THE INFLUENCE OF LEADERSHIP BEHAVIOR ON EFFECTIVE
CHANGE AND THE ROLE OF ARX ON THIS RELATION
A study into the perspectives of both change agents and change recipients
Author: N. H. Perton Student number: 2566435
University of Groningen Faculty of Economics and Business Master thesis BA, Change Management
Koeriersterweg 69 9727 AE Groningen 06 - 49 72 26 71 n.h.perton@student.rug.nl Supervisor: Dr. C. Reezigt Co-assessor: Dr. J.F.J. Vos May 12th, 2016
ABSTRACT
Many organizational change initiatives fail. Change leaders’ behavior has been suggested to enhance change effectiveness. Agent-recipient exchange (ARX) is identified as a moderator on the relation between leadership behavior and change effectiveness. This paper deals with the different perceptions of this relation of both change agents and change recipients in change situations and whether the interaction of ARX has different effects for the change recipient and the change agent. This study contributes to literature on ARX, change leadership behavior and change effectiveness. Results provide evidence of the simultaneous existence of three leadership behaviors. Shaping and framing behavior were found to have a significant positive effect on change effectiveness from both perspectives. Creating behavior has only a positive influencing effect on change effectiveness from a change recipients’ perspective. Furthermore it was found that ARX negatively moderates the relation between framing behavior and change effectiveness from a change recipients’ perspective. The effect of the relation between all leadership behaviors and change effectives is stronger for change recipients than for change agents. These results support the notion that investigating change from two perspectives is of great importance. No evidence was found of the differences between change agent’s and change recipient’s perspectives on the moderation effect of ARX on leadership behaviors and change effectiveness. Future researchers are urged to investigate the possibility for a balanced set of behaviors which fits both the change agent’s and the change recipient’s perspective.
Keywords: Organizational change, Perspectives, Agent-recipient exchange, Change leadership
ACKNOWLEDGEMENT
CONTENT
INTRODUCTION ... 5
Literature review ... 7
Change leadership behavior... 7
Agent-recipient exchange ... 9
Different perspectives of change agents and change recipients ... 10
Effective change ... 11
HYPOTHESIS DEVELOPMENT... 13
The influence of leadership behavior on change effectiveness ... 13
The moderating role of ARX on leadership behavior and change effectiveness ... 14
The different perspective of change agents and change recipients and the moderating effect of ARX ... 15
METHODS ... 16
Sample and procedure ... 16
Measures ... 18
Change agent leadership behavior. ... 20
Agent-recipient exchange. ... 20
Perceived change effectiveness ... 20
Control variables ... 21
Data analysis ... 22
RESULTS ... 25
Correlations from the change agent’s perspective ... 25
Hypotheses testing change agent’s perspective ... 25
Correlations from the change recipient’s perspective ... 30
Hypotheses testing change recipient’s perspective ... 30
Comparing the change agent’s view to the change recipient’s view ... 35
DISCUSSION ... 41
Discussion of the findings ... 41
Theoretical contributions ... 44
Practical implications ... 44
Limitations and future research ... 45
REFERENCES ... 46
APPENDICES ... 53
INTRODUCTION
Leading a change to success is often more difficult than most leaders think. Literature reports failure in up to ninety percent of all change initiatives (Crosby, 1979; Rogers, Meehan, & Tanner, 2006). Although some researchers claim that this rate is exaggerated and that it is not clear how these percentages are measured (Burnes, 2011), these results do point out the serious nature of the matter. Failing change initiatives cost organizations much money, time and effort. Not every failed change can be traced back to organization’s handling, but most of them have human aspects which can be influenced by the organization. Therefore, it is not surprising that ingredients for successful change are subject to much academic literature (Al-Haddad & Kotnour, 2015). Burnes (2011) addresses researchers to foresee what organizations can do to improve their success rate during organizational change. It can be of great value to managers to know antecedents of effective change. Moreover, research into this topic can reduce the huge costs of failing change initiatives.
Researchers have found evidence that leaders have a significant effect on the success of change (Eisenbach, Watson, & Pillai, 1999; Gill, 2003; Higgs & Rowland, 2011; Mehta, Maheshwari, & Sharma, 2014). Even incremental change requires leadership behavior in order to make it happen (Mehta et al., 2014). More specific, a growing evidence indicates that specific leadership behaviors influence the success of organizational change initiatives (Eisenbach et al., 1999; Higgs, & Rowland, 2000, 2005, 2011). A leader’s behavior has a high influence on for example: firm performance (Wang, Tsui, & Xin, 2011), achieving subordinates to move beyond their normal tasks (Bass, 1985), influence subordinates behavior (Oreg & Berson, 2011), but also in successful change (Higgs & Rowland, 2005, 2011).
beneficial in change situations (Bernerth, Armenakis, Field, Giles, & Walker, 2007; Self, Armenakis, & Schraeder, 2007). Rogulic (2013) recognized this and developed agent-recipient exchange (ARX). In which, leaders who are guiding in a change are called change agents and their subordinates are called change recipients. The change agent and change recipient hold a dyadic relation during times of change. Rogulic (2013) found that if the relationship is of high quality, than ARX influences the effectiveness of change initiatives positively. The influence of ARX as a moderator of the relation between leadership behavior and effective change could be an important variable in the search for effective ways to change.
Many researchers claim that, in order to define the quality of a relationship between change agent and change recipient, both perspectives have to be consulted (e.g. Gerstner & Day, 1997; Liden, Sparrowe, & Wayne, 1997; Scandura & Schriesheim, 1994; Gils, Quaquebeke, & Knippenberg, 2010). It is known that change recipients have different perceptions of change related matters and experience relationships in different ways than change agent do (Vos, Rupert, & Eseryel, 2015). In that line of reasoning leadership behavior and effective change cannot be observed objectively and have to be viewed from both change agent’s and change recipient’s perspective (Higgs & Rowland, 2005; Gils et al., 2010). This is in agreement with the suggestion of Herold, Fedor, Caldwell and Yi (2008), who state that future research should capture leadership as an individual-level variable, varying as a function of the leader’s behavior toward a particular follower. Furthermore, it is in consonance with the suggestion of Higgs and Rowland (2005) who state that future research should use both the change agent’s and change recipient’s views on change.
To the researcher’s knowledge, no previous research exists which investigates the effect of ARX as a moderator on the relation between change agents leadership behavior and change effectiveness. Based on the above stated argumentation, the following research question has been formulated: How does ARX moderate the relationship between leadership behavior and change effectiveness and do change agents and change recipients perceive this relation and the effect of the moderator differently?
effect for the change recipient than the change agent. This research contributes to the work of Higgs and Rowland (2005, 2011) by performing a quantitative research on change leader’s behavior. Overall it aims to give organizations better insight in how to change successfully.
LITERATURE REVIEW
Change leadership behavior
There has been a lot of research on the subject of leadership. Much research deals with leadership in general rather than in change situations. However, specific literature in change situations is growing. Herold et al. (2008) mention that specific change-oriented leadership does not come from individual traits, but focuses on the trait and involvement of employees during a change. Leaders who execute change leadership formulate a vision for the change at hand (Herold et al., 2008). To improve the understanding and ownership of the change they involve subordinates in the change process, thereby aiming to improve their motivation to enact the change (Herold et al., 2008). Change leadership is a gathering of used behaviors which are specified to the context of change. Higgs and Rowland (2000, 2005, 2011) conceptualized in three papers the behavior of the leader during times of change. They researched the impact of leader’s behaviors on the successful implementation of change and found that a leader’s behavior in a change process has a significant effect on the success of change (Higgs & Rowland, 2005, 2011). In their 2005 paper, they categorized three types of leadership behavior during organizational change; shaping behavior, framing behavior and creating behavior. Although these three leadership behavior types have been described separately, Higgs and Rowland (2005) found that leaders can use those behaviors interchangeably.
(Higgs & Rowland, 2005). It is about the journey rather than about what the change stands for (Eisenbach et al., 1999). The leader establishes a starting point for change and helps subordinates to get there (Higgs & Rowland, 2011). The leader challenges subordinates to deliver the change for instance by going beyond their own tasks (Kotter, 1995). In framing behavior, a change agent sets clear expectations, boundaries, and hard rules (Higgs & Rowland, 2011). The change agent communicates guiding principles in the organization (Higgs & Rowland, 2011). The last leadership behavior, creating behavior, is about giving people individual and organizational capabilities, physical, emotional and temporal space to act and think differently (Higgs & Rowland, 2011). The behavior consists of high communication and creating connections with the subordinates (Higgs & Rowland, 2005). It was found that persistent communication within the organization enhances the change orientation by fostering of shared attitudes and values (Schaap, 2006).
Later, Higgs and Rowland (2011) refined the three types of leadership behavior into four more specific categories: attractor, edge and tension, container, and transforming space. The old behaviors are partly covered by the new categories. Framing behavior can be seen as a combination of attractor, edge and tension, and container behavior (Higgs & Rowland, 2011). Creating behavior contains components of container and transforming space behavior (Higgs & Rowland, 2011). Shaping was not covered in the new categories.
A leader with attractor behavior pulls subordinates towards the change, not towards himself. When a leader uses edge and tension behavior, he increases the disturbance that is generated by the change. He tests and challenges the organization by giving the people insight in repeated and unhelpful behavioral patterns while he keeps the change process on course. When a leader shows container behavior, the leader provides calmness in complex and uncertain times of change. Through their behavior leaders provide subordinates a way to find sense and positive meaning during times of change. Lastly, transforming space, a leader communicates that this is the moment in which change has to happen. Change is only possible in the present moment. Higgs and Rowland (2011) found in their data analysis that all four behavioral components combine into one behavior, framcap.
researches, it is chosen to fall back on the earlier defined three categories in the paper of Higgs and Rowland (2005); shaping, framing and creating (Jansons, 2015; Rogulic, 2013; Vos et al., 2015).
The next sections pay attention to the relation between leader and subordinate.
Agent-recipient exchange
ARX is defined as “the perception held by the change agent/recipient as to what extent voluntary actions on one’s part will be returned by the other party in some way during the change initiative” (Rogulic, 2013, p. 11). The quality of ARX depends on social exchange within a dyadic relation. It is high when voluntary actions are returned or deemed to be returned by the other party. In order to put ARX in perspective, it is important to discuss the history and development of ARX. ARX has its foundation in the leader-member exchange (LMX) literature. By discussing the evolution of LMX, the development of ARX and how it functions becomes more clear. This evolution starts with leadership.
Many studies in current academic literature about leadership assume leader’s behavior and relations with leader’s subordinates is consistent across all subordinates (Schriesheim, Wu, & Cooper, 2011). However, some researchers believe there is a unique relationship between a leader and each subordinate (Bernerth et al., 2007; Schriesheim et al., 2011). Superiors develop leadership exchange with a select group of subordinates, with others they develop only superficial relations based on authority (Dansereau et al., 1975). According to early LMX literature there is a dyadic relationship between a leader and each individual follower (Dansereau et al., 1975). The quality of this relation affects different attitudinal and behavioral outcomes of subordinates (Gerstner & Day, 1997). A high quality LMX relationship enhances organizational citizenship behaviors of organizational members (Ilies, Nahrgang, & Morgeson, 2007). It also creates positive change process characteristics, like information sharing, participation, and trust in management (Dam, Oreg, & Schyns, 2008). High-quality relationships result in employees engaging in behaviors beyond their prescribed roles, this contributes to organizational effectiveness (Ilies et al., 2007).
based on vertical dyad linkage (VDL), in which subordinates had to prove their competence or trustworthiness before engaging in exchanges with their leader (Bernerth et al., 2007). Bernerth et al.’s (2007) leader-member social exchange (LMSX) theory stresses the importance of the reciprocity in social interaction. Social exchange is based on the thought that individuals try to balance the inputs and outputs of their social interactions (Blau, 1964). Bernerth et al.’s (2007) variant does not place restrictions on the leader-member relations as the previous literature did, in which exchange has to be specified beforehand.
A low quality LMSX is caused by distant and more formal interaction based on the transactional part of the employment contract (Bernerth et al., 2007). Relationships of high quality are based on mutual liking, trust, respect, and influence (Bernerth et al., 2007) and enhance subordinates performance (Kuvaas, Buch, Dysvik, & Haerem, 2012). After Bernerth et al. (2007) emphasized the usefulness of LMSX during times of change, Rogulic (2013) tailored the concept specifically to change situations. He conceptualized this adaption into agent-recipient exchange (ARX). Rogulic (2013) focused on the effects of agent-recipient dyadic relationship quality on change effectiveness and found evidence for this relation. Later, other studies into ARX have shown evidence of high quality ARX relations causing low resistance (e.g. Mellema, 2015).
The change agent and change recipient do not have to agree on the level of quality of ARX. This different perception on ARX, also referred to as the ARX gap, has been proven to influence change effectiveness (Rogulic, 2013). Therefore, in this research, the perceptions of change agent and change recipient will be dealt with separately.
Different perspectives of change agents and change recipients
recipient perceives it. Even though sensemaking is not in the scope of the current research, the former argument points out that sensemaking may well lead to different views on change (Thurlow & Mills, 2009). That is why both change agent’s and change recipient’s perspective of leader’s behavior is needed in order to get a comprehensive view on the leader’s behavior influence on change effectiveness.
The different perceptions of the relationship between leader and subordinates occur frequently in literature (Gerstner & Day, 1997; Schriesheim, Neider, & Scandura, 1998). One of the reasons why these different perceptions exist is because of the role change agent and change recipients hold (Zhou & Schriesheim, 2009). Change agents are responsible for establishing the urgency for change, specifying the desired outcome, forming a vision, and making the change happen (Ford et al., 2008; Higgs & Rowland, 2011). They are the leaders in the change project and establish narratives about the change (Sonenshein, 2010). Each change recipient makes sense of these narratives in their own way and with their own behavioral reaction (Sonenshein, 2010). Change recipients are responsible for adapting to the change and implementing it in their work (Kanter, Stein & Jick, 1992). It is likely that these different roles lead to different perspectives on change related matters (Vos et al., 2015).
It was found that change agents evaluatetheir own leadership behavior and the quality of their relationship with the subordinates higher than change recipients with their leader (Jansons, 2015). Furthermore, Rogulic (2013) found that there is a gap in the perceptions of ARX between change agent and change recipient. Several aspects which influence these different perceptions of LMX (also applicable for ARX) are identified, examples are the follower’s interaction with the leader, leader’s self enhancement and subordinates perception of leadership (Kam, Vegt, Janssen, & Stoker, 2015). Bernerth et al. (2007) focused in their research on LMX only on the view of the subordinates, it therefore lacks the different perspectives. ARX takes both change agent’s and change recipient’s perspective in order to assess the quality of the relation.
To understand the impact of the moderating effect of ARX on the relation between leadership behavior and change effectiveness is investigated, effective change has to be defined.
Effective change
or quantifying the concept of change effectiveness (By, 2005). That makes it hard to find a clear definition and measurement (Elving, 2005). In assessing the effectiveness of change, some researchers base their quantification on figures provided by the organization under investigation (e.g. Eitzen & Yetman, 1972; Higgs & Rowland, 2011). A personal view on change effectiveness, by examining the satisfaction of stakeholders with the change process and change output, will help researchers gaining a more realistic image of change effectiveness (Burnes & Jackson, 2011; Pinto, Slevin, and English, 2009; Oreg, 2006; Vos et al., 2015). Moreover, both leadership behavior and ARX are constructs which cannot be objectively observed. They have to be experienced during change. This is in line with Burnes and Jackson’s (2011) view on change effectiveness. They argue that the quantification of successful change has to be assessed based on what the change brings and how it was implemented.
HYPOTHESIS DEVELOPMENT
The influence of leadership behavior on change effectiveness
Higgs and Rowland (2011) found evidence of specific leadership behavior influencing change success in a qualitative study. This research will build upon that premises in a quantitative study, using a multiple perspective to research leadership behavior in relation to effective change. Higgs and Rowland (2011) found that shaping behavior is a leader-centric behavior. Leader-centric behavior has a negative effect on change success (Higgs, 2003; Higgs & Rowland, 2011), because a lack of consideration by the change agent tends to bring up resistance with the recipients (Groot, 2013). Shaping behavior is also identified as behavior in top-down change. Top-down change often fails, because it does not account for the impact of unexpected outcomes resulting from unintended interactions throughout the system (Harris & Ogbonna, 2002). Framing and creating behavior on the other hand seek interaction with the recipient and try to facilitate recipient’s actions (Higgs & Rowland, 2005; 2011). Because they are more facilitating and engaging, these behaviors have a positive effect on the success of change (Higgs & Rowland, 2011). The interaction between agent and recipient results in a better relationship between them; literature on communication during change confirms this (Elving, 2005). It also states that rich communication is preferred during times of change (Yazici, 2002). Furthermore, subordinates have more freedom in their behavior when a leader uses framing or creating behavior.
The current research therefore argues that on the one hand, high levels of shaping behavior have negative influence on the perception of change success. On the other, framing change and creating capacity behaviors positively influence the perception of change success. These relations have been conceptualized in Figure 1. Change agent’s and change recipient’s perspectives on the change agent’s behavior and on change effectiveness will be researched separately in order to see if the relations will indeed be different as argued in the literature section. Based on this argumentation the following hypotheses are proposed:
H1a/b: The higher the change agenta/recipientb’s perceived level of shaping behavior, the lower the change agenta/recipientb’s perceived level of change success.
H3a/b: The higher the change agenta/recipientb’s perceived level of creating behavior, the higher the change agenta/recipientb’s perceived level of change success.
FIGURE 1 Conceptual model
Viewed from a/b perspective
Shaping behavior
Creating behavior
Change
effectiveness
-Framing behavior
+ +The moderating role of ARX on leadership behavior and change effectiveness
Since the ARX literature is relatively novel, and since ARX is closely related to LMX and LMSX, LMX and LMSX literatures are used to support possible moderating relationship for ARX. From the LMX literature, we take that LMX has a mediating role between leadership behavior and employee performance (Dulebohn, Bommer, Liden, Brouer, & Ferris, 2012; Wang, Law, Hackett, Wang, & Chen, 2005). Wang et al. (2005) mention that effective leaders develop high quality exchange relationships with their subordinates. A leader’s behavior is dominant in this development (Dienesch & Liden, 1986). Rogulic (2013) already found evidence of ARX positively influencing change effectiveness. Therefore, it is argued that if ARX is high, it will influence the effect of change agent’s behavior on effective change positively. Despite this potential relation, the empirical relation among leadership behavior, change effectiveness, and ARX has not been documented before. The propositions have been conceptualized in Figure 2. This is formulated in hypothesis H4, H5 and H6. The two groups will be investigated separately to find out whether the moderating effect of ARX holds different implications on the relationship between the different leadership behaviors and change success for change agents and change recipients.
H5a/b: From a change agenta/recipientb’s perspective, ARX will have a positive moderating effect for the relation between perceived level of framing behavior and change success. H6a/b: From a change agenta/recipientb’s perspective, ARX will have a positive moderating
effect for the relation between perceived level of creating behavior and change success.
FIGURE 2 Conceptual model
Viewed from a/b perspective
Shaping behavior
Creating behavior
ARX
Change
effectiveness
+Framing behavior
+ +The different perspective of change agents and change recipients and the moderating effect of ARX
Literature pointed out that change agents and change recipients hold different views on change related matters. In hypotheses 1-3, the change agent’s and change recipient’s perspective on the relation of leadership behavior and effective change have been separated. It is argued that because of the different perspectives both change agents and change recipients have, the effect of the relation between leadership behavior and effective change will differ significantly between the groups. Hypothesis 7 conceptualizes this proposition.
Therefore, a change agent often sees change related matters more positive than the recipients (Sonenshein, 2010). Furthermore, Rogulic (2013) found evidence that change recipients see ARX as an important influencer of change effectiveness, but change agents do not. Therefore, it is argued that the effect of ARX as a moderator on the relationship between leadership behavior and change effectiveness is different for change agents than for change recipients. This proposition is formulated in hypothesis 8.
H7a/b/c: The relation between shapinga/framingb/creatingc behavior and change effectiveness has a significantly different effect from the change agent’s and the change recipient’s perspective.
H8a/b/c: The moderating effect of ARX on the relation between shapinga/framingb/creatingc behavior and change effectiveness has a significantly different effect from the change agent’s and the change recipient’s perspective.
METHODS
The hypotheses have been developed based on a rich field of academic literature on all variables. In order to test the hypotheses, a theory testing approach is used (Aken, Berends, & Bij, 2012). In this chapter, the sample and procedure are described first. Next, the measurements of the main variables and the control variables are discussed. Lastly, data analysis is mentioned.
Sample and procedure
For this research the dataset from the paper of Vos et al. (2015) is used to find an answer to the research question. The data gathering of this dataset is discussed. The questionnaire was developed and respondents were selected three years ago. To get a multiple view on the change, both change agent’s and change recipient’s perspective had to be included in the questionnaire. Therefore, two versions were made. A web-based version of the questionnaires was used to lower the entry threshold.
respondents had to be removed because of sabotaging answering, 361 recipient respondents remained. Sample descriptives can be found in Table 1.
TABLE 1 Sample descriptives Change agents Change recipients Total sample
Age (mean) in years 43 39 40
Gender Male 70,9% 54,0% 58,2% Female 29,1% 56,0% 41,8% Nationality Dutch 99,1% 99,7% 99,6% German - 0,3% 0,2% Other 0,9% - 0,3% Education level Primary education 2,6% 8,3% 6,9% Secondary or intermediate education 14,5% 33,8% 29,1% Polytechnic or University 81,2% 56,0% 62,1% Other 1,7% 1,9% 1,9% Work experience (mean) Current position 7 7 7 Total experience 18 17 17 N 117 361 478
N = 117 change agents, 361 change recipients, total sample 478 respondents
The match between change agent and change recipients is known as a ‘reciprocal one-with-many’ design (Kenny, Kashy & Cook, 2006). Each member of the dyadic relation was asked questions about the other (Kenny et al., 2006). One-with-many-designs are suitable for analyses which require multiple perspectives on a relationship (Krasikova & LeBreton, 2012). Both change agent’s and change recipient’s view on leadership, ARX and change effectiveness are measured in order to see if the relations between those variables differ between the two groups.
Measures
This research relies on existing scales and constructs. This section highlights the variables used in the survey and provides background for each variable. Besides the main constructs, control variables were used in order to be able to better interpret the results. Furthermore, the validity of each construct is tested. The assessment of the quality of the used scales was done in two steps.
The second step was a reliability analysis to assure internal reliability. The Cronbach’s Alphas of the main constructs and the control variables were measured. Because no interference was expected, the demographic control variables were not tested on internal reliability. For a construct to pass this test, Cronbach’s Alpha needs to be above a threshold of 0.7 (Nunnally, 1970). Except for scope breadth, all Cronbach’s alphas were above this threshold. However, because scope breadth had a significant (p < 0,01) correlation with change outcome of 0,557 and it was really close to the threshold (0.695), it was still deemed appropriate to use. Table 2 provides the Cronbach’s alphas of each construct, the items used, their means and standard deviations. Values of the variables are calculated as the mean scores of the items.
TABLE 2
Construct development and descriptives
Cronbach’s
alpha Items Mean
Standard Deviation Mean constructs Leadership behavior Shaping behavior 0.860 1-5 4,897 1,1416 Framing behavior 0.863 1-5 5,075 1,0556 Creating behavior 0.886 1-5 4,608 1,1818 Agent-recipient exchange 0.903 3-8 4,445 1,1410 Change effectiveness Change process* - - - - Change outcome 0.875 1-3 5,420 1,216 Control variable Readiness behavior 0,908 1-3 4,962 1,44806 Resistance behavior 0,786 1-3 2,997 1,50447 Scope breadth 0,695** 1,3 5,083 1,59991 Scope depth 0,785 1-6 4,286 1,19965
Change agent leadership behavior
In order to measure leadership behavior, the construct developed by Vos et al. (2015) was used. They based their quantitative measurement on the qualitative work of Higgs and Rowland (2005, 2011). Vos et al. (2015) designed and tested a set of quantitative measures for the three types of change-leader behavior, shaping, framing and creating, for both change agent and change recipient. This resulted in a five item construct per type of leadership behavior which were measured on a seven-point Likert scale. All items for each of the three leadership behaviors loaded correctly into their own component.
Agent-recipient exchange
Agent-recipient exchange (ARX) was measured by using an adapted version of the Leader-member social exchange (LMSX) scale developed and validated by Bernerth et al. (2007). Rogulic (2013) adjusted these items to fit an organizational change context. Furthermore, he developed an agent’s version, because the original scale of Bernerth et al. (2007) was designed to measure LMSX from a member’s perspective only (Rogulic, 2013). The construct consists of eight items, measured with seven-point Likert-scales. Two items of ARX (ARX_1 and ARX_2) did not meet the predefined requirements and were removed for further analysis.
Perceived change effectiveness
Control variables
Control variables were subdivided in change-related variables and general variables. Using control variables mitigates the risk of overestimating or underestimating the main variables. In line with earlier research on leadership behavior, ARX and change effectiveness, readiness behavior and resistance behavior and the scope of change were added to the research as control variables (Jansons, 2015; Rogulic, 2013; Vos et al., 2015).
Readiness behavior was measured using Vos et al.’s (2015) readiness scale. This scale is based on the intentional readiness scale of Bouckenooghe, DeVos, and Broeck (2009). Vos et al. (2015) adjusted the items slightly to be a positive counterpart of resistance behavior. Items were duplicated and adjusted to fit a change agent’s perspective. To measure resistance behavior, Oreg’s (2006) behavioral resistance scale was used. His scale consists of fifteen-items, containing an affective, cognitive and behavioral component. Only the behavioral component was used because in a dyadic relation it is only possible for both two groups to describe the other’s behavior towards a change. The behavioral component of resistance behavior consisted out of a five-item scale. However, one of those items holds a positive view on resistance and was therefore already covered by the items of readiness behavior (Groot, 2013; Vos et al., 2015). Another item of the original scale is based on cognitions (Groot, 2013; Vos et al., 2015) and therefore not useable to measure behavior. Three items remained. The original items developed by Oreg (2006) only measured resistance behavior from a recipient’s perspective. Since this current research includes both perspectives on the change recipient’s behavior, items for the perspective of the change agent had to be duplicated and adjusted from the recipient’s version to a change agent’s version. Both readiness behavior and resistance behavior were measured using a three-item construct, all measured on a seven-point Likert-scale. The control variables’ items readiness behavior and resistance behavior loaded correctly in separate components as was expected from literature.
scope of the change, scales were adopted from Vos and Brand (2012). Scope breadth was measured using their four item scale and scope depth was measured using their six item scale, all measured on a seven-point Likert-scale. Two items of scope breadth did not meet the preset requirements of 0,5 and were therefore removed from the factor analysis. The remaining two loaded into one component. The items of scope depth all loaded into one component.
Besides the change-related constructs as control variables, more general variables were considered as well. The respondents’ age, gender, educational level, experience in their current position and total work experience were taken into account. Respondent’s age, gender, experience in their current position and total work experience are all one-item constructs. Respondent’s education level was measured using a four item categorical construct. Because of the categorical data of education level, three dummy variables were computed (Field, 2009).
Data analysis
TABLE 3
Kolmogorov-Smirnov test
Variables Statistic Significance Skewness Kurtosis
Shaping behavior ,098 ,000 -,744 ,414 Framing behavior ,111 ,000 -,790 ,495 Creating behavior ,075 ,000 -,507 ,030 Agent-recipient exchange ,089 ,000 -,348 ,077 Change outcome ,162 ,000 -1,036 ,935 Readiness behavior ,128 ,000 -,771 ,079 Resistance behavior ,116 ,000 ,337 -,733 Scope breadth ,177 ,000 -,795 -,277 Scope depth ,075 ,000 -,296 -,441 N = 478
the assumed moderator by centering the mean of each variable. In order to create an interaction effect, they multiply the centered independent variable with the centered moderating variable. To give a clear overview of this interaction between ARX and each leadership behavior, all leadership behaviors were loaded simultaneously for hypotheses 4-6. The first step was to add the control variables in the first model of the hierarchical moderating ridge regression. In the second step all leadership behaviors were added. The last step consisted of adding ARX and the interaction effect of leadership behavior and ARX into the model.
RESULTS
This section first discusses the results from a change agent’s perspective, followed by the results from the change recipient’s perspective. Lastly, a comparison is presented between the change agent’s and change recipient’s models.
Correlations from the change agent’s perspective
A correlation analysis shows whether the independent variables correlate with the main variable and the control variables. All correlations from the change agent’s perspective are shown in Table 4. In accordance with hypothesis 1a and 2a, shaping and framing behavior correlate significantly with change outcome. Creating behavior does not correlate significantly with change outcome. In congruence with hypothesis 4a up to 6a, ARX significantly correlates with change outcome, shaping, framing and creating behavior. Regarding the control variables, the results show that readiness behavior, resistance behavior, scope depth, scope breadth, age, primary education, experience in current position and total experience were all found to have a significant correlation with one or more of the main variables.
Hypotheses testing change agent’s perspective
outcome (B = ,096, p < 0,05). Hypothesis 2a is accepted. Creating behavior does not significantly influence change outcome (B = -,008, p > 0,05). Hypothesis 3a is rejected.
TABLE 4
Correlations variables from agent’s perspective
TABLE 5
Table 6
Hierarchical moderating ridge regression analysis from agent’s perspective on change effectiveness
Model 1 Model 2 Model 3
B Sign. B Sign. B Sign.
Readiness behavior ,212** ,000 ,184* ,000 ,186* ,000 Resistance behavior -,055 ,154 -,056 ,121 -,060 ,117 Scope depth ,035 ,462 ,006 ,893 ,013 ,764 Scope breadth ,061 ,156 ,051 ,209 ,048 ,212 Age -,024 ,499 -,030 ,394 -,029 ,357 Gender ,047 ,158 ,049 ,132 ,044 ,160 Primary education ,012 ,496 ,003 ,857 ,000 ,999 Secondary education ,075* ,038 ,076* ,037 ,071* ,038 Polytechnic University ,042 ,130 ,049 ,090 ,051 ,076 Experience in current position -,003 ,931 -,004 ,903 -,002 ,943 Experience in total ,014 ,653 ,002 ,958 ,000 ,999 Shaping behavior ,096* ,029 ,098* ,014 Framing behavior ,096* ,035 ,099* ,019 Creating behavior -,008 ,865 ,000 ,992 ARX ,025 ,556
Shaping behavior * ARX ,029 ,491
Framing behavior * ARX -,107** ,001
Creating behavior * ARX -,016 ,682
R square ,257 ,294 ,344
R square change ,037 ,050
Adjusted R square ,179 ,197 ,223
F 2,494** 2,441** 2,172**
Correlations from the change recipient’s perspective
All correlations from the change recipient’s perspective can be found in Table 7. The correlations show that all main constructs correlate significantly with each other. In line with hypotheses 1b-3b, shaping, framing and creating behavior correlate significantly with change outcome. For hypotheses 4b-6b, which assume that ARX has a positive moderating effect on the relation between leadership behavior and change effectiveness, results show that ARX correlates significantly with all of the main variables. Some of the control variables also correlate with one or more of the main variables. It was found that readiness behavior, resistance behavior, scope depth, scope breadth, age, primary education and experience in current position significantly correlate with the main variables. Gender, secondary education, polytechnic university and experience total did not load to one of the main variables.
Hypotheses testing change recipient’s perspective
change recipient’s perspective. Results show a positive significant causal relation between creating behavior and change outcome (B = ,079, p < 0,01). Hypothesis 3b is accepted.
Hypothesis 4a up to 6a describe an interaction effect of ARX on the relation between shaping, framing or creating behavior and change effectiveness from change recipient’s perspective. In order to test this a hierarchical moderating ridge regression analysis was used. The results of the analysis are depicted in Table 9. Because model 1 and model 2 are identical to the previous analysis for hypothesis 1, 2, and 3 only model 3 is described. Model 3 has a significant F value: F = 9,098 (p < 0,01). About 32,1 % of variance in change effectiveness in the change recipient’s data can be explained by variables in model 3 (R2
TABLE 7
Correlations variables from recipient’s perspective
TABLE 8
TABLE 9
Hierarchical moderating ridge regression analysis from recipient’s perspective on change effectiveness
Model 1 Model 2 Model 3
B Sign. B Sign. B Sign.
Readiness behavior ,128** ,000 ,087** ,000 ,074** ,001 Resistance behavior -,165** ,000 -,138** ,000 -,132** ,000 Scope depth ,042 ,098 ,039 ,100 ,042 ,063 Scope breadth ,050 ,061 ,036 ,135 ,039 ,119 Age -,043** ,008 -,046** ,002 -,042** ,005 Gender ,022 ,239 ,030 ,122 ,030 ,112 Primary education ,035 ,097 ,036 ,070 ,039 ,059 Secondary education ,027 ,101 ,025 ,095 ,029 ,064 Polytechnic University ,000 ,987 ,005 ,598 ,003 ,722 Experience in current position -,097 ** ,000 -,095** ,000 -,093** ,000 Experience in total -,002 ,899 -,005 ,761 -,004 ,810 Shaping behavior ,088** ,000 ,075** ,000 Framing behavior ,126** ,000 ,110** ,000 Creating behavior ,079** ,000 ,059** ,005 ARX ,109** ,000
Shaping behavior * ARX -,006 ,737
Framing behavior * ARX -,006 ,769
Creating behavior * ARX ,032 ,090
R square ,250 ,336 ,355
R square change ,086** ,019*
Adjusted R square ,226 ,309 ,321
F 8,043** 10,661** 9,098**
Comparing the change agent’s view to the change recipient’s view
Table 10 shows the results of the correlation analysis are showed. All main variables correlate significantly among each other. All control variables correlate significantly with at least one of the main variables.
Hypothesis 7 states that the relation between leadership behavior and change effectiveness has a significantly different effect for change agents and change recipients. The different effect of the two perspectives was tested by analyzing the interaction between leadership behavior and the change-recipient variable in a hierarchal moderating ridge regression analysis. Table 11 provides the results for this test. Model 1, model 2, and 3 have significant F values: F = 8,710 (p < 0,01), F = 14,182 (p < 0,01), and F = 12,647 (p < 0,01). About 19,3 % of variance in change effectiveness in the combined perspectives data can be explained by the control variables in model 1 (R2 (adj.) = ,193). Model 2, with all leadership behaviors, explains 31,1 % of variance in change effectiveness in the combined perspectives data (R2 (adj.) = ,311). About 33 % of variance in change effectiveness in the combined perspectives data can be explained variables in model 3 (R2 (adj.) = ,330). The R² change in model 2 indicates that model 2 is significantly different from model 1 (R2 change = ,120, p < 0,01), indicating that model 2 significantly differs from model 1. Model 2 explains 12 % of the variance more than model 1. The R² change in model 3 indicates that model 3 is significantly different from model 2 (R2 change = ,023, p < 0,01), indicating that model 3 significantly differs from model 2. Model 3 explains 2,3 % of the variance more than model 2.
Hypothesis 7c describes that the relation between creating behavior and change effectiveness has a significantly different effect for change agents than for change recipients. In model 3, creating behavior has a positive significant influence on change outcome (B = ,043, p < 0,05). The interaction effect of creating behavior and the agent-recipient variable has a modest positive significant influence on change outcome (B = ,033, p < 0,05). Therefore, hypothesis 7c is accepted.
In hypothesis 8, it is expected that the positive moderating effect of ARX on the relation between leadership behavior and change effectiveness will have a significantly different effect for change agents than for change recipients. This was tested using a hierarchical three-way interaction ridge regression analysis. In Table 12 the results of this analysis are provided. Model 1 and model 2 are identical as with hypothesis 7. Model 3 and model 4 have significant F values: F = 11,490 (p < 0,01) and F = 9,144 (p < 0,01). About 31,1 % of variance in change effectiveness in the combined perspectives data can be explained by the variables in model 3 (R2 (adj.) = ,311). Model 4 explains 33 % of the variance in change effectiveness in the combined perspectives data (R2 (adj.) = ,330). The R² change in model 3 indicates that model 3 is not significantly different from model 2 (R2 change = ,005, p > 0,05). The R² change in model 4 indicates that model 4 is significantly different from model 3 (R2 change = ,030, p < 0,01). Model 4 explains 3 % of the variance more than model 3. In model 4 ARX and the agent-recipient variable have significant positive causal relations with change outcome (B = ,049, p < 0,01 and B = ,117, p < 0,01). Also, the interaction effect of agent-recipient and ARX has a positive and significant causal relation with change outcome (B = ,058, p < 0,01).
TABLE 10
Correlations variables combined dataset
TABLE 11
Hierarchical moderating ridge regression analysis from a combined perspective on change effectiveness
Model 1 Model 2 Model 3
B Sign. B Sign. B Sign.
Readiness behavior ,172** ,000 ,116** ,000 ,101** ,000 Resistance behavior -,083** ,000 -,076** ,000 -,088** ,000 Scope depth ,054* ,017 ,038 ,065 ,039 ,056 Scope breadth ,052* ,019 ,037 ,087 ,037 ,062 Age -,020 ,183 -,028* ,041 -,035* ,011 Gender ,003 ,808 ,015 ,310 ,022 ,167 Primary education ,043* ,029 ,042* ,027 ,037* ,049 Secondary education ,012 ,304 ,008 ,401 ,003 ,783 Polytechnic University ,036* ,019 ,035* ,015 ,027* ,046
Experience in current position -,080** ,000 -,076** ,000 -,074** ,000
Experience in total -,001 ,953 -,006 ,648 -,006 ,649
Shaping behavior ,110** ,000 ,077** ,000
Framing behavior ,147** ,000 ,103** ,000
Creating behavior ,070** ,000 ,043* ,015
Agent-recipient ,115** ,000
Shaping behavior * Agent-recipient ,035* ,017
Framing behavior * Agent-recipient ,063** ,000
Creating behavior * Agent-recipient ,033* ,029
R square ,212 ,332 ,355
R square change ,120** ,023**
Adjusted R square ,193 ,311 ,330
F 8,710** 14,182** 12,647**
TABLE 12
Hierarchical three-way interaction ridge regression analysis from a combined perspective on change effectiveness
Model 1 Model 2 Model 3 Model 4
B Sign. B Sign. B Sign. B Sign.
Readiness behavior ,172** ,000 ,116** ,000 ,107** ,000 ,091** ,000 Resistance behavior -,083** ,000 -,076** ,000 -,071** ,001 -,082** ,000 Scope depth ,054* ,017 ,038 ,065 ,044* ,031 ,044* ,040 Scope breadth ,052* ,019 ,037 ,087 ,040 ,065 ,041* ,037 Age -,020 ,183 -,028* ,041 -,026 ,056 -,031* ,016 Gender ,003 ,808 ,015 ,310 ,015 ,291 ,021 ,175 Primary education ,043* ,029 ,042* ,027 ,044* ,019 ,040* ,031 Secondary education ,012 ,304 ,008 ,401 ,006 ,562 ,001 ,942 Polytechnic University ,036* ,019 ,035* ,015 ,034* ,016 ,025 ,054
Experience in current position -,080** ,000 -,076** ,000 -,075** ,000 -,072** ,000
Experience in total -,001 ,953 -,006 ,648 -,005 ,741 -,005 ,727
Shaping behavior ,110** ,000 ,101** ,000 ,071** ,000
Framing behavior ,147** ,000 ,133** ,000 ,093** ,000
Creating behavior ,070** ,000 ,056** ,003 ,032 ,055
ARX ,079** ,000 ,049** ,002
Shaping behavior * ARX -,002 ,901 ,004 ,808
Framing behavior * ARX -,040* ,024 -,033* ,043
Creating behavior * ARX ,011 ,525 ,013 ,435
Agent-recipient ,117** ,000
Shaping behavior * Agent-recipient ,026 ,074
Framing behavior * Agent-recipient ,051 ,001
Creating behavior * Agent-recipient ,019 ,214
ARX * Agent-recipient ,058** ,000
Shaping behavior * Agent-recipient * ARX -,010 ,417
Framing behavior * Agent-recipient * ARX -,001 ,942
Creating behavior * Agent-recipient * ARX ,016 ,264
R square ,212 ,332 ,337 ,367
R square change ,120** ,005 ,030**
Adjusted R square ,193 ,311 ,311 ,330
F 8,710** 14,182** 11,490** 9,144**
DISCUSSION
Discussion of the findings
There is extensive literature on the relation between leaders and subordinates (e.g. Bernerth, et al., 2007; Self et al., 2007). However, this literature did not asses the influence of leader-member exchange on the relation between leadership behavior and change effectiveness. In this paper the interaction effect of ARX on leadership behavior and change effectiveness has been investigated from both a change agent’s and change recipient’s perspective. This research demonstrates once more that change cannot be viewed from a unilateral perspective (Bouckenooghe et al., 2009).
Hypotheses 1, 2 and 3 assume that shaping, framing or creating behavior influence change effectiveness. From a recipients’ perspective, shaping, framing and creating behavior influence change effective significantly. Interestingly, shaping behavior has a significant positive effect on change effectiveness. This is in contradiction with the hypothesis and with the research of Higgs and Rowland (2005, 2011). It was expected that shaping behavior would negatively influence change effectiveness. An explanation for these results would be that, just as framing and creating behavior, shaping behavior too is a highly visible leadership behavior (Vos et al., 2015). This indicates that, from a change recipient’s perspective, the fact that the change agent is visible and present during times of change rather than specific leadership behavior influences change effectiveness (Vos et al., 2015). Higgs and Rowland (2011) also found that these three leadership behaviors occur simultaneously.
behavior attach less value to their own behavior in relation to change effectiveness, as the recipients play a more important role.
In hypotheses 4, 5 and 6 it is assumed that ARX has a positive moderating role on the relation between leadership behavior and change effectiveness. From both the change agent’s and the change recipient’s data, none of these hypotheses could be confirmed. Only one modest significant causal relation was found, ARX negatively moderates the relationship between framing behavior and change effectiveness.
This result has not been explained in any of the known literature on the subject. However, from a change agent’s perspective, this relationship can be explained by the self-serving bias. The self-serving bias is based on the premise that change agents have the tendency to take credit for successes and to blame situational factors for their failures (Martinko & Douglas, 1999). That is, they tend to attribute much of the change success to their own behavior. Now, when the exchange relationship between an agent and a recipient is of low quality, an agent would be unlikely to attribute much of this success to a change recipient (Dobbins & Russell, 1986). Rather, the agent would be inclined to take credit himself. When the relationship is of higher quality, the role of this bias may well be smaller.
In sum, it seems that agents and recipients have different views on how ARX moderates the relation between leadership behavior and change effectiveness. Even though this may seem contradictory at first, these effects can be explained using the same mechanism. A word of caution is in place however, since a lack of significant results for these hypotheses limit from fully empirically explaining these effects.
In hypothesis 7 and 8, it was expected that hypotheses 1 to 6 hold significantly different effects for change agents and change recipients. Hypothesis 7 states that the relation between leadership behavior and change effectiveness has a significantly different effect for change agents and change recipients. This hypothesis was confirmed. In hypothesis 8 it was expected that the moderating effect of ARX on the relation between shaping, framing and creating behavior and change effectiveness would be significantly different for change recipients than for change agents. No significant relations were found to support this model. The confirmation of hypothesis 7 shows that there is a significantly different effect for change agents and change recipients on the relation between leadership behavior and change effectiveness. Moreover, the positive significant interaction effect on the relation between all leadership behaviors and change effectives is stronger for change recipients than for change agents. This result supports the notion that investigating change from two perspectives is of great importance (Sonenshein, 2010, Vos et al., 2015).
Theoretical contributions
The findings of this research have implications for change leadership behavior, ARX and change effectiveness literature. First, the leadership behaviors of Higgs and Rowland (2005; 2011) have been researched in only a small amount of quantitative studies (e.g. Vos et al., 2015). This study provides further quantitative results on leadership behavior. Furthermore, as suggested by Higgs and Rowland (2011), the current research points in the direction that change leadership behaviors can be used simultaneously.
Second, although out of the scope of this research, results indicate that ARX has a direct positive effect on change effectiveness in the combined dataset. This contributes to the earlier research on ARX by Rogulic (2013). Furthermore ARX has a direct positive effect on change effectiveness from the perspective of the change recipient. This is in the line with the work of Ilies et al. (2007) who found that LMX contributes to organizational effectiveness.
Lastly, this study once more stresses the importance of looking at change related matters from multiple perspectives. It confirms the expectation of several scholars about the differences between change agents and change recipients (e.g. Gils et al., 2010; Sonenshein, 2010; Vos et al., 2015).
Practical implications
The outcomes of this research have a number of useful practical implications. It is known that organizations experience difficulties finding appropriate ways to change effectively (Burnes, 2011). Also, change agents do not always have the insight in the perspective of the change recipient. This study provides such insight.
First, the results show that for change recipients it is important to have a leader who is present in times of change. It is advisable for change agents to be visible during times of change and accompany recipients. Recipients need a leader who leads them through the change.
Limitations and future research
This research provides insights in the perspectives of both change agents and change recipients on change leadership, change effectiveness and the dyadic relationship between agent and recipient. Nevertheless, like any research, it faced several limitations. First, the data consist out of related variables. The high multicollinearity complicates analyzing relations. Although this study uses a ridge regression to overcome this problem, the multiple interaction calculations could be a reason why results did not show significant effects in the three-way interaction of leadership behavior, ARX and agent-recipient.
Second, the significant and non-significant results alone do not allow researchers to draw an encompassing conclusion. This is because quantitative research does not provide context to explain unexpected results. For example, it is remarkable that from the agent’s perspective the relation between framing behavior and change effectiveness is significant negative moderated by ARX, but shaping and creating behavior are not. Adding more context in some cases would help finding additional explanations. A qualitative study could provide additional context which is needed to clarify some unexpected outcomes in this research.
Third, the combined data set consists out of 117 change agents and 361 change recipients. The interaction effects may have been influenced by the low amount of change agents in this data set, compared to the number of change recipients. A larger and equally distributed amount of change agents and change recipients is expected to provide more significant results (Gwowen, 2009; Kenny & Judd, 1984).
Not only the limitations of this study provide future researchers with interesting starting points for their studies, the results do as well. The results showed that all three behaviors can be expressed simultaneously. Future research could investigate how these behaviors can be expressed complementary by change agents. Furthermore, it seems that change agents prefer different leadership behavior than change recipients. It would be interesting to further investigate whether there are indeed conflicting interests, or if it is possible to find a balanced set of behaviors to fit both perspectives best.
REFERENCES
Aken, J. E., Berends, H., & Bij, H. van der (2012). Quality criteria for research. In: Problem solving in organizations – a methodological handbook for business and management studies. New York: Cambridge University Press.
Al-Haddad, S., & Kotnour, T. (2015). Integrating the organizational change literature: a model for successful change. Journal of Organizational Change Management, 28(2), 234-262. Avolio, B. J., Walumbwa, F. O., & Weber, T. J. (2009). Leadership: current theories, research,
and future directions. Annual Review of Psychology, 60, 421-449.
Balogun, J., & Hope Hailey, V. (2008). Exploring strategic change. Harlow: Pearson.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6):1173-1182.
Bass, B. M. (1985). Leadership and performance beyond expectations. New York: Free Press. Bass, B. M. (1996). A new paradigm for leadership: An inquiry into transformational leadership.
Alexandra: U.S. Army Research Institute fir the Behavioral and Social Sciences.
Bernerth, J. B., Armenakis, A. A., Field, H. S., Giles, W. F., & Walker, H. J. (2007). Leader– member social exchange (LMSX): Development and validation of a scale. Journal of Organizational Behavior, 28, 979-1003.
Blau, P. M. (1964). Exchange and power in social life. New York: John Wiley & Sons.
Blumberg, B., Cooper, D. R., & Schindler, P. S. (2011). Business research methods (2th ed.). Berkshire: McGraw-Hill Education.
Bouckenooghe, D., DeVos, G., & Broeck, H. van den (2009). Organizational change questionnaire-climate of change, processes, and readiness: Development of a new instrument. The Journal of Psychology, 143(6): 559-599.
Brown, J. D. (2009). Statistics Corner. Questions and answers about language testing statistics: Principal components analysis and exploratory factor analysis: Choosing the Right Type of Rotation in PCA and EFA. Shiken: JALT Testing & Evaluation SIG Newsletter, 13(3) 20-25. Retrieved March 27, 2016 from http://jalt.org/test/bro_31.htm
Burnes, B., & Jackson, P. (2011). Success and failure in organizational change: An Exploration of the Role of Values. Journal of Change Management, 11(2), 133-162.
By, R. T. (2005). Organizational change management: a critical review. Journal of Change Management, 5(4), 369–380.
Crosby, P. B. (1979). Quality is Free. New York: McGraw-Hill.
Dam, K. van, Oreg, S., & Schyns, B. (2008). Daily work contexts and resistance to organizational change: the role of leader–member exchange, development climate, and change process characteristics. Applied Psychology: An International Review, 57(2), 313-334.
Dansereau, F., Graen, G. B., & Haga, W. J. (1975). A vertical dyad linkage approach to leadership within formal organizations: A longitudinal investigation of the role making process. Organizational Behavior and Human Performance, 13, 46-78.
Dienesch, R. M., & Liden, R. C. (1986). Leader-member exchange model of leadership: A critique and further development. Academy of management review, 11(3), 618-634.
Dobbins, G. H., & Russell, J. M. (1986). Self-Serving Biases in Leadership: A Laboratory Experiment. Journal of Management, 12(4), 475.
Dulebohn, J. H., Bommer, W. H., Liden, R. C., Brouer, R. L., & Ferris, G. R. (2012). A meta-analysis of antecedents and consequences of leader-member exchange integrating the past with an eye toward the future. Journal of Management, 38(6), 1715-1759.
Eisenbach, R., Watson, K., & Pillai, R. (1999). Transformational leadership in the context of organizational change. Journal of Organizational Change Management, 12(2), 80-89. Eitzen, D. S., & Yetman, N. R. (1972). Managerial Change, Longevity, and Organizational
Effectiveness. Administrative Science Quarterly, 17(1), 110-116.
Elving, W. J. L. (2005). The role of communication in organizational change. Corporate Communications: An International Journal, 10(2): 129-138.
Field, A. (2009). Discovering statistics using SPSS. London: Sage publications.
Ford, J. D., Ford, L. W., & D’Amelio, A. (2008). Resistance to change: The rest of the story. Academy of Management Review, 33(2): 362-377.
Garson G. D. (2012). Testing Statistical Assumptions. Asheboro: Statistical Associates Publishing.
Gerstner, C. R., & Day, D. V. (1997). Meta-analytic review of leader–member exchange theory: Correlates and construct issues. Journal of applied psychology, 82(6), 827-844.
Gideon, R. A., & Hollister, R. A. (1987). A Rank Correlation Coefficient Resistant to Outliers. Journal of the American Statistical Association, 82(398), 656-666.
Gill, R. (2003). Change management - or change leadership? Journal Of Change Management, 3(4), 307-318.
Gils, S. van, Quaquebeke, N. van, & Knippenberg, D. van (2010). The X-factor: On the relevance of implicit leadership and followership theories for leader-member exchange agreement. European Journal of Work & Organizational Psychology, 19(3), 333-363.
Gioia, D. A., & Chittipeddi, K. (1991). Sensemaking and sensegiving in strategic change initiation, Strategic Management Journal, 12, 433-448.
Gioia, D. A., Thomas, J. B., Clark, S. M., & Chittipeddi, K. (1994). Symbolism and strategic change in academia: The dynamics of sensemaking and influence. Organization science, 5(3), 363-383.
Groot, E. de. (2013). Resistance to change: a bilateral perspective: The influences of change agents’ leader behavior and recipients’ involvement on recipients’ resistance to change (Master thesis). Groningen: University of Groningen.
Gwowen, S. (2009). Detecting interaction effects in moderated multiple regression with continuous variables: power and sample size considerations. Organizational Research Methods, 12(3), 510-528.
Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006). Multivariate data analysis. Upper Saddle River: Pearson Prentice Hall.
Harris, L. C., & Ogbonna, E. (2002). The Unintended consequences of culture interventions: a study of unexpected outcomes. British Journal of Management, 13(1), 31-49.
Higgs, M. (2003). Developments in leadership thinking, Journal of Organisational Development and Leadership, 24(5), 273–284.
Higgs, M., & Rowland, D. (2000). Building change leadership capability: The quest for change competence. Journal of Change Management, 1, 116-130.
Higgs, M., & Rowland, D. (2005). All changes great and small: Exploring approaches to change and its leadership. Journal of Change Management, 5(2), 121-151.
Higgs, M., & Rowland, D. (2011). What does it take to implement change successfully? A study of the behaviors of successful change leaders. The Journal of Applied Behavioral Science, 47(3): 309-335.
Hoaglin, D. C., and Iglewicz, B. (1987). Fine tuning some resistant rules for outlier labeling. Journal of American Statistical Association, 82, 1147-1149.
Ilies, R., Nahrgang, J. D., & Morgeson, F. P. (2007). Leader-member exchange and citizenship behaviors: a meta-analysis. Journal of applied psychology, 92(1), 269.
Jansons, J. J. (2015). Antecedents of social exchange relationships in organizational change. A bilateral study on the influence of change leadership behavior on perceptions of Agent-recipient exchange (Master thesis). Groningen: University of Groningen.
Jones, E. E., & Nisbett, R. E. (1972). The actor and the observer: divergent perceptions of the causes of behavior. In: Jones, E. E., Kanhouse, D. E., Kelley, H. H., Nisbett, R. E., Valins, S., & Weiner, B. (1972). Attribution: perceiving the causes of behavior (p. 79-94). Morristown: General Learning.
Kam, N. A. van der, Vegt, G. S. van der, Janssen, O., & Stoker, J. I. (2015). Heroic or hubristic? A componential approach to the relationship between perceived transformational leadership and leader–member exchanges. European Journal of Work and Organizational Psychology, 24(4), 611-626.
Kanter, R. M., Stein, B. A., & Jick, T. D. (1992). The challenge of organizational change: How companies experience it and leaders guide it. New York: Free Press.
Kenny, D. A., & Judd, C. M. (1984). Estimating the nonlinear and interactive effects of latent variables. Psychological Bulletin, 96, 201-210.
