Diversity in Teams
The Activation process of a Faultline and the Effect on Team
Conflict During an Organisational Change
16 May, 2014
Master thesis, MSc Business Administration – Change Management
University of Groningen, Faculty of Economics and Business
ABSTRACT
This study examines the relationship between objective and activated faultlines, and the effects of an activated faultline in an organisational change context. A faultline is a relatively new concept in the diversity literature it is defined as an hypothetical dividing line that splits a group into relatively homogeneous subgroups based on group members demographic alignment among different attributes. The attributes for the faultline in this study are resistance to the organisational change and age. Fautlines can be objective or activated. An objective faultline is a potential faultline based on demographic attributes and an activated faultline is that team members perceive the potential faultline based on the demographic attributes. The relationship between the objective and activated faultline is hypothesized to be moderated by the organisational change.
The effects of activated faultlines are investigated by measuring team conflict. Team conflict is divided into three different types: relational, task and process conflict.
Results indicate that the organisational change does not moderate the objective faultlines to be perceived by team members. Although it is found that teams who perceive more activated faultlines will face relational conflict. This results in negative outcomes, because team members are distracted from their tasks.
Keywords:
Diversity, Teams, Faultlines, Subgroups, Activation Process, Team Conflict, Task Conflict, Relational Conflict, Process Conflict
CONTENT
CONTENT ... 3 1. INTRODUCTION ... 4 1.1 DIVERSITY ... 4 1.2 FAULTLINES ... 4 1.3 RESEARCH ... 6 2. THEORY ... 7 2.1 DIVERSITY ... 7 2.2 FAULTLINES ... 72.3 TRIGGERING PROCESS OF FAULTLINES AND FAULTLINE BASE. ... 9
2.4 ACTIVATED FAULTLINES AND TEAM CONFLICT ... 12
3. METHODOLOGY ... 15
3.1 DATA COLLECTION8 ... 15
3.1.1 Data sources ... 15
3.1.2 Data procedure ... 16
3.2 MEASUREMENTS ... 16
3.3 DATA ANALYSIS METHOD ... 18
3.4 DATA ANALYSIS ... 18 3.5 FACTOR ANALYSIS ... 19 4. RESULTS. ... 22 4.1 CORRELATION TEST ... 22 4.2 HYPOTHESIS TESTING ... 23 5. DISCUSSION ... 25 5.1 KEY FINDINGS ... 25 5.2 THEORETICAL IMPLICATIONS ... 25 5.3 MANAGERIAL IMPLICATIONS ... 26
5.4 RESEARCH LIMITATIONS AND FURTHER RESEARCH ... 27
5.5 CONCLUSION ... 28
REFERENCES ... 30
APPENDIX ... 35
APPENDIX 1 QUESTIONS SURVEY ... 35
1. INTRODUCTION
1.1 Diversity
“If we cannot end now our differences, at least we
can help make the world safe for diversity”
This quote of John F. Kennedy (1963) shows that diversity is everywhere. Not one single person is the same, we are all unique and because of that we all think in different ways. According to Kennedy (1963) diversity results in an unsafe world, because of the different thoughts and views of everyone. Looking at organisations the same issues can arise regarding diversity. Because of the changing environment companies have to adapt and change to stay competitive (Klang & Hacklin 2012). Thatchers and Patel (2012) argue that working with diverse teams is essential in adapting to the challenging environment. Williams & O’Reilly (1998) stress that due to the globalising world and significant demographic changes, teams are more diverse.
Organisations want teams to be as effective and efficient as possible (Michael, 2012). Literature shows mixed results about diversity in teams (Williams & O’Reilly, 1998). Some state that more diverse teams are more effective then less diverse teams (Caudron, 1994; Sullivan, 1974). They argue that with more diverse team members, more different views are generated and eventually this will lead to a better outcome than teams with less diverse members. Milliken & Martins (1996) as well as Watson, Kumar & Michaelsen (1993) argue that diversity in teams provides opportunities for creativity and competitive advantages.
On the other hand, diversity in teams can have negative effects. Zhou, Zhang and Sung (2013) argue that gender diversity has negative effects on team outcomes. Jehn (1995) concluded that diverse teams face problems in communication, which resulted in misunderstandings and/or inefficiencies. According to O’Reilly et al. (1989) highly diverse teams are expected to face barriers to social interaction, leading to reduced group cohesion. So as shown here, literature has mixed results regarding diversity in teams.
1.2 Faultlines
more atributes, for example: gender, age and race. Another aspect is that the literature of faultlines focuses on the alignment of multiple attributes instead of the dispersion view of group composition. (Thatcher & Patel, 2012) The faultline concept is tested empirically by several researchers (Phillips et al. 2004; Polzer et al. 2006; Mannix & Neale, 2005; Jehn & Bezrukova, 2010) this is just a small selection of literature that agrees on the existence of subgroups within a diverse team.
A faultline can be distinguished in dormant and activated faultlines (Jehn & Bezrukova, 2010). Dormant faultlines are potential faultlines based on demographic characteristics like: age, gender, education, race etc. An activated faultline refers to group members who actually perceive subgroups based on the demographic characteristics. Riordan (2000) makes the same distinction, only this definition is putting it in other words: objective (e.g. actual) diversity and perceived diversity.
According to Jehn & Bezrukova (2010) activated faultlines are more likely to lead to coalitions and have conflicts among the team. This results in lower level of satisfaction and performance of the team in comparison to groups with a dormant faultline (Jehn & Bezrukova, 2010). The activated faultlines affect the overall group satisfaction in a negative manner, because members of faultline-‐ based subgroups are likely to identify and feel a strong connection to one another but have negative affect toward members who are not part of their subgroup (Hornsey & Hogg, 2000). Mannix and Neale (2005) argue that faultlines have negative effects on the ability of a team to function effectively. Also Van Knippenberg & Schippers (2007) and Lau & Murnighan (1998) agree that diversity faultlines are generally believed to have a negative impact on group processes and performance. When different dimensions of diversity converge (e.g., when all team members with technical expertise are male and those with knowledge about marketing and sales are female), so-‐ called diversity faultlines emerge, can disrupt group processes and can have effect on performance in a negative way. Pearsall et al. (2008) concluded that in a team working with an activated gender faultline, this negatively affected the creativity in comparison to a team with gender faultlines that were not activated.
research when faultlines are triggered and change from objective into activated fautlines and result in team conflict.
1.3 Research
This research extends past research through taking a closer look at when faultlines are triggered. Chrobot-‐Mason et. al (2009) describe triggering as ‘an external force is present, such as an event that serves to polarize or highlight differences based on gender, race, religion, etc., faultlines may become activated’. The external force in this research is an organizational change. Organizational change is defined as understanding alterations within organizations at the broadest level among individuals and groups (Burnes, 2009). During this research a closer look is taken at the triggering process of faultlines during an organizational change and if this results in team conflict.
The next chapter will give more insights on what existing literature states about faultlines, the triggering process of faultlines and the link between faultlines and conflict.
2. THEORY
In this section the existing literature about faultlines will be discussed, resulting in a conceptual model for this research.
2.1 Diversity
As explained in the introduction diversity is a very important aspect in organizations currently. In this research the definition of diversity is: “any attribute people use to tell themselves that another person is different” (Williams & O’Reilly, 1998). This is a very broad definition, but the advantage is that this definition is suitable for multiple categories that are relevant, salient, and critical for the group members (Merton, 1972).
Diversity is very important in today’s global and challenging world. According to Gratton et al. (2007) diversity is the standard at large companies in Europe and the United States, because diverse teams have a range of experience and attitudes to tackle problems that are very difficult. Thatcher and Patel (2012) also state that diverse groups and teams are essential in meeting economic, social, and technological challenges. Other research of Lau & Murnighan (1998); Knippenberg & Schippers (2007); Mannix & Neale (2005) and Pearsall et al. (2008) conclude that diversity in teams can have negative effects. This is a large contrast, because diverse teams are needed to be competitive nowadays, but negative effects can arise due to the diverse teams. Because of this contrast some researchers took another angle of incidence the last decade. Thatcher and Patel (2012) argue that in the past three decades group diversity has been researched over and over again, most of these researches on group diversity focused on the role of group composition in assessing group-‐level outcomes. In past research only one attribute such as race, sex or age was investigated. Lau and Murnighan (1998) researched diversity in a different way. They focused on the alignment of multiple attributes and provided insights to understand the cumulative effects of group member attributes on group outcomes (Thatcher & Patel, 2012). This view of Lau and Murnighan (1998) is known as the faultline theory. A faultline can be defined as: “a hypothetical dividing line that splits a group into relatively homogeneous subgroups based on group members demographic alignment among different attributes” (Lau & Murnighan, 1998). This is the definition used during this research. The next paragraph will elaborate more on the faultline literature.
2.2 Faultlines
whereby a work team is a group (e.g., project team or management team) whose membership and task are formally recognized by the organization (Cohen & Bailey, 1997; Bell & Kozlowski, 2002). Another aspect of a subgroup is that all members in a team have some basic level of interdependence. A subset of members can be considered as a subgroup only if it is characterized by a form or degree of interdependence that is unique when compared to that of other members. For example, a subgroup exists if a set of members interacts differently with each other than with other team members because these members share a cultural value, scarce resource, or knowledge frame that is unique from that shared by other team members. So, subgroups are based on having the same goal in a group and have a basic level of interdependence.
The other element in the definition of Lau and Murnighan (1998) is: “demographic alignment among different attributes”. Demographic attributes are for example race, sex, nationality, age, functional background, education etcetera (Thatcher and Patel, 2012). But Lau and Murnighan (1998) also distinguish non-‐demographic attributes were faultlines can be based upon, for example: personality type or dispositional resistance. Lau and Murnighan (1998) developed the faultline theory based on different theoretical mechanisms. These different mechanisms are:
-‐ Self-‐categorization (Turner, 1985; Turner et al. 1987) -‐ Social identity (Tajfel and Turner, 1986)
-‐ Similarity attraction (Byrne, 1971)
These mechanisms explain how faultlines are formed within a group. The self-‐categorization theory describes how team members classify themselves and other members based on salient characteristics (Thatcher & Patel, 2012). Tajfel and Turner (1986) describe that social identity is about team members who have assumptions about which social group, with the same emotional and value significance, he/she belongs to. The similarity attraction (Bryne, 1971) is a theory explaining that team members that share similar characteristics are attracted to each other to form a subgroup within a team. Research shows that demographic attributes accommodate one means for determining similarity, classification, and identification (Harrison, Price, & Bell, 1998; Horwitz & Horwitz, 2007; Tsui, Egan, & O’Reilly, 1992). Here the alignment of the different attributes result in a faultline (Bezrukova, Jehn, Zanutto, & Thatcher, 2009). Resulting in a homogenous subgroup (Lau & Murnighan, 1998).
example there is a very strong faultline, because there are differences based on four demographic attributes (occupational roles, gender, age and nationality) and the two different groups (females and males) have high alignment between the attributes. Thathcher and Patel (2012) argue that age is the strongest attribute when looking at faultline strength. Followed by race, gender, tenure, functional background and education. Stronger faultlines can result in more team conflict (Li & Hambrick, 2005). This relation will be elaborated in paragraph 2.4.
The other aspect of faultlines is faultline distance: ‘Faultline distance is the extent to which subgroups diverge as a result of accumulated differences between subgroups’ (Bezrukova et al., 2009). When looking at the previous example and zooming in on the attributes age and gender. It will influence the behaviour regarding the faultline if the males are around the age of 35, instead of 50, which is closer to the age of the females (Bezrukova et al., 2009). The faultline distance score would be lower in the second example in comparison to the first example.
2.3 Triggering process of faultlines and faultline base.
Before elaborating on the triggering process of faultlines an important difference has to be clarified. Fautlines can be dormant or activated. Jehn & Bezrukova (2010) extended research on dormant faultline to activated faultlines. A dormant faultline is a potential faultline based on demographic attributes and an activated faultline is that group members perceive the potential faultline based on the demographic attributes.
Chorbot-‐Manson, Ruderman, Weber and Ernst (2009) argued that until 2009 there was very little research conducted on what type of events or situation can activate faultlines. For managers this can be valuable information, because faultlines can also lead to team conflict. If team leaders know which type of events result in activated faultlines, they could take into account the effects of an event or take action before the event takes place (Chorbot-‐Manson et al., 2009).
Chorbot-‐Manson et al. (2009) link the social identity theory described in the previous paragraph to triggering faultlines. They argue that an event makes group members vigilant about the influence on behavior regarding the social identity of other group members. Such an event can be the trigger for dormant faultlines to be activated. Hall (2013) states that the trigger or activation of faultlines depends on the task context a team is in. Because if a team has two subgroups, one with team members who work at a firm for more than 20 years and a subgroup with members who have much less tenure a faultline could be activated by for example seniority privileges.
operational and the strategic level”. Furthermore Burnes (2009) argues that for any organization, its ability to identify where it needs to be in the future and how to manage the changes required for getting there, are of primary importance. According to various studies 70% percent of organizational change result in failure, this statement can be seen as confirmation of the statement made by Burnes (2009) which states that successful change is difficult to accomplish.
When organizations are implementing a change, employees will respond to this. Employees differ in how they respond to change according to Oreg et al. (2008). Some employees’ act very positive to change, others tend to avoid it if possible or even resist to the change (Oreg et al., 2008). In this research the focus is on resistance during an organizational change. But how does resistance link to faultlines? As mentioned in 2.1 a faultline is a homogenous subgroup based on attributes (Lau & Murnighan, 1998). Within this research is conducted if resistance can be an attribute to form a subgroup. For example in a team of seven there can be three employees who score high on dispositional resistance to change and four who score low on this concept. Then there is an objective faultline, which probably no one is aware of. Because of the change the three employees can perceive that they are all against change and form a subgroup on this basis.
Before elaborating on resistance of employees towards change it is important to mention that this topic got very little attention in previous studies according to Oreg & Berson (2009). From 2008 there has been a surge of studies demonstrating a key role of employees’ attitude towards change. (e.g. Fugate, Kinicki, & Prussia, 2008). This is because before 2008 studies had a macro perspective, which means they only focused on factors like organization structure, strategy or environment (Oreg & Berson, 2009). In the conclusion of these studies often is mentioned that the actions of leaders are reflecting the success of a change (Oreg & Berson, 2009). Later studies looked more at the micro-‐level, which mean that they focused more on reactions of employees towards a change. (Fugate, Kinicki & Prussia, 2008).
As mentioned before approximately 70% of changes in organisations fails (Burnes, 2009). Employees’ resistance to change is often cited as the reason why change fails (Ford, Ford & D’Amelio, 2008). Piredit (2000) argues that resistance to change can strictly hamper the change process. So if different sources conclude that resistance to change can lead to failure it is important to know how this resistance arises.
scale to measure the resistance to change. This scale identified four factors that determine if an employees is dispositional resistant to change. These four factors are the following:
-‐ Routine seeking, employees want well defined and familiar tasks and have reluctance to
give up old habits (Watson, 1971). These employees like familiarity, because it makes them feel comfortable. This is why the prefer routine and it is reasonable to expect that these employees resist to an organisational change.
-‐ Emotional reaction to imposed change, researchers suggest that organisational change is a
stressor and that only resilience employees can cope with a the stress involved with a change (Ashforth & Lee, 1990). Less resilient employees are more reluctant to make changes, because in this way they admit that in the past they worked inefficient. Therefore a change means loss of face for them and that is why they will resist.
-‐ Cognitive rigidity, research suggests that the cognitive processes underlying employees
response to organizational change might be predicted by the trait of dogmatism (Rokeach, 1960). Dogmatic employees are characterized by rigidity and close-‐mindedness and they might be less comfortable with new situations and will resist to organisational change.
-‐ Short-‐term focus, an organisational change does not come naturally. Employees have to put
effort into an organizational change to make it work. Kanter (1985) argues that employees can resist, because it often involves more work in the short term to learn new tasks for example. After the short term investment employees will normally have advantages, but they resist putting extra effort in the short-‐term.
Based on the above the link in this research with dispositional resistance and faultlines is that dispositional resistance can be a non-‐demographic base for an activated faultline. But as Lau and Murnighan (1998) state, a faultline divides a group on the basis of one or more attributes into subgroups. The first attribute is resistance and the second attribute used during this study to measure faultlines in a team is age. Weiss and Maurer (2004) state that older employees are more resistant to an organisational change than younger employees. Due to an aging population and relatively low birth rates in most industrialized nations, firms will see an increase of the average age of their personnel (Peeters & van Emmerik, 2009). Burnes (2009) states that organisations are continuously changing due to the rapid changing environment. This can be conflicting if organisations face an older workforce that has to change more often. This is why age is chosen to be the next base for the faultline in this research.
employees, but that hard evidence cannot be found. The same accounts for the resistance to change, the assumption is that older employees are seen as more resistant to change than younger employees (Weiss and Maurer, 2004). Kunze et al. (2013) researched if this perception of the negative relation of age and to resistance to change is true. In this research Kunze et al. (2013) used the Pogson et al. (2003) stages of a career to define older employees.
-‐ Stage 1: trial stage, < 31 years old, employees of this age are less routine seeking, because they have less experience compared to older employees. These employees are changing more often of employer.
-‐ Stage 2: stabilization stage, 31 -‐ 44 years old, employees of this age have mostly found what kind of work is appropriate to them and are changing as little as possible.
-‐ Stage 3: maintenance stage, 45 years and older, these employees want to keep their habits during their work and do not want to change the job they have.
Older employees are the ones who are in their maintenance stage. Kunze et al. (2012) used the same scale for resistance to change as this study does, namely the resistance to change scale of Oreg (2003). The stereotype view on elderly employees is that they are more routine seeking, more short-‐ term focused and more cognitively rigid. Kunze et al. (2012) studied German firms and after reducing the data set, 2.981 participants were analysed. The result of the study is that younger employees were more resistant to change than their elderly colleges. So again the stereotype view is not in line with the findings. But this research show that age and resistance are interrelated to each other. This results in the first hypothesis that is presented below:
H1: An organisational change moderates the relationship between the objective faultline and activated faultline based on resistance and age.
2.4 Activated faultlines and team conflict
subgroups. This results in conflict among subgroups that are formed during the activation of the faultline.
Jehn en Mannix (2001) categorized team conflict into three different types:
-‐ Relationship conflict: This is when a team member feels tension, friction, annoyance, frustration, irritation or dislikes another member within the team. Or that a team member is aware that he or she has interpersonal incompatibilities (Amason, 1996; Pinkley, 1990). These are all interpersonal issues, which distract team members from the task (Jehn, 1997). -‐ Task conflict: This is when team members are aware that there are different thoughts and views to a group task. So this is conflict about how to get the work done. These are issues that are related to tasks.
-‐ Process conflict: Jehn (1997) identified a new type of conflict, named: process conflict. This is when there is awareness of differences among team members about how task are completed. For example if team members have conflict about how resources are divided, who should do what or what the responsibilities are of individual members.
On one hand conflict can be seen as a negative outcome and something that has to be avoided (Stone, 1995). On the other hand there are studies that show benefits regarding group performance. Jehn (1995) argues that task conflict on moderate level improves the decision quality, because there is discussion. Instead of one individual perspective a team makes a combination of more perspectives to complete a task (Schwenk 1990). Rispens (2012) endorses Jehns’ (1995) findings that task conflicts may be beneficial for team performance. In contrast negative issues may pop up during conflict such as: lack of coordination, cooperation and cohesion (Brewer, 1996). Moreover task and relationship conflict correlated within team, so if a team has the ‘benefit’ of task conflict they may also have the negative influence of relationship conflict (Rispens, 2012). Research indicates that relationship conflict has a negative effect on team performance (Huang, 2012; Rispens, 2012). De Dreu and Weingart (2003) show that all conflict types are detrimental to group outcomes. Jehn (1995) endorses these outcomes, because she states that conflict is detrimental to individual and group performance and member satisfaction. This is why a closer look is taken to team conflict, because it can lead to less performing teams.
Based on the above the following hypotheses regarding activated faultlines and team conflict are developed:
H2a: The more faultlines are perceived, the higher the degree of relationship conflict within this team. H2b: The more faultlines are perceived, the higher the degree of task conflict within this team.
All the literature is summarized in the conceptual model presented below.
Figure 1: Conceptual Model
3. METHODOLOGY
The following chapter of this thesis presents how this research is designed. The methodology is divided into three parts: the data collection, measurement and data analysis for this study.
3.1 Data collection
This study will use quantitative research to look at faultlines during or after an organizational change. A quantitative study has the advantage that opinions, attitudes and behaviour can be measured in an objective manner. In this research it is important to find out how a team perceives the subgroup/faultlines and the processes in a team during an organisational change. Because of the literature gap with faultlines during an organisational change quantitative research is used to obtain a lot of data and to see what is happening in teams. Further advantages of this cross sectional study is that the data can be gathered quickly and in a reliable way by using a standardized questionnaire so the data is easy comparable.
As mentioned before the data is collected with a survey that is distributed among teams of organizations in the Netherlands. The organisations are mainly of my own network. Organisations were contacted through e-‐mail and most of the time a face-‐to-‐face meeting followed. In this conversation the objectives of the research are explained and explored if the team that was researched meets the criteria. This is because all the teams in this research have to meet three criteria. The first is that they are a team according to the criteria of Hackman and Wageman (2005). These criteria are that two or more individuals have the same goal(s) and interact interdependently. The second criterion is that the teams consisting of three to seven members are ideal for this research. Carton and Cummings (2012) state that people strongly prefer groups composed of two or three members, so in groups of three to seven persons it is more obvious that there are faultlines. It is not reliable to research larger groups, because Levine & Moreland (1998) argue that people rarely try to integrate in groups of more than six members. So larger groups are more likely to contain subgroups and because this research also looks at the triggering process during an organisational change it is not reliable to look at larger teams, because despite of the change they could contain subgroups. The last criterion of a team to fill in the survey is that they are in an organisational change or faced one maximal a year before filling in the survey. In this way the team members still have notion of what the change was about and how it affected their team.
3.1.1 Data sources
two different data sources are collected and can be compared. By using two different sources the effect of common method bias is less than only using one source (Hirunyawipada and Vahie, 2005). So in this way the research is more valid and reliable than only having one source of data. Further on a managerial grid was collected of all team members with this demographic information: gender, age, function, education and tenure. Because for measuring the faultlines in a group it is important to have data from every team member. If one team member is missing the faultline strength and distance could still be calculated, because of the information in the managerial grid.
3.1.2 Data procedure
The survey starts with an introduction about the purpose of the research and it is mentioned that it is on an anonymous base to prevent that team members are not completely honest, for example if they are having problems with other members or their team leader. Most questionnaires are handed out on paper, because in this way the response is higher than using the online survey according to Semler (2010). The questionnaires were handed out on the desks of the team members and they could fill it in whenever they wanted. A deadline was set within two or three days after distributing the surveys to prevent that the survey ended at the bottom of the to do list. As mentioned before it is very important to let all team members participate, because if a team consist out of six members and only four respond than it is really difficult to have reliable results. The two members that are missing could form a faultline and by missing the data of those two members it will not make sense to use that team. So if the response is low a second deadline is set to gather as much data as possible.
Some teams preferred to participate via the online variant of the survey. These teams received an e-‐mail with an introduction to the research and a code that they had to fill in before answering the questions. In this way it is clear which team did fill in they survey. To increase the response a reminder was send to the teams one week after the first e-‐mail.
The managerial grid was collected was collected when visiting the organisations to hand out the surveys. Together with the team leader the grid of the team was filled in. With the organisations of the online surveys the grid was collected by e-‐mail.
3.2 Measurements
because the survey is only conducted among Dutch teams. All questions elaborated on below can be found in appendix 1.
This research attempts to conduct if an organisational change activates a faultline in teams based on age and resistance and if this activated faultline does have effect on team conflict. To measure this, first an objective faultline score has to be obtained. This is measured by the faultline construct of Thatcher et al. (2003). Thathcher et al. (2003) developed the following algorithm (figure 2) that measures the faultline strength. This algorithm presents a score between zero and one where 1 implicates a really strong faultline.
In this algorithm the dispositional change questions of Oreg (2003) are used. This in combination with the age of team members. By combining these two variables a score of the possibility of forming subgroups based on those two attributes is obtained. Which is the start of the conceptual model: objective faultline.
Before elaborating on the activated faultline a moderating variable is presented in the conceptual model: the scope of the organisational change. The scope of the organisational change is measured by the scales of Balogun and Haily (2008). The definition of the scope of an organizational change is: ‘The degree that an organisational change affects the organisation’ (Balogun and Haily, 2008). The scope of the organisational change is an moderating variable, because it effects the relation between the objective and activated faultline as assumed in the theory.
The activated faultlines are measured with the scales of Jehn and Bezrukova (2010) and Rupert and Jehn (2008). These scales are about the perception team members have of their team contains a faultline(s). First there are four questions about faultlines in general and than two about faultlines based on resistance and one about a faultline based on age.
Finally team conflict have to be measured. This is done with the items provided by Jehn & Mannix (2001). Team conflict is divided into three types as mentioned in the theory: task, process and emotional conflict. All different types have three different items to measure the concept.
3.3 Data analysis method
Multiple steps have to be taken to test the hypotheses of this study. The data obtained with the questionnaire is analysed with the statistical software program: SPSS. First the descriptive statistics are calculated to give some basic information about the dataset. Second, a factor analyses is conducted to indicate if items load on the same variable. Based on the factor analysis scales are computed and the reliability of these scale is tested with the Chronbach Alpha score. Third, a correlation analyses is conducted to analyse if there are relationships among the different variables. Finally the hypotheses are tested using a linear regression and multiple regression analysis.
3.4 Data analysis
In total 71 teams, representing 433 respondents, were selected for this study. All teams with more than 30% missing data were deleted. Next to that, the teams where no objective faultline score could be calculated on the base of resistance were also deleted. After this selection the dataset consisted of 22 teams with a total of 164 team members. So on average a team has approximately seven members. The average age is 41,98 years with a standard deviation of 11,58 and 52,1% is man and 47,9% is woman. Most respondents have a Dutch nationality, respectively 95,7% and in table 1 below the different sectors are presented were the teams are operating in.
Sector Teams Percent
Education 5 22,7% Government 4 18,2% Healthcare 4 18,2% Food 2 13,6% Industry 2 9,1% Retail 1 4,6% Missing data 3 13,6% Total 22 100%
Table 1: Sectors Teams Are Operating In
3.5 Factor analysis
To obtain insight in the dataset, a factor analysis is conducted. The factor analysis analyses which items measure the same concept. In this way it is measured if the items are valid and can be combined into one scale. Items can be combined if they load higher than 0,30 on the same component. If items score a double loading than the difference between the loadings should be more than 0,20. In appendix 2 the scores of the first factor analysis are shown of all the items with ‘eigenvalue’ set on 1. This factor analysis is showing that the scales are almost all measuring the same concept as literature encountered. Only item 4 of the dispositional resistance scale loads on a totally different component as the other 3 items of that scale. This item is therefore deleted on account of statistical reasons, because this item loaded 0,01 on the component where the other three items score all higher than 0,75. Because the 0,01 is not even near 0,30 and the item scores really high (0,89) on a totally new component, this item is deleted. Furthermore the ninth item measuring conflict has a double loading on different components. The question in the survey for this item is: How often do you disagree about resource allocation in your work group? It is assumed that this question is interpreted in different ways. Team members could have had different knowledge about the allocation of resources. For example a team member close to the team leader knows more of this topic than other team members. Or the hierarchy within a team is distributed unequally and some members have nothing to say about the allocation and interpreted this question in a way that there never is disagreement. But other members higher in the hierarchy can have tough discussions about the allocation. Based upon these assumptions this item is deleted and a new factor analysis is presented on the next page in table 2.
Table 2: Factor analysis
This factor analysis has a KMO score of 0,82 and is significant at 0,00 on the Barlett’s Test of Sphericity. Based on the results of the factor analysis items can be combined into scale to reduce the data and make it suitable to test the hypotheses. First the scale of dispositional resistance can be made by combining the three items. This scale has a Cronbach Alpha of 0,78. To be reliable a scale has to have a Cronbach Alpha higher than 0,70, so this scale is reliable and can be used to measure the objective faultline score. The next variable is the scope of the organizational change, these six items are combined into one scale and are reliable according to the Cronbach Alpha with a score of 0,91. To combine items to measure the activated faultline the general items (FAULT1-‐4) about faultline perception and the items about the perception of faultlines based on age (FAULT_AGE) and resistance (FAULT_RESIS1-‐2) are taken together to obtain a good measure of activated faultlines based on age and resistance. There are some double loadings on the items of faultlines based on
Rotated Component Matrix
resistance and age, but the differences are more than 0,20 so the items can be combined into one scale. Also this scale is reliable approved by the Cronbach Alpha with a score of 0,83.
According to the theory conflict has three different types. In the first factor analysis this is not shown. This is in contrast with the theory, because according to Jehn & Mannix (2001) there are three types of conflict: task, relationship and process. This is way a second factor analysis is conducted and forced into three components. This factor analysis has a KMO score of 0,87 and is significant at 0,00 on the Barlett’s Test of Sphericity. The results are shown in table 3.
Rotated Component Matrix
Component 1 2 3 CONFLICT1 ,854 ,242 ,273 CONFLICT2 ,781 ,411 ,256 CONFLICT3 ,817 ,329 ,321 CONFLICT4 ,479 ,725 ,193 CONFLICT5 ,258 ,893 ,235 CONFLICT6 ,312 ,806 ,389 CONFLICT7 ,282 ,278 ,888 CONFLICT8 ,314 ,271 ,879
Table 3: Second factor analysis
Before combining the items into scales, the reliability of these scales are tested. Relational conflict is measured with the first three items and has an Cronbach Alpha of 0,92, which means it is a very reliable scale. Task conflict is measured with the next 3 items and is reliable as well with an Alpha of 0,92 and process conflict is measured with the last two items and has an Alpha of 0,94.
Now that is proven that the scales are validated and reliable, the items are combined into scales. The combined items are presented in appendix 1. After combining the scales the data is aggregated on team level. Until now the analysis was on individual level, but the measurement of Thatcher (2003) is used to obtain the objective faultline (faultline strength) of a team. This measurement is on team level and van Aken et al. (2012) argue, preferably all measurements have to be aligned at the same level.
4. RESULTS.
Before testing the hypotheses of this study the relationship between the variables are analyzed with a correlation test.
4.1 Correlation test
A Pearson correlation test is conducted, the results are presented in table 4 below:
Table 4: Pearson Correlation Test
The correlation test shows that there is a relationship between the perceived faultline and relational conflict. This correlation is between the perception of a faultline based on resistance and age and relationship conflict r = 0,490 p<0,021. The positive relations between these two variables indicates that when a team perceives more of the faultline there is a higher degree of relational conflict. Further on task conflict and relation conflict do correlate. This is the strongest correlation with r = 0,636, p<0,001. There is a positive relation between the two scales, which means that if one conflict arises, the other type of conflict also is perceived. The same accounts for process conflict in relation to relationship conflict (r = 0,608, p<0,003) and task conflict (r = 0,439, p<0,041).
One unexpected correlation did arise between process conflict and the scope of the organisational change (r = 0,523, p<0,021). This relation can be clarified in a way that an organisational change can mean that tasks or responsibilities alter within a team. If this changes process conflict can arise.
Mean SD N 1 2 3 4 5 6 1. Objective_Faultline 0,763 0,021 22 -‐ -‐0,09 0,28 0,08 0,33 0,14 2. Scope_Change 4,045 1,355 22 -‐ -‐ 0,17 0,33 0,36 ,523* 3. Activated_Faultline 3,180 0,769 22 -‐ -‐ -‐ ,490* 0,16 0,28 4. Relational_Conflict 2,418 0,748 22 -‐ -‐ -‐ -‐ ,636** ,608** 5. Task_Conflict 2,962 0,887 22 -‐ -‐ -‐ -‐ -‐ ,439* 6. Process_Conflict 2,942 1,154 22 -‐ -‐ -‐ -‐ -‐ -‐
