EMOTIONAL INVOLVEMENT
A RESEARCH INTO EMPLOYEE PERCEPTIONS
DURING PARTICIPATIVE, COMPRESSED CHANGE
PROJECTS
Master thesis – September 2007
Msc Business Administration, specialization Change Management
University of Groningen, Faculty of Management and Organization
Martijn van Ooijen Studentnumber: 1503685 De Laan 2A8 9712 AV Groningen tel.: +31(0)6 1440 8000 e-mail: martijnvanooijen@gmail.com Supervisor/ university B.J.M. Emans/ K.S. Prins
EMOTIONAL INVOLVEMENT
A RESEARCH INTO EMPLOYEE PERCEPTIONS
DURING PARTICIPATIVE, COMPRESSED CHANGE
PROJECTS
ABSTRACT
This article discusses employee perceptions towards participative, compressed change projects from the project’s start until the implementation phase. The aim was to explain how emotional involvement of employees to a change project is influenced by content (what the change is about), context (wherein change is realized) and process (how change is realized) and how it fluctuates during the project’s life time. A barometer of emotional involvement was used as a research instrument in two Process Innovation Team Projects of Pentascope at two Dutch healthcare organizations. The hypothesis predicting a collapse of emotional involvement during the calendar midpoint, derived from the punctuated equilibrium theory, was not accepted. It is argued that the involvement of consultants prevented the occurrence of collapses or at least reduced their depths. This outcome points to the manageability of emotional involvement in the pre-implementation phase; content, process, and context need to be managed to ensure a positive perception of the change project. Outcomes of regression analyses indicate that process, as independent variable, and content, as mediator, are the most important variables in determining the emotional involvement. Context is a weaker independent determinant of emotional involvement.
Key Words: Emotion, Emotional involvement, participative change, punctuated equilibrium
INTRODUCTION
For organizational change to be successful, support of the employees concerned is a requirement (Kanter, 1985; Wissema, Messer, & Wijers, 1991; DeVos, Vanderheijden, & Van den Broeck, 2002). Especially in the implementation phase the active support of employees is a critical ingredient (Armenakis, Harris, & Feild, 1999). Naturally, the characteristics of employee perceptions and behaviors regarding change initiatives have drawn the attention of scholars in the field of change management such as Armenakis, Harris and Mossholder (1993) who developed a model of ‘readiness for change’ to measure perceived readiness of an organization to undergo a large-scale change. Several other scholars revealed variables that explain employee support, commitment (see Eby et al., 2000; DeVos et al. 2002) or resistance (see Trader-Leigh, 2001; Oreg, 2003).
While acknowledging the importance of the active support of employees during the
implementation phase, these findings do not explore the influence of employee perceptions from the start of a change project until the implementation phase. Therefore, research was conducted to explore these employee perceptions regarding change initiatives before actual implementation, when employee support or perceived readiness may not be prevalent yet. Mossholder, Settoon, Armenakis and Harris (2000) portray discrete emotions and related constructions as ideal variables to measure employee reactions to change initiatives (e.g. Gilmore, Shea, & Useem, 1997; O’neill & Lenn, 1995).
Emotional aspect of change
So far, scholars have paid limited attention to the emotional side of organizational change (Huy, 1999; Seo, 2003; Ashford & Humphrey, 1995). This is striking, considering the evidence offered that emotions drive the actions of individuals in organizations (Huy, 1999; Seo, 2003; Gersick, 1991; Mossholder et al. 2000; DeVos et al. 2002; Kayzel, 2002). Huy (1999) explains that by paying attention to emotions on the micro dynamic level, it is possible to generate macro changes. Mossholder et al. (2000) state that positive emotion is likely to be related to commitment, while negative emotion is likely to be related to resistance. Seo (2003) explains that some scholars like Argyris (1990) treat emotions as expressions that have to be met with rationally to find the reasons behind the emotions. Actually, emotions play a key role in the cognitive process of perception, decision and behavior itself, Seo argues (2003). The problem with using reasoning to deal with emotions is that individuals turn on different sets of emotions with their automatic nervous system discharge, when they perceive the situation as a threat or opportunity. Seo (2003) argues that these biological reaction patterns of humans are not influenced by reason. In line with Seo, Kayzel (2002) explains that emotions during change initiatives have the tendency to appear on moments that you do not expect them, asking for special attention and a proper insight in the origin of the specific
emotion. As a result, emotions can have an unexpected effect on organizational change and to
prevent emotions to become a surprise, change managers need to recognize the emotional component of a change initiative, Kayzel (2002) argues. Gersick (1991) acknowledges the importance of
emotions as well and states that emotions have a motivational role; they drive the individual’s actions. They could very well be one of the main predictors of radical change because emotions interrupt the regular behavior patterns (Isen, 1984).
Focus of this study
periodic event, while several factors have influence on the success of change initiatives. Studies that only provide snapshots of change fail to provide insights in the dynamic nature of change (Pettigrew, 1988). To prevent this research to solely provide a snapshot of change, the researcher took a
longitudinal perspective. Second, because several studies have focused on emotions during the implementation phase (e.g. Mossholder et al. 2000; DeVos et al. 2002), this study aimed at exploring employee perceptions before the implementation of change. Underlying the choice to focus on the pre-implementation phase is the assumption that building commitment for change should start as early as possible to prevent employee resistance to change in the implementation phase (see Kanter; 1985; Armenakis et al., 1993). To clarify emotional involvement, three decisive categories can be distinguished, i.e. content, process and context. Firstly, this study aimed at explaining how emotional involvement of employees to a change project is influenced by content, context and process.
Secondly, this study aimed at explaining how and why emotional involvement of employees fluctuates during the project’s life time of the observed projects. Commitment for the change initiative or actual change effectiveness was outside the scope of this study, partly because of temporal constraints (this study was focused on the pre-implementation phase) and because this study aimed at explaining the emotional involvement itself.
To gather the necessary data, a barometer of emotional involvement was used as a survey instrument during two participative change projects. The first organization was a hospital in one of the province capitals in the North of the Netherlands and the second organization was a psychiatric hospital in the West of the Netherlands. In the following sections the theory, the methods and the results of the research on this participative change method will be elaborated on. Last but not least, a discussion section has the purpose to expand upon the practical and theoretical consequences of the research.
THEORY
To answer the first question (how is emotional involvement of employees to a change project likely influenced by content, context and process variables?) it is necessary to explain why context, content and process matters. In addition, one may wonder which variables lie beneath these three factors and why. A conceptual model shows the variables per factor that determine the level of emotional involvement (see figure 1). The first part of the theory chapter is devoted to the theoretical underpinnings of the variables in the conceptual model. To answer the second question (in which manner and why emotional involvement of employees fluctuates during the project’s life time of the observed projects?) a distinction was made between project group members and ‘regular’ employees from the departments. For project group members a prediction of a collapse in the emotional
involvement on the calendar midpoint could be derived from Gersick’s theory about group development. To clarify the prediction of a collapse at the calendar midpoint of the project, the Punctuated Equilibrium theory (Gersick, 1988; 1989) is introduced in the second part of the theory section. The assumption that the emotional involvement of all employees (participators or non-participators) fluctuates was not specifically grounded in the literature because of the scarcity of related studies. The exploratory question how and why the emotional involvement of employees fluctuates is conferred in the result and discussion section.
Content, context and process
conditions of the surroundings in which the change unfolds (wherein change is realized). The following three paragraphs will introduce the role of content, process and context in this study. The role of content
Content of the change refers to what the change is about. Two variables determine the appraisal of the content of change: the level of organizational and personal benefit.
Organizational benefit. Employees may believe some changes are necessary but may differ in their idea of the subject to change or what to do with the change subject. Armenakis and Harris (2002) refer to this as “appropriateness of the change” and state that the perception of appropriateness is a possible source of resistance; if there is a large number of employees who are not convinced of the change message, then the subject of the change should be reassessed. Central in the appropriateness is whether the employee believes that the organization will have long-term benefits from the proposed changes (Holt, Armenakis, Field, & Harris, 2007). Therefore, Organizational benefit is assumed to have a significant effect on the content appraisal of organizational change.
Personal benefit. Personal benefit is the second variable that fits with the content elements. The personal interest of an employee regarding the change initiative colors the employees perception of the change initiative; an employee who beliefs the change will be a threat to his/her self-interest will probably resist the initiative (Armenakis & Harris, 2002). Personal benefit therefore examines the feeling that an employee has regarding whether or not he or she will benefit from the change.
In this study the influence of the two content variables on emotional involvement was tested with the following hypothesis:
Hypothesis 1:
The higher the values on content variables (organizational benefit & personal benefit), the higher the value of emotional involvement
The role of process
As discussed earlier, process refers to the manner in which change is realized. Three interconnected process variables influence the emotional involvement of employees: the level of participation, the level of communication and the level of procedural justice. These three variables have a self-sufficient role in the process appraisal of organizational change, be they also show some
relationships with each other. In these paragraphs, both the independence and the interdependence of the process variables will be elaborated on.
Participation. Participation is derived from the wish of us humans to control our own situations; the more we have influence on the events and decision-processes around us, the more we feel good about them (Kanter, 1985). Reicher, Wanous and Austin (1997) explain that employees must have the feeling that their ideas about the change are perceived to be relevant and have been seriously considered to feel good about the change initiative. DeVos et al. (2002) link participation to communication when they state that to be involved, employees need sufficient information.
Moreover, participation enhances the possibility to gather more information about the change. This additional information leads to a better evaluation and validation for the changes and their
underlying purposes (Holt et al., 2007).
Communication. The statement that communication during change projects is important is commonly accepted among scholars (Lewis, 1999) and is stated to be a critical element for the project’s success (Kanter, 1985). Difonzo and Bordia (1998) explain that poor communication within changing organizations leads to uncertainty of employees and creates rumors that go hand-in-hand with negative emotions. Armenakis and Harris (2002) argue that the communicated information about the change initiative shapes the feeling that employees have about this initiative, which in turn determines their way of responding to the change initiative. Elving (2005) adds to this that
Procedural justice. Procedural Justice relates to the fairness of a decision-making process in the perception of individuals involved (Daly & Geyer, 1994). Korsgaard, Schweiger, & Sapienza (1995) argue that a fair treatment has a high influence on the reaction to the decision made. It is even more difficult to question the outcome of whatever is decided upon if the procedure is fair, according to Tang and Sarsfield-Baldwin (1996). Schneider, Brief, & Guzzo (1996) state that a fair decision-making process in the change program is an essential requirement for successful, embedded change. Korsgaard et al. (1995) found evidence in line with the study of Schneider et al. (1996); they proved that participation (another process variable in this study) increases the perception of a fair process, which in turn creates commitment to the decision made; however, employee’s perception of a fair process is only influenced by participation if employees believe it would be appropriate for them to be involved in the decision-making process (Daly & Geyer, 1994). Daly and Geyer (1994) found evidence that explaining the necessity for change and the effects of the changes on the working life of employees (which relates to the level of communication, also present in this study) adds to the perception of a fair process.
In this study the influence of the three process variables on emotional involvement was tested with the following hypothesis:
Hypothesis 2:
The higher the values on process variables (participation, communication & procedural justice), the higher the value of emotional involvement
The role of context
Context deals with the surroundings wherein the change unfolds. According to Pettigrew (1990), context variables consist of internal and external conditions; external conditions include government decisions and market forces, while structural, cultural and political surroundings are part of internal conditions. Due to the focus of this study being on the emotional involvement of employees,
essentially a personal feeling, context was dealt with individually rather than the approach Pettigrew describes. This choice is in line with Weick’s (1979) sense-making process, in which individuals develop a frame of reference to structure and comprehend their own context. Therefore, Pettigrew’s (1990) division of internal and external context was not applicable to this study. The individual context is shaped by its actors, which in this study consists of the managers and colleagues (or peers).
Management support. Management support is defined as the support of the leaders of the company for the change project and its initiatives and is a necessary precondition for change; it feeds the belief that the new direction will be the right one (Kanter, 1985). According to Elving (2005) trust between management and employees could help in this respect; trust shapes the perceptions and guides the actions of individuals in uncertain times. Nutt (1986) studied planned change initiatives in 91 hospitals and found evidence that the most successful change tactics were those in which early and active support of leaders or project sponsors was created. This implies the active involvement of high managers by, for instance, taking part in the steering group. DeVos et al. (2002) state that if project members of change projects get the feeling that those who lead the company are not interested in the progression and outcome of the project, their belief in the success of the change initiative disappears. The opposite occurs when management support is present; it motivates employees to focus on innovations (Zika-Viktorsson & Ingelgård, 2006). However, not only managers shape the context of employees.
which could be achieved by a feeling that the team member is capable, reliable or because it just feels right (Pinto, 2007). The focus in this study was on the level of trust derived from the project-member’s perceived abilities and skills. The belief that the project members have the appropriate skills contributes to the collective belief of the project’s capacities which could help the group when they have to deal with issues like increased time pressure (Nordqvist, Hovmark, & Zika-Viktorsson, 2004). On the work floor level, employees cannot see how a project member functions in the project team, but they will probably have an opinion about the capabilities of the participant: is he or she skilled enough to take part in the project group?
In this study the influence of the two context variables on emotional involvement was tested with the following hypothesis:
Hypothesis 3:
The higher the values on context variables (management support & trust in peers), the higher the value of emotional involvement
Relevance of this study
The abovementioned studies provide the relevance of taking content, process and context into account when studying emotional involvement. While building on the studies of Eby et al. (2000) and DeVos et al. (2002), some differences in the research approach of this study can be noted, which will be discussed in the following paragraph.
Eby et al. (2000) did a study in which employee perceptions of organization’s readiness for change was examined in an organization that changed into a team-based organization. To explain the organization’s readiness, Eby et al. (2000) identified three classes of variables, to be exact:
individual attitudes and preferences (e.g. self-efficacy for change), work group and job attitudes (e.g. participation at work), and contextual variables (e.g. trust in management). At first, a limitation is that the data was gathered on one occasion, which makes it impossible to observe the dynamic nature of the perceptions of the employees regarding change readiness. Second, the model has limited attention for content aspects; just one variable measures the preference of working alone or in a group, which is in fact only a content-related variable. Third, from the perspective of this study, the design lacks an emotional component.
DeVos et al. (2002) developed a framework in which process and context variables explain emotional involvement. The population of their cross-organizational research consisted of change project teams of 35 organizations that underwent an “important change trajectory”. At first, a limitation is that the framework lacks content aspects, while the emotional involvement is likely to be influenced by the substance of a change initiative. Second, solely project members filled out the questionnaire, which means the organization-wide commitment to the change initiatives remains unknown. Third, the data was gathered on a single point in time, which makes it impossible to study the dynamics of the emotional involvement of the respondents.
Content - Organizational benefit - Personal benefit Process - Communication - Participation - Procedural justice Context - Management support - Trust in peers Emotional involvement
Figure 1: conceptual model
Punctuated Equilibrium Theory
To describe the development of groups, several models are offered (e.g. Tuckman, 1965; Hare, 1976; LaCoursiere, 1980; McGrath, 1986). According to Gersick (1988) the mother of all is the model of Tuckman (1965). According to this model, groups undergo the following sequential steps: forming, storming, norming and performing. Tuckman & Jensen (1977) added a final stage of adjourning. In short, forming consists of the first stage in which team members are assigned to create a coherent team; storming is a phase in which conflict can occur and interpersonal behavior is challenged; norming consists of the creation of common ways of working and attitudes; performing consists of the actual production according to the project plan, which can only occur if the first three stages have passed and the project team has reached a certain maturity level; adjourning is the final stage in which the project is terminated (Pinto, 2007). Gersick’s (1988) criticism on this and other related models is that these models assume project groups develop gradually though a universal sequence. Gersick (1988) explains that these models lack insight in the factors that trigger development and the dynamics of the development system itself, and that these models forget to explain the role of the environment on the development of groups.
The second research question deals with how and why emotional involvement of employees fluctuates. Gersick (1988; 1989) describes the fluctuations of project teams (which is one type of group) as atypical. However, timing of when groups shaped, continued and changed the way they worked was similar in all the groups, Gersick observed. In the first few seconds of a project start-up, project group members start implicitly building a common framework of the perceived
circumstances, which includes the way the group will behave. This framework stays in condition until the midpoint of the project. This does not mean all groups in Gersick’s research did the same things during this first phase; some groups experienced a harmonious atmosphere and/or had direction right away, while other groups had no direction and/or experienced internal struggles. At the calendar midpoint a collapse occurs, which is visible through feelings of frustration with the progress so far or with each other. This midpoint is characterized by the team’s contact with the project’s environment (e.g. external supervisors, steering groups, outside experts), the taking on of new viewpoints and impressive steps forward. The original framework of the situation, the common behaviors and the team’s direction are all revised after the collapse, Gersick (1988; 1989)
demonstrates. The trigger that causes the midpoint transition is time awareness or more specifically the feeling of urgency about the impending deadline, which is in line with other theories like Levison’s adult development theory (Gersick, 1991).
by an external focus (explicit attention for outside requirements and revising the project material for exterior use) and an emotional move towards very positive or very negative feelings towards each other and about the project results (Gersick, 1988).
Although Gersick’s model seems to provide a realistic, in-depth insight in the development of project teams, some scholars have criticized the model (e.g. Seers & Woodruff, 1997). Seers and Woodruff (1997) prove that the temporal pattern of midpoint transitions in groups is caused by the deadline itself and not by a group process. Another complaint of Seers and Woodruff (1997) is the fact that there is much attention for individual aspects like attitudes and behaviors in Gersick’s theory but less for group aspects like the development of cohesiveness and the structure of social positions and interpersonal relations. This devalues the ‘group development’ label of Gersick’s model and reduces its comparability with the model of Tuckman (1965). Moreover, the broad applicability of the model is questioned because only for groups with a specific task and a concrete deadline, like a project team, the model is likely to be relevant (Seers and Woodruff, 1997).
For this study the limitations to Gersick’s Punctuated Equilibrium model did not reduce its applicability because the Process Innovation Teams Approach fits the specific characteristics of a project group. Nevertheless, in this study the timing of the midpoint transition is questioned. From earlier experiences with the Process Innovation Teams Approach, the consultants of Pentascope expect a collapse around the tenth week (of the twelve week project). Why should the moment of deadline urgency solely occur at the calendar midpoint of the project? The hypothesis that was used in this study is as follows:
Hypothesis 4:
A collapse in the emotional involvement of project members occurs around the calendar midpoint of the project
Research questions
With the subjects discussed in the theory section the theoretical foundation of the research is established. From the abovementioned hypotheses four sub-questions arise that deal with the effect of content, process and context on emotional involvement (see sub-questions 2-4) and the collapse during the calendar midpoint (see sub-question 5). However, additional questions can be derived from these four questions. Before determining the effect of content, process and context on emotional involvement, it is necessary to check if content, process and context are reliable factors (see question 1). To not only look into the calendar midpoint collapses within the project groups but to fluctuations of emotional involvement in general of all groups under study, the sixth sub-questions was added:
1. To what extent do the variables within the content, process and context factors connect? 2. To what extent is the level of emotional involvement explained by content variables? 3. To what extent is the level of emotional involvement explained by process variables? 4. To what extent is the level of emotional involvement explained by context variables? 5. Does a collapse in the emotional involvement of project participants on the calendar
midpoints occur at the objects under study?
6. How and why does the level of emotional involvement of project participants and non-participators fluctuate at the objects under study?
METHOD
Process Innovation Teams Approach
To study the dynamics of emotional involvement, a new method for participative and compressed change projects, called “Process Innovation Teams Approach”, developed by Pentascope, was chosen. This approach suits the research objective because the approach only covers the pre-implementation phase of change. The approach prescribes that every 12 weeks a small group of representatives (both workers and managers) of one of the departments of the organizations investigates and redesigns their own work process. Every team needs to create a redesign of work that either ensures 10% more production with the same expenses, or saves 10% on expenses while ensuring the same production. The first three weeks are devoted to bringing up the current situation (costs and benefit measurements, inefficiencies etc.) which is validated by the department and the board of directors. From this validated picture of the current situation, a new work process is designed, which is presented for the board of directors and other stakeholders at the end of the project. After termination of the project, the implementation of the redesign starts. Three days a week the project members work together in “the war room” while the consultants of Pentascope facilitate this change process. The combination of soft and hard elements that come together during a tight project schedule make the Process Innovation Teams Approach a unique venture.
Data collection
The study covered two projects in two healthcare organizations. Selection of these organizations was based on the requirement that the Process Innovation Teams Approach was used in their change trajectory. Of the three organizations that underwent a change trajectory based on the Process Innovation Teams Approach, two organizations were willing to participate in the research. The first organization was a hospital in one of the province capitals in the North of the Netherlands. In this hospital two departments took part in the change project, which were the ICT department and
Physiotherapy. For these two departments the questionnaire was filled out weekly by members of the steering group, project group and employees of the participating departments. Next to these two teams, a third team participated that consisted of employees that were designing the Hospital Control Centre. Since this was a newly to be formed department, there were no department employees; only steering group and project group member filled out the questionnaire. The second organization was a psychiatric hospital in the West of the Netherlands. In this hospital, the treatment center took part in the change project. At this location only the project groups filled out the questionnaire on a weekly basis.
The questionnaires were handed out regularly (of the 12 project weeks in the hospital
respectively in weeks 2, 4, 5, 6, 8, 9, 10, and 12 and in the psychiatric hospital in weeks 2,3,4,6,8 and 10) on paper, filled out on a voluntarily basis and respondents were promised that their responses would remain anonymous. To respect confidentiality, but still be able to follow the development of a single respondent’s values during the project, the birthday of one’s mother was used as a ‘secret code’.
Barometer of emotional involvement
research was created. Additionally, all items in the questionnaire needed to be written in Dutch, often requiring translation from English.
The item for the dependent variable emotional involvement was derived from Metselaar (1997) (“I experience –name change project- as something positive”). Metselaar (1997) originally used the term ‘process’ instead of the project name in this item, but since this study assumes emotional involvement is explained by several factors (process is only one aspect of these factors), process was replaced by the project’s name.
The items for the independent content variables were both derived from scales developed by Holt et al. (2007). The first content variable is organizational benefit (“-name change project- will benefit our organization”). The second variable is personal benefit (“-name change project- will benefit me”). The scales developed by Holt et al. (2007) were different in that they both focused on the effect of the changes, while in this study the change project is the subject. Additionally, the scale for personal benefit of Holt et al. (2007) focused on financial benefit while in this study the type of benefit was meant to be open to suggestion, which is why ‘financial’ was removed from the item.
For all independent process variables new scales were developed. The first process variable is participation (“-name change project- seriously considers my input”). The second process variable is communication (e.g. “I am being well informed of the progress of –name change project-”). The third process variable is procedural justice (“Significant decisions of –name change project- are taken in a fair way”).
For the independent context variables, scales from other studies were used. The first context variable is management support wherefore a scale of Boonstra and Bennebroeck Gravenhorst (1998) was used (“Relevant managers are positively involved with –name change project-”). Boonstra and Bennebroeck Gravenhorst (1998) originally used the term ‘top of the organization’ in his item, which was not applicable to the specific characteristics of the Process Innovation Teams Approach, in which both high management as well as middle management are involved. In addition, active involvement was replaced with positive involvement for the sound of it. The second context variable is trust in peers wherefore an item of Cook, Hepworth, Wall, and Warr (1981) was used
(“Participants of -name change project- have the appropriate skills”). The original item of Cook et al. (1981) was in the I-form and used the level of trust explicitly while in the current item trust is more implicitly dissolved in the item.
Data analyses
To answer the first sub-question (to what extent the three factors connect), two assessments were done. At first, for the entire population correlation analyses were used in week 2 to check whether the items within the factors correlated. Items that did not correlate with the other items in the factor were removed. Although the scales were on an ordinal scale, the data showed normal distributions. Therefore, the factors content, process and context were treated as interval data and computed by taking the average of the underlying variables. Second, reliability analyses were carried out on the entire population every week to check whether the factors content, process and context are
homogeneous and reliable. Items that reduced the reliability of the factors were removed.
To answer the second, third and fourth sub-question two analyses were executed. Firstly, to find out how the factors content, process and context correlate with each other and with emotional involvement, correlation analysis was conducted. Secondly, regression analyses were conducted to explain what independent factors determine emotional involvement. The population at the
psychiatric hospital was too small for regression analysis (i.e. a population mean of 4), that is why solely data gathered at the hospital were used for the statistical analyses.
RESULTS
In order to answer the research questions a few topics will be discussed. First, a response table will be presented. Second, outcomes of correlation and regression analyses will be dealt with. Third the fluctuations per survey group will be discussed.
Response rates
Seven groups that filled out the barometer of emotional involvement can be distinguished. In the hospital a steering group, three project groups and two departments took part in the study and in the psychiatric hospital one project group was involved. In the project groups response rates were high (i.e. 50-100%) and in the departments the response rates were at a sufficient level (i.e. 25-64%).
Table 1: Response rates per group
Number
of members Response Response rates in %
Hospital steering group 5 2 ≤ N ≤ 5 40 ≤ R ≤ 100
Hospital project group Physiotherapy 3 2 ≤ N ≤ 3 67 ≤ R ≤ 100
Hospital project group ICT 4 2 ≤ N ≤ 4 50 ≤ R ≤ 100
Hospital project group HCC 4 2 ≤ N ≤ 4 50 ≤ R ≤ 100
Physiotherapy department 28 7 ≤ N ≤ 18 25 ≤ R ≤ 64
ICT department 40 12 ≤ N ≤ 16 30 ≤ R ≤ 40
Psychiatric hospital project group 5 4 ≤ N ≤ 5 80 ≤ R ≤ 100
Correlation and reliability analysis
The first sub-question deals with the connection of the variables within the content, process and context factors. First, correlation analyses were used to check whether the items within the factor correlated (see table 2). The Pearson correlation coefficients within the squares disclose acceptable values to assume correlation within the factors.
Table 2: Correlations (Pearson) between study variables (week 2)
2 3 4 5 6 7 8 1. Emotional involvement ,79** ,70** ,72** ,65** ,61** ,59** ,56** 2. Organizational benefit ,75** ,70** ,67** ,51** ,62** ,54** 3. Personal benefit ,76** ,65** ,61** ,63** ,49** 4. Participation ,67** ,73** ,60** ,60** 5. Communication ,60** ,77** ,53** 6. Procedural justice ,64** ,55** 7. Management support ,61** 8. Trust in peers Notes: 42 ≤ N ≤ 44; ** p < .01 (2-tailed) Squares mark the factors content, process and context
Second, the reliability of these factors was tested (see table 3). The alpha’s differ, ranging from α = .65 (content, week 5) to α = .88 (process, week 8) but are on a sufficient level.
Table 3: Reliability (Cronbach´s Alpha) of factors
Week 2 4 5 6 8 9 10 12
Content ,84 ,75 ,65 ,77 ,71 ,84 ,79 ,79
Process ,82 ,81 ,77 ,82 ,89 ,88 ,81 ,83
Context ,75 ,58 ,62 ,71 ,77 ,80 ,79 ,85
content compared to process and context with emotional involvement. A related finding is that every week the correlations of content with process and context are higher than the correlations of process and context with emotional involvement. To further explore this finding, regression analysis was conducted.
Table 4: Descriptive statistics and correlations (Pearson) between factors
Week N Mean SD Content Process Context
Emotional involvement 4,9 1,3 ,79** ,74** ,63** Content 5,2 ,92 ,79** ,67** Process 4,6 1,1 ,78** 2 42 ≤ N ≤ 44 Context 4,8 ,98 Emotional involvement 4,8 1,3 ,87** ,72** ,61** Content 4,8 1,0 ,76** ,73** Process 4,8 1,2 ,81** 4 41 ≤ N ≤ 43 Context 4,9 ,92 Emotional involvement 4,9 1,2 ,66** ,58** ,39* Content 4,9 ,93 ,74** ,67** Process 4,9 ,79 ,63** 5 39 ≤ N ≤ 41 Context 5,0 ,76 Emotional involvement 4,5 1,1 ,78** ,53** ,42** Content 4,6 1,1 ,59** ,51** Process 4,6 ,92 ,82** 6 37 ≤ N ≤ 37 Context 4,8 ,97 Emotional involvement 4,5 1,3 ,79** ,42* ,47** Content 4,7 1,0 ,71** ,68** Process 4,7 1,1 ,75** 8 35 ≤ N ≤ 37 Context 4,8 ,82 Emotional involvement 4,8 ,98 ,74** ,68** ,69** Content 4,9 ,83 ,66** ,71** Process 4,7 ,86 ,76** 9 35 ≤ N ≤ 37 Context 4,9 ,79 Emotional involvement 4,8 1,0 ,77** ,59** ,65** Content 4,9 ,79 ,68** ,65** Process 4,9 ,79 ,73** 10 37 ≤ N ≤ 38 Context 5,0 ,75 Emotional involvement 4,9 1,2 ,88** ,68** ,55** Content 4,8 1,1 ,74** ,61** Process 4,9 ,90 ,82** 12 33 ≤ N ≤ 34 Context 5,0 ,86
Notes: ** p < .01 (2-tailed); * p < .05 (2-tailed)
Regression analysis
A regression analysis was carried out to check how the factors content, process and context explain emotional involvement (see table 5). In week 2 content is the only factor with a significant beta (ß = .502, sig = .003). Process and context have a significantly lower influence on the level of emotional involvement in this week. The factor correlations (see table 4) and the regression analysis (see table 5) of week 2 indicate a mediation pattern in which content mediates between emotional involvement and process and context factors. Therefore, after one week the conceptual model was adjusted to a model of two independent variables (i.e. process and context), one mediator (i.e. content) and the independent variable emotional involvement (see figure 2). The adjustment of the conceptual model also implied an adjustment of the hypotheses based on this mediation pattern.
Table 5: Outcome regression analysis (week 2)
ß Sig.
Content ,502 ,003
Process ,271 ,161
Context ,089 ,572
Notes: N = 41; β = standardized coefficient; sig. = significance
Content - Organizational benefit - Personal benefit Process - Communication - Participation - Procedural justice Context - Management support - Trust in peers Emotional involvement
Figure 2: conceptual model with a mediatior
Regression analysis with a mediator
Bouma and Emans (2005) explain that to assume mediation four requirements have to be met; at first there has to be a direct relation between the independent (i.e. process and context) and dependent (emotional involvement) factors. Second, the positive mediator (i.e. content) needs to have a positive relation with the independent factors (i.e. process and context). Third, the positive mediator (i.e. content) needs to positively affect the dependent variable (i.e. emotional involvement). Fourth, the influence of the independent variables (i.e. process and context) on the dependent variable (i.e. emotional involvement) needs to disappear or at least decrease in the model with the mediator, which becomes visible through the reduced significance of context and process.
To test for mediation in the eight weeks that the data was gathered, both the correlation table (see table 4) as well as a regression table was used (see table 6). To check whether the first
requirement of Bouma and Emans (2005) was met with, model 1 in table 6 shows the betas of the independent variables on the dependent variable. Notable is the finding that process seems to affect emotional involvement more strongly (beta’s of respectively .62, .54 and .56 with strong significance in weeks 4, 5 and 6) than context (beta’s of respectively .13, .05 and -.03 and weak significance in weeks 4, 5 and 6). Conversely, in week 9 and 10 context affects emotional involvement more
strongly than process (betas of respectively .46 and .47 with strong significance of context versus .33 and .24 with no significance of process). Additionally, table 4 provides the correlations to verify the influence of process and context on emotional involvement.
To be able to check whether the second requirement is met with, the correlation table (see table 4) was used. As was discussed earlier a positive relation between content and context and process is prevalent in every week the data was gathered. To be able to check whether the third requirement is met with, model two in table 6 shows the beta’s of the connection between mediator and dependent variable. Clearly, the beta’s of the mediator expose a strong connection (.66 ≤ ß ≤ .87) between content and emotional involvement. The correlation table (see table 4) confirms this
relation.
The fourth requirement is the decline of the influence of the independent variables on the dependent variable when the mediator is used. The occurrence of this phenomenon should be visible in the betas of process and context in model 3, compared to the values of process and context in model 1 of table 6. In week 2, 9 and 10 the decline of the betas and significance of context and process is obvious. In week 4, 5, 6 and 12 solely process seems to meet with the fourth requirement. To estimate the percentage of the total effect that is mediated by the content variable, the Sobel test, written by Dudley, Benuzillo and Carrico (2004) was used. The mediation percentage per
Table 6: Outcomes of regression analyses with content as mediator
Model 1 Model 2 Model 3
Week Categories
ß Sig. ß Sig. ß Sig. Sobel %
Content -- -- ,79 ,00 ,50 ,00 -- Process ,62 ,00 -- -- ,27 ,16 55 2 Context ,15 ,38 -- -- ,09 ,57 68 R2 ,55* ,60* ,65* Content -- -- ,87 ,00 ,84 ,000 -- Process ,62 ,00 -- -- ,24 ,10 82 Context ,13 ,50 -- -- -,20 ,16 100 4 R2 ,53* ,76* ,78* Content -- -- ,66 ,00 ,58 ,01 -- Process ,54 ,00 -- -- ,24 ,21 65 Context ,05 ,76 -- -- -1,5 ,39 100 5 R2 ,58* ,43* ,67 Content -- -- ,78 ,00 ,72 ,00 -- Process ,56 ,04 -- -- ,17 ,40 80 Context -,03 ,90 -- -- -,09 ,66 100 6 R2 ,28* ,60* ,61* Content -- -- ,78 ,00 ,95 ,00 -- Process ,17 ,48 -- -- -,26 ,16 100 Context ,34 ,15 -- -- ,02 ,93 100 8 R2 ,23* ,60* ,63* Content -- -- ,73 ,00 ,41 ,02 -- Process ,33 ,08 -- -- ,23 ,20 49 Context ,46 ,02 -- -- ,24 ,22 53 9 ,55* ,54* ,62* Content -- -- ,77 ,00 ,61 ,00 -- Process ,24 ,20 -- -- -,02 ,92 79 Context ,47 ,02 -- -- ,26 ,12 61 10 R2 ,45* ,59* ,63* Content -- -- ,87 ,00 ,83 ,00 -- Process ,71 ,01 -- -- ,50 ,62 90 Context -,04 ,86 -- -- -,22 ,83 100 12 R2 ,46* ,76* ,77*
Notes: β = standardized coefficient; sig. = significance; R2 = determination coefficient; * p < .05; Sobel % = percentage of the total effect that is mediated by the content variable – a percentage of
100% indicates a lack of mediation of the content factor
Fluctuations in the emotional involvement of participants
Sub-question 4 was about the predicted collapse of emotional involvement of project participants during the calendar midpoint and sub-question 5 was about the fluctuations of emotional
involvement in general. In this paragraph both the predicted collapse as the fluctuations of the emotional involvement and underlying variables of project participants will be discussed. Since in this study three project groups at the hospital and one project group at the psychiatric hospital were studied, four distinct figures can be visualized. To enhance the clarity of the figures, a percentage was computed by dividing the item mean by seven (the total of the scale). This implies that a scale mean of 4 (which corresponds with “neutral”) equals 50% and a scale mean of 7 (which corresponds with “to a very strong extent the case”) corresponds with 100%. Although the 7-point Likert-scale is an ordinal scale, the data was treated as interval because of its normal distribution. The percentage shows the extent to which an item was filled out positively (high percentage) or negatively (low percentage).
The first figure (see figure 3) reveals the fluctuations of the emotional involvement of the three project groups at the hospital. The three lines show different patterns and while there are moments that the emotional involvement significantly decreases, no real collapse at midpoint is apparent. Furthermore, the percentages of the emotional involvement during the collapses stay well above 70%, which corresponds with “to a certain extent the case”. Not only emotional involvement was illustrated, all items in the questionnaire were visualized (see figure 4, 5 and 6) to check if these underlying variables show extraordinary changes or if they collapsed. Figure 7 shows the
filled out. Per project group both fluctuations in emotional involvement as well as fluctuations of underlying variables are discussed.
74% 76% 78% 80% 82% 84% 86% 88% 2 4 5 6 8 9 10 12 A) Physiotherapy B) ICT C) HCC
Figure 3: Fluctuations emotional involvement of hospital project participants A
B C
In the Physiotherapy department a small collapse is visible during the fourth week (76%), as well as in the twelfth week (76%). The underlying variables (see figure 4) show a different pattern; in general the items trends tend to stay closely together except for procedural justice and
management support. Procedural justice has a low point in the sixth week. Management support has a low point in the eighth week. Remarkable is the low start of trust in peers (67% in week 2). Additionally, all items (except for communication) decrease dramatically in the last week.
60% 65% 70% 75% 80% 85% 90% 2 4 5 6 8 9 10 12 Weeks P e rc e n ta g e Em. Involvement Org. Benefit Pers. Benefit Participation Communication Proc. justice Man.support Trust peers
Figure 4: Fluctuations emotional involvement of hospital project participants Physiotherapy The emotional involvement of project group ICT fluctuates but does not show a dramatic collapse. However, in week 10 the value of emotional involvement is the lowest (77%). The
70% 75% 80% 85% 90% 95% 2 4 5 6 8 9 10 12 Weeks P e rc e n ta g e Em. Involvement Org. Benefit Pers. Benefit Participation Communication Proc. justice Man.support Trust peers
Figure 5: Fluctuations emotional involvement of hospital project participants ICT The emotional involvement of project group of the Hospital Control Center shows a collapse in the sixth week (76%). This collapse is also visible in some underlying variables; with personal benefit, communication and procedural justice having a value of 71%. Remarkable about figure 6 is that every item fluctuates highly during the project’s life-time. Another noteworthy thing is that all items seem to have the highest percentages in the last week, especially organizational benefit (89%).
70% 75% 80% 85% 90% 2 4 5 6 8 9 10 12 Weeks P e rc e n ta g e Em. Involvement Org. Benefit Pers. Benefit Participation Communication Proc. justice Man.support Trust peers
50% 55% 60% 65% 70% 75% 80% 85% 90% 2 3 4 6 8 10 Week P e rc e n ta g e Em. Involvement Org. Benefit Pers. Benefit Participation Communication Proc. justice Man.support Trust peers
Figure 7: Fluctuations emotional involvement project participants in psychiatric hospital Fluctuations in the emotional involvement of non-participants
Sub-question 5 deals with the fluctuations of both participants as well as non-participants. Since the data of project participants was introduced in the previous paragraph, this paragraph will deal with the data of non-participants. Two departments at the hospital were studied, that is the ICT
department and the physiotherapy department.
The items of the ICT department (see figure 8) start off rather low (e.g. communication 46% in week 2 and 37% in week 4) but slowly rise throughout the weeks (e.g. communication 69% in week 12). From week 8 on the values of all items start to rise quickly. Another finding is that some items have the highest values every week (e.g. management support and organizational benefit) and other items have the lowest values every week (e.g. participation, communication, procedural justice and personal benefit). Three out of four of these lower items belong to the process group. This entails that the way the change unfolds was the least satisfying. Emotional involvement rises gradually in the middle of the stream of items (from 56% in week 2 to 69% in week 12). Items tend to follow the same patterns; as one item rises, the others rise as well. The only exception is communication, which is out of step with the other items, especially in the beginning.
35% 40% 45% 50% 55% 60% 65% 70% 75% 2 4 5 6 8 9 10 12 Weeks P e rc e n ta g e Em. Involvement Org. Benefit Pers. Benefit Participation Communication Proc. justice Man.support Trust peers
This implies the way the change unfolds is appreciated better by the physiotherapy department compared to the ICT department. Personal benefit is always on the bottom of the stream (e.g. 57% in week 6). 50% 55% 60% 65% 70% 75% 80% 85% 90% 2 4 5 6 8 9 10 12 Weeks P e rc e n ta g e Em. Involvement Org. Benefit Pers. Benefit Participation Communication Proc. justice Man.support Trust peers
Figure 9: Fluctuations emotional involvement of non-participants in Physiotherapy department DISCUSSION
In the previous sections a study of emotional involvement of employees during a participative, compressed change project (Process Innovation Team Approach) was introduced, theoretically grounded and methodized, and its results were presented. In this section the outcomes of the study and the theoretical and practical consequences will be discussed. Additionally, the research will be reflected upon. This study aimed at explaining how emotional involvement of employees to a change project is influenced by content, context and process variables and how it fluctuates during the project’s life time. A barometer of emotional involvement was used as a measurement instrument at two Dutch healthcare organizations. Recall that to answer the two main questions six sub-questions were posed:
1. To what extent do the variables within the content, process and context factors connect? 2. To what extent is the level of emotional involvement explained by content variables? 3. To what extent is the level of emotional involvement explained by process variables? 4. To what extent is the level of emotional involvement explained by context variables? 5. Does a collapse in the emotional involvement of project participants on the calendar
midpoints occur at the objects under study?
6. How and why does the level of emotional involvement of project participants and non-participators fluctuate at the objects under study?
Three hypotheses were formulated that predicted a positive influence of content, process and context on emotional involvement. Furthermore, a hypothesis for the occurrence of a collapse of emotional involvement of participants during the calendar midpoint of the project was formulated. Whether or not these hypotheses were accepted will be discussed. First, the collapse of emotional involvement of project participants will be discussed. Second, the fluctuations in emotional involvement of non-participants are dealt with. Third, the effect of content, process and context on emotional involvement will be elaborated on. Fourth, the research will be reflected upon.
Collapse of emotional involvement of participants
The fourth hypothesis predicted a collapse of emotional involvement during the calendar midpoint of the projects, which will be discussed in this paragraph. Four project groups at two healthcare
emotional involvement did not collapse at all, but four underlying variables showed a small collapse in week 10. Emotional involvement in the third project group (Hospital Control Center) showed a small collapse during calendar midpoint and ended low. The emotional involvement at the project group of the psychiatric hospital showed a dramatic collapse at the calendar midpoint. Therefore, the fourth hypothesis is not accepted; only two out of four project groups experienced a collapse in emotional involvement at the calendar midpoint. The reasons for the collapse of emotional
involvement differ, but often both personal benefit and procedural justice score low when there is a collapse in emotional involvement. I label some of the collapses as small because during the collapses in the first three project groups, the mean of emotional involvement stay well above 5 of the 7-point scale, which is still a very positive response (corresponds with ‘to a certain extent the case’). Gersick (1988) describes the collapse as moments with feelings of frustration with the
progress so far or with each other. I expected more dramatic movements in the barometer. One of the main differences with Gersick’s study is the involvement of consultants; this is most likely the main cause of the unexpected research outcomes. For that reason, the experiences of the consultants will be added to the discussion.
Generally, the three consultants at the hospital experienced the projects as relatively smooth because no real problems occurred. According to the consultant that leaded the Physiotherapy team, the collapse at this team was less heavy because he learned to anticipate on it; since the collapse was foretold, the project participants expected it and developed a sensible approach towards a collapse (“this is something we have to go through”). According to the consultant, anticipating was also enhanced with the barometer of emotional involvement; it created an opportunity to talk about serious issues and helped to address participant’s emotions. The consultant recognized the collapse in week 12, for which data was gathered during the final meeting. He explains that one of the
outcomes of the project was that the work of the physiotherapist at the treatment centers turned out to be financially unattractive. This finding created a huge setback in the morale of the participants, because the work at the treatment centre is cherished by the physiotherapists; removing it from the work description would not be appreciated. Furthermore, two of the project members disagreed about the direction of the redesign, which fueled the negative emotions in the end. Gersick (1988) describes the final meeting as a moment with heavy emotions, either positive or negative which is in line with data found in this study. The consultant of the ICT project group explains that there was some internal friction because of a mismatch between the goals of the participants and the method Pentascope prescribes, which created some tensions. In line with the consultant of the project group of Physiotherapy, he argues that talking about the possible occurrence of a collapse reduces its depth. The consultant of project group Hospital Control Center recognizes the collapse on the calendar midpoint; the steering group was not satisfied with the amount and the concreteness of the
information they presented during the validation presentation. Additionally, these project members were already working together for more than two project rounds (about eighteen weeks) which may have enhanced a feeling of weariness, the consultant argues.
The overall message of the consultants at these three projects is that talking about the
in line with Gersick’s portrayal of the collapse; seeking outside contact could help to establish a new framework (Gersick, 1988). Furthermore, the consultant explains that the leaving of one of the group members and the joining of a new member helped with establishing new group norms (in this case a dominant member was objected more often). The consultant experienced this as a positive
development.
The first theoretical implication of not accepting the hypothesis is that a collapse in emotional involvement does not only occur during the calendar midpoint, but could also appear on other
moments in time. However, Gersick (1988) did also suggest the possibility of dissimilar timing of collapses. Second, the depth of the collapse could possibly be reduced by forecasting and discussing it. Note that in Gersick’s studies there were no consultants involved and the effect of a consultant was not part of the theory. In this study consultants turn out to have an influence on the project members’ emotional involvement with the project.
Practically, these findings indicate that a collapse in emotional involvement could happen on several moments during the project, and that the depths of these collapses are manageable. Gersick’s group development model emphasizes two phases in which project teams are open to influence on core characteristics: the initial meeting, in which the basic framework is created, and the transition phase, in which the team is not only open to influence but is also seeking outside help to alter the original framework (Gersick, 1988). Although this study may prove the timing of collapses could differ, their characteristics, as introduced by Gersick (1988; 1989), were not challenged. Both the initial meeting as the transition phase after the collapse should have the attention of the consultant (or project manager). During the collapse, the creation of a revised, productive framework is very important to ensure the project’s success. Seeking outside contact (for instance with the steering group) and the rotation of project members could possibly influence the creation of this new
framework. Further research should explore the exact influence of consultants in reducing the depth of a collapse, communication about collapses, rotating project members and seeking outside contact during the transition. Using a longitudinal interventions approach (Bijleveld & van der Kamp, 1998), the effect of the just discussed interventions on the emotional involvement of project participants could be explored. A longitudinal interventions approach prescribes test and control groups to study the effect of interventions over a longer period of time.
Fluctuations of emotional involvement of non-participants
For project participants, the fluctuations were fairly discussed in the previous paragraph. In this paragraph, fluctuations in the data of non-participants are dealt with. Non-participants did not seem to experience a calendar midpoint collapse; instead in these groups the curves show more radical rises and falls based on their reaction towards for instance the project’s progress or the way the participants achieve their goals. Two departments at the hospital were studied, namely the ICT department and the physiotherapy department. At the ICT department all variables had low values in the beginning and rose during the following weeks (in particular from week 8 on). Especially
communication is valued very low in the beginning of the project. While handing out the barometer in week 2, one respondent said that he had no idea what his colleagues were doing in the “war room”. Another said that she did not receive any information about the project. The following week’s respondents kept telling me they lacked sufficient information. In general, the process variables (communication, participation and procedural justice) and personal benefit were valued the lowest week after week, which means the way the change unfolds was the least satisfying.
handing out the barometer, the most remarks I got were on the change initiative itself (content) like: “the American model of efficiency and cost-saving measures in hospitals are criticized, even in the USA, and does not contribute to a positively-loaded work environment” or “the project goal to save €100.0000 is already determined, which means this is not going to be fun!”.
To conclude the story above, one department (Physiotherapy) is relatively satisfied with the way the change is managed, while the other is not (ICT). Zooming in on the two project teams related to these departments, a difference in approach is observable; from the start the physiotherapy team visits their department weekly during the coffee break to inform and involve their colleagues, while project team ICT had only sent emails about their progress up until week 5 (they presented a draft of the current situation in week 5). In week 6, 9 and 11 the ICT project team also organized so called “walk-in-meetings” to give information and to involve employees in finding efficiencies. After this turnaround in the approach of the ICT project team, not only the process but also the other variables started to rise. These findings imply that real life communication and participation is preferred above sending e-mails. It must be stressed that this finding is not empirically tested; further research should explore the effect of the type of communication on the emotional involvement. As was stressed earlier, a longitudinal interventions approach (Bijleveld & van der Kamp, 1998) would probably help to further explore these findings.
Influence of content, process and context on emotional involvement
The preceding paragraphs explain that collapses of emotional involvement could happen at several moments in the project’s life time, and that a consultant could possibly reduce the depth of the collapse. Additionally, in the ICT departments an important rise in the emotional involvement occurred due to a switch in the change process. These findings indicate that emotional involvement is manageable. In this study three hypotheses predicted a positive influence of respectively content, process and context on emotional involvement; these three factors need to be managed to create a positive perception of the change project. That is why the first three hypotheses will be answered in this paragraph.
At first, correlation and reliability analyses have provided the evidence that the factor content, consisting of organizational and personal benefit, the factor process, consisting of participation, communication and procedural justice, and the factor context, consisting of management support and trust in peers were reliable and homogeneous scales. Second, after correlation and regression on the effect of the factors content, process and context on emotional involvement of the data of the first week, a mediation pattern was observable; the correlations of content with process and context were higher than the correlations of process and context with emotional involvement. Content thus seemed to mediate between the independent variables process and context and the dependent variable
emotional involvement. The theoretical implication of this finding is that the way change is managed, and the context wherein the change unfolds influences the way the change itself is perceived, which in turn determines the emotional involvement. The results indicate that this model worked well for 3 of the 9 times it was tested. In 4 of the 9 times process was the single significant independent variable. However, in some of the weeks that both context and process were used as independent variables, context was more significant in determining the value of content compared to process. These outcomes indicate that process, as independent variable, and content, as mediator, are the most important variables in determining the emotional involvement. Context is therefore a less strong independent determinant of emotional involvement, but should not be removed from the model. The initial hypotheses that predicted a positive influence of the three independent variables are accepted but with the comment that process and context indirectly influence emotional
involvement; only content directly influences emotional involvement.
(DeVos et al., 2002); the current study explains how a positive or negative perception of a change project is created. The factors underlying emotional involvement have proven to be relevant subjects to take into account if one wants to manage the emotional involvement of an employee that is one way or another, related to the change initiative.
Reflection
At first the role of the researcher will be reflected upon, secondly the research itself will be reflected upon. I chose a very interactive way of doing research; instead of keeping enough distance to not influence the respondents and the data, I jumped right in. Firstly, at the departments and in the war room, a physical barometer was present, which pointed to the level of emotional involvement per week and the values of underlying variables. In the department the barometer was highly
appreciated, I assume because employees wanted a tool for their grief. Often they put the physical barometer in the red zone (underneath 50%) themselves to send out a signal to their managers that they were not happy. Secondly, to the consultants I weekly sent an e-mail in which I summed up the most remarkable fluctuations and my interpretation of it. For instance, at the hospital I reported the consultants about the low values of the process variables in the ICT department and the difference in approach of the two teams. This implies that the data of this research was influenced by the research itself, which reduced the validity and the reliability of the research. However, the researcher believes the effects of his interaction with the respondents and consultants only faintly blew up the outcomes but did not significantly alter them. This does not reduce the value of the data but these remarks should be considered when interpreting them.
There are a few limitations to this research. Firstly, the variables that determine emotional involvement were measured on single-item scales, while multi-item scales would increase the validity of the model. Secondly, Holt et al. (2007) included individual attributes as a fourth factor (next to content, process and context) in their model for organizational readiness. Individual differences have an effect on the way change initiatives are perceived; the conceptual model of this study would probably benefit from this fourth factor as well. People tend to feel differently about change in general; whether changes in general are perceived to be positive or negative, influences the employees’ perceptions of the change initiatives they are involved in (Holt et al., 2007). I would expect positive individual attributes to have a positive influence on the emotional involvement of employees. Thirdly, in this study emotional involvement was not linked to success or failure of the organizational change initiative. This limitation was caused by the characteristics of the change projects that were studied; the projects were all in the pre-implementation phase. Further research should, while building on the conceptual model of this study, explore the influence of emotional involvement on the success of the change initiative.
REFERENCES
Argyris, C. 1990. Overcoming organizational defenses: Facilitating organizational teaming. Boston: Allyn & Bacon.
Armenakis, A. A., & Bedeian, A. G. 1999. Organizational change: A review of theory and research in the 1990s. Journal of management. 25 (3): 293-315.
Armenakis, A. A., & Harris, S. G. 2002. Crafting a change message to create transformational readiness. Journal of organizational change management 15 (2): 169-183
Armenakis, A. A., Harris, S. G., & Feild, H. S. 1999. Making change permanent: A model for institutionalizing change interventions. In W. A. Pasmore & R. W. Woodman (Eds.),
Research in organizational change and development, Vol. 12: 97-128. New York: JAI.
Armenakis, A.A., Harris, S.G., & Mossholder, K.W. 1993. Creating readiness for organizational change. Human relations, 46 (6): 681-704
relations, 48 (2): 97-125.
Boonstra, J. J., & Bennebroeck Gravenhorst, K. M. 1998. Power dynamics and organizational change. A comparison of perspectives. European journal of work and organizational
psychology, 7: 97-120.
Bouma, J.T., & Emans, B.J.M. 2005. Participatief leidinggeven aan organisatieverandering : Een onderzoek rond de implementatie van customer relationship management. Gedrag en
organisatie, 18 (2): 122-138.
Van der Kamp, L.J., & Bijleveld, C.C.J.H. 1998. Methodological issues in longitudinal research. In C.C.J.H. Bijleveld & L.J. Van der Kamp (Eds.), Longitudinal data analysis: Designs,
models, & methods 1-45. Thousand Oaks, CA: Sage Publications.
Cook, J. D., Hepworth, S. J., Wall, T. D., & Warr, P. B. 1981. The experience of work. London: Academic Press.
Daly, J. P., Geyer, P. D. 1994. The Role of Fairness in Implementing Large-Scale Change: Employee Evaluations of Process and Outcome in Seven Facility Relocations. Journal of
organizational behavior, 15 (7): 623-638.
Devos, G., Vanderheyden, K., & Van den Broeck, H. 2002. A framework for assessing
commitment to change: Process and context variables of organizational change. Working Paper Series 2002-11, Vlerick Leuven Gent Management School, Gent, BE.
DiFonzo, N. & Bordia, P. 1998. A tale of two corporations: managing uncertainty during organisational change. Human resource management, 37 (3): 295-303.
Dudley, W.N., Benuzillo, J.G., Carrico, Mineh S. 2004. SPSS and SAS Programming for the Testing of Mediation Models. Nursing research, 53 (1): 59-62.
Eby, L. T., Adams, D. M., Russell, J. E. A., & Gaby, S. H. 2000. Perceptions of
organizational readiness for change: Factors related to employees’ reactions to the implementation of teambased selling. Human relations, 53 (3): 419-442.
Elving, W.J.L. 2005. The role of communication in organisational change. Corporate
communications: An international journal, 10 (2): 129-138.
Gersick, C. J. G. 1988. Time and transition in work teams: Toward a new model of group development. Academy of management journal, 31 (1): 9-41.
Gersick, C. J. G. 1989. Marking time: Predictable transitions in task groups. Academy of
management journal, 32 (2): 274-309.
Gersick, C. J. G. 1991. Revolutionary change theories: a multilevel exploration of the punctuated equilibrium paradigm. Academy of management review, 16 (1): 10-36 Gilmore, T. N., Shea, G. P., & Useem, M. 1997. Side effects of corporate cultural
transformations. Journal of applied behavioral science, 33 (2): 174-189. Hare, A. P. 1976. Handbook of small group research (2d ed.). New York: Free Press Holt, D. T., Armenakis, A. A., Field, 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
Huy, Q. N. 1999. Emotional capability, emotional intelligence, and radical change. Academy
of management review, 24 (2): 325-345.
Isen, A. 1984. Toward understanding the role of affect in cognition. In R. Wyer & T. Srull (Eds), Handbook of social cognition, vol 3: 179-236. Hillsdale, NJ: Earlbaum. Kanter, R. M. 1985. Managing the human side of change. The management review, 74 (4):
52-56.
Kayzel, R. A. 2002. Veranderen met verstand en gevoel: Een inleiding in de kunst van
Organisatieverandering (2nd ed.). Bussum: Coutinho.
Korsgaard, M. A., Schweiger, D. M., Sapienza, H. J. 1995. Building commitment, attachment, and trust in strategic decision-making teams: The role of procedural justice. The academy of
