The Cultural Influence on
Faultlines
The impact of the cultural context on perceived team faultlines
within an organizational change context
February 14
th2014
University of Groningen
Faculty of Economics and Business
MSc Business Administration: Change Management
2
Abstract
3
Content
Abstract ... 2 Content ... 3 Introduction ... 4 Relevance ... 7 Literature ... 8 Team Diversity ... 8 Faultlines ... 9 Context Cultures ... 11 Organizational Change ... 14 Hypothesis Development... 15 Conceptual Model... 18 Methods ... 18 Research Design ... 18 Data Collection ... 19Sample & Procedure ... 19
Measures ... 21
Data Analysis ... 23
Results ... 28
Correlations ... 28
Hypotheses Testing ... 29
Conclusion & Discussion ... 31
Conclusion ... 31
Discussion ... 32
Bibliography ... 36
4
Introduction
''We inhabit a universe that is characterized by diversity.''
Desmond Tutu
Archbishop Emeritus of Cape Town & 1984 Nobel Peace Prize winner
Following this inspiring statement, this exploratory qualitative research will set about the major differences between the two predominant types of culture as described by E.T. Hall (1976). Hall describes cultures on a context continuum, ranging from High to Low. Following this typology, Low context cultures are explicit and direct when it comes to the causes of events; the reason for certain actions and communication. On the other side: High context cultures can be described as a culture in which events; actions and communication are highly implicit; thus most of the explanatory value lies within the context. Therefore, both cultural types have significant different approaches to their environment; including organizations and team work. As such, this research will emphasize on an important practical and real-life issue. Cultural diversity is becoming increasingly important in today's global environment with many firms not only operating internationally but additionally globalization, and related workforce migration, also leading to a more multicultural workforce even for firms who operate domestically. In the Netherlands for example, CBS data (Appendix A) shows that the workforce migration more than tripled between 1995 and 2007. OECD (Appendix B) data indicates that currently the Dutch population now even consists of more than 10% foreign born, which is even below the average of OECD-nations. These figures show that multiculturalism of the workforce is increasing. This will most likely have a considerable impact on most firms. Therefore, this research, aims to identify what the differences and related implications between these two types of culture are to enhance the understanding of both firms and researchers dealing with multiculturalism. As such, this paper definitely has a practical purpose, although much research has been conducted on multiculturalism and its impact on organizations, the specific research focus on organizational teams from and their dynamics makes it worthwhile.
5
affects group-level outcomes. The focus on the aggregation of atomistic individual characteristics to the group level has led many diversity researchers to neglect the joint effects of multiple types of diversity on group interactions'' (Thatcher & Patel 2012:6). It is this neglection, argued by Thatcher & Patel, which lead Lau & Murnighan in 1998 to develop a new stream in diversity literature, their so-called faultline theory. Faultlines are ‘'hypothetical
dividing lines that may split a group into subgroups based on one or more attributes’’ (Lau & Murnighan 1998:5) And thus, ''faultline theory offers insights to understanding the cumulative effects of group member attributes on group outcomes'' (Thatcher & Patel 2012:6). Naturally, however, there are various attributes which could lead to a faultline. Generally, research has identified three main types of attributes that could lead to faultlines. First there are so-called demographic or social category attributes, examples of such attributes are age; gender and race. Secondly, there are non-demographic attributes; these generally focus on the personality or preferences of an individual (e.g. Gratton, Voigt & Erickson 2007; Molleman 2005). Thirdly, there are informational attributes, these are attributes such as education; experience; position etc. (e.g. Bezrukova, Jehn, Zanutto & Thatcher, 2009; Molleman 2005; Zimmermann, 2011) In this research, we aim to identify whether high and low context cultures have a different perception of faultlines on the basis of all three attributes. However, we will additionally focus on faultlines that could be perceived on the basis of organizational change. Gover and Duxbury (2012) identified that organizational change was a common reasons for faultlines to emerge. Following the significant differences in communication; actions; relationships; environmental perception and additionally their ideas concerning change between both cultural types these can be expected to perceive team dynamics differently. As such this paper sets out to identify whether a different perception of faultlines can be seen when comparing both cultural types.
6
processes and its related impact on organizational performance. By enhancing the knowledge on cultural diversity and especially its impact on organizations and teams and thus enable us to further optimize their potential. Since diverse teams are often found to be more effective (Caudron 1994; Sullivan 1974) and additionally allow for more opportunity creation and creativity. (Martins & Milleken 1996; Watson, Kumar & Michaelsen, 1993) Even though a wide variety of researchers has acknowledged this and even though a wide debate on cross-cultural management and its impact on teams exist, the cross-cultural influence on team dynamics and faultlines, especially within in organizational change context, is hardly investigated.
Therefore, although a considerable amount of literature on faultlines exist, the cultural context in which those faultlines are perceived is often neglected; a context which is especially relevant in todays’ globalizing environment. The diversity that national cultures offer us should not be neglected. Therefore this paper will search whether there are difference between high and low context cultures in the faultlines they perceive. Additionally, it will aim to identify what the implications are of faultline differences between these two cultural types and as such provide implications for both theorist and practitioners dealing with multicultural organizations. The research will compare the perception of individual team members with a high context culture with individuals low context culture. All teams are located within the Netherlands to ensure institutional similarity. Given the fact that usage of teams among organizations is increasing (Bettenhausen, 1991; Cooke Rosen & Salas 2008; Hollenbeck, Ilgen, Major & Sego, 1993) and due to a more globalized and international workforce (Williams & O’Reilly 1998) this research should not only provide a relevant extension to the current debate on subgroups and their establishment as well as an extension to the managerial issues regarding international; multicultural teams.
This exploratory research aims to enhance the understanding of the influence of national culture on faultline theory.
To live up to this aim, this paper will seek to find an answer to the following research question:
7
Relevance
Academic Relevance. This research aims to identify whether there are considerable
differences between people from both high and low context culture with regard to the perception of objective faultlines. By doing so, this paper aims to make a connection between the cultural context and their impact on the theories regarding intra team faultlines and subgroups. As such, the paper aims to extend the theories on faultlines by providing the influence of culture on these faultlines. So far, research on faultline theory has identified that cultural dissimilarities can result in a faultline. (Jehn & Bezrukova 2010) Thus, cultural difference being a faultline basis in itself. Yet, considering the fact that these two cultural types are fundamentally different in their respective underlying perceptions regarding their environment there might be an underlying reason for this seeming division on the basis of culture. As a result, it seems wise to examine the cultural context prior to identifying faultlines. Cultural differences might not be faultlines themselves but the respective different perceptions on organizations and team dynamics between different types of cultures could be the underlying reason for a seeming division on the basis of culture. As such, this paper will try to identify whether this assumption holds, thus whether these cultural types perceive faultlines differently and on which attributes exactly. This should prove especially relevant for future researchers when examining the impact or presence of faultlines in multicultural organizations or organizations operating in a variety of different cultures.
Practical Relevance. The practical relevance is especially aimed at managers who have to
8
Literature
Team Diversity
In this paper organizational teams are often referred to, teams can be defined as: a group whose membership and task are formally recognized by the organization (Cohen & Bailey, 1997; Kozlowski & Bell, 2003). Teams can be crucial to organizational ability to tackle a variety of different issues, ranging from merely economical issues to technological or social issues. However, being jointly involved in such a team does not necessarily involve a collective team approach. Teams operate in a dynamic way, involving many social actions. Additionally, teams often consist of diverse populations. We define diversity in line with (Mannix & Neale, 2005:3) as ''variation based on any attribute people use to tell themselves that another person is different.'' Since every individual team member will have his/her individual background and related perceptions and opinions regarding the teams functioning; focus and processes; highly diverse teams can face more barriers to social interaction. (O'Reilly, Caldwell & Barnett, 1989); resulting in a decrease in team cohesion. Generally, diversity literature focuses on the variation between individual team member characteristics and its related effect on teams. (Thatcher & Patel, 2012) Now although this point of view most certainly has its benefits, by focusing on a meso-level, thus the focusing on the communalities that arise within a team; the joint effects of intra-team dynamics allow us to identify what enhances or hampers team performance. Or as put by Thatcher and Patel (2012:6):''it is likely that the multidimensional nature of individual characteristics has cumulative meso-level interactions that influence group outcomes.'' Thus, focusing on a meso-level provides the opportunity to examine the influences that go beyond the individual level and allows us to identify the wider impact of diversity onto team dynamics.
9
Figure 1: Kelly (2009) on the influence of diversity on performance.
It is this negative route which especially needs consideration. If the divisional impact can be reduced, performance may be enhanced. However, teams do not necessarily have a collective approach. Commonly within teams certain subgroups are formed. According to Carton & Cummings (2012:2) a subgroup is a ''subset of members of the same team that is unique when compared to that of other members.'' It is this divisional character of organizational diversity that leads us to question the underlying causes to team malfunctioning. Because if the ''value in diversity'' hypothesis holds; diversity should enhance organizational performance. Therefore it is important to enhance our understanding of the divisional forces within an organization since such, negative, dynamics carry the high potential for discrimination and suboptimal utilization of group member’s diverse knowledge, skills, and abilities (Tesluk & Mathieu, 1999). This suboptimal utilization of organizational resources is therefore a highly interesting topic of consideration, especial to profit-oriented organizations. To that end, this paper focuses on a literature stream which specifically focuses on this divisional character and its causes within diverse organizations to ultimately optimize organizational diversity: faultline theory.
Faultlines
10
(Byrne, 1971). Both social identity and self-categorization, as indicated by their names, explain how individual team members perceive and classify themselves as well as their fellow members. Additionally, the similarity-attraction framework by Byrne (1971) explains that individuals tend to identify and aligns themselves with team members who possess similar perceptions and characteristics. Although these theories aid in creating a better understanding of subgroup formations, faultline theory differs from these literature streams in the sense that it advocates the meso-level approach of investigating the cumulative effects of individuals on group outcomes (Thatcher & Patel, 2012)
Lau & Murnighan were the first to develop this relatively new stream in diversity literature. According to their definition: ‘’Faultlines are hypothetical dividing lines that may split a group into subgroups based on one or more attributes.” (Lau & Murnighan, 1998:5) One of the main strengths of faultline theory lies in its ability to consider and identify behavioral asymmetries within organizational teams. Faultline theory then identifies which of these asymmetries (can) cause subgroup commencement. These hypothetical lines which can result into separate subgroups are named faultlines. Furthermore it is important to consider that these subgroups will be homogeneous when activated by a similar faultline (Mannix & Neale, 2005) However, as multiple faultlines can be present within an organizational team, this can result in a variety of subgroups and more importantly result in team members being in multiple subgroups, depending on which faultline perspective one examines them. Now it is important to understand that faultlines can be merely hypothetical. Faultlines do not necessarily always have to emerge in practice. Lau en Murnighan divided faultlines into two main types: dormant faultlines and active faultlines.
Dormant faultlines are purely hypothetical; these are the possible dividing lines that can result in subgroup formation. Dormant faultlines may be recognized from an external point of view, yet within the team they have not caused any subgroup formation yet. However, in case of a trigger, an event or situation that activates a dormant faultline, thus making it an active faultline, (Rink & Jehn, 2010) these faultlines can become activated.
11
a members' identity to relate more with a certain subgroup than with the team as a whole. (Lau & Murnighan, 1998) It is this different sense of belonging which then can result in a variety of team dynamics, with both positive and negative effects. Yet, in any case, it poses several issues onto the managerial team, issues that were going unnoticed on forehand. It is important to consider that this research focuses on the different perceptions regarding faultlines between both cultural types. As such, this exploratory research focuses on the opinion of teams. Therefore, wat are commonly considered ''activated faultlines'' will be often referred to as faultline perception.
Literature has derived a variety of attributes on which faultlines are perceived. As previously described three main categories can be identified (Thatcher & Patel 2012) 1) demographic or social category attributes, examples of such attributes are age; gender and race. 2) non-demographic attributes, these generally focus on the personality or preferences of an individual (e.g. Gratton et al. 2007; Molleman 2005). 3) informational attributes, these are attributes such as education; experience; position etc. (e.g. Bezrukova et al. 2009; Molleman 2005; Zimmermann, 2011) However, it is has to be understood that within teams, faultlines can be present on all three of these attributes and that all attributes can result in subgroup formation. Furthermore it is important to understand that each faultline can result in a different subgroup and that as a result, depending on which faultline perspective is taken, team members can belong to a variety of subgroups. However, it depends on the team characteristics; individual perceptions and trigger event which category and which faultlines become activated.
In this research we aim to identify whether the cultural background of individuals matters on their perception of subgroup emergence. Thus, we examine whether individuals with different cultural background perceive different activated faultlines in regard to the previously described typology. However, to that end we first need to describe the cultural continuum among we will differentiate the investigated national cultures.
Context Cultures
12
it is highly plausible they influence the individual perspective on teams; teamwork and team dynamics as well. Therefore, following Hall, individuals ranging from a variety of cultures are likely to perceive different faultlines and possibly separate subgroups due to their cultural lens. Therefore, this paper will investigate what the differences are among cultures with regard to their faultline perception. This is especially important considering the potential that a cultural diverse team holds. According to Richard (2000) when operated in a proper manner, cultural diversity does contribute to competitive advantage creation. To that end we will examine these perceptual differences among the high-low context culture continuum as advocated by Hall (1976). The continuum Hall describes the different thought patterns between the two ultras of his cultural typology. These thought patterns center on the individuals perceptions of its environment and especially its social interactions. Thus the context continuum as described by Hall offers a suited match with faultline theory, as activated faultlines differ based on the perception of members and the perspective that is chosen.
13
implications of an event are to be understood by all individuals involved without explicitly mentioning this. In fact, it is considered impolite to be too specific since it limits an individual’s ability to derive his/her own personal meaning out of the context (Hall, 1976) However, this only holds for individual behavior. Being highly collective, in a high-context setting, an individual tends to be ignored, but as a person’s behavior comes into conflict with group values, it will be engaged. (Choe, 2002) Therefore, cultures which tend to be more high-context will have different opinions on social actions than do lower context cultures. As a result they can be expected to have considerable differences in faultline perception.
Alternatively Low-context cultures are the theoretical opposites of high-context cultures. 'In low-context cultures, people rely more on the explicit verbal content of messages.'' (Choe, 2002:8) With a rigid emphasis on results and outcomes this type of culture highly values individualism; effectiveness; target orientation and a ''can-do'' mentality. (Choe, 2002; Würtz, 2005) ''Low-context cultures prefer directness, specificity, frankness in stating demands, confrontation, and open self-disclosure.'' (Choe, 2002:8) Being implicit and indirect is therefore commonly considered unnecessary; inefficient or complex for individuals with low context culture. The surroundings are, in contrast to to high context cultures, recognized but only as a factor to benefit from or alternatively to fight or change. The result orientation is key in low-context cultures. As such the individual should adapt to its aim and adapts its environment accordingly, even if this would imply a change of personal or relational environment. As such, personal ties; hierarchical levels and power distances are less valued among these cultures.
In short, the main differences between High and Low context cultures can be summarized as in table 1 as found in Ting-Toomey (1985). Following this summarizing table, it becomes even clearer how these two types of culture are fundamentally different and clearly opposites.
Context Cultures
High Context Low Context
Homogeneity Encouraging Heterogeneneity Encouraging
Group Oriented Individual Oriented
Deviance discouraged Encourages variation
People Oriented Active and Solution Oriented
Non-confrontational Conflict encouraged
Humanistic Procedural
Risk-averse Welcome Risk
14
Now although Hall (1976) specifically stresses that it is a high - low context continuum and that each respective national cultures has some degree of high context aspects and some low context aspects, literature fortunately provides us with a graphical representation on how cultures can generally be ranked among the continuum. Following Hall & Hall (1990) and Usunier (1993) we are able to comprise the following graphical representation that can classify nationalities among the context continuum:
Figure 2. National cultures on the high-low context continuum. Adapted from: Hall, E.T. & Hall, M. R. (1990). & Usunier, J.C. (1993)
Organizational Change
Kochan et al (2003) found that, the predominant psychological concepts behind diversity initiatives revolve around two main principles. The first centers on recognizing, understanding, and overcoming in-group-out-group dynamics, while the second aims to improve attribution and the resulting attitudinal-behavioral linkages of the majority group. According to the authors, workplace related in-group-out-group dynamics occur whenever the cohesion of an existing group is challenged by the arrival of some outside influence. This outside influence, or trigger, as Lau & Murnighan (1998) describe it plays an important role in the framing of the faultline activation. As such it is important to describe it and especially its likely effects on team dynamics.
15
approach which considers organizational change involves a series of predictable, reducible steps that can be planned and managed. (Collins, 1998 in Greatz & Smith, 2010) This theory of thought approaches changes a quantifiable; manageable and with clear outputs and targets. However, many authors have argued the traditional approach is too uni-dimensional and therefore too limited. They argue that alternatively organizational change can also be seen as natural process. Something that arises spontaneously within organizations and develops along a more natural path. According to Greatz & Smith (2010), this other approach includes the human factor, which is often forgotten in the traditional approach. They argue that individuals are automatons rather than active agents in the change process.
Thus, this research will highlight the differences between high-low context cultures and their faultline perception within an organizational chances setting. However, it is additionally interesting to examine whether an organizational change could influence team dynamics. Thus, whether faultlines could arise on the basis of organizational change. Especially considering the fundamental differences between high and low context cultures and their perception on change. According to Pfeiffer (1993) high context cultures seem to consider change as slow. Things are rooted in the past, slow to change, and stable. Whereas low context cultures perceive change as fast. One can make change and see immediate results.
Being so fundamentally different in their perceptions regarding organizational change, this should be an ideal context to identify whether high- and low-context cultures are indeed different in their perception regarding team dynamics.
Hypothesis Development
Now when we reexamine the previous descriptive section it is important to compare both ends of the context continuum and especially question what their respective influences are on the three types of faultline attributes and possible change related faultlines we previously described.
16
''high-context cultures tend toward indirect, ambiguous, cautious, non-confrontational, and subtle ways of working through communication'', (Hall, 1976; Ting-Toomey, 1985; Choe, 2002) they have a cautious and non-confrontational nature. Furthermore, with their emphasis on the interpretation of the environment, ''the messages which are explicitly spoken can be elliptical'' (Choe, 2002:8) Therefore, high context cultures can be expected to derive perceive faultlines on more indirect bases. With low context cultures being direct, specific, confrontational, and open (Choe, 2002), it can be argued that members with a lower context culture are more likely to perceive subgroups based on demographic attributes which are directly perceivable.
Hypothesis 1: Individuals from a lower context culture are expected to perceive more faultlines on the basis of demographic attributes than individuals from a higher context culture.
Non-demographical attributes are according to Thatcher and Patel (2012) oriented at the individual preferences and individual personality. However, as Hall en Würtz argued: ''collectivistic cultures prioritize group welfare over the goals of the individual.'' (Hall, 1976; Würtz, 2005) Hight context cultures are unlikely to place significant focus onto the individual needs and therefore, high context cultures are most likely to focus on a wider collectivist, or team, perspective. On the contrary lower context are highly individual (Ting-Toomey, 1985; Choe, 2002; Würtz, 2005) and individual needs and related development are essential in the low context perspective. Thus since low context cultures place considerable more attention and thought into individual needs and since the individual perspective and preferences are part of low context team dynamics, it is likely to assume that individuals from a lower context culture tend perceive perceive subgroups based on non-demographical attributes more.
Hypothesis 2: Individuals from a lower context culture are expected to perceive more faultlines on the basis of non-demographical attributes than individuals from a higher context culture.
17
egalitarianism''. (Augsburger, 1992:32) Therefore, informational attributes such as experience; education etc. are unlikely to be considered relevant within a low context culture; whereas they are essential in high context cultures. Since individuals from a high context culture are likely to place considerable more attention on informational attributes, it can be hypothesized that high context cultures are more likely to perceive faultlines on the basis of informational attributes.
Hypothesis 3: Individuals from a higher context culture are expected to perceive more faultlines on the basis of informational attributes than individuals from a lower context culture.
Finally, when focusing on faultlines on the basis of change, we’ve already established that a considerable difference can be expected in perception regarding change between the two types of culture High context cultures considers changes a slower more continuous process which follows a more natural evaluation and is hard to measure. On the other hand, Low context cultures can be argued to be more traditional in their viewpoint, considering change as makeable, fast and measurable. Considering the fact that the individuals we investigated were located in the Netherlands, undoubtedly a low context culture following figure 2, teams were selected on the basis of a definable and measurable change process, we can assume that among the respondents from a high context considerable more discomfort regarding these measurable and definable change processes exists. As such we can hypothesize that, since it has a larger impact on people from high context cultures, individuals from a high context culture are more likely to perceive subgroups according to the attitude towards change. Especially, since a negative opinion facilitates more concrete thought on the discomforting issues. (Kuvaas & Selart, 2004; Tversky & Kahneman, 1986)
18
Conceptual Model
Summarizing the previously described theory, the conceptual model for this paper is similar to this:
Figure 3: Conceptual Model
Methods
Research Design
19
answers of individual respondents and so that individuals from a high context culture would be less influenced by the low context environment the Netherlands provided. Although teams were asked to participate in this research, the focus will be on the individual level. Thus focusing on the individual responses of the respondent since the aim of this research is to identify whether individuals from separate cultural perspectives perceive faultlines differently.
Data Collection
The teams that provided the data for this research were each contacted by means of visits to identify whether they were willing to contribute. Primarily these visits entailed a conversation regarding the purpose of the questionnaire and to assess the suitability of the teams for this research. Teams were suitable for further research when they met the following criteria. Primarily they should collaborate in a team. The first criterion is that employees must collaborate in team composition. According to Hackman and Wageman (2005), a unit of two or more individuals who interact interdependently of achieve a common objective can be called a team. Therefore, teams are considered teams not when they are labelled as such by the organizations but when a group of collaborating individuals meets this definition. Secondly, teams should consist of at least 3 and maximum 15 persons to correctly depict faultlines. According to Hogg & Cooper (2003) individuals prefer social interaction within small groups, ideally containing two or three people. Hardly are their individuals who prefer interaction with groups above five or six people. Therefore, ideally, teams consisted of 3 till 7 persons, because larger teams should, following Hogg & Cooper, result in faultlines due to the team size in itself. Thirdly, participating teams should be in a state of organizational change, or should have experienced an organizational change in the past year to ensure they have clear and vivid memories and opinions regarding the change and its influence on the teams' dynamics.
Sample & Procedure
Procedure: To provide a reliable and valid answer to the presented hypotheses the surveys
20
Surveys were handed out after an initial explanatory sessions with respondents. The aim of these sessions was to explain the underlying aim of the research; to establish which groups qualified as a team and to ensure that full participation was achieved. Afterwards, all respondents individually filled out the presented questions. Ideally surveys were held on hardcopy reprints however, in the teams specifically preferred a digital version, digital copies were provided for. Each team was given a specific code just as every member to be able to comprehend the source of the answered questionnaires. In case of any missing team members a second survey moment was planned to ensure that the missing team member was included into the question list. The team members were asked to answer the questions at times that they were free to do so, ensuring that they could have their full attention on the questionnaire and preventing that answers would be answered incorrectly due to time pressure. These measures were instigated to enhance an individual; reliable and valid answer to the posed questions.
Sources: Following this gathering procedure, a total of 72 teams and in total 426 respondents
were collected to be used for further analysis. Yet, many of these teams consisted of more than 7 members. As Hogg & Cooper (2003) mentioned, teams consisting of more than 7 members are less interesting for this research since due to the number of members inevitably divisions will arise. As a result, all teams which had more than 7 respondents were discarded. Similarly, there were a variety of teams which had only 2 respondents. These teams were also deleted because 2 respondents are unable to provide a conclusive answer on the perception of faultlines within a team. With a team comprising at minimum 3 members, 2 respondents never equal the minimum response rate, which was set at 75% to ensure a valid understanding of the faultlines. Finally, additional teams which lacked the 75% response rate benchmark were also deleted.
21
However, after the deduction teams on the bases of response rates and size, further steps had to be taken that the data provided by the remaining respondents ensured reliable and valid results.
Measures
To test the hypothesis certain questions needed to be answered in the questionnaire. These questions and related measurement scales were derived from existing literature. To enhance the validity of replies multiple questions were used to construct a single variable. Each specific question was a specific item to enhance clarity. Table 4 summarizes the questions and related items (Appendix C) that were assigned to answer the variables in the conceptual model. However, an additional variable had to be included, which measured to what extent subgroups were perceived within teams. Whilst most questions in the questionnaire we closed questions, being answered on a 1-7 likert scale, the questionnaire included open questions to establish the origin and nationality of a respondent. For these open questions, respondents were encouraged to provide all relevant information. In case of multiple replies, multiple answers were taken into account. Furthermore, teams were encouraged to provide additional data regarding team performance and results to enhance insights into the provided answers.
Scales Measurement Technique
Corresponding Items Source
Cultural Context 2 question in questionnaire
Country of Birth; Nationality
Subgroup Perception
4 questions in questionnaire
FAU: 1;2;3;4 Jehn & Bezrukova
(2010) Demographic Attributes 4 questions in questionnaire FAU: 5;6;8 Non-Demographic Attributes 3 questions in questionnaire FAU: 14;15;16 Informational Attributes 6 questions in questionnaire FAU: 9;10;11;12;13;17 Perception towards Change 2 questions in questionnaire FAU: 21;22 (Oreg 2003)
Table 4. Variable Measurement Table1
Cultural Background: To analyze whether a culture represents a certain cultural type 2 open
questions were asked. The respondents were asked to indicate their country of birth and nationality. Furthermore, respondents were additionally asked to provide multiple nationalities if applicable. Finally, to ensure full understanding into the cultural background of
1
22
the respondents, they were asked to provide their descend additionally to the questionnaire since both country of birth and nationality does not necessarily represent the cultural background of a respondent. Although neither of these two questions fully explains the cultural type of an individual respondent, following figure 2 we were able to derive the common cultural typology for that nationality or descend.
Faultline Perceptions: To establish the faultline perception, the questions used to establish on
what bases faultlines are perceived describe the 12 most common faultline bases. From these 12 faultline bases, culture was discarded as a basis in this report since teams were of similar cultural descend; the ambiguity that could arise from different uses of the word culture. All of these questions were measured on a 1-7 likert scale, ranging from totally no division within a team (1) to a very strong division (7).
For the faultlines based on informational attributes the following items were measured in the questionnaire: Division on: education; expertise; experience; status; employment duration and fulltime/part-time employment. These were coded as FAU 9-13 and FAU 17. (Appendix C) These items were considered as informational attributes following Bezrukova et al. (2009) Molleman (2005) and Zimmermann, (2011)
Regarding the faultlines based on non-demographic attributes 3 questions were selected, which aimed at identifying faultlines on the basis of: personality; private situation and work style of the respondents. These items correspond with codes FAU 14- 17. These items are considered as non-demographic since following e.g. Gratton et al. (2007) & Molleman (2005) ''these generally focus on the personality or preferences of an individual.''
Thirdly, the demographic attributes which could be a faultline basis were identified following three related questions in the questionnaire. These questions aimed to identify the degree to which gender; age and nationality were considered as faultlines. These items were considered demographic following Thatcher & Patel (2012) These items correspond with codes FAU 5;6;8
23
Finally, to control whether faultlines were indeed present within the investigated teams they were asked four questions to establish the presence of subgroups. These questions were coded as FAU 1-4 and were derived on the basis of Jehn & Bezrukova. (2010)
Data Analysis
Data Processing: Firstly, most questions were held on a 1-7 likert scale. Yet, in case a
question remained unanswered it was coded as ''999''. To check whether no other numbers were given, the scores were tested on minimums and maximums. Several other answers could be identified following this assessment and all these were recoded as ''999''. Then all respondents were checked on their response rate to the investigated questions. If the respondents failed to answer at least 50% of all the questions regarding faultlines, these respondents were discarded.
For the open questions regarding the cultural background, all nationalities were given a single code, resulting in 14 countries of birth and 5 nationalities, additionally a 15th country of birth was added since some respondents were of a descent indicating a different cultural type even though they were born within the Netherlands. Whenever an answer was not provided by a respondent, the code ''999'' was used. Also these codes were checked for other values, fortunately none were found. To ensure that a valid assumption regarding the cultural background of a respondent could be made, respondents who failed to answer both questions were discarded.
Following these measures, the data set still consisted of 186 respondents divided among 47 teams.
Data Recoding: However, to be able to measure the differences between high and low context
cultures these countries of birth and nationalities had to be recoded into either high or low. To do that, we used the continuum as sketched following Hall & Hall (1990) and Usunier (1993) and classified the nations and nationalities into either low (1) or high (2). For those respondents which had only answered one of these two questions, remaining ''999'' were recoded as 1,5. Thus ensuring, that whenever a new variable concerning the cultural type had to be computed, low context cultures would on average still be, since (1 +1,5)/2 is rounded to 1, and similarly (2+1,5) /2 would be rounded as a 2, being a high context culture.
24
method was chosen. Thus, the missing values were given the mean of the corresponding variable to ensure that the differences between the cultural types would not be reduced. This implies that respondents from a high context culture were given the mean of the high context respondents on that variable and vice versa for low context respondents. As a result, the any differences between the two cultural types that were investigated remained.
Validity of measurements: To ensure further validity of the data a Principal Component
25 Rotated Component Matrix
Component
Item 1 2 3
During work subgroups are formed ,891
During work several subsets of people ,896
Subgroups in meetings ,786
During work subgroups in entire team ,880
Gender ,692 Age ,759 Nationality ,662 Education ,658 Experience ,570 Status ,632 Employment Duration ,607 Work Style ,676 Private Situation ,552
Comfort vs Discomfort towards Change ,881
Positive vs Negative towards Change ,867
Eigenvalues 3,91 3,44 1,84
Percentage of Variance Explained 26,03 22,91 12,4
Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. Table 6. Rotated Component Analysis.
26
Data Reliability. This analysis identifies how reliable the scales are and checks for each item
whether the reliability is increased whenever an item is deleted. Generally, literature agrees that if a scale has a Cronbach alpha above 0,6 the scale can be considered reliable. Furthermore, SPSS indicates for each scale that if the deduction of a variable indicates a considerable improvement in reliability, this should be considered. The results indicated that when including all the scale it was considerably reliable with a Cronbach alpha of 0,912. Furthermore, the assessment of the individual scales indicated that the exclusion of a single variable would not significantly enhance the reliability.
So therefore, following the principal component analysis and reliability analysis, it was chosen to include all items that remained after the principal component analysis and use these for further testing of the hypotheses. Although the personal attributes according to the factor analysis represent a single scale, following the theoretical foundations of Thatcher & Patel (2012), in addition to the expected differences among the cultural types among the three types of personal attributes, it was chosen to compute the three different variables: Demographic Attributes; Non-Demographic Attributes and Informational Attributes. Furthermore, although the items concerning Team Division loaded on a separate factor, to ensure that all perceptions on faultlines indeed matched with the subdivision within teams, each variable was merged with the Subgroup Perception items. Thus, the Subgroup Perception items were incorporated within the variables Demographic Attributes; Non-Demographic Attributes; Informational Attributes and Perception towards change. Concerning the items that should derive the cultural context variable, they remain, and together form the variable Cultural Context since no theoretical basis was found why the remarkable common factor loadings with the faultline perception items occurred. Since they are theoretically fully independent, the cultural context items remained and the variable was computed. Together this results in a total of 5 variables.
Also the Context Culture variable could be computed. By taking the averages of the two variables Country of Birth and Nationality, 167 low context respondents and 19 high context respondents were identified.
Correlation. To gain insight into the relationship among the variables the correlations
27
available. Probably the most important normality test is the Kolmogorov-Smirnov test. After conducting this test on all variables, the following results were provided:
Kolmogorov-Smirnov Test
Demographic Informational Non-Demographic Perception Change Kolmogorov-Smirnov
Z
1,012 1,120 1,464 1,065
Asymp. Sig. (2-tailed) ,257 ,162 ,028 ,206
Table 7. Normality test. N=186. Significance at α = 0,05
The Kolmogorov-Smirnov test assumes normality for variables. At N=186 a significance level of α = 0,05 was chosen to determine whether this assumptions holds for the selected variables. From table 7 we can conclude that for three of the variables have a p-value above the critical α of 0,05. Implying that the scores for these variables are significant and that these variables can be regarded as normally distributed. However, the p-value of the Kolmogorov-Smirnov Z score for the Non-Demographic Attributes variable is with just 0,028 below the required α of 0,05, implying that the variable is not normally distributed.
Since this results in 3 normal variables and 2 non-normal variables, since the cultural context variable is ordinal, we cannot use the Pearson correlation test, but have to use the Spearman-rank correlation test as a result. (Huizingh, 2007)
Hypothesis Testing. Finally, following these initial steps to ensure reliability and validity the
hypothesis that were derived from literature could be tested. To test the hypothesis, we want to know whether the means of the two cultural types differ significantly. Generally, for an interval variable this implies using a t-test. (Cooper & Schindler, 2008 & Huizingh, 2007) However, as with the Pearson correlation test, t-tests require a normally distributed variable. As already identified, the variables used in this analysis are not all normally distributed. Therefore, a non-parametric alternative to the t-test had to be chosen. The so-called Mann-Whitney was selected since. Like the t-test it identifies whether potential differences between variables differ significantly (Huizingh, 2007). Thus the Mann-Whitney test matches with the hypotheses stated in this research.
28
consider, that since the hypotheses tested in this research questions whether a group scores higher than the other group, this implies that one-tailed testing applies to the p-value. (Huizingh, 2007) Since the p-value that is provided for by the Mann-Whitney test considers a two-tailed significance (implying that there is a difference of some sort, without assuming direction) this p-value has to be divided by two. If the one-tailed p-value then remains above α = 0,05 (p > α), the difference between variables is insignificant. If the one-tailed p-value is below α = 0,05 (p < α) the difference is significant. By then comparing the mean rankings that the test provides, it can be concluded what a significant difference implies, thus whether the assumed direction of the difference as stated in the hypotheses holds.
Results
Correlations
After analyzing all variables on correlations, the following results were provided:
Spearman Correlations Test
Mean Standard
Deviation Demographic Informational Non-Demographic Perception Change Context Culture Demographic Correlation Coefficient 2,9641 1,18182 Sig. (2-tailed) . Informational Correlation Coefficient 3,0308 1,24918 ,916 Sig. (2-tailed) ,000 . Non-Demographic Correlation Coefficient 3,3447 1,39587 ,929 ,914 Sig. (2-tailed) ,000 ,000 . N 186 186 186 Perception Change Correlation Coefficient 3,4126 1,28919 ,858 ,832 ,893 Sig. (2-tailed) ,000 ,000 ,000 . Context Culture Correlation Coefficient 1,10 ,304 ,140 ,099 ,050 ,104 Sig. (2-tailed) ,056 ,180 ,498 ,159 .
Table 8. Correlation test. Significance at α = 0,05
29
implies that when the score of variable X increases, this also results in an increase of variable Y. This can be explained considering the fact that the variables are to a large extent computed with the same Team Division items, as such they are to a large extent similar. However, the correlation coefficients between the independent variable Context Culture and the dependent variables are far from the required significance level of α = 0,05. Only the relationship between Context Culture and the Demographic Attributes is close to linear. Now this does not imply that there is no relationship between the dependent and independent variables, it only assumes that there is no linear relationship. This thus implies that another factor may influence the relationship.
Hypotheses Testing
To examine whether the hypotheses hold, or to identify other remarkable differences between the two types of culture, the variable were compared by the Mann-Whitney test. The results were as follows: Mann-Whitney Test Dependent Variable Context Culture N Mean Rank Sum of Ranks Mann-Whitney U Z score p-value (2-tailed) Hypothesis Demographic Low 167 90,96 15190,50 1162,50 -1,908 ,028 (,056) Rejected High 19 115,82 2200,50 Informational Low 167 91,71 15316,00 1288,00 -1,343 ,0895 (,179) Rejected High 19 109,21 2075,00 Non-Demographic Low 167 92,60 15463,50 1435,50 -,680 ,2485 (,497) Rejected High 19 101,45 1927,50 Perception Change Low 167 91,62 15301,00 1273,00 -1,411 ,079 (,158) Rejected High 19 110,00 2090,00
Table 10. Test results. N total =186. Significance at α = 0,05
Let's examine these results per hypothesis. Hypothesis 1 stated that: Individuals from a lower
context culture are expected to perceive more faultlines on the basis of demographic attributes than individuals from a higher context culture. The first step is to examine whether
30
therefore have to examine the one-tailed p-value for this Z-score which with 0,028 is below the critical significance α of 0,05. Thus, following this assessment of the Z-value we can see that there is indeed a group that will score significantly higher when it comes to the perception of division on demographic attributes. However, when we then focus on the mean rank scores, we can see that the mean rank of high context cultures is significantly higher than the mean rank of low context cultures. This in contrast to the hypothesis that was stated on the basis of literature. Therefore, we have to reject hypothesis 1, since although one type of culture is perceives significantly more than the other, it are higher context cultures which score significantly higher on the perception of division on the basis of demographic attributes whereas lower context cultures were hypothesized to score significantly higher.
Considering hypothesis three: Individuals from a lower context culture are expected to
perceive more faultlines on the basis of non-demographical attributes than individuals from a higher context culture, when we examine the Z-value of -0,680 we can see that this is
considerably closer to zero than the Z-value on the previous hypothesis. Thus apparently the difference between the two cultural types is lower. To examine whether the difference is still significant, we have to evaluate the one-tailed p-value of 0,2548 against the α of 0,05. As we can see p > .05 thus we can reject hypothesis 2 since no significant difference exists between the two types of culture.
When focusing on hypothesis three, which states that: Individuals from a higher context
culture are expected to perceive more faultlines on the basis of informational attributes than individuals from a lower context culture we can see that with a Z-score of -1,343 there is a
considerable difference between the two cultural types. However, when determining the significance of this relationship, thus to establish whether the difference also holds in a wider population, we can see that with a p-value of, 0895 this is higher than the required α of 0,05. Thus p > .05 and therefore we have to reject hypothesis three.
Finally, hypothesis four argues that: Individuals from a higher context culture are expected to
perceive more faultlines on the basis of the perception towards change. With a p-value of
0,079 for the Z-score of -1,411 we can say that although there is a considerable difference between high and low context cultures on the extent to which they perceive faultlines on the basis of attitudes towards change, the results from this sample do not hold in a wider since p
31
Thus, summarizing this results section, we can see that all hypotheses were rejected. Even though some considerable differences exist between the two types of culture, they are either insignificant and thus cannot be generalized onto a wider population or the expected direction of the difference was the opposite as hypothesized. Nevertheless, although all hypotheses were rejected, an important implication of the results is the fact high context cultures perceive significant more faultlines on the basis of the demographic attributes.
Conclusion & Discussion
Conclusion
Based on past research it was expected that both cultural types had their respective differences regarding their perception of team work; social procedures and change. Following this theoretical basis, four hypotheses were chosen, each with an expected direction as to which faultline attribute would be considerably more perceived by a single cultural type. Yet, all of these hypotheses were rejected following a statistical assessment of a data set incorporating both cultural types. Even though this data set was not equally distributed among both cultural types, following reliability and validity assessments it was suitable for further research, especially considering the exploratory nature of this paper.
Now, in conclusion we have to return to the initial research question: What are the differences
on faultline perception between high and low context cultures within an organizational change environment?
32
Discussion
When interpreting the results of the statistical analysis of the data, some remarkable findings should receive further attention. Primarily it is interesting to observe that in this sample, individuals from a high context culture scored higher on each variable that indicated a faultline perception. Although not significantly on all variables (only on the demographic attributes it scores significantly higher) it is remarkable to see this tendency. The immediate question ''why?'' springs to mind. Now although a conclusive answer to this requires further research, this might be a result of the more inward and collective focus of high context cultures. As Ting-Toomey (1985) argues, high context cultures are both more group oriented and deviance discouraged. The group orientation, or collective attitude of high context culture might result in a closer attention to the team dynamics. It might be that any possible deviations from individuals that could conflict with the collective or team are therefore more identified by high context cultures than the more individually and more variation encouraged (Ting-Toomey 1985) low context cultures.
Furthermore, it might be important to consider what is perceived as a team. Individuals prefer social interaction within small groups, ideally containing two or three people (Hogg & Cooper 2003) Now although this assumption will probably hold for both types of cultures, the collective; group oriented nature of high context cultures might result in more ''stronger'' faultlines between small intra-team groups. This concept of ''strong'' or ''weak'' faultlines was argued by Lau and Murnighan (1998) faultlines are argued to become stronger as more attributes are highly correlated, thus increasing the homogeneity of the subgroups. (Lau and Murnighan 1998) Considering the homogeneity, it seems likely that a collective group oriented culture enhances intra-team bonding and as such results in a stronger homogeneity and thus probably a stronger perception of faultlines. These issues, considering the collectivism vs. individualism between the two cultural types, and its influence on the fautline strenght, are an interesting question that could be derived from this research that requires further research.
33
than low context cultures, it seems important to investigate whether this indeed results in more relational and task conflict and as a result a lower performance and satisfaction. This is especially an interesting point for further consideration since Choe (2002:8) argues that when ''behavior comes into conflict with group values, it will be engaged.'' Thus indicating that the collectivist nature of high context cultures should result in team trying to resolve such issues when actions would hamper the team’s or group’s interests. Therefore it seems desirable that further research is conducted on the dynamics that result from faultline perception.
When we return to the topic of correlations, further thought seems required as to why no linear relationship could be identified between the cultural variable and the variables regarding faultline attributes. Although the cultural variable and faultline attribute variables are theoretically completely independent constructs, a linear relationship seemed logical following Hall (1976) who argued that culture designates what individuals consider important or not. However, in this data set such a linear relationship could not be identified. Thus, it might be that the differences between the types of culture we identified are not only due to the cultural perspective but are additionally to some extent influenced by other moderating variable(s). However, since the data set used to determine this influence was constructed to a large extent of low context cultures and thus limiting the variation within the variable, this may have hampered the correlation test. Further research into the relationship between culture and faultline attributes, specifically concerning possible moderating variables, seems desirable.
Finally, in general it seems advisable to, in addition to further qualitative research to determine the impact of cultural influence on the faultline perception, perform qualitative research into why exactly the cultural influence has an influence on team dynamics and to what extent the differences in faultline perception between different cultures result in differences in team performance. Especially when it comes to the inherent standpoints on teamwork and their influence further in-depth clarifications seem to be relevant. Furthermore it seems desirable to investigate teams in which both cultural types are present to identify what the impact of such diversity is on faultlines.
Practical Implications. As this is an exploratory research, the main implications that could be
34
context culture will tend to perceive more faultlines on the basis of demographic attributes in particular and can be assumed to do so on several other attributes. It seems fair to argue that practitioners dealing with both cultural types should acknowledge the fact that high context cultures are more likely to perceive subgroup formation and will deal accordingly. Thus they are more likely to form subgroups within their team. Whether this has negative or positive effects, the perception of subgroups is something practitioners should monitor, ideally by means of engaging in discussions regarding the faultline perceptions, in order to allow for an optimal utilization of the groups performances.
Limitations. However, although this exploratory research has identified some points that
require further academic consideration, it is important to discuss its limitations and areas for further improvement. Primarily it is important to identify that this research took place in the Netherlands. Now although this might lead to similar circumstances for all variables, i.e. ceteris paribus, it is likely to assume that it has had its influence on the results. With the Netherlands being undoubtedly a low context culture, the environment in which most respondents were questioned was generally low context as well. Although several of the high context respondents were operating in mainly ''high context teams'' their daily environment can be expected to be more ''low context'' and this could have had its impact on the results. Since these teams operated within a low context setting, they are likely to deal with partners; customers and other stakeholders from a low context setting. Furthermore, since most members are, in their private life, also engaging in a generally low-context society the responses of the high context cultures might have been ''low-contextualized.'' Therefore, a research that was held in multiple cultural environments seems advisable.
35
people from a Germanic background, a sample with multiple low-context cultures seems desirable.
Additionally, the results were derived from a single questionnaire, taken at a single moment. As a result, the responses identify the opinion of individual respondents at that specific moment. Needless to say, this opinion might not fully represent the opinions these respondents would have had when investigating their average opinion on team dynamics as a result of the organizational change over a longer period. By focusing on a single measurement, respondents could have been influenced by a variety of different factors, both professional and private. As a result it is hard to identify whether the responses were only due to the organizational change environment or whether additional influences were important as well. A more longitudinal study with multiple measurements might have enhanced the representation of the actual opinions of respondents. Yet due to limited time and by specifying questions specific on the impact of organizational change the impact of these issues were minimalized.
36
Bibliography
Amaram, D. I. 2007. Cultural diversity: Implications for workplace management. Journal of
Diversity Management, 2(4:, 1-6.
Augsburger, D. 1992. Conflict Mediation across Cultures: Pathways and Patterns. Louisville, KY: WJK.
Berger, M. C. 1984. Critical thinking ability and nursing students. Journal of Nursing
Education, 23(7): 306- 308.
Bettenhausen, K. L. 1991. Five years of groups research: What we have learned and what needs to be addressed. Journal of Management, 17: 345-381
Bezrukova, K., Jehn, K. A., Zanutto, E. L., & Thatcher, S. 2009. Do workgroup faultlines help or hurt? A moderated model of faultlines, team identification, and group performance.
Organization Science, 20: 35-50.
Brewer, M. B. 2001. Ingroup identification and intergroup conflict. In R. D. Ashmore, L. Jussim, & D. Wilder (Eds.), Social identity, intergroup conflict, and conflict reduction: 17-41. Oxford, UK: Oxford University Press.
Burnes, B. 2009. Managing change. 5th edition. Harlow: Prentice Hall
Byrne, D. E. 1971. The attraction paradigm. San Diego, CA: Academic Press.
Caldwell, S., Fedor, D., Herold, D. M., and Liu, Y. 2012. When Does Management’s Support for a Change Translate to Perceptions of Fair Treatment? The Moderating Roles of Change Attributions and Conscientiousness. Journal of Applied Behavioral Science, 48(4):441-462
Carton, A. M., and Cummings, J. N. 2012. A theory of subgroups in work teams. Academy of
Management Review, 37(3): 41–470.
Caudron, S. 1994. Diversity ignites effective work teams. Personnel Journal, 73(9): 54. Centraal Bureau voor de Statistiek.(CBS) Bevolking: Migratie/Verhuizingen. Accessed, September 2013.
Centraal Bureau voor de Statistiek.(CBS) Beroepsbevolking in twee decennia ruim 5 jaar ouder. (2013). Accessed, September 2013.
Choe, Y. (2002) Intercultural Conflict Patterns and Intercultural Training Implications for Koreans. International Area Review 5(1): 111-128
Cohen, S. G., & Bailey, D. E. 1997. What makes teams work: Group effectiveness research from the shop floor to the executive suite. Journal of Management, 23: 239–290.
Collins, D. (1998). Organizational Change: Sociological perspective. London. Routledge.
37
Cooper, D.R. & Schindler, P.S. (2008). Business Research Methods. Singapore:McGraw-Hill
Gover, L. and Duxbury, L. 2012. Organizational Faultlines: Social Identity Dynamics and Organizational Change, Journal of Change Management, 12(1): 53-75
Greatz, F. Smith, A. (2010). Managing Organizational Change: A Philosophies of Change
Approach. Journal of Change Management. 10(2): 135-154.
Gratton, L., Voigt, A., & Erickson, T. J. 2007. Bridging faultlines in diverse teams. MIT
Sloan Management Review, 48(4): 22-29.
Gross, Y. T., Takazawa, E. S., & Rose, C. L. (1987). Critical thinking and nursing education.
Journal of Nursing Education, 26(8): 317-322.
Hackman, J. R., & Wageman, R. 2005. A theory of team coaching. Academy of Management
Review. 30: 269-287
Hair, J.F., Black, W.C., Babin, B.J., & Anderson, R.E. (2010). Multivariate Data Analysis. Seventh Edition. Prentice Hall, Upper Saddle River, New Jersey.
Hall, E.T. (1976). Beyond Culture, New York: Doubleday
Hall, E. T., & Hall, M. R. (1990). Understanding cultural differences. Yarmouth, ME: Intercultural Press Inc.
Hofstede, G. (1980). Culture's consequences: international differences in work-related. Beverly Hills, Sage Publications.
Huizingh, E. (2007) Applied Statistics with SPSS. London: Sage
Hogg, M A & Cooper, J (eds) (2003) The SAGE Handbook of Social Psychology. London: Sage
Hollenbeck, J.R., Ilgen, D.R., Major, D.A., and Sego, D.J. 1993. Team research in the 1990s. In M. M. Chemers and R. Ay- men (Eds.), Leadership theory and research: Perspectives and
research: 245-270.
Homan, T. (2008). De Binnenkant van Organisatieverandering. Alphen aan den Rijn. Kluwer.
Jehn, K. A., & Bezrukova, K. (2010). The faultline activation process and the effects of activated faultlines on coalition formation, conflict, and group outcomes. Organizational
Behavior & Human Decision Processes. 112(1): 24-42.
Kelly, P. (2009) International Business and Management, Andover, Cengage Learning EMEA
38
Kuvaas, B., & Selart, M. (2004). Effects of attribute framing on cognitive processing and evaluation. Organizational Behavior and Human Decision Processes. 95: 198–207. Lau, D. C., & Murnighan, J. K. 1998. Demographic diversity and faultlines: The
compositional dynamics of organizational groups. Academy of Management Review. 2: 325-340.
Mannix, E., & Neale, M. A. 2005. What differences make a difference? The promise and reality of diverse teams in organizations. Psychological Science in the Public Interest. 6(2): 31–55
Martins, L. L., and Milliken, F. J. 1996. Searching for Common Threads: Understanding the Multiple Effects of Diversity in Organizational Groups. Academy of Management Review. 21(2) 402-433
Molleman, E. 2005. Diversity in demographic characteristics, abilities and personality traits: Do faultlines affect team functioning? Group Decision and Negotiation, 14: 173-193.
Kozlowski, S. W. J., & Bell, B. S. 2003. Work groups and teams in organizations. In W. C. Borman, D. R. Ilgen, & R. J. Klimoski (Eds.), Handbook of psychology: Industrial and organizational psychology, vol. 12: 333–375. London: Wiley.
Pfeiffer, J W (1993) The 1993 Annual: Developing Human Resources. San Diego, Pfeiffer & Co.
Organisation for Economic Co-operation and Development (OECD). International migration policies and data. Accessed September 2013
Oreg, S. (2003). Resistance to change: Developing an individual differences measure.
Journal of Applied Psychology, 88(4): 680-693
O’Reilly III, C. A., Caldwell, D. F., Barnett, W.P. (1989). Work group demography, social integration and turnover. Administrative Science Quarterly, 34, 21-27.
Richard, O. C. (2000). Racial diversity, business strategy and firm performance: A resource-based view. Academy of Management Journal, 43, 164-177.
Rink, F. A., & Jehn, K. A. 2010. How identity processes affect faultline perceptions and the
functioning of diverse teams. In R. Crisp (Ed.), Psychology of social and cultural diversity:
281-296. Malden, MA: Wiley-Blackwell.
