efficacy and work-related conducted in the absence of any commercial or financial relationships that could performance: a meta-analysis. Psychol. Bull. 124, 240–261. doi: 10.1037/0033- be construed as a potential conflict of interest. 2909.124.2.240 Copyright © 2019 Woodley, McLarnon and O’Neill. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Frontiers in Psychology | www.frontiersin.org 1733 May 2019 | Volume 10 | Article 992

CONCEPTUAL ANALYSIS published: 09 May 2019 doi: 10.3389/fpsyg.2019.01006 Teams in a New Era: Some Considerations and Implications Lauren E. Benishek1* and Elizabeth H. Lazzara2 1 Johns Hopkins School of Medicine, Armstrong Institute for Patient Safety and Quality, Baltimore, MD, United States, 2 Department of Human Factors and Behavioral Neurobiology, Embry–Riddle Aeronautical University, Daytona Beach, FL, United States Edited by: Teams have been a ubiquitous structure for conducting work and business for most Michael Rosen, of human history. However, today’s organizations are markedly different than those of Johns Hopkins Medicine, previous generations. The explosion of innovative ideas and novel technologies mandate changes in job descriptions, roles, responsibilities, and how employees interact and United States collaborate. These advances have heralded a new era for teams and teamwork in which previous teams research and practice may not be fully appropriate for meeting current Reviewed by: requirements and demands. In this article, we describe how teams have been historically Gro Ellen Mathisen, defined, unpacking five important characteristics of teams, including membership, University of Stavanger, Norway interdependence, shared goals, dynamics, and an organizationally bounded context, and relating how these characteristics have been addressed in the past and how they Paul B. Paulus, are changing in the present. We then articulate the implications these changes have on University of Texas at Arlington, how we study teams moving forward by offering specific research questions. United States Keywords: teams and groups, teamwork, team performance, team dynamics, team membership, team interdependence, team goals, team context *Correspondence: Lauren E. Benishek INTRODUCTION [email protected] Today’s organizations are markedly different than previously established. With the explosion Specialty section: of innovative ideas and novel technologies, organizations are redesigning the way work is This article was submitted to accomplished (Wageman et al., 2012). This new redesign is mandating a change in job descriptions, roles, and responsibilities as well as how employees interact and perform collaborative work. Organizational Psychology, According to Graesser et al. (2018), collaborative work can have potential disadvantages: ineffective a section of the journal communication, social loafing, diffusion of responsibility, and conflict. When harnessed correctly, Frontiers in Psychology though, collaborative work can entail division of labor, multiple perspectives, emergent ideas, and multi-source evaluation which enhances quality (Graesser et al., 2018). Collaborative work, Received: 30 November 2018 as the name would suggest, involves collaborations. Collaborations manifest differently with Accepted: 15 April 2019 the rise of geographic dispersion, working remotely, and collaborative technologies. Essentially, Published: 09 May 2019 collaborations entail teams and teamwork that have evolved and resemble a new era. Citation: The original conceptualization of teams considered them to be intact, tightly bounded, and Benishek LE and Lazzara EH coupled with members from a single organization who are co-located, interacting face-to-face to (2019) Teams in a New Era: Some generate an identifiable product, service, or solution (Hackman, 2012; Tannenbaum et al., 2012; Considerations and Implications. Wageman et al., 2012). Conversely, teams today consist of members from multiple organizations shifting in and out of the team while relying heavily on technology to complete a variety of tasks Front. Psychol. 10:1006. (Hackman, 2012; Tannenbaum et al., 2012). To illustrate, previous research has found that up to doi: 10.3389/fpsyg.2019.01006 84% of teams experience change (Espinosa et al., 2012), and another study found that the number of members from different countries was the same compared to the number of members located in Frontiers in Psychology | www.frontiersin.org 714 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era the same room (Cummings and Haas, 2012). These studies in a team’s core characteristics is permissible? And, what are simply illustrate that the archetype of teams is changing, and the implications of these characteristics with regard to team fluidity is increasingly prevalent. composition, process, and performance? Organizations are relying on fluid teams for several reasons. Recognizing the reality of teams, researchers are beginning to One, organizations must remain agile, responding quickly to advocate for more novel yet realistic approaches to theorizing and opportunities and market changes, and strive for strategic investigating teams (Harrison and Humphrey, 2010; Hackman, and operational innovation (Tannenbaum et al., 2012; Chiu 2012; Tannenbaum et al., 2012). Some have proposed that the et al., 2017). Two, organizations are often using independent concept of teams should be modified to reflect “teaming” – a contractors to execute work, which traditionally entails a finite continual process where teams are constituted and reconstituted duration and potentially limited involvement (Chiu et al., (Edmondson, 2012). Others have suggested the idea of team 2017). Three, organizations rely on such teams to stimulate fluidity to address this evolution of teams. However, many and energize members (Mortensen and Haas, 2016). Fourth, define team fluidity as simply changes in team membership organizations are designing teams according to the specific (e.g., Dineen and Noe, 2003; Bushe and Chu, 2011). More skills and expertise needed to execute particular tasks, and the recently, though, researchers contend that team fluidity is more requisite knowledge and skills may vary as the tasks fluctuate than membership change because that does not accurately (Tannenbaum et al., 2012). depict today’s teams and experiences (Chiu et al., 2017). Understanding that teams are in a new era, the purpose of Because organizations have new demands that leverage fluid this paper is to dissect each of the fundamental components teams, the implicit assumptions surrounding teams are not of teams – membership, dynamics, interdependence, goals, necessarily applicable. Previously, the well-established definitions and boundaries, – delineate the implications of how these assumed that the long-standing characteristics of teams (i.e., components are conceptualized, and recommend avenues for multiple members interacting dynamically and interdependently future research that will better capture the current nature of team working in a bounded context toward a shared goal; Hackman, membership, contexts, and dynamics. 1987; Salas et al., 1992; Cohen and Bailey, 1997; Kozlowski et al., 1999) remain stable and consistent throughout the team’s life CONSIDERATIONS REGARDING THE span (Tannenbaum et al., 2012). See Table 1 for a list team CORE CHARACTERISTICS OF TEAMS characteristics highlighted in various definitions of teamwork. These characteristics, though, as originally conceptualized may The fluidity and versatility of how actual teams operate not be accurate as the landscape of organizations, work, in real-world settings present serious challenges for those and teams has evolved considerably to be much more scientists and practitioners attempting to understand teams. diverse and heterogeneous (Harrison and Humphrey, 2010; We suggest that placing careful limits and boundaries on how Mathieu et al., 2017). we qualify ‘real’ teams is not the path forward if we are to provide research insights applicable to these real-world teams. Understanding there is a need to discern teams differently, Instead, we suggest we may find more practical direction researchers have argued that there is a notable distinction with in a comprehensive/integrated deconstruction of the defining teams either being “real” or “pseudo” (Hackman, 2002; Wageman features of teams that seriously considers how fluctuations within et al., 2005; Richardson, 2010; West and Lyubovnikova, 2012). each feature practically affect our approach to studying and Hackman (2002) suggested that real teams consist of four improving teams. primary elements: clear boundaries, established interdependence, moderately stable members, and authority. Wageman et al. Membership (2005) contended that real teams are comprised of three features: clear boundaries, collective responsibility for shared goals, and Perhaps the most defining characteristic of teams is membership. moderate membership stability. A more recent update posited After all, what makes a team recognizable as a specific team is that real teams are hallmarked by six dimensions: tightly coupled its members. Extending even further, team composition, team interdependence, agreed upon objectives, systematic reflex or size, and team tenure have team membership as the foundation. review of performance, clear boundaries, high autonomy, and According to many definitions, teams must be comprised of two specified roles (Richardson, 2010; West and Lyubovnikova, or more members (Kozlowski et al., 1999; Salas et al., 2005; 2012). Pseudo teams, on the other hand, are defined as a group Kozlowski and Ilgen, 2006; Rousseau et al., 2006; Salas et al., of people who call themselves a team and work independently 2007). Although these definitions do not imply that the members or interdependently toward a potentially different perception have to remain the same, it is often assumed that they are of their goal while having permeable boundaries (Richardson, consistent (Wageman et al., 2012), and research has traditionally 2010; West and Lyubovnikova, 2012). While we understand the treated teams as stable entities (Hirst, 2009). desire for a distinction and applaud those trying to more aptly apprehend and examine teams, we contend that this division Teams that have stable membership are considered to may not be totally suitable. The idealized conceptualization of be intact or closed (Ziller, 1965); meanwhile, teams that teams is a rare reality; rather, most teams are messy. That is, have fluctuating membership are thought to be open (Ziller, most teams in today’s climate are emergent social systems that 1965) or fluid (Bushe and Chu, 2011), and membership are fluid in various aspects (Chiu et al., 2017). With this fluidity in changes with regards to addition, subtraction, or substitution. mind, it begs the question – how much variation and fluctuation Frontiers in Psychology | www.frontiersin.org 725 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era TABLE 1 | Team characteristics. Two or more Inter- Shared Social Distinct Embedded in members dependent goal interactions roles larger entity Source x x x x x x Alderfer, 1977 x x x Anderson and West, 1998 x x x x x Cannon-Bowers and Bowers, 2011 x x x x x Cohen and Bailey, 1997 x x Devine et al., 1999 x x x x Francis and Young, 1970 x x x x Gladstein, 1984 x x x Guzzo and Dickson, 1996 x x x x x Hackman, 1990 x Hollenbeck et al., 1995 x x x xx x Katzenbach and Smith, 1998 x x x x Kazemak and Albert, 1988 x x x x Kozlowski et al., 1999 x x x xx x Kozlowski and Ilgen, 2006 x x x Kozlowski and Bell, 2003 x x x x Lanza, 1985 x x x x x McGrath et al., 2000 x x Rasmussen and Jeppesen, 2006 x x x x Richardson, 2010 x x x Rousseau et al., 2006 x x Salas et al., 1992 x x Salas et al., 2007 x x Salas et al., 2005 x Schippers et al., 2007 x x Shea and Guzzo, 1987 x Sundstrom et al., 1990 x Wageman et al., 2005 West, 2004 x West et al., 1998 Zander, 1977 Bedwell et al. (2012) ascribes this fluidity to three scenarios: and recent evidence suggests that ‘job hopping’ is on the rise (1) integrating a new member to an existing team, (2) losing a (Robert Half, 2018). member of an existing team without replacing the lost member, or (3) losing a member and integrating a new member to an Adding another layer of complexity to membership change is existing team. Such changes can occur at the simple level of a the consideration of ‘multi-team’ or multiple team membership single member to the complex level of an entire cohort (Mathieu (MTM), which refers to members serving on multiple teams et al., 2014). Consequently, team membership can range from simultaneously (van de Brake et al., 2018). MTM adds complexity “frozen rigidity” to “radical discontinuity” (Arrow and McGrath, in that there are two relevant considerations: context switching 1995) with changes in frequency (i.e., turnover) and duration and temporal misalignment (O’Leary et al., 2011). Context (i.e., tenure) serving as two major indices (Chiu et al., 2017). switching occurs when members shift their focus from one team Within an organization, such membership change occurs for context to another, and temporal misalignment occurs when seven reasons: (1) desire to have different skills through various there is a gap in time from focusing on tasks. Understanding stages of work, (2) need for flexible allocation of personnel, these considerations is important since some estimates indicate (3) drive to provide developmental career opportunities, (4) 81% of individuals have MTM (O’Leary et al., 2011), and response to high turnover, (5) need for organizational upsizing others suggest 94.9% of members serving on multiple teams or downsizing, (6) desire to promote effective communication, (Martin and Bal, 2006). The pervasiveness of MTMs is due and (7) motive to avoid collusive behaviors among employees to particularly skilled individuals being a desired commodity, (Bushe and Chu, 2011). In addition to within organization, teams being project-centered that necessitate individuals with membership change occurs across organizations. The philosophy specialized expertise, and work that has shifted toward being flat of maintaining the same employment and retiring from the and dispersed (O’Leary et al., 2011). same organization is becoming an old adage (Landrum, 2017), Recognizing the prevalence of MTM or even membership change within a single team, researchers have begun to Frontiers in Psychology | www.frontiersin.org 736 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era investigate the potential implications for taskwork, teamwork, symbiotically with interdependence. Interdependence moderates and team performance. Two opposing views regarding the role the relationship between team processes (i.e., cognitions and of membership change have emerged. One school of thought behaviors) and team performance (Gully et al., 1995, 2002; Beal frames such changes as being disadvantageous. Membership et al., 2003). As such, it is a critical team feature that is almost change results in a loss in individual knowledge and shared ubiquitously included in every team definition. The underlying knowledge, a diverted focus away from the task, lowered tenet of interdependence is that the more interdependent team member commitment, and lack of cohesion (Bushe and Chu, members are with one another, the closer they approach “real” 2011; Bedwell et al., 2012) as well as diminished coordination team status while lower interdependence is more indicative of a (Summers et al., 2012) and reduced cooperation (Arrow and “working group” as opposed to a “real” team (Katzenbach and Crosson, 2003). Additionally, such teams have poorly developed Smith, 1993, 1998; Wageman et al., 2005). shared mental models and transactive memory systems making them unable to orient quickly to new tasks and transitions The source of interdependence can be multifaceted. It may (Bush et al., 2018). Finally, there is evidence that demonstrates be determined by the nature of the task, the manner in which member instability detrimentally impacts performance (Argote goals are defined, the process through which those goals are et al., 1995; Lewis et al., 2007); while, team familiarity strengthens achieved, and the method for assessing team performance team processes and states (Mathieu et al., 2014). The other school (Wageman, 1995; Campion et al., 1996; Van der vegt et al., of thought frames membership change as beneficial, citing as 2001). Task interdependence refers to the degree of task-driven evidence an increase in breadth of knowledge (Bedwell et al., interaction among team members (Shea and Guzzo, 1987). 2012), transfer of knowledge and resources (Tannenbaum et al., Stated differently, task interdependence is the level to which 2012), and the number and diversity of ideas generated (Choi colleagues must rely on one another in order to effectively and Thompson, 2005) as well as increased productivity (Choi and perform their individual roles and job responsibilities (Saavedra Thompson, 2005) and heightened team learning (Savelsbergh et al., 1993). As task interdependence increases, demands for et al., 2015). Moreover, such fluidity may help maintain a team’s coordination, communication, and cooperation also tend to flexibility, which is particularly beneficial in emergent situations increase. Consistent with the idea that interdependence exists in and circumstances (Tannenbaum et al., 2012). To illustrate, degrees, task interdependence has been conceptualized as existing teams that are more fluid may be equipped to address task in different forms that range from a lower degree of integration conflict, which can be a beneficial catalyst for communication to a much higher and more complex degree. as well as a tool for mitigating groupthink (Bush et al., 2018). Therefore, the objective of team membership decisions should be Pooled interdependence can be summarized as a performance- to strategically support the organizational mission and promote sum relationship where each member contributes to the group organizational flexibility in competitive environments (Bell et al., without needing to directly interact with other group members. 2018b). Regardless of the school of thought, membership change Naturally, this is the lowest level of interdependence because it undoubtedly has an impact on processes, states, and outcomes. simply means that team performance is the simple sum of each individual’s performance. When task interdependence is pooled Given that research has repeatedly indicated that a team’s each team member contributes his/her own work to the final members substantially influence teamwork (Mathieu et al., product without being reliant on any other member. A loose 2017; Bell et al., 2018a) and performance (Bell, 2007) and the example of pooled interdependence might be an edited textbook ever-changing nature of membership, new questions start to wherein each chapter an author contributes content based on surface. If membership is so fluid, how should measurement his or her own expertise without needing to consult the authors be implemented to accurately reflect the state of the team as of other chapters. The final publication is the result of multiple well as the dynamism of the team? Is resorting to traditional authors’ contributions and would not have been possible without cross section or correlational designs still appropriate? Can we each, but the chapters within are individual products. Sequential make fair comparisons longitudinally if the composition of the task interdependence occurs when one group member must team is different? These questions simply scratch the surface complete his task before another member is able to complete hers as the dynamism of membership does not only affect the team and different parts of the task must be completed in a prescribed dynamics, but it also influences the team’s interdependence, goals, order. The classic example is of a car assembly line where each and boundaries. Additionally, such fluid membership also raises employee performs a specific action that contributes to the final questions about selection, interventions, and work design that product. In this example, interdependence is a bit stronger than merit investigation. in the pooled example because members are dependent on others to complete their work. Reciprocal interdependence is the next Interdependence conceptualization and occurs when team performance requires individuals to hand tasks back-and-forth between one another. Interdependence refers to the level or sequencing of interaction These “temporally lagged, two-way interactions” (Saavedra et al., required of team members in order to complete a given task 1993, p. 63) generally exist when team members have different or achieve a particular goal or outcome and is often the reason specialty roles that can be completed in a flexible order. For why teams are formed in the first place (Campion et al., 1993). instance, two colleagues co-authoring a paper may write different The nature of what a team is trying to accomplish can be sections and then go back and edit one another’s work until the characterized by a two-dimensional framework – scope and manuscript is ready for review. Finally, team interdependence complexity (Mathieu et al., 2017). A team’s objectives work exists when members jointly diagnose, problem solve, and Frontiers in Psychology | www.frontiersin.org 747 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era collaborate to complete a task. There is considerable freedom individuals to collaborate. They would instead be engaged within this level of interdependence to design your own job in separate pursuits. However, once two or more individuals responsibilities, but the final product requires mutual interaction. are united in the attainment of the same objective, they An example may be a design team working together to co-create become interconnected. While the pathway to goal attainment a redesign of a gaming platform. can vary (see task interdependence), the unity between them manifests as goal interdependence and guides their The problem with conceptualizing interdependence in this performance (Saavedra et al., 1993). Thus, goals direct the way is that it is probably more complex in reality than how attention, effort, and persistence of group members (LePine, it is presently conceived. In fact, many modern teams are 2005) while also influencing interactions within teams (Hu involved with multiple tasks simultaneously, and each of these and Liden, 2011). Specifically, goals direct teams on how tasks might be associated with different levels of collaboration to define individual responsibilities, coordinate actions, and (Bell et al., 2018a). Similarly, the longer the lifespan of the develop efficient work procedures (Klein and Mulvey, 1995). team, the more likely it is that a work group moves between This influence manifests through planning, cooperation, different levels of interdependence. A development team, for mutual support, and member interactions (Mitchell and example, may demonstrate high-levels of interdependence when Silver, 1990; Weingart, 1992; Weldon and Weingart, 1993; they are steeped in the divergent and convergent thinking Crown and Rosse, 1995). stages of development, but as they navigate other stages of creation, they may find that their interdependencies become less The effort extended by group members in the pursuit of complex. Furthermore, this team may find that they regularly shared goals creates variance in the rewards, punishments, and shift between interdependencies as they come together for intense feedback teams receive. Competitive and individual distribution brainstorming and co-creation then somewhat disband to work of outcomes can inhibit team effectiveness through blocking, on individual tasks then come together again to assemble and test undermining, and hindering behaviors (Miller and Hamblin, their prototype. 1963). Alternatively, shared goals can create shared responsibility for outcomes among team members (Shea and Guzzo, 1987) The questions we as researchers must ask is, if which are likely to enhance effectiveness by motivating members interdependence is an organic moving target, how does that affect to cooperate and assist in the performance of other members our definition and conceptualization around ‘team’? Can ‘real’ (Gully et al., 2002). teams be considered teams if the level of their interdependencies changes over their life span? Does the shift in interdependency The interplay of shared goals and responsibilities of outcomes within a working group affect the ‘teamness’ of that group? Can clearly has implications for how teams perform. They affect team a single team be more or less of a team throughout its lifespan as motivations, work distribution, and team member interactions its interdependent nature fluctuates? Importantly, what does that because they set the direction in which the team is moving mean for teams operating in the real world? and serve as glue cementing the team together. Thus, teams are partially defined by the goal(s) they are harmonized in striving If we conclude that teams can and do in fact fluctuate with toward. The implication is that if team goals change, then the regards to interdependencies and this affects their ‘team status,’ team may also be qualitatively different. Work may need to there are clear implications for research. Measurement becomes be restructured. Team members may need to be subtracted or more challenging because we will need to consider what level added. The dynamics of work and team member interactions or even combination of interdependence teams are experiencing may be substantively different. In sum, the morphing of team at the time of measurement, and if their interdependence goals and responsibilities for outcomes should be considered profile is different across measurement timepoints, we may as a parameter for determining whether a work unit can need determine whether fair comparisons can be drawn or be meaningfully compared over time or across performance whether we need to develop more sophisticated methods to contexts. Changes in team processes and emergent states may understand the impact on their performance. We must also not be the result of team learning, for example, so much as they consider more practical concerns such as how do we make may be natural reactions to changes in goals. From a practical sure that team members are selected and/or trained to be standpoint, the implications may be that lessons team members able to navigate these fluctuations. It is quite possible that learned by working with one another toward different objectives the team member who operates best when interdependence may not entirely translate into the current project. Teams may more closely reflects pooled or sequential process will find discover growing pains resulting from goal shifts. It may require periods of more intense interdependence difficult to maneuver additional work for teams to readjust to changing demands and vice versa. Thus, we will need to find better ways and goals, so special efforts may be necessary as a team strives to support employees as they engage in various forms toward a new goal, even when there is a history of collaboration of collaboration. among its members. Goals/Shared Responsibility for Of course, from a theoretical and research perspective it means Outcomes that we may have to be cautious about how we define and measure teams. If goals are substantially changing and work flows are Another defining feature of teams is the existence of at least also changing as a result, it may no longer be appropriate to one shared goal. This feature is central because without consider a specific collection of individuals to be the same team. a shared objective, there would be no reason for multiple In that instance, we have to be careful about the inferences Frontiers in Psychology | www.frontiersin.org 758 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era we are drawing from assessment of these groups over time or TABLE 2 | Matrix of teamwork competencies. across circumstances. Team Team Dynamics Task Specific Specific Generic Clearly, team members must interact with one another in order Generic to pursue shared goals and manage task interdependencies. Context-driven Task-contingent However, team dynamics and interactions vary greatly and are Team-contingent Transportable moderated by a number of attitudes and behaviors. The leading taxonomy for characterizing team interactions and dynamics Context-driven competencies are specific to both the task and comes from Marks et al. (2001, p. 357) who describe processes the team, making them highly specialized. As such, these and emergent states. Processes are “members’ interdependent competencies are generally best developed within in-tact teams acts that convert inputs to outcomes through cognitive, verbal, trained or practiced in realistic settings. They are not especially and behavioral activities directed toward organizing taskwork good candidates for selection since it is difficult to understand to achieve collective goals.” More simply, team process is in advance how a team member may integrate his or her KSAs the interaction of members with each other and their task into an existing team. Task-contingent competencies are specific environment. Processes are the means through which team to the task but not to the team. These are best trained in a members use essential and varied resources such as experience, realistic task environment and may be useful for selecting new expertise, equipment, and financial support to garner team team members. Team-contingent competencies are specific to the outcomes. Thus, it is team process (i.e., action and interaction) team but not to the task. It is generally unhelpful to select team that drives accomplishment of team goals (LePine et al., 2008). members based on these competencies and instead they are better developed with intact teams across a variety of tasks. Finally, Of course, teamwork involves more than simple behaviorally transportable skills are the most flexible. They are applicable to based action. Teamwork also consists of attitudes, values, all teams across all tasks. cognitions, and motivations (Morgan et al., 1993; Salas et al., 2011). Marks et al. (2001, p. 357) call these affective and Irrespective of the specifics of the task or team, all teams cognitively oriented qualities of teamwork emergent states. experience transitions as they evolve from one task to the Emergent states are “constructs that characterize properties of the next. Marks et al. (2001) clearly outlined the relevant processes team that are typically dynamic in nature and vary as a function for transition periods (i.e., mission analysis, goal specification, of team context, inputs, processes, and outcomes.” These team and strategy formulation) as well as the processes for action properties are states in that their quality is not guaranteed to periods (i.e., monitoring goal progress, systems monitoring, team be stable. As such, emergent states can influence how team monitoring, and coordination); however, modern teams likely process unfolds while themselves changing in response to team do not experience these clear delineations as postulated by member interactions. For example, teams low in psychological Marks et al. (2001). In other words, modern teams experience safety (an emergent state) may struggle to ask each other for task transitions along a continuum of length and punctuated between assistance (a process), which might result in performance delays tasks that range on a continuum from similar to dissimilar or errors, which could stir conflict or discord (another emergent (Bush et al., 2018). Consequently, the requisite processes may state) within the team where none previously existed. In other differ given the temporality of the transition periods and the words, emergent states are products of the team experience that similarity (or dissimilarity) of the tasks surrounding those also can impact the way in which team members interact, be it transition periods. positive or negatively. As teams maneuver these phases, they must make decisions Certain competencies are needed to manage these evolving in an evolving world, requiring them to be flexible in the processes and dynamic emergent states. According to Cannon- presence of change. Team adaptation is the process through Bowers et al. (1995), teamwork-related competencies vary on two which teams respond cognitively, affectively, and behaviorally domains – task and team – that span a continuum of specificity to change (Baard et al., 2014), which can stem from internal that ranges from specific to general. Task specific competencies (e.g., membership turnover) or external (resource availability) are those that are applicable only in a specific task or type of task. sources (Frick et al., 2018). Successful adaptation has beneficial For example, aviation skills are highly task-specific. Task generic outcomes for teams; however, it may also manifest maladaptively competencies are those that are applicable across a variety of task for numerous reasons (Frick et al., 2018). Frick et al. (2018) settings. For instance, project management skills are applicable describe the Four Rs heuristic to explain how team adaptation across a variety of projects and contexts. Correspondingly, team occurs and explain the points of failure in this process that competencies can also be categorized as specific or general. Team could result in maladaptation. The stages include recognize (i.e., specific competencies require the team members to know one noticing and acknowledging a change), reframe (i.e., shifting another well and have experience working together whereas team cognitions about the situation as a result of the change), respond generic competencies are applicable across different teams with (modifying behavior), and reflect (i.e., contemplating the change different team members. and the team’s subsequent response). When considered on a matrix (see Table 2) these domains Affecting these tasks and transitions while constraining and combine into four distinct types of competencies: Context- influencing team dynamics is team structure. Team structure driven, task-contingent, team-contingent, and transportable. encompasses the team relationships that drive the assignment Frontiers in Psychology | www.frontiersin.org 769 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era of tasks, roles and responsibilities, and leadership (Bresman and criterion). Although these criteria may provide clarity for Zellmer-Bruhn, 2013; Chiu et al., 2017). Like many other defining how boundedness is determined, and literature often assumes elements of teams, structure has historically been assumed to clear team boundaries are the norm, the actual boundaries in be somewhat stable in nature. That is, task assignments are real-world teams are often less clear (Tannenbaum et al., 2012). pre-defined, roles and responsibilities are clear and consistent Some have long proposed that boundedness may be actually be a over time, and leadership manifests as command and control spectrum with highly permeable boundaries (i.e., underbounded) (Chiu et al., 2017). However, in modern teams these traits are to highly impermeable boundaries (i.e., overbounded; Alderfer, also increasingly fluid. Task assignments occur on an as-needed 1980); meanwhile, others have posited that boundaries are basis and are given to team members with the ability and more dynamic and fluid and are constituted and reconstituted bandwidth to perform them. Roles and responsibilities, therefore, (Edmondson, 2012). In fact, there is so much fluctuation that become more blurred (Dube, 2014) as team members coordinate it is often difficult to determine who comprises the team to move with greater adaptability and agility. Leadership is (Hackman, 2012). increasingly self-directed (Aime et al., 2014) and shared across team members (Carson et al., 2007). Team member status Such ambiguous boundaries are a result of team fluidity, emerges quickly based on observable characteristics and expected overlap, and dispersion (Mortensen and Haas, 2016). Fluidity performance but is subject to change if those in positions of entails members who are dynamically moving in and out of the authority fail to perform adequately (Driskell T. et al., 2018). The team. Overlap involves members who work on multiple teams result is that modern teams rely less on stringent pre-defined simultaneously, and dispersion refers to members working from plans, rules, procedures, and communication norms (Malone different organizations or geographic regions. Such fluctuations and Crowston, 1994) and more on informal and emergent in team membership are often arranged and coordinated rather coordination (Okhuysen and Bechky, 2009). than being chaotic and impromptu (Tannenbaum et al., 2012). Because the boundary is being reshaped with such fluctuations, While these frameworks help us to organize the way we it impacts shared identity and shared understanding. With every approach, think about, and manage team dynamics, they fluctuation, the team must rebuild its identity and must update somewhat fail to account for the complexity that are real world the shared understanding based upon the member’s mental teams. For example, it is likely that many teams require some models of the team, task, and context. combination of specific, general, team, and task competencies to support team emergent states and processes. These competencies Fluidity, overlap, and dispersion affects boundedness between are further influenced by the transitions teams experience the team and the outside context, but it also affects boundaries between tasks. In fact, even the dynamics themselves as well as within the team and the tasks. Team members create boundaries the transitions may be contingent upon the tasks. within the team based upon the extent that they perceive themselves to be similar to one another. That is, team members Team Boundaries rely upon surface-level cues (i.e., attributes that are easily accessible and detectable) and deep-level cues (i.e., psychological The final defining characteristic of a team is the idea that a characteristics) to inform categorization. Categorization enables team does not exist in a vacuum but rather is influenced by team members to rely upon heuristics which can serve as context. According to Bell et al. (2018a), context shapes the an impetus for subsequent attitudes and behaviors (Feitosa team in three ways. One, the context influences the salience of et al., 2018). Consequently, such categorization and perceptions a particular attribute. Two, the context can alter the relevance influence the roles, interactions, and structures (Bell et al., 2018a; and importance of an attribute. Three, the context ignites which Feitosa et al., 2018; Graesser et al., 2018). To elaborate, teams attributes are of value. Some conceptualize context broadly by often develop a core and a periphery structure (Tannenbaum making a distinction between external, influences mostly outside et al., 2012; Mortensen and Haas, 2016). Albeit colloquially, the of the control of the team, and internal, influences within the concept of a core and a periphery is analogous to an inner team (Bell et al., 2018a). Meanwhile, others conceptualize context and outer circle. The core structure is comprised of members more granularly by referring to context as the characteristics who perform a “major” role; whereas, the periphery structure of the task, the timeframe of the performance episode(s), the includes members who perform a more “minor” role. Similarly, governance structure over the team, and a team being embedded tasks can also manifest as a central working sphere and a within a larger entity or context (Edmonson and Harvey, 2018). peripheral working sphere (Gonzalez and Mark, 2004). A central In essence, the context functions by providing boundaries. working sphere is considered important and urgent; whereas, a peripheral working sphere is deemed to be less important and Boundaries in a general sense facilitate togetherness and critical. Additionally, members dedicate more time on central serve as a distinction between what something is versus what working spheres yet allocate minimal time toward peripheral it is not (Alderfer, 1976). Within the team context, a team working spheres. has boundedness with boundedness being a delineation between members and non-members, and individuals use three criteria Given the dynamism of boundedness between entities, within to identify boundedness (Mortensen and Haas, 2016). One, teams, and tasks, it is evident that boundary clarity is integral. members rely on an official team roster (formal criterion). Two, When teams experience boundary clarity, members experience individuals receive the label of team member by themselves individual certainty, and the team experiences a collective or someone else (identity-based criterion). Three, members are agreement (Mortensen and Haas, 2016). Conversely, teams identified through a pattern of interactions (interaction-based that have poor boundary clarity are comprised of members Frontiers in Psychology | www.frontiersin.org 870 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era with individual uncertainty and an overall sense of collective for team type taxonomies, it does portray that there is no disagreement. Members are unsure of who is considered a consensus regarding how teams should be classified and that member of the team, and members have opposing views on who many taxonomies approach classification based primarily on is an actual member of the team. task type. Teams have greater distinctions beyond task type, so such categorization actually limits our apprehension of Regardless of the clarity or ambiguity, the boundedness team effectiveness (Tannenbaum et al., 2012). Recognizing the of a team has implications for researchers and practitioners limitations of instituting a categorical classification system for (Mortensen and Haas, 2016). That is, researchers may need team types, Hollenbeck et al. (2012) created a dimensional to alter their theorizing and measuring depending upon the scaling framework to describe teams positing that teams varied stability and clarity of the boundedness. For example, many on authority differentiation, skill differentiation, and temporal team processes or states are grounded in the idea that teams stability. The dimensional scaling approach is closer to potentially are tightly coupled and bounded (e.g., transactive memory representing teams; however, the theory might need to be altered systems), but how do these manifest if teams have loose and further to account for the dynamism of all facets of modern permeable boundaries? Similarly, roles and responsibilities are teams. For example, because boundaries can be ambiguous and often theorized based on the assumption that they remain membership can be fluid, the team type may also change with consistent, but if a team’s boundaries are fuzzy, the idea of a time and as the team progresses and transitions between tasks. If boundary spanner needs revisiting (Mortensen and Haas, 2016). the team type does in fact change and is in fact dynamic, what are Practitioners, similarly, may need to select, design, and support the implications for teamwork and taskwork as well as team and teams differently depending upon the consistency and certainty task performance? Do variations on teams (e.g., virtual teams; of the boundedness. Gilson et al., 2015 and multi-team systems; Shuffler and Carter, 2018) impact teamwork, taskwork, and outcomes? Essentially, IMPLICATIONS what constitutes different teams may need to be updated with the changes to reflect contemporary work and organizations. For decades, the needs and experiences that teams faced in Understanding what constitutes such teams as well as what the real-world as well as the policies and procedures that conditions are most important helps lead to greater insights practitioners used to manage teams corresponded to the studies regarding team effectiveness (Hackman, 2012). Questions for that researchers were conducting (Tannenbaum et al., 2012; future research include: Wageman et al., 2012). However, the organizational landscape that has manifested is not always aligning with prevailing • What features beyond the task constitute team types? How research; therefore, research and even practice needs to evolve does the evolution of these features impact the team type? according to current needs to advance team effectiveness. Although “old questions” become relevant again when the very • What approaches are most suitable for characterizing teams nature of teams has changed (Wageman et al., 2012), others (e.g., categorical or dimensional scaling)? argue that the questions should actually shift given the gravity of changes (Mathieu et al., 2017). Below we discuss implications of • What other categorizations or dimensional scaling factors the evolution of teams in the modern era for research. need to be considered and included? • How do variants on traditional team types impact teamwork? Team Types Models and Frameworks When attempting to understand what constitutes a team, many Perhaps the foundation of most team theorists is the depiction have theorized about team types. For example, Sundstrom of team effectiveness models. Many team models are rooted et al. (1990) postulated that there are four main team types: in the input–process–output (IPO) foundation put forth by advice/involvement, production/service, action/negotiation, and McGrath (1964). Inputs are the antecedents that influence the project/developmental teams. Cohen and Bailey (1997) followed dynamics of team members. The processes are the interactions suit by suggesting there are project teams, traditional work that team members undertake to achieve the desired goal, and the teams, parallel teams, and management teams. Devine et al. outputs are the outcomes or results accomplished by the team. As (1999) created another taxonomy to include four team types: Hackman (2012, p. 431) says, “the core idea of the model is that ad hoc project teams, ongoing project, ad hoc production, input states affect group outcomes via the interaction that takes and ongoing production and actually modified the taxonomy place among members.” to include 14 different team types (Devine, 2002). Even still, De Dreu and Weingart (2003) created their own team Despite being a valuable infrastructure, the IPO framework type taxonomy, which included project teams, production has several limitations leading others to modify the original teams, decision making teams, and mixed teams. Finally, conceptualization. Many adaptations have included an Wildman et al. (2011) presented a team type taxonomy based environmental or contextual component since teams do upon tasks: managing others, advising others, human service, not operate in a vacuum and are certainly influenced by negotiation, psychomotor action, defined problem solving, and contextual factors (e.g., Cohen and Bailey, 1997). Introducing a ill-defined problem solving. Although these are simply several contextual component lead to the realization of the multilevel examples demonstrating various interpretations and suggestions nature of teams – individuals are nested within teams, and teams are nested within organizations which exist within even Frontiers in Psychology | www.frontiersin.org 881 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era broader environments (Klein and Kozlowski, 2000). A second • What are the relationships between all of the contextual limitation of the IPO approach is the narrow focus on process. factors as well as the team processes, states, and outcomes Processes are interdependent cognitive, verbal, and behavioral if they are not linear? activities that convert inputs to outputs (Marks et al., 2001), but not all teamwork components are simply processes. Teamwork • What are the unique contributions of team attitudes, is also comprised of emergent states, which are properties that behaviors, and cognitions, and how do these vary over time? represent the attitudinal and cognitive properties of the team (Marks et al., 2001). Further, not all “processes” are mediators • How can temporal components be best incorporated as originally depicted in the IPO organization; unpacking and depicted? teamwork entails that some processes and emergent states can be moderators as well as mediators. Understanding these Measurement conceptual limitations, Ilgen et al. (2005) delineated “process” by presenting the input–mediator/moderator–output–input As we have indicated, previous thinking depicted teams with (IMOI) framework. A third limitation in the IPO approach is relatively stable factors (e.g., goals and roles). Because team the lack of temporality, noting the limitations of suggesting that factors were primarily theorized as being stable, they are often teams operate linearly and not episodically (Marks et al., 2001). only measured once or used as correlates (Tannenbaum et al., To address the temporality of teams, there are two prominent 2012). Such data is often collected at the individual level, approaches: developmental and episodic (Mathieu et al., but because it is evident that individuals are nested within 2008). The developmental approach suggests that teams have teams, individual data is often aggregated to the team level. differential influences and qualitatively change over time. The Some argue, though, that simple linear aggregations are not episodic approach posits that teams exhibit different processes appropriate since the inputs, processes, and states are not and states at different times. See Mathieu et al. (2008) for a review perceived similarly across members and are not interchangeable of team effectiveness. (Murase et al., 2012). Aggregates represent compositional characteristics, and compositional thinking assumes the content All of the models that attempt to address previous limitations and structure are created linearly and represented similarly certainly advance our understanding of the complex phenomena (Bell, 2012). The characteristics of today’s teams are much of team effectiveness; however, we argue that more work more fluid and dynamic. Therefore, teams and the factors regarding the theoretical nature of teams is still needed. The that comprise teams need to be studied and measured with influences and the underpinnings of teams do not reside in regards to patterns over time (Bell, 2012; Mortensen and clear and distinct packages, but rather the effectiveness of Haas, 2016). More specifically, measures of patterns could teams lies in the complex web inherent within teams and include: density, reciprocity, transitivity, and centrality. Studying teamwork (Hackman, 2012). Modern teams are likely not networks and patterns is more representative since extensive well-represented within simple cause-effect models because what fluidity raises the question of whether the relevant team ensues as teams strive to accomplish their goal(s) is not a linear members are being measured across time and whether the progression; instead, a complex combination of factors varying multilevel nature of teams is being captured. Simply stated, differentially is a more accurate representation. Modern teams are researchers are at risk of comparing different sets of team likely to juggle tasks over time, experience membership churn, factors (e.g., membership) when only using cross sectional coordinate with other teams, and reconfigure throughout its measurement (Murase et al., 2012). A network approach acquires lifecycle (Driskell J.E. et al., 2018). Future approaches should information about individuals and their attributes as well as consider more sophisticated frameworks that move beyond the team-level properties, and it captures the nature of the causal models and involve an analysis of all factors and conditions interactions. This approach is useful for understanding where (Hackman, 2012). Additionally, other processes, such as team an individual is embedded within a larger team or multiteam creativity and innovation, and emergent states, such as team system (Bell et al., 2018b), helping to identify essential players well-being, may become more central drivers of modern team and create a more comprehensive understanding of teamwork effectiveness and should situate more prominently in team through a relational lens. Ultimately, theorizing and researching performance frameworks and research (Driskell J.E. et al., current teams requires a shift from the old fashioned to a 2018). With these considerations, we put forth the following more modernized approach. Mathieu et al. (2017) posit that research questions: a more modernized approach likely means that there is no standard set of measures for team research. The specifics that • What approach(es) are more suitable if the input-process- influence or are inherent within each team vary too greatly state-output structure is outdated and not applicable? between and across teams making them markedly different and necessitating more nuanced metrics. Consequently, we propose • How should modern team effectiveness models and the following questions: frameworks be represented to accurately depict contemporary teams? • What research designs should be leveraged to most appropriately generalize from lab settings and study teams • How can team models and frameworks correctly depict the in applied settings? fluidity of all of the characteristics of modern teams? • What statistical techniques should be employed to • What team attitudes, behaviors, and cognitions should be correctly represent the complex web of teamwork and considered when creating team effectiveness frameworks? team performance? Frontiers in Psychology | www.frontiersin.org 892 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era • What approaches are more suitable if central tendencies no have a strong identification with the company culture so longer provide a clear picture and understanding? that these individuals are better able to collaborate and coordinate with other members of the organization even as • What tools and metrics are most suitable to best understand their teams assemble, disband, and reassemble in different and unpack the simultaneous and interrelated nature of configurations and navigate changing expectations, goals, and teamwork? What tools and metrics are appropriate for demands. Organizational value congruence is expected to time-dependent constructs? reduce both task and relationship conflict between team members (Chuang et al., 2004), therefore, selecting staff for • How can illustrative case studies be leveraged to highlight congruence with organizational values will help team members novel constructs and relationships? subject to participating on multiple teams or teams that quickly configure and disband work collaboratively with their • What continual streaming metrics (e.g., wearable sensors) colleagues. With this in mind, we present the following can be utilized to address longitudinal issues? research questions: Staffing Teams • What competencies or values should be considered when staffing teams? The dynamism and fluidity of today’s teams present special challenges for staffing teams. In some cases, such as in surgical • How does staffing influence the manifestation of team units, teams may come together for fairly brief periods of time, competencies or values? even just a few hours to complete a single surgery. These teams might complete multiple projects or cases in rapid succession, or • How does the fluidity of team characteristics they may disband after just one project together. It is possible (e.g., interdependence) impact staffing decisions? for these short-duration teams to reconfigure, sometimes with Conversely, how does staffing impact the dynamism a majority subset of the original team and other times with a of team characteristics? composition of team members that barely resembles the original team. This is a stark difference from the “traditional” team, • What are the qualities of a flexible team member, and with its longer-duration lifecycle and mostly stable membership. how can team members help one another to become Traditional teams have the benefit of time, allowing them to more more flexible? deeply develop critical emergent states like trust, psychological safety, and transactive memory systems. Because members of • How do team member characteristics impact how traditional teams have to work with one another for longer work is completed? durations, it is possible that individual idiosyncrasies and work habits are more important in these contexts. However, for the • What task characteristics should be considered when rapid cycle teams that are appearing with greater regularity in staffing modern teams? How does the evolution of the task the modern workplace, these individual differences may be less and its characteristics influence staffing decisions? important to staffing a team. Instead, it is likely that who is on the team is less important than what knowledge, skills, and Team Interventions attitudes they contribute. Selection, therefore, may require less attention on compatibility of team members’ personality and Of course, it does not make logistical, practical, or even work preferences and instead emphasize the compatibility of conceptual sense to rely on selection as the main source of team member strengths and competencies. controlling team membership and performance during dynamic situations. However, team-based interventions (i.e., systematic We must also consider how dynamism in roles, goals, and activity aimed at strengthening team competencies and dynamics tasks can impact selection of team members. Teams should be and improving team performance; Lacerenza et al., 2018) staffed based on members’ value to organizational competitive are also employed. As described above, interventions are advantage (Bell et al., 2018b). Changing needs, which may or especially useful for modifying context-specific and team-specific may not be anticipated, adds complexity to the issue of staffing competencies (Cannon-Bowers et al., 1995). However, given teams. While each team member may still bring a specific the faster pace of change in modern organizations and background or expertise to the team, expertise particulars may the agility that many teams must demonstrate in order to not be able to be successfully anticipated. It, therefore, may be perform well, traditional approaches to interventions may also more appropriate to look for individuals with certain attributes need to be re-thought. Known hallmarks of well-designed that might facilitate adaptability to changing circumstances and training include communicating information, demonstrating demands. Such attributes may include learning orientation, the principles, skills, or behaviors to be learned, providing self-directed motivation, tolerance for ambiguity, and willingness opportunities for students themselves to practice, and providing to empathize, brainstorm, and prototype. It is also likely subsequent feedback on their performance (Salas et al., 2012, that selection should focus on identifying candidates with 2015). Typically, instruction has been constrained to formal transportable teamwork competencies as opposed to those that classroom style approaches wherein participants come together are task-contingent. in a face-to-face setting and learn from an instructor. In these sessions, participants usually must plan in advance to As always, identifying team members for cultural fit is attend, register, commute to the classroom, and have protected paramount when staffing teams, even as teams become more time in their calendars to participate in training, all of which and more dynamic and short-lived. It is potentially even can present as barriers to the communication of necessary more important for staff working on rapid-cycle teams to information. Traditional approaches may be suitable for teaching Frontiers in Psychology | www.frontiersin.org 1830 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era transportable competencies but may no longer be sufficient for • How do we embed interventions within the team imparting other types of knowledge, skills, and attitudes when performance context so they are available precisely needed. They also may not reflect how learning actually takes when needed? place. Much of learning is experiential, occurring in informal or on-the-job real-world settings (Shank, 2012). • What tools and resources beyond training (whether classroom-based or on-the-job) can help teams respond On-the-job training is not a new idea. Adult learners want flexibly to changing demands? to have access to information, practice, and feedback when they need it most and since experience plays a major part in how we • How should interventions be employed given the learn and perform, the thought of incorporating these lessons movement toward globalization and the rapid advent of at work, where they are most relevant is attractive. Compound new technology? this proclivity for convenience and applicability of materials to current work processes with a dynamic environment and the • To whom should interventions target when attempting to implication is that interventions intended to improve teamwork maximize teamwork and team performance? and team performance may be better presented on-the-job, to actual team incumbents. Furthermore, with changing conditions, Digitization and Technology whether they are reconfigured team membership, a new team goal, changing interdependencies, or new task assignments and Perhaps the biggest factor influencing how we define, work responsibilities, there is greater need to have access to relevant in, and study teams is digitization. Digital technologies information and interventions real-time. Teams may not have the are radically changing the world and the ways we live and time or resources to schedule formal training off-site, taking time work. Face-to-face meetings have given way to phone and away from their jobs and the work that needs to be accomplished. video conferencing; paper-based mail has been replaced by email; typewriters exchanged for laptops and smart The natural next question is what delivery mechanisms phones; wired connections substituted for wireless always- can be used that would be suitable for these demands? connected devices. Team members are able to communicate Technology is likely to play a large role. Access to online with each other across time and distances in ways that were repositories of information that teams and individual team previously impossible. Tannenbaum et al. (2012) outline members can access and download at will would be ideal. many of the advantages and pitfalls that technology has for Of course, incorporating demonstration, practice, and feedback teamwork. Among the perceived advantages are greater ability opportunities into these materials will be equally important and to collaborate over distance (enabling the collaboration of may require more creative approaches when a knowledgeable experts across the globe), automatization of routine tasks, instructor, mentor, or coach is unavailable to provide teams swifter communication, and flexibility in scheduling. These with direction and sensemaking of content. Teams would also characteristics have the potential to enhance teamwork and need access to reliable equipment like internet and computers team effectiveness, but they come with their own set of capable of presenting content. For many teams, these materials challenges that may off-set potential gains. For example, it are easily accessed but for teams such as military, medical, is easier than ever to work non-traditional hours. While the and construction teams that work in a variety of settings flexibility afforded by technology may be believed to facilitate equipment of this type may not be already provided. Take individual employee productivity, it can also invade personal construction teams as an example; these teams may not have time for non-work activities and create dissatisfaction with access to reliable WiFi while on a job site. However, most work-life balance (Barber et al., 2019). While employees are Western employees do have access to smartphones and data working longer hours, this does not necessarily translate plans. Practitioners and researchers should consider how these into greater productivity as they forego necessary rest technologies can be tapped as a platform for accessing team and down time needed for renewal (Fritz et al., 2010). resources real-time. Furthermore, technology-mediated collaboration can create lags in information exchange, more misunderstandings, fewer Finally, interventions and delivery of content for modern information seeking attempts, and less coherent messages teams needs to be bite-sized and digestible, with only relevant (Andres, 2012). information being presented. Teams operating in dynamic environments may not have the time to muddle through Digitization and technology may underlie most, if not all, excessive content when there is work to be completed and of the challenges and advancements we see in modern teams. very little, if any, dedicated time for additional learning. However, much like Pandora, we cannot put our digital tools Couple that with recent estimates that the adult attention back in their boxes. The world in which we collaborate is not like span may be as short as eight seconds (Microsoft, 2015), and the world of yesterday; but neither does tomorrow’s world look there is further evidence for the need to keep intervention like that which we see today. The technologies we are using now and informational content brief. Finding the balance will likely be outdated within a decade, and teams will continue between how little content is essential and how much to evolve. Future teams research centered on technology and content is excessive will be a challenge moving forward, digitization might explore: especially when individual learner needs will naturally vary. Regarding team interventions, we propose the following • How can technology be leveraged to facilitate increasingly research questions: important team processes, such as creativity? • What impact does technology have on team and team member well-being? Frontiers in Psychology | www.frontiersin.org 1841 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era • Does technology need to mimic the advantages of Wageman et al., 2012). Additionally, as Tannenbaum et al. face-to-face interactions or can teamwork be organized to (2012) indicated, the need for future research is exacerbated by better leverage the advantages of technology? conflicting evidence (e.g., membership fluidity). To understand what novel thinking and research is necessary, we must first • How do technologies impact the emergence of team states unpack the defining components of teams. Thus, the purpose such as cohesion, trust, identity, and adaptation? of this paper was delineate how the traditional defining characteristics of teams are actually being represented in the • How do we limit the invasiveness of technology within real working environment and offer avenues for investigators our collaborations? to conduct future research to better unpack the theorizing and implications surrounding teams. We hope that future CONCLUSION researchers begin to dissect the theory regarding and surrounding teams with finer detail to advance an accurate depiction of Teams have been ubiquitous, so there have been longstanding contemporary teams. theories and research. However, teams are very different given the macro trends in organizations and tasks. Consequently, these AUTHOR CONTRIBUTIONS well-established theories and methodologies may necessitate some modernizing as the landscape of teams looks very Both authors contributed substantially to the development of differently in today’s society. Contemporary teams and ideas, drafting, and preparation of the final manuscript. collaborations require new thinking and approaches to gain real insights and answer enlightening questions (Murase et al., 2012; REFERENCES Bell, S. T. (2007). Deep-level composition variables as predictors of team performance: a meta-analysis. J. Appl. Psychol. 92, 595–615. doi: 10.1037/0021- Aime, F., Humphrewy, S., DeRue, D. S., and Paul, J. B. (2014). The riddle of 9010.92.3.595 heterarchy: power transition in cross-functional teams. Acad. Manag. J. 57, 327–352. doi: 10.5465/amj.2011.0756 Bell, S. T., Brown, S. G., Colaneri, A., and Outland, N. (2018a). Team composition and the ABCs of teamwork. Am. Psychol. 73, 349–362. doi: 10. Alderfer, C. P. (1976). “Change processes in organizations,” in Handbook of 1037/amp0000305 Industrial and Organizational Psychology, ed. M. Dunnette (Chicago, IL: Rand McNally), 1591–1638. Bell, S. T., Brown, S. G., and Weiss, J. A. (2018b). A conceptual framework for leveraging team composition decisions to build human capital. Hum. Resour. Alderfer, C. P. (1977). “Group and intergroup relations,” in Improving Life at Manag. Rev. 28, 450–463. doi: 10.1016/j.hrmr.2017.06.003 Work, eds J. R. Hackman, and J. L. Suttle (Santa Monica, CA: Goodyear), 227–296. Bresman, H., and Zellmer-Bruhn, M. (2013). The structural context of team learning: effects of organizational and team structure on internal and external Alderfer, C. P. (1980). “Consulting to underbounded systems,” in Advances in learning. Organ. Sci. 24, 1120–1139. doi: 10.1287/orsc.1120.0783 Experiential Social Processes, eds C. P. Alderfer, and C. L. Cooper (Chichester: John Wiley and Sons), 267–295. Bush, J. T., LePine, J. A., and Newton, D. W. (2018). Teams in transition: an integrative review and synthesis of research on team task transitions and Anderson, N. R., and West, M. A. (1998). Measuring climate for work group propositions for future research. Hum. Resour. Manag. Rev. 28, 423–433. innovation: development and validation of the team climate inventory. J. doi: 10.1016/j.hrmr.2017.06.005 Organ. Behav. 19, 235–258. doi: 10.1002/(SICI)1099-1379(199805)19:3<235:: AID-JOB837>3.0.CO;2-C Bushe, G. R., and Chu, A. (2011). Fluid teams: solutions to the problems of unstable team membership. Organ. Dyn. 40, 181–188. Andres, H. P. (2012). Technology-mediated collaboration, shared mental model and task performance. J. Organ. User Comput. 24, 2558–2580. Campion, M. A., Medsker, G. J., and Higgs, A. C. (1993). Relations between work group characteristics and effectiveness: implications for designing effective Argote, L., Insko, C. A., Yovetich, N., and Romero, A. A. (1995). Group learning work groups. Pers. Psychol. 46, 823–847. doi: 10.1111/j.1744-6570.1993. curves: the effects of turnover and task complexity on group performance. tb01571.x J. Appl. Soc. Psychol. 25, 512–529. doi: 10.1111/j.1559-1816.1995.tb01765.x Campion, M. A., Papper, E. M., and Medsker, G. J. (1996). Relations between work Arrow, H., and Crosson, S. (2003). Musical chairs: membership dynamics in team characteristics and effectiveness: a replication and extension. Pers. Psychol. self-organized group formation. Small Group Res. 34, 523–556. doi: 10.1177/ 49, 429–452. doi: 10.1111/j.1744-6570.1996.tb01806.x 1046496403254585 Cannon-Bowers, J. A., and Bowers, C. (2011). “Team development and Arrow, H., and McGrath, J. E. (1995). Membership dynamics in groups at work: a functioning,” in APA Handbooks in Psychology. APA Handbook of Industrial and theoretical framework. Res. Organ. Behav. 17, 373–411. Organizational Psychology, Vol. 1, Building and developing the organization, ed. S. Zedeck (Washington, DC: American Psychological Association), 597–650. Baard, S. K., Rench, T. A., and Kozlowski, S. W. (2014). Performance adaptation: a theoretical integration and review. J. Manag. 40, 48–99. doi: 10.1177/ Cannon-Bowers, J. A., Tannenbaum, S. I., Salas, E., and Volpe, C. E. (1995). 0149206313488210 “Defining competencies and establishing team training requirements,” in Team Effectiveness and Decision Making in Organizations, eds R. Guzzo and E. Salas Barber, L. K., Conlin, A. L., and Santuzzi, A. M. (2019). Workplace telepressure and (San Francisco, CA: Jossey-Bass), 333–380. work life balance outcomes: the role of work recovery experiences. Stress Health doi: 10.1002/smi.2864 [Epub ahead of print]. Carson, J. B., Tesluk, P. E., and Marrone, J. A. (2007). Shared leadership in teams: an investigation of antecedent conditions and performance. Acad. Manag. J. 50, Beal, D. J., Cohen, R. R., Burke, M. J., and McLendon, C. L. (2003). Cohesion and 1217–1234. doi: 10.5465/amj.2007.20159921 performance in groups: a meta-analytic clarification of construct relations. J. Appl. Psychol. 88, 989–1004. doi: 10.1037/0021-9010.88.6.989 Chiu, Y.-T., Khan, M. S., Mirzaei, M., and Caudwell, C. (2017). “Reconstructing the concept of team fluidity for the digitized era,” in Paper Presented at ISPIM Bedwell, W. L., Ramsay, S., and Salas, E. (2012). Helping fluid teams work: a Innovation Summit, Melbourne. research agenda for effective team adaptation in healthcare. Transl. Behav. Med. 2, 504–509. doi: 10.1007/s13142-012-0177-9 Choi, H., and Thompson, L. (2005). Old wine in a new bottle: impact of membership change on group creativity. Organ. Behav. Hum. Dec. Process. 98, Bell, B. S. (2012). Three conceptual themes for future research on teams. Ind. 121–132. doi: 10.1016/j.obhdp.2005.06.003 Organ. Psychol. 5, 45–48. doi: 10.1111/j.1754-9434.2011.01403.x Frontiers in Psychology | www.frontiersin.org 1852 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era Chuang, Y., Church, R., and Zikic, J. (2004). Organizational culture, group diversity of analysis and moderators of observed relationships. J. Appl. Psychol. 87, and intra-group conflict. Team Perform. Manag. 10, 26–34. doi: 10.1108/ 819–832. doi: 10.1037/0021-9010.87.5.819 13527590410527568 Guzzo, R. A., and Dickson, M. W. (1996). Teams in organizations: recent research on performance and effectiveness. Annu. Rev. Psychol. 47, 307–338. doi: 10. Cohen, S. G., and Bailey, D. E. (1997). What makes teams work: group effectiveness 1146/annurev.psych.47.1.307 research from the shop floor to the executive suite. J. Manag. 23, 239–290. Hackman, J. R. (1987). “The design of work teams,” in Handbook of doi: 10.1016/s0149-2063(97)90034-9 Organizational Behavior, ed. J. Lorsch (Englewood Cliffs, NJ: Prentice-Hall), 315–342. Crown, D. F., and Rosse, J. G. (1995). Yours, mine, and ours: facilitating Hackman, J. R. (1990). Groups that Work (and those that don’t). San Francisco, CA: group productivity through the integration of individual group goals. Jossey-Bass. Organ. Behav. Hum. Dec. Process. 64, 138–150. doi: 10.1006/obhd.1995. Hackman, J. R. (2002). Leading Teams. Setting the Stage for Great Performances. 1096 Brighton, MA: Harvard Business School Press. Hackman, J. R. (2012). From causes to conditions in group research. J. Organ. Cummings, J., and Haas, M. (2012). So many teams, so little time: time allocation Behav. 33, 428–444. doi: 10.1002/job.1774 matters in geographically dispersed teams. J. Organ. Behav. 33, 316–341. Harrison, D. A., and Humphrey, S. E. (2010). Designing for diversity or diversity for doi: 10.1002/job.777 design? Tasks, interdependence, and within-unit differences at work. J. Organ. Behav. 31, 328–337. doi: 10.1002/job.608 De Dreu, C. K. W., and Weingart, L. R. (2003). Task versus relationship conflict, Hirst, G. (2009). Effects of membership change on open discussion and team team performance, and team member satisfaction: a meta-analysis. J. Appl. performance: the moderating role of team tenure. Eur. J. Work Organ. Psychol. Psychol. 88, 741–774. 18, 231–249. doi: 10.1080/13594320802394202 Hollenbeck, J. R., Beersma, B., and Schouten, M. E. (2012). Beyond team types Devine, D. J. (2002). A review and integration of classification systems relevant and taxonomies: a dimensional scaling conceptualization for team description. to teams in organizations. Group Dyn. Theory Res. Pract. 6, 291–310. Acad. Manag. Rev. 37, 82–106. doi: 10.5465/armr.2010.0181 doi: 10.1037//1089-2699.6.4.291 Hollenbeck, J. R., Ilgen, D. R., Sego, D. J., Hedlund, J., Major, D. A., and Phillips, J. (1995). Multilevel theory of team decision making: decision performance Devine, D. J., Clayton, L. D., Philips, J. L., Dunford, B. B., and Melner, S. B. (1999). in teams incorporating distributed expertise. J. Appl. Psychol. 80, 292–316. Teams in organizations: prevalence, characteristics, and effectiveness. Small doi: 10.1037/0021-9010.80.2.292 Group Res. 30, 678–711. doi: 10.1177/104649649903000602 Hu, J., and Liden, R. C. (2011). Antecedents of team potency and team effectiveness: an examination of goal and process clarity and servant leadership. J. Appl. Dineen, B. R., and Noe, R. A. (2003). “The impact of team fluidity and its Psychol. 96, 851–862. doi: 10.1037/a0022465 implications for human research management research and practice,” in Ilgen, D. R., Hollenbeck, J. R., Johnson, M., and Jundt, D. (2005). Teams in Research in Personnel and Human Resources Management, eds J. Martocchio organizations: from input-process-output models to IMOI models. Annu. Rev. and G. Ferris (Bingley: Emerald Group Publishing), 1–38. Psychol. 56, 517–543. doi: 10.1146/annurev.psych.56.091103.070250 Katzenbach, J. R., and Smith, D. K. (1993). The discipline of teams. Harv. Bus. Rev. Driskell, J. E., Salas, E., and Driskell, T. (2018). Foundations of teamwork and 71, 111–120. collaboration. Am. Psychol. 73, 334–348. doi: 10.1037/amp0000241 Katzenbach, J. R., and Smith, D. K. (1998). The Wisdom of Teams. Berkshire: McGraw-Hill. Driskell, T., Salas, E., and Driskell, J. E. (2018). Teams in extreme environments: Kazemak, E., and Albert, B. (1988). Learning the secret to teamwork. Healthc. alterations in team development and teamwork. Hum. Resour. Manag. Rev. 28, Financ. Manag. 42, 108–110. 434–449. doi: 10.1016/j.hrmr.2017.01.002 Klein, H. J., and Mulvey, P. W. (1995). Two investigations of the relationships among group goals, goal commitment, cohesion, and performance. Organ. Dube, L. (2014). Exploring how IT professionals experience role transitions at the Behav. Hum. Dec. Process. 61, 44–53. doi: 10.1006/obhd.1995.1004 end of successful projects. J. Manag. Inf. Syst. 31, 17–46. doi: 10.2753/mis0742- Klein, K., and Kozlowski, S. W. J. (2000). Multilevel theory, Research and Methods 1222310102 in Organization. San Francisco, CA: Jossey-Bass. Kozlowski, S. W. J., and Bell, B. S. (2003). “Work groups and teams in Edmondson, A. C. (2012). Teaming: How Organizations Learn, Innovate, and organizations,” in Comprehensive Handbook of Psychology: Industrial and Compete in the Knowledge Economy. San Francisco, CA: Jossey-Bass. Organizational Psychology, Vol. 12, eds W. Borman, D. Ilgen, and R. Klimoski (New York, NY: Wiley), 333–375. Edmonson, A. C., and Harvey, J.-F. (2018). Cross-boundary teaming for Kozlowski, S. W. J., Gully, S. M., Nason, E. R., and Smith, E. M. (1999). innovation: integrating research on teams and knowledge in organizations. “Developing adaptive teams: a theory of compilation and performance across Hum. Resour. Manag. Rev. 28, 347–360. doi: 10.1016/j.hrmr.2017.03.002 levels and time,” in The Changing Nature of Work Performance: Implication for Staffing, Personnel Actions and Development, eds D. R. Ilgen and E. D. Pulakos Espinosa, J. A., Cummings, J. N., and Pickering, C. (2012). Time separation, (San Francisco, CA: Jossey Bass), 240–292. coordination, and performance in technical teams. IEEE Trans. Eng. Manag. Kozlowski, S. W. J., and Ilgen, D. R. (2006). Enhancing the effectiveness of work 59, 91–103. doi: 10.1109/tem.2011.2126579 groups and teams. Psychol. Sci. Public Interest 7, 77–124. doi: 10.1111/j.1529- 1006.2006.00030.x Feitosa, J., Grossman, R., and Salazar, M. (2018). Debunking key assumptions about Lacerenza, C. N., Marlow, S. M., Tannenbaum, S. I., and Salas, E. (2018). teams: the role of culture. Am. Psychol. 73, 376–389. doi: 10.1037/amp0000256 Team development interventions: evidence-based approaches for improving teamwork. Am. Psychol. 73, 517–531. doi: 10.1037/amp0000295 Francis, D., and Young, D. (1970). Improving Work Groups: A Practical Manual for Landrum, S. (2017). Millennials Aren’t Afraid to Change Jobs, and Here’s why. Teambuilding. La Jolla, CA: University Associates. Forbes. Available at: https://www.forbes.com/sites/sarahlandrum/2017/11/ 10/millennials-arent-afraid-to-change-jobs-and-heres-why/#2688a85719a5 Frick, S. E., Fletcher, K. A., Rasay, P. S., and Bedwell, W. L. (2018). Understanding (accessed November 10, 2017). team maladaptation through the lens of the four R’s of adaptation. Hum. Resour. Lanza, P. (1985). Team appraisals. Pers. J. 64, 47–51. Manag. Rev. 28, 411–422. doi: 10.1016/j.hrmr.2017.08.005 LePine, J. A. (2005). Adaptation of teams in response to unforeseen change: effects of goal difficulty and team composition in terms of cognitive ability and Fritz, C., Yankelevich, M., Zarubin, A., and Barger, P. (2010). Happy, healthy, and goal orientation. J. Appl. Psychol. 90, 1153–1167. doi: 10.1037/0021-9010.90.6. productive: the role of detachment from work during nonwork time. J. Appl. 1153 Psychol. 95, 977–983. doi: 10.1037/a0019462 Gilson, L. L., Maynard, M. T., Young, N. C. J., Vartianinen, M., and Hakonen, M. (2015). virtual teams research: 10 years, 10 themes, and 10 opportunities. J. Manag. 41, 1313–1337. doi: 10.1177/0149206314559946 Gladstein, D. L. (1984). Groups in context: a model of task group effectiveness. Adm. Sci. Q. 29, 499–517. doi: 10.2307/2392936 Gonzalez, V., and Mark, G. (2004). “Constant, constant, multi-tasking craziness: managing multiple working spheres,” in Paper Presented at the SIGCHI Conference on Human Factors in Computing Systems, New York, NY. Graesser, A. C., Fiore, S. M., Greiff, S., Andrews-Todd, J., Foltz, P. W., and Hesse, F. W. (2018). Advancing the science of collaborative problem solving. Psychol. Sci. Public Interest 19, 59–92. Gully, S. M., Devine, D. J., and Whitney, D. J. (1995). A meta-analysis of cohesion and performance: effects of level of analysis and task interdependence. Small Group Res. 26, 497–520. doi: 10.1177/1046496495264003 Gully, S. M., Incalcaterra, K. A., Joshi, A., and Beaubien, J. M. (2002). A meta- analysis of team-efficacy, potency, and performance: interdependence and level Frontiers in Psychology | www.frontiersin.org 1863 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era LePine, J. A., Piccolo, R. F., Jackson, C. L., Mathieu, J. E., and Saul, J. R. (2008). Salas, E., Benishek, L. E., Coultas, C., Dietz, A., Grossman, R., Lazzara, E. H., A meta-analysis of teamwork processes: tests of a multidimensional model et al. (2015). Team Training Essentials: A Research-Based Guide. New York, NY: and relationships with team effectiveness criteria. Pers. Psychol. 61, 273–307. Routledge. doi: 10.1111/j.1744-6570.2008.00114.x Salas, E., Dickinson, T. L., Converse, S. A., and Tannenbaum, S. I. (1992). “Toward Lewis, K., Belliveau, M., Herndon, B., and Keller, J. (2007). Group cognition, an understanding of team performance and training,” in Teams: Their Training membership change, and performance: investigating the benefits and and Performance, eds R. W. Swezey and E. Salas (Norwood, NJ: Ablex), detriments of collective knowledge. Organ. Behav. Hum. Dec. Process. 103, 3–29. 159–178. doi: 10.1016/j.obhdp.2007.01.005 Salas, E., Shuffler, M. L., Thayer, A. L., Bedwell, W. L., and Lazzara, E. H. (2011). Malone, T. W., and Crowston, K. (1994). The interdisciplinary study of Understanding and improving teamwork in organizations: a scientifically based coordination. ACM Comput. Surv. 26, 87–119. doi: 10.1145/174666.174668 practical guide. Hum. Resour. Manag. 54, 599–622. doi: 10.1002/hrm.21628 Marks, M. A., Mathieu, J. E., and Zaccaro, S. J. (2001). A temporally based Salas, E., Sims, D. E., and Burke, C. S. (2005). Is there a “big five” in teamwork? framework and taxonomy of team processes. Acad. Manag. Rev. 26, 356–376. Small Group Res. 36, 555–599. doi: 10.1177/1046496405277134 doi: 10.5465/amr.2001.4845785 Salas, E., Stagl, K. C., Burke, C. S., and Goodwin, G. F. (2007). “Fostering team Martin, A., and Bal, V. (2006). The State of Teams. Greensboro, NC: Center for effectiveness in organizations: toward an integrative theoretical framework of Creative Leadership. team performance,” in Modeling Complex Systems: Motivation, Cognition and Social Processes, eds R. A. Dienstbier, J. W. Shuart, W. Spaulding, and J. Poland Mathieu, J., Maynard, M. T., Rapp, T., and Gilson, L. (2008). Team effectiveness (Lincoln, NE: University of Nebraska Press), 185–243. 1997-2007: a review of recent advancements and a glimpse into the future. J. Manag. 34, 410–476. doi: 10.1177/0149206308316061 Salas, E., Tannenbaum, S. I., Kraiger, K., and Smith-Jentsch, K. A. (2012). The science of training and development in organizations: what matters in Mathieu, J. E., Hollenbeck, J. R., van Knippenberg, D., and Ilgen, D. R. (2017). A practice. Psychol. Sci. Public Interest 13, 74–101. doi: 10.1177/152910061243 century of work teams in the Journal of Applied Psychology. J. Appl. Psychol. 6661 102, 452–467. doi: 10.1037/apl0000128 Savelsbergh, C. M. J. H., Poell, R. F., and van der Heijden, B. I. J. M. (2015). Mathieu, J. E., Tannenbaum, S. I., Donsbach, J. S., and Alliger, G. A. (2014). Does team stability mediate the relationship between leadership and team A review and integration of team composition models: moving toward a learning? An empirical study among Dutch project teams. Int. J. Proj. Manag. dynamic and temporal framework. J. Manag. 40, 130–160. doi: 10.1177/ 33, 406–418. doi: 10.1016/j.ijproman.2014.08.008 0149206313503014 Schippers, M. C., Hartog, D. N. D., and Koopman, P. L. (2007). Reflexivity in teams: McGrath, J. E. (1964). Social Psychology: A Brief Introduction. New York, NY: Holt, a measure and correlates. Appl. Psychol. 56, 189–211. doi: 10.1111/j.1464-0597. Rinehart, & Winston. 2006.00250.x McGrath, J. E., Arrow, H., and Berdahl, J. L. (2000). The study of groups: Shank, P. (2012). Smart Companies Support informal Learning. Available past, present, and future. Pers. Soc. Psychol. Rev. 4, 95–105. doi: 10.1207/ at: https://www.elearningguild.com/insights/index.cfm?id=159&action= S15327957PSPR0401_8 viewonly&utm_campaign=research-inf12&utm_medium=link&utm_source= lsmag&_ga=2.249953734.1734585252.1543552949-1329011644.1543552949 Microsoft (2015). Attention Spans. Canada: Consumer Insights, Microsoft Canada. (accessed August 16, 2012). Miller, L. K., and Hamblin, R. L. (1963). Interdependence, differential Shea, G. P., and Guzzo, R. A. (1987). “Groups as human resources,” in Research in rewarding, and productivity. Am. Sociol. Rev. 28, 768–778. doi: 10.2307/208 Personnel and Human Resources Management, Vol. 5, eds K. M. Rowland and 9914 G. R. Ferris (Greenwich, CT: JAI Press), 323–356. Mitchell, T. R., and Silver, W. S. (1990). Individual and group goals when workers are interdependent: effects on task strategies and performance. J. Appl. Psychol. Shuffler, M. L., and Carter, D. R. (2018). Teamwork situated in multiteam 75, 185–193. doi: 10.1037//0021-9010.75.2.185 systems: key lessons learned and future opportunities. Am. Psychol. 73, 390–406. Morgan, B. B. Jr., Salas, E., and Glickman, A. S. (1993). An analysis of team doi: 10.1037/amp0000322 evolution and maturation. J. Gen. Psychol. 120, 277–291. doi: 10.1080/ 00221309.1993.9711148 Summers, J. K., Humphrey, S. E., and Ferris, G. R. (2012). Team member Mortensen, M., and Haas, M. R. (2016). Rethinking Team Boundaries. Available change, flux in coordination, and performance: effects of strategic core roles, at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2841150 (accessed information transfer, and cognitive ability. Acad. Manag. J. 55, 314–338. September 21, 2016). doi: 10.5465/amj.2010.0175 Murase, T., Doty, D., Wax, A., DeChurch, L., and Contractor, N. S. (2012). Teams are changing: time to “think networks”. Ind. Organ. Psychol. 5, 41–44. Sundstrom, E., De Meuse, K. P., and Futrell, D. (1990). Work teams: applications doi: 10.1111/j.1754-9434.2011.01402.x and effectiveness. Am. Psychol. 45, 120–133. doi: 10.1037//0003-066x.45. Okhuysen, G. A., and Bechky, B. A. (2009). 10 coordination in organization: 2.120 an integrative perspective. Acad. Manag. Ann. 3, 463–502. doi: 10.5465/ 19416520903047533 Tannenbaum, S. I., Mathieu, J. E., Salas, E., and Cohen, D. (2012). Teams are O’Leary, M. B., Woolley, A. W., and Mortensen, M. (2011). “Multiteam changing: are research and practice evolving fast enough. Ind. Organ. Psychol. membership in relation to multiteam systems,” in Multiteam Systems: An 5, 2–24. doi: 10.1111/j.1754-9434.2011.01396.x Organization form for Dynamic and Complex Environments, eds S. J. Zaccaro, M. A. Marks, and L. A. DeChurch (New York, NY: Routledge), van de Brake, H. J., Walter, F., Rink, F. A., Essens, P. J. M. D., and van der Vegt, 141–172. G. S. (2018). The dynamic relationship between multiple team membership and Rasmussen, T. H., and Jeppesen, H. J. (2006). Teamwork and associated individual job performance in knowledge-intensive work. J. Organ. Behav. 39, psychological factors: a review. Work Stress 20, 105–128. doi: 10.1080/ 1219–1231. doi: 10.1002/job.2260 02678370600920262 Richardson, J. (2010). An Investigation of the Prevalence and Measurement of Teams Van der vegt, G. S., Emans, B. J. M., and van de vliert, E. (2001). Patterns of in Organisations: The Development and Validation of the Real Team Scale. interdependence in work teams: a two-level investigation of the relations with Doctoral dissertation, Aston University, Birmingham. job and team satisfaction. Pers. Psychol. 54, 51–69. doi: 10.1111/j.1744-6570. Robert Half (2018). Does Job Hopping Help or Hurt Your Career? Available 2001.tb00085.x at: http://rh-us.mediaroom.com/2018-04-05-Does-Job-Hopping-Help-Or- Hurt-Your-Career?utm_campaign=Press_Release&utm_medium=Link& Wageman, R. (1995). Interdependence and group effectiveness. Adm. Sci. Q. 40, utm_source=Press_Release (accessed April 5, 2018). 145–180. doi: 10.2307/2393703 Rousseau, V., Aubé, C., and Savoie, A. (2006). Teamwork behaviors: a review and an integration of Frameworks. Small Group Res. 37, 540–570. doi: 10.1136/ Wageman, R., Gardner, H., and Mortensen, M. (2012). The changing ecology bmjopen-2015-007685 of teams: new directions for teams research. J. Organ. Behav. 33, 301–315. Saavedra, R., Earley, P., and Van Dyne, L. (1993). Complex interdependence in doi: 10.1002/job.1775 task-performing groups. J. Appl. Psychol. 78, 61–72. doi: 10.1037/0021-9010. 78.1.61 Wageman, R., Hackman, J. R., and Lehman, E. (2005). Team diagnostic survey: development of an instrument. J. Appl. Behav. Sci. 41, 373–398. doi: 10.1177/ 0021886305281984 Weingart, L. R. (1992). Impact of group goals, task component complexity, effort, and planning on group performance. J. Appl. Psychol. 77, 682–693. doi: 10. 1037/0021-9010.77.5.682 Weldon, E., and Weingart, L. R. (1993). Group goals and group performance. Br. J. Soc. Psychol. 32, 307–334. doi: 10.1111/j.2044-8309.1993.tb01003.x Frontiers in Psychology | www.frontiersin.org 1874 May 2019 | Volume 10 | Article 1006

Benishek and Lazzara Teams in a New Era West, M. A. (2004). Effective Teamwork: Practical Lessons from Organizational Conflict of Interest Statement: The authors declare that the research was Research. Oxford: Blackwell. conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. West, M. A., Borrill, C. S., and Unsworth, K. L. (1998). “Team effectiveness in organisations,” in International Review of Industrial and Organisational The handling Editor declared a shared affiliation, though no other collaboration, Psychology, eds C. L. Cooper and I. T. Robertson (Chichester: Wiley), 1–48. with one of the authors LB. West, M. A., and Lyubovnikova, J. (2012). Real teams or pseudo teams: the Copyright © 2019 Benishek and Lazzara. This is an open-access article distributed changing landscape needs a better map. Ind. Organ. Psychol. 5, 25–55. under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original Wildman, J. L., Thayer, A. L., Rosen, M. A., Salas, E., Mathieu, J. E., and Rayne, author(s) and the copyright owner(s) are credited and that the original publication S. R. (2011). Task types and team-level attributes: synthesis of team classification in this journal is cited, in accordance with accepted academic practice. No use, literature. Hum. Resour. Dev. Rev. 11, 37–129. distribution or reproduction is permitted which does not comply with these terms. Zander, A. (1977). Groups at Work. San Francisco, CA: Jossey-Bass. Ziller, R. C. (1965). Toward a theory of open and closed groups. Psychol. Bull. 64, 164–182. doi: 10.1037/h0022390 Frontiers in Psychology | www.frontiersin.org 1885 May 2019 | Volume 10 | Article 1006

SYSTEMATIC REVIEW published: 15 May 2019 doi: 10.3389/fpsyg.2019.00811 What We Know About Team Dynamics for Long-Distance Space Missions: A Systematic Review of Analog Research Suzanne T. Bell 1*, Shanique G. Brown 2 and Tyree Mitchell 3 1 Department of Psychology, DePaul University, Chicago, IL, United States, 2 Department of Psychology, Wayne State University, Detroit, MI, United States, 3 School of Leadership & Human Resource Development, Louisiana State University, Baton Rouge, LA, United States Edited by: Background: To anticipate the dynamics of future long-distance space exploration Eduardo Salas, mission (LDSEM) teams, research is conducted in analog environments (e.g., Antarctic Rice University, United States expeditions, space chamber simulations), or environments that share key contextual features of LDSEM such as isolation and confinement. We conducted a systematic Reviewed by: review of research conducted on teams in LDSEM-analog environments to identify which Mathias Basner, factors have been examined with quantitative research, and to summarize what the University of Pennsylvania, studies reveal about team dynamics in LDSEM-analog environments. United States Methods: We used a comprehensive search strategy to identify research on teams that Gloria Rakita Leon, lived and worked together. Data on team dynamics were extracted where possible, and University of Minnesota Twin Cities, sources were coded for key contextual features. The data did not lend themselves to traditional meta-analysis. We used two approaches to summarize the data: a weighted United States averages approach when the study reported enough data to calculate an effect size, and descriptive figures when data across studies were directly comparable. *Correspondence: Suzanne T. Bell Results: Seventy-two sources met our inclusion criteria, yielding 253 effect sizes and 1,150 data points. Results from our weighted averages approach suggested [email protected] that the team cohesion and performance relationship may be operating differently in isolated and confined environments than other teams that lived and worked together Specialty section: (e.g., military teams), and that, given the available data, we can say very little about This article was submitted to the magnitude and direction of the relationship. Our descriptive figures revealed important trends: (a) team members in longer missions generally spent less social time Organizational Psychology, together than shorter missions; (b) consistent team efficiency over time was typical, a section of the journal whereas decreased team efficiency over time was atypical; (c) by 40% of mission Frontiers in Psychology completion or 90 days, all teams reported at least one conflict, (d) commanders’ written communication with mission control decreased in length over time, and (e) Received: 01 December 2018 team mood dynamics did not consistently support the third-quarter phenomenon. Accepted: 26 March 2019 Published: 15 May 2019 Citation: Bell ST, Brown SG and Mitchell T (2019) What We Know About Team Dynamics for Long-Distance Space Missions: A Systematic Review of Analog Research. Front. Psychol. 10:811. doi: 10.3389/fpsyg.2019.00811 Frontiers in Psychology | www.frontiersin.org 819 May 2019 | Volume 10 | Article 811

Bell et al. LDSEM Team Dynamics and Performance Conclusions: There are inherent limitations to our study, given the nature of the analog research (e.g., correlational studies, small sample size). Even so, our systematic review provides key insights into team dynamics in LDSEM-analog environments. We discuss the implications of our research for managing future space crews. Importantly, we also provide guidance for future research. Keywords: team dynamics/processes, space exploration, astronaut, conflict, small sample, analog, over time changes, teams and groups INTRODUCTION Russian Academy of Science’s Institute of Biomedical Problems (e.g., Ushakov et al., 2014; Binsted, 2015; Roma, 2015). Extreme teams help to solve complex problems outside of traditional performance environments and have significant Analog settings share similar characteristics of LDSEMs consequences associated with failure (Bell et al., 2018). As a expected to challenge crews and possibly impinge on team type of extreme team, astronaut crews will be expected to dynamics. As examples, analog crews live in a confined space (i.e., live and work under psychologically and physically demanding small living and working spaces with minimal privacy, physical conditions for future long-distance space exploration missions discomfort), are isolated from others (i.e., limited interaction (LDSEMs), such as missions to Mars (Salas et al., 2015b). For with others outside the crew, difficulty in communicating with example, LDSEM astronaut crews will be required to function family), are surrounded by a harsh physical environment (i.e., effectively as a team in isolated and confined environments for an environment in which survival is not possible without up to 30 months (Human Exploration of Mars Design Reference special equipment), have variable workload (i.e., a high and Architecture [DRM] 5.0; Drake, 2009). LDSEMs will require low volume of work at different periods), and have long- crews to operate more autonomously as their communication duration missions (i.e., the team works together for an extended with mission control (MC) will be delayed up to 22 min (DRM; period of time). Each analog may have its strengths and Drake, 2009). Crewmembers will switch between periods of high weaknesses given that not all of the environmental factors and low workload, as well as between individual and team tasks. It may be present in a particular analog. For example, crews in will be necessary for the LDSEM crew to work together seamlessly Antarctic stations experience physical confinement and isolation, for demanding team performance situations such as landing but are typically isolated as smaller crews for shorter periods on Mars, keep conflicts manageable, and provide one another than is expected for LDSEMs. They also have environmental with social support as crewmembers deal with the stressors of cues not available in spaceflight (e.g., daylight). Crews in space prolonged space flight. simulations (e.g., HUBES, SFINCSS) may experience isolation and confinement but are typically not surrounded by a harsh Rationale physical environment. The National Aeronautics and Space Administration (NASA) Research on teams in analog environments has a rich and other space agencies seek to optimize team performance to history. In fact, a number of factors (e.g., compatibility and minimize the risk of mission failure, and work with researchers cohesion, mood, communication, conflict, performance) have from various scientific disciplines to prepare for future LDSEM been investigated in natural analogs (e.g., Antarctic; Wood et al., missions. While meta-analytic investigations of important team 1999; Steel, 2001), space simulations (e.g., HUBES, Mars 105, relationships exist (e.g., team cognition, cohesion, composition, SFINCSS; Gushin et al., 2001; Sandal, 2004; Nicolas et al., 2013), and performance), these investigations include traditional work and isolated and confined laboratory settings (e.g., Emurian team samples and findings may not necessarily generalize to the et al., 1984) dating back to at least the 1960s (e.g., Gunderson LDSEM context (Beal et al., 2003; Bell, 2007; DeChurch and and Nelson, 1963; Altman and Haythorn, 1965; Gunderson and Mesmer-Magnus, 2010; Bell et al., 2011). As such, researchers Ryman, 1967). This research suggests several dynamics unique to collect data in spaceflight and Earth-based analog environments, the LDSEM-analog settings. which are thought to mimic the challenges crews will encounter in LDSEM, to best design, prepare, and support future LDSEM As examples, while a meta-analysis of the traditional team crews and mission teams. Research on natural analogs examines literature suggests that the team cohesion and team performance teams that exist outside of research purposes; examples include relationship is generally small (Beal et al., 2003), team cohesion polar stations in the Antarctic, where teams conduct scientific may be of particular significance when crewmembers live and research while living in an isolated and harsh environment work together and rely on one another for social support (Landon (e.g., Leon et al., 2011). Research in controlled analogs includes et al., 2015). Astronaut journals collected in the International teams that exist specifically for research purposes; examples Space Station (ISS) reveal a decreasing number of positive include teams in HI-SEAS, Human Exploration Research Analog comments about team interaction over the course of a mission (HERA) at Johnson Space Center, and the NEK facility at the (Stuster, 2010). Further, problems associated with poor unit- level team cohesion such as subgrouping and isolation can occur, which have implications for conflict, information sharing, and team performance (Kanas, 1998; Kanas et al., 2009). Frontiers in Psychology | www.frontiersin.org 920 May 2019 | Volume 10 | Article 811

Bell et al. LDSEM Team Dynamics and Performance The psychological health of the crew is likely to be important reviewed quantitative research conducted on teams in LDSEM- for LDSEMs as crews will be living and working in an extreme analog environments. We answer two primary questions with our environment for an extended duration. Communication between systematic review: (1) which factors have been examined with space crews and MC is thought to provide information about quantitative research, and (2) what do these studies reveal about the crew’s psychological health and the crew’s psychological team dynamics in LDSEM-analog environments? climate. Analysis of a space crew’s communication with MC is the standard operating procedure of the psychological support group METHODS in Russian MC and is used to examine crews’ emotional status and the communicators’ coping strategies (Gushin et al., 2012, Study Design and Inclusion Criteria 2016). Among other things, research by Gushin et al., 1997, 2012 indicated that crews decreased the scope and content of their Typically, meta-analysis is preferred for integrating estimates communication to outside personnel over time—a phenomenon of the same relationship of interest across studies; it allows called psychological closing. us to generate cumulative knowledge about a set of studies. The benefits of meta-analysis over narrative reviews have been Some crews have reported changes in mood over time. The widely noted (see Glass, 1976; Schmidt and Hunter, 2015). third quarter phenomenon is the tendency for positive mood Early in our review process, however, we suspected that most levels to decrease while negative mood levels and conflict increase studies conducted on teams in analog environments would not after the midpoint of the mission (Bechtel and Berning, 1991; lend themselves to traditional meta-analysis. Frequentist meta- Steel, 2001; Dion, 2004; Kanas, 2004; Wang et al., 2014). Though analytic techniques can be inappropriate when a limited number mood is typically measured in LDSEM-analog research as an of studies have examined a particular relationship or when individual-level variable, researchers sometimes use the team sample sizes or data do not permit the calculation of an effect mean of individual-level mood scores to represent team mood. size, for example, when data are only reported for a single Team mood is important because it contributes to team emotion, team. Further, a review of the analog research at the individual- which is defined as a team’s affective state that arises from level determined that traditional meta-analytic techniques were bottom-up components such as affective composition, and top- inappropriate (e.g., Shea et al., 2011). Given this, our general down components such as affective context (Kelly and Barsade, approach (e.g., search strategies, coding) was consistent with 2001). Team emotion starts with individual-level moods and best practices in meta-analysis in organizational psychology (e.g., emotions and is then shared with the team either implicitly Schmidt and Hunter, 2015); however, we retained a broader set through emotional contagion or explicitly through means such of studies and ultimately used alternative analytic approaches as affect management. Environmental context such as lighting to summarizing the data. Our reporting is consistent with the and physical layout can affect moods (see Kelly and Barsade, PRISMA guidelines (Moher et al., 2009) to the extent that they 2001). Thus, a better understanding of how team mood changes apply to non-medical systematic reviews. over time is necessary, especially given the extreme conditions expected for LDSEMs, such as living in a small transit vehicle with We sought to be as inclusive as possible while also striving no natural light. The aforementioned evidence on team cohesion, to ensure that the data were relevant to understanding team communication, and mood are examples of findings that may be dynamics in an LDSEM environment. We applied three general unique to the LDSEM context; this underscores the importance inclusion criteria. First, we retained sources that reported of examining team phenomena in LDSEM-analog environments. quantitative data from teams in LDSEM-analog environments, however, we excluded descriptive case studies and narrative While a body of research examines teams in analog reviews. Second, we identified and included only team-level data environments, to date, it has not been quantitatively summarized. (as opposed to individual-level data). We excluded articles that A quantitative summary of the analog team research is important reported individual-level data that were not tied to a particular for several reasons. First, it summarizes what we know about team (e.g., Bartone et al., 2002), or that were tied to a large polar teams in LDSEM analog environments, given the available data. station (>40 people) but not to a team or a small station (e.g., Doll Specifically, it can provide insights into how team dynamics may and Gunderson, 1971; Palinkas et al., 1989). Third, we included unfold over time for LDSEM teams, and be used to benchmark research in which members of the focal team (e.g., the “crew” typical and atypical team dynamics in the LDSEM environment. analog) live and work together for a period. We provide more It also can identify potential threats to LDSEM team dynamics detail on this decision next. and performance. Second, it can help guide future research in analog environments by identifying what areas are in need of Defining an LDSEM-analog environment has challenges more research, new areas for research, and strategies that aid with because a particular extreme environment (e.g., Antarctic winter- knowledge accumulation over time. Guidance for future research overs) may only share some of the same characteristics expected is particularly important given the expense and time required to of LDSEM. All analogs are imperfect approximations of LDSEM, collect analog research. and researchers must weigh the importance of different features of the context in understanding the phenomena of interest. Objectives and Research Questions Because of this, we broadly defined LDSEM-analog research as research in which members of the focal team (e.g., the “crew” The primary purpose of our research was to provide an overall analog) live and work together for a period. We included military picture of the available data on team dynamics and performance teams when they were expressly described as intact teams (e.g., in LDSEM-analog environments. To do this, we systematically combat teams; Ko, 2005; Lim and Klein, 2006) even if the Frontiers in Psychology | www.frontiersin.org 931 May 2019 | Volume 10 | Article 811

Bell et al. LDSEM Team Dynamics and Performance research did not explicitly mention that the unit lived together. from further review. The decision to exclude an article was We did not include military or firefighter training exercises when agreed upon by at least two members of the research team. it was unclear whether the team lived together either while at Seventy-two sources were retained for inclusion and coded for training or while not at training (e.g., Oser et al., 1989; Hirschfeld fidelity characteristics and other moderators, and the quantitative and Bernerth, 2008). We excluded sources that included data data on team dynamics. Of these, 11 different sources (e.g., on children (e.g., Tyerman and Spencer, 1983). We coded journal articles, technical reports) provided enough information features of the analog environment and sample characteristics to calculate effect sizes representing the relationship between a as moderators, rather than excluding studies based on specific predictor and a criterion related to team functioning, and 61 features of the analog (e.g., mission length, autonomy). We chose different sources reported quantitative data on team dynamics this approach so that we could make comparisons across different over time in LDSEM-analog environments but did not include conditions (e.g., in isolated and confined setting, non-isolated enough data to calculate an effect size. confined settings; how phenomena change over time), as opposed to designating arbitrary cutoffs related to fidelity. It is important To extract data, two coding forms were created: one for coding to note, however, that our decision criteria led to the inclusion effect sizes and one for coding data (e.g., means and standard of some missions in which teams lived and worked in an isolated deviations) related to team dynamics over time. When a source and confined setting for shorter-durations (e.g., 6 and 10 days). reported data on 5 or more teams and a predictor and team We retained these in order to be able observe any potential outcome relationships, we coded or calculated an effect size, changes over time, but note that they have lower fidelity in either r or Spearman’s r. For sources with a team sample size regards to duration. <5 we coded quantitative data such as means (or another team- level representation) and within-team standard deviation, when Search Strategy available, for team dynamics across time. We included data that were presented numerically as well as those presented in figures, We used a comprehensive search strategy to obtain quantitative except when the approximate value reported in the figure could research on teams in LDSEM-analog environments. Our efforts not be reasonably estimated (e.g., due to ambiguities in labeling included: (1) searches of 13 databases that ranged from general of the axis). databases such as Google Scholar and EBSCOhost, specialized databases such as the Military and Government Collection and Coding forms were similar in that both captured space agency databases and technical report repositories (e.g., characteristics of the source, the sample, fidelity characteristics, NASA, ESA, JAXA); (2) searches of specific journals such as Acta and information about the predictor and/or criteria. In addition, Astronautica, Aerospace Medicine and Human Performance, a codebook with definitions of the variables and descriptions Human Factors; (3) contacting 29 researchers that we identified for the different categories for each variable was developed. through the NASA taskbook, our project contact at NASA, or We coded fidelity characteristics when they were described or because they frequently publish in the area (e.g., Vinokhodova, could be reasonably assumed by two independent coders, given Leon); (4) posts to listservs (e.g., Science of Team Science, the descriptions provided in the sources. We used the Internet INGRoup, relevant Academy of Management area listservs); to locate information about specific simulations or Antarctic and (5) a review of reference lists of key articles, including stations to complete missing fidelity information, where possible. those from which we were able to obtain an effect size (e.g., Gunderson and Ryman, 1967; Emurian and Brady, 1984), reviews We coded study design as: (a) descriptive, (b) correlational, (c) of similar domains (e.g., Schmidt, 2015), and recent technical quasi-experimental, and (d) experimental. We coded the degree reports on team research funded by NASA (e.g., Bell et al., 2015b; of similarity between the sources’ samples and the anticipated Burke and Feitosa, 2015; DeChurch and Mesmer-Magnus, 2015; characteristics of LDSEM crews in terms of demographic Gibson et al., 2015; Smith-Jentsch, 2015). The search process differences (e.g., gender, national background). We coded included research published until November 2016. Researchers the fidelity of the team to the characteristics expected for were contacted in May 2015. LDSEM crews. Studies were coded as occurring in dangerous environments when the setting had features that required Data Sources, Studies Sections, and individuals to use special equipment (e.g., winter-overs in Antarctic) or posed an imminent threat (e.g., polar bear threat). Data Extraction Studies were coded as isolated when team members were limited in physical interaction with outside parties for a substantial In total, we identified approximately 309 sources (e.g., books, period of time during the study, and confined when they technical reports, dissertations, journal articles, and conference primarily operated in a highly restricted space. For example, papers) for possible inclusion. To better understand the nature of winter-overs in small Antarctic stations or space simulations were the available data, we sorted the 309 sources into three categories: coded as an isolated and confined environment. Autonomy was (1) sources that included quantitative data with a team-level coded as high, moderate, low, or not reported. Many studies did sample size of 5 or greater, for which a team-level effect size not describe the level of autonomy in detail and were coded between a predictor and criteria related to team functioning as “not reported.” Mission length was coded as the total of could be generated; (2) sources that included quantitative data number of days in the team’s life span. Ongoing teams such as on fewer than 5 teams or only data for one variable over time; firehouses (e.g., Kniffin et al., 2015) were coded to the max of the and (3) sources that did not provide relevant data for our distribution (e.g., 730 days). We also coded crew size, workload quantitative review. Sources in the third category were excluded amount and variability, how the crew communicated with those Frontiers in Psychology | www.frontiersin.org 942 May 2019 | Volume 10 | Article 811

Bell et al. LDSEM Team Dynamics and Performance outside of the focal crew (e.g., mission control) and whether there meeting between coders, was relatively high (mean agreement of was a time delay in the communication. 87% on the variables that were coded). Discrepancies between the two coders were discussed and agreement was reached using Coder Training and Agreement a consensus approach. When consensus could not be reached with certainty between the two coders, the coders met with the The second and third authors served as coders for this study. primary author to discuss how the characteristic in question The primary author trained the coders on the coding scheme should be coded. After the coding was completed, we inspected described in the previous section. Coders first received a coding the data sets to better understand the nature of the data, to sheet and a codebook that provided descriptive information determine the appropriateness of meta-analysis for summarizing about each category of variables. All three authors then used the the data, and to determine the best way to summarize the codebook and coding sheet to independently code three articles. available evidence. The three authors met to discuss the coding, observe areas of agreement and disagreement, and make modifications to the Analytical Strategy coding sheet and codebook. Next, all three authors recoded the initial set of articles to help establish a frame of reference that Although we were able to locate a relatively large amount of incorporated the modifications made to the coding documents. data for our review, the small sample sizes in most studies (e.g., Disagreements about the coding were resolved during a follow- <5 teams) and the variety of relationships examined in the up meeting using a consensus approach. After the second round effect size studies, suggested the majority of the data did not of coding, a common set of 5 articles was coded to determine lend themselves to traditional meta-analytic techniques. Thus, the efficacy of the coding process and to establish decision we used the following approaches. First, when the team factor rules. When there was little disagreement (i.e., <3 disagreements and team outcome relationship could be represented using an across the variables coded in the studies), two coders coded the effect size, we calculated a weighted average of the effect size remaining articles. A randomly sampled common set of coded from the local (analog) population and the relevant meta-analytic articles indicated that initial agreement, prior to the consensus FIGURE 1 | PRISMA 2009 Flow diagram. Moher et al. (2009). 953 May 2019 | Volume 10 | Article 811 Frontiers in Psychology | www.frontiersin.org

Bell et al. LDSEM Team Dynamics and Performance estimate from the traditional teams literature as a minimum- dynamics over time via a series of figures. We plotted team variance estimate. We used this approach as a means of balancing dynamics over time when data were comparable (e.g., similar the precision that meta-analysis can provide in estimating a scales, similar response formats), and reported for at least three relationship across multiple settings with the high uncertainty different teams across at least two different data sources (e.g., (especially due to small sample sizes, etc.) but localness that a articles, conference presentations). We plotted team dynamics specific effect size generated in an LDSEM-analog environment over time in terms of mission days and over relative time. Relative can provide. We also calculated the average inaccuracy of the time was calculated as the mission day divided by the total estimates and used these to create 95% credible intervals to mission length. Relative mission time was examined given that quantify the uncertainty of the estimates. some effects for factors such as team cohesion and conflict, are thought to different because of the point of the team in the We used equations 1, 2, 3, and A12 from Newman et al. (2007) lifespan (e.g., third quarter phenomenon) rather than the mission in forming our weighted averages. We used estimates from meta- day itself. analyses in the extant literature (e.g., Beal et al., 2003; LePine et al., 2008; Bell et al., 2011) to inform the prior probability RESULTS distribution. We only generated an estimated distribution of the true population local validity when there was a relevant meta- Flow Diagram of the Studies Retrieved for analytic effect reported in the extant literature that could inform the Review our prior distribution. This limited the number of relationships we estimated and narrowed the effects to team performance as Figure 1 depicts the flow diagram of the studies retrieved the outcome. Further, even with performance as the outcome, for review. there were a number of relationships for which we could not locate relevant meta-analyses; the relationships between leader- Study Selection and Characteristics member exchange—the idea that leaders have relationships with their followers that vary in quality (Graen and Uhl-Bien, 1995)— Eleven sources (e.g., journal articles, technical reports) provided and team performance, and between personality characteristics enough data (team n ≥ 5) to generate 253 team-level effect (e.g., conscientiousness) of the leader and team performance sizes that represent a team factor (e.g., team cohesion) and team are examples. We also did not locate relevant meta-analyses outcome (e.g., team performance) relationship. We refer to this for many of the personality and needs variables examined by as our effect size data set. Sixty-one sources included data on Gunderson and Ryman (1967), such as wanted affection and team functioning from fewer than 5 teams; from these sources we nurturance personality. Finally, there were two estimates from were able to glean 1,150 data instances (i.e., data collected on one military teams [e.g., shared mental models from Lim and Klein or more variable at a particular time point) to benchmark team (2006) and collectivism from Ko (2005)] that were already dynamics in LDSEM-analog environments over time. We refer to included in meta-analyses that would have been used in the this as our benchmarking data set. We provide a summary of the calculation of the weighted averages [(DeChurch and Mesmer- fidelity characteristics of our samples in Supplementary Table 1. Magnus, 2010); and Bell (2007) respectively]; we did not estimate local validity of these two estimates. We corrected the observed Synthesized Findings correlations in a given analog study for unreliability of the predictor and criterion in order to match the corrections used Our first research question asked: what factors related to in meta-analyses that were used to inform our prior distribution. team dynamics has quantitative research examined in analog Although we would have preferred not to correct the local validity environments? In the effect size data set, the majority of estimates for unreliability because of the small sample sizes on effects (i.e., 102 effects across 9 studies) represented the which they were based, the majority of the variances used to relationship between a predictor and team performance. Forty- inform the prior distributions were corrected for unreliability. seven effects across 6 studies represented the relationship Newman et al. (2007) indicate the importance of ensuring that between a predictor and cohesion or compatibility, and the the prior and local effects have the same corrections. When remaining effects represented a variety of outcomes that differed reliability was not reported, we used the closest estimation across studies. The specific predictor and criterion relationship of reliability from the most similar research in our data set. examined varied across studies. Predictors included inputs, When the correction resulted in an estimate >1, we did not emergent states, and team process variables (see Marks et al., compute a weighted average. This is because the weighted 2001), personality (e.g., Gunderson and Ryman, 1967), values, averages approach relied on the z transformation, which for leader-member exchange, and team-member exchange (e.g., Ko, values over 1 is undefined. Values exceeded 1 for correlations 2005), compatibility and cohesion (e.g., Gunderson and Nelson, from Gunderson and Nelson (1963) and Gunderson and Ryman 1963), mental models (e.g., Lim and Klein, 2006), conflict (1967), which were based on the same source data (e.g., self- (e.g., Seymour, 1970), leadership (e.g., Lim and Ployhart, 2004), report cooperation and performance) and exceeded 0.90 prior ability, experience, mood, exploratory search, and planning to correction. (e.g., Knight, 2015). Outcome variables included performance effects (e.g., accomplishment, accuracy, time to completion, Second, when the number of teams included in the study was efficiency, and quality), emergent states, team processes, and too few to generate an effect size, and when data across studies other team dynamics such as cohesion, team mood, egalitarian were comparable, we descriptively summarized the data on team atmosphere, viability, team-member exchange, leader-member exchange, exploratory search, and cooperation. The data were Frontiers in Psychology | www.frontiersin.org 964 May 2019 | Volume 10 | Article 811

Bell et al. LDSEM Team Dynamics and Performance largely dependent (i.e., the 253 effects came from only 11 different falls between −0.46 and 0.28. This is rather imprecise, as sources), and a variety of predictor and outcome relationships the prediction interval includes large, moderate, and small were examined. Only the relationship between measures of negative effects, no effect, and small and moderate positive cohesion (e.g., compatibility, spending time together) and effects. Conversely, with 95% certainty, we can describe the team team performance was examined in more than 3 independent cohesion and team performance relationship in the firehouses samples (k = 6). studied as positive and small to moderate (i.e., estimate 7), and in the special operations military teams studied, as positive and In the benchmarking data set, team factors included emergent moderate to large (estimate 13). states, team processes, outcomes, and additional team dynamics markers. For example, emergent states included team cohesion Data for a few additional relationships other than team (e.g., Allison et al., 1991; Vinokhodova et al., 2012), and team cohesion and team performance were also available. The age processes included conflict and interpersonal relations (e.g., Leon homogeneity and team performance (Figure 2, Estimate 2) et al., 2004; Šolcová et al., 2014). Outcomes included performance and the educational level homogeneity and team performance (e.g., Emurian and Brady, 1984) and more subjective outcomes relationships (Figure 2, Estimate 4) in an ICE (e.g., Antarctic such as satisfaction (e.g., Bhargava et al., 2000; Leon et al., station winter-over) were estimated with a large degree of 2004). Finally, other dynamics markers, such as team mood (e.g., imprecision; the prediction interval included positive, negative Kahn and Leon, 2000; Steel, 2001; Bishop et al., 2010), were and no effect. Conversely, with 95% certainty, the true population commonly reported in analog studies. A full list of all team effect between cooperation and team performance is estimated to factors examined for the effect size and benchmarking data sets be positive and large (Figure 2, Estimates 11, 12) for firehouses is available in Supplementary Table 2. and special operations teams. Finally, with 95% certainty, the true population effect between transformational leadership and Our second research question asked what quantitative team performance for special operations teams, and the true research reveals about team functioning in LDSEM-analog population effect between team task-relevant expertise and team environments. We discuss the results of the weighted performance for military training teams are positive and exceed averages approach, and descriptive figures benchmarking a small effect (Figure 2, Estimates 9, 10). team dynamics next. Taken together, there is a high degree of imprecision Weighted Averages Approach associated with estimates of the true predictor and team performance relationships from studies with teams in ICEs. We used our weighted averages approach to provide the Specifically, unlike most of the estimated relationships from best possible estimate of the magnitude and direction of the teams in non-ICE, given the current data, if we retain a 95% relationships between team factors and team outcomes in level of certainty, we have limited to no understanding of the the analog environments, given the available data. Figure 2 size or direction of the relationship of team cohesion and team summarizes the weighted averages results, the credible intervals performance observed in multiple ICE, age homogeneity and around the estimates, and displays the forest plot. Specific team performance in an ICE, and educational homogeneity and information about the local validity information obtained from team performance in an ICE. LDSEM-analog studies, the meta-analytic effects that we used in the calculation of the weighted averages, and the estimated Benchmarking Team Functioning posterior distributions are provided in Supplementary Table 3. Over Time Local validity estimates include team performance with cohesion, age homogeneity, education homogeneity, team learning, Next, we benchmarked team dynamics over time in studies with planning, team task-relevant experience, cooperation, and sample sizes too small to generate a between-team effect size, transformational leadership. but for which data were comparable (e.g., similar measures, similar response formats) on at least three different teams from First, we discuss the team cohesion and team performance at least two different data sources (e.g., articles, conference relationships. Studies 1, 3, and 5 (as noted in Figure 2) were presentations). With this requirement, we were able to generate conducted on teams in isolated and confined environments figures on cohesion, efficiency, team conflict, communication (ICE); each of these studies measured team cohesion and team with MC, and team mood. performance with different operationalizations. Estimates 7 and 13 reflect the team cohesion and team performance relationships Team Cohesion for teams that are sometimes used as LDSEM-analogs but which are not isolated or confined for extended periods (non-ICE). While we identified several studies with cohesion data reported over time from 5 or fewer teams, these data were collected using That data suggest that with 95% certainty, we cannot speak to a variety of cohesion operationalizations making it difficult to the direction or size of the team cohesion and team performance directly aggregate and make for meaningful comparisons across relationship in ICE. For example, estimate 1 reflects the estimated settings. We were able to benchmark a subset of this data by results for the team cohesion and team performance relationship identifying 3 sources with data from 11 teams spending time for data collected in Antarctic stations where team cohesion was together (e.g., social activities, eating meals). We classified these operationalized as self-rated compatibility of station members, activities as evidence as social cohesion. Figure 3A illustrates and team performance was operationalized as self-rated station team cohesion across mission days. Figure 3B plots team achievement. The mean estimated validity is −0.10, and with 95% certainty, we estimate that the true population validity Frontiers in Psychology | www.frontiersin.org 975 May 2019 | Volume 10 | Article 811

Bell et al. LDSEM Team Dynamics and Performance FIGURE 2 | Estimated distributions for the predictor and team performance relationship in analog environments. ICE = analog team was in an ICE environment (e.g., Winter-overs in Antartica) NO = team was living and working together but not in an ICE. Hmgnt = Homogenity. Task-rel Exp = task-relevant experience. Trnsfrm Ldr = Transformational Leadership. The square represents the weighted average local validity population estimate (ρposterior) and the bar represents the 95% credible interval. Specific estimates are provided in the right column as per ρposterior [95% credible interval]. The credible interval can be interpreted as follows: there is a 95% chance that the true population predictor and team performance relationship (ρ) is between the first number and the second number. Number in the left column indicates the analog data source. 1. Gunderson and Nelson (1963), Outcome = self-report team achievement, Antarctic stations; 2 and 4 from Gunderson and Ryman (1967), Outcome = team accomplishment, mixed sources, Antarctic stations; 3. Emurian and Brady (1984), outcome = performance on lab task; 10-day isolated lad experiment; 5. Nelson (1964), outcome = supervisor ratings of individual performance aggregated within station, Antarctic station; 6, 8, and 9 Knight (2015), outcome = team’s time and number of obstacles completed in a final challenge task, military training; 7 and 11 Kniffin et al. (2015), outcome = supervisor rating of performance, firehouses; 10, 12, 13. Ko (2005), outcomes = team performance, mixed sources, special operations teams. cohesion over relative time (i.e., the mission day divided by the team (Eskov, 2011). A number of metrics can be assessed using total mission length). Homeostat, including an efficiency metric (Csh) and leadership tactics. Figure 4A is a plot of team efficiency across mission days. The data reported suggests some fluctuations in cohesion over Figure 4B is a plot of team efficiency over relative time. time. However, two patterns are present. First, it appears team members spend more time together during shorter missions. The data suggest that three teams (i.e., a team in EXEMSI The Concordia, Tara Drift, and Mars 500 missions lasted for and two of the teams in SFINCSS simulations) were relatively 268, 507, and 520 days, respectively. In comparison to shorter consistent in terms of efficiency over time. The HUBES team missions [i.e., Emurian et al., 1978, 1985], which lasted for 6, decreased steadily in efficiency over time. One of the SFINCSS 10 and 12 days, team members in longer simulations generally teams (Group 3) had a sharp decline in efficiency early in the spent less social time together. There was one exception to this: simulation and then steadily increased during the remainder of time together increased sharply at certain points for a team at the simulation. Concordia station. These instances could have been the result of significant events at the station during those periods (Tafforin Descriptive information on team dynamics in the HUBES et al., 2015). It is important to note that we included shorter- and SFINCSS simulations implicate ineffective role structure duration missions to avoid an arbitrary cut off and to observe and conflict as possible triggers of the performance decrements changes over time. The stark contrast between shorter-duration of HUBES and SFINCSS–Group 3. Specifically, in addition to and longer-duration missions on time spent together suggest measures of efficiency, the Homeostat also collects information limited usefulness of shorter-duration studies in understanding on leadership tactics by individual team members as a team cohesion for LDSEM. means of understanding the leadership structure used while completing the task. For SFINCSS group 3, Vinokhodova Team Performance et al. (2002) indicated that the data did not suggest that a role distribution structure had sufficiently developed. Further, Homeostat was used to collect data on team performance the SFINCSS simulation also included a New Year’s Eve across a number of space simulations (e.g., HUBES, SFINCSS). incident between a member of another group and a woman Homeostat is a computer task in which, under time pressure, a in Group 3 of the simulation, which led to tension between team solves tasks that require the coordinated action of the whole crews (Sandal, 2004). The sharp decrease in effectiveness in Frontiers in Psychology | www.frontiersin.org 986 May 2019 | Volume 10 | Article 811

Bell et al. LDSEM Team Dynamics and Performance FIGURE 3 | (A) Team cohesion over time. (B) Team cohesion over relative time. Emurian et al. (1978), Emurian et al. (1985), Tafforin (2015). the SFINCSS Group 3 depicted in Figure 4A also happened summarize data that were comparable across multiple teams around this time. For HUBES, Sandal (2001) reports that from different analog environments for the total number of there was evidence of an unstable crew structure; specifically, conflicts reported within crews. Data do not show a consistent the commander’s leadership was challenged during the first trend across teams. Some teams are more variable than others 8 to 10 weeks of the mission. Further, crew relations in in the number of conflict incidents per month, while others the simulation were marked by interpersonal tension and are more stable. Some teams report conflict early on, while alienation of one crew member during later parts of the others do not. By 40% of the mission completion (with this experiment. Taken together, this may suggest ineffective role data the equivalent of at least 90 days) all teams had reported structure, conflict, and alienation as possible threats to a least one instance of conflict. No team had more than six team efficiency. instances of conflict per month with a given target (i.e., the crew or MC). Team Conflict Communication With Mission Control (MC) A few sources (k = 4) reported conflict scores over time for 8 different teams using 2 types of conflict metrics (e.g., Gushin et al. have examined crew communication with MC total number of conflicts reported, Likert scale). Figures 5A,B in several studies (e.g., Gushin et al., 1997, 2001; Gushin Frontiers in Psychology | www.frontiersin.org 997 May 2019 | Volume 10 | Article 811

Bell et al. LDSEM Team Dynamics and Performance FIGURE 4 | (A) Team efficiency over time. (B) Team efficiency over relative time. Le Scanff et al. (1997), Vinokhodova et al. (2001), Eskov (2011). and Yusupova, 2003) and have reported comparable data, were inconsistent across teams. It is interesting to note, that which allowed us to plot the total duration of crew–MC the HUBES crew that had decreasing efficiency over time audio-communication sessions (in seconds) over time (see (Figure 4A) also had shorter audio communication with MC Figures 6A,B), as well the average report length per week of over time (Figures 6A,B). As depicted in Figures 7A,B, average the commander’s end-of-day report to MC (see Figures 7A,B). mission report length to MC per week decreased over the course For the SFINCSS, HUBES, and ECOPSY simulations, audio of the mission in SFINCSS, EXEMSI, and ECOPSY. Gushin communication paralleled the standards of Mir in that 30 min et al. (2012) describe this as the closing of a communication were made available for audio communication every 90 min in channel, or psychological closing. Psychological closing can the daily schedule but use of the time was not required. At the end include a decrease of the communication volume throughout of each day, the commander submitted a written report to MC isolation, decrease in the issues discussed, and preference for on mission status and fulfillment of the daily schedule (Gushin communication partners. et al., 1997, 2001). Data in Gushin and Yusupova (2003) was collected by researchers listening to crew-MC communication It should be noted that there is a wealth of specific details (e.g., once a week (for ISS mission 1) and twice a week (for ISS negative statements, jokes) that can be gleaned and assessed via mission 2). content analysis of within- and between-group communications. Our figures here only reflect report length and total time for As depicted in Figures 6A,B, patterns of average audio- audio communication, which were reported in the same format communication length between the commander and MC across multiple teams. We refer the interested reader to Gushin Frontiers in Psychology | www.frontiersin.org 1980 May 2019 | Volume 10 | Article 811

Bell et al. LDSEM Team Dynamics and Performance FIGURE 5 | (A) Team conflict over time. (B) Team conflict over relative time. Steel (2001), Basner et al. (2014). et al. (2012) and Tafforin (2015) for more detail on the range of LDSEM-analog environments, although it should be noted that communication parameters that have been examined. the scaling reported for Scott Base was 0 to 4 instead of 1 to 5 as in the other simulations. Thus, the winter-over at Scott Team Mood Base may have ratings more similar to Mars 500. Both studies that included teams in ICE for a year or more (e.g., Mars 500, Multiple studies reported the affect of team members using an Antarctic winter-over) showed a spike in team total mood Profile of Mood States (POMS; Shacham, 1983; Curran disturbance around the 1-year mark, and this was confirmed in et al., 1995). POMS captures individuals’ mood via self-report the text of the studies reporting the data (e.g., Steel, 2001; Wang ratings on six dimensions using a 5-point Likert scale. The et al., 2014). Figure 8B, which shows total mood disturbance over dimensions are tension-anxiety, depression-dejection, anger- time relative to the proportion of the mission complete, does not hostility, fatigue-inertia, confusion-bewilderment, and vigor- support a clear third-quarter phenomenon at the team level. activity. To arrive at an overall total mood disturbance score, the first five subscales listed are summed and then the vigor- Team mood also has been operationalized in LDSEM-analog activity subscale is subtracted. Team mood is captured with the environments as the team mean of self-report ratings on the average total mood disturbance across the team. Figures 8A,B positive and negative mood components of the Positive and show team mood over time and team mood over relative time, Negative Affect Schedule (PANAS; Watson et al., 1988, see respectively. Figure 8A shows that the MARS 500 crew reported Leon et al., 2004, 2011, for examples). Figures 9A,B,10A,B, elevated total mood disturbance compared with teams in other show the relationship between affect operationalized as the team Frontiers in Psychology | www.frontiersin.org 1991 May 2019 | Volume 10 | Article 811