Leadership Support and Personality Characteristics: do they moderate the influence of Participation on gaining Commitment to a change process?

Leadership Support and Personality Characteristics: do they

moderate the influence of Participation on gaining

Commitment to a change process?

Research performed at Philips during a change process to improve the market orientation

Master thesis, MSC BA, specialization Change Management University of Groningen, Faculty of Economics and Business

February 2012 Patricia Woudstra Student number: 1618555 Dennenstraat 43 8924 CG Leeuwarden Tel.: +31 (0)6-46592919 E-mail: patriciawoudstra@hotmail.com University supervisor: Dr. J.C.L. Paul Company supervisor: Lotte de Koning

PRC Personal Care, Philips, Drachten

Leadership Support and Personality Characteristics: do they moderate the

influence of Participation on gaining Commitment to a change process?

ABSTRACT

Purpose – This research seeks to explore whether employee’s personality characteristics, composed out of Need for Achievement and Need for Independence, and a supportive leadership style moderate the relation between participation and commitment. And if a moderating relationship exists, in what way and to what extent these factors moderate the relation between participation and commitment. Design/methodology – A questionnaire is designed to statistically examine the effects of personality characteristics and Leadership Support on the relation between participation and commitment. This questionnaire is distributed among employees of a private company undergoing a participative change program.

Findings – The results show that participation, as was found in literature, indeed has a positive influence on commitment, but that the predictors hypothesized to be moderators of this relation, have no

significant moderating effect. Leadership Support proved to have a mediating role instead of a moderating role, whereas Need for Achievement and Need for Independence have no significant relation with Commitment.

Research limitations/implications – Caution is warranted when generalizing the findings, as the sample was quite small and came from only one organization. To increase the generalizability of the results, further research should include multiple organizations. Although Leadership Support and Personality Characteristics are not moderators of the relation between participation and commitment, there could still be other factors that moderate this relation. Further research should investigate this. Further research should also focus on a valid and reliable scale for measuring Leadership Support and Need for Achievement.

Originality/value – This research adds knowledge to the field of organizational change. It sheds light on the complexity of the relationship between Participation and Commitment and investigates several possible moderating factors for this relationship.

Table of contents

1. INTRODUCTION 4

1.1 Theoretical Relevance 4

1.2 Management problem 4

1.3 Company History 5

1.4 Research objectives & questions 6

2. THEORATICAL FRAMEWORK 7

2.1 Organizational Change 7

2.2 Participation 7

2.3 Commitment 8

2.4 Relation participation – commitment 9

2.5 Identification of Moderating Factors 9

2.6 Conceptual Model 15

3. METHODOLOGY 16

3.1 Data Collection 16

3.2 Measures 16

3.3 The change project 17

3.4 The sample 18 3.5 Data analysis 18 3.6 Models 19 4. RESULTS 20 4.1 Preliminary analyses 20 4.2 Descriptive Statistics 21 4.3 Correlations 22 4.4 Regression analyses 23

5. CONCLUSION & DISCUSSION 30

5.1 Acceptation/rejection of hypotheses 30

5.2 Discussion 31

5.3 Implications for the organization 32

5.4 Limitations and implications for future research 33

REFERENCES 35

1.

INTRODUCTION

As organizations strive to remain as effective and efficient as possible, change initiatives are essential and ongoing. Given the accelerated rate and complexity of changes in the workplace, it is not surprising that there is a large and growing literature on the causes, consequences, and strategies of organizational change (for reviews, see Armenakis & Bedeian, 1999; Porras & Robertson, 1992). What is surprising, however, is the paucity of research on employee reactions to change. Judge, Thoresen, Pucik, and Welbourne (1999) noted that much of extant theory and research on organizational change takes a macro, or systems-oriented, approach. However, change efforts often stir up unforeseen challenges, and employee responses to changes ultimately determine whether the change efforts succeed or fail. Porras and Robertson (1992) suggest that for sustainable organizational change to occur, the behavior of organizational members has to change. And for their behavior to change, one must first alter individual cognitions as they regulate the organizational members’ behavior. Beer (1976) also states that the members of an organization are the key source of energy for organizational change processes, and he states that for this reason their commitment and involvement is the crucial factor for successful organizational change. From this it can be concluded that for change to occur, managers will first need to gain employee’s commitment and involvement to the change process.

1.2 Management problem

Gebhardt, Carpenter and Sherry (2006) found in their literature review that the concept of a market orientation has become increasingly important to the study and practice of management, and that it is recognized as significant for organizational success. Increasing competition, shortening product life cycles and the need for successful new products have put market orientation on the agenda of

innovating firms. Product innovation is a core process for creating superior customer value through new products (Day, 1994; Srivastava et al., 2001), by making sure the innovation process is market-oriented. This research is carried out at Philips – Personal Care, part of Royal Philips Electronics NV. Philips is a company that designs and manufactures electronics, to be more specific, the company focuses on innovative solutions in the areas of healthcare, lighting and consumer lifestyle. This study has been performed during a change process to improve the market orientation of the Innovation & Development process within Consumer Lifestyle - Personal Care.

The higher management of Philips found, by analyzing the situation and holding interviews with

managers in different functions that one of the problems that lies at the heart of the current situation is ‘Consumer centricity’: While ‘NPS as a metric’ has been adopted, it appears that a true ‘consumer centric’ attitude has not yet cascaded down through the organization.

Innovation and Development can contribute by making sure I&D is consumer-driven, by developing products that satisfy the consumer’s needs and wants and delight consumers better than competition. However, with the current way of working within I&D, products often do not align well enough with what consumers need. I&D will have to become more market oriented and proactive.

Management of Philips has stated that for true Consumer centricity to cascade down the organisation, people’s mind sets have to be changed. An important question that needs answering is then what factors help to change people’s mind sets.

The Theoretical relevance – section above (p.4) suggests that to change the mind sets of employees, as is the case with Philips, commitment to the change program is needed. The way they have chosen to gain this commitment is by participation: letting employees think along and have an input in the changes. Important to know in this respect is whether there are factors that influence the relationship between participation and commitment. Are there certain conditions that will decide whether

participation actually leads to commitment?

The aim of this research was to investigate what factors influence the relation between participation and commitment to the change, in order to contribute to a better understanding of how organizations can make change efforts more successful.

1.3 Company history

Philips was founded in 1891 in Eindhoven. The company started as a manufacturer of light bulbs. In 1914 Philips established a research laboratory to study physical and chemical phenomena and stimulate product innovation. In 1918 it began diversifying its product range and in 1939 it launched the first Philips electric shaver. Examples of inventions done by Philips from the 1940s to 2000 are the rotary heads that led to the development of the Philishave electric shaver, the Compact Audio Cassette, the Compact Disc and the DVD.

In the 1990s the company carried out a major restructuring program, simplifying its structure and reducing the number of business areas. Moving into the 21st century, Philips has continued to change and grow. Long aware that for many people it is no more than a consumer electronics producer, it has dedicated itself to projecting a new and more representative image that reflects the products it offers in the areas of Healthcare, Lifestyle and Technology.

The organizational structure was again simplified per January 1, 2008 by forming three sectors:

1.4 Research objectives & questions

The influence of participation on commitment has been discussed and proven in literature, but the factors that moderate this relationship are not extensively investigated and there is no clarity on what factors do and do not moderate the relationship.

This research pursues two objectives, a practical one and a theoretical one. The practical objective is to investigate the commitment to the change of the employees and to give advice how the management of PRC – Personal Care can potentially ensure participation will lead to commitment. When the right context is created for participation, the employees will be more willing to adopt a more proactive and consumer-centered mind set. The theoretical objective is to investigate whether certain factors moderate the relationship between participation and commitment to a change process.

The assumptions of this research are that commitment will help to successfully change the mind sets of employees, and that participation will, depending on some moderating factors, positively influence commitment.

This paper will investigate what factors moderate the relation between participation and commitment. To be specific, it will be investigated whether and how strong employee’s personalities and supportive leadership moderate the influence of participation on commitment

Reasons for focusing on these relationships:

1. Management wants to use participation as an intervention in the change program. This means it is important to know how participation relates to successfully changing people’s mind sets.

2. According to literature, participation plays a crucial role during a change project.

3. What factors moderate the relationship between participation and commitment has not been extensively investigated, while it can give very helpful insight with regard to in what situations to use participation and what factors to take into account when using participation.

4. In a practical sense, personal characteristics is chosen because this can show management whether participation is more or less helpful when their employees have certain preferences. And leadership support is chosen because management can directly influence this variable.

The research questions are formulated as follows:

- Do employee’s personality characteristics and leadership support moderate the relation between participation and commitment?

- If a moderating relationship exists, in what way and to what extent do employee’s personalities and supportive leadership moderate the relation between participation and commitment?

2.

THEORETICAL FRAMEWORK

Metselaar (1997) defines organizational change as “the planned modification of an organization’s structure or work and administrative processes, initiated by the organization’s top management, and which is aimed at improving the organization’s functioning” (Metselaar, 1997, p. 5). Change is a phenomenon that individuals and organizations face on a daily basis, and because of this, questions about its nature and about responses to change continuously emerge. In addition, the difficulties associated with understanding change have contributed to many failed change efforts that can cost organizations a great deal of time, money, and other resources (Kotter, 1995).

In their review of organizational change theory and research, Armenakis and Bedeian (1999) identified five themes common to change efforts: (a) content issues, which deal with the substance and nature of a particular change; (b) contextual issues, which deal with forces and conditions existing in an

organization’s internal and external environments; (c) process issues, which deal with the actions taken in the implementation of an intended change; (d) criterion issues, which deal with outcomes assessed in organizational change efforts; and (e) affective and behavioral reactions to change. This study focused primarily on aspects of change that are captured by the themes process, contextual and affective and behavioral reactions. In this case the content of the change is improving the market orientation in an I&D process. The change process is characterized by participation, which is the independent variable in this literature. The affective and behavioral reaction is commitment to the change effort which is the mediator in this research. The internal contextual factors are moderators that are hypothesized to influence the relation between participation and commitment. At last, the criterion issue, the intended outcome, is a more market-oriented mindset among the employees.

The successfulness of change efforts is due to not only their content or substantive nature, but also the processes followed or actions undertaken during their implementation (e.g., Hendry, 1996). Content and process considerations must thus be viewed as complementary elements in planning and monitoring an organizational change (Armenakis & Bedeian, 1999). Lewin (1951), in his Three-Step model, divides the change process into three phases: unfreezing, moving and refreezing. Parallels that Armenakis and Bedeian (1999) noted between Lewin’s model and other models of change included: (a) unfreezing corresponds with readiness and resistance to change, (b) moving corresponds with adoption and exploration, and (c) refreezing corresponds with commitment and institutionalization.

2.2 Participation

participation of organizational members in the change process: The more they support and contribute in this process with their ideas and suggestions, the more likely it is that they will get involved in the change. Lewis, Schmisseur et al. (2006) found in their review that one of the general change implementation strategies promoted is the need for change agents to empower and encourage stakeholders to take ownership and participate in the change effort, particularly when it comes to evaluating and implementing various change processes. They argued that participation gives members a sense of control and reduces uncertainty about their changing circumstances, such that if people can feel a part of the implementation process, they will be more committed to the change and thus less resistant to its goals and objectives. Bouma & Emans (2005) define participation in organizational change as management involving the staff in decision making about content and form of the organizational change. This definition of participation is used in the current study.

2.3 Commitment

Many authors have argued that individual commitment to change is a necessary component in garnering support for planned organizational change efforts (Armenakis & Bedeian, 1999; Judge, Thoresen, Pucik & Welbourne, 1999; Klein & Sorra, 1996; Kotter, 1995). Armenakis and Bedeian (1999) emphasized commitment in their framework of successful organizational change, noting that commitment is clearly a relevant criterion variable to be considered in change implementation. Herscovitch and Meyer (2002) and Fedor et al. (2006) identified commitment to change as an important precursor to the behavioral intention to support change. Because of the importance of commitment demonstrated by these authors, it is important for managers to ensure the presence of this phenomenon.

Herscovitch and Meyer (2002) define commitment to change as “a force (mind-set) that binds an individual to a course of action deemed necessary for the successful implementation of a change initiative” According to Neubert & Wu (2009), commitment to change is the level of an employee’s attachment to the implementation of new work rules, policies, programs, budgets, technology, and so forth. What these definitions share, is the notion that commitment to change reflects some kind of attachment to and involvement in the change initiative (Jaros, 2010).

Different forms of commitment have been identified in the literature including, for example, affective, continuance and normative commitment (Meyer and Allen, 1991). The focus of this study is on the more affective aspects of commitment. The definition of affective commitment to change used in this study is a desire to provide support for the change based on a belief in its inherent benefits (Herscovitch and Meyer, 2002).

2.4 Relation participation – commitment

Herscovitch and Meyer (2002) consider that many of the most widely recommended strategies for implementing change such as training, participation and empowerment are likely to impact on support for, and identification with, the change and thus promote affective commitment to change. Peccei, Giangreco and Sebastiano (2011) argue that involvement in the change process helps to reduce resistance to change by reducing anxiety, creating a stronger sense of ownership of the change, and enabling individuals actively to contribute to the shaping of the change. More generally, as argued by Giangreco and Peccei (2005), because of the above reasons, involvement in the change process can generate a more positive attitude towards the change. Lewin (1948, 1951) was the first one to discover the importance of participation in changing attitudes. His group decision experiments stressed the role of group consensus in getting members to commit themselves to subsequent actions. Meyer and Allen (1990) also found that participation is an antecedent of affective commitment. Their study shows that participation had a positive significant correlation of .65 with affective commitment. Herscovitch and Meyer (2002) also found a positive relationship between involvement with the implementation of a change effort and affective commitment to a change.

Foster (2010) states that fair practices in the implementation of a change have significant associations with employee commitment to change. He explains that managers and supervisors can promote fair implementation processes in many ways. One way he mentions is by allowing each individual who is involved to have influence in the process as well as the opportunity to appeal decisions. In other words, to ensure commitment, all employees who are affected should be involved in decision making about the change: they should participate.

2.5 Identification of Moderating Factors

In the search for moderating factors, the author broadened the search field by searching for factors that might moderate the relation between participation and some positive outcome. In this research, it will be investigated whether the moderating factors that are found influence the causal relation between participation and commitment.

Pasmore and Fagans (1992) state that simply involving people in decision making will not automatically produce positive benefits to either those involved or the organization as a whole. They argue that currently popular interventions, such as self-directed work teams and quality programs, continue to fall short of their full potential in many instances due to a failure to recognize the complexity of the

found that participation led to lower productivity. A plurality (49%) found that the results either were the same in both cases or depended on certain contingencies (subordinate knowledge, for instance). Their conclusion is that sometimes participation is useful and sometimes it is not, and that we are now beginning to understand some of the conditions under which it will be effective. From the work of the above authors it can be concluded that there are factors that moderate the relation between

participation and positive outcomes for the change process. A moderator is a variable that affects the direction and/or strength of the relation between an independent variable and a dependent variable (Baron and Kenny, 1986)

Pasmore and Fagans (1992) further state that many interventions are affected by the relative abilities of individuals to engage in participative activities. For this reason, more attention should be focused on preparing people to participate in change. These authors believe there are three variables that influence the effectiveness of participation in organization change. The first is organizational receptivity, which they define as the readiness of the organization to support authentic participation. The second is individual ego development, which deals with to what extent people are psychologically prepared to participate. The third variable is knowledge availability, which they define as a person’s knowledge pertinent to the decision under consideration. These three factors affect a person’s ability and inclination to participate in a given situation. In figure 1, their ideas are shown in a model.

2.5.1 Organizational support

The factor that Pasmore and Fagans (1992) call organizational receptivity has also been described by other authors. Argyris (1964) for example states that traditionally designed organizations conflict with the needs of adults. Individuals want to be active, independent and strive for equality. To close the gap between individual needs and organizational experiences, Argyris advocates changing the structure of the organization and increasing opportunities for meaningful participation. Likert (1967), in his proposals for “systems 4” organizational design and effective group decision making, offered the first direct link between organization design and the management of participative processes. The “systems 4” organization is based on the principle of supportive relationships, which essentially says that the leadership and other processes of the organization must be such that they ensure a maximum

probability that in all interactions and relationships within the organization, each member will view the experience as supportive. Like Argyris and Likert, McGregor (1960) maintained strongly that managerial beliefs and choices influence the performance of their subordinates. Neumann (1989) argues that certain situational factors influence choices individuals make regarding their level of involvement. Two deterrents to participation she proposes are: 1. Structural (organizational design, work design, HR Figure 1: Pasmore & Fagans, three variables that influence the effectiveness of participation in organization change.

Organizational receptivity Individual ego development Knowledge availability

management policies) and 2. Relational (how participation is managed, the dynamics of hierarchy). Zammuto and O’Connor (1992) state that organizational climates with flexible and supportive structures are conducive to establishing a positive attitude toward change. Wagner and Gooding (1987) note that participation programs in sociotechnical system interventions demonstrated positive results because other changes in compensation and job design accompanied and reinforced the introduction of

participation. So participation should be supported by other organizational changes if improvements in outcomes are to be achieved. The thinking of these authors is similar to Pasmore and Fagans: The more supportive the organization is with regard to participation of its employees, the bigger the effect on performance will be. These ideas are summarized in figure 2.

2.5.2 Employee’s readiness for participation

Vroom (1960) was the first to propose that personality might moderate the effects of participation on satisfaction and productivity. Specifically, he suggested that participation will positively influence only employees having personalities with low authoritarianism and high needs for independence. Vroom found some support for his hypotheses. However, further studies have provided mixed support (Miller and Monge, 1986). He also argued that participation was most beneficial when combined with programs designed to increase individuals’ skills and abilities. Locke, Schweiger and Latham(1986) name among the contextual factors that determine the effectiveness of participation: knowledge of participants (having the right knowledge or ability) and motivation (need for independence, need for achievement, job involvement, and commitment to organizational goals). Singer (1974) commented that to assume that all workers desire participation opportunities is to lack sensitivity to individual needs: many workers may not value participation to the extent that academicians do. According to Pasmore and Fagans (1992), participation in complex decisions is difficult if one lacks relevant information or knowledge. They derive this from Hackman and Oldham (1976), who include the skills and knowledge pertinent to the work to be performed as one of the attributes of individuals that moderate the effect of enriched job design on motivation. So it seems that personal variables and the availability of knowledge and skills determine whether employees value participation, and whether participation leads to a positive

outcome. Miller and Monge, in their meta-analysis, state that it was not possible to provide a test of contingency predictions referring to personality, because very few studies provided subgroup analyses considering individuals with different personality types. And studies that have considered

authoritarianism and need for independence have provided conflicting conclusions, meaning that further research into these factors is needed. These ideas are summarized below, in figure 3. Figure 2: Elements and moderating role of organizational support.

Organizational support - Supportive leadership

- supportive structure & processes

2.5.3 Organizational context

Miller and Monge (1986) state that contingency models of participation are based on the belief that no single model of participation is appropriate for all employees in all organizations: it is dependent on the individual and the context. Singer (1974) predicted that participation may only be effective for

employees in certain types of organizations—such as research or service organizations, rather than manufacturing organizations—or only for middle- or upper-level employees. Employees who value participation will be the most positively influenced by it, and this is likely to differ between individuals. For example, higher-level employees might value participation more than lower-level employees, or individuals working in research or service industries might value participation more than manufacturing employees. Locke and Schweiger (1986) name among the contextual factors that determine the

effectiveness of participation: organizational factors (all factors external to the participating employee), task attributes (routine versus non-routine), leader attributes (skill in the use of participative techniques, leader personality, level of interaction with employees), and other organizational factors (time available, size, rate of change). Miller and Monge, in their meta-analysis, found that the possible moderator variables organizational type, job level, and type of decision were not useful in reducing variance in outcomes or in differentiating among effect sizes. So none of these three sub groupings seemed to moderate the relation between participation and positive outcomes. These authors comment that though there seemed little support for contingency models of participation, the lack of measures for several contingency variables could have affected findings. Also, their meta-analysis did not allow for a complete test of the models presented, as they lacked data on several intervening variables in these models, such as upward and downward sharing of information. They encourage researchers to measure these variables in future investigations of participation. They state researchers should extend their consideration of contingency variables. Figure 4 summarizes these findings.

Figure 3: Elements and moderating role of readiness for participation

Readiness for participation Personal factors

- Authoritarianism - Need for independence - Need for achievement - Job involvement

- Commitment to organizational goals

Relevant skills and knowledge - programs to increase skills and abilities

- having the right knowledge or ability - availability relevant information/ knowledge

2.5.4 Synthesis

Research up till now has mentioned a lot of factors that affect the relation between participation and some positive result. From these many possibly moderating factors, research to date has given

conflicting results or has not tested them properly at all. Research has also never investigated whether these factors moderate the specific relation between participation in a change program and affective commitment to this change program.

In this research, the positive result I will focus on is commitment to the change. In this research it is assumed that this variable will help to achieve a more market oriented mind set, as the eventual goal is successful change in people’s mind sets. The independent variable is participation. The moderators that affect this relation are divided into three categories: organizational support, readiness for participation, and contextual factors. This study investigates the moderating role of one factor of organizational support: Leadership Support, and of one factor of readiness for participation: personality characteristics (Need for Achievement and Need for Independence). These moderators are chosen because it can be reasonably expected that there will be differences between people with regard to Need for

Achievement and Need for Independence, and differences with regard to leadership support between the three PRC departments within Philips that take part in this research. Also, it is believed that it is relatively easy for a manager to influence the concept of Leadership Support. They can improve the amount of support if this is needed. It is not easy for management to influence employee’s personality characteristics, but this moderator can give information on when it is useful to encourage participation in a change program. Investigating the moderating role of Need for Achievement and Need for

Independence can show whether it is always positive to use participation, or whether it only has a significant positive effect when people have a high Need for Achievement/ Need for Independence.

Leadership Support

Burnes (2009) describes the role of leadership as establishing a direction (vision & strategy), aligning people (getting people to understand and believe the vision), and motivating and inspiring. Antoni (2004) defines the role of leadership as encouraging participation, Kanter (1985) stresses the

importance of making standards and requirements clear: tell exactly what is expected of people in the change. At last, Osarenkhoe (2008) found that it is important that management sets the examples, to show everybody that you mean what you say. These authors argue that employees need support and inspiration during change, and that leadership should provide this.

Figure 4: Elements and moderating role of contextual factors Contextual factors

- Job level

- Type of problem (routine vs. complex) - Type of organization

- Leader attributes

- Change characteristics (time available, size, rate of change)

The assumption that leadership has impact on how employees view change initiatives is anything but new. Kotter (1995) states that leadership is essential for the successful implementation of any change initiative. According to Devos, Vanderheyden & Van den Broeck (2002), management with a supportive leadership style during a change process enables employees to encounter new challenges. On the other hand, if managers do not support the change initiative actively, employees will consequently not take it serious (Devos et al., 2002).

In literature, two types of leadership can be distinguished: transformational and transactional leadership. Transformational leadership is about inspiring followers to do more than originally expected and includes the following behavioral dimensions: individualized consideration; intellectual stimulation; idealized influence; and inspirational motivation. Transactional leadership involves an exchange process to motivate follower compliance. The concept of transformational leadership corresponds with behaviors that support organizational change, since it includes behaviors as providing vision and meaning and giving support, encouragement and coaching.

In this research, supportive leadership includes the behavioral dimensions of transformational leadership and one dimension of transactional leadership: provision of feedback.

Hypothesis 1: The more supportive the leadership, the stronger is the positive influence of participation on commitment

Personality Characteristics

Heckert et al. (2000) define the need for achievement as the desire to excel, and state it involves attempting to improve on past performance. They define the need for autonomy as the desire for self, rather than other, direction.

Spector (1985) found that the strength of higher order needs moderates the relation between job scope and employee outcomes. The results of his meta-analysis support the idea that employees with high levels of these needs will be most likely to demonstrate the hypothesized relationship between scope and outcomes. He also found that low need strength individuals are not adversely affected or

unaffected by job scope. Both high and low need strength individuals respond favorably to high scope jobs, higher-order need strength only affected the magnitude and not the direction of the

relationship.

From the above it is eminent that Need for Achievement and Need for Autonomy are two higher order needs that may influence individuals' affective reactions to characteristics of their jobs. From this, it can be concluded that these needs may also influence individuals’ affective reactions to characteristics of a change program.

The assumption in this research is that persons high and low in Need for Achievement and Need for Independence possess unlike personality characteristics, which results in them responding differently to certain participation in a change program.

Need for Achievement

feel empowered than individuals with low achievement need, especially when the workers are allowed to participate in setting performance goals. Their study shows that personal needs and participation in goal-setting reinforce each other.

Orpen (1985) explains that high need achievers are stimulated by tasks that are inherently meaningful and also provide both ample knowledge of results and opportunities for independent action and thought. From the above, It can be expected that persons with high Need for Achievement will react more positively to participation because these people really want to make a difference. They want to help and think along to improve performance. On the other hand, persons with low Need for

Achievement may not believe that they have much control over outcomes anyway, and so may not feel as positive about participation.

Hypothesis 2: The higher a person’s need for achievement, the stronger is the positive influence of participation on commitment

Need for Independence

From the work of Maslow (1943) and Vroom (1964) it can be found that highly independent persons are motivated to put forth a lot of effort, when they are given tasks to perform that give them scope for the exercise of this need.

Taking the above into the context of participation, you will probably not get the commitment of these persons when you tell them how to change without them having a say in this. It would seem logical that letting people with high Need for Independence participate in deciding how to bring about the change and so appealing to their desire for self direction, would lead to higher commitment. On the other hand, persons with low Need for Independence are persons that value someone telling them what to do. They may dislike having to make decisions by themselves, and this may lead to lower commitment.

Hypothesis 3: The higher a person’s need for independence, the stronger is the positive influence of participation on commitment

2.6 Conceptual model

The conceptual model that will be tested in this research is shown below in figure 5.

Participation

Commitment

- Leadership Support

- Personality Characteristics

- Need for Independence

+

+

3.

METHODOLOGY

The research consists of three interrelated steps. First a desk research was done, in which the concepts of participation, commitment and its possible moderators are defined, relationships between these variables are described and measures for the variables are identified. For the desk research the change literature was analyzed using electronic databases ((Business Source Premier, Google Scholar, Academic Search Premier and databases on psychology). Second, explorative observations, interviews and reading of documents has been done to understand the situation at the case company. This way, the change they face and the management problem are understood and the compatibility of the research with the specific case has been ensured. The third step deals with answering the research questions formulated in chapter 2. Collecting data to do so was done by means of empirical research in the form of a

questionnaire. This questionnaire and the accompanying letter can be found in Appendix 1.This form of quantitative research helps to find out whether the above-stated moderators indeed affect the relation between participation and commitment, and if so, in what way.

3.2 Measures

To measure the amount of commitment, participation, need for achievement and need for

independence, already existing measurement scales were used. This ensures compatibility with earlier studies and has the advantage that the reliability of the scales has already been proven to be sufficient. For the moderator leadership support there was no fitting scale. A new scale has been constructed by using parts of different existing scales that fit the operationalization of supportive leadership in this research.

To measure commitment to the change, the Herscovitch and Meyer (2002) scale is used as it is the only scale the researcher was aware of specifically designed to measure commitment to change. This 18-item scale measures three dimensions of commitment to change, including affective, continuance, and normative. For this research, only the 6 items that measure affective commitment were used. Originally, a 7-point Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree) was used to measure these items (Herscovitch & Meyer, 2002). Because the other variables in this study use 5-point Likert scales, affective commitment will also be measured using a 5-point scale for the sake of keeping

conformity in the questionnaire. The affective commitment to change subscale measures an individual’s desire to be committed to a change. This is a psychological conceptualization that refers to an

individual’s emotional attachment or identification with the target of the commitment (Meyer & Allen, 1991). The validity of this measure has already been proven (Herscovitch and Meyer, 2002).

Participation measurement is done by use of the measurement scale formulated by Bouma (2009). This participation scale captures the extent to which employees participated in the change process.

Dutch): have been passed (1), have been informed (2), have been consulted (3), have been involved (4), being a decision maker (5).

To measure need for achievement and need for independence the scales of the Manifest Needs Questionnaire, developed by Steers and Braunstein (1976) to assess the manifest levels of these needs among persons in work contexts, was used. This scale has been specially developed for work contexts, and has already proven its usability. The responses were originally recorded on seven-point Likert scales with the following response categories: always, almost always, usually, sometimes, seldom, almost never, and never. However, for this research the scale has been re-coded into a 5-point Likert scale, ranging from (1) never to (5) always, to keep uniformity in the questionnaire.

For measuring Leadership Support, some questions from the measurement scale of Bouckenooghe et al. (2009) are used and some questions from the scale of Podsakoff et al. (1990). The questions from Bouckenooghe come from their scale that measures management’s ability to lead change. Podsakoff measures leadership on two dimensions: transformational leadership and transactional leadership. They measure transformational leadership by the use of 22 items, the present study uses only two items. From the six items with the highest loading in the factor analysis (Podsakoff et al., 1990: 118‐119), two were chosen that best fit the goal of this research. Transactional leadership is measured using one of the original five items of Podsakoff et al. (1990:118‐119), this one is among the factors with the highest loading in the factor analysis ( ‘I believe my leader always gives me positive feedback when I perform well’), and best fits the goal of this research. Again a 5 point Likert scale was used, ranging from (1) Strongly Disagree to (5) Strongly Agree.

To determine whether the statements could be combined into one concept, a factor analysis and a reliability analysis were performed. When the total variance percentage and Cronbach’s Alpha were satisfactory, the items could be joined.

3.3 The change project

Within the I&D departments of Philips a change program called Accelerate! has been set up. This program is founded on the principle that to satisfy today’s demanding consumers, Philips has to focus on the needs of those consumers and connect these insights to their innovation process. This change program is based on three pillars: Speed, Open Innovation and Consumer Centricity.

Consumer Centricity is defined within Philips as ‘working on a product that delivers a key benefit to customers, better than competitors’. The goal of ‘Consumer Centricity’ is to reduce the reliance on technical competence and previous own products and increase the reliance on consumer needs, wants and delight factors and competitor products. The departments that have to drive Consumer centricity are the Product Research Centers (PRC). The PRC represents the Voice of the Consumer in I&D. It is their responsibility to find out what consumers want and what competitors are making, and then how to excel.

this information, they need to be effective at disseminating market information across business functions and to be able to translate market information into technical requirements.

3.4 The sample

For collecting the needed data, questionnaires were distributed among three Product Research Departments of Royal Philips Electronics NV. The questionnaires were made and distributed among employees by using a program the company uses to administer online questionnaires to respondents. In total, the three departments had about 70 – 75 respondents that I asked to participate in the

questionnaire. The amount of filled in questionnaires I received was 61, and after eliminating

questionnaires that were not filled in completely, I had 51 respondents left. This yields a response rate of about 70 %.

3.5 Data Analysis

To investigate whether the aforementioned hypotheses are supported by statistical evidence, SPSS 16.0 was used to analyze the gathered data. To test the hypotheses an alpha level of 0,05 was used. The following steps were taken during the data analysis in order to answer the research questions: - The variables Commitment, Participation, Leadership Support, Need for Achievement and Need for Independence were measured using different statements in the questionnaire. The negatively formulated statements were inversely coded so that recoding is not needed.

- By checking the data in SPSS, questionnaires that were not filled in completely were eliminated, so that the data would not contain any missing values.

- A factor analysis and reliability analysis were done to see whether the items could be combined into the variables they were intended to represent. Detailed information of the factor and reliability analysis can be found in Appendix 4. After this, the items were summarized into the variables they represented, using the average of responses.

- A box-plot was then used to check the data for extreme values/outliers.

- In order to investigate whether the sum scales of the concepts are normally distributed, the

Kolmogorov – Smirnov test was performed. The findings of this test determine whether nonparametric or parametric tests will be used. The null-hypothesis of this test assumes that the distribution is normal. If the P-value is >= 0,05, the null hypothesis will not be rejected, and the null-hypothesis will be rejected when the P-value is < 0,05.

- After that, bi-variate correlation analysis is used to see whether there are relations between Participation, Commitment, Leadership Support and Personality Characteristics.

- To investigate the moderating effect of Leadership Support and Personality Characteristics, a linear regression analysis is done. For analyzing the effect of the moderators, three steps are taken: first, the independent and moderator variables have to be centralized by calculating the mean of each variable and then subtracting the mean from every measurement. Second, an interaction variable,

interaction path (path c in figure 6) is significant (Baron and Kenny, 1986). In this research, we keep with the convention that the P-value should be smaller than 0.05 in order to be able to reject the

null-hypothesis and so conclude that a factor delivers a significant contribution to the variance in the dependent variable.

Figure 6: Moderator model by Baron and Kenny (1986). 3.6 Models

To make the regression analyses that will be done more clear, the following models are made and will be tested in this order:

The first model is to test the main effect of Participation on Commitment: Model 1: C = β0 + β1P

These models are to test the three hypotheses: Hypothesis 1 Model 2: C = β0 + β1P + β2L + β3PL Hypothesis 2 Model 3: C = β0 + β1P + β4A + β5PA Hypothesis 3 Model 4: C = β0 + β1P + β6I + β7PI

See table 1 for linking the variables in the models to the factors under investigation. Variable Factor under investigation

P Participation L Leadership Support

PL Product variable of Participation and Leadership Support A Need for Achievement

PA Product variable of Participation and Need for Achievement I Need for Independence

PI Product variable of Participation and Need for Independence c b a Predictor Moderator Predictor X Moderator Outcome Variable

4.

RESULTS

In this section, the moderating effects of Leadership Support, Need for Achievement and Need for Independence are investigated. But before the moderator analysis can be done, preliminary analyses will be done to ensure the quality of the results, and descriptive statistics and correlations will be given to provide more elaborate insight into the data.

4.1 preliminary analyses

4.1.1 Validity and Reliability

A factor analysis was done on the items to ensure validity of the results, and the Cronbach’s Alpha was calculated to check for reliability. The factor analysis helps you to see whether the items (the questions asked) were recognized by the respondents as belonging to the same variables that you intended. In this research, five variables were distinguished. So we would also like the items to load on five factors, and each item to load on the correct factor. When executing the factor analysis in SPSS, eight factors were distinguished and the only items that all loaded on the same factor were those for Participation and Commitment. Then items that loaded on the wrong factor, did not load high enough on all the factors or loaded high on two factors were deleted stepwise until five factors were left that all represented one variable. This meant that the variable Leadership Support now contains just two items, the variable Need for Achievement also contains just two items, and the variable Need for Independence contains four items. For the items that were still present when only five factors were left, see Table 2. From now on a respondent’s score on these variables is the average of his scores on the items that remain. The total explained variance was 70,34 %, the variance explained by the individual factors can be found in the table below.

Variable Participation Commitment Leadership

Support Need for Achievement Need for Independence Items left in the variable

Purpose Value Encourage Improve performance Own boss Meaning Good strategy Feedback Perform better My own way

Work with Mistake Personal freedom

Goal Purpose Work alone

Optimizing Better without Design Necessary Explained

varaince

24,9% 17,54% 7,96% 7,79% 12,15%

Table 2: Items left after factor analysis

To check for reliability of the results, the Cronbach’s Alpha was calculated for each of the variables. The Cronbach’s Alpha for Commitment, Participation and Need for Independence are quite high, which means the outcomes based on these variables are reliable. However, the Cronbach’s Alpha for

Variable Cronbach’s Alpha

Participation 0,945

Commitment 0,85

Leadership Support 0,559 Need for Achievement 0,536 Need for Independence 0,748 Table 3: Cronbach’s Alpha’s for all variables

4.1.2 Extreme Values

By using box plots, extreme values and/or outliers can be located. The concepts Participation, Need for Independence and Commitment had outliers, but none of the variables had extreme values (see Appendix 3 for the box plots). Adding to this that none of the respondents that had an outlier in one of the concepts also had an outlier in another concept, it is safe to say that these respondent’s extreme answers are probably not due to uncarefully reading of the statements or not paying attention to the answer scale, and therefore will be kept in the dataset.

4.1.3 Normal distribution

The Kolmogorov – Smirnov test was done and Q-Q Plots were made to investigate whether the sum scales of the concepts were normally distributed. Depending on the findings of the test and plots, parametric or non-parametric tests will be used for data analysis. As can be seen in the table below, the P-values were all > 0,05 except for participation and leadership support, so the outcomes of the

Kolmogorov – Smirnov test indicate that the distribution is normal. Additionally all Q-Q plots also demonstrate that the factors are normally distributed. The Q-Q Plots can be found in Appendix 4. As a consequence of the normal distribution, parametric tests will be used for further analysis of the concepts.

Kolmogorov – Smirnov test Kolmogorov – Smirnov Z Asymp. Sig. (2-tailed) Participation LeadershipSupport NeedforAchievement NeedforIndependence Commitment 1,690 0,642 1,147 0,817 1,256 0,007 0,015 0,220 0,543 0,085 Table 4: Outcomes of the Kolmogorov – Smirnov Test

4.2 Descriptive Statistics

Descriptive Statistics

N Minimum Maximum Mean Std. Deviation

commitment 51 3,00 5,00 4,2190 ,45276 participation 51 1,00 4,67 2,4542 ,95515 Leadershipsupport 51 2,50 5,00 4,0098 ,62841 NeedforAchievement 51 2,00 5,00 4,0000 ,74833 NeedforIndependence 51 1,00 4,00 2,7843 ,69466 Valid N (listwise) 51

Table 2: Average values and standard deviations of the variables

From the table it can be seen that Commitment to the change program is quite high among the respondents (mean score is 4.22), which means that the employees questioned in this research are committed to the change program their company is undergoing. They belief in the inherent benefits of the change program and want to support it. However, Participation scored quite low among the respondents (mean score is 2,45), which means that although the company has chosen for a participative change program, the employees do not experience high levels of participation. The next step is to check for correlations between the different items, to see if relationships exist between the items.

4.3 Correlations

First of all, the correlation between Participation and Commitment was calculated to check if the

assumption that participation in the change program contributes to commitment in the change program is indeed true. Then, the correlation of the moderators with Participation and Commitment was

calculated. It is best when the moderator variables are not significantly correlated to the independent and dependent variable, in order to provide a clearly interpretable interaction term. For tables showing all the correlations, see Appendix 6.

Participation – Commitment: With a positive correlation of 0.485 which is significant at the 0.005 level, there is a clear moderate-strength relationship. This was also expected as current literature states that participation contributes to commitment (Herscovitch and Meyer, 2002; Lewin, 1948; Giangreco and Peccei, 2005). This finding is proof for the assumption that when participation is higher, commitment to the change is also higher.

Commitment – Leadership Support: With a correlation of 0.297 which is significant at the 0,05 level, there is a positive, moderate-strength relation. This correlation could mean that Supportive Leadership has its own significant effect on Commitment.

Commitment – Need for Independence: With a correlation of 0.01 which is not significant, it can be assumed that there is no relation.

Participation – Leadership Support: With a correlation of 0.323 which is significant, there is a moderate-strength positive relation between these two factors. This means another relation apart from a possible moderating one might exist.

Participation – Need for Achievement: With a correlation of 0.324 which is significant, there is a moderate positive relation.

Participation – Need for Independence: With a correlation of 0.315 which is significant, there is a moderate positive relation.

The correlation between Participation and Leadership Support results in the presence of

multicollinearity when the effects of the independent variable and the moderator on the dependent variable are estimated. This results in reduced power in the test of the coefficients of the interaction term (product variable). To overcome this multicollinearity issue, the independent and moderator variables were first centralized. Centralization avoids the occurrence of high correlations between the interaction term on the one side and both predictors that make up the product on the other side

4.4 Regression Analyses

Two different regression analyses were done: simple and multiple regression analysis. The simple regression analysis to examine the causal relationship between the independent variable (participation) and the dependent variable (commitment). The multiple regression analysis is done to examine the moderator role of Leadership Support and Personality Characteristics. To test whether Supportive Leadership and Personality Characteristics moderate the relationship between Participation and Commitment, it needed to be tested whether the interaction term (or product) of the independent variable and a possible moderator has a significant effect on Commitment (Table 3). The independent variables were first centralized because this makes interpretation of the regression coefficients more easy and it also avoids the occurrence of high correlations between the interaction term on the one side and both predictors that make up the product on the other side.

Overview interaction terms

Participation Leadership Support

Need for Achievement Need for Independence

Participation*Leadershipsupport Participation*Achievement Participation*Independence Table 3: Overview interaction terms

First, a regression analysis is done for the main effect: Participation > Commitment, to see whether the positive correlation also exists when a causal relation is added.

R square shows the percentage of variance that can be explained by the independent variable and is 0.235. this means that 23,5% of the variance in Commitment can be explained by Participation.

The F-value is 15,057, which means the model variance is 15 times bigger than the residual variance. The significance (which means the chance that the F-value will be exceeded) is 0,000. The null hypothesis (that the model variance is equal to the residual variance) is not true. The results of this simple regression analysis show that the independent variable (Participation) has a causal relationship with Commitment. See table 4 and Appendix 7A.

Model 1: C = β



β1P

Commitment Participation Regression coefficient

t-value

P-value (significance)

R square (percentage explained variance) F-value p-value (significance) 0,485 3,88 0,000 0,235 (23,5%) 15,057 0,000 Table 4: Simple regression of participation on commitment

The next step is to investigate whether Leadership Support, Need for Achievement and Need for Independence moderate this causal relation.

4.4.1 Leadership Support

To test Hypothesis 1, A hierarchical analysis was done which means that first only the centralized predictors ‘Participation’ and ‘Leadership Support’ are in the analysis, and in the second step the product variable is added. The advantage is that you get an indication of the strength of the interaction effect by looking at the R Square Change of the second model.

The significance of t, which means the chance that the t-value will be exceeded (two-tailed), is larger than 0,05. From this we can conclude that the weight of the interaction term does not significantly differ from 0. This means that an interaction effect is not present. R square change shows the extra variance that can be explained by the interaction term and is 0.045. This means that only 4,5% extra variance can be explained by the interaction term.

The results of this multiple regression analysis show that the effect of Participation on Commitment does not differ with different levels of Leadership Support. See table 5 and Appendix 7B.

Model 2: C = β0 + β1P + β2L + β3PL

Commitment Participation*Leadership Regression coefficient

t-value

P-value (significance)

Because there is a significant moderate-strength relation between Leadership Support and Commitment and a significant moderate-strength relation between Leadership Support and Participation, Two other possibilities should be investigated: Whether Leadership Support has its own independent effect on Commitment, and whether Leadership Support mediates the relation between Participation and Commitment. These possibilities are dealt with in paragraph 4.4.5.

4.4.2 Need for Achievement

For hypothesis 2 the significance of t, which means the chance that the t-value will be exceeded (two-tailed), is > 0,05. From this we can conclude that the weight of the interaction term does not

significantly differ from 0. This means that an interaction effect is not present. R square change shows the extra variance that can be explained by the interaction term and is 0.001. This means that only 0,1% extra variance can be explained by the interaction term.

The results of this multiple regression analysis show that the effect of Participation on Commitment does not differ with different levels of Need for Achievement. See table 6 and Appendix 7B.

Model 3: C = β0 + β1P + β4A + β5PA

Commitment Participation*Achievement Regression coefficient

t-value

P-value (significance)

R square change (extra percentage explained variance by interaction term) F-change P-value (significance) -0,026 -0,248 0,805 0,001 (0,1%) 0,061 0,805 Table 6: Multiple regression for Participation*Achievement

Because the correlation coefficient between Need for Achievement – Commitment was only 0.162 and also not significant, it makes no sense to do a regression analysis to find out whether Need for

Achievement has its own influence on Commitment.

4.4.3 Need for Independence

At last, hypothesis 3 was tested. The significance of t, which means the chance that the t-value will be exceeded (two-tailed), is > 0,05. From this we can conclude that the weight of the interaction term does not significantly differ from 0. This means that an interaction effect is not present. R square change shows the extra variance that can be explained by the interaction term and is 0.002. This means that only 0,2% extra variance can be explained by the interaction term.

The results of this multiple regression analysis show that the effect of Participation on Commitment does not differ with different levels of Need for Independence. See table 7 and Appendix 7B. Model 4: C = β0 + β1P + β6I + β7PI

Participation*Independence Regression coefficient t-value

P-value (significance)

R square change (extra percentage explained variance by interaction term) F-change P-value (significance) 0,055 0,395 0,695 0,002 (0,2%) 0,156 0,695 Table 7: Multiple regression for Participation*Independence

Because the correlation coefficient between Need for Independence – Commitment was only 0.01 and also not significant, it makes no sense to do a regression analysis to find out whether Need for

Achievement has its own influence on Commitment.

4.4.4 Standardized/unstandardized values

For the multiple regression analyses concerning an interaction term, the unstandardized regression coefficients were used. Standardized values were not computed because the weight of the interaction terms was not significant so the effect sizes did not matter; insight into the importance of the

interaction effects was not needed.

4.4.5 Other possible effects of Leadership Support

Correlation analysis showed a significant moderate-strength relation between Leadership Support and Commitment, which means it is possible that Supportive Leadership has its own independent effect on Commitment. To test this a regression analysis was done with Commitment as dependent variable and Leadership Support as predictor to see whether the positive correlation also exists when a causal

relation is added. The significance of t is < 0,05. From this we can conclude that the weight of Leadership Support significantly differs from 0 and that Leadership Support delivers a significant contribution to the variance of Commitment. R square is 0.088, which means that 8,8% of the variance in Commitment can be explained by the amount of Leadership Support. The F-value is 4,737, which means the model

variance is 4,737 times bigger than the residual variance. The chance that the F-value will be exceeded is 0,034. The null hypothesis (that the model variance is equal to the residual variance) is not true.

The results of this simple regression analysis show that the independent variable (Leadership Support) has a causal relationship with Commitment, though not a very strong one. See table 8 and Appendix 7C.

Commitment Leadership Support Regression coefficient

t-value

P-value (significance)

R square (percentage explained variance) F-value p-value (significance) 0,297 2,177 0,034 0,088 (8,8%) 4,737 0,034 Table 8: Simple regression for Leadership Support

focal independent variable (Participation) is able to influence the dependent variable (commitment). So more participation in the change process would lead to more support from the leader, which would in turn lead to higher commitment. This would logically wise seem possible: Participation means that employees are more involved in the change process, and in order for them to contribute to the change the best they can, employees need to be given the right direction, motivated, encouraged, told what is expected and inspired. So participation leads to more support from the change leader, and through this more supportive structure, commitment to the change can become higher. To test this line of reasoning, a new hypothesis is added and will be tested:

Hypothesis 4: Participation has a positive effect on Commitment through the (partially) mediating factor Leadership Support

For a mediating relationship to exist, Baron & Kenny (1986) described conditions which should be met: a. Variations in levels of the independent variable significantly account for variations in the presumed mediator, b. Variations in the mediator significantly account for variations in the dependent variable, and c. When paths a and b are controlled, the relation between the independent and

dependent variables is significantly decreased (Baron and Kenny, 1986 p.1176), see figure 7. A significant reduction would demonstrate that Leadership Support is indeed a potent mediator.

Models

To make the regression analyses that will be done more clear, the following models will be tested to test for a mediating effect.

Model 1: C = β0 + β1P Model 2: L = β0 + β1P Model 3: C = β0 + β2L Model 4: C = β0 + β1P + β2L

The effect of P on C must be less in the fourth equation than in the first

See table 1 for linking the variables in the models to the factors under investigation. Variable Factor under investigation

c b a Mediator Independent Variable Outcome Variable

C Commitment P Participation L Leadership Support Table 9: Appointment of variables

First of all, we already know that the independent variable (Participation) has a causal relationship with Commitment. The simple regression analysis on page 23 showed that Participation has a statistically significant effect on Commitment. See table 4.

Second, Leadership Support can only explain this relation between Participation and Commitment, when Participation has a causal relationship with Leadership Support. A regression analysis was done with Leadership Support as dependent variable and Participation as predictor.

The results show that the significance of t is smaller than 0,05. From this we can conclude that the weight of Participation significantly differs from 0 and that participation delivers a significant contribution to the variance in Leadership Support.

R is 0.104 which means that 10.4% of the variance in Leadership Support can be explained by

Participation. The F-value is 5.702, which means the model variance is 5.7 times bigger than the residual variance. The significance (which means the chance that the F-value will be exceeded) is 0,021. The null hypothesis (that the model variance is equal to the residual variance) is not true.

The results of this simple regression analysis show that the independent variable (Participation) has a causal relationship with Leadership Support. See table 10.

Leadership Support Participation Regression coefficient

t-value

P-value (significance)

R square (percentage explained variance) F-value p-value (significance) 0,323 2,388 0,021 0,104 (10,4%) 5,702 0,021 Table 10: Simple regression for Leadership Support

Third, Leadership Support can only explain the relation between Participation and Commitment, when Leadership Support has a causal relationship with Commitment. This relationship has already been established, the simple regression analysis on page 25 showed that Leadership Support has a statistically significant effect on Commitment. See table 8.

At last, if Leadership Support can (partially) explain the relation between Participation and Commitment, then Commitment should have a significantly weaker relation with Participation when Leadership Support is kept constant. To check this, a regression analysis was done with Commitment as dependent variable and both Participation and Leadership Support as predictors. The partial regression coefficient of Participation should be less when Leadership Support is kept constant.

Commitment when Leadership Support is kept constant, the regression coefficient, t-value and

significance are lower than when Leadership Support is not kept constant, so the relation is weaker. See table 11. This means that Leadership Support is a potential (partial) mediator of the relation

Participation > Commitment. Commitment Participation when Leadership Support is kept constant Regression coefficient t-value P-value (significance)

R square (percentage explained variance) F-value p-value (significance) 0,434 3,3 0,002 0,257 (25,7%) 8,304 0,001 Table 11: Multiple regression for mediating effect of Leadership Support

5.

CONCLUSION & DISCUSSION

Based on the results of the statistical analysis, conclusions are drawn with respect to the moderating role of the factors investigated. Furthermore, the results will be discussed using additional literature. Last, implications and limitations of this study will be discussed and suggestions for further research will be given.

The results did not show statistical evidence for accepting the hypotheses, which means no moderating role of the factors Leadership Support, Need for Achievement and Need for Independence was found. However, because positive correlations exist between Participation and Leadership support and

between Leadership Support and Commitment, and because Regression Analysis shows that Leadership Support has a positive causal relation with Commitment, the mediating role of Leadership Support was also investigated and conclusions will also be drawn with respect to this. The results showed statistical evidence for accepting this fourth hypothesis, which means a mediating role of Leadership Support was found.

Hypothesis 1: The more supportive the leadership, the stronger is the positive influence of Participation on Commitment

Regression analysis demonstrates that the positive influence of Participation on Commitment does not depend on /become stronger with the presence of a supportive change leader. Hypothesis 1 is therefore rejected.

Hypothesis 2: The higher a person’s Need for Achievement, the stronger is the positive influence of Participation on Commitment

Regression analysis demonstrates that the positive influence of Participation on Commitment does not depend on /become stronger when an employee has a high Need for Achievement. Hypothesis 2 is therefore rejected.

Hypothesis 3: The higher a person’s Need for Independence, the stronger is the positive influence of Participation on Commitment

Regression analysis demonstrates that the positive influence of Participation on Commitment does not depend on /become stronger when an employee has a high Need for Independence. Hypothesis 3 is therefore rejected.

Hypothesis 4: Participation has a positive effect on Commitment through the (partially) mediating factor Leadership Support