The impact of status hierarchy on individual behavior and team processes
Doornenbal, B.M.
2021
document version
Publisher's PDF, also known as Version of record
Link to publication in VU Research Portal
citation for published version (APA)
Doornenbal, B. M. (2021). The impact of status hierarchy on individual behavior and team processes.
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal ?
Take down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
E-mail address:
CHAPTER 3
WHEN AND WHY DOES HIERARCHY AFFECT TEAM COORDINATION: THE ROLE OF HIERARCHY STRUCTURE AND DEFERENCE.
ABSTRACT
Status hierarchy is theorized to benefit teams by enhancing coordination-enabling processes. Although a handful of studies have found support for this claim, a recent meta-analysis suggested an overall opposite effect. In response to this theory-inconsistent finding, this chapter explores when and why hierarchy is more likely to produce its purported benefit. In an archival study (n = 102) and a vignette study (n = 84), we aim to uncover the positive impact by adopting a more fine-grained conceptualization and measurement of status hierarchy and by testing deference as a mediator of the relationship between status hierarchy and coordination. Deference is often argued to mediate the impact of status hierarchy on team outcomes, but is rarely tested as such. In this chapter, we find that teams are less efficient when their hierarchies are shaped as inverted pyramids ( ). We do not find support for the positive impact of status hierarchy and for the mediating role of deference. Collectively, our findings suggest that a more fine-grained conceptualization and measurement of status hierarchy helps to study the relationship between status hierarchy and coordination-enabling processes, but that status hierarchy affects teams differently than expected.
INTRODUCTION
Teams are the building blocks of modern organizations (Mathieu et al., 2014). An important aspect that affects teams is status hierarchy – which we define in this chapter as vertical differences between individuals in their esteem, prestige, and respect that these receive (Anderson & Willer, 2014; Magee & Galinsky, 2008). Status hierarchy affects teams such that individuals with more status receive more attention from their peers. The hierarchy literature is rather unanimous that the greater attention for higher-ranking members affects team processes and team performance, but contains opposing accounts (Anderson & Willer, 2014; Bunderson et al., 2016; Greer et al., 2018). Critical accounts argue that the greater attention for higher-ranking members highlights inequalities among team members that result in conflict-enabling states (Bunderson et al., 2016). In contrast, functionalist accounts argue that the centralized attention for higher-ranking members results in coordination-enabling processes (Halevy et al., 2011). Thus far, most empirical support has been found for a negative impact of status hierarchy: that hierarchy results in conflict enabling states and worse coordination-enabling processes (Greer et al., 2018).
Despite the greater support for the critical accounts (Greer et al., 2018), a handful of studies have pointed towards a positive relationship between status hierarchy and coordination-enabling processes (De Kwaadsteniet & Van Dijk, 2010; Halevy et al., 2012). Scholars argued that the impact of status hierarchy may be contextually dependent (Bunderson et al., 2016; Greer et al., 2018). In support of this argument, hierarchy was found to have a more positive impact on teams with more ambiguous tasks and less task interdependence (Greer et al., 2018).
Scholars proposed that the impact of hierarchy also depend on the conceptualization and measurement of hierarchy and the choice of the team process measure (Bunderson et al., 2016; Greer et al., 2018). Some
conceptualizations and measurements of hierarchy might yield more negative effects (Bunderson et al., 2016). Across closely related team processes, such as types of conflict or coordination, hierarchy may have a different impact (Greer et al., 2018). Hence, narrowing down on aspects or types of team processes may reveal the more precise impact of hierarchy. To further uncover the impact of hierarchy, scholars have called for research on when and why hierarchy is more likely to produce its purported benefit (Greer et al., 2018).
This chapter responds to this call by studying conditions under which status hierarchy may help coordination-enabling processes (i.e. when) and by examining a central mechanism through which status hierarchy is argued to produce its purported benefit (i.e. why). As outcome measure, we focus on coordination – that is, the alignment of actions, knowledge, and objectives of interdependent members (Rico et al., 2008). Coordination is found to help teams to perform better through enabling all members to contribute to performance while preventing process losses (LePine et al., 2008; Rico et al., 2008). Understanding the relationship between status hierarchy and coordination can thus help teams to perform better. We study the impact of status hierarchy on coordination in settings that require efficiency. By narrowing down the concept of coordination, our goal is to help create a fine-grained understanding of the impact of status hierarchy.
In examining when status hierarchy affects coordination, we build on Chapter 2. In Chapter 2, we demonstrated a shortcoming in how scholars
conventionally measure status hierarchy. In measuring status hierarchy, scholars often focus on either how widely status levels are distributed among team members (Bunderson et al., 2016), which we conceived as status hierarchy steepness, or how asymmetrically status levels are distributed among team members (Groysberg et al., 2011; Swaab et al., 2014), which we conceived as status hierarchy skewness. When scholars focus on status hierarchy steepness only, they largely neglect the proportion of team members at the top and bottom of the hierarchy. Analogously, when scholars focus on status hierarchy skewness only, they largely neglect the status difference between the top and the bottom of the hierarchy. Both these properties affect, however, the team functioning (Bunderson et al., 2016; Groysberg et al., 2011). To gain a more complete understanding of the impact of hierarchy, we proposed in Chapter 2 to measure status hierarchy as the interaction between status hierarchy steepness and status hierarchy skewness. We showed that combining these hierarchy properties helps explaining when hierarchy affects team outcomes. Relevant to coordination, we found that teams with pyramid shaped hierarchies ( , which are steeper and have a more positive skewness) involve less in information elaboration – that is, the exchange, discussion, and integration of individuals’ input (Van Knippenberg et al., 2004). Information elaboration can benefit the performance of knowledge-intensive teams conducting ambiguous tasks (Mesmer-Magnus & DeChurch, 2009; Resick et al., 2014), but impair teams conducting unambiguous tasks requiring efficient coordination (Halevy et al., 2011; Joshi & Knight, 2015). Therefore, in examining under what conditions status hierarchy affects coordination, it is crucial to study
status hierarchy as the interaction between hierarchy steepness and hierarchy skewness.
In order to unravel why status hierarchy benefits coordination, this chapter focuses on a mediator related to team information elaboration: deference – that we denote in this chapter as the tendency of individuals to yield to the opinions, needs, beliefs, and decisions of others (Joshi & Knight, 2015). Deference has been theorized as a mediator of the relationship between status hierarchy and
coordination because pairs of individuals often coordinate their efforts such that they go along more with the higher-status team member (Correll & Ridgeway, 2006; Halevy et al., 2011). Lower-status team members often go along with the input of higher-status members because they believe this helps in both using more valuable insights and preventing conflicts and competition with more competent peers (De Kwaadsteniet & Van Dijk, 2010; Joshi & Knight, 2015). The focus on higher-status team members results in an aligned effort that is theorized to improve coordination (Halevy et al., 2011; Joshi & Knight, 2015). Thus far, however, no studies examined deference as a mediator of the relationship between status hierarchy and coordination. To further unravel why status hierarchy affects coordination, this chapter examines deference as a mediator of the status hierarchy-coordination relationship.
In this chapter, we study the relationship between status hierarchy and coordination is as the interaction effect between status hierarchy steepness and status hierarchy skewness on coordination through deference (see Figure 1). We expect a more positive indirect impact of hierarchy steepness when the team has a small hierarchy top (i.e. pyramid-shaped hierarchy ), whereas we expect a more
negative indirect impact when the team has a larger hierarchy top (i.e. hierarchy shaped like an inverted pyramid ). The goal of this chapter is to contribute to the hierarchy literature by explaining when status hierarchy is more likely to produce its purported positive impact on coordination and by showing the process through which hierarchy affects coordination. This insight is valuable in reconsidering the functionalist accounts of status hierarchy (Halevy et al., 2011) and might help organizations in improving coordination, which is critical to performance. The impact of status hierarchy is tested in an archival study and a vignette study. This multi-study design is important to ensure both internal and external validity of the findings.
Figure 1: Conceptual model.
Status hierarchy in teams
Teams – which are groups of three or more individuals that exhibit interdependence, have at least one shared goal, and perform organizational tasks (Kozlowski & Ilgen, 2006) – almost always have a status hierarchy (Magee & Galinsky, 2008). A status hierarchy develops spontaneously because individuals ascribe each other different levels of status (Barton & Bunderson, 2014; Correll & Ridgeway, 2006). Individuals ascribe more status to others they perceive as more
valuable (Anderson et al., 2015). Individuals perceive others as more valuable when they have characteristics that signal competence (Barton & Bunderson, 2014; Correll & Ridgeway, 2006). In work contexts, for example, team members with higher income often receive more status because individuals consider income as an indicator for prestige (Stewart & Moore Jr, 1992).
Studies on status hierarchy focus mainly on two distinct hierarchy properties: status hierarchy steepness (Bunderson et al., 2016; Gajendran, 2009), which is referred to in this chapter as steepness, and status hierarchy concentration (Groysberg et al., 2011; Swaab et al., 2014), which is referred to as hierarchy skewness. Steepness refers to the aggregated differences (i.e. dispersion) in status and is greater when team members differ more in the status they hold (Anderson & Brown, 2010; Bunderson et al., 2016). The lowest steepness is present when each member as an equal amount of status, whereas a maximum steepness is present when half the team has the lowest possible status, and the other half has the highest possible status (Harrison & Klein, 2007). In contrast, hierarchy skewness refers to the asymmetry in the status distribution and is greater when there are fewer high-status team members (Harrison & Klein, 2007). In terms of hierarchy skewness, a minimal hierarchy implies that each member has an equal amount of status, whereas a maximum hierarchy – that has the largest impact – implies that the hierarchy is positively skewed, with one person with the a lot of status and the rest of the team with little status (Harrison & Klein, 2007). As illustrated in Figure 2, and further explained in Chapter 2, steepness and hierarchy skewness are distinct in that the hierarchy skewness of teams with a similar steepness can be positive (i.e. pyramid-shaped hierarchy ) or negative (i.e. hierarchy shaped like an inverted
pyramid ), and the steepness of teams with a similar hierarchy skewness can range from small to large.
Figure 2: Status hierarchy configurations.
Note: The numbers in the circles denote individual’s status level.
Despite the conceptual distinction between steepness and hierarchy skewness, there are clear similarities in both the theorization about steepness and hierarchy skewness and how to measure these properties (Bunderson et al., 2016). Regarding the theorization, scholars argued that both steepness and hierarchy skewness have the potential to help coordination because they suggest status-differences (Bunderson et al., 2016; Groysberg et al., 2011). Status status-differences can benefit coordination because they motivate lower-status team members to align their actions, knowledge, and objectives with others (Anderson & Willer, 2014; Halevy et al., 2011). Besides the strong theoretical similarity, there is overlap in the way scholars have measured steepness and hierarchy skewness (Bunderson et al., 2016). Conventionally used indices of steepness and hierarchy skewness
(conceived as hierarchy concentration) are strongly correlated when applied to small teams (Bunderson et al., 2016).
The theoretical and empirical similarity between steepness and hierarchy skewness has led scholars to question whether the study of both hierarchy properties is necessary (Bunderson et al., 2016). Although studying hierarchy properties that are theoretically and empirically similar may be redundant, Chapter 2 showed that steepness and hierarchy skewness are theoretically and empirically distinct when hierarchy skewness is measured as skewness instead of more conventional measures such as the Gini-coefficient and the Coefficient of Variation. The skewness correctly operationalizes the conceptualization of hierarchy skewness and moderates the impact of steepness (see Chapter 2). Steepness and hierarchy skewness also help to measure what Harrison and Klein conceived as the concept of disparity (2007). More disparity means a larger steepness and a more positive hierarchy skewness (top-right of Figure 2). Despite the criticism about the discriminant validity of steepness and hierarchy skewness (Bunderson et al., 2016), measuring both steepness and hierarchy skewness is important in studying hierarchy.
Integrating two hierarchy properties in predicting coordination
As demonstrated in Chapter 2, measuring status hierarchy as the interaction between steepness and hierarchy skewness helps to capture the hierarchy configuration (see Figure 2). More specifically, hierarchies with a larger difference between the top and the bottom of the hierarchy are steeper. Hierarchies that are more pyramid-shaped ( ) have a more positive hierarchy skewness, whereas
hierarchies that are more shaped like an inverted pyramid ( ) have a more negative hierarchy skewness.
The configuration of the hierarchy is important for upward deference (Bunderson et al., 2016). If lower-status team members defer to higher-status team members, such as theorized (Correll & Ridgeway, 2006; Halevy et al., 2011), deference depends on hierarchy steepness, because steepness implies that the hierarchy has team members that are lower in status. When a steeper hierarchy is more positively skewed (i.e. in pyramid-shaped hierarchy ), more deference can be expected because the hierarchy has more team members that are lower in status. Thus, more upward deference can be expected in hierarchy configurations such as in the top-right of Figure 2.
In Chapter 2, in which we conceptualized hierarchy as the interaction between steepness and hierarchy skewness, we found an interaction effect on information elaboration. Steeper hierarchies impaired information elaboration when the hierarchy was more positive skewed (i.e. in pyramid-shaped hierarchy
). These hierarchies may have impaired information elaboration because they motivated more individuals to defer to higher-status partners. Although this deference may hinder the performance of knowledge-intensive teams because they need to elaborate on information (Mesmer-Magnus & DeChurch, 2009; Resick et al., 2014), it can benefit teams that require smooth coordination of routine and predictable tasks (Halevy et al., 2011; Joshi & Knight, 2015). A hierarchy that hurts knowledge-intensive teams through impaired information elaboration could thus help teams focus on efficient coordination of predictable tasks.
We propose that steepness and hierarchy skewness have an interaction effect on coordination. As previously described, more deference can be expected in steeper hierarchies that are more positively skewed (i.e. in pyramid-shaped hierarchy ), because these hierarchies have a relatively larger hierarchy bottom consisting of individuals with substantially less status than the individuals at the small hierarchy top. The upward deference results in alignment (with the higher-status members) of actions, knowledge, and objectives (Halevy et al., 2011; Joshi & Knight, 2015) – that is, coordination (Rico et al., 2008). In addition, in steeper hierarchies that are more positively skewed (i.e. in pyramid-shaped hierarchy ), the small hierarchy top with clear high-status members is likely to create clarity as to which persons are high-status. This clarity is helpful for coordination, because high-status members usually take on coordinating tasks and roles (Halevy et al., 2011; Magee & Galinsky, 2008). Thus, in steeper hierarchies with a more positive skewness (i.e. pyramid-shaped hierarchy ) coordination is likely to benefit from the clarity as to who takes on coordinating tasks and roles.
In contrast to the positive influence, we propose a negative influence of steepness on coordination when the hierarchy is more negatively skewed (i.e. hierarchy shaped like an inverted pyramid ). In these hierarchies, the larger status differences are distributed such that there is a larger hierarchy top that has substantially more status than the (smaller) hierarchy bottom. We expect that teams with such hierarchies hurt coordination for two main reasons. First, these teams have fewer low-status team members, the members who typically align their actions, knowledge, and objectives with others. Second, we expect that the larger hierarchy top leads to ambiguity about with whom the low-status team members
should align their behavior. Furthermore, we expect the team members to experience ambiguity about who has or should take on coordinating roles and tasks because the team has multiple high-status members – the members who usually take on such tasks and roles (Halevy et al., 2011; Magee & Galinsky, 2008). A negative influence of hierarchy is in line with meta-analytical findings (Greer et al., 2018). In sum, we hypothesize the following relationship.
Hypothesis 1: Steepness and skewness interact such that the relationship between steepness and coordination is positive when the status hierarchy is more positively skewed (i.e. pyramid-shaped hierarchy ) and negative when the status hierarchy is more negatively skewed (i.e. hierarchy shaped like an inverted pyramid ).
METHODS STUDY 1
This chapter is composed of two studies. We examine Hypothesis 1, which predicts that hierarchy skewness moderates the relationship between steepness and team coordination, in an archival study. In selecting the context of the study, we sought to ensure that teams had a comparable task that relies heavily on coordination efficiency. We focus on the dispatch of firefighters.
Data and sample
This study is based on archival data from the Los Angeles Fire Department (LAFD). The analysis is based on 905 employees from 102 fire station teams in 2014. As an indicator of coordination efficiency, the analysis focuses on the turnout
time of firefighters. To mitigate validity issues related to the use of archival data (Barnes et al., 2015), the analysis was discussed with a Battalion Chief, who was part of the department Fire Statistical Analysis Section. Based on the discussion with the Battalion Chief, we chose to focus on the status hierarchy among the team members that are most responsible for the turnout time: apparatus operators, fire department engineers, and fire captains. Apparatus operators and fire department engineers are the drivers of respectively the aerial ladder trucks and the standard fire engines. Fire captains monitor the turnout time. The Battalion Chief confirmed that the task (i.e. the dispatch) was unambiguous and predictable. During the dispatch, the team members coordinate their physical positioning and the speed at which the task is carried out, but the task itself is formalized. For example. the captains always ride the right front seats of the aerial ladder trucks and the standard fire engines.
Measures Status hierarchy
We measured the status of individuals using their income as a proxy. Income helps to measure status in field settings where it is not always possible to measure status directly (Christie & Barling, 2010; Halevy et al., 2012). Higher income levels tend to result in more status (Stewart & Moore Jr, 1992). Income is stronger related to status when the distribution of income is known (Pfeffer & Langton, 1993; Shaw, 2014). The distribution of income at the fire stations was available on a (public) website. Team members were thus able to
compare their income easily with the income of others members of the fire station team.
. In line with conventional practices (Bunderson et al., 2016), and similar to those in Chapter 2, we operationalized steepness as the standard deviation of the status of team members.
Similar to in Chapter 2, we operationalized hierarchy skewness as the skewness of the status of team members. A more positive skewness value suggests that the hierarchy is more positively skewed (i.e. pyramid-shaped hierarchy ), whereas a more negative skewness value suggests that the hierarchy is more negatively skewed (i.e. hierarchy shaped like an inverted pyramid ).
We used turnout time as a proxy for coordination. The LAFD measured the turnout time as the average (for 2014) time interval between the activation of the station alerting devices and the moment the first fire unit was en-route. We argue that smoother coordination leads to lower turnout times. Before a fire unit can be dispatched from its station, it is necessary for the team members to wear protective equipment and to board the apparatus. Faster turnout means that team members are better at coordinating their actions. At the fire department, teams were required not to take longer than 80 seconds. No data were available on the turnout time that exceeded 300 seconds. These data points were deleted by the department Fire Statistical Analysis Section because they were due to a failure in the deactivation of the time measurement system.
Based on recommendations of Bernerth and Aguinis (2015) to control for reliable, measurable factors related to the focal variable, we controlled in the analysis for four variables: mean income, team size, amount of
incidents, and fire captains ratio. First, we included mean income as a control variable because higher average team income suggests that the team has used more financial resources to attract better performing team members (Bloom, 1999). Furthermore, as income increases with age (and we did not have data on age distribution), controlling for mean income helps to control for age. Age may have affected coordination in opposing ways. Age could harm coordination, because younger members could have a better physical condition, but age could also benefit coordination, because it brings experience that leads to efficiency (Rico et al., 2008). Second, we controlled for team size because of several reasons. A larger team implies a larger possible dispersion (Lindell & Brandt, 2000), and thus larger possible hierarchy. A larger team implies a reduced bias in the measurement of hierarchy properties because there are more data points (Biemann & Kearney, 2010). A larger team has more members who need to coordinate their turnout, which means more potential for process losses. Third, we controlled for the number of incidents that teams had to turn out for because experience can benefit coordination (Rico et al., 2008). Finally, we controlled for the proportion of fire captains because these individuals had more power in teams. Although we expect that controlling for the proportion of captains reduce the
effects we find1 because being a captain might result in more status (correlation
between power and our status measure is .37, p < .01), we consider controlling for power hierarchy to be necessary because status and power independently affect teams (Hays & Bendersky, 2015). Without controlling for power hierarchy, we
1 The analysis yielded stronger results when we did not control for power
hierarchy.
might overestimate the effects of status hierarchy, because the effects could be partly due to the influence of power hierarchy.
RESULTS STUDY 1
The means, standard deviations, and the correlations among the variables of interest are presented in Table 1. The average turnout time ranged from 61 to 87 seconds. Thus, 11 fire stations exceeded the maximum turnout time of 80 seconds.
To test Hypothesis 1, we conducted a hierarchical multiple regression analysis. In the first step of this analysis, coordination was regressed to the control variables. In the second step, we added the independent variables steepness and skewness to the model. Finally, the interaction effect between steepness and hierarchy skewness was included. Although scholars often mean-centered variables in creating cross-product terms (Dalal & Zickar, 2012), it would harm in the interpretability of hierarchy skewness in our case. As we argued in Chapter 2, by mean-centering skewness, a positive skewness sign (+) would no longer accurately indicate that the hierarchy is pyramid-shaped ( ) and a negative skewness sign (-) would no longer accurately indicate that the hierarchy is shaped like an inverted pyramid ( ). Therefore, we refrained from mean-centering. Note that this decision does not affect the direction or significance of the interaction in any way or form (Dalal & Zickar, 2012).
Table 1: Means, standard deviations, and correlations among the variables of Study 1 Variables Mean SD 1 2 3 4 5 6 1. Team size 8.87 4.20 2. Mean Income a 17.75 1.59 .16 3. Incidents 618.57 383.20 .29** .02
4. Fire Captain Ratio .51 .14 -.45** .08 .02
5. Steepness 0.23 0.07 .03 .34** -.02 -.10
6. Skewness 0.25 0.80 .17 .20* .08 .06 .34**
7. Coordination b 74.20 5.32 .08 -.11 .63** .05 -.16 .08
Note. N = 102 teams. a Yearly average in 10,000 USD per team member b
Coordination is reported as reverse coded seconds. p < .10. * p < .05. ** p < .01.
The results of the regression analyses are presented in Table 2. As reported in this table, the results suggest a significant positive interaction effect between steepness and hierarchy skewness on coordination (B = 0.81, SE = 0.36, p < .05). This interaction effect helps to predict coordination above and beyond the direct
effects of steepness and hierarchy skewness ( R2 = .03 (F (1, 95) = 5.31, p < .05).
To further explore the interaction effect, we plotted the interaction for teams with a positively skewed hierarchy (i.e. +1 SD skewness; pyramid-shaped hierarchy ) and teams with a negatively skewed hierarchy (i.e. -1 SD skewness; hierarchy shaped like an inverted pyramid ). As illustrated in Figure 3, steepness has a negative influence on coordination when the hierarchy skewness is negative (i.e.
hierarchy shaped like an inverted pyramid ), whereas the influence seems unclear when the hierarchy skewness is positive (i.e. pyramid-shaped hierarchy ). To compute the value interval for which the interaction effect between steepness and hierarchy skewness on coordination was significant (p < .05), we conducted a regions of significance analysis (Preacher et al., 2006). As plotted in Figure 3, this analysis showed that the influence of steepness on coordination ranges from non-significant to negative, depending on the hierarchy skewness. The relationship between steepness and coordination was only significant when the skewness was lower (i.e. more negatively skewed) than 0.24. For the lowest observed hierarchy skewness (-1.78), the relationship between steepness and coordination was strongest (B = -2.33, SE = 0.92, p < .05). Hence, we did not find support for a positive association between steepness and coordination when the hierarchy was positively skewed (i.e. pyramid-shaped hierarchy ) and reject Hypothesis 1. Consistent with meta-analytical findings (Greer et al., 2018), however, we found support for a negative relationship between hierarchy and coordination. More specifically, we found a negative relationship between steepness and coordination when the hierarchy is negatively skewed (i.e. hierarchy shaped like an inverted pyramid ). Coordination suffered from steepness when the proportion of high-status team members was high.
Beyond the hypothesized interaction effect, Table 2 shows that the control variable incidents had a significant positive influence on coordination (B = 0.01, SE = 0.00, p < .001). Steepness had a significant negative influence on coordination (B = -0.89, SE = 0.41, p < .05), whereas hierarchy skewness had a marginal influence on coordination (B = -2.61, SE = 1.55, p < .10). In pointing out these findings, we
focus on the full model. Although scholars often interpret lower-order terms before entering the interaction terms in the model, the predictors tend to vary depending on the moderating variables (Aguinis et al., 2016). Considering the lower-effects based on the full model, thus including the interaction effects, is therefore more meaningful (Aguinis et al., 2016).
Table 2: Results of regression analysis Study 1.
Dependent Variable: Coordination
Step 1 Step 2 Step 3
Variables
Team Size -0.10 0.12 -0.14 0.12 -0.12 .14
Mean Income a -0.38 0.27 -0.26 0.28 -0.17 0.27
Incidents 0.01*** 0.00 0.01*** 0.00 0.01*** 0.00
Fire Captains Ratio 0.61 3.27 -0.33 3.29 -0.04 2.95
Steepness -0.63 0.36 -.89* 0.41
Skewness 0.74 0.55 -2.61 1.55
Steepness x Skewness 0.81* 0.36
Adjusted - R2 .42 .44 .47
F 17.51*** 12.54*** 11.10***
Note. N = 102 teams. Unstandardized regression coefficients (i.e. B) were used.
Performance is reported as reverse coded seconds. a Yearly average in 10,000 USD
per team member.
p <.10. * p < .05. ** p < .01. *** p < .001.
Figure 3: Interaction and regions of significance plots Study 1, interaction effect of steepness and skewness on team coordination.
Note. Steepness and Skewness have a significant interaction effect on coordination (95% Confidence Interval) for the region on the left side of the confidence band. This region denotes more negatively skewed hierarchy (i.e. hierarchy shaped like an inverted pyramid ). Coordination is reported in reverse seconds (turnout time) and mean-centered. -3 -2 -1 0 1 2
low steepness high steepness
Te am C o or d in at io n negative skewness no skewness positive skewness
DISCUSSION STUDY 1
Although previous studies on the relationship between status hierarchy and coordination have captured hierarchy as either steepness (Bunderson et al., 2016; Gajendran, 2009) or hierarchy skewness (Groysberg et al., 2011; Swaab et al., 2014), Study 1 suggests that the interaction between these hierarchy properties predicts coordination above and beyond steepness and hierarchy skewness on their own. In line with our prediction, steepness had a more negative relationship with coordination when the hierarchy was more negatively skewed (i.e. hierarchy shaped like an inverted pyramid ). However, in contrast to our hypothesis, we did not find a more positive relationship between steepness and coordination when the hierarchy was more positively skewed (i.e. pyramid-shaped hierarchy ). These findings suggest that focusing on the interaction between steepness and hierarchy skewness can lead to a more complete understanding of the impact of status hierarchy. It does not unravel, however, when status hierarchy has its purported benefit.
Although archival research offers a “unique avenue for tying psychological phenomena to important real-world outcomes” (Barnes et al., 2018, p. 145), our study warrant validity considerations. In our study, we used income as a proxy for status. Although income is found to be positively related to status (Stewart & Moore Jr, 1992) and previously studied as a status proxy (e.g., Christie & Barling, 2010; Halevy et al., 2012), it does not reflect all the characteristics that determine the status of a person (Christie & Barling, 2010). Thus, we did not capture all variance in status by operationalizing status as income. Another measurement validity issue is that we used turnout time as a proxy for coordination. Coordination
is an important antecedent of the turnout, but other aspects, such as physical strength and obstacles in the operating environment, are also essential for turning out. It is therefore likely that the variance in turnout time is partly due to
phenomena unrelated to coordination. These other phenomena could have weakened the effects of hierarchy. Thus, with a more valid measure of coordination, we might have found stronger effects of hierarchy.
In addition to the measurement validity issues, the findings warrant internal and external validity issues. Regarding internal validity, the correlational nature of the study does not provide a rigorous test of this causal claim that status hierarchy affects coordination. In fact, status hierarchy might be shaped by turnout time. For example, when teams fail to meet their coordination goals, moderate-status team members may be promoted in the hierarchy, resulting in a more negatively skewness (hierarchy shaped like an inverted pyramid ). Regarding external validity, the fact that we did not find support for a positive effect of hierarchy might be due to that the teams executed a highly formalized routine task. Possibly, the formalized roles were more important for the coordination than status hierarchy. If no formalized roles were present, the status hierarchy could have had more functional effects. To be better able to generalize our findings, it is therefore important to test our hypothesis in a context in which teams do not have formal roles.
STUDY 2
In Study 2, we examine the interaction effect between steepness and hierarchy skewness on coordination in an experimental vignette design. This
design provides a more rigorous test of our causal claims (Aguinis & Bradley, 2014). Furthermore, it allows to take into account the previously described validity issues. In Study 2, status is more directly measured, the teams do not have a power hierarchy or formal roles, and coordination is measured using a validated measure for implicit coordination. Implicit coordination – that is, coordination in which individuals adapt to needs or demands without verbal communication (Rico et al., 2008) – underlies more general coordination tendencies (De Kwaadsteniet & Van Dijk, 2010; Rico et al., 2008) and is a logical results of deference – the degree to which individuals go along with someone else’s input (Joshi & Knight, 2015).
Beyond accounting for methodological issues, this study responds to a call for more research on the mechanisms through which status hierarchy affects the team functioning (Greer et al., 2018). By responding to this call, we help to reveal to what extent the theory of the relationship between status hierarchy and coordination-enabling processes should be revised. Across previous studies, including Study 1 of this chapter, status hierarchy did not have the theorized positive impact on coordination-enabling processes (Greer et al., 2018). Possibly, studies did not find support for a positive impact because status hierarchy does not produce the theorized mechanism(s) through which status hierarchy is argued to benefit teams.
Scholars have theorized that status hierarchy benefits coordination through deference (Anderson & Willer, 2014; Halevy et al., 2011). Status hierarchy is argued to affect deference because hierarchy implies having lower-status and higher-status individuals. When teams have higher-status individuals, they have members that are perceived to be more valuable (Barton & Bunderson, 2014;
Correll & Ridgeway, 2006). By deferring to these members, individuals intend to capitalize on more valuable input (Barton & Bunderson, 2014; Correll & Ridgeway, 2006). Individuals also defer to higher-status members to meet behavioral expectations (Cooper & Withey, 2009; Neeley & Dumas, 2015). By deferring to higher-status members, lower-status individuals believe that they convey a message of acceptance of their current position in the hierarchy that will help them prevent conflict and competition (Halevy et al., 2011; Joshi & Knight, 2015).
Deference results in the alignment of actions, knowledge, and objectives (Halevy et al., 2011; Joshi & Knight, 2015) – which is conceived as coordination (Rico et al., 2008). A few studies have found some support for deference as a mediator of the relationship between status hierarchy and (team) outcomes. De Kwaadsteniet and Van Dijk (2010), showed that status hierarchy benefits coordination in dyads because the lower status individuals adjust their behaviors to the higher-status individual. If dyads need to coordinate where to meet, they prefer to meet at a location close to the higher-status member. Halevy and colleagues (2012) built on this finding while studying coordination in basketball teams. Although they did not examine deference explicitly, they showed a positive association between income hierarchy and team outcomes closely related to deference and coordination.
We propose that the interaction effect between steepness and hierarchy skewness affects deference. More specifically, we expect more deference in teams with a larger hierarchy bottom that has substantially less status than the small hierarchy top (i.e. pyramid-shaped hierarchy ). First, because these teams have more low-status team members who tend to defer to high-status members.
Second, we expect more deference in these teams because it is clearer to the low-status team members who the high-low-status team members are. We expect less deference in teams have a larger hierarchy top that has substantially more status than the hierarchy bottom (i.e. hierarchy shaped like an inverted pyramid ). First, because these teams have more high-status team members that do not tend to defer to others, but expect others to defer to them. Second, we expect less deference in these teams because individuals who want to defer to a higher-level partner are more likely to experience ambiguity about who to defer because there are multiple high-level members. Thus, we hypothesize the following relationship.
Hypothesis 2: The interaction between steepness and skewness on coordination is mediated by deference, such that the indirect relationship between steepness and coordination is more positive when the status hierarchy is more positively skewed (i.e. pyramid-shaped hierarchy ) and negative when the status hierarchy is more negatively skewed (i.e. hierarchy shaped like an inverted pyramid ).
METHODS STUDY 2 Participants and design
The participants of this study were 252 full-time employees (34.5% women; age M = 34 years, SD = 10 years) in the United States. We recruited these participants via the online platform Prolific Academic and received each £ 0.30 (approximately $ 0.40) for taking part in the study. Although scholars have questioned the validity of findings from studies conducted using online samples,
meta-analytical evidence suggests that they yield similar results as more traditional studies (Walter et al., 2016). The quality of data gathered through Prolific
Academic is similar to the data collected through Amazon Mechanical Turk (Peer et al., 2017). An advantage of Prolific Academic is that participants are more naïve, which reduces the impact of learning effects and dishonesty (Peer et al., 2017). In recruiting participants, we focused on employees and excluded individuals younger than 18 years and older than 65 years from participation because this enhances generalizability to the work context (Aguinis & Bradley, 2014). Before participants began their participation, we randomly assigned them to a team role and to one of the four conditions of the two (low steepness vs. high steepness) by two (negative skewness vs. positive skewness) design. Next, we assigned the participants that completed the study per team role and condition to one of the 84 three-person teams based on the latitude of their computer location. We assigned participants with a computer location more south to a lower team number. Assigning participants based on the latitude of their location was largely
independent of the variables of interest (deference: r = -.02, p = .76; the choice for meeting at someone else’s location: r = .00, p = .99) and is a systematic and verifiable approach.
Procedure
We invited participants to a team study on decision-making, similar to Study 1 of De Kwaadsteniet and Van Dijk (2010), adjusted to coordination in teams. Each participant was presented with a scenario in which they had a coordination problem within their team. In this scenario, the teams planned a meeting but did
not decide the location of the meeting. Without communicating (e.g. leaving a note or phone call), participants had to decide where to meet their team. We offered three alternatives: the office of team member A, the office of team member B, or the office of member C. Participants were part of a team in which we
randomly assigned them as team member A, B, or C. In addition, we instructed the participants that the three offices were at the same distance from each other and that they were in the exact middle of those locations, at equal distance from each possible meeting location. Participants did not meet each other, and we instructed them that they did not know where their fellow team members were. To help participants understand the scenario, we illustrated it with a graphical depiction of the three offices (see Figure 4). Before we asked the participants where they were planning to go, we manipulated the team's status hierarchy and then measured their deference trends. By ordering the procedure as such, we are better able to test the causal influence of hierarchy on deference and coordination. Similar to De Kwaadsteniet and Van Dijk (2010), we measured (implicit) coordination as the extent to which team members agreed upon a meeting location.
Figure 4: Graphical depiction of offices A, B, and C.
Manipulations
We manipulated the team's status hierarchy by informing participants about the status of team members, including themselves. Similar to a manipulation used by Blader and Chen (2012), we manipulated individuals' status by informing them about their status in the group. The instruction was: "Team member A (/B/C) has attained very little (/somewhat little /somewhat a lot/a lot of) status within the group. Group members ascribe this person very little (/somewhat little/an average amount of/somewhat a lot/a lot of) respect, esteem, and prestige.” In every hierarchy condition, the team members had a status combination that represented a low (1.15) or high (3.46) steepness and a negative (-1.73) or a positive (1.73) hierarchy skewness (see Table 3).
Note in Table 4 that there is a high correlation between average status and hierarchy skewness (r = -.89, p < .01). Average status was not constant across the experimental conditions, because in 3-person teams that can only have a unimodal distribution (i.e. with a single peak), the skewness of the team distribution is related to the average of that distribution (Von Hippel, 2005). An overrepresentation of low-status members is the reason for a lower average status in teams with a positive hierarchy skewness (i.e. pyramid-shaped hierarchy ). A lower average status in hierarchies with a more positive hierarchy skewness is, however, consistent with the concept of disparity (Harrison & Klein, 2007). More specifically, conceptualized as disparity, pyramid-shaped hierarchy is argued to have a stronger impact because it has more members that are disadvantaged in their status (Harrison & Klein, 2007).
Table 3: Overview status hierarchy manipulations Negative skewness (-1.73), pyramid-shaped hierarchy ( ) Positive skewness (1.73), inverted pyramid-shaped hierarchy ( ) High Steepness (3.46)
Member A: very little status (1). Member B: a lot of status (7). Member C: a lot of status (7).
(Average status: 5.00)
Member A: very little status (1). Member B: very little status (1). Member C: a lot of status (7).
(Average status: 3.00) Low
Steepness (1.15)
Member A: somewhat little status (3).
Member B: somewhat a lot of status (5).
Member C: somewhat a lot of status (5).
(Average status: 4.33)
Member A: somewhat little status (3). Member B: somewhat little
status (3).
Member C: somewhat a lot of status (5).
(Average status: 3.67) Note. Status levels, presented in between brackets, were not reported to participants.
Manipulation check
During the status manipulation check, we asked participants to indicate how much status each team member has on a scale ranging from 1 (=very little) to 7 (=very much). We then calculated for each respondent the steepness and skewness of the status hierarchy they perceived. As intended, the perceived steepness was lower in the low steepness conditions (M = 1.95, SD = 0.48) than in
the high steepness conditions (M = 3.17, SD = 0.36). To test these differences, we conducted a one-way ANOVA. This analysis suggested that the differences between the conditions were significant, F(1, 250) = 166.66, p < .001. We conducted a two-way ANOVA to test the degree to which these differences were moderated by the skewness condition. This test revealed that the skewness condition did not significantly affect the influence of the steepness condition on the perceived steepness F(2, 248) = 0.32, n.s.
As intended, the perceived hierarchy skewness was negative in the negative skewness condition (M = -1.50, SD = 0.62; hierarchy shaped like an inverted pyramid ) and positive in the positive skewness condition (M = 1.45, SD = 0.74; pyramid-shaped hierarchy ). To explore these differences further, we conducted a one-way ANOVA. This analysis suggested that differences between the conditions were significant, F(1, 250) = 1185.71, p < .001. We conducted a two-way ANOVA to test the degree to which these differences were moderated by the steepness condition. This test revealed that the steepness condition did not significantly affect the influence of the skewness condition on the perceived steepness hierarchy skewness F(2, 248) = 3.25, n.s.
Measures
We assessed deference using a four-item scale adapted from Joshi and Knight (2015). This scale comprised four statements: "I yield to someone else", “I go along with someone else’s preference”, “I respect other’s needs”, and “I defer to someone else’s choice” We instructed participants that they needed to respond to these statements regarding their decision about where to go to meet
fellow team members. Our assumption was that they estimated that all team members preferred to meet at their own office, rather than going to someone else’s office. We did not instruct the participants with this preference because it might have steered their response to the four statements too much. We asked the participants to respond to the statements (Cronbach alpha = .86) on a 7-point scale (1 = "strongly disagree", 7 = "strongly agree") and computed team deference by mean-aggregating individuals' deference scores. Usually, a low within-team variance justifies mean-aggregation because it implies that the aggregated scores are a manifestation of the full team (Bliese, 2000; James et al., 1984). We
considered mean-aggregation to be appropriate regardless of the within-team variance, because higher-status individuals feel less pressured to engage in deference in teams with a greater status hierarchy, even if all other members defer to others. Note that our deference measure is not limited to deference to a higher-status peer. If we had asked for deference to a higher-higher-status peer, we would have introduced bias, because less upwards deference is possible in the negative skewness condition (i.e. hierarchy shaped like an inverted pyramid ). In this condition, fewer participants have a higher-status peer.
In line with De Kwaadsteniet and Van Dijk (2010), we measured coordination as the extent to which individuals agree on a meeting location. We measured the inter-rated agreement by computing Fleiss’ Kappa, a variant of Cohen’s Kappa that is appropriate for more than two raters that give categorical ratings (Fleiss, 1971).
RESULTS STUDY 2
The means, standard deviations, and the correlations among the variables of interest are presented in Table 4. As reported in this table, hierarchy skewness had a significant association with coordination (r = .37, p < .01). Average status, which is – as previously argued – negatively related with hierarchy skewness in our design, also had a significant association with coordination (r = -.23, p < .05). The association between deference and coordination was non-significant (r = .04, n.s.).
Table 4: Means, standard deviations, and correlations among the variables in Study 2 Variables Mean SD 1 2 3 4 5 6 1. Gender a .35 .30 2. Age 33.90 5.00 .03 3. Average Status 4.00 0.75 .10 .12 4. Steepness 0.50 0.50 .04 -.14 .00 5. Skewness 0.00 1.01 -.09 -.06 -.89** .00 6. Deference 4.37 .72 .09 .20 .12 -.11 -.08 7. Coordination .46 .34 -.06 -.06 -.23* -.01 .37** .04
Note. N = 84 teams. a 0 = man, 1 = woman
p <.10. * p < .05. ** p < .01.
A multiple regression analysis was conducted to test the hypotheses. In this analysis, we regressed both deference and coordination on steepness, hierarchy skewness, and the interaction effect between steepness and hierarchy skewness
(see Table 5). To test Hypothesis 1, we first regressed coordination on steepness, hierarchy skewness, and its interaction. This regression suggested that steepness and hierarchy skewness have a significant interaction effect on coordination (B = -0.15, SE = 0.07, p < .05). Moreover, this interaction effect helped to predict coordination above and beyond the direct effects of steepness and hierarchy skewness ( R2 = .04 (F (1,82) = 4.89, p < .05).
To further examine the interaction effect, we plotted the simple slopes for the two skewness conditions (see Figure 5). A simple slope analysis indicated that the simple slope was not significant for both the negative skewness condition (B = 0.14, SE = 0.09, n.s.; hierarchy shaped like an inverted pyramid ) and the positive skewness condition (B = -0.16, SE = 0.10, n.s.; pyramid-shaped hierarchy ). We did thus not find support for Hypothesis 1, which predicted that steepness and hierarchy skewness interact such that the relationship between steepness and coordination is more positive when the status hierarchy is more positively skewed (pyramid-shaped hierarchy ) and more negative when the status hierarchy has a more negative skew (hierarchy shaped like an inverted pyramid ).
Although we did not find support for Hypothesis 1, an interesting finding of our regression analysis (see Table 5) is the significant positive influence of hierarchy skewness on coordination (B = 0.20, SE = 0.05, p < 0.01). This finding implies that, across the two steepness conditions, teams coordinated better when their hierarchy was positively skewed (pyramid-shaped hierarchy ), with one member at the top the hierarchy. Across the two skewness conditions, in contrast, steepness did not significantly influence coordination (B = -0.01, SE = 0.07, n.s.).
W H EN A N D W H Y D O ES S TA TU S H IE RA RC H Y A FF EC T TE A M C O O RD IN A TI O Ta b le 5 : R es ul ts o f r eg re ss io n an al ys is M ed ia to r: D ef er en ce D ep en d en t V ar ia b le : C o o rd in at io n St ep 1 St ep 2 St ep 1 St ep 2 St ep 3 V ar ia b le s St ee p ne ss -0 .1 5 0. 16 -0 .1 5 0. 16 -0 .0 1 0. 07 -0 .0 1 0. 07 -0 .0 1 0. 07 Sk ew ne ss -0 .0 6 0. 08 0. 03 0. 11 0. 12 ** 0. 04 0. 20 ** 0. 05 0. 20 ** 0. 05 St ee p ne ss x Sk ew ne ss -0 .1 7 0. 16 -0 .1 5* 0. 07 -0 .1 5* 0. 07 D ef er en ce 0. 02 0. 05 A d ju st ed - R 2 .0 0 .0 0 .1 1 .1 5 .1 4 F 0. 74 0. 88 4. 95 ** 5. 65 ** 4. 49 ** N ot e. N = 8 4 te am s. * p < .0 5. * * p < .0 1
Figure 5: Interaction effect of steepness and skewness on team coordination
To further unravel these effects, we conducted a post-hoc analysis comparing coordination differences between hierarchy conditions. First, we compared coordination in the low steepness condition. In this condition, teams coordinated significantly (Bonferroni-corrected p < .05) better in the positive skewness condition (pyramid-shaped hierarchy ; M = .27, SD = .13) than in the negative skewness condition (hierarchy shaped like an inverted pyramid ; M = .67, SD = .37). In the condition with the smaller status differences, teams thus coordinated better when the hierarchy top is small. In the high steepness conditions, however, there were no significant (Bonferroni-corrected p < .05) differences in coordination between teams in the positive skewness condition (pyramid-shaped hierarchy ; M = .51, SD = .33) and in the negative skewness condition (hierarchy shaped like an inverted pyramid ; M = .41, SD = .36). In the condition with the larger status differences, teams did thus not coordinate significantly better with more or less members at the top of the hierarchy. These
0.00 0.25 0.50 0.75 1.00
low steepness high steepness
Te am C o or d ia nt io n negative skewness positive skewness
findings thus seem to suggest that hierarchy skewness is less of importance for coordination when the average status differences are large.
To test Hypothesis 2, in which we predict that deference mediates the interaction effect between steepness and skewness on coordination, we computed the index of moderated mediation following recommendations of Hayes’ (2015). Testing moderated mediation effects through this index is appropriate because it reflects the association between the indirect effect and the moderator (Hayes, 2015). A significant index suggest that “any two conditional indirect effects estimated at different values of the moderator are significantly different from one another” (Hayes, 2015, p. 2). To compute the index, we first examined the indirect effect of steepness on coordination. We secondly assessed the difference in the indirect effect for the different skewness conditions by conducting a 95% bootstrap estimation. This analysis indicated an index of -.01 (SE = .02), with a confidence interval including zero [-.08, .02], suggesting no support for deference as a mediator of the interaction between steepness and skewness on coordination (H2).
To further explore the role of deference, we regressed deference on the independent variables (see Table 5). We found that the independent variables do not significantly influence deference (F(3, 80) = 0.88, n.s.). We did thus not find significant support for a role of deference in the influence of (the interaction between) steepness and hierarchy skewness.
DISCUSSION STUDY 2
Study 2 shows that the interaction effect between steepness and hierarchy skewness predicts coordination above and beyond steepness and skewness
independently. However, the pattern of interaction found in this study does not support our hypotheses and, as discussed in the overall discussion, differs from the effects found in Study 1. More specifically, we found that steepness did not have a significant influence on coordination. Skewness influenced coordination such that more positive hierarchy skewness (pyramid-shaped hierarchy ) helped teams to coordinate. The results of Study 2 do not support that deference mediates the interaction effect between steepness and hierarchy skewness on coordination. This finding thus does not suggest that deference is a significant mechanism through which status hierarchy affects coordination.
Although we did not find significant support for our hypotheses, it is important to consider that the perceived steepness in our low steepness condition was high (M = 1.95, SD = 0.48) in comparison with the average steepness measured in similar hierarchy studies (Bunderson et al., 2015: M = 0.23, SD = 0.11; Chapter 2: M = 0.48, SD = 0.32). At the same time, we found an overall influence of skewness across the two conditions, rather than just in a large steepness condition. We may not have found a significant interaction effect because the participants also perceived steepness as high in the low-steepness conditions. It is also important to note that deference was part of the coordination measure. More specifically, coordination was better when more individuals deferred to the same team member. In the paper of De Kwaadsteniet and Van Dijk (2010, p. 516), the authors described choosing someone else’s location as deference. If we had measured deference as the degree to which individuals chose someone else's location to meet, we might have found a stronger relationship between hierarchy and deference.
In addition to these issues, our study warrants a number of other
considerations. First, as explained in the methods section, hierarchy skewness was related to average status. Status is related to team commitment in such a way that low-status individuals may experience less urge to repay the respect they receive by putting more effort into their work than is least needed (Anderson et al., 2015; Sleebos et al., 2006). In the high steepness positive skewness condition, the low-status team members were lower in low-status than in the low steepness positive skewness condition. Possibly, we did not find better coordination in the high steepness positive skewness condition, because the lower-status team members felt unmotivated to repay the respect they received by going to someone else's office proactively to meet for an appointment. Second, similar to Blader and Chen (2012), we manipulated individuals' status by informing participants about their and others’ status in the group. We did not use concrete characteristics to substantiate individuals’ status because it might have introduced ambiguity about the drivers of the effects. In Study 1, for example, we cannot exclude that the effects are more due to income than to status. By not describing characteristics other than status in Study 2, however, participants may feel less urge to defer to higher-status peers.
GENERAL DISCUSSION
This chapter aims to increase the understanding of the relationship between status hierarchy and coordination-enabling processes by studying conditions under which status hierarchy may help coordination (i.e. when) and by examining a central mechanism through which status hierarchy is argued to produce its
purported benefit (i.e. why). A better understanding is needed because examinations of the relationship between status hierarchy and coordination-enabling processes do not support the current theoretical understanding of this relationship (Greer et al., 2018). Whereas scholars have theorized that status hierarchy helps coordination-enabling processes (Anderson & Willer, 2014; Halevy et al., 2011), meta-analytical findings suggest that status hierarchy harms
coordination-enabling processes (Greer et al., 2018). By studying the impact of status hierarchy as the interaction effect between steepness and hierarchy skewness, we aim to uncover when status hierarchy affects coordination. By studying deference as a mediator of the relationship between status hierarchy and coordination, we aim to uncover why status hierarchy benefits coordination. Knowing when and why status hierarchy helps coordination benefits organizations because status hierarchy is ubiquitous to teams (Magee & Galinsky, 2008) and coordination has a positive effect on team performance (LePine et al., 2008; Rico et al., 2008).
In this chapter, we did not find support for our hypotheses that (H1) steepness and hierarchy skewness interact such that the relationship between steepness and coordination is positive when the status hierarchy is more positively skewed (i.e. pyramid-shaped hierarchy ) and negative when the status hierarchy is more negatively skewed (i.e. hierarchy shaped like an inverted pyramid ), and that (H2) deference mediates this interaction effect. However, the study results show that status differences can harm coordination when the status differences are distributed such that teams have a larger proportion of high-status members. In Study 1, steepness had a negative influence on coordination when the hierarchy
was more negatively skewed (i.e. hierarchy shaped like an inverted pyramid ). In Study 2, teams with a negatively skewed hierarchy (i.e. hierarchy shaped like an inverted pyramid ) were coordinating worse across the two status steepness conditions. In the larger steepness condition alone, however, we did not find a significant effect of hierarchy skewness. Study 2, furthermore, did not provide support for the mediating role of deference in the relationship between status hierarchy and coordination.
Theoretical implications
Scholars have called for more research into when hierarchy benefits coordination because of theory-inconsistent findings (Greer et al., 2018). We aimed to unravel when status hierarchy benefits coordination following our suggestion in Chapter 2 to study status hierarchy as the interaction effect between steepness and hierarchy skewness. By studying status hierarchy as the steepness-skewness interaction, scholars can precisely measure the degree to which hierarchy is pyramid-shaped ( ) or shaped like an inverted pyramid ( ). As demonstrated in Chapter 2, pyramid-shaped hierarchy ( ) can reduce information elaboration. Information elaboration – that is, the process of exchanging,
discussing, and integrating input – helps knowledge-intensive teams conducting ambiguous tasks (Mesmer-Magnus & DeChurch, 2009; Resick et al., 2014), but can hinder teams conducting unambiguous tasks requiring efficient coordination (Halevy et al., 2011; Joshi & Knight, 2015). As argued in this chapter, we expected that pyramid-shaped hierarchy ( ) helps coordination by centralizing attention to the small hierarchy top. Although our two studies suggest that the interaction
effect between steepness and hierarchy skewness affects coordination above and beyond these hierarchy properties independently, they do not explain when status hierarchy benefits coordination. Our findings thus imply that either scholars should revise the theory about the relationship between status hierarchy and
coordination, or that a different approach is needed to explain when status hierarchy benefits coordination.
Our findings yield important implications for the role of task ambiguity in the impact of status hierarchy. Meta-analytical evidence suggests that hierarchy has a less negative impact on teams with a greater task ambiguity (Greer et al., 2018). In Chapter 2, where we studied teams with a high task ambiguity, we found a negative influence of steepness on team performance when the hierarchy was more positively skewed (i.e. pyramid-shaped hierarchy ). In this present chapter, where we studied teams with a low task ambiguity, we did not find a negative influence of steepness on coordination when the hierarchy was more positively skewed (i.e. pyramid-shaped hierarchy ). Thus, if greater hierarchy means a more pyramid-shaped hierarchy, our findings are opposed to the meta-analytical findings. However, as we found in this chapter that hierarchy shaped like inverted pyramid ( ) hurts coordination, the role of task ambiguity may depend on how hierarchy is conceptualized and measured. Possibly, pyramid-shaped hierarchy ( ) hurts teams with a high level of task ambiguity, because it hinders information elaboration, whereas hierarchy shaped like an inverted pyramid ( ) hurt teams with a low level of task ambiguity, because it hinders efficient coordination.
Besides contributing to the understanding of when status hierarchy affects coordination, this chapter contributes to the accumulation of knowledge of why
status hierarchy affects coordination. It does this by examining deference as a mediator of the relationship between status hierarchy and coordination. Scholars have argued that status hierarchy benefits coordination through deference – that is, the tendency of individuals to yield to the opinions, needs, beliefs, and decisions of others (De Kwaadsteniet & Van Dijk, 2010; Halevy et al., 2011; Joshi & Knight, 2015). In a previous study (De Kwaadsteniet & Van Dijk, 2010), scholars found support for the influence of status hierarchy on deference in dyads. However, no studies have examined deference as a mediator of the status hierarchy-coordination relationship. In this chapter, we tested deference as a mediator of the status hierarchy-coordination relationship, but did not support for a mediating role. The lack of support may be due to shortcomings of the deference measure, as discussed in the discussion of Study 2. However, status hierarchy might not always result in deference. If status hierarchy indeed does not always result in deference, this would explain the lack of support for the positive relationship between hierarchy and coordination (Greer et al., 2018) because deference is seen as the main reason why hierarchy benefits coordination (Anderson et al., 2006; Bunderson et al., 2016).
Practical implications
This chapter demonstrates that status hierarchy can affect coordination in teams that have unambiguous tasks and are less able to communicate. We found that the impact of status hierarchy on coordination depends on the shape of the hierarchy. Team coordination suffered from hierarchies shaped like inverted pyramids ( ). Having many team members at the top the hierarchy can thus be
detrimental to teams. As demonstrated in Study 1, income data can be used to compute hierarchy scores. Computing hierarchy scores based on such data is non-intrusive compared to approaches in which hierarchy scores are computed based on survey data – such as in Study 2.
Ensuring that teams have a pyramid-shaped hierarchy ( ), with a clear hierarchy top, can prevent poor coordination. The status of a team member can be enhanced by stressing the social instrumental value of this member (Anderson et al., 2015; Blader & Chen, 2012). By stressing the value of one team member, for example by honoring an “employee of the year”, the status hierarchy could become more pyramid-shaped. In line with previous research (Rico et al., 2008), however, our findings suggest that coordination benefits from having experienced teams. To increase the coordination (in teams with hierarchies shaped like inverted pyramids; ), it is therefore important to practice the task.
Limitations and suggestions for future research
In addition to the more technical limitations discussed in Study 1 and Study 2, this chapter contains a number of theoretical limitations. First, we studied coordination in settings that require efficiency, but the effectiveness of many knowledge-intensive organizations relies on coordination that lead to the
integration of knowledge and insights (Mesmer-Magnus & DeChurch, 2009; Resick et al., 2014). Consequently, the hierarchy that was detrimental to teams in this chapter could not be detrimental to teams operating in knowledge-intensive organizations. The findings of this chapter may therefore not be directly applicable to other types of coordination. In addition, team performance can depend on the
interaction effect between coordination efficiency and knowledge integration coordination (Reagans et al., 2016). As status hierarchy might have a different impact on types of coordination, we suggest future research to study the status hierarchy – team performance relationship while studying coordination efficiency and knowledge integration coordination as interacting mediators. Such research would help to further uncover when and why hierarchy contributes most to the performance teams.
A second limitation is that this chapter focused solely on the impact of status hierarchy on coordination (through deference). Besides affecting coordination, hierarchy is argued to affect team conflict (Anderson & Willer, 2014; Greer et al., 2018). The impact of status hierarchy on conflict can hinder coordination-enabling processes (Bunderson et al., 2016). Therefore, we might have been better able to explain the relationship between status hierarchy and coordination by including conflict in the conceptual model. We chose not to focus on conflict because it is theorized to hurt the team functioning, whereas our aim was to uncover the benefits of status hierarchy. To increase the understanding of the relationship between status hierarchy and coordination, we suggest that future research should consider conflict while studying the status hierarchy-coordination relationship. We suggest focusing on conflict as an outcome of status hierarchy that weakens the positive relationship between status hierarchy and coordination.
Finally, the previously mentioned shortcomings in our status manipulation and deference measure might be the reason why we did not find support for deference as a mediator of the relationship between status hierarchy and coordination. Future research is necessary to examine the extent to which status
hierarchy results in deference. It is possible that status hierarchy does not have the theorized positive impact on coordination because scholars have overestimated the role of deference. Thus, we encourage scholars to further uncover the relationship between status hierarchy and deference.
CONCLUSION
Status hierarchy is theorized to benefit coordination-enabling processes, but meta-analytical evidence suggests an overall opposite effect (Greer et al., 2018). In response to this theory-inconsistent finding, this chapter examines when and why status hierarchy produces its purported benefit. Across two studies, we find that teams coordinate worse when their hierarchy is shaped as an inverted pyramid ( ). We do not find consistent support for the positive impact of status hierarchy. The examination of why status hierarchy benefits coordination did not yield support for a mediating role of deference, a process of yielding to others that is often argued to mediate the impact of status hierarchy on team processes (Halevy et al., 2011). Collectively, this chapter demonstrates that a more fine-grained conceptualization and measurement of status hierarchy helps to study the relationship between status hierarchy and coordination, and that status hierarchy affects teams differently than expected.
CHAPTER 4
DIFFERENT STROKES FOR DIFFERENT FOLKS:
THE ROLE OF GOAL ORIENTATION IN UPWARD STATUS-BASED DEFERENCE The hierarchy literature suggests that individuals tend to go along with the input of higher-status partners. This tendency, which we denote as upward status-based deference, can be seen as the individuals who reach for the apples in the painting below. The apples can be perceived as the ideas of others. In Chapter 4, we find that upward status-based deference can depend on personal characteristics. Some individuals do not reach for the (rotten) apples.
Munch, E. (1910). The Tree of Life [Tempera (?) and crayon on unprimed canvas]. Munch Museum, Oslo, Norway. Munchmuseet (11-23-2004). Retrieved from:
https://foto.munchmuseet.no/fotoweb/archives/5014-Malerier/Arkiv/M0899_20041123.tif.info. Photograph: Sidsel de Jong. CC BY 4.0 Munchmuseet.
