Activated faultlines in the context of change: The
effects and its consequences.
JASPER BELTMAN Student number: 1764225
University of Groningen
MscBa, Faculty of Economics & Business Tel: 0613752951
E-mail: Jasperbeltman@gmail.com
Thesis supervisor: Dr. J. Rupert
Abstract
In this research I hypothesize that change will lead to the activation of change related faultlines and consequently will have a negative influence on the effectiveness of the initiated change project. I further hypothesize that this relationship is mediated by coalition formation, open communication and conflict. By researching 9 companies with 33 teams and 155 participants using a survey, I found that, against what was expected, activated change faultlines made change projects more effective. Activated change faultlines also led to significant levels of coalition formation, open communication and conflict within the group. However, these process variables did not mediate the relation between active change faultlines and change effectiveness. This research extend existing literature by explicitly looking at faultline bases likely to be activated during change Furthermore, instead of researching dormant faultlines, what is the focus of existing research, I am looking at the effects of perceived and therefore activated faultlines.
2 TABLE OF CONTENTS
1 INTRODUCTION ... 3
2 THEORY ... 5
2.1 Activated change faultlines ... 6
2.2 Consequences for change effectiveness ... 9
2.3 Mediators: Coalition formation, open communication and conflict ...11
3. METHOD ...16
3.1 Data collection procedure ...16
3.2 Sample ...17
3.3 measures ...17
3.4 Analyses ...19
5 RESULTS ...21
5.1 Descriptive statistics and correlations ...21
5.2 Hypotheses testing ...23
6. DISCUSSION ...30
6.1 Discussion main findings ...30
6.2 Theoretical contributions ...32
6.3 Practical implications ...32
6.4 Limitations and future research ...33
7. CONCLUSION ...34
REFERENCES ...35
Appendix A: Questionnaire items ...42
3
1 INTRODUCTION
A long history of diversity research has led to mixed results, indicating that diversity can have both positive and negative outcomes. With a labor force that is becoming more diverse, better understanding in which circumstances diversity lead to positive outcomes will be key (Milliken & Martins, 1996; Lau & Murnighan, 2005). However, diversity is one of the most difficult topics of organizational behaviour, because the advantages only become in existence under a very narrow set of contextual conditions (Mannix & Neale, 2005). Therefore, lots of opportunities exist for research in examining the role of diversity by incorporating context in which the diversity becomes apparent (Jackson, Joshi & Erhardt, 2003). Context is imperative, because “context provides the purpose, resources, social cues, norms, and meanings that shape behaviour” (Jackson, Joshi & Erhardt, 2003: p.813). Furthermore, different sets of diversity attributes become more or less important in specific contexts (Mannix & Neale, 2005). Therefore, diversity research should incorporate context, to understand the under what conditions the positive effects of diversity becomes apparent.
The mixed results in diversity research might be explained by how diversity research has been approached. Lau & Murnighan (1998) argue that previous research has focused on using one specific diversity attribute, and has researched the degree of distribution of the attribute among group members. A shortcoming of this approach is that it views the distribution of one attribute independent from other attributes. This approach does not encompass the impact of other attributes and the interaction between them. Additionally, depending on the context, different attributes can become more of less salient. Researching one specific diversity attribute misses the impact of context on different diversity attributes. As a result, by focusing on a single attribute, this approach misses the impact of other attributes and their interaction within context.
Therefore Lau & Murnighan (1998) proposed the concept of faultlines. This approach entails that hypothetical separating lines can form a group of people into subgroups, based on specific attributes (Lau & Murnighan, 1998, p328). They argue that individuals within a group possess different attributes simultaneously and that these attributes may align across several attributes. The alignment among groups members of these attributes can divide them into subgroups, thereby effecting team processes and outcomes (Bezrukova, Thatcher, Jehn & Spell, 2012). These faultlines can be activated depending on the context. For example, when a firm engages in affirmative action, this might activate a racial or age-related faultline. Therefore this approach focusses on more than one attribute and also the interaction between them within a specific context.
4 lacks research, in combination with the faultlines approach, is the context of change. Change has been researched in combination with diversity, but these remain focused on a specific attribute. Examples of these papers researching the effects of age (Caldwell, Herald & Fedor, 2004), professional identity (Callan, et al., 2007; Terry & Callan, 1998), gender (Hornung & Rousseau, 2007), organizational status (Martin, Jones & Callan, 2005) and identity loss (Jetten, O’Brien & Trindall, 2002) on change. These papers focus on the distribution of one diversity attributes, and therefore fail to measure the interaction between several attributes. A faultline approach will help to measure the interaction between diversity attributes that are relevant in the context of change.
Researching the effect of faultlines in the context of change seems important, because from existing research it can be deduced that organizational change is a process with high levels of complexity that often end up in failure (Elrod & Tippett, 2003). The success of planned change has been disappointing. Researches who have looked at success rates of recent change argue that more than half end up in failure (Clark & Koonce, 1997; Kotter, 1995), while Burnes (2009) even suggested it to be more than two third. Diversity faultline research could give insight in, or reveal to be an explanation for these failure levels, especially because organizations have increased their interest in diversity to respond to the changing social and economic factors in today’s business environment (Evans & Glover, 2012). And as the case study of Gover & Duxbury (2012) indicate, diversity and faultlines are potential causes for these change failures, because a change project may indicate team members that they share several characteristics with some members of the team but not with others. The awareness of these differences actives faultlines (Carton & Cummings, 2012) and therefore might influence change effectiveness.
Since the context of change is lacking in faultline research (Gover & Duxbury, 2012), I will argue that in the context of change, activated faultlines could be triggered due to several reasons. For example, different groups might be treated differently during change, which makes them aware of the potential faultlines within the group (Chrobot-Mason, Ruderman, Weber & Ernst, 2009). Furthermore, it is hypothesised that activated faultlines will have a negative relationship with change effectiveness and that this relation is mediated by group-processes like conflict, coalition formation and open communication.
5 unnoticed, and therefore will not directly influence team processes or outcomes (Lau & Murnighan, 1998; Bodenhausen, 2011). Potential practical implication will be that change leaders will get insight in the effects of diversity in teams within the organization on the effectiveness of change. Especially the processes behind the effect of diversity gives insight, and can present itself as a starting point for solutions in organizations with high levels of diversity who wish to initiate change, but encounter high levels of failure.
In sum, I will try to answer the following mean question:
To what extend do activated change related faultlines have a negative influence on change effectiveness and to what extend is this process mediated by conflict, coalition building and open communication?
2 THEORY
In this section I will first further elaborate on the concept of faultlines and why this can be interesting in the context of change. Then I will elaborate to what extend dormant faultlines become activated faultlines in the context of change. Furthermore, I will theorize how this might influence change effectiveness and how this relation is mediated by coalition formation, open communication and conflict. These relations are visualized in a conceptual model (Figure 1).
Faultlines are hypothetical separating lines that can lead to the formation of a group of people into subgroups, based on specific attributes (Lau & Murnighan, 1998, p328). Faultlines can be divided into social category and informational faultlines. Social category faultlines find their bases in attributes that are easily noticed, like age, gender and ethnic background (Lau & Murnighan, 1998; Shaw, 2004). This differs from informational faultlines, because such faultlines are more difficult to perceive, such as experience, personality and educational level (Thatcher, Jehn & Zanutto 2003).
Faultlines can also differ in strength. The key indicators in this respect are the level of homogeneity within and level of heterogeneity across subgroups with the same attributes (Lau & Murnighan, 2005). In other words, when different potential faultlines align to split the group into two perfect subgroups, the faultline is very strong. The faultline is weak when potential faultlines do not align with each other and thus have a low potential to create one strong faultline that splits a group into perfect subgroups.
6 interesting because, Lau & Murnighan (1998) state that faultlines can be hidden, but can become salient over time, depending on the context of these faultlines. This means that faultlines can exist, but not have become salient and are therefore dormant. They can, however be triggered due to context, like an organizational change that severely disrupt habituated work processes (Gover & Duxbury, 2012). For example, Lau & Murnighan (1998) state that organizational change might trigger faultlines that lead to a group of young liberals who see change as positive, and a group of old conservatives who do not like to see things change. A faultline approach is therefore an interesting concept to examine in the context of change, because these triggered faultlines might have severe impact on the dynamics within the group and consequently the effectiveness of change.
Gover & Duxbury (2012) have made a beginning in examining faultlines in the context of change, by doing a case study in a Canadian hospital which faced relocation to a different facility. Although the change was seen as beneficial to both staff and management, the change had a negative outcome. Their case study examined the reactions to change by using the concept of faultlines. They found a strong alignment of faultlines between staff and management across different attributes, which led to negative change outcomes. However, while they have made a start in this research, the impact of faultlines on group dynamics in an organizational change context lacks empirical testing.
FIGURE 1 Conceptual model
2.1 Activated change faultlines
7 2010, p26).” Dormant faultlines become activated when these they are actually perceived and become salient, which makes them subjective. Dormant faultlines can be any potential attribute that can split a group into subgroups. It is imperative that activated faultlines are considered in research, because it has a more direct predictability for measuring causal relationships for group dynamics and outcomes in comparison with objective faultlines that remain unnoticed (Bodenhausen, 2011). However, previous research still imphasizes measuring dormant faultlines, thereby neglecting the percetions of team members (Harrison, et al, 2002)
In this research, I will focus on the activation of faultlines in a context of change. As stated earlier, Lau & Murnighan (1998) pose the example that organizational change might lead to a group of young liberals who see change as positive, and a group of old conservatives who do not like to see things change. In this example, the dormant faultlines of age and political orientation are activated, which strongly align and therefore create a stronger faultline.
Several dormant faultlines are interesting and likely to be activated in a change context. In their meta-analysis to individual’s reaction to change, Oreg, Vakola & Armenakis (2011) have identified several antecedents that lead to different individual reactions to change. They identify individual characteristics that predict and individual’s reaction to change. This highlight that when individuals are exposed to change, they are predisposed to respond in a certain way. These individuals’ characteristics include personality traits like , need for participation, tolerance for ambiguity, need for information, and the extent to which team members can anticipate of negative or positive outcomes and personal consequences (Oreg, Vakola & Armenakis, 2011; van Dam, 2005; Cunningham, et al., 2002; Szabla, 2007). Individuals with different levels of these individual characteristics are likely to respond differently towards the change (Oreg, Vakola & Armenakis, 2011).
In an organization these characteristics, or dormant faultlines, might be activated when these differences of individuals within the group become salient due to forces and events that are external to the group (Lau & Murnighan, 1998); change in this case. Such an event is called a trigger. A trigger is an event that acts as a sign that a specific group may be different on diversity attributes in comparison to another group (Chrobot-Mason, et al., 2009). In such an event individuals, become aware of the social differences that exist within the group (Steele, Spencer & Aronson, 2002).
8 resources and opportunities. The benefiting individuals will perceive it as a demonstration of loyalty, whereas the opposing individuals will perceive it as favouritism. Differences in treatment have to do with “the distribution of resources such as promotions, pay, opportunities or praise or disciplinary actions (Chrobot-Mason, et al., 2009).”
Differential treatment during episodes of change will likely activate dormant faultlines. When the change project that is initiated entails that, for example employees with higher needs for participation will be treated better compared with for example employees with no need for participation, they will perceive this as favoritism, resulting in the activation of the faultline need for participation. Bernerth, Armenakis, Field & Walker (2007) call this distributive justice of a change project which is significantly linked with individuals response to change. Gover & Duxbury (2012) also found this trigger where the non-dominant group talked about how the changes only benefitted the dominant group. Thus when employees are treated differently during episodes of change this might activate dormant faultlines.
Another characteristic of a trigger identified by Chrobot-Mason, et al., (2009) is different values. They define it as “a clash of fundamental beliefs regarding what is right and wrong or normal and abnormal (Chrobot-Mason, et al., 2009, p.1775).” As explained, some individuals are predisposed to have different reactions towards change. When change is initiated these predisposed characteristics are triggered by the change. For example, individuals within a group with high and low tolerance for ambiguity, will likely react differently towards change (Oreg, Vakola & Armenakis, 2011), because they have fundamental different values. And because change is characterized by high levels of uncertainty, this makes the dormant faultline of tolerance for ambiguity salient and is therefore active. It is therefore the characteristics of the change, which is the trigger, that leads to the activation of the dormant change related faultlines.
9 2.2 Consequences for change effectiveness
An interesting mechanism in the context of change is sensemaking. Sensemaking theory recognizes people as agents who create their own reality by constructing meaning (Tsoukas & Chia, 2002; Sonenshein, 2010). According to the theory, this means that within organizations we must focus on constructed meanings and interpretations that individuals have constructed.
Balogun & Johnson, (2005) connect this sensemaking process with change. According to them individual make sense of the change in what they call a social process of interaction. In the context of change, this process of sensemaking entails the engagement of individuals to show agency in order to construct interpretations about the change (Weick, 1995). According to Balogun & Johnson (2005) these interaction processes “are the conversational and social practices that the change recipients engage in as they attempt to make sense of the new structure and their new roles and responsibilities” (Balogun & Johnson, 2005, p.1587). They argue that in the context of change, employees thus try to understand what change is about, by increasing their contact with colleagues and share experiences and interpretations of the change. As they describe it, “individuals engage in gossip and negotiations, exchange stories, rumours and past experiences, seek information, and take note of physical representations, or non-verbal signs and signals, like behaviours and actions, to infer and give meaning (Balogun & Johnson, 2005, p. 1576).” It is through a shift in language and these conversations that change comes about.
In their analysis Balogun & Johnson (2005) argue that through sensemaking individual interpretations are merged into one interpretation for all the group. According to this train of thought this process of interaction leads to a unified interpretation of the change at the group level. They do not assume that sensemaking on the level of the group goes without conflict, but in the end its leads to a unified interpretation of the change. Furthermore, they assume that individuals exchange information freely to establish a unified interpretation. This process of information exchange, the social process of interaction, is thus essential to establish a unified interpretation and key for change to come about.
10 integration within the group (Li & Hambrick, 2005). This fragmentation within the group limits informational resources for the individuals, because activated faultlines leads to a stop in communication and information exchange between subgroups (Lau & Murnighan, 2005; Dyck & Starke, 1999).
Consequently different individuals within the group can construct different meaning towards the same change. Sonenshein (2010), for example identified different constructed meaning by individuals in the same change project. Different employees engage sensemaking differently, because they differ in ways of thinking (Balogun & Johnson, 2005) and therefore interpret the same situation differently. Individuals can construct supportive, subversive, preservation and transformational meanings (Sonenshein, 2010). As a consequence individuals also responded differently towards the change by either resisting, supporting or championing the change. Besides acknowledging that different constructed meanings coexist, Sonenshein (2010) also found that the different meanings compete. When faultlines are active, an individual construct meaning only with information from individuals from one’s own subgroup and thus severely inhibits the social process of interaction in the overall group. Consequently, this might lead to a competition of meanings within the group.
The competition of meanings is also articulated by Burnes (2009). He argues that during change shifting groups and individuals compete in order to let their constructed reality becomes dominant during the change. Furthermore, Handy (1986) also illustrates that individuals during change have different interests and actively pursue these interests, regardless of the harm done for the group or organization.
11 Hypothesis 1: Team members who perceive change faultlines, rate the change effectiveness of the change as more negative.
2.3 Mediators: Coalition formation, open communication and conflict
The next step in the conceptual model is through which mediators the relation between activated faultlines and change effectiveness is mediated. I propose that the mediators are coalition formation, open communication and conflict. I especially look at these mediators, because these constructs are key in facilitating or inhibit the sensemaking process and result in competing meanings and might block to change to come about (Balogun & Johnson, 2004). As Burnes (2009) indicates, it is the change in language and communication that indicates individuals that there is change and it facilitates the construction of new meaning. Processes such as coalition formation, less open communication and conflict will therefore likely influence sensemaking.
How activated faultlines lead to increased coalition formation, less open communication and conflict, can be explained following social identity theory (SIT). The faultline approach to research diversity finds their basis in SIT. SIT explains how we as individuals are attracted towards others with the same attributes (Tajfel, 1978; Tajfel & Turner, 1979). The two mechanisms behind social identification are self-categorization theory (Turner 1985) and self-enhancement (Turner, Hogg, Oakes, Reicher & Wetherell, 1987).
Self-categorization is a process where individuals assign and define themselves according to their group membership (Turner, 1985). Group members are emotionally involved in this definition and share social consensus about the group. According to Terry & Callan (1998) self-enhancement means that “a positive sense of the self, or self-esteem, means that people tend to perceive norms and stereotypes that favour the in-group, and they tend to make intergroup comparisons that achieve this same goal.”
12 Activated faultlines will more likely will lead to coalition formation within the group (Jehn & Bezrukova, 2010). Polzer, Mannix & Neale, (1998) define coalition formation as the process of cooperation between parties “to obtain a mutually desired outcome that satisfies the interests of the coalition rather than those of the entire group within which it is embedded” (Polzer, Mannix & Neale, 1998, p.42). In the context of change, coalition formation can be defined as the forming of partnerships with interests for or against change (Cawsey, Deszca & Ingols, 2012).
Coalitions will more likely be built within the group when group members perceive activated faultlines (Jehn & Bezrukova, 2010). They will form coalitions with individuals from which they perceive demographic similarities trough processes like self-categorization (Eisenhardt & Bourgeois, 1988), which lead to distinction between the sub-groups on social identities (Pelled, 1996). The group-related social identity of the group member thus shifts from the entire group toward the subgroup (Lau & Murnighan, 2005).
Furthermore, people form coalitions on the basis that they agree with each other on the definition of the situation in which they have to perform their tasks and reach goals (Murnighan & Brass, 1991). In the context of change, this agreement might be related to how individuals respond to change. Individuals are predisposed to react to change differently (Oreg, Vakola & Armenakis, 2011). These reactions work as faultline bases that can be triggered by the change.
When these faultlines are activated by change, this will likely create subgroups on the basis of similar responses of individuals towards the change. They will do this, because they will agree with each other on the situation in which they have to perform their tasks and reach goals. When they agree on this, they will likely form coalitions (Murnighan & Brass, 1991). And as explained above, when these faultlines collide with each other the subgroup distinction becomes even stronger (Lau & Murnighan, 1998).
13 The process of coalition building is illustrated by the case study of Glover and Duxbury (2012), were such formation of coalition with activated faultlines was actually found. Two groups, staff and management, constructed competing meanings and an own reality about the change. During change the groups acted according to their reality, and created coalitions with individuals who had the same reactions towards the change. This led to staff blocking the change and management supporting the change. They this this, because they thought it was in their own interest. This led the staff to want to keep the status quo, whereas management wanted to change. In the end this led the change to fail, because of the severe competition between these constructed realities and meanings, thus leading to lower change effectiveness.
Hypothesis 2a: Team members who perceive change faultlines, experience more coalition formation within the group.
Hypothesis 2b: Team members who perceive coalition formation within the group, rate the change effectiveness of the change as more negative.
Hypothesis 2c: Coalition formation mediates the relationship between activated faultlines and change effectiveness.
Activated faultlines might also lead to less open communication between the established subgroups. Communication is the vehicle through which team members are able to share information (Pinto & Pinto 1990). Open communication is the extent to which individuals listen to each other, try to understand and feel free to express ideas and feelings at any moment in time (Earley & Mosakowski, 2000). For such open communication to develop, a safe psychological environment is key (Edmondson, 1999), because individuals will only share information and beliefs in an environment of trust and understanding (Lau & Murnighan, 2005).
However, according to Lau & Murnighan (2005), their faultline model suggests that in groups that are divided by subgroups, communication is characterized by misinterpretations and doubts. For example, when a member of a subgroups discloses constructive critique to a member of the other subgroup, this is likely perceived an attack on the subgroup (Bartel, 2001). This consequently is detrimental for the exchange of information that is negative or sensitive and even can lead to a total breakdown of communication (Dyck & Starke, 1999). This breakdown in communication will accentuate the differences between the subgroups, and also its biases (Sheriff & Sheriff, 1953). Consequently, this leads to an decrease in performance and learning within the group (Lau & Murnighan, 2005)
14 and beliefs in this specific context. Disclosing information about one’s own beliefs and goals in this context will confirm the social identity of the group members and the developed stereotype (Phillips, Rothbard & Dumas, 2009). Subgroup members will more likely not disclose their different information with the other subgroup, because he is certain that the other subgroup member will not understand and that it will be met with conflict and disgust (Hewlett, Luce & West, 2005).
Less open communication will reduce change effectiveness. As Balogun & Johnson (2005) indicate the construction of meaning is dependent on a social process of interaction. Hatch (1997) also indicates this as he states that the change of language and symbols lead to new organizational realties and meanings. Because subgroups only communicate with their own members, this will likely lead to different constructed meanings and realities about the change project, and these meanings will compete for dominance (Burnes, 2009). This will lead to political behavior and will lead to uncertainties and conflict between groups (Thompkins, 1990) and therefore reducing change effectiveness.
Hypothesis 3a: Team members who perceive change faultlines, experience less open communication within the group
Hypothesis 3b: Team members who perceive open communication within the group, rate the change effectiveness of the change as more positive.
Hypothesis 3c: Open communication mediates the relationship between activated faultlines and change effectiveness.
Activated faultlines will likely lead to inter-subgroup conflict. First of all a distinction widely used in defining diversity is between readily detectable attributes and underlying or less visible attributes. This distinction is useful, because when different attributes between people are salient, people far more likely engage in responses as stereotyping and prejudices that when attributes are nonvisible. However, nonvisible attributes also can create large differences in orientations towards styles of interaction and issues (Milliken & Martins, 1996).
15 of task disagreement (Wall & Nolan, 1986). Lastly, Jehn & Mannix (2001) identify process conflict, which entails conflict about how a task is accomplished. It deals with the allocation of resources and who should do what within the group.
Members of sub-groups engage in a process of categorization, which lead to distinction between the sub-groups on social identities (Pelled, 1996). Research has shown that this distinction lead individuals to favor their own subgroup members more the other subgroup members (Abrams, Wetherell, Cochrane, Hoggs & Turner, 1990). As a consequence this favoring of own-group members will increase the status of the sub-group (Tajfel & Turner, 1986).
A group that has activated faultlines and created subgroups triggered by change likely will have higher levels of conflict. Conflict arises when two groups become suspicious about the others motives and behaviour (Li & Hambrick, 2005) and when these suspicions become larger, the level of conflict is likely to rise. Through categorization and self-enhancement processes, the distinction between the sub-groups cause competition and hostility between them (Insko, Schopler, Hoyle, Dardis & Graetz, 1990). When change faultlines are perceived by the members of the group, the differences between the groups have become salient. Both groups likely have motives and interests that contradict each other. For example, the group that support and participate in the change likely wants see current tasks changed, because the change project requires it, while the group that blocks the change is likely opposed to such a change in tasks, which will lead to conflict.
Conflict is especially present in episodes of change, which is characterized as competing coalitions of groups or individuals who act to achieve their goals and objectives (Buchanan & Badham, 1999). Handy (1986), indicates that in such a situation individuals with different interests are in constant conflict to direct the company in favour of their interests. Bradshaw-Camball & Murray (1991) support this by indicating that when during change the conditions of two parties which are interdependent and the divergent interest of at least one of the parties are met, conflict will arise. This behavior is what Zaleznik (1970) calls unconscious collusion, which indicates that this is unconscious behavior, which is inevitable behavior in organizational life.
16 Hypothesis 4a: Team members who perceive change faultlines, experience more conflict within the group
Hypothesis 4b: Team members who perceive conflict within the group, rate the change effectiveness of the change as more negative.
Hypothesis 4c: Conflict mediates the relationship between activated faultlines and change effectiveness.
3. METHOD
3.1 Data collection procedure
For this quantitative study, data was collected using a survey in companies within the Netherlands. The survey was used to measure all constructs that are relevant for this research to test the theory outlined above. The reason for developing a survey is that it allows for a theory to be tested using a large sample that will help generalize the results for a wider population (Cooper & Schindler, 2008). During the development of the survey, it was pilot-tested among several students and professors. The survey was then digitalized in order to increase the response rate, by making ease of entry for participants less difficult. However, companies who indicated to have problems with the digitalized version were given the opportunity to fill in paper versions.
To conduct surveys at teams within companies, acquaintances were approached through convenience sampling. Some acquaintances were serving as the contact of the company, while other acquaintances gave contact information of the team leaders. After collecting the contact information, phone calls were made in order to find out the suitability of the companies and their interest in participating. Suitability was based on a number of conditions. First of all, the companies need to have teams. The companies do not necessarily have to call the team a team, for example, it can be called a department. But in order to be a team, team members and others within the organization need to recognize a social system or team and at least three team members should share a responsibility for and end goal within the organization (Hackman, 1987). In order to have meaningful results in the research of faultlines, team could not be larger than 15. Furthermore, the team needed to be exposed to change in the last half year, in order to have the participant still have a fresh memory about the change project. When these conditions were met, and the company showed interested in participating, the survey could be conducted.
17 distributed via the team leader for the specific teams who filled in the paper version of the survey. All team leaders were also asked to fill in a grid with basic demographic information about all members of the team.
3.2 Sample
A total of nine companies were included, with a total of 33 teams. This means that in some organizations only one team filled in the questionnaire whereas in other organizations more teams filled in the questionnaire. These companies ranged from small to large and from public to private companies. In total 150 participants filled in the questionnaire with an age of 16 to 61 (M = 39.68; SD = 11.20), of which 60% was male and 40% was female and 74.2% was full-time and 29.7% was part-time. 70.7% of the participants finished a vocational study or higher and 23.2% had a university degree. Membership length of the group ranged from 2 months to 18 years (M = 22.20 months; SD = 30.60 months). All companies were undergoing a change project which started within the last half year. Change projects ranged from all sizes like takeovers, mergers, new working procedures and new team members.
3.3 Measures
Scales for the variables were selected from papers of high quality and peer-reviewed journals. All the items within the survey were translated from English to Dutch and were checked by both students and professors. The survey was used to let team members rate their perception of the activation of change faultlines, coalition forming, conflict, openness of communication and change effectiveness. The items could be answers according to a seven-point Likert scale, ranging from ‘strongly disagree’ (1) to ‘strongly agree’ (7). In appendix A, an overview of the survey including all the survey items can be found.
Group faultlines. in order to measure activated group faultlines I used 14 survey items to assess whether change faultlines were activated. First, 8 items developed by Jehn & Bezrukova (2008) and adapted by Rupert & Jehn (2008), entailed to what extend team members have the feeling there are subgroups within their team. For these items the factor analyses showed one underling factor and internal consistency was good (α = 0.954). An example question is: ‘During work my team divides in sub-groups.’
18 are predisposed to respond in a certain way. Internal consistency for these items were good (α = 0.909). An example of an item is: ‘To what extend do subgroups become apparent, on the basis of group members who think positively towards the change against group members who less positively about the change.’
Coalition formation. In order to measure coalition formation I relied on Lawler & Youngs’ (1975) measures of the perceived utility of each alliance and the anticipated conflict within each alliance. Perceived utility reflects whether or not forming a coalition leads to the achievement of goals and is measured with two items. Anticipated conflict reflects to what extend effort has been made to keep the coalition together, which is measured with 3 items. One item was likely misinterpreted as positive while it is a mirrored question, which thus had to be removed. The factor analyses showed two underling factors and internal consistency was acceptable (α = 0.670). An example of an item is: ‘In my team I have a lot to win with the members who I form a coalition with.’
Conflict. This factor was measured on a nine-item scale measuring relationship conflict, task conflict and process conflict. This scale is developed by Jehn & Mannix (2001). The factor analyses showed tree underling factors. However for one item respondents likely confused process with relational conflict, which therefore had to be removed. Internal consistency was good for process conflict (α = 0.920), relational conflict (α = 0.865) and task conflict (α = 0.857) . An example of relational conflict is: ‘How much relationship tension is there in your work group.’ An example of task conflict is: ‘how much conflict is there about task responsibilities within your team.’ An example of process conflict is: ‘how much conflict is there about ideas within your team.’
Open communication. This factor is measured on a three-item scale developed by Earley & Mosakowski (2000). The factor analyses showed one underling factor and internal consistency was good (α = 0.768). An example item in of this scale is: ‘We really listen to one another and try to understand the feelings and points of view of each other.’
Change effectiveness. This factor was measured on a six-item scale, measuring satisfaction with the accomplishment of the change goals, the resources that were needed for the change and the process in with change came about. This scale is developed by Higgs & Rowland (2005) and Yukl & Fu (1999). Factor analyses showed double loadings for two items, which had to be removed. The other items showed one underlying factor (Figure 2) and internal consistency was good (α = 0.771). An example item of this scale is: ‘about the functioning of the organization now and during the change, I am….(Likert scale, very unsatisfied’ to ‘very satisfied).
19 control if the team is actually a team I measured task interdependence (Campion, Medsker & Higgs,1993), goal similarity (Jehn,1995) and team stability (Wageman, Hackman & Lehman, 2005). Lastly we control for team size.
3.4 Analyses
First a rotated Varimax factor analysis was conducted for all items of the constructs concerned. The extraction was based on an eigenvalue that is greater than one. The results of this factor analyses can be found in appendix B. As explained above, some items had to be removed for various reasons. Because the analysis based on eigenvalue resulted in 7 factors, and extraction with 8 underlying factors was performed. The result of this analysis can be found in table 2. Additionally, for all constructs reliability analyses were conducted.
Next to this, a Kolmogorov-Smirnoff test was conducted to test the variables for normality, of which the results can be found in table 1. The variables relational and task conflict deviated significantly from a normal distribution. However, the values of skewness and kurosis from these variables stayed within the range of -1.0 and 1.0, which is acceptable to accept normal distribution of these variables (Cooper & Schindler, 2008).
TABLE 1
Kolmogorov-Smirnov test
Variable Statistic Significance Skewness Kurtosis
Coal_form .154 .192 Open_Comm .123 .200 Conf_rel .109 .038 .764 -.630 Conf_task .192 .035 .845 .921 Conf_proc .143 .200 Act_fau_change .171 .096 Change_eff .099 .200
20 considerably reduced. In case of full mediation, the whole relationship between the independent and dependent variable disappears.
21
5 RESULTS
5.1 Descriptive statistics and correlations
The descriptive statistics of the main variables, which are all measured with a seven point Likert scale, can be found in table 2 below.
TABLE 3 Descriptive Statistics
N Minimum Maximum Mean
Std. Deviation Scope of Change 147 2,20 7,00 4,55 1,04 Task Interdependance 150 1,00 7,00 4,78 1,13 Goal Similarity 148 1,67 7,00 5,57 0,97 Real Team 150 2,00 7,00 5,84 1,30 Group Size 149 3 14 5,85 2,51
Perceived Change Fau 135 1,00 5,81 3,21 1,16
Open Communication 148 2,67 7,00 5,39 0,97 Relational Conflict 150 1,00 6,00 2,14 1,15 Task Conflict 149 1,00 6,33 2,81 1,27 Process Conflict 148 1,00 6,00 2,24 1,33 Coalition Formation 145 1,00 7,00 3,82 1,23 Change Effectiveness 148 2,40 7,00 4,92 0,89
22 TABLE 4
Correlations between all main variables
M SD 1 2 3 4 5 6 7 8 9 10 11 12 1. Scope of Change 4.55 1.04 - 2. Task Interdependence 4.78 1.13 ,289** - 3. Goal Similarity 5.57 0.97 ,136 ,062 - 4. Real Team 5.84 1.30 -,034 -,016 ,090 - 5. Group Size 5.85 2.51 -,029 -,043 ,051 -,149 -
6. Perceived Change Fau 3.21 1.16 ,172* ,223** -,084 -,092 ,075 -
7. Open Communication 5.39 0.97 ,130 ,041 ,593** ,095 ,024 -,246** - 8. Relational Conflict 2.14 1.15 ,046 ,240** -,363** -,121 -,003 ,386** -,416** - 9. Task Conflict 2.81 1.27 ,365** ,283** -,126 -,105 ,021 ,417** -,256** ,463** - 10. Process Conflict 2.24 1.33 ,091 ,124 -,174* -,179* ,137 ,396** -,278** ,607** ,418** - 11. Coalition Formation 3.82 1.23 -,008 ,258** -,022 -,089 ,129 ,434** -,120 ,281** ,224** ,313** - 12. Change Effectiveness 4.92 0.89 ,064 ,008 ,400** -,054 ,179* ,226** ,192* -,013 ,068 ,045 ,002 - **. Correlation is significant at the 0.01 level (2-tailed).
23 5.2 Hypotheses testing
Hypothesis 1 states that group members that perceive change related faultlines, rate the change effectiveness of the change as more negative. To test the hypotheses two steps were taken. First of all, all the control variables that correlate significantly with the main constructs were added in the first model. In the second model, the independent variable were added. The results of the regression analyses can be found in table 5 below. As the table indicates, perceived change faultlines relate significantly to change effectiveness (β = .214, p < .05). Despite that the hypothesis stated that this relation should be negative, the regression analysis shows a positive relationship. Therefore, I reject hypothesis 1, because when team members perceive more change faultlines, they rate change effectiveness more positive. The variance for change effectiveness is for 23.9% explained by this model. Furthermore, the control variable goal similarity has a significant positive relationship with change effectiveness.
TABLE 5
Regression analyses change faultlines on change effectiveness
Model 1 Model 2
Independent
variables β Sig. β Sig.
24 Hypothesis 2a states that group members who perceive more change faultlines, experience more coalition formation. The results of the regression analyses can be found in table 6. A significant positive relationship between experienced change faultlines and coalition formation was found (β = .396, p < .01). Therefore I accept hypothesis 2a. After adding perceived change faultlines in the model, the total variance explained for rose from 7.4% to 25%. Furthermore, when group members perceive faultlines, scope of the change and task interdependence have a negative significant relationship with coalition formation.
TABLE 6
Regression analyses change faultlines on coalition formation
Model 1 Model 2
Independent
variables β Sig. β Sig.
25 Hypothesis 2b states that group members who perceive coalition formation, rate the change effectiveness of the change as more negative. The results of the regression analyses can be found in table 7. As the table indicates, no significant relation can be found between coalition formation and change effectiveness (β = -.002, p > .05). Therefore I reject hypothesis 2b. Lastly, goal similarity has a significant positive relationship with change effectiveness.
Hypothesis 2c states that the relationship between perceived change faultlines and change effectiveness is mediated by coalition formation. However, no significant relation exists between coalition formation and change effectiveness, which is a required condition in order to test for testing mediation (Baron & Kenny, 1986). Therefore, I also reject hypothesis 2c.
TABLE 7
Regression analyses coalition formation on change effectiveness
Model 1 Model 2
Independent
variables β Sig. β Sig.
26 Hypothesis 3a states that group members who perceive change faultlines, experience less open communication. The results of the regression analyses can be found in table 8. A significant negative relationship between experienced change faultlines and open communication was found (β = -.182, p < .01). Therefore I accept hypothesis 3a. After adding perceived change faultlines in the model, the total variance explained for rose from 33.3% to 35.6%. Furthermore, goal similarity has a positive significant relationship with open communication.
TABLE 8
Regression analyses change faultlines on open communication
Model 1 Model 2
Independent
variables β Sig. β Sig.
27 Hypothesis 3b states that group members that perceive open communication within the group, rate the change effectiveness of the change as more positive. The results of the regression analyses can be found in table 6. No significant relationship between experienced open communication within the group and open change effectiveness was found (β = -.078, p > .05). Therefore I reject hypothesis 3b. Furthermore, goal similarity has a significant positive relationship with change effectiveness.
Hypothesis 3c states that the relationship between experienced faultlines and change effectiveness is mediated by open communication. However, no significant relation exists between open communication and change effectiveness, which is a required condition in order to test for testing mediation (Baron & Kenny, 1986). Therefore, I also reject hypothesis 3c.
TABLE 8
Regression analyses open communication on change effectiveness
Model 1 Model 2
Independent
variables β Sig. β Sig.
28 Hypothesis 4a states that group members who perceive change faultlines, experience more conflict. The results of the regression analyses can be found in table 7. A significant positive relationship between experienced change faultlines all three types of conflict was found (relational: β = .287, p < .01; task: β = .341, p < .01; process: β = .359, p < .01). Therefore I accept hypothesis 4a. After adding perceived change faultlines in the model, the total variance explained for rose from 17.1% to 27.7% for relation conflict, 27.2% for task conflict and 20.9% for process conflict. Also interesting to note is that when the scope of a change project becomes larger, group members perceive more task conflict (β = .289, p < .01). Furthermore, task interdependence has a negative significant relationship with relational conflict and a positive significant relationship with task conflict. Goal similarity has a negative significant relationship with both relational and task conflict. Finally, the extend to with a team is actually a real team, has a negative relationship with process conflict.
TABLE 9
Regression analyses change faultlines on conflict
Model 1 Model 2: Relational conflict Model 3: Task conflict Model 4: Process conflict Independent
variables β Sig. β Sig. β Sig. β Sig.
29 Hypothesis 4b states that group members that perceive conflict, rate change effectiveness more negative. The results of the regression analyses can be found in table 8. No significant relationship between conflict and change effectiveness was found (relational: β = .126, p > .05; task: β = .111, p > .05; process: β = .079, p > .05). Therefore I reject hypothesis 4b. Furthermore, in all models goal similarity has a significant positive relationship with change effectiveness.
Hypothesis 4c states that the relationship between experienced faultlines and change effectiveness is mediated by conflict. However, no significant relation exists between conflict and change effectiveness, which is a required condition in order to test for testing mediation (Baron & Kenny, 1986). Therefore, I also reject hypothesis 4c.
TABLE 10
Regression analyses conflict on change effectiveness
Model 1 Model 2: Relational conflict Model 3: Task conflict Model 4: Process conflict Independent
variables β Sig. β Sig. β Sig. β Sig.
30
6. DISCUSSION
6.1 Discussion main findings
In this research I researched the effects of change related faultlines on the effectiveness of a change project. Furthermore, I tested whether this relation is mediated by conflict, coalition building and open communication. The main findings of the research are that when team members perceive more change related faultlines, the effectiveness of the change is rated higher. Additionally, more perceived change related faultlines lead to more conflict and coalition building and less open communication. However, conflict, coalition building and open communication have no significant relationship with the effectiveness of the change project.
The first finding of this research indicates that, against what was expected, when team members experience more perceived change faultlines they rate change effectiveness more positive, and not more negative. These findings suggest that when team members perceive more subgroups on the basis of change faultlines, it has a positive effect on change effectiveness. This might be explained by the potential to see the sub-group that expresses negative reactions towards the change as a resource rather than an obstacle for change effectiveness in the implementation of the change project (Knowles & Linn, 2004). Ford, Ford & D’Amelio (2008), explain this by stating that negative reactions show engagement of the change recipients and have value. They for example state that individuals with high levels of commitment towards the organization who expresses negative reactions, are likely to engage in the change process by pressing concerns. These concerns can be seen as a valuable source of feedback for improving the change project (Amason, 1996). However, whether this valuable source is used as a resource depends on the skills of change agent or team leader to utilize the forces (Ford, Ford & D’Amelio, 2008).
31 While perceived change faultlines have a positive relationship with change effectiveness, perceived change faultlines still lead to significant levels of coalition formation, conflict and less open communication between the experienced subgroups. The significant relationship of the change faultlines with these process variables is in line with the model proposed by Lau & Murnighan (1998), when they proposed their concept of faultlines. However, the results also indicate that no significant relationship exists between these group processes and the effectiveness of the change project. This might be explained by several reasons. First of all, as explained, the change agent or team leader could prevent such processes to harm the effectiveness of the change project (Ford, Ford, D’Amelio, 2008).
Second, the positive relation might be explained by faultline distance. Although it was not an original concept discussed by Lau & Murnighan (1998), research has shown that it provides additional insight in group dynamics that are confronted with faultlines (Bezrukova, Jehn, Zanutto & Thatcher, 2008). Distance entails the extend to with two subgroups differ in terms of their attributes. For example, a group can split on the basis of age. A group with subgroups with individuals of the age of 20 and a of 30, is different from a group with subgroups with individuals with the age of 20 and 50. In this case faultline strength is the same, however faultline distance and its effects on group behavior is vastly different (Bezrukova, 2008). Large faultline distances between subgroups prohibit positive relations between subgroups and lead to more hostility (Jetten, Spears & Postmes, 2004). It may be the case that faultline distance for the groups researched in this paper, was not large enough to create insurmountable differences between the subgroups and thereby not influencing change effectiveness as expected.
Thirdly, it can be questioned to what extend team members, who filled in the questionnaire, have complete knowledge of the effectiveness of the change. Change effectiveness focuses on the satisfaction with the result and process of the change (Higgs & Rowland, 2005; Yukl & Fu, 1999) and the resources that are used (Beck, Brüderl & Woywode, 2008). Especially with regard to resources, it is questionable to what extend the initiators of the change shared for example budget calculations for the specific change project with the team members. Additionally team member possibly do not have insight in the extent to which the result that was achieved actually was intended by those who planned the change.
32 project and not the effectiveness of the specific changes effecting their team and the effectiveness of those changes.
The results indicate that during episodes of change, the change will lead to subgroups on the basis of change characteristics. The bases I researched were, need for participation, tolerance for ambiguity, need for information, need for distributive justice and the extent to which team members can anticipate of negative or positive outcomes. The basis indicate that, as Oreg, Vakola & Armenakis (2011) predicted, when individuals are exposed to change, they are predisposed to respond in a certain way and that different team members respond differently and have different attitudes towards change. The different reaction of individuals towards the change consequently led to the formation of subgroups, which had a positive effect on change effectiveness, but led to more conflict and coalition formation and less open communication between the subgroups.
6.2 Theoretical contributions
The main theoretical contribution of this research is the integration of research in the field of change with the concept of faultlines. Although Gover & Duxbury (2012) have made a beginning with their case study within a single company and two teams, I researched to what extend change leads to subgroups over a range of 9 companies and 33 teams, which initiated wide ranging change projects. In doing so I have contributed by explicitly researching faultline bases that are likely to be activated during episodes of change, namely need for participation, tolerance for ambiguity, need for information, need for distributive justice and the extent to which team members anticipated of negative or positive outcomes. Change has been researched in combination with diversity, but these remain focused on a specific attribute, while this research focusses on different attributes and their interaction.
Furthermore, this research has explicitly looked at perceived faultlines. I specifically focus on activated faultlines, whereas previous research mainly focused on dormant faultline (Harrison, et al., 2002). Dormant faultlines are inactive and unnoticed, and therefore will not directly influence team processes or outcomes (Lau & Murnighan, 1998; Bodenhausen, 2011). Therefore this research has looked at faultlines that are perceived and therefore activated, in this case by a change project.
6.3 Practical implications
33 effective change project, one must select individuals who are to react differently towards the initiated change project.. One can do this by for example selecting both individuals with low and high tolerance for ambiguity to be a team member.
However, composing a group with both individuals who are predisposed to react negatively and positively toward the change, leads to an increase in conflict and coalition formation and a decrease in open communication between these individuals when they are confronted with change. Past research shows that these processes severely disrupt group outcomes, such as learning behavior (Rupert & Jehn, 2008), member satisfaction and performance (Jehn & Bezrukova, 2010). And when the change has more impact these effects will likely be stronger. Therefore, it is desirable to reduce the disrupting effects of conflict and coalition building and improve communication between the subgroups during the change project.
6.4 Limitations and future research
While this research has come up with some important contributions, this research is not without limitations. First of all, because of the cross-sectional nature of this research, we have to be aware that no conclusions can be made about causality. For the future this means that we need research that is longitudinal in order to research the whole process of dormant faultlines, its activation and its consequences for team processes. Adding qualitative methods such as interviews with this longitudinal approach will provide information richness which help to explain to activation process and its consequences.
Additionally, groups that filled in the questionnaire did not perceived high levels of activated faultlines (M = 3.21, SD = 1.16), which reduced the variance levels of perceived faultlines. It is interesting to note that although team members had low levels of perceived faultlines, the effect on conflict, communication and coalition formation were still significant. This emphasises the influence of faultlines on processes within the group. A way of tackling this problem in the future is to select teams with high levels of perceived faultlines.
34 With regard to future research, interesting topics still remain to be researched. First of all, this research was unable to determine the objective faultline bases related to change and therefore was unable to measure the activation process by change. For future research to determine the objective basis of for example tolerance for ambiguity, a complete test of all the individual team members level of tolerance of ambiguity should be made. With these levels one can calculate to what extend subjective faultlines are trigger by change by calculation the interaction between them and their effects on activated faultlines.
Additionally, possible mediators or moderators between activated change faultlines and change effectiveness need to be determined. These variables might explain why activated faultlines with dysfunctional group behaviour still leads to a more effective change projects. The influence of the team leader or the change agent and the environment and culture within the group on this process will likely provide good starting points for explaining the relationship.
Furthermore, I researched to what extend a change led to the activation of change faultlines and subgroups. However, I did not research how long this faultlines and related subgroups remained active. Do faultlines on the basis of change faultlines still remain active once the change has finished or do the subgroups dissolve back into a unified group? This question is important, because when subgroups still exist, the potential dysfunctional group behaviour that comes with activated subgroups also still exist. This dysfunctional group behaviour like conflict, less open communication and coalition formation will likely have a deteriorating effect on the performance of the team. It is therefore important if faultlines are still activated and also what can make faultlines become inactive.
Lastly, as can be seen from the correlation table, scope of the change has a significant positive relationship with perceived change faultlines. This is interesting, because this means that when the change has more impact, the perceived change faultlines become stronger. This indicates that a predisposition to react in a certain way towards the change is more likely leads to subgroups when change has more impact. The impact of the change, along with the characteristics of a change project such as change context, content and process (Bouckenooghe, 2010), provide exelent starting point to research what specific changes lead to subgroups formation on the basis of change faultlines.
7. CONCLUSION
35 155 participants using a survey, I found that, against what was expected, activated change faultlines made change projects more effective. Activated change faultlines also led to significant levels of coalition formation, open communication and conflict. However, these process variables did not mediate the relation between active change faultlines and change effectiveness.
REFERENCES
Abrams, D., Wetherell, M., Cochrane, S., Hogg, M. A. & Turner, J. C. 1990. Knowing what to think by knowing who you are: Self-categorization and the nature of norm formation, conformity, and group polarization. British Journal of Social Psychology, 29, 97-119. Amason, A. C. 1996. Distinguishing the effects of functional and dysfunctional conflict on
strategic decision making: Resolving a paradox for top management teams. Academy of Management Journal, 39, 123-148.
Balogun, J. & Hope Hailey, V. 2008. Exploring strategic change. Harlow: Pearson Education. Balogun, J. & Johnson, G. 2005. From Intended Strategies to Unintended Outcomes: The
Impact of Change Recipient Sensemaking. Organization Studies, 26(11), 1573-1601. Balogun, J. & Johnson, G. 2004. Organizational restructuring and middle manager
sensemaking. Organization studies, 26(11), 1573-1601.
Baron, R. M. & Kenny, D. A. 1986. The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182.
Bartel, C. A. 2001. Social comparisons in boundary-spanning work: Effects of community outreach on members’ organizational identity and identification. Administrative Science Quarterly, 46, 379-413.
Batten, J.D. & Swab, J.L. 1965. How to crack down on company politics. Personel, 42, 8-20. Beck, N., Brüderl, J. & Woywode, M. 2008. Momentum or deceleration? Theoretical and
methodological reflections on the analysis of organizational change. Academy of Management Journal, 51(3), 413-435.
Becker, T.E. 2005. Potential problems in the statistical control of variables in organizational research: A qualitative analysis with recommendations. Organizational Research Methods, 8: 274-289.
36 Bezrukova, K., Jehn, K. A., Zanutto, E. L. & Thatcher, S. M. B. 2008. Do workgroup faultlines help or hurt? A moderated model of faultlines, team identification, and group performance. Organization Science, 20(1), 35-50.
Bezrukova, K., Thatcher, S. M. B., Jehn, K. A. & Spell, C. S. 2012. The effects of alignments: Examining group faultlines, organizational cultures, and performance. Journal of Applied Psychology, 97(1), 77-92.
Bodenhausen, G. V. 2010. Diversity in the person, diversity in the group: Challenges of identity complexity for social perception and social interaction. European Journal of Social Psychology, 40(1), 1-16.
Bouckenooghe, D. 2010. Positioning change recipients’ attitudes toward change in the organizational change literature. The Journal of Applied Behavioural Science , 46 (4), 500-531.
Bradshaw-Campbell, P. & Murray, V.V. 1991. Illusions and other games: A trifocal view of organizational politics. Organization Science, 2(4), 379-398.
Brewer, M. 1991. The social self: On being the same and different at the same time. Personality and Social Psychology Bulletin, 17, 475-482.
Brewer, M. 1995. Managing diversity: The role of social identities. In S. Jackson & M. Ruderman (Eds.), Diversity in workteams, (pp. 131-159). Washington, DC: APA Books.
Brewer, M. B. & Gardner, W. 1996. Who is this "we"? Levels of collective identity and self-representations. Journal of Personality and Social Psychology, 71, 83-93.
Buchanan, D A and Badham, R. 1999. Power, Politics and Organisational Change: Winning the Turf game. London: Sage.
Burnes, B. 2009. Managing change. London: Prentice Hall.
Byrne, D. 1971. The attraction paradigm. New York: Academic Press.
Carton, A. M. & Cummings, J. N. 2012. A theory of subgroups in work teams. Academy of Management Review, 37(3), 441-470.
Caldwell, S. D., Herold, D. M. & Fedor, D. B. 2004. Toward an understanding of the relationships among organizational change, individual differences, and changes in person-environment fit: A cross-level study. Journal of Applied Psychology, 89, 868-882.
