Resistance and readiness: The influence of commitment to change and communication on change recipients’ reactions

Resistance and readiness:

The influence of commitment to change and communication on

change recipients’ reactions

Master thesis, MSc BA - Change Management University of Groningen, Faculty of Economics and Business

October 31, 2013 Justin Beenakker Student number: 1764179 Paterswoldseweg 36c 9726 BE Groningen T: +31 (0)6 24987587 Email: justinbeenakker@live.nl Supervisors - University

1

Resistance and readiness:

The influence of commitment to change and communication on

change recipients’ reactions

Abstract

Resistance to change and readiness for change make up the vast majority of the literature on recipients’ reactions to change. Although a large body of research identified numerous antecedents, their influence is not always as well understood. Using two well-known concepts, this research tests the influence of commitment to change and communication on both reactions to change. Using Herscovitch and Meyer (2002) and Russ (2008) their scales, these concepts were measured in two companies in the Netherlands. It was found that programmatic communication was neither related to commitment to change nor to recipients’ reactions to change. Participatory communication was only related to affective commitment to change and readiness for change. The only significant influence of commitment to change on recipients’ reactions was the affective dimension. Additional findings were the mediating role of affective commitment to change on the relationship between participatory communication and readiness and the lack of evidence for the claim that resistance to change and readiness for change are opposites. This research partially supports the notion that communication plays a crucial role in reactions to change and emphasizes the importance of affectively committed recipients.

2

Content

1. Introduction ... 4

2. Literature review ... 6

2.1 Recipients’ reactions to change ... 6

2.1.1 Resistance ... 6 2.1.2 Readiness ... 7 2.2 Commitment to change ... 9 2.2.1 Affective commitment ... 10 2.2.2 Normative commitment ... 11 2.2.2 Continuance commitment ... 12 2.3 Communication ... 14 2.3.1 Programmatic communication ... 15 2.3.2 Participatory communication ... 20 2.4 Conceptual model ... 23 3. Methodology ... 25 3.1 Data collection ... 26 3.2 Measurements ... 28 3.2.1 Dependent ... 28 3.2.2 Independent ... 29

3.2.3 Factor and regression analysis ... 29

4. Results ... 32

4.1 Factor and reliability analysis ... 32

4.2 Descriptive and correlation analysis ... 36

4.3 Hypotheses testing ... 38

5. Discussion ... 47

5.1 Discussion ... 47

5.2 Implications and limitations ... 53

6. Conclusion ... 56

7. References ... 57

Appendix I – Transcripts of interviews KPN ... 70

Appendix II - Transcripts of interviews KPN ... 73

Appendix III – Transcripts of interviews KPN ... 76

Appendix IV – Factor analysis Part I – All variables ... 78

Appendix V – Factor analysis Part II – Dependent variables ... 79

3

Appendix VII – Scatterplot Programmatic communication – Resistance to change ... 83

Appendix VIII – Scatterplot Programmatic communication – Readiness for change ... 84

Appendix IX – Mediation analysis ... 85

4

1. Introduction

Although there is an increase in large scale change efforts, the success rate is remarkably low (Beer and Nohria, 2000). In a study of over 3,000 executives, Meany and Pung (2008) reported only one-third of the executives perceived their organization as successful in implementing change. In order to increase the success rate, the change management literature tries to identify factors that could increase the chances of success (Raffery, Jimmieson and Armenakis, 2013). It is argued that employee attitudes towards change are critical in achieving successful change (Miller, Johnson and Grau, 1994). A lot of research has been done on attitudes towards change, but an overwhelming ninety percent of all conceptual work on change attitudes deals with just two reactions: resistance to change and readiness for change (Bouckenooghe, 2010).

The research into these reactions to change requires a micro level approach (Cunningham, 2006). This has resulted in research focused on the person-oriented view (Herscovitch and Meyer, 2002; Judge, Thoresen, Pucik and Welbourne, 1999; Neves, 2011; Oreg, et al., 2011). The increasing pace of change has made people more concerned with how these changes affect themselves, their jobs and their colleagues (Neves and Caetano, 2009). So far, research has determined attitudes in general comprise three dimensions; affective, cognitive and behavioral (Elizur and Guttman, 1976; Piderit, 2000). While the affective dimension refers to the feelings about the change, the cognitive dimension refers to the opinion one has about certain aspects of the change. The behavioral dimension refers to actions taken or which will be taken for or against the change. This conceptualization has led researchers to develop measures to capture these three dimensions for both resistance to change (e.g. Oreg, 2006) and readiness for change (e.g. Bouckenooghe, Devos and van den Broeck, 2009). As mentioned, these two concepts make up the majority of the research on attitudes towards change (for a review, see Bouckenooghe (2010)). Nevertheless, there remains disagreement on the role each reaction plays and their position towards each other.

5

Before recipients can connect to any change goal, these goals need to be communicated to them. Providing recipients with more information about proposed changes leads to greater change acceptance (Wanberg and Banas, 2000) and increased commitment (Schweiger and DeNisi, 1991). Moreover, communication plays a crucial role during organizational change (Allen, Jimmieson, Bordia and Irmer, 2007). It can be said that change is implemented and sustained through human communication (Russ, 2008). The research on communication is abundant, as is the number of conceptualizations. A useful categorization is made by Russ (2008), who divides communication into two approaches: programmatic and participatory communication. The former is one-way communication about the change in a top-down manner, while the latter is dialogic communication involving stakeholders of the change in the communication process. According to Elving (2005) communication serves three goals. First, communication should prevent or try to reduce resistance. Second, it should inform employees about their tasks and other organizational matters. The third and final goal is creating a community. Through these three goals, communication should create or increase readiness and commitment (Elving, 2005). These goals, however, reveal nothing on how the communication should be done. The relation between communication and reactions to change is not always as clear as assumed. Oreg et al. (2011) identify the role of information as a complexity. Providing information is often found to reduce resistance through reducing uncertainty, but some find more information leads to increased resistance (e.g. Oreg, 2006). This contradiction could be explained by the content (Oreg et al., 2011) or by the quality of information (e.g. Bouckenooghe and Menguç, 2010). The inconsistent findings of communication remains underexplored.

This study will address the disagreement concerning the role of resistance to change and readiness for change by assessing the influence of commitment to change and communication on both reactions simultaneously. Furthermore, by assessing the influence of commitment to change on these reactions, the study will contribute to the limited understanding on this subject. Finally, by using the categorization by Russ (2008), this study hopes to shed new light on the complex role communication can have on recipients’ reactions to change. In sum, this research addresses the following research question:

6

2. Literature review

2.1 Recipients’ reactions to change

The reactions of change recipients have been a profoundly studied topic. In their 60-year

review, Oreg et al. (2011) summarize the findings on reactions to change in a model of change

recipient reactions to organizational change. These reactions are considered to play a key role

and are therefore a main determinant of the extent to which any change can succeed

(Bartunek, Rousseau, Rudolph and DePalma, 2006; Oreg et al., 2011). With their review,

Oreg et al. (2011) provide clarification on change recipients’ explicit reactions by classifying

them into affective, cognitive and behavioral dimensions. Even though this does answer the

limitation coined by Piderit (2000) about the unidimensionality of reactions to change in

earlier research, it does not address the fact that most research conceptualized reactions as

resistance or readiness. Bouckenooghe (2008) concludes that ninety percent of the literature

refers to either resistance or readiness as a reaction to change. However, there is disagreement

among researchers about which role these reactions play. Armenakis, Harris and Mossholder

(1993) argue readiness is the cognitive precursor of either resistance or openness, while

Bouckenooghe (2008) sees them as opposite poles.

Either way, making the distinction between these two concepts explicit will help refine

the discussions of the implementation of change efforts (Armenakis et al., 1993). Therefore,

both resistance and readiness will be used to capture change recipients’ reactions to

organizational change.

2.1.1 Resistance

7

change. Change recipients might resist change because of this uncertainty for the future

(Coghlan, 1993) or because employees do not perceive the change as feasible (Dent and

Goldberg, 1999). Despite the extensive body of literature on resistance to change, Dent and

Goldberg (1999) argue that the knowledge on resistance to change has not been altered much

over the past decades. Piderit (2000) argues that the concept of resistance has certain

limitations, like the unidimensionality and the fact that it is mostly seen as a negative attitude.

Piderit (2000) argues that attitudes should be conceptualized along three dimensions (Katz,

1960; Ajzen, 1984) and any conceptualization that does not recognize all three dimensions is

incomplete. In a recent meta-analysis, Erwin and Garman (2009) reviewed the literature on

resistance and drew similar conclusions as Piderit. Moreover, they found some researchers

answered the call of Piderit about the multidimensionality when conceptualizing resistance to

change. Both Oreg (2006) and Szabla (2007) operationalize resistance to change as a

multidimensional construct in order to better understand the concept. These authors define

resistance as a tripartite attitude consisting of affective, cognitive and behavioral dimensions.

Piderit (2000) argues that these three dimensions not only operate simultaneously, but change

recipients may also be ambivalent in these three dimensions. This further complicates, but

also better captures the complexity of the concept of resistance (Erwin and Garman, 2009,

Piderit, 2000). Oreg (2006, p.76), in line with Piderit (2000), defines resistance as ‘a

tridimensional (negative) attitude towards change, which includes affective, cognitive, and

behavioral components’. Here, resistance is defined as a negative attitude. While this negative

view implies resistance is bad, some researchers have argued the opposite; it can be a valuable

source for constructive change (Ford, Ford and D’Amelio, 2008; Waddell and Sohal, 1998;

Weisbord, 1987). Despite the negative connotation given to resistance in the definition by

Oreg (2006), it is recognized that assuming resistance is negative and should be resisted

enables change agents to miss opportunities for using it for the better (Ford et al., 2008). How

change agents should act upon the reactions of change recipients is outside the scope of this

current research. For now the definition of Oreg (2006) will be used, since it recognizes the

multidimensionality and is used more often than Szabla’s (2007) definition.

2.1.2 Readiness

8

(1957), it was not until the work of Armenakis et al. (1993) that more serious interest in the

topic awakened (Bouckenooghe, 2008). To satisfy the need for a positive side of attitudes

toward change they researched people’s readiness for change. Armenakis et al. (1993) their

definition of readiness for change has been widely used when investigating readiness for

change and reads as follows: ‘readiness is an organizational member’s beliefs, attitudes and

intentions regarding the extent to which changes are needed and the organization’s capacity to

successfully make those changes’ (p. 681). Although they mention the three components of

attitudes (beliefs, emotions, and intentions), they have a strong focus on the cognitive

component (Bouckenooghe, 2010). Referring back to Piderit’s (2000) tripartite view on

attitudes, this focus on the cognitive dimension compromises the definition by Armenakis et

al. (1993). Bouckenooghe (2008) further developed readiness to incorporate all three

components equally. Similar to the development of a multidimensional concept of resistance

to change, he focused on all three dimensions to better capture the complexity of readiness for

change (Bouckenooghe et al., 2009). Where most others measured readiness for change

focused on a well-defined change (e.g. Holt, Armenakis, Feild and Harris, 2007; Armenakis,

Bernerth, Pitts and Walker, 2007), the scale by Bouckenooghe et al. (2009) also incorporates a

general component. His cognitive readiness scale is more an attitude towards change in

general, while the emotional and intentional readiness are both reactions to a specific change

(Bouckenooghe et al., 2009, p 576). For this research, the widely used definition by

Armenakis et al. (1993) is selected. However, the remark on the emphasis on the cognitive

component made by Bouckenooghe et al. (2009) will be taken into account by selecting their

measure for readiness for change.

9

change might be comparable to each other. Although no specific hypothesis will be made

about the connection between resistance to change and readiness for change, this research will

investigate this connection and discuss the findings. If they truly are opposites, a strong

negative correlation is expected and the results of the other hypotheses should be contrasting.

2.2 Commitment to change

The interest in employee commitment in change situations has been high and according to

various authors it is one of the most important factors influencing employees’ reactions to

change (Coetsee, 1996; Iverson, 1996; Vakola and Nikolaou, 2005). Moreover, it has been

said that commitment and the success of organizational change are related (cf. Herscovitch

and Meyer, 2002; Iverson, 1996; Neves, 2009; Oreg et al., 2011). Nonetheless, commitment is

a hard to define concept and there is little consensus about definitions (Elias, 2009). Until

recently, most research focused on organizational commitment; however, much criticism has

been expressed regarding construct definition (Meyer and Allen, 1991). In their research,

Meyer and Allen (1991, p. 62) define organizational commitment as the ‘feelings and/or

beliefs concerning the employee’s relationship with an organization’ and that it is a mind-set.

With their work they expand upon the concept of organizational commitment by posing that

this mind-set is not restricted to value and goal congruence. According to them, it represents a

desire, a need, and/or an obligation to maintain membership in the organization. They

distinguish three components in organizational commitment and named these affective,

continuance, and normative commitment. In the following years their three-component model

received considerable empirical attention (see Meyer, Stanley, Herscovitch and Topolnystky,

2002). In their meta-analysis, Meyer et al. (2002) found support for the three-component

model. Although some revisions were made to the scales (e.g. Meyer, Allen and Smith, 1993),

the three component model of organizational commitment has been related to various

desirable work outcomes such as attendance, job performance, and organizational citizenship

behavior (Meyer et al., 2002).

10

(2002, p.475) define commitment to change as ‘a force (mind-set) that binds an individual to a

course of action deemed necessary for the successful implementation of a change initiative’.

Furthermore, they refer to Meyer and Allen (1991) when they argue that commitment can take

several forms. First, there is affective commitment to change, which is defined as the desire to

provide support for a change based on a belief of its inherent benefits. Second, they

distinguish continuance commitment to change, or the perceived costs associated with failure

to provide support. Finally, normative commitment to change is the perceived obligation to

provide support for the organization (Herscovitch and Meyer, 2002). Commitment to change

has been subject of study and is associated with several outcomes such as implementation

success and performance improvement (Parish, Cadwallader and Bush, 2008), and turnover

intentions (Cunningham, 2006). Moreover, commitment to change can be seen as an outcome

itself (Bouckenooghe, 2012). It is argued that all three components will have a positive

influence on the chances an employee will stay with the organization, but they are considered

to have quite different implications for on-the-job behavior (Herscovitch and Meyer, 2002).

Nevertheless, research on commitment to change in relation to recipients’ reactions to change

is limited (Conway and Monks, 2008; Meyer, Srinivas, Lal, and Topolnytsky, 2007) and

research specifically linking commitment to change to either resistance or readiness is even

scarcer. Therefore, in this article it will be investigated whether these different types of

commitment to change influence either resistance to change, readiness for change, or both and

to what extent.

2.2.1 Affective commitment

11

also related to these behaviors, but weakly, and continuance commitment was negatively

related. These cooperating and championing behaviors indicate an intentional dimension of

readiness. Affective commitment is often found to have the strongest impact on attitudes of

recipients (Conway and Monks, 2008; Herscovitch and Meyer, 2002; Meyer et al., 2002).

Meyer et al. (2007) replicated this strong relation in their study of commitment to change in

two different cultures and change contexts. On top of that, Meyer et al. (2002) found that

affective commitment is positively related to citizenship behavior; doing more than is

expected or required of you. Yousef (2000) tested the effects of the three dimensions of

organizational commitment on attitudes toward change in a non-western setting. Although he

used organizational commitment, several studies (e.g. Meyer et al., 2007) have shown that

organizational commitment and commitment to change have similar outcomes. Yousef’s

(2000) results suggest that affective organizational commitment is related to the affective and

behavioral tendency attitudes toward change. Recipients having higher levels of affective

commitment will feel and act more positive toward the change.

Based on the above findings it can be argued that affectively committed recipients are

less likely to exhibit resistance, since they are willing support the change and to make

personal sacrifices in order to ensure change success (Herscovitch and Meyer, 2002).

Moreover, since affectively committed individuals show more positive attitudes (Lau and

Herbert, 2001) it can be expected they believe in the organization’s capacity to make the

changes work. Therefore, it is expected that:

Hypothesis 1

: Affective commitment to change has a negative influence on change

recipients’ resistance to change.

Hypothesis 1

: Affective commitment to change has a positive influence on change

recipients’ readiness for change.

2.2.2 Normative commitment

12

found support for this prediction made by Herscovitch and Meyer. Surprisingly, they failed to

provide evidence for a clear distinction between affective and normative commitment to

change (they attribute this discrepancy due to culture differences). Nevertheless, higher levels

of normative commitment to change are associated with a higher willingness to put in effort

beyond minimum requirements to ensure successful change even if it requires some personal

sacrifices (Meyer et al., 2007). This is consistent with Herscovitch and Meyer (2002), who

found that normative commitment to change was also associated with cooperation and

championing behaviors (i.e. intentional dimension), albeit to a lesser extent than affective

commitment to change. On the contrary, Parish, Cadwallader and Bush (2008) found no

significant relationship between normative commitment to change and the successful

implementation of a change initiative or with improved performance. Apparently, in their

study, higher levels of normative commitment did not lead to the favorable behaviors that

promote change. In his study, Yousef (2000) found no significant relationships between

normative commitment and attitudes toward change.

Despite the fact that normative commitment to change is associated with higher effort

and cooperation and championing behaviors, there is evidence that normative commitment to

change has a (much) smaller influence on change recipients’ attitudes than affective

commitment. Since recipients feel an obligation to support the change, resistance is expected

to be lower for higher levels of normative commitment. However, when one is normatively

committed they are less motivated to support the change (no identification with the change, it

is perceived as an obligation), the influence on readiness is expected to be lower. Therefore, it

is hypothesized that:

Hypothesis 2

: Normative commitment to change has a negative influence on change

recipients’ resistance to change.

Hypothesis 2

: Normative commitment to change has a positive influence on change

recipients’ readiness for change.

2.2.2 Continuance commitment

13

words, they assess the perceived switching costs as higher. Herscovitch and Meyer (2002)

found continuance commitment to change to be related with compliance with the change, but

not with cooperation or championing. However, Fable and Yukl (1992) see compliance as a

minimum condition that is often insufficient to ensure success, especially in more complex

situations. Results similar to Herscovitch and Meyer (2002) were obtained by Meyer et al.

(2007). Yousef’s (2000) results support the hypothesis that continuance organizational

commitment (defined as low perceived alternatives) directly and negatively influences

cognitive attitudes toward organizational change. Recipients with high continuance

commitment might think more positive about the change, but do not feel or act more positive

toward it. However, when Yousef defined continuance organizational commitment as high

personal sacrifice, the results did not support any relation between continuance commitment

and attitudes toward change. According to Kalyal, Berntson, Baraldi, Näswall and Sverke

(2010) employees experiencing high continuance commitment can feel trapped in their job

roles due to inability to handle changes successfully. Those individuals commit to the change

while they do not relate to it or find it beneficial and, as a result, show minimal support for it

(Kalyal et al., 2010). Relating this to Yousef’s (2000) findings, recipients with high

continuance commitment are not willing to make high personal sacrifices. Though

continuance commitment does not indicate resistance, it is a less positive form of commitment

(Kalyal et al., 2010).

Given the evidence that continuance commitment is associated with compliance,

minimal support, and unrelated to supportive behaviors (Herscovitch and Meyer, 2002) and

based on the assertion made by Kalyal et al. (2010) that it does not indicate resistance, it can

be expected that higher levels of continuance commitment do not lower resistance to change

either. High levels of continuance commitment to change are often found to be insignificantly

associated with positive attitudes toward organizational change (Herscovitch and Meyer,

2002; Meyer et al., 2007). Moreover, because recipients exhibiting this commitment can feel

trapped in their job roles (Kalyal et al., 2010) readiness is expected to be low. Hence, the

following can be hypothesized:

Hypothesis 3

: Continuance commitment to change has no influence on change

recipients’ resistance to change.

Hypothesis 3

: Continuance commitment to change has no influence on change

14

2.3 Communication

It is generally accepted that communication plays a crucial role during organizational change

(Armenakis et al., 1993; Russ, 2008). Communication is considered a process factor or, in

other words, it relates to how change is dealt with (Bouckenooghe et al., 2009). However, in

developing the Organizational Change Questionnaire (OCQ) Bouckenooghe et al. (2009)

delineated participation, support by supervisors and attitude of top management also as

process factors. These process factors concern the actual approach that is taken on how to deal

with a change project and can thus be managed (Bouckenooghe et al., 2009). Given the

crucial role of communication in change initiatives and the limited scope of this paper,

communication about the change will be the process factor that is examined in this research.

According to Dubrin and Ireland (1993) fear, caused by uncertainty, is a powerful motivator

behind resistance. Communication can reduce this uncertainty and threat associated with

organizational change (Schweiger and DeNisi, 1991; Rafferty and Restubog, 2010) by

providing timely and accurate information concerning the change (Bastien, 1987; Jimmieson,

Terry and Callan, 2004). When change recipients understand why the company is undertaking

a change initiative, they will more readily accept it (Ulrich, 1998). This information can be

communicated by using both formal or informal communication channels (Ashford, 1988).

Using this information, recipients go through a process of sense-making to establish a sense of

prediction and understanding of the situation (Sutton and Kahn, 1986). According to Elving

(2005) communication about organizational change serves three goals; (1) reducing

resistance, (2) informing employees about organizational matters, and (3) creating a

community. Accordingly, communication should lead to effective change resulting in low

levels of resistance, high levels of readiness and high commitment among change recipients

(Elving, 2005).

15

past that is used to prepare for change. Moreover, Armenakis et al. (1993) distinguish a third

approach that is named ‘management of external information’. Using this approach, the

change agent uses external sources of information to strengthen the change message. There

are two ways in which a change agent can use these external sources. One is to provide

information to external sources (e.g. the press), the other is to distribute external sources of

information to recipients (e.g. books or articles).

Despite the insight this conceptualization gives in which role communication can play

during change, it is just one of many approaches to communicate organizational change

(Russ, 2008; Russ, 2009). In his article, Russ (2008) advances the conceptual understanding

of communicative treatments for implementing change by clarifying two broad approaches.

He synthesizes the different conceptualizations that currently exist into two theoretical

categories: programmatic change communication and participatory change communication.

Programmatic change is focused on providing top-down information to recipients and

resembles the ‘persuasive communication’ by Armenakis et al. (1993). Participatory has a

focus on actively involving the recipients in the change, much like the ‘active participation’

by Armenakis et al. (1993). The third approach identified by Armenakis et al. (1993) is not

specifically included in the conceptualization of Russ (2008), but would likely fall under the

programmatic approach, because it consists of the one-way information dissemination typical

for programmatic communication activities. Since Russ builds on a broad basis of literature,

this article will draw on the classification by Russ (2008) for explaining the role of

communication during change. These two approaches will be elucidated below.

2.3.1 Programmatic communication

16

implemented is a key component of programmatic approaches. In order to do so, the message

that is sent should be identical to the message that is received in order to attain a high degree

of fidelity (Russ, 2008). Taking this one step further, influencing the behavior of recipients

can also be done using certain influence tactics (Bennebroek Gravenhorst, and Boonstra,

1998; Falbe and Yukl, 1992; Yulk and Falbe, 1990; Yulk, Fu and McDonald, 2003; Yulk, Kim,

and Falbe, 1996). These influence tactics are associated with varying impacts on commitment,

compliance and resistance (Falbe and Yukl, 1992). For example, more programmatic tactics as

pressure and legitimizing might result in compliance and/or resistance.

In programmatic communication approaches it is assumed the power lies with the

change agent and it is of high importance to gain recipients’ compliance (Russ, 2008). It is

implied that careful and explicit preprogramming the implementation procedures can reduce

or eliminate elimination problems (Bermann, 1980). In addition, communicating the ‘right’

message using the ‘right’ approach will minimize such problems as well (Russ, 2008).

Empirical evidence supports the importance of this need for information by recipients (Covin

and Kilmann, 1990). Armenakis and Harris (2002) mention that, from a change agent’s

perspective, compliance is seen as essential in preventing problems and achieving the desired

vision. However, as mentioned before, compliance is seen as a minimum condition which is

often insufficient to ensure success (Fable and Yukl, 1992). According to Russ (2008)

compliance is often sought from stakeholders who have an interest in the operational

execution of the implementation of the planned change in order to minimize resistance among

recipients (Russ, 2008). Compliance can be seen as preferable to resistance, but not as useful

as commitment to change (Fable and Yukl, 1992).

17

reduced uncertainty and anxiety, respectively, it did not elicit more positive reactions toward

the change. These results might indicate that simply providing information does not result in

reduced resistance, rather employees base their decision to resist or not on their agreement

with the proposed change (Oreg, 2006). As Greenwood and Hinings (1996) mention, interest

dissatisfaction (or disagreement with the change) is a cause of internal pressures and can

obstruct change. This complexity might be dealt with using, what Russ calls, a ‘need-to-know’

strategy: selectively communicating a carefully chosen message to a specific audience that

highlights specific elements to appeal to one’s interests. Furthermore, Oreg (2006) argues that

both too much as well as too little information can be harmful and increase recipients’

resistance to change. Too little information can increase uncertainty and subsequently

resistance, too much information can give all the more reason to resist a change when it is not

in your interest.

Returning to the influence tactics, the one associated most with programmatic

communication is rational persuasion. This is defined as using logical arguments and factual

evidence to persuade recipients (Yukl and Fable, 1990). In their study, Fable and Yukl (1992)

found that rational persuasion led to resistance, but this was not replicated in other studies.

Based on the assertion made in the literature that more programmatic communication (i.e.

more top-down information about the change) reduces uncertainty of change recipients and

this reduced uncertainty is expected to lead to lower levels of recipients’ experienced

resistance. However, taking the findings of Oreg (2006) and Kramer et al. (2004) into

account, this relationship is not linear. Sometimes, learning more about the change can give

change recipients all the more reason to resist it (Oreg et al., 2011). Following Oreg’s

reasoning, it is expected that the relation between programmatic communication and

resistance is U-shaped. Therefore, the following is hypothesized:

Hypothesis 4: The relationship between programmatic communication and resistance

to change is U-shaped.

18

Armenakis et al. (1993) for these components). Along the same lines, Elving (2005) argues

that communication about the motives for change, by reducing uncertainty, will create

readiness. The quality and timeliness of communication is crucial to increase readiness, for it

helps recipients to make sense of changes under way, makes changes more salient and helps

reframe them (Bouckenooghe and Devos, 2008; Weick, 1995). Bouckenooghe (2008)

mentions communication about the change as the primary mechanism for creating readiness

for change. A contrasting finding is provided by Bouckenooghe and Menguç (2010). They

found that recipients had a lower level of cognitive readiness for low levels of programmatic

communication. However, they mention that this can be explained by the fact that politicking

played a vital role, reasoning that a highly political environment combined with strong

programmatic communication is perceived by recipients as persuasive and aiming at

compliance, leading to reduced readiness.

So far, programmatic communication served the first part of the definition of

readiness, in that it clarifies ’the extent to which the changes are needed’. It is harder to find

articles that also explain how (programmatic) communication influences the second part of

the definition; the organization’s capacity in successfully making the change. Many authors

argue that communication is essential in creating readiness (e.g. Armenakis and Harris, 2002;

Elving, 2005; Ford and Ford, 1995; Jones, Jimmieson and Griffiths, 2005; Rafferty et al.,

2013), but none are very specific in how this process is shaped. Based on the argument that

clear communication to change recipients engenders a more favourable attitude toward the

change (Fairhurst, 1993; Lewis, 2006) and the general notion that communication creates

readiness, it is hypothesized that:

Hypothesis 5: Programmatic communication (top down communication) is positively

related to recipients’ readiness for change.

Literature on the possible relationship between communication and commitment to change is

scarce. This is further complicated by the fact that communication is often used in a general

sense or categorized in various ways (e.g. horizontal or vertical), which does not conform to

the programmatic and participatory distinction of Russ (2008). Rafferty and Restubog (2010)

found little research on these approaches. For that reason, this research will also draw upon

literature describing more general concepts and the differences will be made explicit.

19

2001) and as a resource that can strengthen the level of commitment (e.g. Bastien, 1987;

Kotter, 1995; Trombetta and Rogers, 1988). Ulrich (1998) considers communication, defined

as the candidly and frequently sharing of information, as one of the tools for developing

commitment. In their meta-analysis, Postmes, Tanis and de Wit (2000) show a positive

relation between (vertical and horizontal) communication and organizational commitment, but

this is stronger related for vertical communication. It is argued that one’s organizational

commitment is ‘more strongly related to their appreciation of the management’s

communication’ (p.240). These results were later reciprocated by Bartels, Peters, de Jong,

Pruyn and van der Molen (2009). However, a remark should be made, because in both studies

vertical communication also consisted of a participative part. Chawla and Kelloway (2004)

found that communication indirectly influences commitment (measured as turnover

intentions) through openness to change and trust. However, they did not distinguish in forms

of communication either.

A more promising insight is offered by Rafferty and Restubog (2010). They made a

distinction between programmatic and participatory communication in researching whether

these communication approaches influence affective commitment to change. Unfortunately,

the results were not significant, indicating that programmatic communication did not alter the

recipients’ beliefs about the benefits associated with the change. They performed post-hoc

interviews to understand the underlying reasons for this result and concluded that the

employees did not trust senior management (a context variable in the OCQ by Bouckenooghe

et al. (2009)). Bouckenooghe (2012) also tested the influence of programmatic

communication on commitment to change and found no significant relation with affective

commitment to change. He did find that programmatic communication moderates the effect of

organizational politics on continuance commitment to change, strengthening the relationship.

This means that employees will experience more pressure to support the change when it is

pushed downwards by management (Bouckenooghe, 2012). In contrast to Rafferty and

Restubog (2010) and Bouckenooghe (2012), Conway and Monks (2008) did find a positive

relation between programmatic communication and affective commitment to change.

However, it should be noted that the participants in Conway and Monks’ study were rather

negative about communications in relation to that change.

20

Hypothesis 6: Programmatic communication (top down communication) does not have

an influence on change recipients’ commitment to change.

2.3.2 Participatory communication

The second approach identified by Russ (2008) is participatory communication to change. In

her research, Lewis (1999) found that participatory communication (soliciting input from

change recipients) was not the preferred choice of respondents in implementation teams from

a variety of countries around the world. However, participatory approaches account for most

research on communication in change context (Russ, 2008). The difference with

programmatic communication is that the change agent invites recipients to give input (Lines,

2007). Change agents might ask input from recipients during the development,

decision-making processes and/or implementation of organizational change (Russ, 2008). The basic

assumption with this approach is that recipients should be active participants in the change

process, such that recipients have a voice during the change (Russ, 2008). Participatory

communication to change can be defined as ‘leveraging dialogic communication so as to

involve most or all stakeholders through solicitation of their ideas and input about the change

and the implementation process’ (Russ, 2008, p. 200). It invites input from recipients to shape

the change rather than just ‘receiving’ it. By actively participating, recipients shape, construct,

and implement the change themselves, thus contributing to the building of consensus and

galvanizing support (Russ, 2008). As Russ (2008, p. 204) mentions, ‘this approach is not

necessarily about the basic act of participation, but is about whether employees, in the end,

have a voice during planned organizational change’.

21

If change agents hope to gain support for, or at least minimize resistance to, a change

they need to capture and consider change recipients’ perspectives and underlying rationale

(Weisbord and Janoff, 2005). As early as Coch and French (1948) it was concluded that

participation can reduce resistance to change. They conclude that those who participated in

the design and development of the change had much lower resistance. Lewis (2006) sees it as

key during change implementation. Since communication reduces uncertainty because

recipients learn about the change (Jimmieson et al., 2004), participatory communication

should also decrease resistance because recipients gain more knowledge about the change.

Research has confirmed that participation in decision making is beneficial in minimizing

resistance (Bordia et al., 2004; Lines, 2004; Sagie, Elizur and Koslowsky, 1990). However,

participatory communication also has its limitations. It can result in ambiguity about the

original intent of the change, such that people who prefer clarity may grow weary of and

possibly frustrated with the change (Russ, 2008). Moreover, if the participatory

communication to change approach is seen as insincere by recipients it can create resentment

(Russ, 2008). Lewis (2006) confirmed that when employees perceive that their input is

valued, they consider the change as more successful. Based on the evidence that participatory

communication is beneficial in reducing resistance (Lines, 2004), but that not everyone may

like it, it is hypothesized that:

Hypothesis 7: Participatory communication has a negative effect on recipients’

resistance to change.

22

participation would lead to increased readiness for the change through indirectly sending

readiness messages. Recipients place greater trust in information they discover themselves

(through these indirect messages), thereby increasing their readiness (Armenakis et al., 1993).

In later works, they confirmed this argument (see Armenakis and Harris, 2002; Armenakis and

Harris, 2007). Furthermore, soliciting input from recipients in the decision making process

related to the change creates a sense of control (Armenakis et al., 1993; Dirks, Cummings and

Pierce, 1996). According to Cunningham et al. (2002) recipients reported higher levels of

readiness for change when they had more perceived control over their job. Recipients’

involvement in the change is considered central to increasing their acceptance of the change

(Kotter, 1995; Wanberg and Banas, 2000) as well as increasing their openness to change

(Devos, Buelens, and Bouckenooghe, 2007). These participatory approaches elicit positive

reactions by giving recipients greater accessibility to information and a voice in the change

process (Russ, 2008). In addition, Szabla (2007) investigated which influence different

leadership strategies had on change recipients reactions. He concluded that recipients’

perceived participation in the change process elicited positive beliefs, emotions and intentions

and is effective in gaining full support for an organizational change. In their development of a

readiness scale, Holt et al. (2007) found support for the positive relation between perceptions

of communication climate and the readiness-scale. Open communications can enhance

positive attitudes towards change by reducing fear and conveying the organization’s

competence in making the change happen (Mayer, Davis and Schoorman, 1995). Based on the

above reasoning, it is hypothesized that:

Hypothesis 8: Participatory communication has a positive effect on change recipients’

readiness for change.

23

recipients are involved in the change process using more substantive participatory approaches,

their commitment to change tends to be higher (Armenakis and Harris, 2002; Devos et al.,

2002; Lines, 2004). As Devos et al. (2002) note, if recipients want to commit themselves to

the change, they want to be informed about the change and be involved in the development

and implementation of the change. In line with this reasoning, Niehoff, Enz and Grover

(1990) found that communicating a shared vision was significant in increasing

(organizational) commitment. As Maurer (2005) concludes, the interaction between change

agent and recipient is critical in creating and sustaining commitment to change.

Drawing on justice research the influence of participatory communication on

commitment to change can be explained. As Folger (1977) mentions, perceived fairness is

likely to depend on whether those affected by the change are invited to communicate their

opinions. Self, Armenakis and Schraeder (2007) continue this argument and claim that

recipients who perceive that the organization has treated them fairly may respond with loyalty

and commitment. Moreover, communication shapes images, to which recipients subsequently

commit themselves (Fairhurst, 1993; Treadwell and Harrison, 1994). Van Vuuren, de Jong and

Seydel (2007) conclude that through a clear view of values, goals and effectiveness of the

organization, communication strengthens organizational

commitment. Applying this to a

change situation, a positive effect of these participatory communications on commitment to

change is expected. Therefore, it is expected that:

Hypothesis 9: Participatory communication has a positive influence on change

recipients’ commitment to change.

2.4 Conceptual model

24 H7 H6 H9 H1a H6 H2a H6 H3a H6 H6 H9 H8 H1b H6 H2b H6 _H3b H6

Figure 1. Conceptual model including hypotheses. Programmatic communication Participatory communication Commitment to change Affective Normative Continuance Resistance to change H4 Programmatic communication Participatory communication

Readiness for change

25

3. Methodology

In order to find out whether communication and commitment to change influence change

recipients’ reactions quantitative and qualitative research was conducted. Quantitative

methodology can be described as experimental or manipulative; proposing hypotheses, testing

and verifying them while ensuring confounding conditions to prevent outcomes from being

influenced in an improper way (Guba and Lincoln, 1994). Quantitative research has a realist

orientation, which implies reality exists independent of human perception (Sale, Lohfeld and

Brazil, 2002). According to Slevitch (2011, p. 76) ‘[quantitative] research findings are be

viewed as ‘true’ or valid, as long as prescribed procedures are rigorously followed’. However,

a sufficiently large sample size is critical in quantitative research.

On the contrary, qualitative research takes an idealist standpoint claiming that reality is

constructed or interpreted by an individual and that there is no single reality (Smith, 1983). As

Slevitch (2011) puts it; because reality is socially and psychologically constructed, it is

viewed as an intersubjective creation. In other words, it is continuously recreated by recipients

based on their intersubjective understanding (Hellström, 2008). As a result, things can only be

described as how we perceive them and not as a ‘reality’ (Guba and Lincoln, 1994). In

qualitative research sample size becomes irrelevant (Slevitch, 2011). Rather than being

representative of a larger group, samples are evaluated on the ability to provide rich

information (Hellström, 2008).

26

(1989) see as ‘complementarity’. According to Bryman (2006), this is also the most often

cited justification for combining quantitative and qualitative methods.

3.1 Data collection

Two studies were undertaken to investigate the hypotheses as proposed in Figure 1. The first study was performed at KPN consulting, a national consultancy company. The change involved a structure-changing program. The aim of the change was to become more competitive in the Dutch market by changing the way the consultants worked. Before the change, consultants were detached to an external client full-time, performing their work mainly at the client’s location. The new way of working consisted of the creation of subunits focused consulting, rather than detachment. With this focus on consulting, KPN is trying to get ready for the future. Furthermore, the consultants’ work changed from working on one job to working on multiple jobs at the same time, for different clients. When there is time, consultants are expected to work on the development of consulting tools as well.

The second study took place at Qbuzz, a transit company operating mainly in the Northern part of the Netherlands. The change taking place in this company related to a quality improvement program. The transit market in the Netherlands works with concessions, granting a company the monopoly of the transit market in a specific part of the Netherlands. However, every couple of years these concessions are reconsidered and the transit company has to make a new offer. In order to take the next step in quality improvement, Qbuzz started a program to guide this change. This program concerns every indirect staff member at Qbuzz. Meetings are held to discuss how everyone can contribute to quality improvement in order to contribute to the success of Qbuzz.

Study 1 Questionnaire

This study used the software program ‘Qualtrics Survey Software’ to conduct self-administered questionnaire. The link to the questionnaire was delivered through e-mail by the liaison or the researcher to all employees involved in the change. An introduction was given at the opening of the questionnaire, introducing the research and researcher, explaining the purpose of the study and ensuring anonymity and confidentiality. A reminder was sent a week after the first email, again containing the link to the survey.

27

KPN consulting were taken into account before the questionnaire was sent to every employee in the subunits.

Respondents could indicate their answers to the questions on a five-point Likert scale, ranging from strongly disagree (1) to strongly agree (5). Questions were limited to three per page in order to keep an organized overview during the questionnaire. Respondents were invited to give comments and remarks related to the questionnaire. Two respondents made use of this option and gave remarks concerning the content of the questionnaire. These comments were used to improve the questionnaire for the second study.

Interviews

28 Population and responses

Eventually, twenty-six (out of a possible fifty) employees completed the questionnaire, which amounts to a response rate of 52%. The average age of the respondents was approximately 34 years. From the three subunits, roughly half the responses can be attributed to one subunit, while the other responses are spread across the other two subunits. This is inconsistent with the population, where the subunits are approximately the same size. See Table 1 for additional information. A further question considered the time employees joined the subunits

(which are the result of the proposed change). Most employees joined at the formation of the subunits, while only three employees indicated that they joined later in the process.

Table 1. Descriptive statistics study 1. Study 2

Questionnaire

The questionnaire used in the second study was a copy of the questionnaire used in the first study. Only minor adjustments were made to make the questionnaire fit for this population. Moreover, respondents were asked for their tenure. Again, a reminder was sent a week after the first mail to the respondents was sent out. In this study no interviews were conducted.

Population and responses

A different picture emerges from the date of Qbuzz. From the 115 employees approached, 49 employees filled out the complete questionnaire, which constitutes a response rate of 42.6%. A majority of the respondents was male (73.5% versus 26.5% female). The youngest respondent was 23 years old, while the oldest was 64 years old. However, the average was 50.5 years old, indicating a skewed distribution. The tenure differed from just 1 year up to 38 years, with an average of approximately 19 years.

3.2 Measurements

In order to test the proposed relationships, the concepts were measured using existing scales. These scales have all been tested and validated. Since all validated scales were multi-item measures, a factor analysis was conducted to ensure the different items loaded on the corresponding factor, see chapter 4. 3.2.1 Dependent

Resistance to change: To measure the concept of resistance to change, the existing scale of Oreg (2006) was used. This scale comprises three subscales measuring the affective (ResAf), cognitive (ResCog) and behavioral (ResBeh) resistance to change. An example of affective resistance is ‘I was

Frequency Percentage

Subunit 1 11 42.3%

Subunit 2 6 23.1%

Subunit 3 5 19.2%

29

afraid of the change’. Cognitive resistance was measured using, for example, ‘I believed that the change would make my job harder’ and an example item for behavioral resistance is ‘I protested against the change’. Each aspect of resistance consisted of 5 items in total.

Readiness for change: To measure the concept of readiness for change, Bouckenooghe’s (2009) scale was used. This scale encompasses 9 items and was also made up of subscales which measured the emotional (ReaEmo, three items), cognitive (ReaCog, three items) and intentional (ReaIn, three items) aspects of readiness. Examples items are ‘I have a good feeling about the change project’ for

emotional readiness; ‘Plans for future improvement will not come too much’ (reverse coded) for cognitive readiness and ‘I want to devote myself to the process of change’ for intentional readiness. 3.2.2 Independent

Commitment to change: In order to measure commitment to change, the scale developed by Herscovitch and Meyer (2002) is used. To the researcher’s knowledge this is the only scale that specifically measures commitment to change (in contrast to other types of workplace commitment). This scale consists of 18 items measuring the three subscales of affective (ComAf), continuance (ComCon) and normative (ComNor) commitment. For this research, the three highest loading questions (based on Herscovitch and Meyer (2002) their results) were used in the questionnaire. A notion should be made, the original scale was measured using a seven-point Likert scale, but for the conformity of this research the five-point Likert scale was adopted. An example item for the affective commitment to change scale is ‘I believe in the value of this change’. For continuance commitment to change an example item would be ‘I have too much at stake to resist this change’. Finally, normative commitment was measured using items like ‘I feel a sense of duty to work toward this change’.

Communication: For measuring the concept of communication, two scales were used. Since the typology of Russ (2008) is fairly new, there was little research done on these specific types of communication. However, Bouckenooghe (2008, 2012) made an effort by developing two scales to measure these concepts. For participatory communication, Bouckenooghe’s (2008) scale measuring participative involvement (adopted from the OCQ) was used. According to Bouckenooghe (2008) this measures the extent to which the communication about the change was clear and timely. It is about management creating support by using participation and communication. The scale consists of 6 items, of which ‘Departments are consulted about the change sufficiently is an example. To measure programmatic communication, Bouckenooghe (2012) developed a new scale to measure the formal or programmatic communication. The scale consists of four items and an example item would be ‘Information on change is mainly provided by management’.

3.2.3 Factor and regression analysis

30 Factor analysis

There are different types of factor analysis that can be performed. According to Cortina (1993) principal component analysis (PCA) is the recommended method for extracting factors. However, Guadagnoli and Velicer (1988) mention both PCA and PFA usually generate in similar results. Since PCA is concerned with establishing which linear components exist within the data and how a particular variable might contribute, it is deemed more appropriate for this analysis. Field (2005, p. 631) mentions PCA should not be described as factor analysis, but in this research the terms factor analysis and principal component analysis are used interchangeably. In the PCA, Oblimin rotation was used to provide a more accurate and realistic representation of the construct, because it allows factor to be related (Fabrigar, Wegener, MacCallum, and Strahan, 1999).

A general assumption is that a minimum ratio of respondents to variables is necessary and this ratio is about 5:1 (see Floyd and Widaman, 1995). According to Field (2005), most important in factor analysis is the absolute sample size and the absolute magnitude of factor loadings. The smaller a sample is the more frequent and higher the loadings should be in order to be significant. It is generally accepted that factor loadings of .40 are too low (Clarck and Watson, 1995; Floyd and Widaman, 1995). Field (2005) refers to Stevens (1992), who produced a table of critical values against which loadings can be compared. For a sample of 50 the minimum loading should be .722, while for a sample of 100 the minimum loading is .512. In this research a sample of 75 is used, therefore the minimum factor loading used in this research is .617. The sampling adequacy of the analysis is tested with the Kaiser-Meyer-Olkin measure and the factorability of the correlation matrix is tested with Bartlett’s test of Sphericity. The KMO should be at least .5 to be sufficient and at least .7 to be good (Field, 2005). The Bartlett’s test should be significant. Measures of sampling adequacy should exceed .5 as well (Field, 2005). The amount of variance accounted for by a component in the factor analysis is also known as the communality (MacCallum, Widaman, Zhang and Hong, 1999). Again, the sample size plays a role, but according to MacCallum et al. (1999) when communalities are high (> 0.6) this is not necessarily a problem.

In order to assess how many factors to extract, the Kaiser criterion is used. The Kaiser criterion shows how much information is accounted for by an average single item and should be at least 1.0 or larger (Floyd and Widaman, 1999). They also argue the Scree test is a more accurate method for assessing how many variables to retain. However, both methods have their flaws (Fabrigar et al., 1999). In this research, both methods will be used, in combination with the expectations from theory.

Regression analysis

31

mediator. The idea that commitment to change can act as an independent variable and as a mediator is based on the theory and on the conceptual model, respectively.

One of the most popular methods for testing hypotheses about mediation is the causal steps strategy by Baron and Kenny (1986). In this method, the investigator has to meet three conditions and follow three steps to determine whether a variable functions as a mediator. When variable M is the mediator, variable X the independent variable and variable Y the dependent, the steps are as follows, also see Figure 2. The first step is to regress X on M to see if the independent variable accounts for significant variability in the mediator (path a). The second step is to regress X on Y to see if the independent variable also significantly accounts for variability in the dependent variable (path c). The third and final step is to regress X and M on Y, to see if the mediator significantly accounts for variability in the dependent variable when controlling for X (path b) and the independent variable accounts for substantially less variability then before (path c’) (Baron and Kenny, 1986). The first two conditions are that the path of X to M is significant and that the path of M to Y is significant. The third condition entails that path c’ is significantly less than path c. To establish this significance, the Sobel test (Sobel, 1982) is often used. However, this test assumes a normal distribution and is therefore only useful in large samples (Preacher and Hayes, 2008). An alternative to the causal step strategy is multiple mediation. This involves simultaneous mediation by multiple variables (Preacher and Hayes, 2008). This method allows the researcher to determine the relative magnitudes of specific indirect effects associated with all mediators (Preacher and Hayes, 2008). Given the fact that the mediator in this study, commitment to change, encompasses three attitudes which can act simultaneously (Herscovitch and Meyer, 2002), a multiple mediation model offers a broader picture. Another advantage is the fact that bootstrapping (part of the multiple mediation model) is more powerful in testing intervening variable effects than the Sobel test or the causal steps approach (Preacher and Hayes, 2008; MacKinnon, Lockwood and Williams, 2004). For a more thorough explanation of the multiple mediation model and the advantages and disadvantages, see Preacher and Hayes (2008).

32

4. Results

This chapter will provide the results of the factor and the reliability analysis, the correlation analysis and the regression analysis. Furthermore, a qualitative analysis of the three interviews will be given. A summary of the results will be given, including the confirmation, or not, of the hypotheses. Important to mention is that the resistance to change and readiness for change variables will be factor analyzed on the subscale level in order to confirm if the used subscales for the measures hold. However, all hypotheses are on a general level, thus when these subscales are confirmed they will be combined into two aggregate variables (i.e. ResTotal and ReaTotal, also see the reliability analysis).

4.1 Factor and reliability analysis

A principal component analysis was performed on all variables from the dataset. Based on the theory, it was expected that 11 components would be extracted to divide the 43 variables. The PCA was performed using Oblimin rotation. Assessment of the relevant statistics revealed that the Kaiser-Meyer-Olkin measure (= 0.698) exceeded the recommended value of 0.6 (Kaiser, 1974) and the Bartlett’s test of Sphericity was significant, which supports the overall significant of all correlations in the correlation matrix. The anti-image matrix shows many Measures of Sampling Adequacy below the minimum of 0.5 (Field, 2005). All communalities were sufficient, which means they exceeded 0.6 (MacCallum et al., 1999). The Kaiser criterion indicates that 11 components (explaining 76.6% of variance) can be extracted, while the Scree plot is difficult to read. Omitting all loadings lower than 0.4 provides a more organized overview of the different components and relevant loadings, see Appendix IV. However, there are some cross loadings and most variables do not load neatly on the component they are expected to load on. This, combined with the low MSAs supported the decision to analyze the dependent and independent variables separately.

Dependent variables

33

A new principal component analysis was performed on all variables measuring the concepts of resistance using Oblimin rotation. The scale, developed by Oreg (2006), consists of three components, with five variables for each component; cognitive resistance (ResCog1-5), behavioral resistance (ResBeh1-5) and affective resistance (ResAf1-5). The analysis extracted three components (explaining 65.5% of variance) based on the Kaiser criterion. Based on the Scree plot, four components could be extracted. However, guided by theory it was decided that three components were to be extracted. The KMO was 0.723, which is sufficient, and the Bartlett’s test was significant. All MSA’s are sufficient ( > 0.593). The communality of ResCog3 (0.461) is below the threshold of of .6 (MacCallum et al., 1999). Therefore, ResCog3 was deleted from the analysis. ResBeh5 is deleted from the analysis because of a loading on the component of ResCog and because of its low correlation with the other variables measuring behavioral resistance. Now ResAf1 has no high loadings on any of the components and has a low communality (0.504). ResAf3 is problematic due to low and insignificant correlations with the other ResAf variables and the loading on the component of ResCog. ResAf2 is problematic due to a communality which borders the lower side of the threshold (0.609) and a loading on the ResCog component. Thus, ResAf1, ResAf2 and ResAf3 are deleted from the factor analysis. The final solution has three components (explaining 73.7% of variance), based on the Kaiser criterion and guided by theory, and a KMO of 0.749, exceeding the 0.6 threshold (Kaiser, 1974). The Bartlett’s test of Sphericity is significant, supporting the factorability of the model. All MSA are above 0.598, exceeding the 0.5 threshold (Field, 2005) and all communalities are above 0.653 (exceeding the 0.6 threshold mentioned by MacCallum et al., 1999). Table 2 shows the loadings of the variables, which all meet the minimum of 0.617 (Field, 2005).

Behavioral Resistance Cognitive Resistance Affective Resistance ResCog1 -,043 ,760 ,244 ResCog2 ,000 ,755 ,223 ResCog4Rev ,150 ,845 -,201 ResCog5Rev -,012 ,817 -,179 ResBeh1 ,870 ,036 ,076 ResBeh2 ,932 ,013 -,034 ResBeh3 ,876 -,041 -,058 ResBeh4 ,805 ,016 ,079 ResAf4 ,002 ,078 ,859 ResAf5 ,083 -,062 ,845 Eigenvalues 3.998 1.873 1.508 % of variance 39.879 18.729 15.076 Cronbach’s α .877 .811 .700

Table 2. Rotated factor loadings: Resistance to change components