Readiness for change,
A comparison of variables influencing readiness
Master Thesis, MSc. BA, specialization Change Management
University of Groningen, Faculty of Economics and Business
16
thof August, 2012
EMMA C. WEEDER
Student number: 1919814
Lutkenieuwstraat 44
9712AZ Groningen
tel: +31 (0)6 1974 6247
e-mail: emma@weeder.nl
Supervisor/ University
dr. C. Reezigt
drs. J.C.L. Paul
Supervisor/ field of study
ir. W.W.M. Ackermans
drs. D.L.J. Franken
Readiness for change,
A comparison of variables influencing readiness
ABSTRACT
Creating deeper understanding of the relationship of participation, communication, self-‐efficacy, discrepancy and principal support with the tripartite readiness for change construct (Piderit, 2000) is subject to this research. Two propositions are stated: the first proposition states that all five variables have a direct relationship with readiness. The second statement hypothesizes that participation and communication are moderating variables. Notable is the aggregation of the variables communication and principal support in the factor analysis and the aggregation of emotional and intentional readiness. The sample of respondents is taken from Eneco N.V. in Rotterdam. These employees are subject to a strategic change. The research resulted in an affirmation of the first hypothesis stating that several direct relationships exist. The level of connection gives rise to further research matching the second hypothesis. The study sample does not support a positive moderating relationship, several significant negative moderations were found. These results give rise to further research of a larger sample and a repetition of the data gathering along the change process.
Keywords: emotional readiness, intentional readiness, cognitive readiness, participation, communication, self-‐efficacy, discrepancy, principal support, change readiness.
Acknowledgments:
TABLE OF CONTENTS
1. INTRODUCTION...4
2. THEORETICAL FRAMEWORK ...6
2.1 READINESS...6
2.2 INDEPENDENT VARIABLES...9
2.2.1 DEFINITIONS OF THE INDEPENDENT VARIABLES...9
2.2.2 DIRECT RELATIONSHIP OF INDEPENDENT VARIABLES WITH DEPENDENT VARIABLE...11
2.2.3 MODERATING INFLUENCE OF COMMUNICATION AND PARTICIPATION...13
2.3 THEORY WRAP-UP... 16
3. METHOD ... 17
3.1 CASE STUDY: ENECO N.V. ... 17
3.2 DESK RESEARCH... 18 3.3 FIELD RESEARCH... 18 3.4 PROCEDURE... 19 3.4.1 PARTICIPANTS...20 3.5 DATA ANALYSIS... 21 3.5.1 FACTOR ANALYSIS...22
3.5.2 CRONBACH’S ALPHA ANALYSIS...24
3.5.3 REGRESSION AND MODERATOR ANALYSIS...24
4. RESULTS ... 27
4.1 ANALYSIS AND FINDINGS... 27
4.1.1 MULTIPLE REGRESSION ANALYSIS...27
4.1.2 MODERATOR ANALYSIS...31
5. DISCUSSION... 39
5.1 FACTOR ANALYSIS... 39
5.2 HYPOTHESIS 1: DIRECT POSITIVE RELATIONSHIP OF THE INDEPENDENT VARIABLES WITH THE DEPENDENT VARIABLE... 40
5.3 HYPOTHESIS 2: MODERATING RELATIONSHIP OF PARTICIPATION AND COMMUNICATION.. 43
6. CONCLUSION AND RECOMMENDATIONS ... 47
6.1 CONCLUSION... 47
6.2 MANAGERIAL IMPLICATIONS... 48
7. LIMITATIONS AND FURTHER RESEARCH ... 49
7.1 LIMITATIONS... 49
7.2 FURTHER RESEARCH... 50
8. REFERENCES... 52
APPENDICES ... 56
APPENDIX A 1: DUTCH QUESTIONNAIRE ENECO EMPLOYEES... 56
4
1. INTRODUCTION
Readiness for change is an intensively researched topic, originating with the well-‐known change research of Lewin (1947). Recent research has focused on a number of independent variables that influence employee readiness. Resulting from their research, authors have designed numerous models describing the influence different variables have on readiness for change (Armenakis, Harris and Mossholder, 1993; Armenakis, Harris and Feild, 1999; Armenakis, Bernerth, Pitts and Walker; 2007; Armenakis and Harris, 2009; Bouckenooghe, Devos and Van den Broeck, 2009; Holt, Armenakis, Feild and Harris, 2007; Oreg, 2006; Oreg and Berson 2011B).
This thesis will review the relationship of five independent variables with readiness for change. The variables discussed are split into two groups. First, communication and participation as moderating variables (Armenakis et al., 2007; Armenakis and Harris, 2009; Bouckenooghe et al, 2009). Secondly, discrepancy, self-‐efficacy and principal support as independent variables (Armenakis et al., 2007). Authors have contrary opinions on the direct and indirect relationship between these variables and readiness. The concept of readiness is divided into three sub-‐concepts that come into play during the change: intentional, cognitive and emotional readiness (Armenakis et al., 1993; Bouckenooghe et al., 2009; Oreg, 2006; Piderit, 2000). Researching these five variables in accordance to readiness could create a deeper understanding of the value of known models. Additionally it could create new insights on these specific relationships.
change. For this research external factors influencing readiness for change are excluded.
The objectives of this research are twofold. The first goal is to gain greater understanding of the relationship between the selected variables and readiness for change. Two propositions supported by two conceptual models are tested to explain the relationship between the variables and readiness. This is most important considering the academic context this thesis is written in. The second goal is to get a deeper understanding of the success of the implementation of the strategy change at Eneco Holding N.V.. Measuring the level of readiness of the employee in this specific case. The practical results of the study will be reported in a company report. The data gathered at Eneco forms the data set used to test the hypotheses stated.
6
2. THEORETICAL FRAMEWORK
The concepts of several scholars (e.g. Armenakis and Oreg) describe the characteristics and conditions that are associated with successful organizational change (Oreg, 2006). This research tests two hypotheses. First the direct relationship of discrepancy, self-‐efficacy, principal support, communication and participation with the tripartite readiness concept is tested. Second the moderating relationship of communication and participation on the relationship of discrepancy, self-‐efficacy and principal support and readiness is looked at. The test of the direct relationship of the variables with readiness for change is a precondition for the test of the moderating relationship. Armenakis et al. (2007) and Oreg (2006) support the focus on the change recipient. Armenakis et al. (2007) state that: what the recipients believe about the change is an essential influence on the outcome of the change.
2.1 Readiness
Every individual experiences acceptance of change differently. For some it implies a source of joy, benefits, or advantages, whereas for others it is a source of suffering, stress, and disadvantages (Bouckenooghe, 2010). This characterizes the unique experience we all have when encountering a changing situation. In such an organizational transformation, change recipients make sense of the new situation. They hear, see and experience the change. Cognitions, emotions, and intentions become part of their decision processes that result in resistance or readiness for change (Armenakis et al., 2007).
behaviors of either resistance to, or support for, a change effort”. This definition recognizes the three sub-‐concepts intentional, emotional and cognitive readiness as described by Piderit (2000). Holt et al. (2007) and Bouckenooghe et al. (2009) complement each other on the assumption that readiness for change consists of four elements: content, process, context and the individuals engaged in the change. Readiness is described by a number of authors as the individual readiness for change according to three dimensions: emotional, cognitive and intentional readiness (Bouckenooghe et al. (2009), Oreg, Vakola and Armenakis (2011A) and Holt et al. (2007), Piderit, (2000)). Bouckenooghe et al. (2009) derive the following definition from Oreg et al. (2011A) and Piderit (2000): “Emotional readiness for change is the affective reaction towards change. Cognitive readiness for change is the beliefs and thoughts people hold about the (dis) advantages of the change. Intentional readiness for change is the extent to which employees are prepared to put their energy into the change process (Bouckenooghe et al., 2009: 599)”.
Readiness and resistance are concepts that are extensively described by scholars. Coch and French (1948) are commonly known for their early research on influences on resistance and readiness for change. An unambiguous definition of readiness has proven hard to agree upon. The contraposition of readiness and resistance to change and the exact content of both terms are subject to debate. Paul, van Peet and Reezigt (2012) pay attention to the entanglement of the terms readiness and resistance. The concept of readiness is an overarching term divided in: a) antecedents of the recipients’ reaction (e.g. participation, management support), b) the content of the reaction (e.g. readiness for change), and c) the result of the reaction (e.g. performance) (Paul et al., 2012). This research focuses on the antecedents (a) and as a result readiness for change (b).
2.2 Independent variables
The literature study was accompanied by qualitative research in the company. This led to a selection of independent variables. As shown in the introduction the relationship of these antecedents with readiness and in particular the position of communication and participation will be researched. Two propositions are introduced for the positioning of the independent variables but first the definitions of the selected variables are given.
2.2.1 Definitions of the independent variables
To create common ground in terminology first the definitions of the independent variables that are used in this research are stated below.
Communication in the change context, is referred to by Bouckenooghe et al. (2009) as the way of translating the change message. The components of effective communication are clarity of information, frequency of messaging and openness about the change content. Armenakis and Harris (2009: 135) define persuasive communication in a change situation as: “transmitting the message components to change recipients”.
The involvement of the employee in a change process is known as participation. More specifically stated by Whelan-‐Berry and Somerville (2010: 182):
“Participation involves employees in tasks specifically related to the change initiative”. Participation involves the employee in the process, which in turn forces the employee to take part in the change (Bernerth, 2004; Armenakis et al., 2007; Cunningham, Woodward, Shannon, Maclntosh, Lendrum, Rosenbloom and Brown, 2002).
Discrepancy is defined by Bernerth (2004: 41) as: “A gap between the current state and an ideal state”. The employee facing a change should understand why the change is initiated and must see the improvement of the new situation. The change agent needs to communicate and convince the change recipient that this
10
needs (discrepancy) or if they are convinced of the improvement the change will bring (appropriateness) (Bernerth, 2004; Armenakis et al., 1993; Armenakis et al., 2007).
Self-efficacy is the confidence of the employee that he/she possess the competences to successfully enact on the proposed organizational change (Ajzen, 1985; Armenakis et al., 1993; Bernerth, 2004; Paul et al., 2012). Self-‐efficacy is related to the amount of effort employees are willing to put into the change (Bernerth, 2004). Self-‐efficacy is more commonly referred to as change confidence in empirical research (Bouckenooghe et al., 2009).
2.2.2 Direct relationship of independent variables with dependent variable
Independent variables influencing readiness for change directly is shown in figure 1. The model suggests an equal positive relationship of all independent variables with the dependent variable readiness. The strength and direction of the relationship is tested.
FIGURE 1
Conceptual Model direct relationships hypothesis 1
The direct relationship of the independent variables with readiness is a precondition of the moderator proposition later in this research. Previous research on direct relationships of the variables with readiness is discussed in more detail in the following section.
Oreg (2006) describes the direct influence of variables on readiness for change of the employee. Oreg employs the tripartite description of readiness as described by Piderit (2000).
12 interlinked with principal support, communication and discrepancy. Providing information to the recipient via communication by the change agents creates readiness (Coch and French, 1948; Oreg, 2006). Effective communication results in well-‐informed employees (Holt et al., 2007). These employees are more likely to recognize the discrepancy with the status quo (Kotter, 1995; Kotter, 2007) and show increased levels of readiness. Enthusiasm and constructive discussion (participation and communication) have a positive influence on the openness towards the change (Kaplan and Norton, 2006; Oreg, et al. 2011A). Armenakis and Harris (2009), Dunphy and Stace, (1993) and Whelan-‐Berry and Sommerville (2010) complement each other in describing the direct positive relationship of participation with readiness. Coch and French (1948) demonstrate in their early research that participation has a positive effect on adaptation to change. Bouckenooghe et al. (2009) uses a more general dimension, ‘involvement in the change process’, and found that this has a positive influence on the success of the implementation. This suggests that readiness of the employees is higher when they participate actively in the change process. Complementing this statement is the view of Kaplan and Norton (2006) that when participation takes place, people become owners of the change and become more willing to cooperate.
change is that he or she will report higher readiness to the organizational change (Eby, Adams, Russell and Gaby, 2000).
Managers who have a trusting relationship with their subordinates are more effective in creating readiness among their employees. They provide principal support during the uncertain times of a change process, which leads to intentional readiness (Oreg, 2006). The support of change agents and managers is given to the change recipient to fit this person’s needs and personality (Oreg et al., 2011A). “The support by the supervisors is conceived as the extent to which employees experience support and understanding from their immediate supervisor. More specifically it measures their openness to reactions of their staff and their ability to lead them through the change process” (Bouckenooghe et al., 2009: 599). These authors all describe direct relationship of one or more of the selected variables with readiness for change.
The description of the direct relationships with readiness for change leads to the following hypothesis:
Hypothesis 1: Communication, Participation, Discrepancy, Self-efficacy and Principal support all have a direct positive relationship with the - intentional (A), cognitive (B), emotional (C) - readiness of the employee.
2.2.3 Moderating influence of communication and participation
14 Armenakis and Harris, 2009; Holt et al, 2007; Kaplan and Norton, 2006; Oreg, 2006; Oreg et al., 2011A).
FIGURE 2
Conceptual model moderating relationship hypothesis 2
In the model of Armenakis et al. (1993: 684) and Armenakis and Harris (2009) communication and participation are labelled influencing strategies that lead to ‘the message’. The message in their model consists of discrepancy, principal support and self-‐efficacy, resulting in system readiness (Armenakis et al., 2009). The relationship of communication and participation with readiness is also supported by the research of Jimmieson, Peach and White (2008). Their focus is on moderation the of planned behaviour of the employee and thus also describes participation and communication as indirect determinants of readiness.
of Armenakis et al, (1993). This is in correspondence with Oreg et al. (2011A) stating that communication and participation lead to an explicit reaction of the change recipient, indirectly leading to change readiness of the employee.
Communication and participation activities lead to the last step in creating readiness for change, is the change message. The message should contain the need for change (discrepancy) and change confidence to overcome discrepancy (self-‐efficacy) and support for the employee (principal support) (Armenakis et al., 1993). The described indirect relationship of communication and participation with change readiness is in line with the expected moderation of the two variables.
The independent variables for this hypothesis are self-‐efficacy, principal support and discrepancy. Self-‐efficacy of an employee as an individual attribute can be influenced by participation and open communication to serve as a positive influence on readiness for the change (Holt et al., 2007). Self-‐efficacy of the recipient can be negatively influenced by the creation of discrepancy, decreasing confidence. Discrepancy can be a trigger for counterproductive behaviour if not communicated clearly by the change agent and the advantage of the change is vague (Armenakis et al., 1993). This will lead to a decrease of confidence in the individuals coping capabilities, self-‐efficacy. Principal support during the implementation of the change is important. The implementation success depends on both formal and informal leaders (Armenakis et al., 1999) who support the change in their words and actions. Discrepancy communicates the relevance and need for the change and creates a state where the change recipients feel that the current situation is awry (Nadler and Tushman, 1989).
The moderation of communication and participation on the independent variables will be tested using the following hypothesis:
16
2.3 Theory wrap-‐up
Armenakis, Oreg and several other authors conducted extensive empirical
research on the relationship between these independent variables and readiness of employees for a change. However, none of these authors have tested
3. METHOD
This research aims at deepening the understanding of the independent variables influencing the readiness for change of the employee. This research uses a case study at Eneco N.V. Rotterdam.
3.1 Case study: Eneco N.V.
Eneco Holding N.V. (hereafter Eneco) aims to be the most sustainable energy supplier in the Netherlands and pursues a strategy that is in line with this ambition.
The approximately 7000 employees supply over 2 million residential and corporate consumers in the Netherlands and Belgium with electricity, gas, installation services (e.g., boilers) and energy saving products and services. The company owns and contracts electricity-‐producing assets in these countries as well as in France, the UK and Germany. Eneco was established in 1995 as a result of the merger of the energy companies of Rotterdam, The Hague and Dordrecht. In subsequent years thirteen regional energy companies were added to the holding in the first wave of consolidation of the energy market. The shares of Eneco are still held by 61 Dutch municipalities. In 2011 the revenues of the holding were € 5,007 million and a profit of €204 million was recorded (Annual Report Eneco Holding N.V., 2011).
18 Eneco’s business environment has thus changed dramatically over the past decade. Adapting to this changing environment, the company and its employees needed to transform from a technology-‐centric supply-‐push approach to competition on increasingly international markets. Adapting the positioning and product, service offering to customer’s needs.
Eneco Holding now develops its strategy, by aligning core competences and corporate identity with market developments. The implementation of a new strategy entails change. The measured readiness for change of the employees at Eneco serves as input for this research.
3.2 Desk research
The desk research forms the foundation for the explorative research at Eneco. A long-‐list of independent variables that influence change readiness of the employee was defined according to the literature from the field. Several authors define both existing scales of measurements and comparable research setups and results. The concepts as defined in the previous chapter are considered to be the foundation of the field research conducted at Eneco. Relevant literature was found in electronic databases (Business Source Premier, Elsevier, Sagepub) and books available via the University of Groningen. The specification of the used literature can be found in the reference list. All internal documents of Eneco are confidential and cannot be distributed on request.
3.3 Field research
The assessment of the change situation is done using both qualitative and quantitative methods. The qualitative interview techniques provide a rich pool of case specific information. The quantitative questionnaire is an appropriate supplement to create measurable results (Holt et al., 2007). The interviews were conducted using a long-‐list of variables that served as the basis for the variables that were selected for further research.
interviewed. The coded results from the interviews describe the current situation and the viability of the chosen variables of the research. The interviews form the input for the quantitative questionnaire and yielded five independent variables supported by content analysis of the literature.
Subsequently, the questionnaire was designed and distributed. The original can be found in appendix A1 and the English translation in appendix A2. Note that the questionnaire was conducted in Dutch and translated into English for reporting purposes of this thesis.
The questionnaire consists of nine or ten questions per independent variable. The overall opinion of the recipient on the change, their readiness, is questioned. This is done using items based on the research of Bouckenooghe et al. (2009) and the readiness concept of Piderit (2000). Personal details are asked for to serve as control variables. All questions are posed on a 5-‐point-‐type Likert scale with anchors being 1 (strongly agree) to 5 (strongly disagree), as used by Bouckenooghe et al., (2009). The employees were asked to indicate their agreement with the statement on this scale. The questions on personal details of the respondents are closed questions and an open question is posed to create a feedback opportunity for the respondents.
3.4 Procedure
To safeguard the normality of the results of the questionnaire several demographic measures are integrated (Holt et al., 2007; Armenakis et al., 2007; Bouckenooghe et al., 2009). Eight respondents had one missing value in their questionnaire, the average of the answers of the respondent was taken to complete the dataset.
Dependent and independent variable
The independent variables used in previous research were long-‐listed before the interviews took place. This long-‐list was discussed with the interviewees and the most applicable variables were determined.
20 During the item development phase the understanding of the questions in the questionnaire were tested using a content adequacy test (N=8) based on the previously used test by Holt et al. (2007). 63 questions were posted, nine or ten per variable. Items were retained if 60% of the sample identified it to fit the correct variable (Holt et al., 2007). Nine items failed to meet the criteria and were excluded from the final version or rephrased. The wording of several questions was adapted and checked again after inclusion in the questionnaire. To check if the questionnaire was accessible and the results is processed correctly a test round was conducted. The employees were unaware of the fact that the will not be included in the final results (N=9). Five respondents were excluded of analysis because of the great number of unanswered questions. To prevent people from leaving questions unanswered the introduction text was adjusted. The final questionnaire contained a total of 59 questions, nine or ten items for each independent variable and twelve questions on the overall change readiness of the employee according to the tripartite subdivision of readiness by Piderit (2000).
Control variables
To test viability of the questionnaire personal details of every respondent are recorded. Sex, age, tenure, educational level and function within Eneco are the basis for the population analysis. The grouping of the employees is done by age, tenure and educational level. The recording of the position within Eneco checks the spread of employees over the business units. There is a known-‐group of designers with a bias towards the strategy. Designers form a separate testing group as described by Holt et al. (2007).
To ensure anonymity of the employee they will not be grouped according to position and no names or subordinate relationships are recorded.
3.4.1 Participants
All who have participated in this research are employees of Eneco. Designers and receivers are included in the qualitative and the quantitative research part. The questions asked in the questionnaire are on a 5-‐point Likert-‐scale.
To get a representative sample of the Energy Company employees from every business unit are included in the questionnaire. From the total of 2000 employees of the energy company approximately 400 were included in the sample and received a request to complete the questionnaire. The questionnaires were distributed via the manager of the business units. 229 questionnaires were filled out, of which only 122 respondents answered all questions. The survey was distributed in a digital survey program with an invitation via email.
The respondents group existed of 71 (58%) men and 51 (42%) women. The average age of the sample was 37 years. The average tenure of all respondents was 7,5 years. 107 respondents were excluded because their questionnaire or personal details were not completely filled out.
Designers of the change were separated from the receivers of the change. This is to prevent that the bias the designers hold towards the framework influence the results. This bias is the in general more positive attitude towards the change that is recorded for designers that are extensively involved in the design of a change (Holt et al., 2007). The analysis of the data is conducted separating designers and receivers. A one-‐way MANOVA test is used to determine the difference in scores for both known-‐groups (Holt et al., 2007). The designers group reported lower but almost equal mean readiness score than the receivers group, respectively 2,63 and 2,64. The designers group existed of only 4 (3%) respondents and is therefore not effective in discriminating between the two participant groups (Holt et al., 2007, cf. Coch and French, 1948). This means that the designers are included in the total sample and part of the analysis.
3.5 Data Analysis
22
3.5.1 Factor Analysis
Factor analysis was conducted to safeguard the validity of the research and the input for the regression analysis. As a rule of thumb the number of respondents has to exceed the number of items five times (Hair, Black, Babin and Anderson, 2010). To pursue to this rule the dependent and independent variables were split before analysis. The results of the factor analysis can be found in table 1.
TABLE 1
Results factor analysis independent variables
1 2 3 4 SE 4 ,804 SE 5 ,716 SE 7 ,785 PS 1 ,699 PS 3 ,831 PS 8 ,765 P 2 ,716 P 5 ,756 P 8 ,778 D 2 ,836 D 4 ,600 D 7 ,790 C 1 ,797 C 8 ,752 C 9 ,688
Extraction Method: Principal Component Analysis Rotation Method: Varimax with Kaiser Normalization a. Rotation converged in 6 iterations.
The items that are not in table 1 had to be dropped because of their high cross loading.
Notable is that principal support (PS) and communication (C) load in the same factor, which means these two variables, will be pooled in the regression analysis. The analysis had to be brought down to a four-‐factor analysis to create acceptable loadings. This means that the original conceptual model with five independent variables is no longer subject to analysis. Communication/Principal support will be treated as one variable. In the second hypothesis it serves as a moderator in the analysis, because of the fact that principal support is given by using communication. The aggregation of two variables is previously described by Ford, Ford and D’ Amelio (2008).
The dependent variable is split into emotional (E) readiness, cognitive (C) readiness and intentional (I) readiness according to the tripartite approach of Bouckenooghe et al. (2009) and Piderit (2000). The analysis was limited to a two-‐factor analysis. Several items were dropped because of cross loading.
This resulted in a 1-‐item scale for cognitive readiness and an aggregation of the intentional readiness items and the emotional readiness items into one factor. Piderit (2000) does acknowledge that the distinction between the variable is still subject of debate. Even though intentional readiness and emotional readiness load in the same factor the components will serve as separate dependent variables in the moderator analysis. The combination of intentional and emotional in one factor will also be included in the analysis and discussed in the result section.
TABLE 2
Results factor analysis dependent variable
1 2 R1 I ,816 R12 I ,724 R9 C ,930 R3 E ,709 R11 E ,674
Extraction Method: Principal Component Analysis Rotation Method: Varimax with Kaiser Normalization
24
3.5.2 Cronbach’s alpha analysis
To check the validity of the scales the Cronbach’s alpha test was performed. The coefficient of all variables has to exceed 0.6 in this kind of explorative research (Holt et al., 2007). After conducting the factor analysis this was checked for all variables.
TABLE 3
Results Cronbach’s Alpha analysis
Variable Alpha score
Communication/Principal Support .881 Participation .683 Discrepancy .680 Self-efficacy .793 Emotional Readiness .502 Intentional Readiness .547
Cognitive Readiness n.a.
Emotional/Intentional Readiness .741
The values of the independent variable scales were found to be sufficient all exceeding 0.6. Cognitive readiness is a one-‐item scale and can therefore not be included in a Cronbach’s alpha analysis. The separate alpha’s for the emotional and intentional scales do not satisfy the limit of 0.6. The combined alpha for intentional and emotional does meet the criteria.
3.5.3 Regression and moderator analysis
analysed as one moderator. The description of Ford et al. (2008) agrees with the integration of these variables stating that these two are interwoven. To be able to perform regression all variables should be normally distributed. This was tested using the One-‐Sample Kolmogorov-‐Smirnov Test. Significance greater than 0,05 supports the nil hypotheses1 validating a normal distribution. Respectively the following significance was found shown in table 4:
TABLE 4
Significance of variables supporting a normal distribution
Self-efficacy .018 Discrepancy .070 Communication/Principal Support .176 Participation .172 Readiness .098
These results exclude self-‐efficacy of analysis. To check the abnormality the histogram in figure 3 was generated, this showed that data were distributed along the normal distribution line with an accent on the 1.0 answer value. The missing values for zero are caused by the answer option 1-‐5. The analysis show too low significance (.018) but as shown in figure 3 this does not disqualify the variable. Therefore self-‐efficacy was included in the multiple regression analysis.
26
FIGURE 3
Histogram showing normality of Self-efficacy
NB. The histogram shows the frequency of the answers on the 5-‐point Likert scale.
The multiple regressions were performed using centralized data form the dataset gathered at Eneco. Every variable was tested on the relationship with intentional readiness, cognitive readiness and emotional readiness and the combined variable of emotional and intentional readiness. Subsequently the moderator analysis was performed with centralized data and multiplied independent variables and moderators using a regression test in SPSS.
The results of the analysis described above are presented in the result section.
4. RESULTS
This chapter outlines the results of the statistical analysis in SPSS and the explanation of these results. First, the findings from the statistical analysis are discussed. Secondly, the overall findings are summarized.
4.1 Analysis and findings
Based on the factor analysis and the satisfactory Cronbach’s alpha tests, multiple regressions with and without the inclusion of moderating variables were performed. Note that the Cronbach’s alpha’s for emotional readiness and intentional readiness do not satisfy the 0.6 criterion. The result of the factor analysis was that communication and principal support are aggregated into one variable. Intentional and emotional readiness also load on the same factor, the sub-‐concepts are analysed separately and combined as one variable. The regression analysis is performed with this combined variable, resulting in four independent variables. The results are discussed in the following section.
4.1.1 Multiple regression analysis
To satisfy the first assumption that the independent variables explain the value of the dependent variable, a multiple regression was performed. For every readiness component an individual analysis was performed. This results in several output tables summarizing the model scores, ANOVA2 scores, significance and coefficients of the regression.
Intentional Readiness
The coefficient results show that the model is significant for independent variables Self-‐efficacy (.016), Discrepancy (.000) at significance of >0,05. Communication/Principal Support (.058) and Participation (.063) are significant at >0.10. The adjusted R-‐Square is .397 which means that almost 40% of
28 intentional readiness is explained by the independent variables self-‐efficacy, discrepancy and communication/principal support and participation. The ANOVA test confirms the significance of the whole model (.000). The beta’s found for the significant variables show a positive relation with the dependent variable intentional readiness. Self-‐efficacy (.209) and Discrepancy (.346) show a strong positive relationship with intentional readiness. Communication and Principal Support (.162) and Participation (.148) show a positive but weaker relationship with Intentional Readiness in this study sample. Hypothesis 1A is confirmed for the relationship of discrepancy, self-‐efficacy and the combined variable Communication/Principal support and Participation with intentional readiness.
Cognitive Readiness
The ANOVA test confirms the significance of the whole model at .001. The independent variables show a parting in significance between Self-‐efficacy (.049) and Discrepancy (.037) opposing Communication/Principal Support (.688) and Participation (.679). Self-‐efficacy and Discrepancy are significant at >0.05. The adjusted R-‐Square of only 0.149 is mainly explained by the significant variables (Self-‐efficacy and Discrepancy). The contribution of the insignificant Communication/Principal Support (.041) and Participation (.040) is small. The positive significant relationship of Self-‐efficacy (.205) and Discrepancy (.206) do support Hypothesis 1B, however the two other variables do not explain the expected relationship with Cognitive Readiness.
Emotional Readiness
the high explanatory adjusted R-‐Square of almost 50%. The insignificant Self-‐ efficacy (.119) and Participation (.051) do not support hypothesis 1C.
The relationships that support hypothesis 1 with a positive significant relationship with readiness are summarized in figure 4.
FIGURE 4
Independent variable results
NB. Only paths that achieved significance at .05 level or lower are drawn.
30
Table 5
Relationships independent variables
Independent variable Strength of relationship Dependent variable
Self-efficacy .209 Intentional Readiness
Self-efficacy .205 Cognitive Readiness
Discrepancy .346 Intentional Readiness
Discrepancy .239 Emotional Readiness
Discrepancy .206 Cognitive Readiness
Communication/Principal Support .162 Intentional Readiness Communication/Principal Support .479 Emotional Readiness
Participation .148 Intentional Readiness
Emotional/Intentional readiness
The direct relationship with the aggregated Emotional/Intentional Readiness is summarized in table 6. The relationships describe 51% of the dependent variable with an adjusted R-‐Square of .508. The model is significant at .000
TABLE 6
Standardized coefficients direct relationship with Emotional/Intentional Readiness
Intentional and Emotional Readiness Sig.
Independent variable Self-efficacy .180 .021 Discrepancy .321 .000 Communication / Principal Support .352 .000 Participation .109 .130
4.1.2 Moderator analysis
Following the multiple linear regressions, a moderator analysis was performed according to the method of Baron and Kenny (1986). The original model that matches hypothesis 2 states Communication as a moderator and Principal Support as an independent variable. Due to the factor analysis the two variables are grouped and tested as a combined moderator supported by the findings of Ford et al. (2008). Self-‐efficacy and Discrepancy are used as independent variables. The moderators are expected to show positive values increasing the strength of the relationship of the independent variable and the dependent variable, confirming hypothesis 2 (A, B, C). The combined Intentional/Emotional Readiness dependent variable is tested as well as the tripartite division of readiness.
Intentional Readiness
The independent variable Self-‐efficacy shows a positive beta of .369 for the relationship with Intentional Readiness. The moderating effect of Communication/Principal Support is negative with a beta of -‐.218, stating that the relationship between Self-‐efficacy and Intentional Readiness is weakened by the moderation of Communication/Principal Support. The ANOVA test shows that the model is significant at .000. This means that when Communication and Principal Support are experienced to be high, Self-‐efficacy of the change recipient is of less influence on the recorded level of intentional readiness. The adjusted R-‐ Square proves that 31% of the Intentional Readiness is explained by these variables.
32
FIGURE 5
Moderating effect of Communication/Principal Support on Self-Efficacy and Intentional Readiness
In the second model the influence on Intentional Readiness of Discrepancy is strong at a beta of .489. In model with ANOVA significance of .000 and significance of the moderation of .101 the negative beta of -‐.126 implies a weakening of the relationship between Discrepancy and Intentional Readiness when Participation is high. The adjusted R-‐square shows that these variables explain 34% of the intentional readiness perceived.
FIGURE 6
Moderating effect of Participation on Discrepancy and Intentional Readiness
The third model explaining the perceived Intentional Readiness has an ANOVA significance of .000. Self-‐efficacy shows a strong positive relationship with Intentional Readiness (.400) and the moderating relationship with Participation is negative (-‐.078). This moderating relationship is not significant at .333. The adjusted R-‐Square for this model is 28%.
The forth model is also significant (ANOVA .000) and shows similar results. Discrepancy has a strong positive relationship with Intentional Readiness (.435) and Communication/ Principal Support moderate the relationship negatively at the same value as the third model (-‐.078), again this is not significant (.297). The R-‐Square is similar to the other models explaining 35% of the Intentional Readiness.
Two of the four models show a significant moderation of Participation or Communication/ Principal Support. Both moderations are negative and do not support the proposed positive relationship. Therefore, hypothesis 2A is rejected.
Cognitive Readiness
Self-‐efficacy shows a positive relationship with Cognitive Readiness (.272) but the R-‐Square (.093) warns for very low explanation of the variables in the Cognitive Readiness component. The model is significant at .002. Notable is the fact the here, although not significant (.332), a positive moderation (.099) of Communication/ Principal Support is recorded.
Next the relationship of Self-‐efficacy with Cognitive Readiness moderated by Participation is described. Again the adjusted R-‐Square is very low at .103. The model is significant at .002. The positive relationship of Self-‐efficacy and Cognitive Readiness is affirmed (.302) as well as an insignificant negative moderation (-‐.112).
34 Support is not significant but describes a negative moderation (-‐.106). Notable is the low adjusted R-‐Square at .106.
Discrepancy and Cognitive Readiness moderated by Participation is significant (.003) but has a low adjusted R-‐Square (.089). The relationship of Discrepancy and Cognitive Readiness stays positive at .300 and the insignificant (.553) moderation is -‐.053.
None of the relationships researched for Cognitive Readiness prove to have a significant moderating relationship. The adjusted R-‐Square values are clearly lower than for Intentional and Emotional Readiness. The insignificance of all moderating relationships mean that Hypothesis 2B cannot be supported with this research sample. Cognitive Readiness is a 1-‐item scale, which might be the cause for the lower adjusted R-‐Square.
Emotional Readiness
The influence of Self-‐efficacy on Emotional Readiness of the employee moderated by Communication/ Principal Support gives a beta of -‐.119 with and adjusted R-‐Square of .439. The significance of .000 demonstrates the validity of the model. Again this model shows that the relationship of Self-‐efficacy with the readiness component is weakened by the moderator.