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 6
CHAPTER 6 DISCUSSION
This dissertation aims to contribute to the understanding of the impact of status hierarchy on individual behavior and team processes. The chapters of this dissertation are structured around three research questions: (1) “what are the consequences of the choice of the conceptualization and associated measurement in studies on the impact of hierarchy?”, (2) “how does the influence of status hierarchy vary across functional group processes?”, and (3) “when does upward status-based deference occur?” As motivated in the introduction chapter, two types of validity issues are central across my studies: measurement validity matters, which are issues about the degree to which the measurements measure what they intend to measure, and construct validity matters, which are issues about the degree to which the meaning of constructs aligns in their nomological networks of related constructs (Borsboom et al., 2004). I call these issues matters because they are issues that matter to the findings of my studies and therefore to my answers to the research questions. My work reveals a number of insights, which I gained by applying a wide variety of methods and by studying hierarchies with different bases (Table 1). This methodological variety helped me to approach the research questions from a broad perspective and helped to uncover several insights into the validity matters.
In this final dissertation chapter, I discuss the answers I found to the research questions and reflect on the implications and contributions of my work. I will first elaborate on the key findings related to each research question. Drawing on these key findings, I will subsequently reflect on a more abstract level on the validity
DISCUSSION
matters. During this reflection, I discuss the theoretical implications of my answers and point out promising avenues for future research. I conclude this chapter by outlining the practical implications for dealing with team hierarchies.
C H A PT ER 6 19 0 Ta b le 1 . O ve rv ie w C ha p te rs 2 to 5 . C h St u d y P ri m ar y fo cu s M et h o d H ie ra rc h y b as is H ie ra rc h y co n ce p t an d m ea su re m en t D ep en d en t va ri ab le s R es ul ts 2 1 RQ 1 & RQ 2 Su rv ey C on fe rr ed st at us D is p ar ity m ea su re d a s ST *S K In fo rm at io n el ab or at io n - L es s in fo rm at io n el ab or at io n fo r p yr am id -sh ap ed h ie ra rc hi es ( ), w hi ch a re b ot to m he av y. Pe rf o rm an ce L o w er p er fo rm an ce fo r p yr am id -s ha p ed hi er ar ch ie s ( ), w hi ch a re b o tt o m h ea vy . 3 2 RQ 2 A rc hi va l Sa la ry D is p ar ity m ea su re d a s ST *S K Ef fic ie nc y - L es s ef fic ie nt c oo rd in at io n fo r hi er ar ch ie s sh ap ed a s in ve rt ed p yr am id s ( ). 3 RQ 2 & RQ 3 V ig ne tt e In fo rm ed st at us D ef er en ce - N o s ta tis tic al ly s ig ni fic an t r el at io ns hi p b et w ee ST *S K a nd d ef er en ce . Im p lic it co or d in at io n - N o s ta tis tic al ly s ig ni fic an t s up p or t f or hy p ot he si s. H ow ev er , m or e im p lic it co or d in at io n fo r p yr am id -s ha p ed h ie ra rc hi es ( ). 4 4 RQ 3 Ex p er im en t Ed uc at io n an d co m p et iti on Pa rt ne r st at us m ea su re d a s co nf er re d st at us . U p w ar d st at us -b as ed d ef er en ce - M or e up w ar d s ta tu s-b as ed d ef er en ce b y th os th at h av e a PA V G O (v s. P A PG O a nd L G O ) a nd lo w er s el f-ef fic ac y. 5 5 RQ 1 Su rv ey C on fe rr ed in flu en ce D is p ar ity m ea su re d a s ST *S K , a nd A cy cl ic ity m ea su re d a s K H M . ST *S K a nd K H M - W ea k re la tio ns hi p b et w ee n ST *S K a nd K H M 6 RQ 1 Si m ul at io n U p w ar d d ef er en ce ST *S K a nd K H M - W ea k re la tio ns hi p b et w ee n ST *S K a nd K H M N ot e. P A VG O = P er fo rm an ce A vo id an ce G o al O ri en ta tio n. S T* SK = S te ep ne ss -S ke w ne ss in te ra ct io n, K H M = K ra ck ha rd t’s H ie ra rc hy M ea su re .
SUMMARY OF THE MAIN FINDINGS
In this section of the discussion, I describe research findings that provide answers to the three research questions. I summarize these key findings, categorized by research question, in Table 2. Regarding RQ1 (see 2nd column in
Table 2), my findings imply that the choice of the conceptualization and
measurement affects study outcomes. In particular, my findings suggest that (a) the hierarchy conceptualizations disparity and acyclicity, and their associated
measures, represent different concepts, and (b) that combining hierarchy measures in studying hierarchy can help to explain team processes. Regarding Q2 (see 3rd column in Table 2), the findings suggest that the influence of status
hierarchy on learning-related processes and efficiency-related processes – which are team dynamics that I call functional group processes – can be very different. In my studies, pyramid-shaped hierarchies ( ) hampered a learning-related process (information elaboration), but hierarchies shaped like inverted pyramids ( ) hampered efficiency-related processes (coordination efficiency and implicit coordination). Finally, regarding RQ3 (see 4th column in Table 2, the findings
suggest that upward status-based deference depends on the personal
characteristics of the individuals within the hierarchy. I found that low self-efficacy individuals that have a performance-avoidance goal orientation (PAVGO), in contrast to a performance-approach goal orientation (PAPGO) and learning goal orientation (LGO), defer more to higher-status peers. In the vignette study I conducted, I did not find a statistically significant influence of status hierarchy on upward status-based deference. Taken together, my findings contribute to the
CHAPTER 6
understanding of the impact of status hierarchy on individual behavior and team
processes, as I will describe in more detail next. DIS
C U SS IO N 19 3 Ta b le 2 . S um m ar y o f r es ea rc h fin d in g s in re la tio n to th e re se ar ch q ue st io ns . R es ea rc h Q u es ti o n 1. W ha t a re th e co ns eq ue nc es o f t he ch oi ce o f t he c on ce p tu al iz at io n an d as so ci at ed m ea su re m en t i n st ud ie s on th e im p ac t o f h ie ra rc hy ? 2. H ow d oe s th e in flu en ce o f s ta tu s hi er ar ch y va ry a cr os s fu nc tio na l g ro up p ro ce ss es ? 3. W he n d oe s up w ar d s ta tu s-b as ed d ef er en ce o cc ur ? C ha p te r 2 M ea su rin g e ith er h ie ra rc hy st ee p ne ss o r hi er ar ch y sk ew ne ss le ad s to in co ns is te nt s tu d y re su lts b ec au se th es e p ro p er tie s ha ve a n in te ra ct io n ef fe ct o n te am p ro ce ss es . Py ra m id -s ha p ed h ie ra rc hi es ( ) ha m p er in fo rm at io n el ab or at io n an d , su b se q ue nt ly , h ur t t ea m p er fo rm an ce . C ha p te r 3 M ea su rin g th e hi er ar ch y sk ew ne ss al on g si d e th e hi er ar ch y st ee p ne ss is cr uc ia l i n ex p la in in g c oo rd in at io n ef fic ie nc y an d im p lic it co or d in at io n. In ve rs e p yr am id al s ha p ed hi er ar ch ie s ( ) h am p er c oo rd in at io n ef fic ie nc y an d im p lic it co or d in at io n. N o s ta tis tic al ly s ig ni fic an t s up p o rt fo r up w ar d s ta tu s-b as ed d ef er en ce in m y vi g ne tt e st ud y. C ha p te r 4 C on ce p tu al iz in g s ta tu s th ro ug h th e co m b in at io n o f h ig he r ed uc at io n an d g re at er c om p et iti ve ne ss a ffe ct s co nf er re d s ta tu s, b ut n ot th e d ef er en ce to o th er s. Th e re la tio ns hi p b et w ee n p ar tn er st at us a nd d ef er en ce to th e (lo w -q ua lit y) in fo rm at io n of th is p ar tn er d ep en d s o n se lf-ef fic ac y an d g oa l o rie nt at io n. Lo w s el f-ef fic ac y in d iv id ua ls w ith a PA V G O h av e a st ro ng er te nd en cy to d ef er m or e to h ig he r-st at us p ee rs co m p ar ed to lo w s el f-ef fic ac y in d iv id ua ls w ith a P A PG O a nd a LG O . C ha p te r 5 M ea su rin g d is p ar ity a s K H M a nd ac yc lic ity a s ST *S K c an b e p ro b le m at ic b ec au se d is p ar ity a nd ac yc lic ity a re c o nc ep tu al ly d iff er en t. M or eo ve r, S T* SK a nd K H M a re em p ir ic al ly d iff er en t. B ec au se th e hi er ar ch y co nc ep ts (a nd as so ci at ed m ea su re s) d is p ar ity a nd ac yc lic ity a re d is tin ct , t he y w ill h av e an in d ep en d en t i m p ac t o n fu nc tio na l g ro up p ro ce ss es . N ot e. P A VG O = P er fo rm an ce A vo id an ce G o al O rie nt at io n, P A PG O = P er fo rm an ce A p p ro ac h G o al O ri en ta tio n, L G O = L ea rn in g G o al O rie nt at io n, S T* SK = S te ep ne ss -S ke w ne ss in te ra ct io n, K H M = K ra ck ha rd t’s H ie ra rc hy M ea su re .
Research Question 1: What are the consequences of the choice of the conceptualization and associated measurement in studies on the impact of hierarchy?
The first research question addresses the consequences of the choice of the conceptualization and measurement in studies on the impact of hierarchy. Scholars often conceive status hierarchy as disparity, where hierarchy is greater when it is more pyramid-shaped ( , Harrison & Klein, 2007). Studies on the impact of hierarchy have focused on different properties of a hierarchy (Bunderson et al., 2016; Greer et al., 2018). Within the management literature, the status hierarchy properties that have received most attention are status hierarchy steepness and hierarchy concentration: respectively the average status-difference between high-status and low-high-status team members (Anderson & Brown, 2010) and the proportion of high-status members within a team (Groysberg et al., 2011). Management scholars have also conceived hierarchy as acyclicity (Bunderson et al., 2016), where hierarchy is greater when influence cascades down the dyadic relations within the hierarchy, from individuals at the top of the hierarchy to individuals at the bottom of the hierarchy. Because the hierarchy properties are believed to have a different impact on team processes, I argued that scholars capture an incomplete picture of status hierarchy's impact when they study the hierarchy properties separately.
Building on conventional measurements of hierarchy, I proposed to measure hierarchy as the interaction between status hierarchy steepness and status hierarchy skewness (ST*SK). I explained that, compared to frequently used measurements, this measurement is more encompassing in operationalizing
hierarchy conceptualized as disparity. Moreover, I demonstrated across three studies (Chapter 2 and Chapter 3) that modeling the ST*SK can help to explain team processes above and beyond the direct influence of status hierarchy steepness and status hierarchy skewness.
Although I have proposed to measure hierarchy through the ST*SK, I explained that this measurement is a suitable operationalization for the concept of disparity and not for operationalizing the concept of acyclicity (Chapter 5). For operationalizing the concept of acyclicity, Bunderson and colleagues (2016) suggested to apply Krackhardt’s Hierarchy Measure (KHM, Krackhardt, 1994). I demonstrated, in both a simulation study and a survey study (Chapter 5), that their two associated measures ST*SK and KHM are empirically distinct. Additionally, I explained that ST*SK and KHM should be empirically distinct because they are measures of different concepts. Through this explanation and demonstration, I showed that it would be appropriate to focus on both disparity and acyclicity (and their associated measures) within one study. To support others in the choice of the conceptualization and measurement of hierarchy, Chapter 5 provides guidelines.
In proposing the ST*SK as a measurement of disparity, I introduced hierarchy skewness as a measurement of hierarchy concentration – a vital property of disparity. In Chapter 2, I compared hierarchy skewness with conventional hierarchy concentration measures, such as the Gini-coefficient and the Coefficient of variation (Dawson, 2011; Wei et al., 2016), and demonstrated that hierarchy skewness is a more appropriate measure for the concept of hierarchy
concentration. The introduction of hierarchy skewness has two main advantages. First, using skewness helps to overcome a discriminant validity issue between
CHAPTER 6
conventional hierarchy steepness measures and conventional hierarchy
concentration measures (Bunderson et al., 2016). That is, conventional measures of hierarchy steepness and hierarchy concentration are argued to be statistically too related for two conceptually distinct properties (Bunderson et al., 2016). I demonstrated that steepness and skewness have a low statistical relation and explained that the skewness of a hierarchy with either a small or a large steepness can vary from very negative to very positive. Second, using skewness helps to overcome a content validity issue of conventional concentration measures. That is, conceptually, disparity is greater when valued resources such as status are concentrated in a few individuals (Harrison & Klein, 2007), but conventionally chosen hierarchy (concentration) measures are not monotonous in relation to the size of the hierarchy top. A higher value of a conventional hierarchy measure does not always reflect a smaller hierarchy top. Hierarchy skewness denotes the degree to which the valued resources are concentrated in a few individuals. Therefore, hierarchy skewness can help to accurately measure disparity.
Measuring hierarchy skewness is also recently proposed in a paper in which scholars study the impact of formal hierarchy (Wellman et al., 2020). In this paper, hierarchy skewness is found to help explaining team functioning. As I will discuss while answering research question 2, they found that formal hierarchy shaped as inverted pyramids ( ) enhances perceived patient care relatively to pyramid-shaped hierarchy ( ) when task variety is high. This dissertation demonstrates that measuring the concept of disparity as the interaction effect between steepness and skewness helps to understand the impact of status hierarchy.
DISCUSSION
Research Question 2: How does the influence of status hierarchy vary across functional group processes?
The second research question addresses the functionalist accounts of status hierarchy, which conceive hierarchy as an organizational structure that benefits functional processes, such as coordination and collective decision making (Anderson & Willer, 2014). On average, studies provided little empirical support for the functionalist accounts (Greer et al., 2018). I argued that this lack of support is partly due to differences in impacts between types of functional group processes. Drawing on the assumption that having a hierarchy encourages individuals to defer to others higher in rank, I argued for both a positive and negative impact of status hierarchy. I argued that status hierarchy can hurt learning-related processes, because the upward status-based deference results in disproportionate attention for higher-status team members and a neglect for crucial input from the lower-status members (Bunderson & Boumgarden, 2010). In contrast, I argued that status hierarchy can contribute to efficiency-related processes, because the upward status-based deference provides guidance for smooth cooperation (Halevy et al., 2011, 2012).
Across my studies, I did not find clear overall support for the functionalist accounts of status hierarchy (Anderson & Willer, 2014). Instead, I found contrasting outcomes. The study focused on the relationship between status hierarchy and a learning-related process (Chapter 2) suggested a negative influence of pyramid-shaped hierarchies ( ). In contrast, while studying the relationship between status hierarchy and efficiency-related processes (Chapter 3), I found a negative influence of status hierarchy when hierarchy was shaped as inverted pyramid ( ). The results
thus suggest that distinguishing types of functional group processes can help to understand the influence of status hierarchy on functional group processes.
Besides distinguishing types of functional group processes, considering task variety as a moderator might help in understanding the influence of status hierarchy on functional group processes. A main difference between the two types of functional group processes I studied was the task variety they required – that is, the amount of different actions that were needed to fulfill the task responsibilities (Humphrey et al., 2007). The learning-related process I studied (Chapter 2) – i.e., information elaboration: the exchange, discussion, and integration of knowledge and perspectives (Van Knippenberg et al., 2004) – required variety of ideas and perspectives to be exchanged, discussed, and integrated. In contrast, the efficiency-related process I studied (Chapter 3) – i.e., coordination in which individuals have to adapt to needs or demands without verbal communication (Rico et al., 2008) – required standardized, repetitive actions. Conceiving the difference between the functional group processes I studied as a difference in task variety, my findings suggest that pyramid-shaped hierarchies ( ) have a more negative impact on team processes that require a wider variety of tasks, whereas hierarchies shaped as inverted pyramids ( ) have a more negative impact on team processes that require a smaller variety of tasks. As previously mentioned, Wellman and colleagues (2020) found task variety to moderate the relationship between power hierarchy and patient care. Similar to my findings, they found that pyramid-shaped hierarchies ( ) had a more dysfunctional impact when the task variety was larger. Unlike my findings, they did not (hypothesize and) find a more
dysfunctional impact of hierarchies shaped as inverted pyramids ( ) when the task variety was smaller.
Collectively, my findings suggest that status hierarchy can hinder both learning-related processes and efficiency-related processes. In my studies, pyramid-shaped hierarchy ( ) hampered information elaboration, but hierarchy shaped like inverted pyramids ( ) hindered coordination efficiency.
Research Question 3: When does upward status-based deference occur? The third research question focuses on upward status-based deference – that is, the tendency of individuals to give in more to the opinions, beliefs, and decisions of higher-status others. In this dissertation, I conceived upward status-based deference as an efficiency-related process that provides guidance for smooth cooperation in which greater attention is paid to those that rank higher in the hierarchy, but also affects learning-related processes by creating more attention for input from the higher-status team members and a neglect for input from the lower-status members (see Chapters 1 and 2). The hierarchy literature describes upward status-based deference as a key mechanism between hierarchy and (functional) group processes and outcomes (Anderson & Willer, 2014; Bunderson et al., 2016; Halevy et al., 2011). Even though upward status-based deference is an assumed key mechanism through which hierarchy has its purported impact, only little empirical evidence supports that status hierarchy results in upward deference (e.g., De Kwaadsteniet & Van Dijk, 2010; e.g., Joshi & Knight, 2015). Moreover, some scholars suggest that individuals sometimes rather defer to similar others than to higher-status others (Joshi & Knight, 2015). Given
CHAPTER 6
the importance of upward status-based deference in the theory about the impact of status hierarchy, I examined the conditions under which upward status-based deference occurs. Possibly, the reason for the little support for the functionalist accounts is that status hierarchy does not always encourage people to defer to higher-status team members.
I studied upward status-based deference in a vignette study (Chapter 3) and in an experiment (Chapter 4). In the vignette study, I did not find support for a relationship between status hierarchy and upward deference. In the experiment, I found that low self-efficacy individuals that have a performance avoidance goal orientation (PAVGO), in contrast to individuals with either a performance approach goal orientation (PAPGO) or a learning goal orientation (LGO), defer more to the (low-quality) information of higher-status peers. Collectively, these findings suggest that status hierarchy does not always lead to upward status-based deference and that personal characteristics affect the extent to which individuals defer more to higher-status peers.
The studies I conducted on upward status-based deference contribute to a key assumption about the impact of status hierarchy. The hierarchy literature suggests that status hierarchy affects team processes through upward status-based deference (Anderson & Willer, 2014; Halevy et al., 2011). My work suggest that upward status-based deference depends on personal characteristics. These findings do not only support earlier arguments put forward by other scholars (Driskell & Salas, 1991; Joshi & Knight, 2015), but also suggest that personal characteristics may affect the relationship between status hierarchy and team processes by moderating upward-status based deference. For example, the
DISCUSSION
impact of status hierarchy on team processes might be weaker in teams composed of people who are less inclined to defer to high-status peers. To the best of my knowledge, moderators of the relationship between status hierarchy and upward status-based deference have not been tested.
The focus in Chapter 4 on deference to low-quality information has important implications. Previous studies mainly perceived upwards status-based deference as a mechanism that helps teams to utilize more valuable input (Bunderson, 2003; Joshi & Knight, 2015) and to facilitate useful processes such as team coordination (Anderson & Willer, 2014; Halevy et al., 2011). Nuancing this perception, I argued that upwards status-based deference can hurt performance if high-status partners provide low-quality input. This nuanced understanding is important because several studies have showed that higher-status individuals do not always provide higher-quality input (Barton & Bunderson, 2014; van Dijk & van Engen, 2013). Thus, if upward status-based deference occurs, it could hurt the team performance.
THEORETICAL IMPLICATIONS, VALIDITY ISSUES, AND FUTURE RESEARCH A central theme in finding answers to the research questions is validity. Two validity types are especially relevant to my studies, measurement validity, which concerns the degree to which the measurements measure what they aim to measure, and construct validity, which concerns the meaning of constructs in a nomological networks of related constructs. In the introduction chapter (Chapter 1), I presented an overview of how validity is essential in studying – and
apply this overview while reflecting on the study findings. I have structured this section along the three research questions. For each of these research questions, I will first describe the overall theoretical implications of my findings. Second, I will reflect on how validity issues might have affected my (interpretation of the) study findings and discuss how these issues relate to the broader hierarchy literature. Drawing on my reflection of the validity matters, I will provide suggestions I think will help in further uncovering the impact of status hierarchy.
Research Question 1: What are the consequences of the choice of the conceptualization and associated measurement in studies on the impact of hierarchy?
The findings that (a) the hierarchy conceptualizations disparity and acyclicity, and their associated measures, are different and (b) that combining hierarchy concepts in studying hierarchy can help to explain team processes, highlight the importance of being precise in conceptualizing and measuring phenomena of interest. Acknowledgement of the importance of conceptualizations and measurements is not new. Several recent studies call for more attention for measurement practices (Flake & Fried, 2019; Fried & Flake, 2018), even within the hierarchy literature (Bunderson et al., 2016; Wellman et al., 2020). Moreover, the existence of psychometrics, the field of study focused on the theory and technique of psychological measurement, suggests that conceptualization and measurement is crucial within the (management) science (Borsboom et al., 2004). What is new is that I show, through several studies, that conceptualizing and measuring (about) the shape of hierarchy is important in studying the impact of hierarchy. Hierarchy is often
conceptualized as disparity (Greer et al., 2018). Conceived as disparity (thus not as acyclicity), hierarchy is greater when its shape resembles a pyramid ( ) more (Harrison & Klein, 2007). I demonstrated that hierarchy can have a stronger impact when it is not pyramid-shaped ( ) and proposed a way to conceptualize and measure the shape of hierarchy. This advancement is relevant because it implies that the impact of (status) hierarchy can be better understood by broadening the view on hierarchy beyond hierarchy as disparity.
While demonstrating that the impact of hierarchy can better be understood by broadening the conceptual understanding of hierarchy, I introduced a construct validity matter related to disparity. The theoretical understanding of disparity is that a hierarchy is greater when it is more pyramid-shaped ( , Harrison & Klein, 2007). Although my work was mostly based on studies that conceptualized hierarchy as disparity, I introduced a mismatch between the theoretical understanding of disparity and my studies by arguing that hierarchy shaped as inverted pyramids ( ) can affect team processes more than pyramid-shaped hierarchy ( ). Introducing this mismatch was a dilemma. On the one hand, I deliberately chose a concept of hierarchy that did not fully match with the concept of disparity. On the other hand, the mismatch seemed inevitable because there are several reasons (Chapter 3) why the impact of hierarchy is greater when it is more shaped like inverted pyramids ( ). During my research, also other scholars proposed to conceive hierarchies beyond disparity. Bunderson and colleagues (2016) argued for studying acyclicity, in which hierarchy is not greater with it is pyramid-shaped ( ), but when the influence cascades more from the top down to the bottom of a hierarchy. By the end of my research, Wellman and colleagues (2020) introduced a hierarchy taxonomy in which
CHAPTER 6
they argued that the shape of a maximum hierarchy can also be shaped like an inverted pyramid ( ). Wellman and colleagues introduced this taxonomy focused at power hierarchies. Across my studies, I explained and demonstrated that studying the degree to which hierarchies are shaped as inverted pyramids ( ) can help to understand the impact of status hierarchy better. My studies thus suggest that at least parts of the taxonomy of Wellman and colleagues seems applicable to status hierarchy as well. Ultimately, the construct validity matter I introduced resulted in insights that suggest to “move beyond comparing the effects of the presence or steepness of pyramid-shaped hierarchies and towards exploring fundamentally different structures” (Wellman et al., 2020, p. 1018).
In my exploration of the impact of hierarchy, I noted an important measurement validity matter related to the measurement of hierarchy. While measuring hierarchy, I studied hierarchy steepness in combination with hierarchy skewness repeatedly across studies. A consequence of repeatedly applying a similar measurement is greater comparability of studies. However, the comparability of my studies warrants additional consideration because I applied my measurement of hierarchies to different bases. Across the hierarchies I studied (Table 1), one striking difference was the hierarchy mutability – that is, the extent to which the basis of a hierarchy (e.g. conferred status or salary) is open to be contested by its team members (Hays & Bendersky, 2015). The hierarchies studied in Chapter 2 and in Study 2 of Chapter 3 were very mutable, as their bases were respectively conferred and informed status, but the hierarchies studied in Study 1 of Chapter 3 were not, as their basis was salary. As argued by Hays and Bendersky (2015), and further supported by the meta-analysis of Greer and colleagues (2018), more mutable
DISCUSSION
hierarchies impede team processes because they encourage individuals to contest their place within the hierarchy. Thus, across similar concepts and measures of hierarchy, the impact of hierarchy can vary depending on the mutability of the hierarchy. If hierarchy indeed impedes team process when it is more mutable, it would suggest that the hierarchies in Chapter 2 and in Study 2 of Chapter 3 have a more negative effect – as these hierarchies were very mutable. Even though I did not test the difference in effects among the studies, the only negative influence of hierarchy I found in Chapter 3 was in Study 1, in which the hierarchies were less mutable. By pointing this out, I want to illustrate that the complexity of choosing the conceptualization and measurement of hierarchy goes beyond the aspects that I put central in my dissertation (disparity, acyclicity and their associated measures). In applying concepts and measurements of hierarchy, the basis of the hierarchy (e.g. conferred status or salary) is important as well.
For future hierarchy research, Chapter 5 helps to be more precise about the choice of the conceptualization and measurement of hierarchy. Being more precise will help to gain more detailed knowledge of the impact of hierarchy. Wellman and colleagues (2020) proposed six different formal hierarchy structures and associated indicators that can help to measure hierarchy. I am convinced that this taxonomy is useful in advancing the hierarchy measures. As I argued and demonstrated in my dissertation, it can help to consider the combination of indicators (e.g., the standard deviation and skewness as indicators of respectively hierarchy steepness and hierarchy skewness) in order to measure hierarchy. While studying hierarchy, I recommend that scholars consider differences in the bases of hierarchy. As
previously argued, I expect that there will be differences in the impact depending on the mutability of the hierarchy.
Research Question 2: How does the influence of status hierarchy vary across functional group processes?
Although my studies do not suggest that status hierarchy is functional, they demonstrate that hierarchy affects functional group processes. After finding that pyramid-shaped hierarchies ( ) can reduce information elaboration (Chapter 2), I noted that information elaboration may be too time consuming in teams that rely on efficiently completing routines (Resick et al., 2014). Hence, I hypothesized that pyramid-shaped hierarchies ( ) can enable teams to be efficient. In Chapter 3, I uncovered that teams were less efficient when their hierarchies were shaped as inverted pyramids ( ). Collectively, these findings imply that teams involve more in discussions that can be inefficient when their team hierarchy is shaped like an inverted pyramid ( ), as opposed to when the team hierarchy is pyramid-shaped ( ). This implication is consistent with the functionalist accounts arguing that pyramid-shaped hierarchies ( ) can reduce conflicts (Anderson & Willer, 2014; Halevy et al., 2011). While the functionalist accounts propose conflict reduction through explicit hierarchy as a benefit to teams, inhibited information elaboration is detrimental to the performance of knowledge-intensive teams.
In studying the impact of status hierarchy on functional group processes, I conceptualized the group processes in detail. The precise conceptualization is important for the accumulation of knowledge on the relationship between hierarchy and functional group processes. In Greer and colleagues’ meta-analysis (2018), the
relationship between hierarchy and functional group processes was synthesized on the basis of 12 studies (including the study reported in Chapter 2). In order for such syntheses to test differences in the impact of hierarchy between different types of functional group processes, precise conceptualizations are necessary – particularly when the number of studies is still low. I expect future syntheses to point out the previously motivated different impacts on learning-related processes and efficiency-related processes. Precise conceptualization is crucial to test such nuanced views.
Besides this construct validity matter, I dealt with several measurement validity matters that have important implications for the hierarchy literature. One important issue concerns the variety of studies I conducted. In Chapter 3, I tested the impact of hierarchy in both an archival study and a vignette study. In archival data, variables are often proxies of the constructs of interest (Hageman, 2008). In my archival study, turnout time seemed to be an apt proxy for coordination efficiency. However, income can be a problematic proxy for status. As explained in Chapter 3, income relates positively to status (Stewart & Moore Jr, 1992), but does not encompass its complete variance (Christie & Barling, 2010). People with similar incomes can vary widely in status. Therefore, using income as proxy of status will result in less accurate findings. Although I had more control over the measurements in the vignette study, I had to deal with different challenges. As explained in Chapter 3, it was not possible to change levels of steepness and skewness while keeping the average level of status unchanged. As a result, the average status of the participants differed among the conditions I compared. The difference in status might have affected the group processes given that lower status team member experience less team commitment (Anderson et al., 2015; Sleebos et al., 2006). Individuals may have felt less
CHAPTER 6
committed in teams with a lower average level of status. In both Chapter 2 and the first study of Chapter 3, where I controlled for average status, I found statistically significant associations between average status and implicit coordination – that is, coordination in which individuals adapt to needs or demands without verbal communication (Rico et al., 2008).
In testing the influence of status hierarchy on functional group processes, the survey study in Chapter 2 (which showed that pyramid-shaped hierarchies ( ) can hinder information elaboration) seemed to suffer the least from measurement validity issues as the measurements were most consistent with the studied concepts. Nevertheless, I do encourage scholars to not limit themselves to survey research. Archival data have proven to be useful in organizational research (Barnes et al., 2018). An advantage of archival research that is particularly relevant to the hierarchy literature, and perhaps to the team literature in general, is the higher statistical power that is often possible (Barnes et al., 2018). In many hierarchy studies, around 74% of the studies in the meta-analysis of Greer and colleagues (2018), samples were smaller than 100. Testing based on larger sample sizes, especially when expected effects are not very large (and when more complex statistical models are applied), increases the likelihood that the test will be able to reject the null hypothesis when it is false (i.e. the statistical power). In other words, archival data can help prevent scholars from finding false-negative results. In addition to this advantage, archival studies help to test whether hypothesized relationships occur in realistic settings (Barnes et al., 2018) and can inspire practitioners to test the extent to which scientific insights are relevant to their context. My archival study, conducted on the basis of 102 teams (which is still not large from a statistical power
DISCUSSION
perspective), illustrates well how the impact of hierarchy can non-intrusively be tested in realistic settings that are less available/suitable for survey research.
A promising way to further unravel the impact of hierarchy is to test the difference in impact between different types of functional group processes. As argued above, I recommend that scholars test the difference in the impact of hierarchy on learning-related processes and efficiency-related processes. If hierarchy has opposing effects between different functional group processes, meta-analyses focused on the overall relationship between hierarchy and functional group processes might yield misleading (or at least incomplete) results. For example, a non-significant impact of pyramid-shaped hierarchies ( ) on efficiency-related processes might balance out a negative impact of those hierarchies on learning-related processes. For future syntheses that compare and contrast the influence of hierarchy among types of functional group processes, a precisely defined conceptualization of functional group processes is required. Using different terms for similar phenomena may produce confusion that hinders the development of knowledge (Suddably, 2010).
Research Question 3: When does upward status-based deference occur? Across the studies, I found limited support for upward status-based deference. This limited support is striking because upward status-based deference is an assumed key mechanism through which hierarchy has its purported influence on (dys)functional group processes (Anderson & Willer, 2014; Bunderson et al., 2016; Halevy et al., 2011). My finding that goal orientation and self-efficacy affect individuals’ tendency to defer to others higher in status (Chapter 4) has important
theoretical implications. Most importantly, if replication studies also suggest that upward status-based deference depends on the personal characteristics of team members, and hierarchy affects team processes through upward status-based deference, then personal characteristics are likely to moderate the impact of hierarchy. Given that teams can vary from one another in terms of the personal characteristics of their members, it would also mean that a similar hierarchy is unlikely to have a consistent impact across teams. This contextual dependence of upward status-based deference may provide an explanation for the substantial differences found among studies on the impact of hierarchy (Anderson & Willer, 2014; Bunderson et al., 2016; Halevy et al., 2011).
In Chapter 4, where I studied the role of personal characteristics in upward status-based deference, I conceptualized deference as “yielding to one another’s opinions, beliefs, and decisions” (Joshi & Knight, 2015, p. 59). I further
conceptualized deference as instrumental – that is, aimed at improving the performance by drawing on the knowledge and expertise of others (Barton & Bunderson, 2014; Correll & Ridgeway, 2006; Joshi & Knight, 2015). I designed my study in such a way that individuals could only defer to low-quality input. By demonstrating that individuals sometimes defer to low-quality input, I wanted to point out that instrumental deference (thus with the intend of improving
performance) can impair performance. This nuanced understanding of the concept of (instrumental) deference has an important theoretical implication. The
functionalist accounts argue that hierarchy affects teams through upward status-based deference such that it mitigates conflicts and facilitates functional group processes such as collective decision making (Anderson & Willer, 2014; Bunderson
et al., 2016; Halevy et al., 2011). By expanding the concept of upward status-based deference, I created a research space for theory development in which status hierarchy sometimes results in low-quality input selection through upward status-based deference.
In addition to this construct validity matter, I dealt with a measurement validity matter in relation to upward status-based deference. In Chapter 4, I found that individuals – on average – did not defer significantly more to higher-status partners (r = .05, n.s.). This finding differed from the findings of Driskell and Salas (1991), from which I adapted the deference measure. One possible explanation for the absence of support is that participants were first asked to solve a task
themselves and were then offered to defer to poor-quality input from a peer. By asking to solve the task themselves, it might have been easier for participants to uncover the low quality of the peer input. In the study of Driskell and Salas (1991), the peer input was not of low quality. Thus, the measurement in Chapter 4 may have been appropriate and my findings could simply suggest that the relationship between status hierarchy and upward status-based deference is less consistent than assumed. When individuals feel competent enough, which may be more often when task complexity is low, they may feel less motivated to defer to higher-status team members.
Future research on when upward status-based deference occurs should take into account the role of moderators. My work suggests three broad types of moderators. First, my findings suggest that personal characteristics moderate upward status-based deference. I tested the moderating impact of personal characteristics only in one study (Chapter 3). Replication studies may help to test
CHAPTER 6
the robustness of my findings. Second, as elaborated on in Chapter 3 and as discussed above (as a construct validity matter), individuals sometimes defer to others motivated by political pressure. Political pressure could thus moderate the impact of status hierarchy on upward status-based deference. Third, as discussed above (as a measurement validity matter), individuals might feel less motivated to defer to a (high-status) partner when this partner provides low-quality input. The hierarchy legitimacy, which is greater when higher-ranked members are more valuable, might thus moderate the relationship between status hierarchy and upward status-based deference. Given the central role of upward-status based deference in the literature on the impact of status hierarchy (Anderson & Willer, 2014; Bunderson et al., 2016; Halevy et al., 2011), uncovering moderators of upward status-based deference will help to understand the impact of status hierarchy.
MANAGERIAL IMPLICATIONS
This dissertation contains a number of valuable insights for practitioners. First, my findings warn that status hierarchy can hinder individual behavior and team processes. My studies showed that status hierarchy can motivate individuals to defer to low-quality information, and showed a negative impact of status hierarchy on information elaboration (that is, the exchange, discussion, and integration of knowledge and perspectives) and coordination efficiency. However, my findings also show that, depending on the shape of the hierarchy, the negative impact does not necessarily apply. When hierarchy is pyramid-shaped ( ), teams involve in less information elaboration (Chapter 2) and suffer less from inefficient
DISCUSSION
coordination (Chapter 3). In order to intervene with information elaboration and inefficiencies in coordination, practitioners could try to ensure that teams have one person clearly at the top of the hierarchy. When manipulating hierarchy, it should be noted that pyramid-shaped hierarchies ( ) could have a negative impact on overall performance. Less information elaboration often leads to lower
performance in knowledge-intensive teams (Chapter 2). In those teams, it is better for the team performance when the team hierarchy is small and/or shaped like an inverted pyramid ( ). In two of my studies, I manipulated hierarchy – in Study 2 of Chapter 3 and in Chapter 4 – respectively by informing people about the status hierarchy and by emphasizing the social value of persons. Although my manipulations affected hierarchy, these studies did not show strong effects of hierarchy relative to the effects found in others studies. Rather than manipulating hierarchy to intervene in team processes, I recommend practitioners to be aware of (the shape of) team hierarchy when assigning tasks to teams.
Recommendation 1: Be aware of (the shape of) team hierarchy when assigning tasks to teams. When information elaboration is important, pyramid-shaped hierarchy should be avoided ( ). When efficient coordination is important, inverted pyramid-shaped hierarchy should be avoided ( ).
To recognize (and compare) the shape of team hierarchy, I recommend practitioners to compute the hierarchy skewness. This index denotes the relative width of the top and bottom of the hierarchy. Practitioners can compute the hierarchy skewness based on various types of (ordinal) data. Demonstrating the
versatility of hierarchy skewness, I computed this index on the basis of various aspects: conferred status (the prestige, esteem, and respect that individuals assigned to others), informed status (the prestige, esteem, and respect I told that persons received), salary, and conferred influence (the influence that individuals perceived others to have). As showed in Chapter 3, secondary data can be used to measure the hierarchy skewness in a non-intrusive manner. I computed hierarchy skewness on the basis of salary data to test the impact of hierarchy. Such tests help to confirm which hierarchy shape is least desirable in that context.
Recommendation 2: Recognize (and compare) the shape of team hierarchy by computing the hierarchy skewness. This index can be computed on the basis of different data types. Readily available data can be used to measure the hierarchy skewness in a non-intrusive manner.
Although hierarchy skewness plays a central role across my findings, other hierarchy properties are important as well. In testing the influence of status hierarchy, the main effect I focused on was the impact of hierarchy steepness – that is, the distance between the top and bottom of the hierarchy. I found that a larger hierarchy steepness resulted in a stronger impact, moderated by the hierarchy skewness. Regardless of the hierarchy skewness, the impact of hierarchy was weaker when the distance between the top and the bottom of the hierarchy was smaller. In other words, hierarchy skewness is less relevant when the hierarchy steepness is smaller. In addition to considering hierarchy steepness, I recommend that practitioners consider acyclicity – the degree to which influence cascades
down from the hierarchy top to the hierarchy bottom. Acyclicity also affect team processes (Bunderson et al., 2016) and its recommended measurement is, as shown in Chapter 5, largely independent of the measurement of hierarchy steepness and hierarchy skewness. By considering multiple hierarchy measures, practitioners can get a more complete understanding of the hierarchies they are dealing with.
Recommendation 3: Get a more complete understanding of the impact of hierarchy by considering hierarchy properties in addition to hierarchy skewness. Hierarchy steepness affects the functioning of teams, the effect on the team depends on the skewness of the hierarchy. Acyclicity has also been found to affect teams.
When trying to understand team functioning better through recognizing the (shape of) hierarchy, it is important to note my findings about upward status-based deference – the tendency of individuals to rely more on input from those holding higher-status positions. Status hierarchy is assumed to affect team processes such as discussions and coordination through upward status-based deference (Anderson & Willer, 2014; Bunderson et al., 2016; Halevy et al., 2011). Possibly, upward status-based deference was the mechanism through which hierarchy affected information elaboration and coordination efficiency in my studies. However, Chapter 4 showed that upward status-based deference can depend on personal characteristics. I found that particularly low self-efficacy individuals deferred to high-status peers. They did so when they had a performance
CHAPTER 6
avoidance goal orientation (PAVGO) instead of a learning goal orientation (LGO) or a performance approach goal orientation (PAPGO). Although I did not test the moderating impact of self-efficacy and goal orientation on the relationship between status hierarchy and team processes, my findings on upward status-based deference suggest that the impact of status hierarchy is stronger in teams with members that experience low levels of self-efficacy and have a stronger PAVGO. Hence, considering the self-efficacy and goal orientation of individuals can help practitioners to understand the impact of hierarchy. When individuals should not rely more strongly on the input of higher-level individuals, it is important to avoid low self-efficacy and PAVGO.
Recommendation 4: Avoid low self-efficacy and PAVGO across individuals in situations where individuals should not rely more strongly on the input of higher-status individuals. When individuals defer less to higher-higher-status others, hierarchy may have a weaker effect.
