Shared leadership and self-efficacy during organizational change in the COVID-19 crisis

(1)

Shared leadership and

self-efficacy during organizational

change in the COVID-19 crisis

Master Thesis on Pubic Management and Leadership

Ivona Hristova

Supervisor: Joris van der Voet

Ivona Hristova.

(2)

Ivona Hristova. Shared leadership and self-efficacy during organizational change in the COVID-19 crisis.

1

List of Tables and Figures

Figure 1. The target population from the ABC structure. Program Management team. – page 21 Figure 2. Social Network Analysis map of Visionary leadership – page 38

Figure 3. Social Network Analysis map of Directive leadership – page 39

Table 1. Concept operationalization for the interviews – page 29 Table 2. Reliability statistics: Cronbach’s alpha. – page 30 Table 3. Survey findings: population demographics – page 32

Table 4. Survey findings: Perceptions of the impact of the COVID-19 crisis on organizational performance – page 33

Table 5. Survey results: Mean ratings of ABC’s visionary and directive leadership – page 36 Table 6. Survey results Shared leadership behaviors ratings. – page 36

Table 7. Survey results: mean scores from all questions on the concept of self-efficacy – page 40 Table 8. Individual mean scores of directive leadership and self-efficacy used for linear regression. – page 42

Table 9. Regression statistics of directive leadership and self-efficacy – page 42 Table 10. P-value results of directive leadership and self-efficacy correlation – page 42

Table 11. Individual mean scores of visionary leadership and self-efficacy used for linear regression. – page 43

Table 12. Regression statistics of visionary leadership and self-efficacy – page 43 Table 13. P-value results of visionary leadership and self-efficacy correlation – page 44 Table 14. Descriptive findings from interviews. – page 44-45

Table 15. Causal mechanisms influencing self-efficacy identified in the interviews. – page 50

Chart 1. Survey findings: population demographics – page 32

Chart 2. Survey findings: Perceptions of the impact of the COVID-19 crisis on organizational performance – page 32

Chart 3. Survey findings: Individual perceptions of crisis severity – page 34

Chart 4. Survey results: Mean scores of the self-efficacy levels in the ABC team. – page 40 Chart 5. Survey results: Individual self-efficacy levels per member – page 41

(4)

3

Abstract

This case study explores the individual self-efficacy of the members of a single team to cope with the organizational changes caused by the COVID-19 crisis, and how their confidence is shaped by the sharing of two concrete leadership behaviours within the team. Visionary leadership has been proven effective in organizational change, and the directive style in managing crisis, both as a vertically-exerted authority. Instead, this study places these leadership behaviours in a shared leadership context. Through 8 surveys and 3 interviews, the research found that the team experienced initial levels of fears and anxieties regarding the uncertainty of the crisis, which were further transformed into high levels of self-efficacy to cope with those changes. The findings confirm that visionary leadership was indeed effective to stimulate individual confidence for performance, also when horizontally shared between equal peers. The evidence on sharing directive leadership style, however, was insufficient to demonstrate their existing relationship.

(5)

4

Chapter 1. Introduction

Organizations change in order to grow and adapt, but sometimes also to survive unforeseen crisis situations disrupting their environment and challenging their performance. In such times, organizations face a high urgency for a radical action and a quick adaption, but the surprise elements of a crisis event can meet organizational members unprepared and incapable for an immediate and adequate reaction (Sarkar & Osiyevskyy, 2018; Weick & Quinn, 1999). Often, the safeguard to the pressures of a crisis to prevent an organizational failure is the leadership managing the undesirable effects. Usually, as a result of the threat and the uncertainty of a crisis organizational authorities constrict, work flows and communication channels narrow down, structures become rigid, and the change management leadership is concentrated into a single central change agent (Kamphuis et al., 2011; Whelan-Berry & Somerville, 2010).

There are different types of leadership behaviours related to a successful change management. On one side, crisis literature puts forward directive leadership, characterized with a centralized decision-making regarding tasks structing, roles distribution, and performance monitoring (Lorinkova et al., 2013; Pearce & Sims, 2002) as a successful approach for managing organizational transformation during crisis (Staw et al., 1981; Stoker et al., 2019). On the other side, well-established organizational change theories promote for the effectiveness of transformational leadership, defined with the inspiration of individual desires to move beyond and towards an idealized vision for a better collective future (Pearce, 2004; van der Voet, 2016), for a successful change management (Allen et al., 2013; van der Voet, 2016; Witmer & Mellinger, 2016; Yasir et al., 2016). However, the empirical measures of transformational leadership have been criticized for their unclear conceptual definition, and the measurement tools’ failure to “reproduce the dimensional structure specified by theory” and “to achieve empirical distinctiveness from other aspects of leadership” (van Knippenberg & Sitkin, 2013). As a result of this critique, it has been proposed that visionary leadership style, similar to the transformational one, is a more clearly defined and empirically distinct behaviour to overcomes these challenges. Visionary leaders have been related to positive organizational performance (Kasanah, 2019; Kotter, 1995; M. Taylor et al., 2014) by actively promoting a collectively shared vision to guide the collective for a successful change (Kasanah, 2019; Kotter, 1995; Whelan-Berry & Somerville, 2010).

(6)

5

Most of the knowledge on leadership in crisis and change management is derived from studies based on exploring the vertical authority of a single change agent. However, this power centralization has been gradually challenged in favour of shared leadership where the change authority is instead decentralized across and within organizational levels because of its greater potential for a successful change implementation (Canterino et al., 2020; van der Voet, 2016). Albeit limited, the literature on shared leadership has pointed out that distributing authority horizontally among work peers is an important determinant for good group performance Most of the knowledge on leadership in crisis and change management is derived from studies based on exploring the vertical authority of a single change agent. However, this power centralization has been gradually challenged in favour of shared leadership where the change authority is instead decentralized across and within organizational levels because of its greater potential for a successful change implementation (Canterino et al., 2020; van der Voet, 2016). Albeit limited, the literature on shared leadership has pointed out that distributing authority horizontally among work peers is an important determinant for good group performance (Carson et al., 2007; Pearce, 2004; Pearce & Cogner, 2003), especially with complex problem-solving tasks requiring creativity (Pearce, 2004; Wood, 2005). For this reason, this paper deduces its theoretical assumptions on crisis and change management leadership from studies based on a vertical authority, and goes beyond the given context by placing them in a decentralized shared leadership context. The exploration of the horizontal distribution of the two concrete leadership behaviours – directive and visionary, among all members of a team, requires an exhaustive investigation of their interpersonal dynamics. The majority of shared leadership studies, however, treat teams as aggregate, whole units without unwrapping the dyadic (between two people) exchanges (Canterino et al., 2020; van der Voet et al., 2014), which crisis-based organizational change certainly is. This allows the assumption that shared leadership could be a suitable fit for effective organizational transformations, instead of traditionally studied vertical one.

For this reason, this paper deduces its theoretical assumptions on crisis and change management leadership from studies based on a vertical authority, and goes beyond the given context by placing them in a decentralized shared leadership context. The exploration of the horizontal distribution of the two concrete leadership behaviours – directive and visionary, among all members of a team, requires an exhaustive investigation of their interpersonal

(7)

6

dynamics. The majority of shared leadership studies, however, treat teams as aggregate, whole units without unwrapping the dyadic (between two people) exchanges (Canterino et al., 2020; van der Voet et al., 2014). For this reason, literature has encouraged for more focus on the interpersonal relationships, and only then to relate them to performance (D’Innocenzo et al., 2016; Pearce & Cogner, 2003). Such within-team relationships mapping is possible through the application of a novel empirical approach called Social Network Analysis (SNA) which allows for the understanding of a team’s interpersonal dependencies. Inevitably, the way team leadership is shared will have an effect on the individual attitudes, confidence, and motivations about personal performance (D’Innocenzo et al., 2016; Kadushin, 2012).

This study focuses on one of those dimensions, that is the confidence for performance, also called self-efficacy, and its relationship with the concrete shared leadership behaviours in a particular organizational change context. Generally, empirical findings suggest that employees who exhibit strong self-efficacy in dealing with non-routine tasks are also performing better, up to initially-set higher standards. Self-efficacy is worth exploring because it is an important predicator for the commitment and effort intensification that employees invest in the work processes (Bandura & Cervone, 1983; Bandura, 2001), and invites a potential insight about individual performance effectiveness (Bandura & Cervone, 1983; Locke & Latham, 2002; van der Hoek et al., 2018), which in turn would define the quality of their final outputs.

1.1. The knowledge gaps

Based on the theory presented above, current knowledge on crisis-based organizational change and teams’ role in the process is still limited, offering a number of knowledge gaps worth exploring. This paper moves beyond the well-studied central change agent by looking at organizational change in the context of shared leadership. It does so by applying the novel SNA approach to address the knowledge gaps left from aggregate-centred studies on shared leadership.

Moreover, it approaches the knowledge gaps in yet another novel approach for the current crisis and organizational change literature. Most knowledge is based on a for-profit setting, leaving the public and non-profit sector understudied (Crevani et al., 2007). Studying the way public organizations change and the relevant leadership dynamics, however, is crucial since due to their public service role, their direct environment is highly sensitive to social, economic, and political changes. Thus, generating more knowledge on the sectors like the non-profit can

(8)

7

provide crucial support for public leadership for smooth and successful transitions in entities with high social value, which are equally susceptible to external environmental changes, not to say crises. Moreover, academic literature has already encouraged the shared leadership approach to be applied to non-profit organizational changes (Akingbola et al., 2019).

For this reason, the non-profit shared leadership is conceptualized with the two dominant behaviours in their respective fields – directive, a prominent leadership approach in the crisis literature (Kamphuis et al., 2011; Stoker et al., 2019) and visionary leadership from the change literature (Akingbola et al., 2019; Allen et al., 2013; M. Taylor et al., 2014; van der Voet, 2016). The question here is whether those leadership behaviours would be as effective when the authority is shared horizontally among equal peers, as opposed to the traditional centrality of leadership. Their effectiveness is measured in a novel way by exploring how team members experience the exhibition of each behaviour by relating them to member’s individual self-efficacy in performing work-related tasks. This self-efficacy is an indicator for the commitment and effort in one’s performance, which would be an important determinant for expecting positive work results (A. Bandura & Cervone, 1983; Albert Bandura, 2001).

To sum up, an effective organizational change would be possible when team members are confident in the success of their effort (A. Bandura & Cervone, 1983), and directive and visionary leadership behaviours in a shared leadership context have the potential to provide the conditions for stimulating team members’ self-efficacy (Weiner, 2009). The measurement of the concepts takes place in a public non-profit organizational change in the COVID-19 crisis as an ideal context to explore the presented knowledge gaps.

1.2. Research question

To explore all the knowledge gaps above, the study asks the following research question:

How does the sharing of directive and visionary leadership behaviours within a work team influence individual members’ self-efficacy to handle the COVID-19 organizational change?

(9)

8

1) What is the individual perception of directive leadership among the members of the ABC team?

2) What is the individual perception of visionary leadership among the members of the ABC team?

3) What is the individual self-efficacy among the members of the ABC team?

4) What is the relationship between the perceived directive leadership and the individual self-efficacy among the members of the ABC team?

5) What is the relationship between the perceived visionary leadership and the individual self-efficacy among the members of the ABC team?

1.3. The case

Based on the reasons, presented above the research focuses on the case of an non-profit organizational change, and particularly a single team with a shared leadership structure, in the context of the COVID-19 crisis. The study takes a 8-member team of a Dutch-based non-profit organization, called the ABC (name changed) for its case study due to its suitability to the research needs. ABC is an internationally operating NGO, working with partners from the public and private sector from Europe and Africa. Each team member has a non-repetitive role in comparison to their peers, thus they provide a unique leadership input, and can benefit from a variety of expertise. Having faced and dealt with the COVID-19 crisis together, collectively achieving a successful organizational survival, adaptation and transformation, the ABC team is a great case to investigate shared leadership behaviours’ effect on individual self-efficacy in dealing with the situation.

1.4. Research methods

The study relies on mixed research methods approach. Primarily, team’s dyadic relationships are studied through surveys where each member assesses the received influences based on the leadership behaviours exhibited by their peers. Consequently, the strength of the concrete leadership behaviours is assessed for their possible effect on the reported levels of self-efficacy through statistical linear regression analysis.

(10)

9

Then, while there is a certain risk of reverse causality, where self-efficacy might be the one encouraging the leadership behaviours, this risk is averted by the strong theoretical basis, and is also mitigated by a second research method: semi-structured interviews to exhaustively explore the underlying factors stimulating self-efficacy to face organizational changes, and to what extent do the leadership behaviours help this confidence.

1.5. Theoretical relevance

The study contributes to the academia in a number of ways. A major methodological innovation is the deviation from the traditional aggregate approach where teams are treated as a single unit. Instead, it overcomes this limitation through the application of a novel empirical approach - Social Network Analysis, by exploring the team’s interpersonal dynamics in a very new light. This allows an innovative mapping of the leadership network that illustrates a snapshot of all dyadic work-based relationships when facing a crisis.

In addition, there is a number of theoretical contributions. First, it offers an important advancement for the organizational change literature by exploring the potential effects of horizontal leadership distribution in the process of organizational transformation, instead of the traditional vertical authority. Second, it contributes to both organizational change and shared leadership knowledge by conceptualizing and comparing two opposing leadership styles – directive and visionary. Such comparison expands the understanding of the conditions for a successful leadership environment, as well as an effective organizational change. The effectiveness of each leadership style is assessed by correlating the influence of the particular behaviour to the individuals’ confidence and motivation for performance. This approach allows a different angle for judging shared leadership effectiveness instead of asking team members to rate leadership effectiveness directly, which usually risks biased measures and undermining reliability.

1.6. Societal relevance

The empirical nature of the study, its novel research approach, topic relevance to the rapidly changing organizational and team structures, as well as timeframe importance considering the COVID-19 crisis, are antecedents for generating knowledge with a high societal value. The

(11)

10

findings would allow managers, leaders and members from all organizational levels to get familiar with the level of complexity of social relationships within a single work entity. The mapping of the dyadic relationships can educate managers that want to stimulate shared leadership within a team. The insights on the leadership behaviours and their effect on self-efficacy can serve as informative tools in employees’ training and development programs. The knowledge generated can be especially useful for public and non-profit leaders, as the case study is based on a team within such an organization, as due to the fact that their direct work and performance indicators are tightly related to the societal well-being and dynamics, and hence they are also highly vulnerable to the external changes. Thus, public and non-profit managers can benefit use the findings of this study to understand what factors attract employees and stimulate their sense of well-being within the organization. Overall, the research results can be used as deemed suitable for the education and team-related performance improvement on any organizational level, for routine tasks or organizational changes, as well as regardless of sector, but particularly useful for the non-profit field.

1.7. Paper structure

The rest of the paper is organized in the following manner. The second chapter presents the theoretical framework, developing all concepts concerned in this research, and variables to be measured to answer the research question. Consequently, the research design and methodology are exhaustedly described. Chapter 4 objectively presents the empirical findings obtained from the surveys and interviews, followed by their analysis and theory-based interpretation in Chapter 5. The concluding chapter summarizes the new knowledge generated, with all its limitations, and recommendations for future research to fill in the gaps of the study, overcome some of its constrains, and further enrich the literature on shared leadership, self-efficacy and organizational change.

Chapter 2. Theory

This chapter presents the theory based on which the hypotheses are deduced. First, as the study is situated in a very concrete context, it is conceptualized as a “crisis-based

(12)

11

organizational change” in a “shared leadership” setting to allow the reader to understand how this study is different from others.

Consequently, the independent variables are explained: the directive and visionary leadership behaviours, primarily deduced from vertical leadership frameworks, but each style also conceptualized in the organizational change and shared leadership setting. As independent variables, the strength of those leadership behaviours is related to the way they shape the only dependent variable of this study: individuals’ self-efficacy to deal with change process as a result of the COVID-19 crisis.

Based on the theory, the conceptual model assumes that team members’ confidence and motivation for performance are influenced by the exhibited leadership behaviours they receive from their work peers. This way, despite the limitations of a cross-sectional study as opposed to a longitudinal one, the study aims to scratch the surface of individual attitudes in shared leadership, and create grounds for future studies to deepen this knowledge.

The research is primarily based on assumptions derived from existing theory on organizational and leadership literature. Hence, the theory below serves as a base to deduce the hypotheses presented in the end of this chapter, and tested through an empirical approach.

2.1. Crisis-based organizational change

Organizations change constantly, and sometimes they are forced to change to ensure survival when they face turbulences in their direct or indirect environment during unforeseen crisis situations (Weick & Quinn, 1999). The conceptualization of a crisis is well summarized by Sarkar et al (2018), whose working definition illustrates a crisis as an event or series of disruptions, perceived as an organizational threat due their unpredictable character, duration, and impact on performance. Often, crisis brings undesirable outcomes, and creates a high urgency for action while organizational structures stiffen, and employees lack the needed readiness to react (Sarkar & Osiyevskyy, 2018; Staw et al., 1981; Weick & Quinn, 1999). Sarkar et al (2018) operationalize organizational crisis based on Dutton (1986) 3-point framework: 1) the importance of the issue – the extent to which an organization is affected by the crisis in case of inaction;

(13)

12

2) the immediacy – how fast are the undesirable outcomes expected, hence how quick should an organization react in order to mitigate the potential harms;

3) the uncertainty – the inability of a manager to foresee the type, degree, continuation, and scope of the harm on organizational performance.

Based on those three dimensions, the paper assesses the context in which the research is based on – the global pandemic caused by the spread of the novel COVID-19 virus. The rapid rise of infectious and mortality rates pressured virtually all national governments to announce emergency situations, and undertake the relevant to a crisis situations measures – closing borders, businesses, restricting movement (Kuo & Sullivan, 2020). As a result, the crisis has had its impact on established work practices and routines, affecting and altering organizational needs, methods, resources. The handling and rearrangement of the crisis-related changes would require a firm leadership to mobilize the change process, set necessary goals and design relevant strategies (Staw et al., 1981; Stoker et al., 2019). As the change leadership is a key determinant for organizational performance in the crisis context, it is the main lens of investigation of this research, and the more concrete leadership behaviours in the change management are treated as the central independent variables in this study.

2.2. Leadership in organizational change

Traditionally, the literature on change leadership has extensively emphasized on the role of the central change agent as the main authority of enabling and guiding the processes (Armenakis et al., 1993; Weick & Quinn, 1999). However, the narrow focus on the central change agent has left questions about the potential of distributing leadership horizontally to empower organizational members to take an equal lead in the change processes, known as sharing leadership, below also referred to as “horizontal” or “horizontally distributed” authority between the members of a single team.

Shared leadership is a relatively new concept in the organizational and team management literature, however, increasingly relevant due to the growing importance of teams’ social dynamics for the overall team performance (Carson et al., 2007). The way work peers share tasks and responsibilities, and receive support and advice from their colleagues are the key precondition for shared leadership. This structure overrides the traditional vertical authority

(14)

13

through a collective empowerment based on a shared purpose, mutual social support, and individual voices, and as a result stimulates work satisfaction and performance (Carson et al., 2007; Pearce, 2004).

Shared leadership is especially suitable for complex, non-routine tasks requiring interdependency and creativity (Pearce, 2004), as well as problem solving (Wood, 2005), which organizational change certainly is. Dunphy and Stace (1988) early on acknowledge the potential participative organizational change through “overcoming resistance”, and encouraging an “industrial democracy” which involves organizational members in all change stages. Albeit limited, the more recent literature on shared leadership in organizational change has proven the positive effects of distributing authority across organizational levels for effective planning and mobilizing of change. Nevertheless, these studies also point out that more research is needed to understand the group dynamics, and their effectiveness in implementing change (Akingbola et al., 2019; Binci et al., 2016; Canterino et al., 2020, 2020; van der Voet, 2016).

Based on the knowledge and its limitations, the following paper measures the potential effectiveness of two specific and opposing leadership styles, derived from two different fields of studies: directive leadership from crisis literature (Staw et al., 1981; Stoker et al., 2019), and visionary style from the organizational change literature (M. Taylor et al., 2014; van der Voet, 2016). Their assessment in this context is with the purpose of evaluating their effectiveness, already proven in a vertical crisis-based change management, however, through an innovative lens of shared leadership structure. The plotting of these two opposing leadership styles also allows the comparison of the potential benefits for effective change management within horizontally managed teams.

2.2.1. Directive leadership

Directive leadership is the first of the two central leadership concepts in this study due to the solid theoretical base that puts it forward as a suitable and effective approach in the coordination of crisis responses in an organization (D. C. Dunphy & Stace, 1988; D. Dunphy & Stace, 1993; Staw et al., 1981). Based on the Threat-Rigidity theory (Staw et al., 1981), the directive leadership is the preferred crisis response in organizational change management,

(15)

14

specifically in the type of dramatic environmental disturbances which the COVID-19 crisis has caused, due to the narrow constriction of control and focused authoritative style of command in the overall organization of the change process. This narrowing of power and information allows management to effectively channel efforts, communication and resources in responding to the urgent organizational needs for the crisis survival (Kamphuis et al., 2011; Staw et al., 1981; Stoker et al., 2019).

Traditionally, directive leadership is associated with a centralized structure and focus on assigning tasks, goals, and responsibilities to subordinates (Pearce & Sims, 2002), however, often compromising on the individual consideration of and building relationships with work peers (Kerr et al., 1974). Directive leadership is based on a number of theoretical frameworks:

• Theory X by McGregor (1960) that assumes that a worker is unable to perform without externally-imposed motivation, goals, and direction.

• Initiating structure, developed by a series of The Ohio State Studies (Kerr et al., 1974), that entitles the leader with the initiation, planning and coordination of group goals, tasks, activities, and communication, often without a consultation with subordinates or peers.

• Task-oriented behaviour that puts an explicit focus on planning and coordinating all task-related matters (Yukl, 2012).

To summarize, the directive behaviour mainly focuses on a centralized decision-making and goal assignments, single-handed definition of roles and responsibilities in the task organization and distribution, and exclusive issuing of instructions and expectations for performance (Pearce & Sims, 2002). While typically observed in power concentration in the higher levels of a vertical authority (Staw et al., 1981; Yukl, 2012), directive leadership can also be exercised among team members belonging to the same hierarchical level. Pearce and Sims (2002) explore it in the context of both vertical (central leader) and shared (among team members) leadership, proving that work peers can also exhibit those exclusive decision-making behaviours towards their colleagues. Their effectiveness, however, is highly contested. While directive leadership has been proven beneficial in vertically managing organizations during a crisis (D. Dunphy & Stace, 1993; Kamphuis et al., 2011; Staw et al., 1981; Stoker et al., 2019), and even desired by employees in times of uncertainty (Rast et al., 2013), it has been

(16)

15

found rather ineffective in a shared leadership setting (Pearce & Sims, 2002). This duality makes it an intriguing case to measure its effectiveness during a crisis in a horizontal instead of a vertical leadership structure.

To extend the argument on leadership in crisis management in organizational change, the directive leadership is compared to another behaviour, proven to be beneficial in organizational change (Ateş et al., 2020; M. Taylor et al., 2014; Witmer & Mellinger, 2016), as well as in a shared leadership environment (D’Innocenzo et al., 2016; Mayo et al., 2003) – that is the visionary leadership.

2.2.2. Visionary leadership

Visionary leadership has been studied in closeness to transformational one as its variation (Kasanah, 2019; M. Taylor et al., 2014), being the necessary guidance and emphasis on a collectively shared vision for all organizational members (Kasanah, 2019; Kotter, 1995; Whelan-Berry & Somerville, 2010). While transformational leadership has been widely studied in the context organizational change (Allen et al., 2013; van der Voet, 2016; Witmer & Mellinger, 2016; Yasir et al., 2016), a growing critique on its conceptualization and empirical measurement puts forward the idea that the approach should be remodelled to consider more concrete and measurable aspects of this behaviour, such as encouraging and team-oriented communication, which are represented by the visionary leadership style (van Knippenberg & Sitkin, 2013).

Visionary leadership inspires the achievement of an idealized goal with practical and concrete strategies for performance. The visionary style is found to be positively related to team and organizational performance (Ateş et al., 2020; M. Taylor et al., 2014), and is especially suitable for change management, as an organizational transformation begins with a vision beyond the current status quo, and sets a central idea that guide all goals, objectives, activities and inputs (Antonakis et al., 2004; Kotter, 1995; Kouzes & Posner, 2012; M. Taylor et al., 2014).

The visionary leader sets a common ground for a collective vision by challenging existing work and decision-making practices. They encourage new ways of working in order to achieve new performance levels by regularly and clearly communicating the ultimate goal and its milestones with all parties in the change process (Kotter, 1995; M. Taylor et al., 2014). Visionary leaders trust the team, demonstrate confidence and willingness to try-and-fail, and

(17)

16

favour individual contribution. This way, they allow power sharing instead of exerting tight vertical control over the process (Kouzes & Posner, 2012), making it a great fit in a horizontal leadership distribution.

Having explored those two leadership behaviours, their application in organizational change would be futile if they would not bring the necessary benefits and positive performance for the team and the organization. In order to measure their effectiveness, the study assesses their impact on the way each behaviour style stimulates, or not, individual confidence to deal with the challenges of an organizational change. While the confidence levels do not objectively reflect organizational performance, it is a promising indicator for understanding team dynamics in times of crisis.

2.3. Self-efficacy

Person’s confidence in their ability to perform a task successfully is called self-efficacy. The concept of self-efficacy is situated within the Social Cognitive Theory (SCT), developed by Bandura (2001). SCT is constructed on the human agency of “perceived control over the nature and quality of one’s life” (Bandura, 2001), distinguishing three dimensions of such control: direct control, influence over another person’s agency to act accordingly, and the collective abilities. In the following research, the focus is explicitly limited to the concept of self-efficacy regarding the immediate circle of direct control of an individual in the context of work-related activities (Bandura, 2001). Hence, the study is not concerned with general self-efficacy, and its other domains, such as social life.

Perceived self-efficacy is a fundamental precondition for good performance, because without it, failure is much more likely than otherwise (Bandura, 2001). Empirical studies find that employees exhibit strong self-efficacy when they are dissatisfied with current performance, and are willing to set their own ambitious, yet realistic, attainable goals, and would consequently demonstrate high commitment and better strategies (Locke & Latham, 2002; van der Hoek et al., 2018). Confident employees are motivated to invest in their performance, and more likely to accomplish their personally-set ambitious goals (A. Bandura & Cervone, 1983; Albert Bandura, 2001). Thus, the effective organizational change would be influenced by the employees’ confidence in the success of their efforts in change implementation (Albert

(18)

17

Bandura, 2001; Weiner, 2009). The measurement of self-efficacy in this concrete research is placed in the context of the COVID-19 crisis, and the confidence of the employee’s to meet the demand for an organizational change and to ensure organizational survival and effective performance.

2.4. Conceptual model of shared leadership and self-efficacy in organizational change

Based on the knowledge available, a few expectations can be derived about the interactions between the independent and dependent variables for the deduction of hypotheses. As seen, there is a sufficient knowledge base supporting the assumption that directive leadership is effective for managing crisis (Kamphuis et al., 2011; Staw et al., 1981; Stoker et al., 2019), and visionary in managing change (Ateş et al., 2020; M. Taylor et al., 2014; van der Voet, 2016; Witmer & Mellinger, 2016), however, exercised from a central vertical authority in aggregate organizations. (Armenakis et al., 1993; D. C. Dunphy & Stace, 1988; Weiner, 2009). This study builds on this knowledge by instead focusing on sharing the authority among team members. First, the effectiveness of shared leadership, and more specifically the two concrete leadership behaviours in this context, are determined based on the strength of the leadership experienced, and its impact on individual confidence in their performance. This is the concept of self-efficacy, proven to result in better strategies and performance outputs (Bandura & Cervone, 1983; Locke & Latham, 2002; van der Hoek et al., 2018), allowing the assumption that the measurement of team members’ self-efficacy would still provide an insight into their potential levels of performance.

Based on the theory, each leadership behaviour is expected to have different effect on the individual self-efficacy levels. For example, directive leadership has been found to be negatively correlated with motivation in a team with shared leadership (Pearce & Sims, 2002). However, placing this shared leadership structure in the context of crisis-based organizational change, which is very different than the situation in Pearce & Sims (2002) study, could have had different effect on team motivation and productivity. A big multi-level study covering almost 1000 organizations coping which crisis has found directive leadership to be rather effective and beneficial for successfully coping with organizational changes (Stoker et al., 2019). This is because due to the uncertainty and a challenge of a crisis, “workers exhibit

(19)

self-Ivona Hristova. Shared leadership and self-efficacy during organizational change in the COVID-19 crisis.

18

uncertainty to meet work demands on their own, relying on leaders to provide structure and guidance” (Rast et al., 2013). Thus, the latter evidence supports the assumption of a positive effect of directive behaviours on team members’ confidence in times of a crisis-based organizational change. Based on this knowledge, the following hypothesis is formulated:

H1. Sharing directive leadership among work peers stimulates their individual self-efficacy for the implementation of organizational change.

The other change-relevant leadership behaviour standing out in the literature is the visionary style (Antonakis et al., 2004; Kotter, 1995). While also proven effective for a hierarchical, single-agent organizational change, it differs from the expectations laid on the directive leadership with its emphasis on the collective vision, fundamentally opposing the singularity of the directive style. By inspiring team members to align with the collective purpose and internalize the common goal, visionary leaders overcome a potential obstacle for performance where workers lack motivation for achievement because do not see the direct connection between the daily task and future vision (Carton, 2018), especially when leadership is concentrated on the high tiers of vertical authority and alienates lower management and their structures (Ateş et al., 2020). The strong emphasize on the collective goal and efforts implies the importance and recognition of the individual contribution, which stimulates it rather than constricting it (Carson et al., 2007). Hence, the recognition of efforts and achievements encourage motivation, intensify effort, and boost one’s self-efficacy for performance (Locke & Latham, 2002; van der Hoek et al., 2018), allowing the following hypothesis:

H2. Sharing visionary leadership among work peers stimulates their individual self-efficacy for the implementation of organizational change.

Chapter 3. Methodology.

This chapter outlines the research methodology for the measurement of the central concepts and hypotheses testing. First, it presents the team of the case study, with all team members being the target research population. Then, it describes the specific empirical approach applied to the context of shared leadership, and the concrete qualitative and quantitative research methods for the assessment of the concerned concepts. The concept

(20)

19

operationalization is based on established empirical scales from the academic literature to fit the concrete research tool, followed by a description of the analysis strategies applied to each set of data collected. Finally, the research methodology is critically assessed based on the validity of measures and the reliability of its findings.

3.1. The Case

The investigation is structured as a case study, focusing only on the team dynamics in the organization of ABC, name changed for the purpose of the study, in order to exhaustively deconstruct the interpersonal dynamics within a team that has dealt with the COVID-19 crisis. ABC is a non-profit organization that operates on the international stage, facilitating efforts of entrepreneurship and private sector development between numerous countries. While the ABC has a more complex structure, the study focuses on a 8-person Program Managers (PM) team that has initiated, planned and implemented organizational changes in order to meet the new working conditions of the pandemic. The ABC’s Program Managers team operates under the formal authority of a Management Team (MT), but enjoys the flexibility to set goals and strategies within the general mission and vision of the organization. The PM team structure is illustrated in Figure 1.

The PM team is a great case for studying shared leadership because each team member has the equal opportunity to offer support, advice and guidance to their colleagues. They are also representative for a team dealing with the urgent COVID-19 changes due to the nature of their work, which was usually highly dependent on social contacts and personal meetings, and thus radically changed in the wake of the pandemic and social gatherings restrictions. The 8-member team is not an isolated social unit, however, as a team it operates in a horizontal leadership with only essential boundary definition and guidance from the central management team.

Usually characterized with non-routinized and creative problem-solving tasks, as well as knowledge work, ABC’s team adaptation to the pandemic has unfolded the opportunity for each team member to be a valuable contribution in shared leadership dynamic in the process of organizational change. This made it the suitable case to study the leadership’s potential in stimulating individuals’ confidence in facing the inevitable changes in the ABC.

(21)

20 Figure 1. The target population from the ABC structure. Program Management team.

3.2. Empirical approach

The main empirical approach of this study is the Social Network Analysis (SNA) because of its ability to unwrap the interpersonal social dynamics within a team, allowing an in-depth measurement of the decentralization and distribution of shared leadership among all members (Carson et al., 2007; D’Innocenzo et al., 2016; Pearce & Cogner, 2003; Pearce & Sims, 2002). The SNA approach is applied through the design of a survey where team members are asked to rate each of their work peers on the strength of the concrete leadership behaviours they have observed.

The application of the SNA is useful for a number of research objectives. First, the ratings from the survey allow the measurement of the strength of the existing dyadic (between 2 people) leadership-based relationships in the ABC team, which scores are used for testing the hypotheses. Thus, the identification of all significant dyadic relationships represents the quantitative measurement of the independent variables in this study.

Second, the SNA approach also serves a descriptive purpose by providing the means for mapping the distribution of dyadic leadership relationships among individuals (Kadushin, 2012). This builds on the theory that traditionally, decision-making is centralized in a single

Management Team 3 members

(22)

21

person with vertical authority, whereas in teams with shared leadership influence is more evenly distributed among peers (Carson et al., 2007). According to the SNA approach, the equal distribution of dyadic relationships means that the leadership network is fully a decentralized, which is the ideal shared leadership arrangement (D’Innocenzo et al., 2016; Kadushin, 2012). Thus, the Social Network maps provide an illustration of the way the shared leadership in the ABC team is constructed, which allows the contextualization of the findings, and supports the replicability of the research design and data collection process.

The SNA methodology is inspired by a few empirical studies concerned with shared leadership and SNA: Carson et al (2007), Pearce and Sims (2002), Mayo et al (2003), the meta-analysis by D’Innocenzo et al (2016), and the theoretical book on social networks by Kadushin (2012). Pearce and Sims (2002) and Mayo et al (2003) approaches also allow the measurement of the different behaviour styles, such transformational and transactional in their case. The shared leadership mapping strategy in the Methodology section is specifically inspired by Mayo et al (2003).

3.3. Research methods

The research is based on a mixed research methods, combining both qualitative and quantitative research tools. Primarily, hypotheses are tested through a quantitative approach with a survey distributed among all team members, complemented with qualitative interviews with a few representatives of the ABC team. This way of organization offers the great advantage of data triangulation by obtaining information on the same question in more than one way. It allows for a more decisive confirmation or rejection of assumptions, strengthening and more exhaustive justification of evidence (Almalki, 2016).

The collected data focuses on the following: individuals’ perception of the COVID-19 crisis impact on organizational performance; the strength of the perceived visionary and directive leadership behaviours in the organizational change, the level of individual’s self-efficacy to meet those changes, as well as the possible relationships between the dependent and independent variables.

The survey offers a great advantage of collecting answers from all team members in a less time-consuming manner. It benefits the study by offering a sufficient illustration of the general

(23)

22

state of each of the concerned variables. The survey is also a great tool to achieve one of the goals of the study - the mapping of dyadic relationships to draw the Social Network map, and measuring leadership decentralization by identifying all dyadic relationships. However, the limitations of relying on a survey alone is that it would fail to explain causal mechanisms between those leadership behaviours, and may even risk a reverse causality. This is why the interviews, albeit insufficient on its own to map the shared leadership, are applied as a complementary tool to potentially identify the causal mechanisms stimulating self-efficacy of individual workers, and whether those horizontal leadership behaviours are indeed as influential as expected in the hypotheses. The two methods are complementary applied to the target population. They simultaneously investigate the concerned variables by the means of the specific operationalization fit for each research tool. Below are explained each research tool with the applied operationalization and analysis strategy.

3.3.1. Survey

One survey with 15 closed question was distributed among the eight ABC team members. The survey was the primary research tool because it allowed the collection of wide range of information from all team members in a less time and resource intensive manner. The survey covered 4 variables that answer 3 of the main research sub-questions:

• directive leadership • visionary leadership • self-efficacy

• the anticipated impact of the COVID-19 crisis on the organization. All survey questions are included in the Appendix.

The survey measures respondents’ demographics and individual perceptions of the impacts of the COVID-19 crisis on organizational performance to allow study contextualization and replicability.

Then, the different leadership behaviours perceived were measured based on the strength of the reported dyadic relationship between two team members. The measurement approach of the specific leadership behaviours is based on Mayo et al (2003) who quantifies their

(24)

23

strength in a Likert scale format. In this particular survey, the strength is rated from 1 - not at all to 7 - very much. A strong dyadic relationship between two peers is determined when the rating above or equal to 5. This is because while 4 is the midpoint on the scale of 7, a proof for a strong relationship would require a slightly higher score than the middle point. Thus, anything above 5 is considered a “beyond the average”, hence a strong relationship. Consequently, based on the SNA approach, all strong (≥5) dyadic relationships are mapped with one or two-way arrows to illustrate both directive and visionary leadership interactions in two distinct shared leadership maps.

Lastly, the survey answers sub-questions 4 and 5 by exploring individuals’ self-efficacy to deal with the organizational changes brought by the pandemic as a dependent variable in order to identify any statistical correlations to depict a possible relationship between the concrete leadership behaviours and self-efficacy. All variables are operationalized as described below. The internal consistency of those measurements is determined through reliability analysis on the answers given by identifying their Cronbach’s alpha scores.

Organizational change

All independent and dependent variables are measured in the context of the COVID-19 crisis. For this reason, team members are first asked about their perceptions of the severity of the crisis based on Dutton (1986) 3-point framework: 1) the importance of the issue; 2) the immediacy; 3) the uncertainty. Those were measured in a Likert scale format (1 – weakest and 7 – strongest value). The purpose of these descriptive statistic is to provide more understanding about the individual and team work-related mood regarding the possible impacts of the crisis on the organizational performance. This allows placing the main variables in a more concrete context, which serves not only descriptive purposes but also supports the possible replicability of the study.

Directive leadership

The concept of directive leadership is based on a number of distinct theories that promote structure & task-oriented behaviours that focus on instructing and guiding workers in order

(25)

24

to perform (Kerr et al., 1974; McGregor, 1960; Yukl, 2012). The scale of the survey is adapted from the statements developed by Lorinkova et al (2013): “expects me to perform up to a certain standard” ; “clarifies which tasks are expected from me to deliver” ; “defines my work roles and concrete responsibilities”. All statements are rephrased into questions, asking to what extent a concrete teammate engages in the specific behaviour. The survey measures the directive leadership behaviours based on their perceived strength by the receiver in the form of a Likert scale (1 – not at all, to 7 – very much) in the surveys.

Visionary leadership

Visionary leadership has been often studied under the concept of transformational leadership by multiple scholars in the management and leadership literature, as well as in change management specifically (M. Taylor et al., 2014; Mayo et al., 2003; Pearce & Sims, 2002; van der Voet, 2016). The operationalization of visionary leadership is based on the questions on transformational leadership from van der Voet (2016), as they measure leadership effectiveness in organizational change context. The scale was modified and adapted to include only visionary leadership relevant statements: “contributes to my feeling of belonging to the team and its mission” ; “inspires me to work towards the collective vision” ; “challenges me to think about old problems in new ways” (van der Voet, 2016). All statements are rephrased into questions, asking to what extent a concrete teammate engages in the specific behaviour. The survey measures those based on the subjected team member’s perceived strength in the form of a Likert scale (1 – not at all, to 7 – very much).

Self-efficacy

The measurement of individual self-efficacy is based on the Social Cognitive Theory developed by Bandura (2001). The author has largely contributed to the academic field with numerous theoretical and empirical works on the different aspects of self-efficacy (Bandura & Cervone, 1983; Bandura, 2001, 2012). The operationalization of self-efficacy in this survey is based on his “Guide for constructing self-efficacy scales” (Bandura, 2006). All statements were formulated in the context of COVID-19, as following:

(26)

25

• Confidence in handling COVID-19-related uncertainties.

• Confidence to ask for help in meeting the COVID-related work challenges.

• Confidence in their contribution to the COVID-19-related organizational changes. The measurement of self-efficacy is based on a wide Likert scale from 1 (strongly disagree) to 7 (strongly agree) to reflect the variety and the depth of one’s confidence levels ( Bandura, 2006).

Other measurements

The survey includes the measurement of three key variables for describing the survey participants: age, gender and tenure of the members. Those serve as descriptive indicators for explaining on what population sample are the survey findings based on, and allows research replicability.

Data analysis

A few pieces of information are derived from the survey data. First, each variable is measured through its respective mean scores. The analysis strategy for drawing the descriptive statistics for each concept is as following:

• Perceived crisis severity and impact on organizational performance – both with a descriptive purpose:

o individual level: the mean of each team member scores of the three questions on the impact of the COVID-19 crisis;

o team level: the mean score for each question on the impact of the COVID-19 crisis;

• Directive leadership:

o within the team as the average of all scores on reported directive behaviours: serves informative purpose;

o exhibited by each individual as the mean score of all ratings given by their peers: serves to identify an individual’s centrality in the social network, informative purpose.

(27)

26

o perceived by an individual as the mean score of all ratings given to their peers – serve as an independent variable in hypothesis testing;

• Visionary leadership:

o within the team as the average of all scores on reported directive behaviours: serves informative purpose;

o as exhibited by each individual as the mean score of all ratings given by their peers: serves to identify an individual’s centrality in the social network, informative purpose.

o as perceived by an individual as the mean score of all ratings given to their peers – serve as an independent variable in hypothesis testing;

• Individual self-efficacy:

o individual level: the mean score of each team member’s responses on the five statements on self-efficacy to deal with the COVID-19 crisis;

o team level: the mean score for each statement of self-efficacy to deal with the COVID-19 crisis;

The shared leadership is depicted as a Social Network Analysis (SNA) through the mapping of all dyadic relationships (≥5). Thus, between two individuals there can be the following types of relationships:

• Visionary leadership:

o Two-way relationship where both individuals experience leadership from their peer;

o One-way relationship where only one receives leadership, and the other does not;

• Directive leadership:

o Two-way relationship where both individuals experience leadership from their peer;

o One-way relationship where only one receives leadership, and the other does not;

The SNA maps are purely informative to provide an image on the shared leadership dynamics in the ABC team. There are two maps, for each of the two leadership behaviours, illustrating

(28)

27

the distribution of these concrete dyadic relationships among members. Those maps are analysed based on the number of one and two-way relationships: a higher number of dyadic relationships indicates a significant leadership decentralization, which means that there is a stronger practice of shared leadership in the ABC team (D’Innocenzo et al., 2016; Kadushin, 2012). This information serves to illustrate the context in which the study takes place.

Then, the direction and strength of the relationships between the independent (the two leadership behaviours) and dependent (self-efficacy) variables were analysed through linear regressions, which indicates how strongly correlated individuals’ self-efficacy is to each of the exhibited leadership behaviours. Those will be one of the main indicators testing the hypothesis.

3.3.2. Interviews

Semi-structured interviews with 3 out of the 8 team members complemented the survey data. The interviews were held over a phone call in a one-week span. Their duration ranged from 16 to 27 minutes, with an average length of the interviews is 22 minutes. All three interviews were recorded and transcribed. The transcripts can be found in the Appendix.

The interview questions were mainly concerned with the work-related perceptions of the COVID-19 crisis, the level of self-efficacy to deal with the organizational changes, and the factors influencing one’s self-efficacy. The operationalization of the concepts related to the leadership behaviours is based on Yukl (2012) typology of leadership behaviours as such:

• Directive leadership: planning, clarifying, monitoring tasks and work performance. • Visionary leadership: supporting , developing, empowering;

Self-efficacy was measured based on Bandura (2006) “Guide for constructing self-efficacy scales”, opting for statements that demonstrate confidence and motivation for dealing with work-related challenges in the context of the COVID-19 as the main focus of the study. The perceptions for the crisis impact were assessed based on Sorokowski et al (2020) findings regarding COVID-19-related social moods. They find that the exposure to information about the development and consequences of the COVID-19 crisis increase anxiety among individuals. Hence, in this study it is assumed that when faced with the possible crisis impacts,

(29)

28

ABC team members would also experience the fear and anxiety identified in Sorokowski et al (2020) study.

The data was organized based on Binci (2016) methodology, making a category for each concept, and further defining the concepts based on their dimensions deduced from the above-mentioned literature. Table 1 presents the operationalization of each concept.

The interviews explored through questions such as:

• How did you experience the COVID-19 crisis when it started changing the work in the ABC?

• How did your role and work responsibilities change in the context of the COVID-19 crisis?

• How did you first feel in the face of the uncertainty around the pandemic regarding the necessary organizational changes?

• How did your attitude change over time? What influenced this transition? • How did the social environment in your team help you cope with the changes?

The open-ended questions avert the risk of biased answers by allowing the respondents to independently elaborate on their assessment of the concerned variables, as well as to identify the causal mechanisms affecting their self-efficacy, and whether the team leadership was a contributing factor. The interview guide, including the interview questions, is included in the Appendix. It contains the more concrete questions about the perception of the social and interpersonal team environment in relation to one’s confidence in case those were not mentioned at first stages.

Table 1. Concept operationalization for the interviews

Concept

Codes

Example

Directive leadership

Instructing tasks execution Defining roles and responsibilities

Performance expectations

“My team mate instructs me how to complete a task.”

“My teammate defines how the final outcome should look like. “

Visionary leadership

Contributing to the team feeling Inspiring collective vision and effort

Empowering individual contributions Inspiring new ideas

“My teammate(s) encourage the whole team to achieve goals together.” “My teammate shares inspiring new ideas

(30)

29

“My teammate is proactive about improving our working methods.”

Self-efficacy

Comfortable with uncertainty Confident with work improvisation

Feeling of being a valuable contribution to the team work

“I feel motivated about working with my team in this crisis.”

“I feel confident that I/my team can manage the new changes.”

Crisis

severity Feelings of: Anxiety; Stress; Fear

“I was feeling anxious of losing my job.” “I felt very stressed about the changes.”

The data was analysed based on a deductive coding involving reading through each interview transcript and identifying all similar key words and statements as the one mentioned in Table 1. Consequently, for each interview, each concept was rated in the same scale as the surveys - from 1 (weakest) to 7 (strongest) to help to complement the descriptive scores from the surveys. Consequently, the causal mechanisms were identified per individual interviewee, based on the words signifying relationship such as “because of”, “due to”, “helped”, “contributed” etc that connect two concepts. Quotes and examples for each of the variables and causal mechanisms were used for a detailed illustration of the findings. The conclusions from the interview were combined with the survey data in the Discussion section to confirm or reject the hypotheses, and to further explain the relationships between the variables.

3.4. Reliability and validity

Reliability

For a cross-sectional approach, the study strengthens its reliability with the application of mixed research methods. The use of both qualitative and quantitative data compensates for the individual method’s weaknesses, and combines their strengths (Almalki, 2016). While a quantitative tool offers the collection of wider range of data, it impedes a more in-depth understanding of its scores and correlations. It risks overseeing factors such as reverse causality (where the dependent variable is actually independent, and vice-versa), or the omission of any other factors influencing self-efficacy. Moreover, the reliability analysis of the Cronbach’s alpha scores were slightly unsatisfactory at 0,640 (Table 2), where a score of 0,7 and above is the acceptable for proving good reliability (Tavakol & Dennick, 2011).

(31)

30

Those weaknesses were strengthened with a qualitative research tool, which on its own would have either provide more limited data, or require more time and energy to collect all the data, obtained by the quantitative approach. The interviews allowed respondents to elaborate on their own understandings of the concepts, and independently report their perceptions in the given context.

Nevertheless, the cross-sectional character of the study presents a single snapshot in time of the team dynamics, which are otherwise never static. Therefore, while the research strategy and data collection methods are replicable, the reproduction of the same results in the same team in a different moment might be contested.

Reliability statistics Cronbach’s alpha score Cronbach's alpha based on standardized elements N of elements ,640 ,727 4

Table 2. Reliability statistics: Cronbach’s alpha.

Validity

The internal validity of the findings has been optimized by using multiple sources for the operationalization of each variable. While in social sciences words are always a subject of personal interpretation, the formulation of both survey and interview questions has been widely based on well-recognized authors and their contribution to the relevant academic fields. Moreover, the limitations of the closed-questions survey were compensated with open-ended questions interviews that allowed the population sample to independently elaborate on the relationship and possible causal mechanisms between the key variables. The external validity is harder to uphold due to the main characteristic of this research, being a case study that focuses on an extremely limited population sample. The research has a solid theoretical base: the relationships established in the hypotheses are well supported by the academic literature. However, as they are tested in a very novel context, that is shared leadership, their external validity is yet hard to determine.

Shared leadership and self-efficacy during organizational change in the COVID-19 crisis