1
THE INTERACTION PROCESS IN CHANGE
-
The effect of change agents and change recipient behavior on the
interaction process and its influence on the change outcome.
University of Groningen Faculty of Economics and Business
MSC Business Administration – Change Management August 2013 Nick Baring Student number: S2227223 Ambonstraat 12 9715 HC Groningen (06) 13385172 nickbaring@hotmail.com S2227223 Supervisor/ university: Dr. J. Rupert/ Dr. Y. Eseyrel Acknowledgements
2
ABSTRACT
This research investigates whether change agent and change recipient behavior influences the interaction process and whether the interaction process influences the change outcome. Data was gathered trough questionnaires and analysis was performed on 47 change stories. As hypothesized, results show that change agent behavior influences the change outcome and the interaction process, the change agent behavior consist of three leadership styles shaping behavior, framing change and
3
TABLE OF CONTENTS
1. INTRODUCTION ... 4
2. THEORETICAL FRAMEWORK... 6
2.1 Change outcome ... 6
2.2 Change agent behavior ... 6
2.2.1 Leadership styles ... 6
2.3. The change recipient ... 8
2.2 Interaction process ... 9
3. METHOD ... 13
3.1 Procedure and sample ... 13
3.2. Measures ... 14
3.2.1 Measurements (factor analysis & reliability) ... 14
3.2.2 Factor analysis: leadership style ... 15
3.2.3 Factor analysis: Change outcome, change recipient behavior and the interaction process ... 20
3.2.4 Control variables ... 21
3.3 Data Analysis ... 21
3.3.1 Descriptive statistics and correlations ... 22
3.3.2 Regression and mediation analysis ... 22
4. RESULTS ... 24
4.1.1 Descriptive Statistics and Correlations ... 24
4.2 Regression Analysis ... 25
4.2.1 Regression analysis linear relationships ... 25
4.2.2 Multiple regression analysis ... 27
4.2.3 Mediation analysis ... 30
4.2.3.1 The mediation of the change agent behavior ... 30
4.2.3.2 The mediation of the change recipient behavior ... 31
5. DISCUSSION ... 33
5.1 Conclusion ... 33
5.2 Theoretical implication ... 36
5.3 Limitations and future research ... 37
5.4 Practical implications ... 38
REFERENCES ... 40
4
1. INTRODUCTION
A lot of research has already been conducted in the field of change management, most of the studies are focused on two themes (Van de Ven & Huber, 1990). The first theme is about persuading people to join the change and how you can manage people’s attitudes towards the change. The second theme is about how organizational change develops, grows, and terminates over time (Van de Ven & Huber, 1990). Often these studies are conducted to find out why changes fail (e.g., Kotter, 1995) there is substantial evidence that some 70% of all change initiatives fail (Burnes, Jackson 2011). Pettigrew et al. (2001) think differently about why one should do research on change because according to them contextual considerations are important in reaching a higher level in understanding of how change works (or not) in practice. Pettigrew et al. (2001) suggests that research should explore the context, content, and process of change, and the interactions of these dimensions over time. Three of the four recommendations are taken into consideration in this research because both change agent and change recipient data will be collected for the change cases which provides the context, content and process of the change cases.
In almost all organizational changes two types of parties are involved. These are the change agents, which are responsible for creating and directing implementation of change (Higgs & Rowland, 2011), and recipients, which are either the subject of change or have the responsibility for carrying out the change (Ford, Ford & D’Amelio, 2008). The transaction process between these two parties implies their relationship and their subsequent information exchange, the change agents and change recipients need to work together to bring the change to a successful change outcome. Both agents and recipients expose interaction behavior during this process such as participation and involvement in the change process.
Most research focuses on one side of the change process, on the change agents, for example to see how the leadership style affects the success of change (Higgs & Rowland 2005), or it focuses on the change recipients on for example how resistance and commitment influences the success of change Oreg (2006) and Bouckenooghe et al. (2009). Limited research has been conducted on whether these behaviors correlate and influence the interaction process and if these and whether the interaction process correlates and influences the change outcome. This is interesting to know since the results of this research will indicate which agent and recipient behavior influence the interaction process and whether if the interaction process influences the outcome of the change.
5 and if so which change agent and change recipient behavior influences the interaction process and if the interaction process influences the change outcome.
This results in the following research question and conceptual model:
- What change agent and recipient behavior influence the interaction process and how does this interaction process influence the change outcome?
Change recipient
behavior
Interaction
process
Change agent
6
2. THEORETICAL FRAMEWORK
The following subjects come to order in sequence: the change outcome, the change agent behavior, the change recipient behavior and the interaction process.
2.1 Change outcome
The change outcome is determined with the satisfaction about the change. The satisfaction is high when the change has been perceived as successful, the goals that the change was supposed to achieve are realized and when the organization has benefits from the change. Burnes and Jackson (2011) researched what was the reason of why change projects succeed or fail, they state that there is support that successfulness of change is associated with the value alignment of three factors: those involved in the change intervention, the objective of the intervention and the approach to change (i.e. the values underpinning the content of the change and the process by which it is managed) Burnes & Jackson (2011). In this research it is about the combination between the agents and recipient behavior and their influence on the interaction process which leads to a certain satisfaction with the change outcome. So no further investigation of the three factors of Burnes and Jackson (2011) will be conducted.
2.2 Change agent behavior
The change agent behavior has been researched by many authors including Higgs & Rowland (2011) who made a large contribution for the field by adding their leadership styles. They studied leadership behaviors by analyzing 70 change stories and their impact on change success. They are the first to address leadership and the role of that on the change outcome, with data drawn from samples that cover a range of organizations. Their analysis identified three broad sets of leadership behaviors: shaping behavior, framing change and creating capacity. Higgs & Rowland (2011) created four new categories that appeared to capture behaviors exhibited by leaders combining framing and creating. These behaviors were described as attractor, edge and tension, container and transforming space and are part of the framing change and/or creating capacity leadership behavior. These behaviors will be discussed later on. It has been chosen for this research to focus on the three umbrella leadership styles shaping, framing and creating because the four newly created categories are behaviors that are part of framing and/or creating and show smaller differences between each other. Using the umbrella
leadership styles make the distinction between the leadership styles, shaping, framing and creating, more clear. First shaping behavior will be introduced after that framing change and creating capacity, when the leadership styles are introduced their influence on the change outcome will be discussed.
2.2.1 Leadership styles
First shaping behavior will be discussed, it is described in the literature as: “the
7 ‘thinking about change’, and ‘using an individual focus’” (Higgs & Rowland 2011). It is also the most leader-centric oriented leadership style. The recipients are told what to do and they can have not much input (Higgs & Roland, 2011). Another description is that the leader attempts to shape what people do by acting as a role model, holding others accountable, thinking about change, and focusing on
individuals in the change process (Cawsey et al., 2012). Behaviors that correspond to the leadership style shaping are: expressing their own views and beliefs about the change (Gill 2001), using their own experience of change to shape the implementation (Kotter 1995), holding others accountable for delivering allocated tasks (Bass 1996), being persuasive and expressive (Bass et al. 2003, Higgs & Rowland 2005).
Framing change is about creating starting points for change: “designing and managing the journey” and “communicating guiding principles in the organization” Higgs & Rowland (2011). Framing change thus describes the future after the change and how to get there. Framing consists of two styles namely Attractor and Edge and Tension. An attractor is described in the literature as: “creates a magnetic energy force in the organization to pull it toward its purpose. The leaders pull people toward what the organization is trying to do, not toward themselves” (Higgs & Rowland 2011). Behaviors that correspond to the Attractor style are: a shaper focuses on the day-to-day reality, sees themes and patterns that connect to a wider movement and from this creates a compelling story for the organization (Gill 2001). Furthermore he uses this story to set the context of how things fit together, working the story into the life of the organization so that every conversation and decision ‘makes sense’ (Eisenbach et al. 1999). Edge and tension behavior is described in the literature as: “the leader tests and challenges the organization; amplifies the disturbance generated by the change process by helping people see the repeating and unhelpful patterns of behavior in the culture while at the same time staying firm to keep the change process on course” (Higgs & Rowland 2011). Behavior that corresponds to the edge and tension style is: a leader with an edge and tension style tells it as it is, he describes reality with respect yet without compromise (Gill 2001). He sets the bar high and keeps it there, he stretches the goals and limits of what is possible (Higgs 2003).
The third behavior style is creating-capacity which is described in the literature as “creating individual and organizational capabilities and communication and making connections” (Higgs & Rowland 2011). Another description about creating capacity is that it focuses on creating the capacity for change by increasing individual and organizational capabilities and creating and communicating connections in the organization (Cawsey et al., 2012). The creating capacity style is about creating a clear framework for people to work within. Creating capacity combined with framing results in the container style, behavior that corresponds to the container style are: a leader with the container style channels energy, which means that he provides calm, confident, and affirming signals that allow people to find positive meaning and sense in an anxious situation (Higgs & Rowland 2011).
8 encouraging signals which creates ownership, trust, and confidence (Griffin 2002). Creating Capacity has also an own style which is Transforming Space. Transforming space is described in the literature as: “The leader creates change in the ‘here and now’ as the only thing you can change is the present moment” (Higgs & Rowland, 2011). This style provides people the space to think and act differently, this is done by giving them emotional, temporal and physical space. Behaviors that correspond to the transforming space style are: the leader understands what is happening in the company and breaks established patterns and structures in ways that create movement in the ‘here and now’ (Stacey and Griffin 2005). Furthermore he demonstrates a commitment that creates trust, which enables the system to go to new places, learn about itself and act differently (Denis et al. 1996, Goffee and Jones 2000).
According to Higgs & Rowland (2005) from these three leadership styles, shaping is leader oriented and framing and creating are group- and systemic focused styles. Higgs & Rowland (2005) found that leader-centric behaviors had a negative impact on change success. On the other hand the more group- and systemic focused behaviors were positively related to success in most of the context they examined. This is supported by other researchers for example Cooper et. al., (2005) which state that framing behaviors enables followers to develop commitment to goals, this is seen to enable followers to develop their potentiality and contribute to goal achievement which improves the change outcome. This results in the following hypotheses:
Hypothesis 1A: shaping behavior has a negative influence on the change outcome Hypothesis 1B: framing change has a positive influence on the change outcome Hypothesis 1C: creating capacity has a positive influence on the change outcome
2.3. The change recipient
The change recipient is often described in change literature but in most cases the literature is limited to readiness to change or the negative view resistance to change. In this research these two concepts are also important since it is investigated whether the behavior of recipients influences the interaction process between change agent and change recipient. First readiness to change and its effect on the change outcome will be discussed after that resistance to change and its influence on the change outcome will be discussed.
9 behaviors Bouckenooghe et al. (2009). Readiness to change is a positive counterpart of resistance to change.
Resistance to change is defined by Oreg (2006) as a tridimensional (negative) attitude towards change, which includes affective, behavioral, and cognitive components. The affective component is about how someone feels about the change (e.g., angry, anxious); the cognitive component involves how someone thinks about the change (e.g., is it necessary? Will it be beneficial?); and the behavioral component is about actions or intention to act in response to the change (e.g., complaining about the change, trying to convince others that the change is bad) (Oreg 2006). There are multiple ways in which resistance to change can exhibit as discussed above. It is believed that resistance to change emerges when recipients are not informed about a change (e.g., Wanberg & Banas, 2000), an interesting discovery was made by Oreg (2006). He discovered that resistance increased when more information is communicated this seemed strange because it was hypothesized in line with previous findings (e.g., Wanberg & Banas, 2000). Furthermore a result of a higher level of resistance is associated with a higher intention to quit the job (Wanberg & Banas, 2000).
Bouckenooghe et al. (2009) state that the change may be rejected when the recipients are not ready for the change, so when they do not show readiness behavior. Furthermore Bouckenooghe et al. (2009) states that readiness to change has a positive influence on the change outcome and resistance to change has a negative influence. This results in the following hypotheses:
Hypothesis 2A: readiness to change has a positive influence on the change outcome Hypothesis 2B: resistance to change has a negative influence on the change outcome
2.2 Interaction process
The interaction process is about the interaction between the change agent and the change recipients. Both change agent and change recipient show interaction behavior, Balogun & Johnson (2005) argue that recipient processes of interaction are vital during change and Chung & Luo (2008) claim that it is important to understand which type of change agents really make change happen. Although the importance of the interaction process has been acknowledged, by scholars that focused on either inter-personal behavior of change recipients (Balogun & Johnson, 2005) or change agents behavior (a.o. Higgs & Rowland, 2005 & 2011), little is known about the influence of agent and recipients behavior together on the change outcome. Although there is extended research about change agent behavior and change recipient behavior e.g. Higgs & Rowland (2005), Oreg (2006) and
Bouckenooghe et al. (2009), there is a gap in the academic literature that view agents and recipients behavior simultaneously and if this influences the interaction process and whether the interaction process influences the change outcome.
10 are closely related to each other. They will be discussed individually, starting with the involvement in the change process and after that participatory management. When both concepts are introduced the influence of the change agent and change recipient behavior on the interaction process will be
discussed and finally the influence of the interaction process on the change outcome will be discussed. Involvement in the change process is the term used by Bouckenooghe et al. (2009) which can be defined as how much the recipients are involved in the change process. It consists of two parts namely the quality of change communication and whether organizational members participate in the change process. According to Armenakis & Harris (2002) communication is an important part of the interaction process, this is confirmed by Padmashree Radhaswamy & Anjum Zia (2011) who state that communication is an indispensable behavior at the workplace. Furthermore, better and more effective communication is fundamental to success. Ineffective and wrongly timed and conveyed information can cause nothing but communication blunders, leading to escalations and further discomforts
(Padmashree Radhaswamy & Anjum Zia 2011). A part of the involvement in the change process is the quality of the change communication, a good quality of this communication can be achieved by using a change message. A change message is a message which communicates what the nature of the change is, so what is the change about, furthermore it tries to determine the reactions of the recipients towards the change and increase the motivation of the recipients Bouckenooghe et al. (2009). This increased motivation has according to Walton (1985) a positive influence on the change outcome he states that recipients respond best when their responsibilities are extended and when they are motivated to contribute to the change.
Participatory management means joint decision making, either by a change agent and one change recipient or a change agent and a group of change recipients Locke et al. (1986). Participation is the opposite from authoritative decision making, where agents make decisions on their own, and from delegation, where the change recipient got the authority delegated to make the decision alone (Locke, Schweiger and Latham, 1986). Cotton et al (1988) studied participation in decision making by reviewing 91 articles which write about participation. Five dimensions of participation were identified in these 91 articles namely formal-informal, direct in direct, level of influence, content, short-term versus long-term (Dachler & Wilpert, 1978; Locke & Schweiger, 1979; Lawler 1986; Sashkin 1976). Cotton et al. (1988) argues that the participation levels, including forms of information and
11 In the following part will be discussed which change agent and change recipient behavior influences the interaction process and what the influence of the interaction process on the change outcome is and to conclude a mediation effect will be discussed.
Leaders who use a shaping style do communicate with the recipients but according to Higgs & Rowland (2011) this is mostly focused on the change and in a directive way, the recipients are told what to do and how to do it. A shaping leader is focused on his self, leader-centric oriented, (Higgs & Rowland, 2011) he expresses his own views and beliefs about the change (Gill, 2001) and uses his own experience of change to shape the implementation (Kotter, 1995). These kinds of behavior are not based on involving the change recipients therefore it can be stated that shaping has a negative
influence on the interaction process.
Leaders with a framing style pull people towards the organizational goal and help people by showing them what needs to change (Higgs & Rowland, 2011). The leader amplifies the commotion which is created by the change process by helping people see what works for the change and what is obstructive (Higss & Rowland, 2011), so a framing leader connects to the recipients and involves them in the change process and why certain things happen. Therefore it can be stated that framing change has a positive influence on the interaction process.
Leaders with a creating capacity style focus on communication connections, creating individual capabilities and making connections (Cawsey et al., 2012). Furthermore leaders with a creating style provide recipients with the space to think and act differently (Higgs & Rowland, 2011), so to create their own opinion about the change, this means that creating capacity has a positive influence on the interaction process.
Hypothesis 3A: shaping behavior has a negative influence on the interaction process Hypothesis 3B: framing change has a positive influence on the interaction process Hypothesis 3C: creating capacity has a positive influence on the interaction process
Readiness to change and resistance to change are both part of the behavior of the recipient, and both are claimed to influence the change process. Where the literature states that readiness to change has a positive influence on the change outcome, resistance to change has a negative influence (Bouckenooghe et al. 2009, Oreg 2006). According to Cawsey et al (2012) when a recipient shows readiness behavior they will be supportive to the change and they will show preferred behavior
towards the change agent. The opposite can be said from recipients who show resistance to the change, they will put in no effort in the change or worse and show obstructive behavior. Readiness or
12 Hypothesis 4A: Readiness to change has a positive influence on the interaction process
Hypothesis 4B: resistance to change has a negative influence on the interaction process
The interaction process consist of involvement in the change process and participation, both concepts are quite similar to each other. Both concepts are about interaction between the change agent and change recipient and are aimed at making the change recipient join the change and to make him positive towards the change. This suggests that the interaction process has a positive influence on the change outcome because when employees are motivated towards a change this will have a positive influence on the change outcome Walton (1985) and is confirmed by several authors, they state that interaction results in a better change outcome (Ichiniowski et al.1996 and Walton 1985)
Hypothesis 5: the higher the amount of interaction, the better the outcome of the change
Both the change agent and change recipient are hypothesized to influence the change outcome and the interaction process, furthermore the interaction process is hypothesized to influence the change outcome, these are requirements for a mediation as will be explained in the methods section. The interaction process seems to play a mediating role between change agent and change recipient
behavior influence on the change outcome. Because change agent and recipient behavior are suggested to influence the interaction process and the change outcome and the interaction process is suggested to influence the change outcome. A mediating role of the interaction process means that the influence of the change agent and change recipient behavior on the change outcome is explained by the interaction process, so that the influence of for example shaping behavior on the change outcome is explained by the interaction process, this is based on research conducted by Baron & Kenny (1986), the procedure will be further explained in the methods section. The interaction process is influenced by the change agent and recipient behavior and it influences the change outcome, this means that the influence of the change agent and recipient behavior on the change outcome might be explained by the interaction process. If the relation of the change agent and change recipient behaviors influence on the change outcome is mediated by the interaction process the mediation will be positive since the interaction process is supposed to have a positive influence on the change outcome. This results in the following hypotheses:
Hypothesis 6A: the relation of the change agent behavior with the change outcome is positively mediated by the interaction process
13
3. METHOD
This research has been conducted by using the theory testing approach. A theory testing approach with quantitative measurements has been chosen. This research was conducted to find out whether the change agent and recipient behavior influences the change outcome and if the influence of the change agent and recipient behavior on the change outcome is mediated by the interaction process. Furthermore it is conducted to find out whether the interaction process influences the change outcome. The advantage of this approach compared to a qualitative approach is that it can measure correlations and it gives a clear answer to the research question; do the change agent and recipient behavior influence the interaction process and does the interaction process influence the change outcome?
3.1 Procedure and sample
To test the hypotheses presented in the chapter above, a questionnaire was distributed among a total of 30 companies with as a result 47 change cases consisting of 1 change agent and one to six recipients with an average of 2.36 recipients per case.
Firstly, this questionnaire was distributed digitally. To accomplish this, questions were translated in Dutch and with the help of the Qualtrics survey software, the questionnaire was created online. Before actually sending the questionnaire, a small pilot of ten persons was set up to test whether the questions were clear, to see how much time respondents would need to fill in the questionnaire and to determine whether the software was working appropriately. As a result of the pilot, some minor adjustments were made to the formulation of some questions. The data was collected by multiple researchers who individually distributed the questionnaires among the change cases. This resulted in a database with 47 change agents and 111 change recipients.
14 job experience is 18.3 years with a minimum of zero years and a maximum of 49 years with a standard deviation of 11.2 years.
3.2. Measures
The questionnaire before the factor analysis was conducted consisted of 51 items, which represented the agent and recipient behavior, the interaction process and the change outcome. Two databases were used to answer the hypotheses, namely a recipient database which consists of filled in recipients questionnaires and an agent database which consist of filled in agent questionnaires. Each respondent in the agent database has one to six recipients with an average of 2,36 who filled in the questionnaire about the same change case as the change agent. This creates a total of 47 change cases. To execute analyses on all scales, SPSS (version 20.0) was used. This section provides insight in the items that were used for each variable, the factor analysis and the Cronbach’s Alpha values for each scale (the results of the factor analysis is presented in appendix A).
3.2.1 Measurements (factor analysis & reliability)
A factor analysis has been conducted for the change agent and change recipient database combined. The factor analysis is presented in appendix A. Because the amount of items in the questionnaire is quite large and the amount of respondents is limited to 158 two factor analysis are conducted. One factor analysis for the leadership styles; shaping, framing and creating, which is called the factor analysis on the leadership styles, and one for the interaction process, recipient behavior and the change outcome, which is called the second factor analysis. This approach has been chosen because for each item five respondents are needed (e.g., Bryant and Yarnold, 1995), for the leadership analysis this means that the factor analysis consists of 32 items and the second factor analysis consist of 19 items. The amount of respondents is 158, for the leadership factor analysis 200 respondents are according to Byant and Yarnolt (1995) needed and for the second factor analysis 95 respondents are needed.
15
3.2.2 Factor analysis: leadership style
The change agent behavior is measured through the leadership style. The leadership style is measured through a total of 40 items, these items are based on Higgs & Rowland (2011) it measures which leadership style is used: shaping, framing or creating. The items are rewritten into agent items and recipient items and translated into Dutch. Examples questions will be given for each leadership style, all styles are measured by using a seven-point Likert scale.
Shaping behavior consisted of eight items, it includes items like: ‘The change agent regularly used his experience to guide the implementation’, ‘the personal commitment of the change agent was to motivate employees to get involved with the change’ and ‘The change agent shaped the
implementation by using his experience with other changes’.
Framing change is according to the theory measured trough two factors, attractor and edge & tension, this is confirmed by the factor analysis in SPSS. The attractor style consists of eight items, it includes items like: ‘the change leader felt responsible to help the recipients deal with the change’, ‘the change leader kept the overarching organization goal in mind in everything that he does’ and ‘the change leader easily put aside his own ambitions for the sake of the change project’. The edge and tension style consists of nine items including: ‘the change leader did not make things better than they were, he kept realistic’, ‘the change leader did not avoid difficulties’ and ‘the change leader did not make any compromises when it came to the quality’.
Creating capacity consisted of seven items, including items like: ‘the change leader made sure that there was plenty of room for employees to think differently’, ‘the change leader pulled employees out of their daily routine to make them think differently’ and ‘the change leader organized discussions to create different solutions’. Framing change and creating capacity combined also result in a
leadership behavior, namely container it consist of 8 items including: ‘the change leader managed to create certainty and trust in turbulent times’, ‘the change agent was confident that he would bring the
Table 1 Reliability analysis
Variable Amount of items Cronbach’s Alpha Value
16 change to a successful conclusion’ and ‘the change agent was confident in carrying out his vision even in hard times.
When looking at the factor analysis as presented below in table 2 certain things are interesting, all items that belong to the shaping behavior concept load on the same factor. Most of the creating capacity items also load on the same factor, the difficulty is within the framing change style. To start with the container style which in theory is proposed as a combination of framing and creating. It loads on four different factors, it loads on the factor where edge & tension and attractor load on but also on the creating capacity factor, this is in line with the theory. On the other hand this makes it very hard or even impossible to get the container style to load on one factor because it is a combination of framing and creating. Since shaping loads on one factor, creating loads specifically on one factor and the attractor and edge and tension styles both partially load on the first factor the different leadership styles can still be distinguished, shaping, framing and creating, it has been chosen to delete the container style from the database because the container style does not load nicely on one factor.
This results in the following factor analysis as shown in table 3, it shows that all shaping items and some attractor items load on the same factor, the reason for this is that the formulation of these question are all oriented on the change agent and how he does things. For example ‘The change agent regularly used his experience to guide the implementation’ for the shaping style and ‘the change leader felt responsible to help the recipients deal with the change’ for the attractor style. The attractor items that only load on the same factor as shaping were deleted from the database these were item two, four, seven and eight. The reason why they load on shaping varies, item two, four and eight contain some shaping behavior in the formulation, for example item two which states that the leader was constantly aware of his role as change agent which is a leader oriented statement which corresponds to shaping. The loading of item seven cannot be clarified directly by its formulation, but when looking more closely to the questionnaire the items that precede item seven are about the companies vision about the change while question seven is about the change agent behavior, this might be the explanation of the loading on shaping. These items were deleted from the database instead of included with the shaping items because their loading on shaping were all lower than the lowest loading of the items that were intended to load on shaping. Furthermore since they were proposed to load on framing, which they do not, they are deleted. On the shaping side item six also loads on framing but the difference between the shaping factor and the framing factor is higher than .2 so the item was kept. The first item was deleted because it loaded on three factors and it lowered the Cronbach’s Alpha value from .917 to .905. Furthermore on the attractor side item one and six load on two factors namely shaping and framing the loadings of item one are around the same size whether the loading of item six is .15 higher on framing than on shaping. Since it decreases the Cronbach’s Alpha value a lot when the two items were deleted they both have been kept in the database.
17 items are for the attractor style one, three, five and six and for the edge and tension style one, two, three, four and nine. they are all formulated in a small way but more importantly the formulation of the items all involve serving the organizational goal or are about helping the recipients to cope with the change. This means that for edge and tension item five, six, seven and eight are deleted from the database, this because these items all load on the same factor which seems to be the factor for edge and tension but because the umbrella styles shaping, framing and creating are chosen these items cannot be used. The first two items that were intended to load on creating capacity also loaded on framing change but the formulation of these items fits the theory of creating capacity. Both items were derived from the characteristics of creating capacity therefore these items cannot be added to the framing change factor. These characteristics are quite similar to the framing change characteristics and the formulation of these items resembles to the framing change characteristic of creating starting points for the change. But as stated before item one and two both are derived from characteristics of creating capacity namely; ‘the leader understands what is happening in the company and breaks established patterns and structures in ways that create movement in the here and now’ and can therefore not be added to change framing change factor. To sum up framing change consists of nine items with a Cronbach’s Alpha value of .911.
18
TABLE 2
FACTOR ANALYSIS LEADERSHIPSTYLES
1 2 3 4 5 Shaping Item 1 .402 -.414 .518 .225 -.305 Item 2 .136 .016 .673 -.074 .064 Item 3 .074 -.095 .630 .119 .175 Item 4 -.254 .297 .830 -.072 .013 Item 5 .119 .258 .607 -.157 .111 Item 6 .361 -0.84 .637 -.008 -.017 Item 7 -.085 .069 .759 -.003 .144 Item 8 .074 -.120 .653 .130 .156 Attractor Item 1 .390 .249 .406 -.072 -.052 Item 2 .162 .296 .452 .056 -.067 Item 3 .267 .227 .139 .064 .156 Item 4 .090 .227 .560 -.128 .154 Item 5 .447 .131 .069 .021 .201 Item 6 .396 .018 .290 .102 .088 Item 7 -.094 .241 .469 .271 -.011 Item 8 -.020 .226 .516 .157 -.040
Edge & Tension Item 1 .896 .024 .065 -.194 -.012
Item 2 .592 .146 -.053 .183 .050 Item 3 .653 .215 .033 .057 -.012 Item 4 .664 -.089 .150 .141 -.236 Item 5 .164 .179 .064 .593 -.103 Item 6 .117 .254 -.135 .736 -.130 Item 7 -.061 .203 .041 .726 -.063 Item 8 .003 .191 -.024 .711 .027 Item 9 .662 -.298 -.042 .438 .038 Container Item 1 .220 -.016 .030 -.080 .576 Item 2 .847 .208 -.031 -.229 .059 Item 3 .313 .689 -.004 -.131 -.085 Item 4 .830 .007 -.079 -.124 .254 Item 5 .686 -.056 -.007 .132 .025 Item 6 .302 .560 -.111 .229 -.056 Item 7 -.238 -.192 .139 .800 .300 Item 8 .166 .107 -.089 .388 .441 Creating Item 1 .203 -.118 .073 .031 .659 Item 2 -.188 -.090 .224 .060 .759 Item 3 .034 .734 -.085 .080 .174 Item 4 .096 .707 -.118 .143 .094 Item 5 -.062 .563 .112 .221 .069 Item 6 -.060 .833 .157 .012 -.093 Item 7 -.229 .896 .236 .033 -.208
Notes. Extraction Method: Principal Component Analysis. Rotation Method: Promax with Kaiser
19
TABLE 3
FACTOR ANALYSIS LEADERSHIPSTYLES
1 2 3 4 Shaping Item 1 .659 -.476 -.171 .446 Item 2 .624 .047 .236 -.095 Item 3 .649 -.067 .182 .091 Item 4 .785 .345 -.177 -.128 Item 5 .536 .276 .256 -.127 Item 6 .666 -.117 .222 .122 Item 7 .811 .122 -.011 -.080 Item 8 .682 -.068 .136 .098 Attractor Item 1 .354 .214 .311 .080 Item 2 .454 .251 .057 .164 Item 3 .112 .194 .421 .100 Item 4 .507 .266 .221 -.106 Item 5 .018 .107 .548 .173 Item 6 .361 -.033 .310 .231 Item 7 .487 .245 -.059 .201 Item 8 .487 .252 -.019 .133
Edge & Tension Item 1 .003 -.082 .809 .080
Item 2 -.131 .099 .575 Item 3 .119 .082 .420 .329 Item 4 .039 -.189 .428 .422 Item 5 .059 .176 -.029 .713 Item 6 -.168 .280 -.025 .781 Item 7 .097 .241 -.268 .779 Item 8 .059 .235 -.116 .706 Item 9 .030 -.252 .426 .565 Creating Item 1 .066 .006 .836 -.210 Item 2 .208 .096 .589 -.298 Item 3 -.046 .657 .219 .079 Item 4 -.220 .700 .312 .144 Item 5 .034 .638 .095 .142 Item 6 .063 .822 .009 .014 Item 7 .173 .850 -.280 .068
Notes. Extraction Method: Principal Component Analysis. Rotation Method: Promax with Kaiser
20
3.2.3 Factor analysis: Change outcome, change recipient behavior and the interaction process
The second factor analysis consists of the change outcome, change recipient behavior and the interaction process. These constructs are measured through a total of 19 items. The items are rewritten into agent items and recipient items and translated into Dutch. Examples questions will be given for each construct, each construct is measured by using a seven-point Likert scale.
The change outcome is measured with a total of three items. It includes items like: ‘I think the goals that were formulated for the change will be achieved.’ and ‘I think the organization benefits from the change’.
The change recipient behavior is measured through two constructs, readiness to change and resistance to change. Readiness to change is measured through the use of three items. These items are based on Bouckenooghe et al. (2009), it includes items like: ’the change recipients made a significant contribution to the change’ and ‘the change recipients put a lot of energy in the change process’. Resistance to change is measured through three items. These items are based on Oreg (2006), it includes items like: ‘during the change process recipients searched for ways to prevent the change from happening’ and ‘during the change process recipients complained about the change to others’.
The interaction process is measured through the involvement in the change process and participatory management, it consists of a total of ten items. Involvement in the change process is measured with a total of six items. Five of the items are based on Bouckenooghe et al. (2009), one item is based on Bennebroek Gravenhorst, Elving and Werkman (2005) and CATOCQ survey (2008). It includes items like: ‘the change agent provided the recipients with clear information.’ and ‘The change agent consulted the employees regularly about the change’. Participatory management is measured with a total of five items. Four of the items are based on Bouckenooghe et al. (2009) and one item is based on Bennebroek Gravenhorst, Elving and Werkman (2005) and CATOCQ survey (2008). It includes items like: ‘the change agent always discussed important changes with the employees’ and ‘the change agent took the input provided by the employees seriously’.
21
TABLE 4
FACTOR ANALYSIS NON LEADERSHIPSTYLES
1 2 3 4
Involvement in Item 1 .900 -.226 .115 -.002
the change Item 2 .883 -.022 .035 -.023
process Item 3 .712 .348 -.141 -.014 Item 4 .866 -.103 .111 -.017 Item 5 .859 .068 -.066 .038 Item 6 .831 -.056 .080 -.056 Participatory Item 1 .821 .134 -.056 .105 Management Item 2 .857 -.038 -.099 .012 Item 3 .840 -.084 .047 -.075 Item 4 .665 .223 .035 .025 Change Item 1 .047 .900 -.047 .016 Outcome Item 2 -.123 .893 .119 .016 Item 3 .016 .882 .046 -.058 Readiness Item 1 .102 -.006 .839 .063 Behavior Item 2 .016 -.004 .943 .050 Item 3 -.031 .137 .865 -.090 Resistance to Item 1 .003 -.031 -.045 .799 Change Item 2 .021 .031 -.066 .932 Item 3 -.035 -.023 .147 .826
Notes. Extraction Method: Principal Component Analysis. Rotation Method: Promax with Kaiser
Normalization
3.2.4 Control variables
Several control variables were included in the questionnaire; age, gender, nationality, country of birth, education, experience in this function, total job experience and the sector where the company operates. Not all of these control variables will be used in this research, some control variables gave no interesting results like the nationality, since only one respondent is not Dutch. Furthermore other control variables cannot be used because the databases of agent and recipient will be aggregated, this makes the ordinal and nominal control variables useless because for example the education will lie somewhere between primary school and university. The linear control variables like age, experience in this function and total job experience can be used. The control variables will be added in the
correlation analysis. Succeeding the measurements section, the following section describes the data analyses that were performed to test the hypotheses.
3.3 Data Analysis
22 of one case were aggregated with each other, so that each change case now has a change agent and an aggregated change recipient. After that the change agent and aggregated recipients were aggregated so that the database consists of 47 change stories/cases. This way of ordering the data has been chosen because the hypotheses can be answered most objectively, this aggregation makes the results per case most objective since it consists of two to seven opinions about the change with an average of 3.36 respondents per change case.
3.3.1 Descriptive statistics and correlations
Firstly, the mean and standard deviation for dependent, independent and control variables were calculated. Since the items of the questionnaire were tested using a seven- point Likert scale, except for the control variables, these descriptive statistics help to create insight in the overall direction (either positive or negative) of the answers given by the respondents. Following the descriptive statistics, a correlation analysis was performed using the Spearman’s rank correlation coefficients since this test is appropriate for this kind of data.
3.3.2 Regression and mediation analysis
The second step concerns regression analyses. These analyses were used to test the relationships between the variables. After that the third step, the mediation effect of the interaction process on the change agent and recipient behaviors influence on the change outcome is analyzed. According to de Heus, van der Leeden & Gazendam (1995) there are four assumptions that need to be controlled for before performing regression analysis;
1. Both the dependent and independent variable need to be interval or ratio. However, research has shown that quasi-interval data, such as a seven-point Likert-scale, can be used as well since the results remain stable when there is no strong variation in the scales of the variables (de Heus, van der Leeden & Gazendam, 1995). Therefore, this assumption can be verified.
2. The second assumption concerns linearity of the variables, which was tested during the regression analysis if the significance is smaller than 0.05 than there is a linearity of the variable.
3. The third assumption is related to randomness. All independent variables should be fixed (selected by the researcher) whereas the dependent variable should be random. Even though this assumption is mentioned, it is possible to conduct regression analysis without verifying this assumption. As de Heus, van der Leeden & Gazendam (1995) mention; this assumption is almost never verified when surveys are used, and as such, this also account for this specific research.
23 Before mediation can be accepted it needs to fulfill four steps, the change agent and recipient behavior should influence the change outcome, secondly change agent and recipient behavior should influence the interaction process, third the interaction process should influence the change outcome and fourth the effect on the change agent and recipient behavior on the change outcome when controlled by the interaction process should be zero. If all four of these steps are met, then the interaction process completely mediates the effect of the change agent and change recipient behavior on the change outcome Baron and Kenny (1986).
24
4. RESULTS
As mentioned earlier, the data was analyzed in three steps. Firstly, the descriptive statistics and the results of the correlation analyses are discussed. Thereafter, the outcomes of the regression tests are presented and finally the mediation analysis will be conducted.
4.1.1 Descriptive Statistics and Correlations
The mean and standard deviation were calculated for each variable included in the conceptual model and the control variables. Following the descriptive statistics, a correlation analysis using the Spearman’s Rho correlation coefficient was performed. The outcomes are presented in the table below (table 5).
TABLE 5
Descriptive Statistics & Correlations
M SD 1 2 3 4 5 6 7 8 9 1. Change outcome 5.70 0.88 - 2. Interaction 5.05 0.79 .35* - 3. Shaping 5.09 0.83 .50** 0.68** - 4. Framing 5.21 0.66 .53** 0.70** 0.65** - 5. Creating 4.67 0.86 .21 0.75** 0.74** 0.62** - 6. Readiness to change 5.16 0.98 .47** 0.63** 0.68** 0.54** 0.58** - 7. Resistance to change 3.16 1.05 -.50** -0.32* -0.20 -0.32* -0.08 -0.25 - 8. Age 42.02 7.38 -.19 -.04 .13 0 .12 .03 .13 -
9. Experience current job 7.79 4.76 -.25 -.10 -.08 -.07 -.05 -.04 .07 .43** -
10. Total job experience 20.41 6.83 -.12 -.04 .14 .02 .16 .03 .07 .91** .35**
*. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed).
25 the change as high. Resistance to change scores low compared to the other variables (M=3.16) which is quite logical since the questions on this factor are formulated in a negative manner which means that when it has a high score there is a lot of resistance to the change.
As suggested in the theoretical part the relationship between the independent and dependent variables show significant correlations (p<.01 & p<.05).
The correlations show interesting issues. Firstly the correlation between shaping behavior (r=.50, p<.01), framing change (r=.53, p<.01), readiness to change (r=.47, p<.01) and resistance to change (r=-.50, p<.01) are significantly correlated to the change outcome. This means that these variables could have an influence on the change outcome. The interaction process is significantly correlated to the change outcome(r=.35, p<.05). Another notable result is that the three leadership styles shaping, framing and creating all are significantly positively related to each other, for shaping and framing (r=.66, p<.01), for shaping and creating (r=.73, p<.01) and for framing and creating (r-.71, p<.01) this means that there is multicollinearity between these variables. Finally the control variables are not correlated to any variable except for the other variables, and these correlations are quite logical for example the older you are the more job experience you are likely to have. These control variables will therefore not be considered in the regression analyses.
4.2 Regression Analysis
As mentioned in the methodology section, the regression analyses were performed in three steps. First, the simple linear regression analysis outcomes concerning the linear relationships are discussed where after the multiple regression analysis outcomes are presented. Finally, mediation tests that were executed are presented.
4.2.1 Regression analysis linear relationships
Hypothesis one is about the influence of the leadership styles on the change outcome. Hypothesis 1A states that shaping has a negative influence on the change outcome. The regression analysis showed that shaping has a significant positive influence on the change outcome (ß= .57 p <.001, R² = .28). This means that the more shaping the leadership style of a change agent the better the change outcome will be. So hypothesis 1A can be rejected. Hypothesis 1B states that framing has a positive influence on the change outcome. The regression analysis shows that framing has a
26 Hypothesis two concerns the influence of the change recipient behavior on the change
outcome. Hypothesis 2A states that readiness to change has a positive influence on the change outcome. The regression analysis shows that readiness to change has a significant positive influence on the change outcome (ß= .37 p =.004, R² = .17). This means that the more readiness to change the recipients are the better the change outcome will be, hypothesis 2A is accepted. Hypothesis 2B states that resistance to change has a negative influence on the change outcome resistance to change has significantly negative influence on the change outcome (ß=-.39 p =.001, R² = .22). Thus the more resistant the recipients are towards a change the more worse the change outcome will be, also hypothesis 2B is accepted.
Hypothesis 3 concerns the influence of the change agent behavior, the leadership styles, on the interaction process. Hypothesis 3A proposes that shaping behavior has a negative influence on the interaction process. The regression analysis showed that shaping behavior is significantly positive related to the interaction process (ß= .69, p <.001, R² = .53). This means that shaping behavior has a positive influence on the interaction process instead of a negative influence, therefore hypothesis 3A can be rejected. Hypothesis 3B states that framing change has a positive influence on the interaction process. Regression analysis confirmed that framing change is significant positively related to the interaction process (ß= .87 p <.001, R² = .53). That is, when the change agent has a high score on framing change the interaction process will improve as well, hypothesis 3B is accepted. In hypothesis 3C, it was proposed that creating capacity improved the interaction process. Creating has a significant positive relationship with the interaction process (ß= .68, p <.001, R² = .54). This indicates that when a change agent has a high level of creating capacity the interaction process also improves, therefore hypothesis 3C can also be accepted.
Hypothesis 4 is about the influence of the change recipient behavior on the interaction process. In hypothesis 4A it was proposed that a high level of readiness to change influences the interaction process in a positive way. The regression analysis shows that readiness to change is significantly positive related to the interaction process (ß= .53 p <.001, R² = .43). Thus, when a recipient has a high level of readiness behavior the interaction process is also better, thus hypothesis 4A can be accepted. Hypothesis 4B proposes that resistance to change would have a negative influence on the interaction process. The results of the regression analysis show that resistance to change is not significantly related to the interaction process (ß=-.21 p =.052, R² = .08). This means that resistance to change does not significantly influence the interaction process therefore hypothesis 4B can be rejected.
27
4.2.2 Multiple regression analysis
Following the simple regression analysis, this section focuses on the outcome of the multiple regression analyses that was performed to test the influence of the different leadership styles on the interaction process. A Multiple regression analysis has been conducted for the change agent behavior because a change agent can use parts of multiple leadership styles at the same time, a leader will not be 100% shaping, 100% framing or 100% creating. As can be seen from the results the averages of shaping, framing and creating lie fairly close to each other, shaping: 5.09, framing: 5.21 and creating:4.67, this could mean that change agents exhibit behaviors of multiple leadership styles. Therefore a multiple regression analysis is adequate because it tests the influence of the leadership styles, at the same time, on the interaction process. According to the theory readiness to change is the positive counterpart of resistance to change which would suggest that the scores should be opposite from each other. This is not the case, the average score on readiness to change is 2 points higher than resistance to change, but they are not the opposite from each other. Furthermore a recipient can show readiness behavior to some parts of the change and resistance to other parts of the change. Therefore it has been chosen to also use a multiple regression analysis of the recipients behavior influence on the interaction process.
The results for the multiple regression analysis of the leadership styles influence on the change outcome can be found in table 6. Hypothesis 1A states that a high level of shaping behavior has a negative influence on the change outcome. Hypothesis 1B and 1C on the other hand state that framing and creating capacity have a positive influence on the change outcome. The explained variance between these three independent variables and the dependent variable is .673. Based on the outcomes of the regression analysis shaping is significantly positively related to the change outcome (ß= .753 p =.001). Secondly framing change is not significantly positive related to the change outcome (β = .297, p= .234). Thirdly creating capacity is significantly negative related to the change outcome (β = -.501, p= .007). This multiple regression analysis gives some unusual results compared to the linear
regression analysis, for example that creating does influence the change outcome which is strange since it is not significantly related to the change outcome and framing does not significantly influence the change outcome. According to this regression Hypothesis 1A and 1B can be rejected and
hypothesis 1C will be accepted.
-28 .330, p= .003). This means that readiness to change has a significant positive influence on the change outcome and resistance to change has a significant negative influence on the change outcome. The outcomes of the multiple regression analysis are the same as at the linear regression. Hypothesis 2A and 2B are both accepted.
The results for the multiple regression analyses on the leadership styles on the interaction process can be found in table 8. Hypothesis 3A states that a high level of shaping behavior has a negative influence on the interaction process. Hypothesis 3B and 3C on the other hand state that framing and creating capacity have a positive influence on the interaction process. The explained variance between these three independent variables and the dependent variable is .403. Based on the outcomes of the regression analysis shaping behavior is not significantly related to the interaction process (ß= .199 p =.177). Secondly framing change is significantly positive related to the interaction process (β = .394, p= 0.021). Thirdly creating capacity is also significantly positive related to the interaction process (β = .345, p= 0.005). This means that framing and creating have a significant positive influence on the interaction process while the influence of shaping behavior on the interaction process is not significant. Therefore hypothesis 3A can be rejected and hypothesis 3B and 3C can be accepted. This means that contradicting to the linear regression analysis, where each leadership style has a significant positive influence on the interaction process, shaping behavior is not significantly related to the interaction process (hypothesis 3A). This result seems more reliable than the linear regression analysis since in the multiple regression analysis the effect of the three leadership styles are not measured individually but in group. Therefore the influence of each style can directly be compared with the other two styles on whether which style has a significant influence on the interaction process. The results differ because when shaping was assessed individually it seemed to be significantly positive related to the interaction process but when it was assessed with the other leadership styles, shaping compared to framing and creating, did not have a significant influence. On the other hand shaping, framing and creating show multicollinearity to each other which can bias the results but since the results seem quite plausible the results are accepted. Thus hypothesis 3A is rejected while
hypothesis 3B and 3C are accepted.
29 regression analysis resistance to change was close to being significant in the multiple regression it is clearly not significant. The differences in the scores have the same explanation as has been given for the leadership styles. This explanation is that in comparison with readiness behavior resistance to change has no significant influence on the interaction process. Therefore hypothesis 3A is accepted and hypothesis 3B is rejected.
TABLE 6
Multiple Regression Analysis; Leadership styles
B ß Shaping behavior .709 .753*** Framing change .225 .297 Creating capacity -.489 -.501** R² .403 F 9.687 R²∆ .403 F ∆ 9.687
Note ***p<.001, **p<0.01,*p<.05 Dependent variable = Change Outcome
TABLE 7
Multiple Regression Analysis; change recipient behavior
B ß Readiness to change .332 .297*** Resistance to change -.395 -.330** R² .321 F 10.397 R²∆ .321 F ∆ 10.397
Note ***p<.001, **p<0.01,*p<.05 Dependent variable = Change Outcome
TABLE 8
Multiple Regression Analysis; Leadership styles
B ß Shaping behavior .209 .199 Framing change .332 .394* Creating capacity .375 .345** R² .669 F 29.025 R²∆ .669 F ∆ 29.025
Note ***p<.001, **p<0.01,*p<.05 Dependent variable = interaction process
Note ***p<.001, **p<0.01,*p<.05 Dependent variable = interaction process
TABLE 9
Multiple Regression Analysis; Change recipient behavior
30
4.2.3 Mediation analysis
Hypothesis 6A and 6B propose that the influence of change agent and change recipient behavior on the change outcome is positively mediated by the interaction process. To test whether this hypothesis holds, four steps, as described in the methods section, need to be satisfied. For all
regression analysis, if there were multiple variables, except the influence of the change agent behavior on the change outcome, multiple regression analysis were used. First the results of mediation effect of the interaction process on the influence of the change agent behavior on the change outcome will be given. After that the mediation effect of the interaction process on the influence of the change recipient behavior on the change outcome will be given. A summary of the results are given in table 10 for the change agent behavior and table 11 for the change recipient behavior.
4.2.3.1 The mediation of the change agent behavior
To test hypothesis 6A the simple linear regression of the change agent behaviors influence on the change outcome has been used because shaping, framing and creating show multicollinearity to each other and the results of the multiple regression analysis are not in line with the correlation analysis. This results in that creating capacity does not correlate to the change outcome according to the Spearman’s Rho correlation coefficient and according to the linear regression analysis creating has no influence on the change outcome (ß= .16 p =.288, R² = .03) but in the multiple regression analysis creating capacity does have a significantly negative influence on the change outcome β = -.50, p= .007). Furthermore framing change is correlated to the change outcome and it does have a significantly positive influence on the change outcome in the simple linear regression (ß= .60 p =.001, R² = .21) but it does not have a significant influence on the change outcome (β = .30, p= .234) in the multiple regression analysis. Therefore it has been chosen to follow the linear regression analysis.
For the first step the correlation analysis shows that from the agent behavior, shaping and framing correlate with the change outcome and creating does not, the regression analysis shows that creating has no significant influence on the change outcome (ß= .16 p =.288, R² = .03), therefore creating cannot satisfy the first step in the process of testing mediation. Shaping (ß= .57 p <.001, R² = .28) and framing (ß= .60 p =.001, R² = .21) do have a significant influence on the change outcome. This means that the interaction process does not mediate the influence of creating on the change outcome.
The second step is about the influence of the change agent behavior on the interaction process. From the change agent behavior shaping has no significant influence on the interaction process (ß= .20 p =.177), framing has a significant influence on the interaction process (β = .39, p= 0.021) and
31 The third step focuses on whether the interaction process influences the change outcome. The interaction process has a significant influence on the change outcome (ß= .33, p <.05, R² = .09).
The fourth and final step can now be executed with change outcome as dependent variable and the interaction process and framing change as independent variable. The regression analysis shows that framing change (ß= .68 p =.012) still has a significantly positive influence on the change outcome while the interaction process does not (ß=--.09 p =.691). This means that there is no evidence that the interaction process mediates the influence of the change agent behavior on the change outcome. Thus hypothesis 6A can be rejected. Table 10 summarizes the results of the linear regression of the change agent behavior on the change outcome, the multiple regression of the change agent behavior on the interaction process, the linear regression of the interaction process on the change outcome and the mediation test of the change agent.
TABLE 10
Mediation on change agent behavior
Variable Change outcome Interaction process
Step ß R² ß R² 1 Shaping behavior .57*** .28 Framing change .60*** .21 Creating capacity .16 .03 2 Shaping behavior .20 .67 Framing change .39* .67 Creating capacity .35** .67 3 Interaction process .33* .09 4 Framing change .68* .21 Interaction process -.09 .21 Note ***p<.001, **p<0.01,*p<.05
4.2.3.2 The mediation of the change recipient behavior
To test hypothesis 6B from the recipient behavior the first step is to test whether resistance and readiness to change correlate to, and influence the change outcome. Readiness to change has a significant positive influence on the change outcome (ß= .30 p =.012) while resistance to change has significantly negative influence on the change outcome (ß=-.33 p =.003). Thus the first step is satisfied.
The second step is about the influence of change recipient behavior on the interaction process. The multiple regression analysis shows that only readiness to change has a significantly influence on the interaction process (β = .50, p<0.001). Resistance to change has no significant influence on the interaction process (β = -.17, p= 0.183). This means that the interaction process does not mediate the influence of resistance to change influence on the change outcome.
The third step focuses on whether the interaction process influences the change outcome. The interaction process has a significant influence on the change outcome (ß= .33, p <.05, R² = .09).
32 readiness to change (ß= .35 p =.039) still significantly positive influence the change outcome while the interaction process does not (ß=-.05 p =.817). This means that there is no evidence that the interaction process mediates the influence of the change recipient behavior on the change outcome. Thus hypothesis 6B can be rejected. Table 11 summarizes the results of the multiple regression of the change recipient behavior on the change outcome, the multiple regression of the change recipient behavior on the interaction process, the linear regression of the interaction process on the change outcome and the mediation test of the change recipient.
TABLE 11
Mediation on change recipient behavior
Variable Change outcome Interaction process
33
5. DISCUSSION
This research was conducted in order to gain insight into which change agent and change recipient behaviors influence the interaction process and if this interaction process influences the change outcome. Not much research has been conducted towards this subject since most research focuses on two themes, the first theme is about persuading people to join the change. The second theme is about how organizational change develops, grows, and terminates over time (Van de Ven & Huber, 1990). The change outcome is measured as the perceived change outcome of both change agent and change recipient, it is measured as a subjective measurement since it measures the perceived change outcome. The interaction process is measured through two concepts namely involvement in the change process and participatory management, these concepts are closely related according to the theory since both consider the recipients influence on the decision making in the change process. Both concepts were eventually considered as one, since the factor analysis conducted in this research showed that both concepts load on the same factor. The change agent behavior consist of the leadership style that he or she employs: shaping behavior, framing change and creating capacity. As already discussed in the methods section it has been chosen to delete the container style because in the factor analysis it loaded on four different factors which makes it difficult to recognize a pattern. Furthermore a large part of the items of the attractor style and edge and tension style loaded on the same factor in the factor analysis, this is a clear pattern for the framing leadership style, conducted in this research, therefore this research is focused on shaping, framing and creating. The change recipient behavior consists of readiness to change and resistance to change, these two constructs both represent a positive view on recipients namely readiness to change and the counterpart a negative view on recipients namely resistance to change. The change recipient literature is focused on these two constructs and is proposed by several authors as important for the change process Bouckenooghe et al. (2009) and Oreg (2006).
5.1 Conclusion
To answer the research question: ‘What change agent and recipient behavior influences the interaction process and how does this interaction process influence the change outcome?’ The interaction process has a positive influence on the change outcome. From the change agent behavior, shaping behavior and framing change behavior have a positive influence on the change outcome and framing change and creating capacity have a positive influence on the interaction process. From the change recipient behavior, readiness to change has a positive influence on the change outcome while resistance to change has a negative influence on the change outcome. Furthermore readiness to change has a positive influence on the interaction process and resistance to change does not influence on the interaction process. Finally the interaction process does not mediate the influence of the change agent and change recipient behavior on the change outcome.
34 the influence of shaping behavior on the interaction process. Previous research suggest that shaping behavior has a negative influence on the change outcome and the interaction process Higgs & Rowland (2005), Cooper et al., (2005). The results of this research state that shaping behavior has a significant positive influence on the change outcome and no influence on the interaction process. Several explanations can be found for the positive influence of shaping on the change outcome and non influence on the interaction process. The first reason is the formulation of the statements in the questionnaire, the shaping statements are formulated quite positively for example: ‘the personal commitment of the change agent was to motivate employees to get involved with the change’. They resemble to the descriptions of transformational leadership: ‘Transformational leaders engage in behaviors that articulate a clear vision, demonstrate enthusiasm and passion, and inspire and motivate employees to work hard’ (Bass, 1985). Transformational leadership has a positive influence on the change outcome Bass (1985) this could be an explanation for the positive influence of shaping behavior on the change outcome. Furthermore since the shaping items focused a lot on the change outcome, this might be the reason why shaping has no influence on the interaction process instead of a negative influence. This is not strange because shaping behavior does not aim to stimulate the
