LIESELOTTE HOEGEE
---
Effective Change
Communication in Practice
Investigating Communication in a Continuous
Quality Improvement Implementation at a Dutch
Hospital
---
Supervised by Christine Liebrecht
In partial fulfillment of the Master’s Program Corporate
Communication of the University of Amsterdam
Abstract
the field of change communication has focused on the development of theoretical models, yet, relatively few attempts have been made to explore change communication empirically (Kaplan et al., 2010). This is however crucial to the understanding of change communication, as change outcomes may be highly dependent upon the type of change and context in which the change is implemented (Lewis, 2006; Garvin, 2012). More practical research is needed in order to judge the validity of proposed theoretical models and to understand how change communication works out in various contexts. This study aims to fill this research gap by investigating a model for effective organizational change communication by Elving (2005) in the context of a Continuous Quality Improvement (CQI) implementation at a Dutch hospital. The model suggests there are several general communication criteria which should be met to implement organizational changes effectively. These criteria are: a high level information satisfaction, affective commitment and empowerment and a low level of job-related uncertainty. The model also postulates that job-related uncertainty is influenced by the other communication criteria. To determine the effectiveness of the CQI change implementation both the factual achievement as well as the personal experience level were considered. Data were gathered through an employee survey. Results indicate that information satisfaction, empowerment and job-related uncertainty effect change effectiveness. The effect of affective commitment was not affirmed, nor were the effects of the communication criteria on job-related uncertainty. Therefore, this study concludes that the model cannot be completely validated in this research setting, but that communication does play an important role in effective organizational changes.
Key words: Organizational change communication, information satisfaction, affective
Introduction
There is substantial evidence that suggests that over seventy percent of organizational change initiatives fail (Burnes & Jackson, 2011). There are numerous reasons for the failure of so many initiatives, such as the timing of the change effort, the role of change-agents and the organizational culture (Bennebroek-Gravenhorst, Werkman & Boonstra, 1999; Lewis, 2011; Weiner, Amick & Lee, 2008). Holt, Armenakis, Field & Harris (2007) have tried to capture the chances of change success by measuring an organization’s “readiness to change”, which they conceptualize as a multidimensional construct influenced by beliefs among employees that they are “capable of implementing a proposed change, the proposed change is appropriate for the organization, management is committed to the proposed change and, the proposed change is beneficial to organizational members (p.232).” With so many factors at play, it almost seems as if successful change is a matter of luck. How then can one “control” organizational change? One of the tools managers have at their disposal to guide the organization through change is communication (Lewis, 2011). It is argued that to implement changes successfully in an organization communication needs to be an integrative part of the change process (Covin & Kilmann, 1990; Lewis, 2000; Elving, 2005). Much research has demonstrated the importance of communication during change and many authors have proposed criteria for effective change communication (e.g. Elving, 2006; Lewis, 2011).Yet, there is still little empirical research which actually examines effective change communication in practice (Elving, 2005; Kaplan et al., 2010). Various authors point out that more research is needed in different settings, involving different types of change and change contexts to understand the effect of change communication under various circumstances (e.g. Lewis, 2006; Kaplan et al., 2010).
The lack of empirical research on this matter provides a double problem: firstly, there is not enough evidence to judge the validity of the various models proposed in the field of change communication. Secondly, there is insufficient information regarding the way change communication models work in various contexts. As Garvin (2012) puts it: “For real progress to be made [in the field of change communication], the proverbial ‘black box,’ the firm, has to be opened and studied from within.” This is exactly what this study aims to do. By means of a case study on a continuous quality improvement (CQI) implementation in a Dutch hospital this study is aimed at finding empirical evidence to support the validity of a model for effective change communication (introduced on page 7) for a particular type of planned change.
A case study at a hospital provides an opportunity to contrast different change settings – a hospital consists of multiple units - while simultaneously limiting the number of contextual factors, which is often a problem in change communication research (Lewis, 2006). Moreover, the hospital under investigation is an exemplary case as it is one of the first hospitals in the Netherlands to have implemented a CQI program organization-wide instead of only project-based (Morgens, 2014). This study also adds to literature because although previous studies have researched the effects of contextual factors on performance outcomes of CQI (e.g. Kaplan et al., 2011), this is one of the first Dutch studies to empirically investigate the role of communication in relation to a CQI implementation.
The paper is structured as follows: in section one, literature relevant to the topic is reviewed. In section two the applied statistical method is outlined, and in section three the effective communication model under investigation is tested through structural equation modelling. In the final section the findings are discussed and some suggestions for future research are made.
Literature review
Firstly, this section discusses the ideas behind CQI and relates them to the actual implementation at the hospital central in this study. Subsequently, the change communication model under investigation is introduced and its theory is explained. The concepts central to the model are also clarified and related to the practices at the hospital.
Continuous quality improvement
CQI represents a set of methods to improve healthcare, originating from industrial process improvement approaches (O’Neill et al., 2011). It is an ongoing effort to improve products, services and work processes. Maximizing customer value is the chief objective. Processes are constantly evaluated and improved in the light of their efficiency, effectiveness and flexibility to reach this goal (Swor, 2012; Kaplan et al., 2011). It is a data driven approach which makes use of systematically measured “quality indicators” to initiate and “drive organizational changes in a never-ending cycle of continuous improvement” (Kritchevsky & Simmons, 1991, p. 1817). This means that CQI is not a simply one transition from a current state to a desired future state (freeze-unfreeze-refreeze), but a permanent transition (a constant unfreeze mode). Achievements are made and can be measured, but will never be considered final. However, CQI can also be considered a “state-of-mind” or working philosophy (Baird & Reeve, 2011). From this perspective the implementation of a CQI initiative can be regarded as a transition from a current state (non CQI-minded) to a desired future state (CQI-minded). This perspective mainly considers whether employees went along in the change and became convinced of CQI’s effectiveness. This study takes a holistic view, and approaches the effectiveness of a CQI change implementation process from a dual perspective; it considers both the actual achievements made throughout the change process, as well as employees’ personal experiences with the change.
In a recent study, O’Neill et al. (2011) put forward that although much research has been done to assess the effectiveness of CQI, there appears to be lacking a consensus on the definition of CQI and what “effective CQI implementation” entails. In practice, the initiatives vary considerably as organizations tailor their programs to their needs. Therefore, O’Neill et al. (2011) argue that to interpret results across different settings, authors should explicitly describe the key features and aims of the CQI initiative they investigate. The hospital initiative under investigation in this study combines insights from two recognized CQI programs: Lean and Six Sigma. Lean in a healthcare setting focuses on minimizing the amount of time and resources used in healthcare processes and other activities of a hospital, with an emphasis on eliminating all forms of waste (De Souza, 2009). It is about “getting the right things to the right place at the right time in the right quantity to achieve perfect work flow” (Gustashaw & Hall, 2008, p.78). Six Sigma is a data-driven approach, which seeks to achieve quantitative insight in all work processes and to reduce variability in these processes as much as possible (Coronado & Antony, 2002). In the healthcare sector blending these two approaches is fairly common (Koning, Verver, Heuvel, Bisgaard & Does, 2006). Indeed the hospital investigated in this study burrows from both approaches, but does not strictly follow the prescribed approaches of either methodology. The initiative has three main focuses: (1) standardizing care programs (2) care process improvement and (3) tactical planning. These three focus points are aimed at achieving efficiency (eliminating waste and variability) and by doing so, improving patients’ comfort and reducing overall healthcare costs. Another important aspect of the initiative is training; all employees are encouraged to follow “Belt trainings” (a Six Sigma tool) in which they learn more about the change program’s philosophy and can become familiar with tactical tools one can use in improvement processes. The program is implemented on unit level and organized around
small projects, which are related to the three focus points. Units are stimulated to participate in projects and can initiate projects or propose additional ideas for improvement themselves as well. Especially for the latter, a degree of CQI-mindedness at a unit is needed. Overall, the hospital has managed to get all units involved in at least one project since the start of the program in 2012, but it appears some units are much more willing than others to get involved in the CQI initiative and more eager to initiate projects themselves. This study tries to find out whether this discrepancy in the involvement in the change process might be attributable to a discrepancy in the effectiveness of change communication.
The effective change communication model under investigation
Communicating about organizational change is complex. However, attempts have been made to decomplexify the matter by providing communication managers with general guidelines - overarching criteria which should be met under all circumstances for change communication to be effective. One of these attempts is from Elving (2005) who proposed a model for effective change communication.
According to Elving (2005) effective change communication needs to fulfill three criteria: it (1) provides employees with essential information about the change, (2) fosters a sense of community to cope with the change and (3) takes away job-related uncertainty. Its purpose is to reduce resistance and thereby foster readiness towards change, which Elving (2005) considers the main antecedent of effective organizational change. The foundations of Elving’s (2005) change communication criteria are found in earlier works of Francis (1989) and De Ridder (2003) who put forward that in organizational communication a distinction can be made between organizational communication aiming to provide information (communicatio) and organizational communication aiming to create a community spirit (communicare). Elving (2005) applied these
principles of organizational communication to the context of organizational change and extended them with additional insights.
The model displayed in Figure 1 shows the relationship between the three communication criteria and their relation with the CQI change implementation and is an adaptation of the effective change communication model put forward by Elving (2005) – that is to say that it is applied to a CQI setting. The model suggests that communication efforts aimed at informing and creating community are achieved independently, but taking away uncertainty and job insecurity is also dependent upon the other two criteria. Together they foster the effectiveness of the CQI change process, which is defined as a continuum that develops to readiness to become more CQI-minded and an end state in which one actually has become CQI-minded and works according to the philosophy. The latter is a deviation from Elving’s (2005) original model, which considers the two levels of the continuum as separate stages. As the case study to which the model is applied is non-longitudinal and the hospital’s employees could be at various stages of the CQI change process, these two stages were boxed together and considered as separate levels of a single stage. The CQI implementation’s effectiveness is considered on a factual achievement, as well as the personal experience level.
Figure 1: Effective change communication model based on Elving (2005)
Effective CQI change implementation Communication to
inform
m
Effective CQI change implementation
Communication to create community Job-related uncertainty Readiness to change CQI-minded
Elving’s model (2005) has not been verified empirically. This study makes a first attempt by testing the model empirically in a real-life CQI setting. It follows the rationale that when these three communication criteria are met employees are more likely to have change changed effectively (and are thus more ready to change and closer to being CQI-minded). In other words, when employees (1) are satisfied with the information provision about the change, (2) experience a sense of organizational community and (3) feel less uncertain about their job perspectives they are more likely to change effectively. Before conducting the empirical study however, it is important to clarify in more detail what is understood by the three communication criteria in this study and which hypotheses need to be investigated. Regarding the former, it should be noted that the three communication criteria were not fully developed by Elving (2005) and have been slightly altered in accordance with recent academic literature on the topic.
Criterion 1: Communication to inform about change. A key objective of communication during organizational change is to inform employees about the change. Employees need to understand what is expected of them in the change process, but more importantly they need to understand why this change is necessary for the organization (Elving, 2005). Critical aspects which can lead to the success or failure of information provision are the timing, understandability, perceived quality, usefulness and correctness of the information (Elving, 2005; Lewis, 2006). McKay, Kuntz & Näswall (2013) put forward that when employees are satisfied with the information provision, they are likely to evaluate the change more positively and exhibit greater willingness to cooperate. Similarly, Lewis (2006) found that the higher the evaluation of quality of implementation information, the less likely employees are to demonstrate resistance towards a change.
key priority, as the initiative implemented was complex. To provide employees with necessary information the hospital’s communication department for instance used newsletters, intranet and CQI posters. They also organized information sessions in which they explained the program or related issues – e.g. reorganization plans and renovation of the hospital – in more detail. Additionally, the department was in charge of a special CQI preparatory program for new employees. It is expected that employees that are more satisfied with the hospital’s information provision have changed more effectively, therefore the following hypothesis is proposed:
H1: Employees who are satisfied with the information provision about the CQI initiative have changed more effectively than employees who are not satisfied with information provision about the CQI initiative.
Criterion 2: Creating community to cope with change. Elving (2005) suggests that to motivate employees to go along with the change a degree of organizational community is a key ingredient. To create an organizational community, communication has been found to be one of the most important antecedents - it can be regarded as the glue that holds the organization together (Meyer & Allen, 1997). Postmes, Tanis & Wit, (2001) argue that:
[A] sense of belonging to the organization does not primarily depend on the quality of their informal and social-emotional interactions with peers and proximate colleagues, but it is related more strongly to their appreciation of the organization’s formal communication (p. 240).
Evidence of a strong organizational community shows e.g. in high commitment to the organization (Elving, 2005). Especially affective commitment, which is defined as “an employee’s emotional attachment to and identification with the organization” (Allen & Meyer, 1990, p.2), appears to be an important factor fostering positive attitudes and reactions towards
organizational change (Iverson, 1996; Yousef, 2000; Oreg, 2006; Peccei, Giangreco & Sebastiano, 2011). Studies to date suggest that some of the most important (communication) tools to increase affective commitment are organizational rewards, sharing of successes and management support (Meyer & Allen, 1997; Rhoades, Eisenberger & Armeli, 2001; Chen & Indartono, 2011).
In line with other academic findings it was decided to broaden Elving’s (2005) conceptualization of community, as one could argue that an organizational community entails more than emotional attachment. For people to feel part of a community a degree of “empowerment” is needed; people need to feel that they have an influence in the community, that they can make a difference and that the management listen to and trusts them (McMillan, 1996). Especially today’s staff is more likely to expect a participatory style and an environment in which they may contribute to decisions on how their departments are managed (Lewis, 2006). McKay et al. (2013) put forward that “employees who feel like they have an opportunity to participate in change planning and who feel that their input is considered tend to exhibit greater engagement with, and often more support for the change” (p.30). Allowing employees to participate in organizational change reduces resistance by making them more comfortable with the proposed change (Appelbaum & Wohl, 2000; Lewis, 2006).
The hospital under investigation used various tools to create commitment and empower employees in the change process; for instance awards for the best CQI initiative of the year, a wall of fame for exemplary initiatives, an improvement board, special CQI workshops, training programs and CQI lunches. It is expected that employees that feel a greater sense of community have changed more effectively. This sense of community is expected to show in both a strong
sense of affective commitment to the organization and strong feeling of empowerment in the change process.
H2: Employees who feel a greater sense of organizational community have changed more effectively than employees who do not feel a great sense of organizational community
Criterion 3: Take away job-related uncertainty. Many studies put forward that one of the most difficult aspects of organizational change is taking away the uncertainty employees associate with the change. Not knowing how a change may affect job perspectives can lead to stressful situations for employees (Bordia, Hunt, Paulsen, Tourish & DiFonzo, 2004; Mckay et al., 2013). Hellgren, Sverke & Isaksson, (1999) distinguish between quantitative and qualitative job-related uncertainty. The first focuses on the threats to the continuity of the job itself, while the latter reflects a threat to the continuity of important job features. Particularly quantitative job-related uncertainty has been linked to anxiety, job satisfaction and turnover intentions (Bordia et al., 2004). In the case under investigation a certain degree of quantitative job-related uncertainty was considered likely, as the CQI implementation also entailed downsizing of personnel in some units. It is expected that the adverse effects of quantitative job-related uncertainty will show in the CQI change process. Therefore it is hypothesized that employees who experience a low level of job-related uncertainty have changed more effectively than employees who experience a high level of job-related uncertainty.
H3: Employees who experience a low level of job-related uncertainty have changed more effectively than employees who experience a high level of job-related uncertainty.
As uncertainty forms a threat to change implementation, an aim of change communication is reducing uncertainty. Bastien (1987) notes that “personal uncertainty [during
change] is pervasive and must be managed through communication management” (p.28). Previous studies have found that effective communication management reduces uncertainty among employees (Schweiger and Denisi, 1991; Johnson, Bernhagen, Miller & Allen, 1996; McKay et al., 2013). Berger (1987) suggests that people have two fundamental needs regarding uncertainty: predictive needs, concerning the ability to predict what is going to happen next and explanatory needs, concerning the ability to explain why things are as they are. It is vital to satisfy these needs by providing adequate information. By providing change information on time and by answering the questions employees may have, the experienced uncertainty of employees is reduced. Moreover, a sense of organizational community can be helpful in this respect. Especially, the empowerment of employees in the change process seems to be of importance. As they participate in the change they are more likely to feel in control over the change process (predictive needs) and to understand the meaning of the change (explanatory needs) (Holt et al., 2007). Therefore, it is expected that information satisfaction and a greater sense of organizational community decrease quantitative job-related uncertainty.
H4: Employees experience less quantitative job-related uncertainty when they have a high level of (a) information satisfaction and (b) organizational community.
Methodology
In this section the methodology of this study is discussed. Firstly, the research setting, sample, data collection procedure and the research materials are reported. The subsequent sections describe the operationalization of the measures and the applied methods of analysis.
Research setting
into existence after a merger between two small regional hospitals in the early 1980’s. The hospital offers basic medical care and specializes in geriatric care. The hospital provides work for approximately 1500 employees. The hospital has around 300 beds and 140.000 patients yearly. The hospital is comparable to other general hospitals in the Netherlands in terms of size and medical care provided (Dutch Hospital Data, 2013).
Respondents
The respondents who participated in this study were all employees of the Dutch hospital under investigation. Employees of five hospital units were excluded from the study; this were employees of outsourced units (facility, chemical lab) and of units involved in CQI program management (Management Board, Communication Department, HRM). In total 23 hospital units were included of which all 1142 employees were asked to complete the survey - of which 233 did. Three surveys were not used for the study as the respondents completed less than 50% of the survey, which made the total response 230 (response rate 20.14%)1. Of the respondents 19.6 percent was male and 80.4 percent female and age ranged from 21-63 (mean 43.2). This is representative for the hospital’s workforce2. Approximately 10 percent of the respondents had a management function, 10 percent was medical specialist, 40 percent was nurse and 40 percent held other staff functions3.
Data collection and materials
The data were gathered via a survey and were collected in the first three weeks of May 2014. All employees of the hospital received an invitation4 via their corporate email address from the hospital’s Head of Communication to complete the online survey. After one week a
1 In Appendix III the response rate per unit can be found.
2 Male/female division at the hospital is 17.44%/82.56% and mean age is 43.7. 3 In Appendix III the exact division response per function can be found. 4 In Appendix I the invitation can be found.
second email was sent to all employees as a reminder to complete the survey. In addition, all unit mangers received an email in which they were asked to encourage their employees to participate. The survey was administered via the hospital’s own survey software (Explora). This was done to facilitate the response process, as employees were likely to be familiar with the format.
As there was a risk that only the most CQI-minded employees would be willing to participate in the survey – which could have lead to a biased sample - a solution was sought to increase response. It was decided to allot two gift vouchers of €25, as monetary lottery incentives are considered a good way to overcome nonresponse error (e.g. Ryu, Couper & Marans, 2006; James et al., 2011; Laguilles, Williams & Saunders, 2011). Participation in the lottery was voluntarily, to keep the possibility open to remain anonymous.
Given that almost all employees of the hospital are native Dutch speakers, the survey was in Dutch. The survey consisted of 41 close-ended questions divided over 6 categories (demographics, program evaluation, information satisfaction, sense of community, job-related uncertainty and additional contextual factors)5. The survey was pretested on clarity, understandability and duration among nine employees of the hospital’s project/communication management unit, which did not participate in the actual study.
Measures
Dependent variable. The dependent variable in this study is the effectiveness of the CQI change implementation. As indicated previously, this study approaches the effectiveness of the change implementation from a dual perspective by considering both the actual achievements made throughout the change process, as well as employees’ personal experiences with the change. To capture both perspectives in a single factor it was decided to employ two change
effectiveness measures. The first set of measures was a management-derived rating based on verifiable factual achievements. The second set of measures was derived from the employees’ own evaluation of the change initiative. Both sets of measures are explained in the following paragraphs.
1) Management-derived change indicators. For this measure all hospital units were scored on their implementation success and all employees of a unit received their unit’s scores. The advantage of a management-derived measure is that it circumvents common method bias, which is a well-known disadvantage that comes with using data of self-reported surveys only (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). To establish the effectiveness of the change implementation per unit two managers involved with the CQI change program and two external consultants that supported the hospital with the program were interviewed and asked to name success indicators for the change program. From these interviews four success indicators were derived: (1) training level (how many employees followed CQI trainings), (2) participation in projects, (3) results (based on the three aims of the CQI program) and (4) visibility of CQI improvement culture. The four interviewees were asked to score every unit on these indicators on a scale of 1 (not at all) to 5 (very much). To verify whether the indicators were evaluated consistently inter-rater reliability scores were calculated (for all indicators: ICC2 >0.8 and Knippendorff’s alpha >0.7), which were all above the minimum acceptable values (Van Ness, Towle & Juthani-Mehta, 2008; Taylor & Watkinson, 2007). As a final validation, these overall mean scores were evaluated by three Executive Board members of the hospital.
2) Employee-derived change indicators. In addition to the management-derived change
process score, employees’ personal evaluation of the CQI implementation was considered. For this measure the three item change-success score of Lewis (2006) was used. In addition, seven
other questions concerning the achievement of CQI goals (e.g. more efficiency, improved patient care, enjoying improvement) were added to this scale. Items were measured on a five-point Likert scale and some items were reversed-coded. Cronbach’s alpha was used to assess the internal consistency of the measure and indicated that exclusion of one item was desirable (Cronbach’s Alpha increased upon variable exclusion). Alpha-level for this nine-item measure was .90, which is well above the critical cut-off point of .70 (Gliem & Gliem, 2003).
The two sets of measures were used to construct one overarching change factor including both perspectives. Cronbach’s Alpha was also used to evaluate internal consistency of the two sets of items combined and indicated that inclusion of all remaining item was desirable (Cronbach’s Alpha decreased upon variable exclusion). Alpha-level for the combined factor consisting of 13 items was .86, which indicated that the two sets of measures could indeed be combined.
Independent variables. Information satisfaction. To measure the information satisfaction of employees regarding the change implementation a four-item scale designed by Lewis (2006) was used. Based on Elving (2005) one item was added, measuring whether employees considered information to be given on time. Items were measured on a five-point Likert scale and some items were reversed-coded. Cronbach’s Alpha indicated that inclusion of all five was desirable (Cronbach’s Alpha decreased upon variable exclusion). Alpha-level for the five-item factor was .83. Therefore, all items were used to compose a single factor for the structural equation model.
Affective commitment. Employees’ sense of community is assessed by two measures, of
which affective commitment is the first. To measure employees’ degree of affective commitment to the hospital the Dutch commitment measure of Jak and Evers (2010) was used which is based
on Meyer and Allen’s classical three-component commitment measurement instrument. For this study only the five items measuring the affective commitment component were used. The items were measured on a five-point Likert scale. Cronbach’s Alpha indicated that inclusion of all five was desirable (Cronbach’s Alpha decreased upon variable exclusion). Alpha-level for the five-item factor was .84. Therefore, all five five-items were used to compose a single factor for the structural equation model.
Empowerment. Empowerment was the second measure to assess employees’ community
sense. To measure employees’ feelings of empowerment Lewis’s (2006) six-item measure of employee involvement was used, which was also assessed on a five-point Likert scale and some items were reversed-coded. Items both addressed to possibility to provide input as well as the appreciation of this input. Cronbach’s Alpha indicated that inclusion of all six was desirable (Cronbach’s Alpha decreased upon variable exclusion). Alpha-level for the six-item factor was .84. All items were used to compose a single factor for the structural equation model.
Quantitative related uncertainty. To measure employees’ degree of quantitative
job-related uncertainty the quantitative dimension of the job-job-related uncertainty measure of Hellgren, Sverke & Isaksson, (1999) was used. Quantitative job uncertainty was assessed on three items on a five-point Likert scale. Cronbach’s Alpha was indicated that inclusion of all three was desirable (Cronbach’s Alpha decreased upon variable exclusion). Alpha-level for the three-item factor was .85. All items were used to compose a single factor for the structural equation model.
Operationalization
This study used structural equation modeling (SEM) to analyze the theoretical model under investigation. As the model is built up from various factors, multiple dependent variables
and the research had a confirmatory aim, this research method is appropriate. The software package SPSS AMOS by IBM was used.
Recoding of variables, assumption checking and preparation of data was done in SPSS. For the latter it was important to take into account that SEM requires variables to have no missing data. Because variables used for analysis did have some missing data, these observations either had to be omitted entirely or to their values had to be estimated. The choice was made for the latter, as this is the preferable solution for structural equation modeling, which rules out the possibility of deleting respondents of whom data was not missing at random (Allison, 2003). Little’s MCAR test was carried out to make sure the observations were missing at random, which they were (χ2 = 4807.871, DF = 4680, p = .094). The expectation maximization technique was used to estimate the values of the missing data.
The Maximum Likelihood estimation method was used to carry out the SEM analysis as this is the most commonly used and most robust estimation method available for SEM (Sherehiy, 2008). On the basis of modification indices and theoretical validation improvement to the model were made. The robustness of the model was also checked by means of demographic control variables (gender and age).
To evaluate the effects of the individual estimates in the models (H1-4) Beta-coefficients (unstandardized) were assessed using an alpha level cut-off point of .05. Model fit was evaluated using the Chi-square test (χ2), Comparative Fit Index (CFI), the Tucker-Lewis Index (TLI) and the Root Mean Error of Approximation (RMSEA). To measure the goodness of fit the following cut-off points were used: alpha level of .001 for χ2, ≥.90 for the CFI and TLI and ≤.07 for the RMSEA. These are the most commonly used fit indices and corresponding cut-off points used in SEM (Kline, 2011; Browne & Cudeck, 1993). To measure the model’s effect size the R² is
reported, which represents the percentage of variance of the dependent variable accounted for by the model. A relevant note to make is that the Chi-square statistic is known to be very sensitive to large sample sizes (+200) and models containing many variables due to its method of computation (Bearden, Sharma and Teel, 1982; Raykov, 2000). Therefore, a likely result is a significant difference between the implied covariance matrix for the confirmatory model and the sample covariance matrix. It is thus important to see the results of the Chi-square in tandem with the other fit indices.
Results
In this section the results of the structural equation modeling process are presented. After the basic research model is introduced, an improved version of this model is presented. This model is used for the analysis and for the evaluation of H1-4. Additionally, an alternative research model is suggested and discussed.
Testing the research model
On the basis of the theoretical model which was adapted from Elving (2005) a basic research model was set up, which can be found in Appendix IV. However, this basic model did not fit the data well (see Table 1 for model fit indices), and therefore it was decided to make some improvements to the model. In line with methodological habit of many in the field of SEM, covariances were added between all independent factors (Hox & Bechger, 2000). The error terms of the items of the two sets of change indicators were correlated as well. It was expected that both sets of measures would have a common omitted cause due to their methods of measurement which could not be measured by the model and which would therefore be captured in the error terms. In addition, on the basis of the modification indices and theoretical argumentation the
error terms of two measurement items of the factor Empowerment were correlated. It was deemed likely that these items had a common omitted cause as well – the two items of which error terms were correlated were not only about being given the possibility to provide input or the appreciation of this input, but also about whether the hospital considered the input, which is an aspect that was not specifically measured by the model. The improved model showed a significant enhancement of model fit (Δχ2=930.84, ΔDF=42, p <.001). Moreover, model fit and effects were robust when controlled for demographic factors (gender and age). This improved model was used for analysis and can be found in Figure 2.
Model fit. As a first indicator of model fit, the χ2-value was examined. This value shows that the covariance matrix implied by the confirmatory model was significantly different from
the sample covariance matrix (χ2
= 674.92, DF=415, p<.001), which indicates a poor model fit. This was already expected due to the sample size and model complexity. To further examine the fit the values of some other indices were reviewed. These yield more positive results regarding the fit; RMSEA is below the cut-off point and indicates a fair fit (RMSEA=0.052) and also CFI and TLI show reasonable fit (CFI =.93, TLI = .92). To provide an indication of the effect size of the model, the squared multiple correlations were considered. The eventual R square value of the change progress variable showed that 60.6 percent of the variance was explained by its predictors (R2=.606). In other words, this means that 39.4 percent of the variance of change progress is still unexplained, and therefore captured by the error terms. Based on these results, it was concluded that the model fitted the data well enough and that it could be used to further examine the hypothesized individual effects.
Individual effects and hypotheses. After assessing the model fit, the individual effects of the model were examined. The individual estimates of the complete model show firstly that the estimates of the items measuring the factors Information Satisfaction, Affective Commitment, Empowerment and Job-related Uncertainty are all significant (for all items
p<.001). For the factor Change Effectiveness one item did not have a significant loading (p=
.071). The covariances between Information Satisfaction, Affective commitment and Empowerment were all significant (for all covariances p<.001). The covariance between the error terms of two items of the factor Empowerment was significant as well (r=.260, p<.001). Also the direct effects of the communication criteria on Change Effectiveness have proven mostly significant; Information Satisfaction (β=.186, S.E. = .057, p<.001), Empowerment (β = .101, S.E. =.039, p =.009) and Job Uncertainty (β = .030, S.E. =.014, p=.030) all showed significant results. Beta coefficient were positive, which means that the higher one’s level of
information and empowerment, and the lower one’s level of job-related uncertainty the more likely one is to change effectively. This means that for this model H1 and H3 can be affirmed. H2 which concerned employees’ experienced sense of community can be partially affirmed, as only the empowerment component was significant. The effect of the other component, Affective Commitment, did not prove to be significant (β=-.007, S.E. =.022, p=.751).
The direct effects of Information satisfaction (β=-.202, S.E=.187, p=.280) Affective Commitment (β =.252, S.E =.154, p = .103) and Empowerment (β =.002, S.E. =.186, p=.992) on Job-related Uncertainty were all shown to be insignificant. This implies that from this analysis it appears that one’s level on job uncertainty is not influenced by any of the communication aspects under investigation. On the basis of these outcomes H4 cannot be affirmed.
Alternative research model
In line with the principle of parsimony, an additional, more lenient version of the research model was tested which can be found in Figure 3. This alternative model only included the significant paths of the original research model. This meant that the effects of the communication criteria on Job-related Uncertainty were removed and that Affective Commitment was excluded from the model in its entirety. As Job-related Uncertainty was no longer endogenous, it was also correlated to the independent factors. This alternative model had a very good model fit (see Table 1) and showed an improvement compared to the improved research model (Δχ2
=254.36, ΔDF=136, p <.001). The model also confirmed the significant direct effects of the remaining communication criteria (see Table 2). The outcomes of this model were found to be robust when controlled for demographic factors (gender and age) as well.
Table 1: Model fit indices of basic, improved and alternative research model
Model fit indices Basic research model Improved research model Alternative research model χ2 1605.76, DF=457 674.92, DF=415 420.55, DF=279 CFI .71 .93 .96 TLI .68 .92 .95 RMSEA .105 .052 .047 R2 .491 .606 .606
Table 2: Direct effects alternative research model
Factor β S.D p-level Information satisfaction .183*** .056 <.001 Empowerment .100** .038 .008 Job-related Uncertainty .029* .013 .029
Conclusion
This study set out to find empirical evidence to verify the validity of a model for effective change communication for a particular type of planned change. The effective change communication model under investigation was the model from Elving (2005), which was modified and applied to a particular CQI change initiative. From the model four hypotheses were derived of which two were confirmed and one partially confirmed. The first hypotheses for which this study found empirical evidence was the relationship between employees’ information satisfaction and their likeliness to have changed effectively. From this study it appears that the more employees feel informed about the planned change, the more they are willing to go along with the change, which is in line with previous findings (e.g. Lewis, 2006; McKay et al., 2013). With regard to the sense of community, which was also expected to be of influence, only partial evidence was found. This study found evidence for the empowerment aspect, but not for affective commitment, which contests Elving’s (2005) idea of the operationalization of his model. From Elving’s (2005) suggestions for operationalization of the model affective commitment was expected to be a good indicator of sense of community. However, it was decided to take a broader perspective on this construct and also include empowerment in the operationalization. It is interesting to see that in this particular change setting it was actually empowerment which proved to be better indicator of change effectiveness. Yet, it should be noted that in the suggestions for operationalization of his model Elving (2005) also indicated that trust and identification could be good indicators of sense of community. Due to this study’s limited scope not all these aspects could be included.
Also one’s level of job-related uncertainty has been found to influence likeliness to change effectively, which also confirms the findings from previous studies (e.g. Bordia et al., 2004). Yet, no evidence was found for the hypothesized effects of information satisfaction and
sense of community on job-related uncertainty, which suggests that these communication aspects do not influence employees’ level of job-related uncertainty. This goes against the model of Elving (2005) and findings of previous studies which argue that these communicative actions could reduce uncertainty (e.g. Schweiger and Denisi, 1991; Johnson et al., 1996; McKay et al., 2013, Bordia et al., 2004). One of the reasons that these relationships were not found in this particular case may be that not all employees were confronted with reorganization of their unit and that the people that were confronted with it no longer work at the hospital due to the reorganization. The sample may have been too small to find a relationship between these aspects – there might simply not have been enough contrast between the respondents. Yet, the standard deviation of the job-related uncertainty items does suggest that there was a considerable disparity between respondents (for all items: 5 item scale, means between 2.48-2.53, and standard deviations between 1.05-1.07). Therefore, it may be more realistic to argue that information satisfaction and sense of community are simply not very good predictors of job-related uncertainty in this particular case – as was projected by the alternative research model which was derived via a more data-driven approach.
Discussion
Although much research has been devoted to define effective change communication theoretically, this is one of the few studies to have investigated change communication empirically. The case study approach at a single organization with multiple units proved to be a good way to find contrasts, while at the same time allowing for the limitation of contextual factors (e.g. different types of change, different organizational culture, etc.). Although a single case study is of limited value for generalizations, this approach can certainly be considered for future studies as the synthesis of multiple case studies may provide a better insight in change
communication in practice. Also the CQI implementation added depth to the study as it is a type of change that requires a factual as well as mental transformation, which makes it a complex, but also very interesting type of change to explore. Still, this study does come with some limitations and suggestions for improvement on which the subsequent paragraphs elaborate. The theoretical and practical implications of this study are discussed in the final paragraphs of this section. Regarding the limitations it should firstly be noted that the model under investigation is a highly simplified model. There are many important aspects during change and it may very well be that the exclusion of some of these aspects may have lead to the overestimation of the effects of the aspects present in the model. As mentioned before, Elving (2005) suggested that both identification and trust would have been interesting aspects to consider, but there are many more aspects which could have been further explored. For example, the content and context of the change were not considered in this study which could be a treat for internal validity. Although it was assumed that the case study approach would make the content and context factors relatively stable, they might still have been of major importance. There may be differences between units that could explain their high score on the communication criteria and on change effectiveness altogether that were not considered in this study. Moreover, this study disregarded the influence of communication media and use of language, which are of increasing interest in research and also believed to be very important in communication (e.g. Schultz, Utz & Goritz, 2011; O’Sullivan, Raguillat, Stern & Willner, 2012). It is therefore important to interpret the findings with caution and for future studies it would be interesting to include possible omitted causes - such as employees’ trust and type of language used – to increase internal validity.
Something else which is important to mention in the context of internal validity is that this study investigated the communication criteria on a general level, meaning that it did not
isolate communication on various levels within the hospital. It is very likely that part of employees’ information satisfaction and feeling of empowerment is derived from other (informal) sources than the formal communication by the hospital’s management (Downs and Adrian, 2012). For instance, unit management, but also colleagues might play an important role in the communication process. It is still interesting to see that higher scores on these factors may lead to more effective change, but it is unrealistic to assume that these effects can solely be achieved via formal change communication. The difference between the effects of formal and informal communication during change is definitely interesting to explore in future studies. Another aspect to mention is the composition of the survey and the reliability of the data. In particular for the items measuring commitment respondents indicated in the comment section of the survey that they felt as if the items did not fit the survey, as they were to general and not focused on the change initiative. This may have influenced their responses (increase “average answering”), which could explain the insignificant effect found for affective commitment. The match between types of questions used in the survey is something to take into account in further research. It appears that in empirical change communication research it is advisable to keep the questions in the survey within the domain of change communication.
Moreover, for the factor measuring change effectiveness, very few scales were available and none approached change effectiveness from a dual (both factual and mental) perspective. Therefore it was decided to design a new measure. Various actions were taken to assure the reliability of the measure (e.g. inter-rater scores, Cronbach’s alpha). Still, the construct remains very difficult to measure. It would be interesting for future studies to explore different ways to operationalize change effectiveness. One could for instance interview more managers or use more quantitative data, to get more specific insights in the factual change progress per unit. But
also qualitative approach consisting of in-depth interviews with employees to better understand their personal experience of the change could be a valuable attribution. These findings could later on be used to compose more specific survey questions to measure personal experience of change. Another interesting angle to explore would be the difference between factual change and personal experience. For this study they were combined, but there is still little known about the overlap between these two “sides” of change (Vakola, Armenakis & Oreg, 2013).
A final critical note concerns the usage of SEM. SEM is a flexible research method that makes it possible to test complex relationships among both latent and observed variables. Though, in this flexibility lies also a hazard; one is likely to draw invalid conclusions if the model under investigation is not specified well enough and too many assumptions are violated. This study tried to make sure all requirements for the SEM analysis were considered. The model was well-grounded in theory and most assumptions were met. Yet, the multivariate normality assumption was violated. This study relied on previous findings that the Maximum Likelihood method which was used for analysis tends to be rather robust against violations of this assumption and would therefore provide reliable results (Sherehiy, 2008; Ping, 2010). A possible explanation for the violations of the assumption may be the use of ordinal variables. For these types of variables, it is hard and according to some even meaningless to establish normality and therefore the violation of assumptions is likely (Yang-Wallentin, Jöreskog, & Luo, 2010). The views on how to model ordinal variables in SEM differ enormously. Some argue that asymptotically distribution free estimation is in fact more suitable (Şimşek & Noyan, 2012). For future studies it is advisable to investigate which method would be most appropriate.
With these limitations in mind, it is still possible to address the theoretical and practical implications of this study. How do these findings affect change communication theory and how
can they be used in practice? Most importantly, this study signifies the importance of communication during organizational change - especially for CQI implementations - and also highlights specific focus points for communication. Providing information, empowering employees and limiting the level of uncertainty during change appear to be key criteria for an effective change communication strategy In line with previous findings this study reaffirms that the specific aspects to take into account for information provision are correctness, perceived quality, usefulness, understandability and providing the necessary information on time (Elving, 2005; Lewis, 2006).
To strengthen employees’ sense of organizational community empowering employees appears to be more important than creating affective commitment to the hospital. For empowering employees, both giving the possibility and showing appreciation of input, as well as doing something with their suggestions seem to be crucial. The attribution of empowerment to Elving’s (2005) model proved to be of much value. From a theoretical point of view this study showed that empowerment might be the preferred operationalization of the communicare concept from Francis (1989) and De Ridder (2003 which was used by Elving (2005) for his model. The idea of empowerment stands in fact much closer to the original Latin meaning of
communicare, which means something as “to participate in” or “to partake” (Tate, 2008). From a
practical perspective the fact that empowerment seems to be more important than affective commitment during change might come as a welcome insight for communication professionals, as it is much more concrete, seemingly more feasible and in fact maybe better viewed as a possible antecedent of commitment (Laschinger, Finegan, Shamian & Casier, 2000).
With regard to uncertainty the outcomes of this study are only of limited value. Although the findings do show that lower levels of uncertainty increase likeliness to change, this study did
not found conclusive evidence on how to reduce this uncertainty exactly via communication, as the effects of the other communication criteria on job-related uncertainty were insignificant. In fact the alternative research model which did not include any direct effects on job-related uncertainty proved to be a better representation of the case under investigation. How one can actively combat uncertainty through communication thus remains something to explore in future studies. Meaningful angles to explore based on previous studies could be top-management involvement in communication (Bordia, Hobman, Jones, Gallois & Callan, 2004), involvement of employees in change planning (DiFonzo & Bordia, 1998) transparency, and trust in management (Kramer, Dougherty & Pierce, 2004).
To conclude, the aim of this study was to find evidence in support of a model for effective change communication and to explore how this model could be operationalized for a specific change setting. Regarding the first, this study was able to validate the theoretical model proposed by Elving’s (2005), but did not prove the model to be entirely conclusive for this particular change setting. Regarding the latter, this study has illustrated the workings of the model in a CQI context and has made some meaningful, innovative suggestions to operationalize the models’ constructs for this particular setting. In all, this study shows that it is possible to investigate organizational change by means of empirical research. By doing so, it has hopefully paved to way for future studies in this field to approach organizational change more empirically. It is true that more research is needed in different change settings to fully understand the workings and value of communication during organizational change. However, this study has put forward empirical evidence to help strengthen the belief that communication is of fundamental importance during change. Therefore, it is advisable for managers of organizations to give the communication criteria validated by this study priority during change and embrace their benefits.
References
Allen, N. J., & Meyer, J. P. (1990). The measurement and antecedents of affective, continuance and normative commitment to the organization. Journal of occupational psychology, 63(1), 1-18.
Allison, P. D. (2003). Missing data techniques for structural equation modeling. Journal of
abnormal psychology, 112(4), 545.
Appelbaum, S. H., & Wohl, L. (2000). Transformation or change: some prescriptions for healthcare organizations. Managing Service Quality, 10(5), 279-298.
Baird, K., Hu, K. J., & Reeve, R. (2011). The relationships between organizational culture, total quality management practices and operational performance. International Journal of
Operations & Production Management, 31(7), 789-814.
Bastien, D. T. (1987). Common patterns of behavior and communication in corporate mergers and acquisitions. Human Resource Management, 26(1), 17-33.
Baxter, P., & Jack, S. (2008). Quantitative case study methodology: Study design and implementation for novice researchers. The qualitative report, 13(4), 544-559.
Bennebroek Gravenhorst, K.M., Werkman, R.M. and Boonstra, J.J. (1999). The change
capacityof organisations: general assessment and exploring nine configurations. In L.
Munduate and K.M. Bennebroek Gravenhorst (Eds), Power Dynamics and Organisational Change, Leuven: EAWOP.
Berger, C. R. (1987). Communicating under uncertainty. In M. Roloff & G. Miller (Eds.),
Interpersonal processes: New directions in communication research (pp. 39-62). London:
Bordia, P., Hobman, E., Jones, E., Gallois, C., & Callan, V. J. (2004). Uncertainty during organizational change: Types, consequences, and management strategies. Journal of Business
and Psychology, 18(4), 507-532.
Bordia, P., Hunt, E., Paulsen, N., Tourish, D., & DiFonzo, N. (2004). Uncertainty during organizational change: is it all about control? European Journal of Work and Organizational
Psychology, 13(3), 345-365.
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models (136-162). Newsbury Park, CA: Sage. Burnes, B., & Jackson, P. (2011). Success and failure in organizational change: an exploration of
the role of values. Journal of change management,11(2), 133-162
Chen, C. H. V., & Indartono, S. (2011). Study of commitment antecedents: The dynamic point of view. Journal of business ethics, 103(4), 529-541.
Conner, D. R., & Patterson, R. W. (1982). Building commitment to organizational change.
Training & Development Journal.
Coronado, R. B., & Antony, J. (2002). Critical success factors for the successful implementation of six sigma projects in organisations. The TQM magazine, 14(2), 92-99.
Covin , T. J. , & Kilmann , R. H. ( 1990 ). Participant perceptions of positive and negative influences on large-scale change . Group and Organizational Studies , 15 , 233 – 248
De Ridder, J. (2003). Organisational communication and supportive employees. Human
Resource Management Journal, 13(4), 112-118.
De Souza, L. B. (2009). Trends and approaches in lean healthcare. Leadership in Health
DiFonzo, N., & Bordia, P. (1998). A tale of two corporations: Managing uncertainty during organizational change. Human Resource Management, 37(3-4), 295-303.
Downs, C. W., & Adrian, A. D. (2012). Assessing organizational communication: Strategic
communication audits. Guilford Press.
Dutch Hospital Data. (2013). Kerngetallen Nederlandse Ziekenhuizen 2012. Retrieved 26 May 2014 via http://www.dutchhospitaldata.nl/kengetallen/Documents/Kengetallen%20 Nederlandse%20Ziekenhuizen%202012.pdf
Eisinga, R., Grotenhuis, M. T., & Pelzer, B. (2013). The reliability of a two-item scale: Pearson, Cronbach, or Spearman-Brown? International Journal of Public Health, 1-6.
Elving, W. J. (2005). The role of communication in organisational change. Corporate
Communications: An International Journal, 10(2), 129-138.
Francis, D. (1989), Organisational Communication, Gower, Aldershot.
Garvin, D. A. (2012). The processes of organization and management. Sloan management
review, 39.
Gliem, J. A., & Gliem, R. R. (2003). Calculating, interpreting, and reporting Cronbach’s alpha reliability coefficient for Likert-type scales. Midwest Research-to-Practice Conference in Adult, Continuing, and Community Education.
Gustashaw, D., & Hall, R. W. (2008). From Lean to Green: Interface, Inc.Target, 24(5).
Hellgren, J., Sverke, M., & Isaksson, K. (1999). A two-dimensional approach to job insecurity: Consequences for employee attitudes and well-being. European Journal of Work and
Holt, D. T., Armenakis, A. A., Feild, H. S., & Harris, S. G. (2007). Readiness for Organizational Change The Systematic Development of a Scale. The Journal of applied behavioral science,
43(2), 232-255.
Hox, J.J. and Bechger, T. M. (2000). An Introduction to Structural Equation Modeling. Family
Science Review, 11 (35).
Iverson, R. D. (1996). Employee acceptance of organizational change: the role of organizational commitment. The International Journal of Human Resource Management, 7(1), 122-149 Jak, S., & Evers, A. V. A. M. (2010). Onderzoeksnotitie: Een vernieuwd meetinstrument voor
organizational commitment. Gedrag en Organisatie, 23(2), 158-171.
James, K. M., Ziegenfuss, J. Y., Tilburt, J. C., Harris, A. M., & Beebe, T. J. (2011). Getting physicians to respond: The impact of incentive type and timing on physician survey response rates. Health services research, 46, 232-242.
Johnson, J. R., Bernhagen, M. J., Miller, V., & Allen, M. (1996). The role of communication in managing reductions in work force. Journal of Applied Communication Research, 24, 139 – 164
Kamaruddin, K., Abeysekera, I. (2013). Structural Equation Modelling. Emerald Group
Publishing Limited, 27, 93-123
Kaplan, H. C., Brady, P. W., Dritz, M. C., Hooper, D. K., Linam, W., Froehle, C. M., & Margolis, P. (2010). The influence of context on quality improvement success in healthcare: a systematic review of the literature. Milbank Quarterly, 88(4), 500-559.
Kline, R.B. (2011). Principles and Practice of Structural Equation Modeling. New York: The Guilford Press.
Kolenikov, S. & Bollen, K. A. (2012). Testing negative error variances: Is a Heywood case a symptom of misspecification? Sociological Methods and Research, 41, 124-167.
Koning, H., Verver, J. P., Heuvel, J., Bisgaard, S., & Does, R. J. (2006). Lean six sigma in healthcare. Journal for Healthcare Quality, 28(2), 4-11.
Kramer, M. W., Dougherty, D. S., & Pierce, T. A. (2004). Managing uncertainty during a corporate acquisition. Human communication research, 30(1), 71-101.
Kritchevsky, S. B., & Simmons, B. P. (1991). Continuous quality improvement: concepts and applications for physician care. Jama, 266(13), 1817-1823.
Kruskal, J. B., Reedy, A., Pascal, L., Rosen, M. P., & Boiselle, P. M. (2012). Quality initiatives: lean approach to improving performance and efficiency in a radiology department. Radiographics, 32(2), 573-587.
Laguilles, J. S., Williams, E. A., & Saunders, D. B. (2011). Can lottery incentives boost web survey response rates? Findings from four experiments. Research in Higher Education,
52(5), 537-553.
Laschinger, H. K. S., Finegan, J., Shamian, J., & Casier, S. (2000). Organizational trust and empowerment in restructured healthcare settings: effects on staff nurse commitment. Journal
of Nursing administration, 30(9), 413-425.
Lewis, L. K. (2000). Communicating change: Four cases of quality programs. Journal of
Business Communication, 37(2), 128-155.
Lewis, L. K. (2006). Employee perspectives on implementation communication as predictors of perceptions of success and resistance. Western Journal of Communication, 70(1), 23-46. Lewis, L. K. (2011). Organizational change: Creating change through strategic
McKay, K., Kuntz, J. R., & Näswall, K. (2013). The Effect of Affective Commitment, Communication and Participation on Resistance to Change: The Role of Change Readiness. New Zealand Journal of Psychology, 42(2).
McMillan, D. W. (1996). Sense of community. Journal of community psychology, 24(4), 315-325.
Meyer, J.P. & Allen, N.J.(1997). Commitment in the Workplace: Theory, Research and
Application. Thousand Oaks, CA: Sage.
Morgens. (2014). Lean transitie in de zorg: van projecten naar continu verbeteren. Leiden: Berkel, F.
O'Neill, S. M., Hempel, S., Lim, Y. W., Danz, M. S., Foy, R., Suttorp, M. J. & Rubenstein, L. V. (2011). Identifying continuous quality improvement publications: what makes an improvement intervention ‘CQI’?. BMJ quality & safety, 20(12), 1011-1019.
Oreg, S. (2006). Personality, context, and resistance to organizational change. European Journal
of Work and Organizational Psychology, 15(1), 73.
O'sullivan, P. J., Raguillat, F., Stern, E. H., & Willner, B. E. (2012). U.S. Patent Application
13/471,358.
Peccei, R., Giangreco, A., & Sebastiano, A. (2011). The role of organisational commitment in the analysis of resistance to change: Co-predictor and moderator effects. Personnel Review,
40(2), 185-204.
Ping, R.A. (2010). How does one estimate categorical variables in theoretical model tests using structural equation analysis? Retrieved 19 April 2014 via http://www.wright.edu/~robert.ping/ categorical3.doc
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: a critical review of the literature and recommended remedies.
Journal of applied psychology, 88(5), 879.
Postmes, T., Tanis, M. and de Wit, B. (2001), Communication and commitment in organisations: a social identity approach. Group Processes and Intergroup Relations, 4 (3), 207-26.
Raykov, T. (2000). On the large-sample bias, variance, and mean squared error of the conventional noncentrality parameter estimator of covariance structure models. Structural
Equation Modeling, 7, 431-441.
Rhoades, L., Eisenberger, R., & Armeli, S. (2001). Affective commitment to the organization: contribution of perceived organizational support. Journal of applied psychology, 86(5), 825. Ryu, E., Couper, M. P., & Marans, R. W. (2006). Survey incentives: Cash vs. in-kind;
face-to-face vs. mail; response rate vs. nonresponse error. International Journal of Public Opinion
Research, 18(1), 89-106.
Schweiger, D. M., & Denisi, A. S. (1991). Communication with employees following a merger: A longitudinal field experiment. Academy of Management Journal, 34(1), 110-135
Sherehiy, B. (2008). Relationships between agility strategy, work organization and workforce
agility. ProQuest.
Şimşek, G. G., & Noyan, F. (2012). Structural equation modeling with ordinal variables: a large sample case study. Quality & Quantity, 46(5), 1571-1581.
Swor, R. (2012). Continuous Quality Improvement. Prehospital Care Pearls and Pitfalls, 89. Tanaka, J.S. (1987). "How big is big enough?": Sample size and goodness of fit in structural
