Leadership style & change readiness

Leadership Style

&

Change Readiness

Student: Supervisor:

Irma Derksen dr. Nathan E. Betancourt

I

Statement of Originality

This document is written by student Irma Derksen who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

II

CONTENT

INTRODUCTION ... 1

CHANGE READINESS ... 4

Types of change in organisations ... 4

Change in a monopolist vs. a non- monopolist organisation... 5

Change readiness ... 6

Change readiness in a monopolist organisation ... 7

LEADERSHIP... 9

Leadership ... 9

Styles of leadership... 9

Styles of leadership in monopolistic organisations ... 10

STYLES OF LEADERSHIP AND CHANGE READINESS... 12

DEMOGRAPHIC CHARACTERISTICS ... 13

The Moderating Effect of Business Units ... 13

The Moderating Effect of Demographic characteristics ... 14

RESEARCH METHOD ... 18

Research model ... 18

Target group ... 20

Data collection ... 20

Analysis of the results ... 24

III

ANALYSIS OF THE RESULTS ... 27

Preparing the data of change readiness ... 27

Preparing the data of the MLQ ... 28

Combining the data... 32

HYPOTHESIS TESTING ... 37

Style of leadership ... 37

Style of leadership and change readiness ... 39

Business Unit hypotheses ... 41

Demographic characteristics hypotheses... 43

CONCLUSION AND DISCUSSION ... 48

STRENGTHS & LIMITATIONS AND ADVISE FOR FUTURE RESEARCH ... 53

BIBLOGRAPHY... 56

APPENDIX A: Multidimensional Leadership Questionnaire ... 66

APPENDIX B: Change readiness Questionnaire ... 67

APPENDIX C: Ordinal scaling of the Business Units ... 68

APPENDIX D: Questionnaire with style of leadership ... 69

APPENDIX E: Pivot tables ... 70

APPENDIX F: Combined data by means ... 72

APPENDIX G: Combined data by matching ... 80

APPENDIX H: Dataset for ANOVA with method of means ... 81

1

INTRODUCTION

Change readiness of employees involved in the process of change is important for this process of change (Kotter, 2001) and a good indicator for the effectiveness of the style of leadership (Pielstick, 2000). Leaders can assist and motivate employees to be change ready and it is very important to know what factors may influence the individual change readiness of employees. Multiple papers have shown that the style of leadership does make a difference for individual levels of change readiness (Thomas A. B., 1988) (Smith, Carson, & Alexander, 1984) (Church, 1995). However, the type of organisation experiencing the change is an important factor that can impact the effect of leadership styles on change readiness. For example, Yo and Miller showed that different generational groups prefer a different leadership style in various organisations in the manufacturing industry, than they do in organisations in the education sector (Yo & Miller, 2005). While researchers have examined the relationship between styles of leadership and change readiness for many types of commercial organisations, the monopolist is a notable exception from this stream of literature. One reason for this omission may be that monopolists rarely experience change as the absence of competitors removes necessary incentives for change (Porter, 1978). Other reason for this omission may be that monopolist organisations aren’t as occurring as commercial organisation; everyone can start a business, but for a business to become the only supplier of a particular service or good with no close substitutes and without the price being influenced by the market (Baumol & Blinder, 2015) is near impossible because of the market itself where newcomers will occur if there is opportunity (Porter, 1979) and regulations that want to keep that opportunity for newcomers and refrain organisations from becoming monopolist via antitrust laws (European Commission, 2014) or anti-monopoly laws (Li & Li, 2014). Third reason for this omission may be that the monopolists that do exist, are not open for research or

2 simply have not been included in research. But because change does exist in monopolist organisations, it is good to research the styles of leadership within these organisations and the effect these styles have on change and fill this gap.

Therefore, this thesis asks the following research questions: What is the dominant style of leadership in a monopolist organisation? The styles of leadership are taken from the Hartog,

Muijen and Koopman (Hartog, Muijen, & Koopman, 1997) research, which in turn builds on the extensive work that Bass and Avolio did in combination with other researchers ( (Bass, Avolio, & Goodheim, 1987) (Bass, 1990) (Bass & Avolio, 1993) (Avolio & Bass, 1999) and are the Transformational, the Rational-Objective and the Passive style of leadership.

What is the relationship between these different styles of leadership and the individual's

change readiness within a monopolist organisation? This question is asked to fill the gap in

the literature about this relationship. But there are an almost infinite amount of factors that can influence this relationship between the styles of leadership and an individual’s change readiness. To put depth and focus in this research, four possible influencers have been chosen to research further. What is the moderating effect of business Unit and the demographic characteristics on the relationship between style of leadership and individual’s change

readiness? Since groups often have a major influence on the behaviour, morals, and standards

within the group (Bearden & Etzel, 1982) (Marc W. Heerdink, Homan, & Fischer, 2013) (Halevy, Chou, & Murnighan, 2012), it is plausible that characteristics of the business unit could impact the relationship between leadership styles and change readiness. Similarly, prior research has found that demographic characteristics have an important effect on change readiness (Epitropaki & Martin, 2004) (Weber & Weber, 2001) (Dijkstra, Roijackers, & Vries, 1998) (Miller & Spilker, 2003) (Isenhart, 1997).

3 With this research, this gap is being filled for the characteristics age, tenure and educational level. Age and tenure are chosen for two main reasons; the heightened age limit for retirement makes it of high importance to get more insight in the effect these characteristics can have on the relation of the style of leadership. Next to this it is because the findings of Epitropaki and Martin (2004) suggest that employees in different groups of age and tenure hold similar perceptions on leadership, so the assessment of the style of leadership between different groups will be the same. The research of Madsen, Miller and John (2005) shows a slight relationship between educational level and change readiness, indicated by identification, loyalty and organisational commitment in this research. Thus from this the educational level is chosen, next to the assumption that higher educated employees have better understanding of the complex need for change, and if the educational level of employees is higher, the change readiness will be higher.

4

CHANGE READINESS

Types of change in organisations

Change can be categorized in two types, where many definitions have been used; Evolutionary and revolutionary change (Greiner, 1972, p.40), linear quantitative and nonlinear qualitative change (Putney,1972, p. 32), rational and radical change (Grabow & Heskin, 1973, p. 476), and first and second order type of change (Levy & Merry, 1986). All definitions make distinction between a planned, sequential, slow and predictable change with continuity of the organisation versus an abrupt, disruptive, quick and unpredictable change where there is no continuity of the current organisation.

First order changes are relatively minor improvements or adjustments such as changing communication systems, technological systems or personal rewards, also called systematic change. First order change is slower, less ambitious, more focused, and more carefully constructed and sequenced than dramatic change. Often it is promoted by staff groups and consultants who handle planning and organisational development and does not change the system’s core (Levy & Merry, 1986) (Huy & Mintzberg, 2003) (Gareis, 2010). Second order change is the opposite; it is quick, highly ambitious and not focused because most of the time the continuity of the organisation is at stake (Gilbert & Bower, 2002) (Reeves & Deimler, 2011). The organisation decides –or is being forced- to change their core business significantly and fundamentally (Levy & Merry, 1986) (Huy & Mintzberg, 2003) (Bergquist, 1993). Most organisations try to follow first order change to avoid second order due to its disruptive nature (Gilbert & Bower, 2002) (Reeves & Deimler, 2011).

In this study, we follow the definition of first order change, but to initiate change the organisation has to be change ready. As Bartunek and Moch (1987) showed, changes occur

5 with particular steps already shared by members and it is important to understand the differences among the members with orders of change (Bartunek & Moch, 1987)

Change in a monopolist vs. a non-monopolist organisation

A monopolist organisation is an organisation with the exclusive possession or control of the supply of or trade in a commodity or service (Oxford University Press, 2016). A monopoly frequently arises from government support (Friedman, 2009, p. 208). This monopolist organisation has the highest barriers to entry among strategic groups (Porter, 1978), because no competitors are allowed as enforced by law. This way, ‘the presence of multiple strategic groups that affects the process of competitive rivalry’ (Porter, 1978) is absence. So the incentive of developing sources of advantage, developing a competitive advantage, which in turn is necessary for achieving superior performance (Slater, 1996) is not present. Also, the monetary incentive to maximize profit by reducing costs or increasing value is not present because the regulator defines the earnings of the organisation on a yearly basis. So there is little incentive to change from the nature of the monopolist organisation.

For non-monopolist organisations, change is an ongoing process in organisations. A subject that has been given a lot of attention by scholars and the need for change is described in many articles; to gain competitive advantage as an organisation (Barney, 1991) or for an organisation to stay in ‘fit’ with the ever changing environment (Venus, Stam, & Knippenberg, 2015) (Quinn, 1989). This need for change in commercial organisations is always build on the basis that organisations need to maximize profit by increasing value or reducing costs and finding commonalities of being unique (Liu, 2013), but has always to do with gaining, enlarging or keeping a spot in an industry (Porter, 1978).

6 Change is one reality with which individuals, groups and organisations must constantly cope in order to survive and the most critical aspect of effective management (Singh, 2005). Change is defined as ‘the coping process of moving from the present state to a desired state that individuals, groups and organisations undertake in response to dynamic internal and external factors that alter current realities’ (Singh, 2005, p. 4). In a (by law) regulated monopolist organisation the environment is very stable. Following the rationale of Gareis (2010) the speed of learning and changing should be low because the speed of learning and changing has to meet the dynamics of the environment.

Change readiness

So the frequency of change differs between a non-monopolist organisation and a monopolist organisation, but both have change efforts to undertake. These efforts all too frequently fail to achieve their intended aims (Beer & Nohria, 2000) and organisational change efforts are often so poorly managed that they accelerate organisational crises (Probst & Raisch, 2005). This points to an important role of the style of leadership. Rafferty et al. (2012) state that few issues are as critical as employees’ attitude towards change. Change readiness is the most prevalent positive attitude toward change that has been studied in the organisational change literature.

The definition of Armenakis et al. (1993, p. 681) to define change readiness as “an

individual’s beliefs, attitudes, and intentions regarding the extent to which changes are needed and the organisation’s capacity to successfully undertake those changes” is used in the

7 Change readiness is important to get to the process of change. The process of change was drawn up by Lewin in 1947 where he proposed that the process of change consisted of unfreezing, change and refreezing (Lewin, 1947). Kotter expanded on this model, which resulted in the detailed version of Lewin’s model, the N-step model. This is a combination of Lewin Unfreeze-Change-Refreeze model (Lewin, 1947) and the eight-step model of Kotter (Kotter, 1995), where the importance of change readiness is incorporated in the fifth step (see table 1). To get to the change part of the model, empowering others to act on the vision of the change is a necessary step (Kotter, 1995), and change readiness is an indicator to see if empowerment has happened.

Table 1: N-step model Lewin and Kotter (2001)

Change readiness in a monopolist organisation

Change readiness is influenced by structural factors that reflect the circumstances under which change is occurring and the extent to which these circumstances enhance or inhibit the successful implementation of a change (Holt & Vardamann, 2010). The circumstances under which change is occurring is different between a monopolist organisation protected by law; the organisation has no competition and a set income so the circumstances are very stable. Scholars researched change readiness in commercial organisations (Armenakis, Harris, & Mossholder, 1993) (Armenakis, Feild, & Harris, 2007) (Holt & Vardamann, Readiness for Change, 2010) or in personal healthcare situations (Dijkstra, Roijackers, & Vries, 1998)

1. Establishing a sense of urgency 2. Forming of a powerful coalition 3. Creating a vision

4. Communicating about that vision 5. Empowering others to act on that vision 6. Planning and creating short-term wins

7. Consolidating improvements and producing still more change

Refreezing 8. Institutionalizing new approaches Unfreezing

8 (Musselwhite & Plouffe, 2010), where change is needed to adapt or remain fit to the environment or simply for survival.

Change readiness is theorised for very stable environments (Porter, 1978) (Rafferty, Jimmieson, & Armenakis, 2012). Because of minor environmental pressure to change, the change rate is low, but the knowledge about this can have a side effect. The position of a monopolist organisation is known by its employees and is expected to have an influence on the beliefs, attitudes, and intentions regarding to the extent to which changes are needed. More specific; the employees know that the environment is very stable and thus may not see a need to change and thus have a low change readiness level. With the Armenakis et al. (1993, p. 681) definition used that change readiness is a bundle an individual’s beliefs, attitudes, and intentions regarding the extent to which changes are needed and the organisation’s capacity to successfully undertake those changes, it is of utmost importance that a leader can help the employee see this need to change. The task of the leader to help the employee get change ready and thus using the right style of leadership is very important. The relation between the three styles of leadership on change readiness within a monopolist organisation is a gap in the literature that has to be addressed.

9

LEADERSHIP

Leadership

Formal leaders are those in a “position of leadership, where informal leaders are not in positions of leadership but recognized as leaders nevertheless (Pielstick, 2000). Leadership in this thesis is confined to the formal leadership roles, on all hierarchical levels. An organisation can benefit by having the right leaders in the right place in the right time and thus makes the role of formal leader of great importance for an organisation. The importance of formal leadership roles is a much covered subject by scholars. From the top management of an organisation (Hambrick, 1995) (Kotter J. P., 2001) (Fiol, 2002), to the middle managers of an organisation (Huy, 2002) (Caldwell, 2003) (Balogun, 2003), where both the strategic direction of the organisation (Dunphy & Stace, 1988) as well as managing employees (Huy, 2003) (Caldwell, 2003) are important. For this research we focus on how leaders vary in how they manage their employees, as captured by the styles of leadership construct.

Styles of leadership

The style of leadership is the manner and approach of providing direction, implementing plans, and motivating people. As seen by the employees, it includes the total pattern of explicit and implicit actions performed by their leader (Newstrom, Davis, 1993). To assess the style of the leader, the Multifactor Leadership Questionnaire was developed by Bass, Avolio and Goodheim (1987). They examined how world-class leaders differ on five transactional/transformational leadership factors. In following research in 1993, Bass and Avolio showed that the best relation with change readiness typically displayed both transformational and transactional leadership, as evidenced by the positive correlations between ratings of these two leadership styles (Bass & Avolio, 1993).

10 Later on, the model was expended to 3 styles of leadership with an adapted Multi Factor Leadership Questionnaire, which covers three scales of leadership; transformational, rational-objective and passive style of leadership.

The transformational leadership style defines the need for change, creates new visions,

mobilize commitment to these visions and leaders utilizing this style can ultimately transform the organisation (Hartog, Muijen, & Koopman, 1997).

The rational-objective leadership style is about motivating the employees to perform as

expected where the leader clarifies the performance criteria and objectives, employees’

responsibilities, the rewards for meeting the objectives, and correcting measures for failing to meet objectives (Eagly, Johannesen-Schmidt, & Engen, 2003).

The passive style of leadership, also known as the laissez-faire, is the frequent absence and

lack of involvement during critical junctures (Hartog, Muijen, & Koopman, 1997).

Both the transformational as rational-objective style of leadership are active styles, where passive style of leadership is –logically– a passive style. Every leader has a dominant style of leadership, but this does not exclude elements of other styles and the elements of every style of leadership can be learned (Jehn, Northcraft, & Neale, 1999).

Styles of leadership in monopolistic organisations

The styles of leadership as found in commercial organisation is assumed to be applicable in a monopolist organisation the same way as in commercial organisations, because there is no reason in the literature why this would not be the case. This research will determine if all the

11 styles are represented within the monopolist organisation to see if the prior assumption holds true. Next is the spread of the styles of leadership. Because of the nature of the monopolist organisation and its stable environment, the need for visions and transformations of the organisation is low. The organisation has a clear raison-d’être as described in the laws and regulations protecting the organisation, but also the goals and penalties are described in the regulations. Therefor it is expected that the style of leadership that is similar to this, will be the most dominant one. This means that it is expected that the rational-objective style of leadership will be the most dominant one. This leads to Hypothesis 1; The three styles are all represented in the monopolist organisation and the dominant style of leadership within a

12

STYLES OF LEADERSHIP AND CHANGE READINESS

Vision on the change, and thus positive beliefs, attitudes, and intentions towards this change, is the most common, distinguishing characteristic identified with leadership (Pielstick, 2000), which makes change readiness a good indicator if the style of leadership has an effect. From the studies of Bass and Avolio and Goodheim, (1987), and the research building on this (Bass & Avolio, 1993) (Yammarino & Dubinsky, 1994) (Hartog, Muijen, & Koopman, 1997) (Antonakis, Avolio, & Sivasubramaniam, 2003) (Muenjohn, 2008) (Eagly, Johannesen-Schmidt, & Engen, 2003), comes the conclusion that transformational leadership shows the highest positive effect on the change readiness. The rational-objective style of leadership has a positive effect on the change readiness (Bass B. M., Avolio, Jung, & Berson, 2003). The passive style of leadership is a form of absence of leadership and there is a negative association between the style and employee performance, effort and attitudinal indicators (Bass, 1990). It implies that this style is not suitable to lead for change readiness, but Kerr & Jermier (1978) propose several employee, task and organisation characteristics that could reduce the importance of leadership and lead to ‘empowerment’ of followers which could even make a passive style of leadership useful for change readiness. All these studies show that there is a relation between style of leadership and change readiness, but the question remains if that holds for every type of organisation and specifically if it holds for a monopolist. So Hypothesis 2: Every dominant style of leadership has a different level of change readiness within a monopolist organisation

13

DEMOGRAPHIC CHARACTERISTICS

The Moderating Effect of Business Units

A Business Unit is ‘a logical element or segment of a company (such as accounting, production, marketing) representing a specific business function, and a definite place on the organisational chart, under the domain of a manager. Also called department, division, or a functional area’ (Business Dictionary, 2016) The influence of Business Unit Level is chosen for two very different reasons. The monopolist has processes that the organisation is structured in a departmentalized way, where processes flows run within departments of Business Units but more often across two or more Business Units. To implement change within in processes which run across various Business Units, the differences between change readiness of these Business Units is of high importance for developing and testing the effectiveness of transfer strategies between those Business Units (Lehman, Greener, & Simpson, 2002).

Across different working positions, employees hold similar perceptions of ideal leadership (Epitropaki & Martin, 2004). But the effect of the leadership style on change readiness can differ with the “openness” of the Business Unit with the environment. If a Business Unit characteristically has more interaction with external competitive environment than another Business Unit, e.g. procurement versus control, one can expect that the first Business Unit will be more prone to change readiness than the second. This because the importance to keep aligned with the environment and thus the need for change is more evident (Morgan, 2006). The influence of the Business Unit is tested for all styles of leadership with Hypothesis 3: Business Units with more interaction with the external competitive env ironment have positive

14

The Moderating Effect of Demographic characteristics

A lot of literature covers the effect of leadership on organisational change (Hambrick & Mason, 1984) (Hill & Levenhagen, 1995) (Herold, Fedor, Caldwell, & Liu, 2008) (Fiol, 2002). In the research, the demographic characteristics within organisations are not often taken into account. Despite the important, sometimes critical, role of such additional theoretical concepts, researchers often leave them loosely specified and unmeasured, creating a "black box" filled with vague, untested theories (Lawrence, 1997). That the style of leadership does make a difference is proven in multiple researches (Thomas A. B., 1988) (Smith, Carson, & Alexander, 1984) (Church, 1995) but with which factors the style of leadership resonates is not often clear, whilst the organisations all differenced in performance levels. In this research the demographic characteristics age, tenure and educational level are chosen because of their indisputability and their constant nature. Yo and Miller showed that different generational groups prefer a different leadership style in the manufacturing industry, but not in the education sector (Yo & Miller, 2005), Jehn, Northcraft and Neale show that diversity in gender as a social category and educational background and work experience as informational category as factors for improved performance (Jehn, Northcraft, & Neale, 1999).

The hypothesis about the different demographic characteristics falls apart in the following various different demographic characteristics from the Epitropaki and Martin (2004) research, because these characteristics are proven to have no effect on the assessment of the style of leadership. So the labelling of the style of leadership is not influenced by the demographics. Reason for the distinction between age and tenure is the explosive growth-rate of the monopolist organisation in the last 8 years, and tenure longer than 8 years has a maximum of 10% of the population. So there can be relatively young employees with a long tenure, but

15 also relatively old employees with a short tenure. This population gives a chance to test both variables as moderator on the relationship of the style of leadership on change readiness. This provides variable information about hiring old(er) employees or making sure your employees get a long tenure with the organisation.

Influence of age and tenure

With the heightened age limit for retirement, it is of utmost importance to investigate the effects coming with employees of older age. From the Upper Echelons Theory, age is named as observable characteristic that have a significant influence on the style of leadership (Hambrick & Mason, 1984). It is expected that these characteristics also have a significant influence on the effectiveness of the style of leadership on change. The effect that age can have on the assessment of the style of leadership is researched by Epitropaki and Martin (2004) and has showed no significant difference. This reduces the possibility that one group would assess a style of leadership of one leader differently than another group. So the employees of different age group will assess a style of leadership similar, but the question is if the same employees of different age groups experience the same effect of the style of leadership on change readiness. This is a gap in the literature that will be filled with this research. Weber and Weber (2001) found no significant influence of age on change readiness, but a behavioural slowing is proven with age in the research of Birren, Woods and Williams (1980). And for change readiness, a change in attitude is needed which is expected to be slower with an increasing average age level. The assumption is that with growing age, the change readiness declines, so the hypothesis is thus; Hypothesis 4: Age has a negative moderating effect on the relation of the style of leadership on change readiness.

16 The findings of Epitropaki and Martin (2004) also suggest that employees in different groups of tenure hold similar perceptions on leadership. But they tend to follow more persistent strategies, strategies that conformed to central tendencies of the industry, and exhibiting performance that closely adhere to industry averages. Organisational routines are a central feature of organisations ( (Feldman & Pentland, 2003). Routines can create an ongoing opportunity for new practices and patterns of action, but also for creating inertia (Hannan & Freeman, 1984), inflexibility (Weiss & Ilgen, 1985), and working on the ‘auto-pilot’ (Ashforth & Fried, 1988). A long tenure in the same organisations indicates multiple repeated routines, where the social stigma direction with multiple repeated routines points in the direction of change averseness. Teams with long tenure tend to follow more persistent strategies, strategies that conformed to central tendencies of the industry, and exhibiting performance that closely adhere to industry averages (Epitropaki & Martin, 2004). So the literature points towards a more change resistance attitude with a lengthening tenure, so the hypotheses are formulated in the same direction. This leads to the following; Hypothesis 5: Tenure has a negative moderating effect on the relation of style of leadership on change

readiness.

Proverbs as “Better the devil you know than the devil you don’t know” (English proverb) and “You can’t learn an old crow new tricks” (Dutch proverb) points in the direction that an old(er) age as well as an long tenure makes one ‘rusty’ and not change ready. Although the two will be correlated to some extent, simply because a long tenure can only be achieved with a certain age, both are a variable here. It is important to make the distinction, because it can be that a long tenure has a negative effect and age has a positive effect, which would speak for hiring older employees. It can also be the other way around, which would speak for hiring young employees and make sure they won’t leave the company until a certain amount of years of tenure is reached.

17

Educational level

The Knowledge-Based Theorists propose that knowledge is the most important organisational capability and source for competitive advantage for every business (Grant, 1996) (Teece, Pisano, & Shuen, 1997) (Sveiby, 2001) (Conner & Prahalad, 1996) (Nickerson & Zenger, 2004). From a multitude of research done on change readiness in Medical Sciences, mostly on addictive behaviour, education on the reasons to change is key (Dijkstra, Roijackers, & Vries, 1998) (Miller & Spilker, 2003) (Isenhart, 1997). The assumption may be made that employees of higher educational levels are more understanding of the complex needs of an organisation and the environment has on the organisations, so their need for change is bigger. So if employees are educated in the need for change, the change readiness grows and if the educational level of employees is higher, the change readiness will be higher. The research of Madsen, Miller and John (2005) shows a slight relationship between educational level and change readiness, indicated by identification, loyalty and organisational commitment in this research.

When employees are educated on the need for change, the change readiness is higher. This is done in the transformational style of leadership, where the biggest change readiness is assumed. But because change can be complex to understand, the question is raised if higher educated employees resonate better with the message of need for change. The slight relation has been proven in the Madsen, Miller & John research (2005) and will be tested. Change readiness is a cognitive state that has a multilevel and complex build-up (Rafferty, Jimmieson, & Armenakis, 2012). It is expected that higher educated employees are better capable to understand and process this complexity, thus a positive direct interaction is presumed. This leads to Hypothesis 6: A higher educational level has a positive moderating influence on relation of the style of leadership on change readiness.

18

RESEARCH METHOD

Research model

With the research question and underlying hypotheses, the research model can be drawn up.

Figure 1: Research Model

The change in performance in this study is focussed on the change in attitude and behaviour of people (PEOPLE), one of the 3 performance areas defined in the change programme of the monopolist. The purpose is given form by communicating the new strategy presentation and the process is given form with the assignment of six major processes to be investigated and in detail written down by a specially formed team. But the main interest is people; there are 6 specific parts in this performance area that are selected to be improved over time. Change readiness is one of the facets of the change in attitude.

19

Figure 2: Performance areas change programme

Translation of the Dutch enumeration from the PEOPLE column; 3. Change of attitude 4. Decision making 5. Keeping word 6. Feedback Culture 7. Communication 8. Assessment Process

The programme should help the organisation to analyse existing structures and processes and identify and strengthen drivers for change. The second step is to identify and assess new patterns and to initiate the necessary changes. Third step is to fixate the new structures, processes and behavioural patterns. This is following the model of Lewin of unfreezing, change and refreezing (Lewin, 1947)

20

Target group

The target group for this research are the employees posted in the Netherlands in the monopolist in the energy transmission sector, TenneT TSO B.V. , where only a newly erected Business Unit was excluded due to is international character and the fact that it was still in the development phase when the research started. This results in a total population of 1.448 internal and external employees (per date 31-12-2014).

Data collection

The research is qualitative with a deductive approach, to test existing theories and enrich the knowledge on the influential factors on change readiness. The data collection is done by a survey that can be filled in anonymously via Qualtrics, after receiving an invitation. The survey questions are all forced-choice questions and besides the demographic questions all on the Likert-scale from 1 (not at all) to 5 (totally agreeing). This scale gives the respondents the self-assessment of the level of agreement with the statements given.

Multidimensional Leadership Questionnaire1

To assess the leadership style within the monopolist organisation, the Multidimensional Leadership Questionnaire X5 (MLQ) is used. To deduct any form of socially desirable answers or an Asch situation and get the truest answers possible, the survey will be completely anonymous. In the Asch situation, the subject knows the unanimous, incorrect judgements of the other members of the group before he makes his own response. Thus he may give the same answer as the others because he takes their answers as evidence about

1

21 reality (the informational conformity to the group). The subject gives his response publicly hence his response may be a function of concern with the evaluation of his behaviour by the group (normative conformity to the group) (Schulman, 1967).

In consideration with the organisation, it was requested to inform and ask for concurrence from the managers of the employees in all layers of the organisation. The invitation to participate in the survey was not obligatory but voluntary. Because of the concurrence of the manager, the sample is distorted by the gatekeeper function the manager has by judging if his group can be part of the sample. There were 111 e-mails send in which the research was explained, what was asked from the employees and if the employees could be addressed for this survey.

After the manager’s concurrence, the survey was sent by individualized and anonymized links (1.178) to all the employees who are hierarchically directly connected to this manager. The .invitation to participate in the survey was not obligatory but voluntary so employees who received the survey could choose not to participate.

Complementary questions for change readiness2

The questions determining the dependent variable change readiness should have come from the Employee Satisfaction Survey of Tower Watson. There were 5 questions about change readiness in the survey, related to the Power to Perform change programme where change readiness is measured as variable of change in attitude on a scale of 1 to 10. The raw from the

2

22 Employee Satisfaction Survey had the huge advantage that there would be a high response rate, high reliability and validity rates and the data supplied would fit this research perfectly so the measurement validity was high. Next to this, it would show the knowledge of the use of secondary data in research. Because of the knowledge about the survey in form of the type of questions, the scaling and the response rates, there was confidence that when the data would be released, it would be usable. Because the process of releasing the data would take a long time, the decision was made to release the MLQ already. This way, the data of both sources would be available for analysis at the same time.

However, after going through the process and signing a statement of secrecy, the released secondary data proved to be unusable. Not the raw data was provided, but the data aggregated to a level of percentage “favourable”, “unfavourable” and “indecisive” on Business Unit level and organisation level, without amount of respondents or boundaries of unfavourable and indecisive (favourable was a 7 or higher). The raw data was not available for research, in contrary to prior understandings. This meant that the dependent variable of this research had no valid ground and there was need for a solution. Multiple solutions are possible; using questions from the MLQ with risk of measuring something the MLQ is not designed and validated for, working around it and being unable to test anything within this thesis but have arguments for this, or doing an additional short survey.

By using questions from the MLQ, the benefit is that the answers are from original respondents, but the cost is that there is high multicollinearity because the questions from the MLQ are not designed to measure change readiness (Williams, 2015). The second option is redoing the data collection by issuing a new questionnaire where the style of leadership and change readiness are both measured with their own questions in one go. Benefit of this is that the data for both variables would be from the same respondent and there is minimal change of

23 multicollinearity because both variables have their own questions. The cost is that the organisation had to approve another big survey. The third option is issuing an additional questionnaire about change readiness to combine with the MLQ questionnaire. Benefit is that this would take a short amount of time of the organisation and change readiness is measured with its own questions so minimal change of multicollinearity. Cost is that there is no way to match individual respondents so the outcomes have to be matched by means. This reduces the validity, but if the respondent rate is high enough this should be of minimal impact with the hypothesis testing because higher response rates assure more accurate survey results (Parker & Rea, 1997) and a high response rate reduces the impact of individual answers because the missing data from non-respondents is not random (Altman & Bland, 2007, p. 424).

The choice is made for the last option, because the ambition is to measure style of leadership on change readiness, so a dependent variable is needed. The measurement for this dependent variable has to be valid and reliable, so the option is to make a new short and personal survey based on the research of Holt, Armenakis, Field and Harris (Holt, Armenakis, Feild, & Harris, 2007)

So a supplementary set of questions has to be administered within the organisation to measure change readiness. This measurement is subjective, but provides a good image about the readiness for change within departments. This survey is done with the same method as the distribution as the MLQ, via a web-based fully anonymized questionnaire. All the employees who received invitation for the MLQ received an e-mail for the second questionnaire with 15 questions, of which four about age, tenure, Business Unit and Educational level. The other eleven questions are selected from the research of Holt, Armenakis, Field and Harris (2007). The questions selected are all the questions with an eigenvalue of .8 or higher from their research and were asked on the same Likert-scale of 1 to 7 (where 1 is strongly disagree and 7

24 strongly agree) as used in their research. Reason for this selection is that the second survey has to be very short, so only the strongest items are incorporated. The questions are correlated with each other and were determined as indicators for change readiness (Holt, Armenakis, Feild, & Harris, 2007).

Due to the short time frame, the survey is conducted in three days and thus could influence the results (Field, 2013). But with the goal-amount of 200+ surveys, the risk of variance and Skewness is reduced (Tabachnick & Fidell, 2001, p. 75). With this survey the strength is also that it was sent by individualized and anonymized links. The invitation to participate in the survey was not obligatory but voluntary so employees who received the survey could choose not to participate. Also the very short amount of time it would take to complete the survey is a strength.

Analysis of the results

The response rate of the MQL is analysed to determine the validity of the usage of the data to say something about the organisation. Of the whole population of 1.448, there were 1.178 send to the employees were the manager concurred. Of this, 532 started the questionnaire and 492 completed it. This is a response rate of 45,16% and a ‘completed’ response rate of 33,98%. This is a good response rate, following the results of the study of Baruch and Holtom (2008) where the average response rate for studies that utilized data collected from organisations is 35,7% with a standard deviation of 18,8.

The response rate of the change readiness survey is analysed to determine the validity of the use of the data to say something about the organisation. Of the whole population of 1.448, there were 1.178 send to the employees were the manager concurred. Of this, 297 started the

25 questionnaire and 297 completed it. This is a response rate of 20,51% and a ‘completed’ response rate of 20,51%. This is an acceptable response rate, following the results of the study of Baruch and Holtom (2008) where the average response rate for studies that utilized data collected from organisations is 35,7% with a standard deviation of 18,8.

The results will be analysed using IBM SPSS. The data will consist of:  The independent variable is style of leadership, nominal variable  Assumed moderator is the Business Unit, a ordinal variable3

 Assumed moderators are the demographic characteristics age, tenure, and educational level, two discrete continuous variables and an ordinal variable

 The dependent variable is the change readiness, a discrete continuous variable

Reliability and validity

Validity is whether an instrument actually measures what it sets out to measure. Reliability is whether an instrument can be interpreted consistently across different situations (Field, 2013). The use of the Multidimensional Leadership Questionnaire to determine the style of leadership has both a high validity and reliability, because it has been frequently tested and validated by scholars to assess style of leadership. Proof of validity and reliability is provided by Hartog et al (1997) and improved to the smaller X5 version. In the last 15 years, 354 articles in multiple languages were the MLQ X5 was used, are published and accessible via the Library of the University of Amsterdam (2016). The acquiescence bias is reduced by the sequence of the questions for the three different types of style of leadership, these are mixed.

3

26 The reliability and validity of the complementary questions to determine change readiness is acceptable. The survey originally contained 25 questions, which the researchers have tested and retested for reliability and validity (Holt, Armenakis, Feild, & Harris, 2007, p. 248). This turned out to be high.

With the deduction of the amount of questions, selected by an eigenvalue of .8 or higher, the validity stays high. The reliability will be high because the questions are straightforward and not very prone to other ways of interpreting the questions, tested by Holt, Armenkis, Field and Harris (2007, p. 241).

This study will use a probability of .95 to deduct statistically valid conclusions, in custom with Behavioural Science Studies.

27

ANALYSIS OF THE RESULTS

Preparing the data of change readiness

For the data from the change readiness survey, the same steps are followed. First, the data was checked for missing responses and errors. There were neither errors nor missing data in the set of 297 respondents. The counter indicative questions, number 6.10 and 6.11 are recoded. Now the new variable change readiness can be created from the means of the questions 6.1 to 6.9 and recoded questions 6.10 and 6.11.

Normality check

Then, the normality check is done by using descriptive statistics. Change readiness has a normal distribution for Skewness (-.061) as well as Kurtosis (-.263) so no recoding is needed.

Reliability check

The Cronbach’s Alpha for change readiness is ,863 with n=11. Recoded question 6.11 has a corrected Item-Total Correlations were lower than ,30 so this item is deleted. After this, the Cronbach’s Alpha for change readiness is ,877 with n=10 with all corrected Item-Total Correlations bigger than ,30 and no items that would cause a positive difference in the Cronbach’s Alpha bigger than ,10 if that item was deleted, so these are the items that will make up the change readiness variable.

Correlations

The correlations between the variables can be tested after the new item is calculated. The variables gender, age, educational level, Business Unit, and tenure are included in the correlation analysis.

28 The correlation shows whether and if so, how strong pairs of variables are related, to better understand the data and deduct from conclusions that are caused by correlated effects.

Table 2: Means, Standard Deviations and Correlations

The correlations for age and tenure (,558) indicate also in this dataset a high positive relation. As with the first dataset, this multicollinearity is expected with age and tenure, because you need a certain age to reach certain tenure, where having a longer tenure than age is impossible. With the testing of the hypotheses about age and tenure the zoom will be on the difference between the two.

Preparing the data of the MLQ

First, the data was checked for missing responses and errors. There were no errors in the set, but there were multiple missing answers on multiple rows. Of the 532 respondents, there were 40 cases were one or more answers were missing. The choice is made to exclude all cases with one or more answers missing, so exclude the cases listwise. This way, only cases without any missing data in any variable will be included. It may reduce the effective sample size and bower, next to the chance of introducing bias into estimates, but with the amount of valid responses (492) this risk is accepted as minor.

Variable Mean SD 1. 2. 3. 4. 5. 6. 1. Age 45,515 9,739 2. Gender 1,212 0,410 -,181** 3. Educational Level 5,949 1,059 -,242** ,071 4. Business Unit 6,030 3,926 -,145* ,282** ,168** 5. Tenure 10,606 10,348 ,558** -,182** -,216** -,236** 6. Change Readiness 4,668 0,970 -,038 ,100 -,043 ,054 -,085 (,877)

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). c. Listwise N=297

29 Of the 34 questions of the MLQ specifically designed to inquire the style of leadership, 18 questions where about transformational leadership, 9 questions were about rational-objective leadership and the 7 questions about passive leadership4. From these questions, the three new variables were created for the styles of leadership. These were built up from the mean of the questions that were assigned to each of the styles of leadership from the improved design of the Multidimensional Leadership Questionnaire by Hartog, Muijen and Koopman (1997).

Normality check

Then, the normality check is done by using descriptive statistics. The styles of leadership, transformational, rational-objective and passive, have a normal distribution for Skewness (-,426,- ,254, -,355) as well as Kurtosis (-,182, -,199, -,246). So recoding is not necessary.

Reliability checks

Next step is the analysis of the reliability of the items, to see if the newly created variables -three styles of leadership; transformational (n=18), rational-objective (n=9) and passive (n=7)- are reliable. The Cronbach’s Alpha has to be at least ,70 to be reliable (Bland & Altman, 1997) so the data proves consistent findings.

The Cronbach’s Alpha for transformational leadership is ,935 with n=18. All corrected Item-Total Correlations were bigger than ,30 except question 9.3 so this question was excluded. After deletion, the Cronbach’s Alpha is ,940 with n=17. No items would cause a positive

4

30 difference in the Cronbach’s Alpha bigger than ,10 if that item was deleted, so these are the items that will make up the transformational leadership style variable.

The Cronbach’s Alpha for rational-objective leadership is ,664 with n=9. The questions 7.10 with a corrected Item-Total Correlations of ,153 and 9.1 with ,242 were excluded. This led to a Cronbach’s Alpha of ,691 with n = 7 and question 8.6 now shows a corrected Item-Total Correlations of ,258 so this item is deleted. Now the Cronbach’s Alpha is ,694 with n=6 and all corrected Item-Total Correlations were bigger than ,30 and no items would cause a positive difference in the Cronbach’s Alpha bigger than ,10 if that item was deleted. This means that this variable is unreliable, because it did not meet the reliability threshold of ,70 on the Cronbach’s Alpha.

The Cronbach’s Alpha for passive is ,728 with n=7 and all corrected Item-Total Correlations were bigger than ,30 except question 7.9, so this item was excluded. Now the Cronbach’s Alpha is ,752 with n=6 with all corrected Item-Total Correlations bigger than ,30 and no items would cause a positive difference in the Cronbach’s Alpha bigger than ,10 if that item was deleted, so these are the items that will make up the passive leadership style variable.

Correlations

The correlations between the variables can be tested after the new items are calculated. Now, the control variables gender, age, educational level Business Unit, and tenure are included in the correlation analysis. The correlation shows whether and if so, ow strong pairs of variables are related, to better understand the data and deduct from conclusions that are caused by correlated effects. The rational-objective style of leadership is excluded because the reliability check did not meet de boundary of ,7. The data is calculated with the listwise exclusion of missing values, which leaves a number of n=492 cases.

31

Table 3: Means, Standard Deviations and Correlations

Age correlates high with tenure and the passive style of leadership correlates high with the transformative style of leadership. High is when the Pearson r is bigger than ,5 but smaller than 1 (Pajic, 2015). The correlation for age and tenure (,563) is significant and positive. The multicollinearity is expected with age and tenure, because you need a certain age to reach a certain tenure, where having a longer tenure than age is impossible. This relation is visible though the positive correlation of .563, which means that if age grows with one year, the tenure will grow with .563.

The transformational style of leadership and passive style of leadership (-,650) indicate a high negative relationship, which is logical. If a leader has a style of leadership which tends to be more transformational, there is less room to use passive style of leadership, because a leader can use both styles but not at the same moment. If one is more used, this cannibalises the other and this is visible by the negative correlation of -.650.

The dominant style of leadership is added by examining per respondent which style scored the highest. If there is a tie between two or three styles of leadership, no dominant type is added with the abbreviations of the styles that are tied.

Variable Mean SD 1. 2. 3. 4. 5. 6. 7. 1. Age 44,874 9,604 2. Gender 1,205 0,404 _-,140** 3. Educational Level 5,870 1,087 -,229** -,004 4. Business Unit 5,823 3,967 -,112* ,162** ,190** 5. Tenure 9,877 10,387 ,563** -,137** -,286** -,212** 6. Passive Style 2,390 0,711 ,085 -,058 ,012 ,002 ,118** (,752) 7. Transformative 3,363 0,722 -,147** ,112* -,026 ,007 -,126** -,650** (,940)

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). c. Listwise N=492

32

_

_

_

_

_

cr_

+ cr_

+ cr_

+ cr_

+ cr_

46 Female IT WO 7

5

Mean change readiness

MLQ respondent x

=

To take the step towards hypothesis testing, the data from the change readiness survey has to be combined with the data of the Multidimensional Leadership Questionnaire. Because both surveys are filled in completely anonymously, there is no certain way to match the respondents from the first survey with the second survey. Other options to combine the two datasets have to be explored. The option to use propensity score matching was researched. This method achieves balance on observed covariates through careful matching on a single score (Thoemmes, 2012). The method uses a binary indicator, which makes it unsuitable on the data in this research.

Another option is to enrich the data of the MLQ with the data of the change readiness survey. To get an as close as possible match, this can be used by combining the variable from the MLQ with the change readiness mean of that variable. So if respondent x of the MLQ is a female of 46 years old and working in the IT department with her academic educational level has a tenure of 7 years, the mean change readiness of respondents from the change readiness survey with a tenure of 7 years is added to the MLQ data, as well as the mean change readiness of the respondents with the academic educational level, as well as the mean change readiness of the respondents working in the IT-department, as well as the mean change readiness of the female respondents and as well as the mean change readiness of the 46 year old respondents. In formula:

33 So the mean of each value for every variable from the change readiness survey is calculated and then added to the respondent in the MLQ. All these change readiness means together and divided by 5 give the weighed mean for that respondent. With this method the change readiness best fitting with the different characteristics of a respondent is added to the data of the MLQ. The respondent rate on both questionnaires makes this an acceptable method (Baruch & Holtom, 2008).

Third option is to match the data-sets based on matching. If all variables from a respondent of the MLQ fit all the variables of a respondent of the change readiness survey, the data is matched. This method shall reduce the amount of usable data, but the data that is left will have a higher probability that the data that is combined, belongs to the same respondent.

Combining the data by method of means

First, with Pivot-tables in Excel the change readiness means of all the values in the variables are calculated. These are added in the appendix5. Some change readiness means were not present for each value in a variable. This means that not all ages, tenures, educational levels or Business Units which are in the MLQ-dataset are present in the change readiness dataset. To match each change readiness mean for each value of a variable present in the MLQ, the need to compute a change readiness mean for each value of every variable is present. This is solved for each variable in a fitting manner:

5

34 CR-Age: Because the year above and under the missing age are represented for the ages 63 and 24, the mean of those surrounding years are used as value. For 21, the same value as for the youngest age group 23 is used.

CR-EdLev: The educational level of VWO is not represented in the CR-data. VWO is a middle school level like HAVO is, but it takes 6 years to finish instead of 5. But because there is no further study choice involved like with an MBO, the mean of HAVO is used here.

CR-BU; There are three Business Units not represented in the CR-survey; Treasury, Public Affairs and Corporate Regulation. Here is looked at the nature of the Business Units. Treasury is a financial Business Unit which demands the same competences from their employees as Corporate Control/Tax. The hierarchical lines and layers are similar and the Business Units both report to the CFO. So the mean value of that Business Unit is taken for Treasury. Almost the same is the situation with Corporate Regulation and Legal Affairs; these Business Units are even more entwined due to the nature of their work. With just one layer and both report to the CEO, the value of Legal Affairs is used for Corporate Control.

For Public Affairs the value of Communications is copied. Although communications has an internal and external part and Public Affairs are exclusively external oriented, these Business Units are the most comparable within the organisation. The competences are the same, the nature of the work is comparable, and they are connected with all the other Business Units within the organisation. They both report directly to the CEO.

CR-Tenure: For tenure, the missing value of 38 is calculated from the mean CR of the values of the year 37 and 39. For the missing value of 45, the value of the highest tenure (43) is used. Now the data from the MQL is enriched with the new variables by combining the five change readiness variables for each value of age, gender, tenure, educational level and Business Unit with the same, correspondent value in the MLQ. With this, the change readiness of each

35 different moderating variable is coupled with the respective variables of each respondent. An overall image will be formed from the mean of all change readiness variables together, which is the ultimate variable for change readiness for each respondent. This is not as powerful as with one dataset, because the perceptions of individuals who regard style of leadership in a certain way can regard change readiness differently, although it is proven that age and tenure have no influence on this (Epitropaki & Martin, 2004). Yet this method is acceptable, because of the amount of respondents in both samples and so the risk of variance and Skewness is reduced (Tabachnick & Fidell, 2001, p. 75) and with the means of every variable the most fitting overall change readiness is calculated.

Second, all the means for each value of the variable from the change readiness survey calculated in the pivot tables are combined with their mutual variables from the MLQ. This results in 5 different change readiness values per respondent in the MLQ-data. From these 5 different change readiness values comes a mean change readiness value that is best fitting for each respondent6.

Combining the data by method of matching

The data from the change readiness survey is the smallest set, so this set will be the basis for the matching. Then the set variables per respondent from this change readiness survey is manually matched with the set variables from the MLQ. If the data is not exact the same, the row is discarded. From this a set of 22 respondents remain7.

6 _{The tables with this result are provided in appendix F.} 7

36 The hypotheses will be tested with both combined datasets, so with the data combined by method of means as well as the data combined by the method of matching. Reason for this testing, is to give the results more robustness.

37

HYPOTHESIS TESTING

Style of leadership

Hypothesis 1 evolves around the most dominant style of leadership. To assess this, the dominant value in style of each row is labelled and the frequency is counted. The 40 respondents that did not complete the survey are kept empty (=”leeg”). The pie-chart makes it in one glance visual that the transformative style of leadership is the most occurring dominant style of leadership, followed by the passive style and then the rational-objective style.

Table 4: Dominant style of leadership Chart 1: Distribution of dominant style

From the data combined by means, with a total of n = 492 of completed cases, the transformational style of leadership is most dominant in the cases with n = 323, or 65,65%. The passive style of leadership is dominant in n = 95 cases or 19,31%. The rational-objective style of leadership is dominant in n = 55 cases or 11,18% of the total cases, but this style of leadership has a reliability that did not meet the threshold of the Cronbach’s Alpha of ,7 and is thus rejected. In n=19 of the completed cases no dominant style came forth. This is 3,86% of the total of completed cases.

Dominant style # of style

No dominant (P/R) 9 No dominant (P/R/T) 3 No dominant (P/T) 3 No dominant (R/T) 4 Passive 95 Rational Objective 55 Transformative 323 (leeg) 40 Total 532

38 In the data combined by matching, with a total of n=22 cases, the transformational style of leadership is also the most dominant with n=15 or 68,18%. The passive style of leadership is dominant in n=7 cases or 31,82%. The rational-objective style of leadership is missing in this dataset.

The hypothesis tested on the dataset combined by means proved that all the styles are represented within the monopolist organisation, but rational-objective style of leadership is not the most dominant style of leadership, but the transformative style of leadership is. And the test with the dataset combined by matching contains only the transformative and the passive style of leadership, where the transformative is the most dominant. Thus hypothesis 1: The three styles are all represented in the monopolist organisation and the dominant style of