The relationship between leadership behaviour and resistance to change: the moderating role of leadership behaviour asymmetry.
Nathalie Smit Student number: S2208814 MSc BA Change Management
Supervisor: dr. J.F.J. Vos Co-assessor: drs. H.P. van Peet
Date: January 23rd, 2017
Word count: 8090
2 ABSTRACT
The aim of this study was to investigate the influence of agents’ leadership behaviour on recipients’
resistance to change, and the moderating role of leadership behaviour asymmetry. Three types of leadership behaviour were studied, namely framing change, creating capacity, and shaping behaviour.
This study contributes to the literature by investigating how change agents’ leadership behaviour influences recipients’ resistance to change. Moreover, this study contributes by examining how the differences between agent and recipient perspectives regarding change agents’ leadership behaviour (asymmetry) influences this relationship. A field study was conducted involving agents and corresponding recipients from 103 organizations. The results of the statistical analysis showed significant differences between the perceptions of agents and recipients perceptions of the agents’
leadership behaviour. However, the results did not provide support for the relationship between leadership behaviour and resistance to change. The relationship between framing change and resistance to change became significant when framing change was added to the model. Moreover, framing change asymmetry significantly moderated this relationship. In contrast, no support was found for the moderating effects of creating capacity and shaping behaviour. Because of these contrasting results, future research is recommended.
Keywords: Organizational change, agent, recipient, resistance to change, leadership behaviour, framing
change, creating capacity, shaping behaviour, asymmetry.
3
INTRODUCTION
Organizational change became increasingly important since organizations have to adapt to the dynamic environment they operate in (Nging & Yazdanifard, 2015). Despite the focus on change management in the past decades, it remains a challenge to successfully implement change (Wanberg & Banas).
Organizational change is influenced by change recipients’ attitudes to change (Agote, Aramburu
& Lines, 2016; Greaz & Smith, 2010). One of these attitudes is resistance to change. Resistance to change is often defined as ‘the intentional or behavioural component as a driving force behind maintaining the status quo, and hindering successful implementation of change’ (Bouckenooghe, 2010:504). When recipients resist organizational change, the change process will be delayed and obstructed (Ansoff, 1990).
To date, there is a lack of empirical research that focuses on the role and behaviours of leaders in a change context (Eisenbach, Watson, & Pillai, 1999). Higgs and Rowland (2005; 2011) contributed to the existing literature by investigating leadership behaviours in a change context. They identified three different leadership behaviours in a change context. First, framing change is characterized as
‘establishing starting points for change (Higgs & Rowland, 2011: 312, second, creating capacity is defined as ‘creating individual and organizational capabilities’ (Higgs & Rowland, 2011:312), and third, shaping behaviour is characterized as ‘the communication and actions of leaders related directly to the change (Higgs & Rowland, 2011:312). However, they did not study these leadership behaviours in relation to resistance to change.
Several authors argue that change agents contribute to recipients’ resistance to change (e.g. Ford, Ford & D'Amelio, 2008). However, it remains unclear how leadership behaviour of change agents influences recipients’ resistance to change. Therefore, this study investigates the influence of change agents’ leadership behaviour on recipients’ resistance to change.
Furthermore, research often focused solely on the change agent or on the change recipient,
instead of these two actors together (Sonenshein, 2010). It is important to include both a change agent
and change recipient perspective, since agents and recipients can have different perceptions of a change
process (Bartunek, Rousseau, Rudolph & DePalma, 2006). Therefore, this study also investigates the
4
moderating role of this difference in perspectives. This difference between the perceptions of the change agent and change recipient regarding the change agents’ leadership style will be labelled as leadership behaviour asymmetry.
Agents and recipients can have different perceptions of change interventions (Bartunek, et al., 2006; Atkins & Wood). This is interesting, since the perceptions agents and recipients have of each other influence their interaction (Williams, Spiro, & Fine, 1990). To date, it remains unclear how these differences in perceptions effect the interaction between change agents and recipients, since research often focused solely on the change agent or on the change recipient, instead of these two actors together (Sonenshein, 2010).
In this study, the difference between the perceptions of the agent and the recipient regarding the agents’ leadership behaviour is labelled asymmetry. The moderating role of leadership behaviour asymmetry on the relationship between leadership behaviour (framing change, creating capacity, and shaping behaviour respectively) and resistance to change will be investigated.
As mentioned above, this study focuses on three styles of leadership behaviour, which provides the opportunity to measure the differences between the perceptions of agents and recipients regarding these three styles. Because of that, a distinction was made between framing change asymmetry, creating capacity asymmetry, and shaping behaviour asymmetry.
Therefore, the aim of this study is to investigate the influence of change agents’ leadership behaviour on change recipients’ resistance to change, and the moderating role of leadership behaviour asymmetry.
This study contributes to existing literature by studying the effects of the leadership behaviours
by Higgs and Rowland (2005; 2011) on resistance to change, and provides better understanding of the
influence agents’ leadership behaviour has on recipients’ resistance to change. This can help change
agents to identify and respond to recipients’ resistance to change (Sonenshein, 2010; Klonek, Lehmann-
Willenbrock & Kauffeld, 2014). Moreover, this study contributes to the literature by investigating how
different perceptions of agents and recipients influence the relationship between relationship between
leadership behaviour and resistance to change. The results of the study will provide interesting insights
5
that can help change agents to understand how both their behaviour and leadership behaviour asymmetry contribute to resistance to change of change recipients.
The remaining part of the study proceeds as follows: first, existing literature will be reviewed and hypotheses will be formulated. Second, the methodology of the study will be presented. Third, the results of the statistical analysis will be presented. Fourth, the results of the study will be discussed, as well as the theoretical contributions, managerial implications, limitations and suggestions for future research. And lastly, the study ends with the conclusion.
LITERATURE REVIEW Resistance to change
Organizational change involves moving from the known to the unknown, which can be very challenging (Wanberg & Banas, 2000). Therefore, resistance is a natural response to change. Organizational change leads to uncertainty and stress among employees (Callan, 1993; Terry & Jimmieson, 2003; Mintzberg
& Waters, 1985). Since recipients fear uncertainty and unpredictable outcomes, this often leads to resistance to change (Lee & Corbett, 2006; Jansen, 2004), which in turn can delay and obstruct organizational change processes (Ansoff, 1990).
Thomas and Hardy (2011) distinguish two different approaches to resistance to change, namely the demonizing and the celebrating of resistance to change. The demonizing approach to resistance to change is the most common approach. This approach views resistance to change as something that has to be avoided or has to be overcome (Oreg, 2006; Jermier, Knights & Nord, 1994; Nadler, 1993; Ansoff, 1990), since it is a reason for failure of organizational change (Armenakis, Harris & Mossholder, 1993).
According to Bouckenooghe (2010:504), most definitions of resistance to change include ‘the intentional or behavioural component as a driving force behind maintaining the status quo, and hindering successful implementation of change’. Furthermore, change agents often use resistance to accuse change recipients for disappointing change results (Krantz, 1999).
The celebrating approach to resistance to change on the other hand views resistance as a valuable
attitude in a change process (Thomas & Hardy, 2011; Ford et al., 2008; Piderit, 2000). For instance,
Ford et al. (2008) argue that the reactions of recipients can be important for the existence, engagement,
6
and strength of a change (Ford et al., 2008). Resistance to change can be valuable in a change process since it keeps the conversation going (Barrett, Thomas & Hocevar, 1995). This gives change agents the chance to clarify and further legitimize the change, which contributes to giving recipients a better understanding of the change. Consequently, this may increase the acceptance of the change (Barrett et al., 1995). Moreover, in this approach resistance is perceived as a driver of organizational learning (Msweli-Mbanga & Potwana, 2006) and improves the quality of decisions in a change process. As a result, recipients are more committed to those decisions and their implementations (Amason, 1996). In addition, this approach perceives resistance to change a sign of engagement. Recipients that are highly dedicated to the success of an organization, are more likely to resist a change compared to people that are less dedicated to the success of an organization (Ford et al., 2008). Others state that it can be a driver of organizational learning (Msweli-Mbanga & Potwana, 2006) and should therefore not be avoided.
The demonizing and celebrating approach are contrasting approaches, but do have something in common. Both demonizing and celebrating approaches make a distinction between change agents and change recipients, and both approaches focus on the interests and assumptions of the change agent (Thomas & Hardy, 2011). The change agent ‘decides what constitutes resistance, who can be resistant, and how resistance should be dealt with’ (Thomas & Hardy, 2011:325). They argue that change agents should not be the only actors who decide what resistance to change consists of. Instead, resistance to change should be viewed from multiple perspectives. Jermier et al. (1994) share this view. In addition, they state that the interests of change agents should not outweigh the interests of change recipients when assessing resistance to change.
Several researchers supporting the celebrating approach argue that resistance to change is
influenced by agents. For instance, Ford et al. (2008) argue that resistance to change is a result of change
agents’ sensemaking. They argue that agents who expect recipients to resist, will make sense of
situations in a way that confirms their expectations. So when change agents expect recipients to resist
change, change agents will behave in a manner that causes those expectations to be confirmed (Kanter,
Stein & Jick, 1992). Another way change agents sensemaking influences resistance to change is by a
self-serving bias (Ford et al., 2008). Then, change agents take credit for positive outcomes, and blame
others (recipients) for negative outcomes. Moreover, change agents can contribute to resistance by
7
breaking agreements and violating trust, communication breakdowns, and by resisting resistance (Ford et al., 2008). Thus, change agents’ sensemaking influences resistance to change. Another factor that influences resistance to change is leadership behaviour (e.g. Nemanich & Keller, 2007). Since resistance to change can negatively influence the change process (e.g. Ansoff, 1990), it is important to study how change agents’ behaviour influences recipients resistance. This will be discussed in the following section.
Leadership behaviour
Despite the considerable amount of literature that has been published on leadership behaviour (e.g. Bass, 1995), only a few studies have investigated leadership behaviours in relation to organizational change processes. For instance, the studies by Higgs and Rowland (2000; 2001; 2005; 2011) investigated leadership behaviours connected to organizational change. Higgs and Rowland (2005) identified three sets of leadership behaviour in a change context. First, shaping behaviour is characterized as ‘the communication and actions of leaders related directly to the change’ (Higgs & Rowland, 2011:312). Of the three sets of leadership behaviours, shaping behaviour is the most leader-centric (Higgs & Rowland, 2005), which means that the individual leader is the focal point of the change process. This type of leaders manage the change process through personal involvement, persuasion and influence. Second, framing change is characterized as ‘establishing starting points for change’ (Higgs & Rowland, 2011:312). These leaders design and manage the change process and communicate guidelines for the organization. Third, creating capacity is characterized as ‘creating individual and organizational capabilities and communication and making connections’ (Higgs & Rowland, 2011:312).
In their case study, Higgs and Rowland (2011; 2005) found that shaping behaviour was a less successful leadership behaviour compared to framing change and creating capacity. Moreover, they found that the success of a change initiative is inhibited by shaping behaviour. However, Higgs and Rowland (2005; 2011) did not study these leadership behaviours in relation to resistance to change.
Therefore, leadership styles with comparable characteristics will be used to formulate hypotheses.
A leadership style with similar characteristics as framing change and creating capacity that has
been investigated in relation to resistance to change is transformational leadership. Transformational
8
leaders appeal to followers’ sense of values and inspire their followers to get and pursue a higher vision (Burns, 1978). These leaders are known to inspire their followers and make their followers advance a higher level of motivation and morality (Burns, 1978). Furthermore, transformational leaders are perceived as change drivers since they create an organizational culture that encourages growth and change (Avolio & Bass, 1995). They move their followers through idealized influence, inspiration, intellectual stimulation or individual consideration (Bass, 1999). Idealized influence and inspiration are comparable to framing change, whereas individual consideration is comparable to creating capacity.
Since transformational leadership is comparable to both framing and creating capacity, it is expected that both framing behaviour and creating capacity influence resistance to change in a similar manner.
Several studies examined the relationship between transformational leadership and resistance to change.
Nemanich and Keller (2007) found that transformational leadership significantly reduced employees’
resistance to change. Moreover, Oreg and Berson (2011) found that transformational leadership negatively influenced employees’ intentions to resist an organizational change.
In addition, framing behaviour and creating capacity have similar characteristics as participative leadership. Participative leaders encourage employees to discover new opportunities and challenges (Druskat & Wheeler, 2003) and stimulate them to learn through acquiring, sharing, and combining knowledge (Edmonson, 1999). Furthermore, participative leadership is associated with consensus, consultation, delegation and involvement (Bass, 1981). This leadership style increases employee involvement, commitment and loyalty to the organization (Bass, 1981). Moreover, they take the opinions and suggestions of subordinates into account before making a decision (Mullins, 2005) and because of that, employees are more committed to those decisions and less likely to resist (Peccei, Gianreco & Sebastiano, 2011).
Based on the studies mentioned above, it is expected that change agents’ framing change and creating capacity behaviours will have a lowering influence on recipients’ resistance to change.
Hypothesis 1: An increase in change agents’ framing change behaviour will lead to a decrease in change
recipients’ resistance to change.
9
Hypothesis 2: An increase in change agents’ creating capacity will lead to a decrease in change recipients’ resistance to change.
Shaping behaviour has similar characteristics as directive leadership. Directive leadership is characterized as autocratic, task-oriented, persuasive and manipulative (Muczyk & Reimann, 1987;
Yukl, 1989; Bass, 1981), and is negatively related to the acceptance of managerial decisions, employee involvement, commitment and loyalty (Bass, 1981). Therefore, it is likely that shaping behaviour positively influences resistance to change, which leads to the following hypothesis:
Hypothesis 3: An increase in change agents’ shaping behaviour will lead to an increase in change recipients’ resistance to change.
The moderating effect of leadership behaviour asymmetry
Interactions between change agents and recipients depend on the perceptions they have of each other (Williams, Spiro & Fine, 1990). Therefore, this study includes both agent and recipient perspectives regarding the agents’ leadership behaviour.
Several authors emphasize the differences between the perceptions of change agents and recipients regarding change interventions (e.g. Bartunek et al., 2006; Atkins & Wood, 2002). For instance, Bartunek et al. (2006) argue that change agents and recipients can have different perceptions of an intervention. This is because each actor assigns different meanings and values to seemingly mutually positive initiatives (Bartunek et al., 2006; Moch & Bartunek, 1990). These differences even exist between middle and higher managers (Balogun & Johnson, 2004). Moreover, managers tend to overestimate their own competences (e.g. Atkins & Wood, 2002; Herbst, Maree & Sibanda, 2006).
Several studies investigated the differences between the perceptions of agents and recipients.
For instance, Herbst and Conradie (2011) investigated managers’ self-perceptions and the perceptions
of others regarding managers’ leadership effectiveness. They defined the difference between the ratings
of others and managers’ own ratings as self-perception accuracy. They found a significant difference
between the perceptions of the managers and the perceptions of others regarding the managers’
10
leadership effectiveness. Moreover, they argue that self-perception accuracy is a prerequisite for effective organizational change. Other authors defined the difference between the perceptions of agents and recipients as leaders’ self-awareness. Leaders’ self-awareness improves the mechanisms required for organizational change (Whetten & Quinn, 1994). The findings of Fletcher (1997) provide an explanation for this, since it was found that self-awareness was positively associated with managerial competences. Moreover, self-awareness is also positively associated with individual and organizational performance (Moshavi, Brown & Dodd, 2003; Church, 1997).
A limited amount of studies investigated the influence of the differences between the perceptions of agents and recipients in a context of organizational change. For instance, Atwater and Yammarino (1997) studied the influence of self-other rating agreement in relation to recipients’ attitudes to change. They argue that self-other rating agreement influenced organizational outcomes such as the attitudes of their followers. Moreover, Braver (1995) argues that leaders’ self-awareness negatively influences recipients’ resistance to change. Therefore, this study will examine the moderating effect of the difference between the perceptions of change agents and recipients. This difference in perceptions will be labelled leadership behaviour asymmetry. Leadership behaviour asymmetry reflects the difference between the perceptions of the change agent and recipient regarding the agents’ leadership behaviour. It is expected that high leadership behaviour asymmetry leads to more resistance to change, since the perceptions agent and recipient are not in agreement with each other, whereas low leadership behaviour asymmetry leads to less resistance to change.
In this study, three types of leadership behaviour are investigated, namely framing change, creating capacity and shaping behaviour. Therefore, three corresponding types of leadership behaviour asymmetry are studied. The three types of leadership behaviour asymmetry that will be investigated are framing change asymmetry, creating capacity asymmetry, and shaping behaviour asymmetry.
As mentioned before, hypothesis 1 expected that an increase in framing change leads to a
decrease in resistance to change. It is expected that high framing change asymmetry will lead to more
resistance, whereas low framing change asymmetry will lead to less resistance, thus it is expected that
framing change asymmetry positively influence the relationship between framing change and resistance
11
to change. Hypothesis 2 predicts that an increase in creating capacity will lead to a decrease in resistance to change. Since high asymmetry is expected to lead to more resistance to change, whereas low asymmetry is expected to lead to less resistance, it is predicted that creating capacity asymmetry strengthens the negative relationship between creating capacity and resistance to change. Moreover, hypothesis 3 expects that shaping behaviour positively influences resistance to change. It is predicted that high shaping behaviour asymmetry strengthens this relationship. This leads to the following hypotheses:
Hypothesis 4: The relationship between framing change and resistance to change will be negatively moderated by framing change asymmetry.
Hypothesis 5: The relationship between creating capacity and resistance to change will be negatively moderated by creating capacity asymmetry.
Hypothesis 6: The relationship between shaping behaviour and resistance to change will be positively moderated by shaping behaviour asymmetry.
The conceptual model is presented below.
FIGURE 1: conceptual model
12
METHODOLOGY
Sample and procedure
A field study was conducted to identify the effects of change agents’ leadership behaviour on recipients’
resistance to change and the moderating role of leadership behaviour asymmetry on this relationship.
Change agents and recipients who had recently been involved in a change project participated in the study. Recipients were linked to their corresponding agent in order to investigate change leadership asymmetry.
In this study, both primary and secondary data were used. Secondary data from 96 organizations were complemented with primary data from 8 other organizations following the same procedures. The secondary data consisted of 117 agents, 361 recipients and 4 middle managers. For 2 respondents it was unclear whether they were agents or recipients. The primary data consisted of 8 agents and 55 recipients.
In total, 18 respondents from the secondary data were removed from the sample. First, 15 respondents from four teams in one organization were removed. From this organization, four top managers filled in the agent version of the questionnaire, whereas four middle managers and seven recipients filled in the recipient version of the questionnaire. However, it was unclear whether the top managers answered the questions about the recipients with the middle managers or the recipients in mind. Moreover, it was unclear whether the recipients answered the questions about the leadership style of their agent with the middle manager or the top manager in mind. In this study, recipients had to be assigned to the corresponding change agent in order to measure leadership behaviour asymmetry. Since it could not be guaranteed that each recipient was assigned to the right agent and vice versa, it was decided to remove all respondents from this organization. Moreover, two respondents from two different organizations were removed because they answered ‘4’ which corresponds to ‘neutral’ to all questions measured on a Likert scale. The last respondent that was removed did only partially answer the questionnaire.
The sample used for the analysis consisted of 121 agents and 407 corresponding recipients. In
total, 121 teams from 103 organizations participated in the study. From the 121 teams, the average
13
number of recipients per agents was 3 to 4 (SD = 1.85). The smallest number of recipients per agents was 1, and the largest number of recipients per agent was 11.
Of the 407 recipients, 56,3 percent were male and 43,7 percent were female. The average age of the recipients was 39 years (SD = 12.26). All recipients were Dutch. Primary education was the highest level of education for 9,6 percent of the recipients, secondary and / or intermediate vocational education for 33,4 percent of the recipients, and 54,5 percent of the recipients completed polytechnic and / or university. The remaining 2,5 percent had a different kind of education. On average, the recipients worked for 7,79 years in their current position (SD = 7,47), and had a total work experience of 17,47 years (SD = 11,35).
Of the 121 agents, 71,1 percent were male and 28.9 percent were female. The average age of the agents was 43,39 years (SD = 10.61). All agents were Dutch. Primary education was the highest level of education for 2,5 percent of the agents, secondary and / or intermediate vocational education for 14,0 percent of the agents, and 82,6 percent of the agents completed higher vocational training or held a university degree. The remaining 0,8 percent had a different kind of education. On average, the agents worked for 7,70 years in their current position (SD = 7,19), and had a total work experience of 18,57 years (SD = 10,75).
Both convenience sampling and snowballing were used to identify potential respondents. In most cases, the initial contact was established with the change agent. Some change agents chose some of their team members to participate in the study, while others asked all team members to participate.
Each participant received an e-mail with a description of the project, a unique code and the link to the
survey. Every change agent was asked to give a short description of the chosen change project. This
description was sent to the recipients by e-mail to make sure that both the agent and recipients thought
of the same project when answering the questionnaire. Moreover, to assign the recipients to the correct
change agent, both change agents and recipients received a unique code. In the questionnaire, each
respondent had to fill in his or her code.
14 Measures
Two existing surveys were used to measure the constructs from both the change agent and change recipient perspective. One survey was used to measure the constructs from the change agent perspective, and the other survey was used to measure the constructs from the perspective of the change recipient.
Exploratory factor analyses were performed to examine whether the questions from the questionnaire empirically reflected the intended constructs. Items that correlated with multiple constructs, correlated to wrong constructs or did not correlate with any construct were deleted. Separate analyses were performed for the recipient data and agent data. The used rotation method was Oblimin with Kaizer normalization. Coefficients with an absolute value below .40 were suppressed.
The factor analyses did not identify the subcategories of the ‘framcap’ behaviours, but did identify framing change, creating capacity, shaping behaviour, and resistance to change. A detailed description of the followed procedures and the outcomes of the factor analyses are presented in Appendix A and B The constructs that were identified in the factor analyses and their measurements are described below.
Resistance to change. Resistance to change can be viewed as a multidimensional attitude
towards change, which consists of affective, cognitive, and behavioural components (Piderit, 2000).
These three components depend on each other but are not necessary in line, since some sources of resistance have a stronger impact on recipients’ feelings, while others have stronger impact on recipients’ thoughts or behaviours (Oreg, 2006). Next to the recipients that had to answer questions regarding their own resistance to change, change agents had to answer questions about the resistance to change of the recipients as well. Since it is problematic to answer questions about the feelings and thoughts of someone else, this study focuses on the behavioural component of resistance to change. The behavioural component of resistance to change ‘involves actions or intentions to act in response to the change’ (Oreg, 2006:76).
To measure resistance to change, three items from the scale of Oreg (2006) were used. These
items are presented in appendix B. An item of the scale was ‘During the change process the employees
15
sought ways to obstruct the change’ for the agent version of the questionnaire, and ‘During the change process I sought for ways to obstruct the change from happening’ for the recipient version of the questionnaire. A seven-point Likert scale was used to measure resistance to change, where 1 represented
‘strongly disagree’ and 7 represented ‘strongly agree’. Reliability analysis on the 3 questions measuring resistance to change showed that the three questions together had a Cronbach’s alpha of α = .77 for the recipient version and α = .78 for the agent version.
Leadership behaviour. In their study, Higgs and Rowland (2005) made a distinction between three
leadership behaviours. These three leadership behaviours are shaping behaviour, framing change, and creating capacity. In another study, Higgs and Rowland (2011) created a coding frame that appeared to capture behaviours exhibited by leaders combining framing and creating. This new category was called
‘framcap’ and consisted of four subcategories. These subcategories are attractor, edge and tension, container, and transforming space. The survey included the category shaping behaviour, and the four subcategories attractor, edge and tension, container, and transforming space. The factor analyses identified the three main categories, shaping behaviour, framing change, and creating capacity, but failed to identify the four subcategories. A detailed description of the factor analyses is presented in the appendix A.
Framing change. Framing change was measured by five items. One item of the scale was
‘During the change process I set clear rules and boundaries so that the employees knew where they stood’ for the agent version, and ‘The change agent set clear rules and boundaries so that the employees knew where thy stood’ for the recipient version of the survey. The other items measuring framing change are presented in appendix B. All items were measured on a seven-point Likert scale and were based on Higgs and Rowland (2008). The Cronbach’s alpha of the recipient version was α = .82, and for the agent version Cronbach’s alpha was α = .73.
Creating capacity. Creating capacity was measured by six items based on Higgs and Rowland
(2008). An example of an item measuring creating capacity is ‘During the change process I organized
discussions with the employees to come up with different solutions’ for the agent version of the survey,
and ‘The change agent organized discussions with the employees to come up with different solutions’
16
for the recipient version of the survey. The remaining five items measuring shaping behaviour are presented in appendix B. These items were also measured on a seven-point Likert scale. This scale had a Cronbach’s alpha of α = .90 for the recipient version, and α = .87 for the agent version.
Shaping behaviour. To measure shaping behaviour, five items were used, based on Higgs and
Rowland (2008). One item of the shaping behaviour scale was ‘During the change process I pointed the employees at their responsibilities regarding their role in the change’ for the agent version of the questionnaire, and ‘The change agent pointed the employees at their responsibilities regarding their role in the change.’ for the recipient version of the questionnaire. The remaining items measuring shaping behaviour are presented in appendix B. Shaping behaviour was measured on a seven-point Likert scale.
The Cronbach’s alpha of this scale was α = .87 for the recipient version, and α = .82 for the agent version, indicating that the internal consistency of the scale was good.
Leadership behaviour asymmetry. Framing change asymmetry was calculated by subtracting
the individual agents’ framing change score from the individual recipients’ framing change score. This
lead to both positive and negative scores. A positive score meant that the recipient perceived the framing
behaviour of the agent more framing than the agent perceived his or her own behaviour framing. A
negative score meant that the agent perceived his or her own behaviour as more framing than the
recipient perceived the behaviour of the agent as framing. The aim of this study was to investigate
whether small amount of asymmetry had a different effect on the relationship between leadership
behaviour and resistance to change than a large amount of asymmetry, without taking the direction of
asymmetry into account. Therefore, all scores were transformed to positive scores. A mean-split was
performed to make a distinction between asymmetry scores higher than the average amount of
asymmetry, and lower than the average amount of asymmetry. All framing change asymmetry scores
below average were coded as 0, while all scores above average were coded as 1. In this way, two distinct
categories were created, namely low and high asymmetry. Creating capacity asymmetry and shaping
behaviour asymmetry were calculated in the same manner.
17
Control variables. Control variables rule out alternative explanations for the findings (Schmitt,
Klimoski, Ferris & Rowland, 1991), and can be used to increase statistical power of the findings (Schwab, 1999). Therefore, this study includes gender and age as control variables. Besides questions regarding agents’ leadership behaviour and recipients’ resistance to change, the questionnaire also included questions regarding respondents’ gender and age.
RESULTS
Correlations and associations between the variables
A correlation analysis was performed to see whether the variables were associated with each other or not. The means, standard deviations, and correlations between the variables are presented in table 1. As expected, resistance to change is negatively correlated with framing change (r = -.14, p < .01). This indicates that resistance to change increases when framing change decreases and vice-versa (Taylor, 1990). Resistance to change is also negatively correlated with creating capacity (r = -.14 , p < .01), indicating that resistance to change and creating capacity move in opposite directions (Taylor, 1990).
Surprisingly, resistance to change was not significantly associated with shaping behaviour. Shaping behaviour, framing change, and creating capacity are positively associated with each other, indicating that the variables move in the same direction (Taylor, 1990). Moreover, shaping behaviour asymmetry, framing change asymmetry, and creating capacity asymmetry positively correlate with each other, and are also positively associated with shaping behaviour, framing change, and creating capacity. Age and gender do not significantly correlate to any of the variables.
Unpaired t-test
Unpaired t-tests were performed to investigate whether there were significant differences between the perceptions of the agents and the recipients. For framing change, a significant difference was found between the means of the agent (M=5.41, SD=.79) and the recipients (M=4.75, SD=1.05)
[t(407)=10.13, p <.001). For creating capacity, a significant difference was also found between the
means of the agent (M=5.28, SD=.99), and the recipients (M=4.44, SD=1.22) [t(407)=10.70, p <.001].
18
For shaping behaviour, a significant difference was also found between the means of the agent
(M=5.60, SD = .76) and the recipients (M=4.86, SD = 1.20) [t(407)=10.39, p < .001].
TABLE 1: correlations and associations between constructs
M SD 1. 2. 3. 4. 5. 6. 7. 8. 9.
1. Gender
11.44 .50 -
2. Age 39.38 12.13 -.08 -
3. Resistance To Change 2.77 1.43 -.08 .09 -
4. Shaping Behaviour (MC) .000016 1.20 .00 .00 -.10 -
5. Framing Change (MC) .000037 1.05 .04 -.04 -.14** .60** -
6. Creating Capacity (MC) .000008 1.22 .02 .00 -.14** .71** .58** -
7. Shaping Behaviour Asymmetry
2-.34 .90 .00 -.01 -.08 .85** .51** .58** -
8. Framing Change Asymmetry
2-.21 0.79 .03 -.06 -.05 .52** .80** .50** .53** -
9. Creating Capacity Asymmetry
2-.28 .89 .04 -.07 -.09 .57** .48** .80** .57** .51** -
Note. N = 407.
1
Dummy coded, 1 = male, 2 = female
2
Dummy coded, 0 = asymmetry below mean, 1 = asymmetry above mean
*** p <.001, ** p <.01, * p <.05
20 Hypotheses tests
A hierarchical linear regression analysis was performed to test the hypotheses. Recipients’ gender and age were used as control variables, and therefore added to the model in the first step. In the second step, the independent variables were added to the model. These are shaping behaviour, framing change, and creating capacity, and were mean-centered to reduce the likelihood of multi collinearity (Aiken & West, 1991). In the third step, the interactions between shaping behaviour and shaping behaviour asymmetry, framing change and framing change asymmetry, and creating capacity and creating capacity asymmetry were added.
The results of the analysis are presented in table 2. Hypothesis 1 predicted that an increase in change agents’ framing behaviour will lead to a decrease in change recipients’ resistance to change. A regression analysis was performed to test the hypothesis. The results of the regression analysis were not significant (B = -.13, SE = .09, p = .13). Accordingly, hypothesis 1 was not supported. Hypothesis 2 predicted that an increase in change agents’ creating capacity will lead to a decrease in change recipients’
resistance to change. The results of the regression analysis were not significant (B = -.12, SE = .09, p = .17). Therefore, hypothesis 2 was also not supported. Hypothesis 3 predicted that an increase in change agents’ shaping behaviour will lead to an increase in change recipients’ resistance to change. The results of the regression analysis were not significant (B = .04, SE = .09, p = .66), thus, no support was found for hypothesis 3. Hypothesis 4 predicted that framing change asymmetry moderates the relationship between framing change and resistance to change. The results of the regression analysis were significant (B = .39, SE = .16, p = .02), indicating that framing change asymmetry moderates the relationship between framing change and resistance to change. Thus, hypothesis 4 was supported. In addition to that, the relationship between framing change and resistance to change also becomes significant when adding framing change asymmetry to the model (B = -.35, SE = .13, p < .01), indicating that an increase in framing change leads to a decrease in resistance to change, when asymmetry is taken into account.
Hypothesis 5 predicted that creating capacity asymmetry moderates the relationship between creating capacity and resistance to change. The results of the regression analysis were not significant (B=.04, SE
= .15, p =.78). Thus, no support was found for hypothesis 5. Hypothesis 6 predicted that shaping
21
behaviour asymmetry moderates the relationship between shaping behaviour and resistance to change.
The results of the regression analysis were not significant (B = -.13, SE = .17, p = .46). Accordingly, no support was found for hypothesis 6.
TABLE 2: regression analysis
RESISTANCE TO CHANGE
M1 M2 M3
Step 1
Gender
1-.16 -.12 -.10
Age .00 .00 -.00
Step 2
Shaping Behaviour .04 .12
Framing Change -.13 -.35**
Creating Capacity -.12 -.16
Step 3
Shaping Behaviour Asymmetry
2-.13
Framing Change Asymmetry
2.38*
Creating Capacity Asymmetry
2.04
Total R
2(ΔR
2) .00 .03 (.02*) .04 (.01)
Notes. N =407; Unstandardized regression coefficients are displayed.
1
Dummy coded, 1 = male, 2 = female
2