Team Flow Among Firefighters: Associations with Collective Orientation, Teamwork-Related Stressors, and Resources

Flow is often experienced under stress-related conditions and associated with improved performance and well-being. Different workplace stressors and resources demonstrably affect individual flow. Considering the team level, our study explores the associations between typical teamwork-related stressors and resources and team flow. Firefighters work under extremely challenging conditions and rely on functioning teamwork processes, thereby constituting a promising sample for examining team flow and its associated factors. An online questionnaire was implemented in Germany examining frequency of team flow, experienced teamwork-related stressors, and resources in firefighting missions. Collective orientation was also measured as a factor potentially associated with team flow. The final sample consisted of 317 experienced firefighters. Multiple regression analyses showed that the more often firefighters experienced stressors of teamwork (e.g., problems regarding leadership or lacking shared mental models) the less often they experienced team flow. Conversely, the experience of resources of teamwork (e.g., good communication and shared mental models) was associated with more frequent team flow. Collective orientation was positively associated with team flow. As team competencies are trainable, the study offers concrete approaches for interventions strengthening teamwork and team flow, thereby benefitting firefighters but also teams in other professions.


Introduction
A second-floor apartment is on fire. Cries for help can be heard. Passers-by alert the fire department. Within minutes firefighters arrive. The situation is hectic and stressful, but the firefighters stay calm. Interdependent tasks are allocated. Clear orders are given, and professionals cooperate. Within minutes, all in the building are rescued and the fire is extinguished.
Such an ideal mission cannot be taken for granted. Many different processes must be coordinated to ensure a good result.
Firefighters are part of so-called High Responsibility Teams (HRTs; Hagemann et al., 2012), working interdependently and under high pressure, handling demanding situations without interruptions, termination, or correction of these. Firefighters bear major responsibility for people's health and safety (Hagemann et al., , 2012, hence they do stressful and onerous work. Flow is a frequently experienced state in stress-relevant situations (e.g., Rheinberg & Manig, 2003;Weimar, 2005) and demonstrable associated with problem-focused coping (Pinquart & Silbereisen, 2010). Flow can be defined as a state in which individuals feel fully absorbed in a challenging task (Csíkszentmihályi, 1975) and is associated with improved performance (Christandl et al., 2018;Demerouti, 2006;Engeser & Rheinberg, 2008;Rijavec et al., 2017) and well-being (Rijavec et al., 2017;Rivkin et al., 2018). Due to its positive consequences, understanding which factors during work strengthen or hinder flow is vital. Interest is rising in transferring the concept of flow to the team level (van den Hout & Davis, 2019). Team flow has been described as a shared experience of flow within a team occurring while working on interdependent personal tasks (van den Hout & Davis, 2019). As described above, teamwork among firefighters occurs in stressful and dynamic environments and interdependent tasks must be accomplished quickly and efficiently. Firefighters thus serve well to investigate team flow and to identify factors related to team flow. This may therefore help to enhance future team performance among firefighters as flow in teams is associated with better team performance (Aubé et al., 2014;Heyne et al., 2011;van den Hout et al., 2019). While various workplace stressors and resources reportedly affect flow experience on an individual level in different ways (for an overview see Peifer & Wolters, 2021), research on team flow remains scarce and heterogeneous . This study aims to reduce this research gap by investigating the associations of team flow with typical teamwork-related stressors and resources in firefighting missions. By examining this particular sample, in which teamwork is a very high priority, we aim to identify starting points for examining associated factors of team flow in more detail, with a view to a later expansion to other contexts.
An attitude known to be significant in successful teamwork is collective orientation (Driskell et al., 2010;Hagemann & Kluge, 2017;Hagemann et al., 2021;Kilcullen et al., 2022). Defined as the "propensity to work in a collective manner in team settings" (Driskell et al., 2010, p. 317), it is amenable to experience and training. As people with a high collective orientation pay more attention to their team members and show better coordination processes (Hagemann et al., 2021;Salas et al., 2008), we further aim to investigate if collective orientation is associated with the experience of team flow.

Team Flow
For clarity, we use the term team flow here, although other terms, such as social flow (Walker, 2010), group flow (Sawyer, 2017) or collective flow (Quinn, 2005) occur. Even if these phenomena are not identical, they have a solid common basis (van den Hout et al., 2018) and conditions and characteristics of team flow can be derived . Team flow is defined as "a shared experience of flow during the execution of interdependent personal tasks in the interest of the team, originating from an optimized team dynamic" (van den Hout & Davis, 2019, p. 28). Besides the characteristics of individual flow -absorption, perceived demand-skill balance, and enjoyment (Peifer & Engeser, 2021) -that are also part of team flow, team flow includes additional characteristics and has additional conditions based on the social nature of the construct and the special demands of teamwork Pels et al., 2018;van den Hout et al., 2018). The four meta-categories derived from the analysis of various descriptions of team flow characteristics and conditions by  are: communication and feedback, shared goal commitment, equal participation, and trust in each other's knowledge, skills, and attitudes.
Team flow is associated with better team performance. For example, van den Hout et al. (2019) found a positive relationship between team flow prerequisites and team performance. Moreover, Heyne et al. (2011) found a positive relationship between flow aggregated at the team level and performance in complex planning tasks. Similar results were reported by Aubé et al. (2014) in project managing tasks. Despite the different survey methods, similar results emerged. Due to the positive effects of flow in teams on team performance it is worthwhile to investigate conducive and inhibiting factors of team flow, especially in the field of firefighting, where poor performance could have far-reaching consequences. Van den Hout and Davis (2022) collected data from workers describing why they experienced team flow and what prevented them from doing so. In our study, we want to apply these generally collected results more specifically to the context of firefighting.
A cooperative firefighting mission entails various task demands. Tasks necessitating management by firefighters throughout a mission include, securing the fire sight, setting up the water supply or rescuing victims. Although taskwork refers to work content (Marks et al., 2001), teamwork is indispensable (Salas et al., 2008). Hence a team, defined as "a set of two or more individuals who interact interdependently and adaptively toward a common goal or objective" (Cannon-Bowers & Salas, 1998, p. 83) performs certain processes together. These team processes are characterized as "members' interdependent acts that convert inputs to outcomes through cognitive, verbal, and behavioral activities directed toward organizing taskwork to achieve collective goals" (Marks et al., 2001, p. 357).
Well-functioning processes within the team perceived as good by the team members, are among the characteristics of team flow . Sawyer (2017) describes the conditions of group flow with different facets of well-functioning teamwork, e.g., close listening to one another or constant and spontaneous communi-cation. Successful coordination of activities, a collective goal (Quinn, 2005), and agreement about roles and procedures (Walker, 2010) are additional conditions of team flow  that are also mentioned as components of process dimensions in the taxonomy of team processes (Marks et al., 2001). Accordingly, well-functioning team processes seem to be important for team flow. Van den Hout et al. (2018) additionally mention sense of unity, sense of joint progress, mutual trust, and a holistic focus as characteristics of team flow. In spite of a variety of descriptions for the characteristics and conditions of team flow, these conditions and characteristics of team flow as a whole exceed those of individual flow . Accordingly, stressors and resources possibly arising in teamwork should be scrutinized more closely as potentially inhibiting or reinforcing team flow.

Stressors and Resources of Teamwork Among Firefighters
Individuals in their work face different work-relevant stressors described as "the degree to which the work environment contains stimuli that require sustained cognitive, emotional, or physical effort" (de Jonge & Dormann, 2006, p. 1359). On the other hand, job resources can be referred to as options for coping (de Jonge & Dormann, 2006). Heinemann et al. (2023) investigated the specific stressors and resources relevant in the field of firefighting. They conducted expert interviews and analyzed accident reports to identify different stressors and resources of teamwork in firefighting missions. Six core categories were identified representing different relevant aspects of teamwork in firefighting: communication, supporting behavior, leadership, shared mental models, organization/coordination, and decision-making. In each of these categories, team member behaviors are described that can be interpreted as stressors or resources of teamwork. For example, possible stressors appearing in firefighting missions are delayed disclosure of information (communication), not pointing out errors of team members (supporting behavior), hectic and uncertain behavior of leaders (leadership), lacking information on the approach and objective of the mission (shared mental model), failure to meet safety standards (organization/coordination) or no clear decision-making processes (decision-making). On the other hand, potential resources within this context are timely transfer of information (communication), checking on each other (supporting behavior), structured and calm leaders (leadership), knowing mission tactics (shared mental models), adherence to standards and safety measures (organization/coordination) or clearly defined decision structures (decision-making) (Heinemann et al., 2023).
The specific stressors and resources of teamwork for firefighters described by Heinemann et al. (2023) can be integrated in a well-known framework of teamwork processes by Marks et al. (2001). They argue that different team processes occur at different points of the task execution, distinguishing between the action phase (e.g., coordination and monitoring) and the transition phase (e.g., evaluation of performance and planning). There are moreover interpersonal processes (e.g., conflict and affect management) influencing team performance. The stressors and resources relevant in firefighting missions can be attributed to all three process dimensions. For example, passing on information or supporting behavior could be included in the action phase, while decision-making about the strategy belongs to the transaction phase. Marks et al. (2001) moreover emphasize that different processes are used simultaneously to manage multitasking. We therefore aim to investigate effects of teamwork-related stressors and resources in the firefighting mission on team flow.

Effects of Stressors and Resources on Team Flow
An approach to understand how stressors and resources could be related to team flow is the Transactional Model of Stress and Flow (Peifer, 2012;Peifer & Tan, 2021). Based on Lazarus and Folkman (1984) the model describes the transactional process possibly leading to stress or flow. According to the Transactional Model of Stress and Flow, influential factors are the demands of the environment, as well as the person's goals, needs, and resources. Stress evolves when an individual perceives a situation as a threat and the demands of the situation exceed the resources available to cope with it. On the other hand, in a situation perceived as a challenge and for a person with enough resources to cope with it flow can be experienced (Peifer, 2012;Peifer & Tan, 2021). Through reappraisal it is possible for a threat to become a challenge (Lazarus & Folkman, 1984).
This model can be applied to team flow. First, teamwork entails demands. In a firefighting mission, different aspects of teamwork (e.g., communication, coordination) are needed to achieve the mission objective. Second, teamwork entails resources which can help to cope with a stressing situation: If team members e.g., give timely adequate information and coordinate actions effectively, the team members feel able to effectively handle the demands of the environment and will more likely interpret the situation as a challenge than as a threat. This contributes to a perceived balance of skills and demands, a core component of team flow . Accordingly, the team members are more likely to experience team flow than stress. Conversely in the absence of such teamwork resources or -even worse -if teamwork stressors appear (e.g., poor communication or confusing coordination), the demand-skill balance is reduced, and stress may occur.
Based on the reported literature, we propose that: Hypothesis 1 Frequency of perceived teamwork-related stressors is negatively associated with frequency of team flow in firefighting missions.
Hypothesis 2 Frequency of perceived teamwork-related resources is positively associated with frequency of team flow in firefighting missions.
As an additional exploratory research question, we further aim to investigate which teamwork stressors and resources have an association with team flow and which do not. Since Heinemann et al. (2023) assigned the stressors and resources found in teamwork in firefighting missions to six subcategories, it seemed reasonable to examine the influence of these subcategories in addition to the overall occurrence of stressors and resources. Accordingly, we examine whether the subcategories communication, supporting behavior, leadership, shared mental models, organization/ coordination and decision-making are related to the team flow experience.

Collective Orientation
In addition to the situational conditions potentially associated with the team flow experience, whether personal attitudes are also related to team flow also merits examination. One attitude related to aspects that are partially described in the conceptual work on team flow is collective orientation (Driskell et al., 2010;Hagemann & Kluge, 2017;van den Hout et al., 2019). It is defined as "the propensity to work in a collective manner in team settings" (Driskell et al., 2010, p. 317). The enjoyment of team membership is one core element of collective orientation. Moreover, highly collectively oriented people focus on the achievement of team goals, value the opinion of others, and are willing to cooperate. They prefer working in a team to working alone (Driskell et al., 2010;Hagemann, 2017) and teams with members with high collective orientation values demonstrate more effective coordination behavior, also in very complex situations (Hagemann & Kluge, 2017;Hagemann et al., 2021). We assume that collective orientation is positively associated to team flow: behavioral consequences of collective orientation are among the conditions and characteristics of team flow. Examples include having a collective goal and successful coordination of activities (Quinn, 2005) or listening closely to one another (Sawyer, 2017). Barthelmäs and Keller (2021) argue that subjective attachment to an activity is important to experience flow. People who like the activity they are involved in are more likely to have an intense flow experience. Regarding collective orientation, people with a high collective orientation should experience team flow during a teamwork activity more frequently and intensely because they are more closely attached to teamwork-related tasks. Conversely, people who enjoy working in a team less should experience team flow less frequently and intensely, being subjectively less attached to the task at hand. In line with that we propose that: Hypothesis 3 Collective orientation is positively associated with frequency of team flow in firefighting missions.

Participants and Design
The sample was recruited Germany-wide via a nationwide mailing-list of the fire brigade in Cologne, distribution through the fire brigade country schools, through advertisement in a firefighter magazine, advertisement on the project homepage, social networks, and personal contacts. The questionnaire was online from January 18 through May 25, 2021. With 1,135 people accessing our questionnaire online, 755 attempted the questionnaire, 438 were excluded due to (1) early termination of the questionnaire and missing values on the study variables (n = 329), (2) insufficient firefighting experience (n = 9), or (3) illogical responses (n = 100).
The final sample consisted of 317 participants, 301 men and 16 women (= 5%), which is fairly representative of firefighters in Germany (approx. 10% of woman in the voluntary fire brigade and approx. 2% in the professional fire brigade; cf. Deutscher Feuerwehr Verband, n.d.). Mean age was 40.71 years (SD = 11.18), range 18-70 years and 65.3% worked in the voluntary fire brigade, 30.6% in the professional fire brigade, and 4.1% in other fire brigades. The participants had worked in their profession for 21.54 years on average, minimum two years, maximum 49 years (SD = 11.09, n = 308).

Stressors of teamwork in firefighting missions
were measured with the recently developed questionnaire by Heinemann et al. (2023) consisting of 26 stressors typical of the firefighting context. For each stressor there was a brief description of a situation in a firefighting mission in which the respective stressor appeared. Participants had to rate on a 5-point rating scale how often (1 = never to 5 = always) they had experienced the respective situations in past firefighting missions. The stressors belong to the six categories communication (COM, 6 items, e.g., 'The information passed on was incorrect, inaccurate and/or incomplete.'), supporting behavior (SUP, 6 items, e.g., 'Firefighters did not pay attention to each other (e.g., did not check on each other; did not point out mistakes and/or hazards to others; did not intervene even when hazards were identified)'), leadership (LS, 4 items, e.g., 'Potential hazards and/or operational standards were not considered by one or more leaders (e.g., no functioning respirator monitoring; inadequate scene reconnaissance).'), shared mental models (SMM, 3 items, e.g., 'Firefighters lacked information on the course of action and objective of the mission (e.g., sequence of steps; mission strategy).'), organization/coordination (O&C, 4 items, e.g., 'Tasks were not distributed at all or not evenly, so some firefighters had to do too many tasks at the same time (e.g., listening in on the radio, issuing materials, and monitoring breathing apparatus).') and decision-making (DM, 3 items, e.g., 'It was not clearly defined who had to make this decision and/or the decision was made by someone who did not have the necessary information/skills.'). Total score was calculated using the mean of the items. This reflects a value that summarizes the occurrence of all the stressors of teamwork. A confirmatory factor analysis (CFA) was used to check the theoretically derived structure with 26 items loading on six subcategories, yielding in turn the total score. Furthermore, relevant correlations (indicated by the modification indices) were integrated in the model. The model which proved robust to violation of the multivariate normal distribution assumption by using a robust estimator (MLM) had an acceptable fit (TLI: 0.92, RMSEA: 0.05, SRMR: 0.05) referring to the recommendations of Vandenberg and Lance (2000). Furthermore, internal consistency was checked. Cronbach's α for the total score was 0.92. The subcategories had a Cronbach's α between 0.47 and 0.79 (see Table 1). As mentioned by Ziegler et al. (2014), low internal consistency is a common problem in short scales but does not necessarily mean that reliability is low. Since the items represent relevant aspects of each subconstruct we avoided excluding single items from subcategories. The model fits when subcategories with low reliability are included so they are used in the analysis. Heinemann et al. (2023). Using the same structure as for stressors, a total of 23 teamwork resources were assessed in terms of frequency of occurrence. Again the six subcategories were communication (COM, 6 items, e.g., 'Information was passed on in time and/or brief consultations took place.'), supporting behavior (SUP, 5 items, e.g., 'The firefighter was responsive to the other person's needs and attempted to provide support.'), leadership (LS, 4 items, e.g., 'One or more leaders behaved in a calm and level-headed, structured and/or supportive manner. This reduced stress and transferred to others.'), shared mental models (SMM, 3 items, e.g., 'The firefighters knew the tasks in the other functions well enough that they could assess what the other person just needed (e.g., information) and/or how their own actions affected the others.'), organization/coordination (O&C, 3 items, e.g., 'The firefighters recognized possible tasks, acted independently and, if necessary, coordinated work steps on their own initiative (e.g. information about their own course of action).') and decision-making (DM, 2 items, e.g., 'Firefighters used new information to reassess the situation and decided to change the plan if necessary.'). The model with the 23 items loading on the six subcategories yielding in turn the total score had negative variances in a CFA with a robust estimator (MLM). By excluding the subcategory decision-making these problems were resolved and the model had a good fit (TLI: 0.97, RMSEA: 0.03, SRMR: 0.05; relevant correlations [indicated by the modification indices] were integrated). As the subcategory decision-making had a very low reliability (Spearman-Brown coefficient: 0.24) and, as mentioned above, no single items should be excluded from a subcategory, it was decided to exclude both items from interferential statistics to analyze the data by using a good fitting model. In the other subcategories Cronbach's α ranged from 0.52 to 0.80. The total score representing the occurrence of teamwork-related resources (not including decisionmaking) in firefighting missions in general had a Cronbach's α of 0.89.

Resources of teamwork during firefighting missions were also measured with the instrument developed by
Team flow was assessed with 12 self-constructed items inspired by the Flow-Short-Scale  and the Team-Flow-Monitor (van den Hout et al., 2019). Assuming that team flow results from characteristics of the individual flow experience and contains additional components , we decided to adapt the FKS used to measure the characteristics of individual flow experience to the team flow and to add items assessing the additional conditions and characteristics of team flow (see Supplementary Material, Table S1). Thus, the various items were supplemented by specifically eliciting experiences in the team. This means that the object of measurement was team flow, but it was measured using individual selfreports. Participants rated on a 6-point rating scale (1 = never to 6 = [almost] always) how often they experienced team flow in firefighting missions. One example item is 'We felt optimally challenged as a team'. A total score was used since the team flow experience represents an interplay of the different aspects and can only be assumed to be a team flow experience if they occur simultaneously. Cronbach's α was 0.93. The structure was checked with the help of a CFA with a robust estimator (MLM). The model fit is good (TLI: 0.97, RMSEA: 0.06, SRMR: 0.03) assuming that the 12 items load on the general factor of team flow and relevant correlations (indicated by the modification indices) were considered. 5.6 6 6.1 6.2 6.3 6.4 6.5 6.   The factor loadings of the CFAs can be found in the Supplementary Material - Table S2, S3 and S4.
Collective orientation was measured with the German-language version of the Collective Orientation Scale (Hagemann, 2017). The scale consists of 16 items (e.g., 'I like working with others') rated on a 5-point Likert scale from 1 = totally disagree to 5 = totally agree. Cronbach's α was 0.83.
Control variables. Age, years of experience of firefighting missions, and number of firefighting missions experienced were collected as control variables. The aim was to check whether experience of missions influenced the team flow experience. This assumption is based on the expertise effect (Rheinberg, 2006), according to which in complex activities, basic skills must first be automated before flow can be experienced. In the context of firefighting missons, this could mean that experienced firefighters experience more flow due to their expertise. In the study by van den Hout et al. (2019), age was also included as a control variable with regard to team flow experience, so it seemed reasonable to also include it here. Furthermore, age is often associated with more experience potentially supporting the expertise effect.

Data Analysis
The robust CFAs (using the MLM estimator) were performed with the software program R and the package "Lavaan" (Rosseel, 2012). Data were analyzed with the IBM SPSS statistics package V28. First, we excluded outliers (> 2.5 SD) on the study variables (not for age and years of working experience). Therefore, n varies for the different analyses. Next, we checked with regressions if the control variables age, years of working experience and number of firefighting missions were associated with team flow. As the three control variables were highly correlated, we used three separate regressions to avoid problems with multicollinearity. Neither age (F(1, 309) = 0.39, p = 0.535), nor years of working experience (F(1, 300) = 0.03, p = 0.866) or number of firefighting missions (F(1, 300) = 0.18, p = 0.674) predicted team flow. Furthermore, the control variables were not correlated with team flow (see Table 1), so they were not considered in the later stages of the analysis.
To investigate Hypotheses 1 and 2 we computed a multiple regression. To explore which stressors or resources were associated with team flow we also used multiple regressions. Requirements for the analyses were met for the most part (outliers, linearity, multicollinearity, normal distribution). We used bootstrapping (5,000 samples) to control the results of the analysis as slight violations of the homoscedasticity assumption occurred and to generate more robust parameter estimators.
To investigate the relationship between collective orientation and team flow (Hypothesis 3) we calculated the regression coefficient.
A priori power analysis for the regression model was conducted using G*Power (Faul et al., 2009). Since to the best of our knowledge there are so far no studies examining the relationship of team flow with specific teamwork stressors and resources in the firefighting context, we assumed medium effect sizes. A power of 0.95 and a significance level of 0.05 were defined for all analyses. For the regression with two predictors and an assumed effect size of f² = 0.15, 107 participants were required. This requirement was met. Table 1 presents the means, standard deviations, internal consistencies and intercorrelations between the study variables.

Descriptive Results
In terms of collective orientation, the sample reaches a mean of 3.65 (SD = 0.46). This value is slightly above the mean values of past samples (e.g., Hagemann et al., 2021: M = 3.12 [0.46], N = 116). With regard to stressors and resources of teamwork, it can be seen that, on average, resources are perceived more frequently than stressors. This is true for the total score as well as for the subcategories. The correlations in Table 1 further show descriptively that stressors are negatively associated with team flow while the experience of resources is positively associated with team flow. Once again, this applies to the total scores as well as to the various subcategories.

Testing Hypotheses
To test Hypotheses 1 and 2 we investigated the relationship between the perceived teamwork stressors and resources and team flow. The linear model which regresses team flow on experience of stressors and resources was significant, F(2, 304) = 100.03, p < 0.001. R 2 for the model was 0.40 (adjusted R 2 = 0.39), which indicates a high goodness-of-fit according to Cohen (1988). Frequency of stressors as well as frequency of resources experienced in firefighting missions significantly predicted team flow, suggesting that the more often firefighters experience teamwork stressors in firefighting missions the less often they experienced team flow. Frequency of experiencing resources of teamwork was positively associated with frequency of team flow. An overview is given in Table 2. Regression with bootstrapping confirmed the results. Hypotheses 1 and 2 were supported by these findings.
To explore which stressors and resources were related to team flow, we regressed team flow on the subcategories of the questionnaire of stressors and resources of teamwork in firefighting missions separately for teamwork stressors and resources. The decision-making subcategory of resources was not considered (see description of measures).
The stressor subcategories shared mental models, leadership, supporting behavior and organization/coordination significantly predicted less frequent team flow. The model explained R 2 = 0.30 (adjusted R 2 = 0.28) and was significant, F(6, 298) = 21.10, p < 0.001. Stressors in the subcategories communication and decision-making were not significant predictors of team flow. Table 3 gives an overview of the multiple regression analyses. The results of the regression were confirmed by the bootstrapping analyses.
The model regressing team flow on the different subcategories of resources was significant, (R 2 = 0.35 [adjusted R 2 = 0.34]; F(5, 287) = 31.29, p < 0.001). The resource subcategories communication, shared mental models, supporting behavior and organization/coordination significantly predicted more frequent team flow. Resources regarding leadership did not predict team flow. Table 4 gives an overview over the results of the multiple regression. The significant results were confirmed by the bootstrapping analyses.

Discussion
This study aimed to investigate the relationships of collective orientation, and teamwork stressors and resources in firefighting missions with team flow. Using a questionnaire, we elicited firefighters' experiences regarding teamwork stressors and resources in missions, team flow, and collective orientation. The results of our multiple regression analysis show that frequency of teamwork stressors experienced was negatively associated with team flow. Experiencing stressors in the subcategories leadership, shared mental models, supporting behavior and organization/coordination was found in particular to be negatively related to team flow. On the other hand, frequency of experiencing teamwork resources was positively associated with frequency of team flow. Specifically, we found that experiencing resources in the subcategories communication, shared mental models, supporting behavior and orga- nization/coordination was significantly positively associated with team flow. Moreover, we found a positive association between collective orientation and team flow.

Stressors and Resources of Teamwork Among Firefighters are Associated with Team Flow
We proposed that experience of teamwork-related stressors is negatively associated with team flow (Hypothesis 1), while experience of teamwork-related resources is positively associated with team flow (Hypothesis 2). Both hypotheses were supported by the findings in this study. These results are in line with those of Heyne et al. (2011), who found positive associations of flow perceived in teams with different team processes (i.e., perceived sharedness of information, process of team knowledge building). Russell (2001), in his qualitative analysis of interviews with athletes likewise found that positive team interaction is an antecedent of flow. Moreover, Aubé et al. (2018) reports that teamwork behaviors are positively associated with individual flow. These results may be transferable to the team level. The relationships found in our study moreover concur with the Transactional Model of Stress and Flow by Peifer and Tan (2021) and expand it to the team level. Teamwork initially creates demands (e.g., communication, coordination) that must be met. Well-functioning processes within the team (e.g., good communication with third parties, supporting behavior) are resources that help to evaluate the situation as a resolvable challenge and to experience team flow. However, in the presence of teamwork-related stressors (e.g., inadequate leadership), the individuals concerned may no longer perceive the situation as challenging, but as stressful. This would lead to reduced team flow. By including team-related stressors and resources in the Transactional Model of Stress and Flow, consideration of the team level succeeds. In this regard, it also supports the assumption that there are additional conditions and characteristics of team flow compared to individual flow .
Accordingly, stressors and resources regarding teamwork seem to be important factors that are related to team flow and merit more detailed discussion: Inadequate leadership was one of the stressors negatively associated with team flow. The subcategory included, for example, insufficient consideration of hazards and operational standards as well as inappropriate behavior by leaders and was related to less frequent team flow. These findings concur with those of van den Hout and Davis (2022), who in an online survey found that inadequate leadership was a factor that hinders team flow. These findings moreover support the assumption that, especially under stress, team management, evaluation of the situation, stress management of oneself and others are important leadership competencies (Flin, 1996). Our results underline that leadership is an important non-technical skill in the context of High Reliability Organizations (HRO). It is well documented that leadership is important for a safe and effective performance in HROs (Flin et al., 2008) and that lack of leadership behavior is related to poorer team performance (Marsch et al., 2004). Our study showed that the effects of leadership also apply to team flow and not only performance. This corroborates findings that flow in teams and team performance are positively related (Aubé et al., 2014;Heyne et al., 2011;van den Hout et al., 2019). Future studies should investigate the extent to which leadership, team performance, and team flow are interrelated and whether team flow could possibly serve as mediating factor between leadership and team performance.
On the other hand, good communication in a firefighting mission was one of the resources positively associated with team flow. For example, the right amount of information and good communication with third parties (e.g., police, railroad employees) were considered in this subcategory, which is related to a more frequent team flow. Again, this finding corroborates that of van den Hout and Davis (2022), who asked participants in an online survey why they experienced team flow and found that open communication was extremely relevant. Moreover, this resonates with earlier descriptions of communication as one of the basic non-technical skills (Flin et al., 2008) for effective teamwork. Furthermore, Nieva et al. (1985) reported that amount of communication is positively associated with team performance in unstructured problem-solving tasks and that intragroup communication does not result in slower problem solving. As flow in teams and team performance are related (Aubé et al., 2014;Heyne et al., 2011;van den Hout et al., 2019), we suggest that the positive effects of communication on team performance may also apply to team flow. Team flow could possibly act as a mediating factor between successful communication and team performance, which should be investigated in future studies. The finding that team flow is related to communication processes also supports the assumption that communication and feedback are team flow characteristics .
Other teamwork-related stressors and resources associated with team flow were found in the subcategories shared mental models, supporting behavior and organization/coordination. For example, awareness of a common course of action and skills (items in the subcategory shared mental models) and behaviors of team members (items in the subcategory supporting behavior) seem to be important to experience team flow. This is in line with the findings of Heyne et al. (2011), who found an association between flow in teams and perceived sharedness, a construct closely related to shared mental models. Also, shared mental models are positively associated with team performance (Espevik et al., 2006;Jo, 2012;Johnson & Lee, 2008) and team flow could mediate the relationship between the perception of shared mental models and team performance. This assumption should be investigated in future studies. Moreover, trust in each other's knowledge, skills, and attitudes is another core char-acteristic of team flow introduced by , resembling shared mental models and underlining that shared mental models could be related to team flow. Aubé and Rousseau (2005) found that supportive behavior mediates the relationship between team goal commitment and quality of group experience. This is in line with our finding that supporting behavior is positively associated with experience of team flow.
The effective interaction is described by Quinn (2005) as an aspect of collective flow. The subcategory organization/coordination in the questionnaire used in this study asks, among other things, to what extent tasks are well distributed so that no individual is overburdened, and to what extent people act on their own initiative. This reflects the requirements for good interaction and emphasizes that coordination can be an important factor in the team flow experience.
Some subcategories were not significantly associated with team flow in the multiple regressions. In the interpretation, care should be taken that the bivariate correlations between the subcategories and team flow become significant in the suspected direction (see Table 1). This may suggest that the different subcategories also overlap and accordingly some variance components were better predicted by other categories, e.g., decision-making could be dependent on leadership and shared mental models or coordination overlaps with communication.
We suggest investigating the relationships between those teamwork processes, team flow, and team performance more closely to find underlying mechanisms. For this purpose, practice missions could be used in which the team processes are evaluated by external observers. Subsequently, both the perceived team flow and the performance would have to be assessed in order to examine correlations more closely.

Collective Orientation is Associated with Team Flow
In line with Hypothesis 3, we found that collective orientation is positively associated with team flow. To the best of our knowledge, there is so far no study investigating the influence of collective orientation on team flow. However, it is well known that collective orientation leads to better performance in interdependent tasks (Driskell et al., 2010;Hagemann, 2017). Given that flow perceived in teams is also positively associated with team performance (Aubé et al., 2014;Heyne et al., 2011;van den Hout et al., 2019) our results offer an explanation why collective orientation can lead to better performance: Team flow could be a link between collective orientation (along with well-functioning team processes) and team performance. Furthermore, collective orientation is positively associated with other teamwork-related constructs like cooperativeness, social interdependence or agreeableness (Driskell et al., 2010;Hagemann, 2017). We propose that these teamwork-related constructs could be connected with some of the characteristics of team flow, like fluent, positive interactions (Pels et al., 2018) and agreement on goals, procedures, and roles (Walker, 2010). Therefore, we assume that collective orientation can enhance team processes and thus could lead to more frequent team flow.
A more advanced approach could be based on the interaction of collective orientation with the stressors and resources of teamwork. As described by Salas et al. (2008), people with a high collective orientation are more aware of their team members' needs and task inputs. Hagemann et al. (2021) demonstrated empirically that teams of people having high collective orientation were more effective in coordination behavior, resulting in better team performance. Therefore, highly collectively oriented individuals contribute through their attitude and their behavior to successful teamwork and should experience more situations in which tasks can be accomplished using teamwork as a resource. On the other hand, if collective orientation is low, the teamwork may be less effective and the supportive character of teamwork lower. Individuals may consequently experience team resources less often and interpret teamwork situations more often as a stressor. In sum, collective orientation could be positively associated with the perception of teamwork-related resources and negatively with the perception of teamwork-related stressors. These theoretical considerations are supported by the results of this study, as collective orientation is positively correlated with the experience of teamwork-related resources and negatively correlated with the experience of stressors (see Table 1). Considering that, according to this study, teamwork resources and stressors are related to the experience of team flow, a mediating effect could be hypothesized. Highly collectively oriented people could experience the resources of teamwork more often, which could lead to more frequent team flow. Conversely, people with low collective orientation could experience fewer resources and more stressors during teamwork-related activities like firefighting missions. This could have a negative effect on team flow. As our data are cross-sectional, we were only able to test these assumptions exploratorily and provide initial circumstantial evidence. It appeared that teamwork stressors and resources could mediate the effect of collective orientation on team flow. However, longitudinal data are needed to adequately examine these effects. Longitudinal data are further necessary to investigate possible other directions of causality of effects, or bi-directional effects. For example, the experience of team flow could also lead to more positive perceptions of team processes. A randomized controlled trial in which one group receives training in non-technical skills and is compared to a group not receiving such training would be highly informative regarding the effects of improved non-technical skills or enhanced collective orientation on team flow. However, acquiring such a specific sample in a longitudinal design is very challenging and was not feasible in the context of this study.

Practical Implications
The results of our study point to two approaches for increasing team flow: training in non-technical skills and collective orientation. Both approaches are described below.
Training in non-technical skills can be found in Crew Resource Management (CRM) training. Starting in aviation, CRM training addressed the knowledge, skills, and attitudes required for successful teamwork and was adapted to various contexts (Hagemann, 2011;Salas et al., 1999Salas et al., , 2006. For example, Okray and Lubnau II (2004) adapted CRM training to the firefighting context and give an overview of how non-technical skills can be acquired. Hagemann and Kluge (2013) confirmed the effectiveness of a team resource management training for firefighters regarding declarative knowledge acquisition and change in safety-relevant attitudes and behaviors. Accordingly, such existing concepts of non-technical skills training in the context of firefighting could enhance team flow by reducing the appearance of teamwork-related stressors and strengthen resources.
Since we surveyed very specific teamwork-related stressors and resources in the firefighting context, concrete recommendations for training content to foster team flow (e.g., clear and unambiguous communication strategies) can be derived from these results. The content to be addressed in the training based on the findings in our study is: (1) Leadership competencies: In German fire departments, training and continuing education for leadership positions are clearly regulated.  & Thielsch, 2019). Accordingly, it is possible that certain training content important for increasing team flow is taught according to the curriculum but is not adequately internalized by the addressees. Only through systematic evaluation is it possible to determine whether the content taught leads to the desired changes at the knowledge, attitude, and behavioral level. Therefore, FIRE could help to evaluate the quality of leadership training (Schulte & Thielsch, 2019). If necessary, the design could be modified to be more readily internalized, thereby ultimately enhancing team flow.
(2) Coping with poor leadership: While the focus of leadership training is on the leaders themselves, it is also necessary to teach subordinates how to deal with poor leadership behavior. In line with this, Smith et al. (2016) suggest that an autocratic leadership style in the fire service was often previously perceived as insufficient because subordinates were not involved. Accordingly, subordinates should be taught how to report such deficiencies to the leaders to reduce stressors caused by inadequate leadership behavior. Conversely, leaders should enable subordinates to provide feedback and propose corrections -which is important for successful leadership (Okray & Lubnau II, 2004). (3) Development of shared mental models: Potential approaches developing shared mental models include guided practice, and feedback mechanisms, e.g., debriefings (Cannon-Bowers et al., 1993;Mathieu et al., 2000). Moreover, job rotation could be another tool to support the development of shared mental models (Mathieu et al., 2000). In firefighting, practice missions focusing on special procedures and discussed in detail afterwards would be appropriate. As in our study, for example, knowledge about skills and behavior of team members proved critical for team flow, this knowledge should be strengthened. One approach could be conversations among colleagues in which special skills are reported and procedures for missions discussed on a theoretical level. (4) Supporting team members: To give team members adequate support the needs of the others must be recognized and training in supporting behavior connected to knowledge about team members' skills and behaviors. Once team members know each other, they can assess whether the other person's behavior is appropriate or whether the other person seems overwhelmed and deviates from their usual patterns of behavior. In this case they can give support. Given the close connection with the development of shared mental models, the same training methods merit consideration, namely guided practice, job rotation, and debriefing (Cannon-Bowers et al., 1993;Mathieu et al., 2000).  (Flin et al., 2008). Furthermore, with regard to coordination, fair distribution of tasks/division of labor should be ensured and no one should be overloaded. Debriefings after missions could be helpful to identify potential for improvement in these areas (Hagemann et al., 2022).
These implications apply not only to firefighting; HROs share similar processes and the consequences of teamwork failures are comparable , thus transferability from the firefighting context to other HROs seems reasonable. For example, transfer to emergency medical teams or the police seems logical as similar team processes occur. Nevertheless, these concepts need to be adapted to the specific context of application. The interest in a better team performance in HROs is immense as the consequences of failure can be dire . Different ways to enhance performance should therefore be considered. Increase in team flow that is positively associated with performance (van den Hout et al., 2019) is therefore desirable.
The second approach based on our results is training in collective orientation. Since collective orientation belongs to an approach based on knowledge, skill and attitudes, it can be cultivated via training (Driskell et al., 2010;Hagemann, 2017). Therefore, training in collective orientation could lead via better team processes to more team flow and so to better team performance. This entails finding decisive factors in collective orientation (Hagemann et al., 2021). For example, Hagemann et al. (2021) propose studying the effects of group incentives systems, leadership style, and team size.

Strengths, Limitations, and Future Directions
To the best of our knowledge our study is the first to investigate how specific teamwork-related stressors and resources influence the experience of team flow in firefighting missions. Since the factors influencing team flow have only rarely been studied, our study provides valuable information on possible contributary factorslike appropriate leadership behavior, good communication, supporting behavior, and shared mental models.
A further strength of our study is its large sample size enabling us to test our assumptions. In addition, we had a highly specific sample that included only firefighters, making it very informative for this specific HRO context.
In the survey, firefighters had to retrospectively answer questions about firefighting missions already experienced. They were asked to aggregate their experiences. With a high level of professional experience and having completed many missions, bias may have occurred during mental aggregation. Future studies should examine whether similar relationships occur when reference is made to a specific situation with no aggregation across different missions. In firefighting, practice missions could be evaluated immediately after execution in terms of the experience of team flow, teamwork-related stressors and resources and could be linked to collective orientation.
Moreover, team flow was assessed on an individual level eliciting participants' experiences of the situations as a team. No values of team members were aggregated or compared to other team members' perceptions, so team flow was defined as a shared experience (van den Hout & Davis, 2019) but assessed at the individual level. Future research should compare the experienced team flow of individual team members to better ascertain whether the entire team perceives the experience similarly or if individual perceptions vary. For example, multi-level models could be used in which the individual's data are nested at team level.
As mentioned above, the data in this study are cross-sectional, so no causal interpretation can be assumed. Accordingly, in this paper we report only associations of teamwork-relevant stressors and resources with team flow. For causal conclusions, longitudinal data or experimental settings would be necessary. Furthermore, common method biases may have occurred because the subjects rated both the stressors and resources and the team flow with self-reports (Podsakoff et al., 2003). To reduce the common method bias in this study, the anonymity of the survey was emphasized and the order of the items on stressors, resources, and team flow was partly randomized to control for reciprocal influences. Nevertheless, future research should use different sources of data to prevent the common method bias even more (Podsakoff et al., 2003). For example, practice missions could be rated by an external observer regarding different stressors and resources of teamwork (e.g., lack of information, hectic behavior of leaders) that occurred.

Conclusion
Our study investigated the associations between teamwork-related stressors and resources, team flow and collective orientation in the context of firefighting. We provide empirical evidence that team flow is associated with the experience of stressors and resources of teamwork, suggesting additional conditions and characteristics in team flow compared to individual flow. Conditions of team flow therefore need further investigation in different contexts. Collective orientation is positively associated with team flow. Again, this relationship should be addressed in future research with an experimental setting and possible mediating effects. Funding Open Access funding enabled and organized by Projekt DEAL. This research was funded by the German Social Accident Insurance (DGUV) grant number FP-433.

Data Availability
The data presented in this study are available on request from the corresponding author.

Declarations
Institutional Review Board Statement The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Ethics Committee of the Faculty of Psychology, Ruhr-University Bochum (no. 482, approval issued on: 3 August 2018).

Informed Consent Statement
Informed consent was obtained from all subjects involved in the study.

Conflict of interest The authors report no conflicts of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/.