A Field Study of Group Faultlines, Team Identity, Conflict, and Performance in Organizational Groups

A Field Study of Group Faultlines, Team Identity, Conflict, and Performance in Organizational

Groups

December 2001

Katerina Bezrukova

The Wharton School

University of Pennsylvania

Philadelphia, PA 19104-6370

Tel: (215) 573-5726 Fax: 215-898-0401

E-mail: [email protected]

Karen A. Jehn

The Wharton School



Tel: 215-898-7722 Fax: 215-898-0401

E-Mail: [email protected]

Elaine Zanutto

The Wharton School



E-Mail: [email protected]

1


Groups

Abstract

We further the theoretical understanding of group faultlines (hypothetical lines that split a group into subgroups based on two or more demographic attributes; Lau & Murnighan, 1998) in demographically diverse organizations by also conceptualizing the distance of difference between the two subgroups. We create a faultline algorithm to capture not only how cleanly a group splits into subgroups (faultline strength) but also how far apart these subgroups are (faultline distance) on demographic differences (e.g., two members of age 20 are closer in age to two members of an opposing faultline of age 21 than of two members of age 71). To examine how group faultlines affect conflict and performance, we develop and test hypotheses in an organizational setting of 78 groups. We also investigate the moderating effects of team identity on the faultline–conflict link. The results show that groups with strong group faultlines (both high on faultline strength and distance) are more likely to have high levels of task, relationship, and process conflict. Furthermore, group faultlines are found to be detrimental to performance. Finally, a strong sense of team identity among group members minimizes the detrimental effects of group faultlines. Future research directions are presented.

Key words: group faultlines, team identity, and inter-subgroup conflict

2


Groups

Recent reviews of both group diversity and relational demography research have stressed

extremely mixed, and often, contradictory empirical results (Williams and O’Reilly, 1998). Some

researchers have reported the positive effects of diversity on process and performance variables

(Hoffman & Maier, 1961; Hoffman, 1978; Jackson, 1992; Jehn, Northcraft, & Neale, 1999). Others

have demonstrated the negative effects of diversity on performance (Ancona & Caldwell, 1992;

O’Reilly & Flatt, 1989; Steiner, 1972). The remaining research has often shown no relationship

between diversity and different performance outcomes (e.g., O’Reilly, Williams, & Barsade, 1997).

Given no consistent main effects of diversity on groups, recent studies have focused on exploring the

alternative ways of studying diversity. Specifically, one way to look at the group composition and

explain the contradictory findings is to apply the group faultline theory developed by Lau and Murnighan

(1998). This approach reconceptualizes a traditional understanding of diversity as a dispersion of a

population on a single demographic characteristic (Blau, 1977) into a more sophisticated consideration

of all the potential dynamics that many characteristics can activate (Lau & Murnighan, 1998).

Specifically, group faultlines take into account more than one demographic characteristics at a time, the

way these characteristics align, and the number of possible subgroupings that emerge (Thatcher, Jehn, &

Zanutto, 2000). We believe this new perspective on group diversity can more adequately explain the

effects of diversity on performance. It also has been shown to have a more powerful predictive ability

than other prevailing measurements of demographic differences (Bezrukova, Jehn, & Zanutto, 2001).

We attempt to further the theoretical understanding of group faultlines in demographically

diverse organizations by exploring the applicability of a new way of conceptualizing faultlines in

workgroups. We adapt a definition of group faultlines as hypothetical dividing lines that split a group into

subgroups based on two or more characteristics (Lau & Murnighan, 1998). We argue that group

faultlines can be viewed as a multidimensional concept that reflects at least two properties of the

demographic composition of a group: deepness (e.g. the more demographic characteristics that align,

the deeper the faultline) and wideness (e.g. the more different subgroups are from each other on the

demographic characteristics, the wider the faultline) of group faultlines. Therefore, we distinguish

3

between the faultline strength (how cleanly a group splits into subgroups) and faultline distance (how far

apart these subgroups are from each other). Conceptual and operational issues regarding the faultline

strength are discussed in more details in Thatcher, Jehn, and Zanutto (2001). In this paper, we

specifically focus on defining and measuring faultline distance, which has not, to our knowledge, been

done previously, and in addition to faultline strength.

Briefly, the more characteristics align themselves in the same way, the stronger the group

faultline and hence, the greater the faultline strength. Faultline distance goes further and reflects how far

apart possible subgroups are from each other. In other words, the more different the subgroups are on

the characteristics, the greater the distance between them. For instance, a group with 2 members who

are 50 year- old computer programmers and 2 members who are 20 year- old musicians will have

extremely high faultline strength (fau strength score = 1). On the other hand, a group with 2 members

who are 25 year-old computer programmers and 2 members who are 20 year-old system

administrators will also have extremely high faultline strength based on past conceptualizions and

operationalizations of faultlines (fau strength score = 1). Both groups score similar and extremely high on

faultline strength and both groups may split perfectly into two homogeneous subgroups. However, the

group composition, and potentially the group processes, of these two groups are very different.

Following the above example, the two subgroups in the first group are farther apart than the two

subgroups in the second group, based on age in this example, as profession is held constant (50 vs. 20

in group 1, and 25 vs. 20 in group 2). Thus, this difference might have a different impact on group

process and performance outcomes of these groups. To capture these faultline distance effects, we

develop the distance measure which allows us to compare groups based on how different homogeneous

subgroups are from each other (i.e., how wide the group faultline is in a group), rather than just

assessing if they are different or not.

In the next sections, we develop our hypotheses to examine the effects of group faultlines in the

connection to three types of conflict - task, relationship, and process conflict (Jehn, 1995). Further, we

link group faultlines to different performance outcomes (e.g. group bonuses, stocks, performance

ratings, and satisfaction). Additionally, we examine the mediating role of conflict in the relationship

between group faultlines and group performance. Finally, we look at the moderating effects of team

identity on the group faultlines and conflict link. We develop our hypotheses separately for faultline

4

strength and faultline distance to investigate whether one of the two is more influential and would have

more predictive ability in explaining the diversity effects in groups (See Figure 1). We consider six

demographic characteristics: age, gender, race, tenure with the company, level of education, and

functional background in our analysis. We chose these variables based on the previous research on

group diversity that showed these variables were potentially relevant for self-categorization and had

significant results in assessing relational effects (Tsui, Egan, & O’Reilly, 1992). We have collected data

on a large sample in an organizational setting with over 78 groups to test our hypotheses.

---------------------INSERT FIGURE 1 ABOUT HERE --------------------

Linking Faultline Strength and Faultline Distance to Conflict

Using the coalition theory (Caplow, 1956; Komorita & Kravitz, 1983; Mack & Snyder, 1957;

Murnighan, 1978) and Schneider’s attraction-selection-attrition model of organizational membership

(1983), we expand group faultline theory and propose that if more demographic attributes align in the

same way (faultline strength), group members in each subgroup will perceive the similarity within their

subgroup. Since similar members are likely to interact with each other more often and find their

interactions pleasant and more desirable, they will be likely to form coalitions (Byrne, 1971; Pool, 1976;

Roger & Bhowmik, 1971; Stevenson, Pearce, & Porter, 1985). Due to the similarity among group

members involved in coalition formation, the conflict within subgroups is apt to decline. However, the

existence of coalitions is likely to amplify the salience of in-group/out-group membership causing strain

and polarization between subgroups (Hogg, Turner, & Davidson, 1990). Once coalitions are formed,

the negative effects of stereotyping, in-group favoritism and out-group hostility are likely to sharpen the

boundary salience around coalitions and strengthen conflict between them. These group processes are

likely to lead to intensification of conflict between subgroups and therefore, promote or activate

intergroup conflict. In particular, we discuss and examine three types of conflict that have been identified

in working groups, bicultural teams, and organizing entities (Amason, 1996; Jehn, 1997; Jehn and

Jageuri, 2001; Jehn, Northcraft, and Neale, 1999; Pelled, 1996; Shah and Jehn, 1993). We add to the

existing literature by looking at the three types of conflict in the context of inter-subgroup relationships in

groups.

5

Because of negative categorization processes, subgroups are likely to experience frustration,

discomfort, hostility, and anxiety that can result in animosity and annoyance between individuals

belonging to different subgroups, and hence, relationship conflict is likely to emerge between two or

more subgroups (Jehn, 1997; Tajfel & Turner, 1986). We expand Jehn’s (1997) concept of

relationship conflict within groups by defining relationship conflicts as disagreements and incompatibilities

between two subgroups within a group about issues that are not task related, but that focus on personal

issues. Furthermore, the more that demographic attributes align in the same way (faultline strength), the

more salient the perceived similarities within subgroups, and the more salient the perceived differences

between subgroups. The greater salience of these out-group differences is likely to facilitate the more

intense polarization between subgroups, which will inevitably result in more fights over non-task related

issues. We argue that the greater the faultline strength, the higher the level of relationship conflict

between the two subgroups will be.

Based on the literature on minority dissent and decision making processes in work groups, we

argue that the very existence of subgroups within a group is a source of divergent thinking (De Dreu &

West, 2001). Specifically, demographically diverse subgroups are expected to have a broader range of

knowledge, experience, and opinion (e.g. Ancona & Caldwell, 1992). Variety in knowledge and

experiences can lead to disagreement among group members about group tasks (Jackson, 1992; Jehn

1995; Jehn 1997; Pelled, Eisenhardt, & Xin, 1997). We similarly expand the concept of task conflict

among all group members to task conflicts as disagreements between members of two subgroups within

a group around ideas and opinions about the task being performed. Furthermore, when subgroups are

formed, subgroup members have a strong tendency toward conformity to the opinion, idea, or

perspective favored by the subgroup (Baron, Kerr, & Miller, 1993). According to social categorization

theory, people confront positions perceived as stereotypical of their subgroup because, in reflecting the

agreement of similar others, such positions provide subjectively valid evidence about external world

(Hogg, Turner, Davidson, 1990). Therefore, the more demographic attributes align in the same way

(faultline strength), the more salient the perceived similarities within subgroups, and hence, the more

intense the polarization between subgroups around ideas and thoughts. We propose that this

polarization will result in high levels of task conflict between subgroups within a group.

6

Based on the discussion above, we argue that the greater the faultline strength, the more intense

the polarization between subgroups around “different ways” of doing things, which will result in a high

level of process conflict between subgroups. Process conflicts between subgroups are about logistical

and delegation issues such as how task accomplishment should proceed in the work unit, who’s

responsible for what, and how things should be delegated (Jehn, 1997; Kramer, 1991). Therefore, we

propose:

Hypothesis 1a (H1a): The greater the faultline strength, the higher the level of

relationship conflict between subgroups.

Hypothesis 1b (H1b): The greater the faultline strength, the higher the level of task

conflict between subgroups.

Hypothesis 1c (H1c): The greater the faultline strength, the higher the level of process

conflict between subgroups.

To conceptualize faultline distance and its effects on group processes and performance, we

draw upon social dominance theory (Sidanius & Pratto, 1999), social identity theory (Tajfel & Turner,

1986), social categorization theory (Turner, 1985; Turner, Hogg, Oaks, Reicher, & Wetherell, 1987),

and similarity-attraction-selection paradigm (Byrne, 1971; Roger & Bhowmik, 1971). Social

dominance theory suggests that human beings naturally form hierarchical groupings that represent a

hierarchical ordering of groups according to their relative position in a larger system of groups (Sidanius

& Pratto, 1999). Furthermore, these researchers suggest that group members (in our case, subgroup

members) share a schema that cognitively constructs other out-group members in a hierarchy by

agreeing what groups are to be kept at a greater distance than others (Mullick & Hraba, 2001). For

example, subjects in the Netherlands, the former Soviet Union, Pakistan, and the United States hold

social-distance hierarchies about ethnic groups in their own countries (Hagendoorn, Drogendijk,

Tumanov, & Hraba, 1998; Mullick & Hraba, 2001). The theoretical and empirical evidence provides

support for the tendency of group members to construct social representations in terms of relative

distances on each demographical characteristic. Faultline distance takes into account all of the potential

distances between group members on demographic characteristics and can be viewed as a distance

between subgroups.

7

We propose that relative distances among group members’ demographic characteristics (e.g., age 20 is

closer to age 25 than to age 50) might influence the salience of categories the group members will align

with. We argue that the less distance between group members’ demographics, the more similar they will

perceive each other, and vice versa. Since similar members are likely to interact with each other more

often, find their interactions pleasant and more desirable, they will be more likely to form coalitions or

cohesive subgroups than those more dissimilar (Pool, 1976; Stevenson, Pearce, & Porter, 1985).

Drawing on the relational demography literature, we propose that the extent to which subgroups are

dissimilar on demographic characteristics will affect the amount of conflict within a group (Tsui, Egan, &

O’Reilly, 1992). We argue that the faultline distance between subgroups of aligned group members is

inherently disruptive by virtue of identification with the distinct subgroup. According to Tajfel and Turner

(1986) this disruption is a sufficient cause for negative emotion, and therefore prejudice, which can lead

to conflict and hostility towards the different subgroup. A number of studies examined whether the social

distance between ethnic groups was associated with perceived out-group threat resulting in aggression

and competition for resources (Sherif, 1966). Furthermore, individuals from each subgroup are likely to

form negative interpersonal relationships across subgroup boundaries (Labianca, Brass, & Gray, 1998).

We argue that these detrimental processes will escalate inter-subgroup conflict over personal issues,

namely, relationship conflicts.

Based on information and decision-making literature, we suggest that the greater faultline

distance between subgroups, the less communication exchanges will be between the subgroups, and

hence, the more exclusive the access will be to informational networks outside the group (Williams &

O’Reilly, 1998). This is likely to lead to a greater difference in the amount and content of information

available across subgroups and, therefore, will stimulate emergence of divergent viewpoints between

subgroups over the group task and the process of how the group work should get done. Furthermore,

members of a distinct subgroup are likely to adapt a particular idea or point of view and define it as a

quality of their subgroup that characterizes and distinguishes it from another subgroup. When subgroups

import divergent views, conflicts around task and process are likely to result (Ancona, 1990; Scott,

1997). Another line of reasoning comes from the social categorization theory which posits that

individuals are more persuaded by the same information from ingroup than from out-group sources, and

from similar others more than from dissimilar others (Hogg, Turner, Davidson, 1990). Therefore, we

8

argue that the greater the faultline distance between subgroups, the greater the subgroup conformity

towards these divergent opinions, and thus, the more intense the task and process conflict between

subgroups.

Hypothesis 2a (H2a): The greater the faultline distance, the higher the level of

relationship conflict across subgroups.

Hypothesis 2b (H2b): The greater the faultline distance, the higher the level of task

conflict across subgroups.

Hypothesis 2c (H2c): The greater the faultline distance, the higher the level of process

conflict across subgroups.

Linking Faultline Strength and Faultline Distance to Performance

As stated earlier, the alignment of demographic attributes based on similarity of group members

is likely to generate coalitional activity (Caplow, 1956; Komorita & Kravitz, 1983; Murnighan, 1978).

Once coalitions are formed, the negative effects of stereotyping, in-group favoritism and out-group

hostility are likely to sharpen the boundary salience around coalitions, intensify conflict between them

that in turn will lead to the split of the group into subgroups. These detrimental social categorization

processes can surface conflict, disliking, and can also lead to decreased cohesion and social integration

(Tajfel & Turner, 1986; Webber & Donahue, 2001). This is likely to lessen the frequency of

interactions and information exchanges between members of different subgroups and, in turn will lead to

performance losses because both access to, as well as amount of, the information in the resource pool is

reduced (Clement & Schiereck, 1973; Freidman & Podolny, 1992). Additionally, the “us versus them”

mentality of subgroups is likely to make it easy for one subgroup to blame the other subgroup for

mistakes. This manifestation of the “sucker-effect” was found to have the detrimental effects on the level

of potential productivity of the entire group (Hogg, 1996). Therefore, we argue that faultline strength will

be negatively related to group performance.

We also propose that the greater the faultline distance across subgroups, the more opposing the

subgroups are and the more incessant the clashes are. When these negative processes get activated,

group cohesion and collective action become impossible (Mackie, Devos, & Smith, 2000). This is likely

to lead to more restricted access to important resources that a subgroup needs to contribute to the

9

accomplishment of a specific task. An entire group might end up paralyzed, as each subgroup will not

cooperate with needed others. Furthermore, once the boundaries around subgroups become salient,

some subgroups can be automatically placed on the status hierarchy relative to the dominant subgroup

(Asante & Davis, 1985). The status differentials may further reduce the level of contribution since the

low-status subgroups are not expected to contribute as much as high-status subgroups (e.g. Kimble &

Musgrove, 1988). This has been shown to result in productivity losses in workgroups (Kirchmeyer &

Cohen, 1992; Plaks & Higgins, 2000). Therefore, we believe that faultline distance will be negatively

associated with group performance. Relational demography research provides some empirical evidence

regarding the effects of employee differences on demographic characteristics on various performance

outcomes. For example, Wagner, Pfeffer, and O’Reilly (1984) found that larger demographic distances

among the vice-presidents in 30 firms were associated with higher levels of turnover. Tsui, Egan, and

O’Reilly (1992) demonstrated that group members were more satisfied with their membership in a

group when they fell within the boundary of a specific age category. Based on the above theorizing, and

adding to the empirical work done on turnover and satisfaction, we predict:

Hypothesis 3a (H3a): The greater the faultline strength, the lower the level of group

performance.

Hypothesis 3b (H3b): The greater the faultline distance, the lower the level of group

performance.

The Mediating Effects of Task, Relationship, and Process Conflict

Based on the previous theoretical discussion and studies that have already investigated the

mediating effect of conflict on the relationship between group diversity and performance (Jehn,

Northcraft, & Neale, 1999), we predict a mediating role of conflict (three types) between faultline

strength and distance, and group performance. Referring to Figure 1, our mediation hypotheses are:

Hypothesis 4a (H4a): Relationship conflict will mediate the effect of faultline strength

on group performance

Hypothesis 4b (H4b): Task conflict will mediate the effect of faultline strength on

group performance.

10

Hypothesis 4c (H4c): Process conflict will mediate the effect of faultline strength on

group performance.

Hypothesis 5a (H5a): Relationship conflict will mediate the effect of faultline distance

on group performance.

Hypothesis 5b (H5b): Task conflict will mediate the effect of faultline distance on

performance.

Hypothesis 5c (H5c): Process conflict will mediate the effect of faultline distance on

group performance.

The Moderating Effects of Team Identity

Team identity refers to an individual's sense that his or her team has a unified identity and reflects

a sense of entitativity, or a common perception of group cohesiveness (definition adapted from Early &

Mosakowski, 2000). Based on social identity and social categorization theories (Tajfel & Turner,

1986), we propose that group faultlines will be less likely to cause inter-subgroup conflict when group

members strongly identify themselves as a single team. According to Lau and Murnighan (1998),

faultlines can lead to salient subgroups that then become a basis for social identification and

categorization. Once group members start identifying themselves with a particular subgroup, the negative

outcomes of categorization (e.g. negative stereotyping and prejudice) are likely to result in conflict

(Thatcher and Jehn, 1999). However, because individuals have multiple identities, the salience of a

particular identity depends on the context where individuals operate (Hogg & Terry, 2000). We argue,

consistent with Thatcher and Jehn’s (1998) theory of intercultural interaction within groups, that when

the group develops a strong sense of group identity, the group members are less likely to categorize

themselves into demographic categories (e.g. race, age, gender) and more likely to conceptualize

themselves as members of this work group. That is, when group members have a super-ordinate, or

meta-identity, focused on the group they will be less likely to split based on other identities such as race,

gender, etc. For example, individuals with a strong group identity are both committed to that group

identity and hold the group identity as primary. Therefore, we propose that:

Hypothesis 6 (H6a): Team identity will moderate the relationship between faultline

strength and relationship conflict; that is, if there is a strong team identity among

11

group members, faultline strength is not likely to result in high level of relationship

conflict. In contrast, if there is a weak team identity among group members, faultline

strength is likely to result in high level of relationship conflict.

Hypothesis 6b (H6b): Team identity will moderate the relationship between faultline

strength and task conflict; that is, if there is a strong team identity among group

members, faultline strength is not likely to result in high level of task conflict. In

contrast, if there is a weak team identity among group members, faultline strength is

likely to result in high level of task conflict.

Hypothesis 6c (H6c): Team identity will moderate the relationship between faultline

strength and process conflict; that is, if there is a strong team identity among group

members, faultline strength is not likely to result in high level process conflict. In

contrast, if there is a weak team identity among group members, faultline strength is

likely to result in high level process conflict.

Hypothesis 7a (H7a): Team identity will moderate the relationship between faultline

distance and relationship conflict; that is, if there is a strong team identity among

group members, faultline distance are not likely to result in high level relationship

conflict. In contrast, if there is a weak team identity among group members, faultline

distance is likely to result in high level of relationship conflict.

Hypothesis 7b (H7b): Team identity will moderate the relationship between faultline

distance and task conflict; that is, if there is a strong team identity among group

members, faultline distance is not likely to result in high level of task conflict. In

contrast, if there is a weak team identity among group members, faultline distance is

likely to result in high levels of task conflict.

Hypothesis 7c (H7c): Team identity will moderate the relationship between faultline

distance and process conflict; that is, if there is a strong team identity among group

members, faultline distance is not likely to result in high level of process conflict. In

contrast, if there is a weak team identity among group members, faultline distance is

likely to result in high level of process conflict.

METHODS

12

Research Site

Our sample is a large Corporate Headquarters with over 26,000 employees at all ranks within

the organization in the computer industry. Employees work in a range of business units (n=12) such as

corporate administration, finance, sales, product development, software systems, and manufacturing.

The workgroups were created using a reporting system developed by the company as well as the

information about the organization/business units’ structures provided by key senior staff. We identified

workgroups from a company listing of who reports to whom as the working groups within the

organization are specified this way. We verified that these were actual working groups (i.e., they

interacted on a day-to-day basis, were task interdependent, identified each other as group members,

and were seen by others as workgroups) by interview and observation. We were informed that

“groups” of 1 or 2 employees (n=973) or groups with over 18 employees (n=291) were not actual

working groups. This is consistent with our definition of group (see above) and with group process

theories regarding group size. In addition, given that the bases of our hypotheses are from social

psychology and organization group theory, we found this appropriate. We were unable to determine

whether the groups of size 8 or over could be broken down into smaller groups that may have then been

appropriate to use in the tests of our research model. This sample includes 518 individuals and 78

groups with complete data who were working full-time for all or part of the time period from January 1,

1999, to December 31, 1999. The age of employees ranged from 26 to 69 years with a mean of 46

years. The employees were 71.2% male, and 28.8 % female. The majority of employees (88%) were

white; 6.9% were African American, 2.7% Asian, 2.3% Hispanic, and there were no Native Americans

in this subsample. The level of education ranged from grade school to the Ph. D. level; the modal level

was a Bachelor’s degree. Tenure in the firm ranged from less than 1 year to 43 years with a mean of 15

years. Work functions included 22 distinct categories (e.g. customer service, finance, marketing).

Measures

Faultline Strength. Potential ethnic faultlines were measured along six demographic

characteristics (race, age, gender, level of education, tenure with the company, and functional

background) using a faultline algorithm and a rescaling procedure developed by Thatcher, Jehn, and

Zanutto (2000). This faultline strength measure calculates the percent of total variation in overall group

characteristics accounted for by the strongest group split, in other words, the faultline strength score

13

indicates how a group splits cleanly into two subgroups. We have calculated the faultline strength

scores excluding subgroups of size one because these subgroups cannot be considered as a group

based on social psychological perspective. Possible values of faultline strength ranged from .33 (weak

faultline strength) to .83 (very strong faultline strength).

Faultline Distance. We measured how far apart two subgroups are from each other on

demographic characteristics (race, age, gender, level of education, tenure with the company, and

functional background). This faultline distance measure differs from the existing measures of distance; it

takes into account multiple characteristics of group members simultaneously. Other distance measures

which have been widely employed in relational demography and diversity literatures (c.f. Williams &

O’Reilly, 1998; Blau, 1977; Tsui & O’Reilly, 1998) assess only one characteristic at a time.

Faultline distance measure was adapted from multivariate statistical cluster analysis (e.g.

Morrison, 1967; Jobson, 1992; Sharma, 1996) and calculated as a distance between centroids (the

Euclidean distance between the two sets of averages):

, where centroid for subgroup 1 = (X , X , X , … , X11 12 13 1P. . . .), centroid for

group 2 = (X , X , X , … , X21 22 23 2P. . . . ).

To rescale the variables so that they can be reasonable combined into one distance measure, we

calculated the scores so that difference in gender = difference in race = difference of 15 years of age

(approx 2sd) = difference of 10 years in tenure (approx 1 sd) = difference of 2 units of education

(approx 1 sd). We have also considered a difference in function is more important than a difference in

gender or race (1.5 times as important) so that difference in function = difference of 22.5 years in age =

difference of 15 years in tenure = 3 units of education. Possible values of faultline distance ranged from

.98 (little faultline distance) to 3.34 (very great faultline distance).

We provide a detailed discussion of faultline strength score and faultline distance measurement

method in Table 1 and below.

---------INSERT TABLE 1 ABOUT HERE-------

Let us consider the first two teams. Team 1 and team 2 have about the same faultline strength

scores (.8057 and .8304). These scores are very high and indicate an almost clean split of these groups

14

into two subgroups (members 1&3 versus 2&4&5 in team one, and members 1&4 versus 2&3&5 in

team two). However, the group members in team one are more different from each other across

subgroups on two demographic characteristics (age and tenure with the company) than the group

members are on these characteristics in team two. Thus, the faultline strength score doesn’t take into

account the fact that the subgroups in team 1 are much farther apart (conceptually different) than the

subgroups in team 2. On the other hand, our distance measure takes this into account and provides a

metric of how far apart (conceptually different) these resulting subgroups are (2.9334 versus 2.0265).

In looking over two other teams (team 3 and team 4), we notice that they have approximately

the same faultline scores (.3589 and .3572). These scores are relatively low and indicate weak faultlines

for both groups. Note team 4 is very diverse on race whereas team 3 is homogeneous on race while

the dispersion of other demographic characteristics is about the same in both groups. This means that

our faultline strength measure captures faultlines being either absent or unlikely at minimum and

maximum diversity, which is consistent with Lau and Murnighan’s (1998) theoretical predictions

concerning the potential effects of diversity and faultlines. Distance scores indicate that resulting

subgroups in team 3 are farther apart than do subgroups in team 4.

Thus the combination of faultline strength and distance measures can provide us with a better

understanding of the possible effects of faultlines in workgroups. The faultline strength scores tell us

how cleanly the group splits into two subgroups, but does not address the distance between the two

subgroups. This limitation can be overcome by using a distance score, which captures the degree of

difference between two subgroups. The combination of the faultline strength score and the faultline

distance score can provide useful information not yet captured in any past demographic or diversity

research that we are aware of. When fitting a linear regression, if we include both the faultline strength

score and the faultline distance score and the interaction between faultline strength score and the faultline

distance score, we can see whether a strong faultline has the same effect regardless of the distance

between the two subgroups.

Content Analyses

As employee survey data (i.e., direct measures of conflict) was not available (and what is most

often used to assess group conflict), we content-analyzed the company’s documents (e.g. the

Leadership Program reports) to generate measures of our variables (see the examples of the LP reports

15

in Appendix 1). The LP reports capture the dominant group processes in work groups including task,

process, and relationship conflict, team identity, as well as group members’ satisfaction. These

documents are a part of a human resources-sponsored application designed to provide the company’s

succession planning process.

We organized the company textual data by work groups and created frequency lists for each

group using the Monoconc content analysis computer program. Then, we developed a list of key words

characterizing each conflict and satisfaction variable based on relevant group and organizational theories

(see details for each construct below), as well as the concepts used in the company’s rhetoric. We

conducted key word searches on all work groups to obtain the number and frequency of key words

mentioned (we set the search parameters to show results with the 70 characters surrounding the search

terms). Following the method of Jehn and Werner (1993), two independent raters reviewed the

surrounding context and coded the text for each workgroup on each variable of interest as defined by

theory. They evaluated the intensity of the conflict in each group on a scale from 1 to 7. The interrater

reliability was quite high and ranged from .89 to .97 on the variables.

To arrive at the score, raters developed a four-step procedure. They began by discussing the

first two pages one group at a time. The discussion helped them formulate scoring rules to guide them

through the rest of the process. After discussing and reaching agreement, they assigned a common

score for the group on each variable. Next, they repeated the above process for four more pages

containing the textual data. They then did six pages individually guided by the rules they had created.

Once completed, they met again to compare the scores they assigned. When discrepancies appeared,

they discussed why and modified their rules if necessary. They also modified their discrepant scores until

they once again had agreement (whereas in the first two steps they had perfect agreement, for the third

step, they allowed for a deviation of one). If discrepancies do not appear, the raters continue to work

independently on the rest of the document. They defined the scale, for example, as (1) – equals 0

conflict. As long as phrase is relevant to conflict type, keep word and assign this if necessary to reflect

the low extent or non-existence of the conflict. (2) – slight difference of opinion – no confrontation. (3)

– clear difference of opinion – no confrontation. (4) – mild or constructive criticism. (5) – harsher

criticism. (6) – strong disagreement – clear confrontation. (7) – approaching or actual gridlock.

16

Conflict. The conflict variables were operationalized as: (1) general conflict: perceived

incompatibilities or perceptions by the parties involved that they hold discrepant views or have

interpersonal incompatibilities (e.g., argue, oppose, compete, agree, concern, confusion, contradict).

(2) process conflict: conflict about how task accomplishment should proceed in the work unit, who’s

responsible for what, and how things should be delegated (e.g. allocate, delegate, assign,

responsibility, who, process, schedule). (3) relationship conflict: interpersonal incompatibilities among

group members, which typically includes tension, animosity, and annoyance among members within a

group (e.g. enemy, fault, personal, backstabbing, complain, pressure). (4) task conflict:

disagreements among group members about the content of the task being performed, including

differences in viewpoints, ideas, and opinions (e.g. discuss, viewpoint, differ, negotiate, perspective,

ends, opinion). Two extracts of the textual data we used to specify conflict variables are included in

Appendix 1.

Team identity. We specified our team identity variable as an individual’s sense that her or his

team had a unified identity and reflects the sense of entitativity, or the common perception of group

cohesiveness (e.g., team, group, communicate, relationship, help, people, person).

Performance. We used group and individual performance ratings, bonuses and stock options

as outcome variables. We also used business unit performance ratings as determined by the Board of

Directors and executive management. Stock options (group records) refer to the number of options

awarded. Bonus amounts (group records) are the actual bonus amounts paid out for the year. Bonus

amounts are calculated by running the bonus calculation module, which is the program code that

performs the actual bonus calculations. The yearly bonus is calculated on total base salary for the year

and includes multiple performance indicators determined by the company. Performance ratings

(individual records) are the codes associated with an employees' performance review (e.g. 5 refers to

outstanding performance, and 1 refers to unsatisfactory).

Satisfaction. Content analyzed data includes indicators of employees’ satisfaction. We

specified two variables to indicate the employees’ satisfaction: (1) positive attitudes toward work-

related issues (e.g., good, well, best, better, improve, win, success, improve, gain, great, happy).

(2) negative attitudes towards work-related issues (e.g., no, not, don’t cannot, bad, loss, poor, fail,

mislead, exploit, ineffective, weak).

17

RESULTS

Table 2 displays means, standard deviations, and correlations among all variables. As expected,

faultline strength and faultline distance were positively correlated with each other. As found in previous

research task and process conflict were highly correlated and positively related to group bonuses and

performance ratings. The performance measures were highly correlated indicating that the bonuses,

stock options, and performance ratings measure similar aspects of performance. We examine the

relationships between faultlines, conflict, and performance further using hierarchical regression analyses.

---------------------INSERT TABLE 2 ABOUT HERE --------------------

Faultlines and Conflict

We conducted hierarchical regression analyses to test our hypotheses predicting the effects of

faultline strength and distance on relationship, task, and process conflict (H1a through H2c). Step 1

includes controls (group size), step 2 includes the main effects of faultline strength and distance. As

shown in table 3, faultline strength was positively and significantly related to task conflict in work groups

as predicted by H1b. H1a and H1c, predicting that faultline strength would increase relationship and

process conflict in work groups, were not supported. Support was also found for H2c: faultline distance

increased process conflict in work groups. Faultline strength and distance explained from 1% to 10% of

the variance in conflict within workgroups.


Faultlines and Outcomes

Hypotheses H3a and H3b were partially supported by the regression analyses (see table 4).

Faultline strength was negatively related to bonuses (beta = -.006, p = n.s.) and negatively and

significantly related to stock options (beta = -.178, p<. 05). Faultline distance was also negatively

associated with bonuses (beta = -.040, p = n.s.) and negatively and significantly related to performance

ratings (beta = -.220, p<.01). Faultline strength was positively and significantly associated with negative

satisfaction (beta = .120, p<.05). Interestingly, faultline strength and distance were positively and

significantly related to performance ratings and negative satisfaction respectively.


Mediators of Faultlines effects

18

The three regression equations suggested by Baron and Kenny (1986) were performed to test

the mediating role of relationship, task, and process conflict. First, relationship conflict (mediator) was

regressed on faultlines (independent variables: strength and distance) and found to be significant for

distance (beta = -.190, p<.01). Second, distance was significantly related to the dependent variables:

Performance ratings (beta = -.220, p<.01) and negative satisfaction (beta = -.331, p<.001). And third,

the effect of distance became nonsignificant or had a substantially lower effect when the relationship

conflict was included in the regression analyses on performance and satisfaction (beta = -.111, p = n.s.;

beta = -.067, p = n.s.). Thus, the mediating role of relationship conflict between faultline distance and

performance was confirmed (H5a). However, results did not confirm hypothesis H4a, that relationship

conflict would mediate the effects of faultline strength on performance.

Task conflict (mediator) was regressed on faultlines (independent variables: strength and

distance) and found to be significant (beta = .197, p<.01 and beta = -.380, p<.001, respectively).

Strength and distance were significantly related to the dependent variables: stock options (beta = .178,

p<.05), performance ratings (beta = .248, p<.001, and beta = -.220, p<.01, respectively), negative

satisfaction (beta = .120, p<.05, and beta = -.331, p<.001, respectively), and positive satisfaction (beta

= .161, p<.05). The effects of strength and distance became nonsignificant or had a substantially lower

effect when the task conflict was included in the regression analyses on stock options (beta = -.149, p =

n.s), performance ratings (beta = .282, p< .05, and beta = -.067, p< .05, respectively), and negative

satisfaction (beta = .061, p = n.s.; beta = -.111, p = n.s., respectively). Thus, the mediating role of task

conflict between faultline strength and distance and performance was confirmed (H4b and H5ba).

Finally, process conflict (mediator) was regressed on faultlines (independent variables: strength

and distance) and found to be significant for distance (beta = .092, p<.05. Distance was significantly

related to the dependent variables: performance ratings (beta = -.220, p<.01) and negative satisfaction

(beta = -.331, p<.001). The effect of distance became nonsignificant or had a substantially lower effect

when the process conflict was included in the regression analyses on performance and satisfaction (beta

= -.167, p = n.s.; beta = -.077, p = n.s.). Thus, the mediating role of process conflict between faultline

distance and performance was confirmed (H5c). However, results did not confirm hypothesis H4c, that

process conflict would mediate the effects of faultline strength on performance.

19

Moderating Effects of Team Identity

We conducted hierarchical regression analyses to test our hypotheses predicting the moderating

effects of team identity on the relationship between faultline strength and distance and conflict (H6a

through H7c). Step 1 includes controls (group size), step 2 includes the main effects of faultline strength

and distance, and team identity, step 3 includes the two hypothesized interactions (faultline strength x

team identity; faultline distance x team identity).


Faultline strength was positively related to task conflict. In support of H6b, team identity

moderated the effect of faultline strength on task conflict: faultline strength was less positively related to

task conflict when team identity was strong in groups. In support of H7c, team identity moderated the

effects of faultline distance on process conflict: faultline distance was less positively related to process

conflict when team identity was strong in groups.

DISCUSSION

This field study examines the effects of group faultlines (i.e. faultline strength and distance) on

three types of conflict (i.e. relationship, task, and process conflict), and multiple group outcome

variables (e.g. bonuses, stock options, performance ratings, employees’ satisfaction). We also

investigate the role of team identity as a moderator of the relationship between group faultlines and

conflict. Many of our hypotheses are supported. In particular, we find support for the positive

relationship between faultline strength and task conflict. Faultline distance is positively associated with

process conflict. Furthermore, both measures of group faultlines are negatively related to various

performance outcomes. Task and process conflict mediates the relationship between group faultlines

and performance. Additionally, the relationship between faultlines and conflict is moderated by team

identity; that is if there is a strong group identity among group members, group faultlines are not likely to

promote task and process conflict in workgroups. However, if there is a weak team identity among

group members, group faultlines are likely to get activated and result in conflict.

The strengths of the current research—data collected from actual social settings—are

accompanied by potential weakness. Because we have not experimentally manipulated the causal

factors (faultline strength and distance) to which we attribute the findings, the possibility remains that our

causal logic could be reversed. Another limitation is the proper treatment of cross-level variables. We

20

recognize that a future possibility is to use the Hierarchical Linear Modeling (HLM), which will link our

group level variables to individual-level outcomes.

A major theoretical contribution of this paper is looking at group faultline as a two-dimensional

concept that includes strength and distance qualities. Specifically, in addition to faultline strength

(Thatcher, Jehn, Zanutto 2001) we conceptualize and define faultline distance as the difference between

two homogenous subgroups. We also add to the existing literature by looking at the different types of

conflict in the context of intergroup relationship.

We contribute to the research by operationalizing constructs that have been theorized about, but

not yet empirically tested (e.g., group faultlines). In particular, we develop the unique measure of

faultline distance to examine how far apart subgroupings are. We use methods from multivariate

statistical cluster analysis (e.g. Morrison, 1967; Jobson, 1992; Sharma, 1996). We refine the fau

algorithm (Thatcher, Jehn, and Zanutto, 2000) and develop the fau distance score to better understand

the effects of faultlines in workgroups. By adding the fau distance measure into the fau algorithm, we

capture not only how cleanly the group splits into two subgroups, but also how far apart two subgroups

are from each other. This combination of the fau score and the fau distance score provides useful

information that has not yet been depicted by past diversity research thus far.

To summarize, this study is fruitful because it provides the empirical test of group faultline theory

and validates our assumption about different effects of faultline strength and distance on conflict and

performance. Evidence of these different effects opens the door to future research of exploring group

faultlines in work groups. There are a number of different models to test in order to see whether

faultline strength or faultline distance is more influential. For example, when fitting a linear regression, if

we include both the faultline strength score and the faultline distance score, and the interaction between

faultline strength score and the faultline distance score, we can see whether a strong faultline has the

same effect regardless of the distance between the two subgroups.

Finally, in this study, we only look at the subgroup’s splits, while excluding the one person’s

(token) splits. Thus, future empirical research is needed to explore the differences between the effects

of the one person’s faultline splits compared to the effects of the subgroup’s splits on the group

processes and performance outcomes in a group. Research on tokens and tokenism can provide a

useful framework for this analysis. We believe that the effects of token will produce and trigger more

21

severe processes in group functioning than the effects of subgroups. The faultline algorithm that allows

calculating all possible splits, specifically, including one person splits and excluding one person splits, can

be a crucial operational tool in conducting an empirical test of these contrasting models.

22

REFERENCES

Allison, P.D. 1978. Measures of inequality. American Sociological Review, 43:865-880.

Ancona, D. & Caldwell, D. 1992. Bringing the boundary: External activity and performance in

organizational teams. Administrative Science Quarterly, 37: 634-665.

Asante, M. & Davis, A. 1985 Black and white communication: Analyzing workplace encounters.

Journal of Black Studies, 16: 77-93.

Baron, R. M., & Kenny, D. A. 1986. The moderator – mediator variable distinction in social

psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and

Social Psychology, 51: 1173-1182.

Blau, P. 1977. Inequality and composition: A primitive theory of social structure. New York: Free

Press.

Boldry, J. & Kashy, D. 1999. Intergroup Perception in Naturally Occuring Groups of Differential

Status: A Social Relations Perspective. Journal of Personality and social Psychology, 77 (6), 1200-

1212.

Byrne, D. 1971. The Attraction Paradigm. New York: Academic Press.

Caplow, T. 1956. A theory of coalitions in the triad. American Sociological Review, 21: 489-493.

Chatman, J., Polzer, J., Barsade, S. & Neale, M. 1998. Being Different Yet Feeling Similar: The

Influence of Demographic Composition and Organizational Culture on Work Processes and Outcomes.

Administrative Science Quarterly, 43: 749-780.

Clement, D. & Schiereck, J. 1973. Sex composition and group performance in a visual signal detection

task. Memory & Cognition, 1 (3): 251-255.

Freidman, R.A. & Podolny, J. 1992. Differentiation of boundary spanning roles: Labor negotiations and

implications for role conflict. Administrative Science Quarterly, 37 (1): 28-47.

Georgopolous, B. S., 1986. Organizational structure, problem-solving, and effectiveness. San

Francisco: Jossey-Bass.

Giles and Johnson, 1981. The role of language in ethnic group relations. In Turner & Giles (Eds.),

Intergroup behavior (pp. 199-272). Oxford: Basil Blackwell.

Hagendoorn, L., Drogendijk, R., Tumanov, S., & Hraba, J. 1998. Inter-ethnic preferences and ethnic

hierarchies in the Former Soviet Union. International Journal of Intercultural Relations, 22: 283-503.

23

Hoffman, L. & Maier, N. 1961, Quality and acceptance of problem solutions by members of

homogeneous and heterogeneous groups. Journal of Abnormal and Social Psychology, 62: 401-407.

Hoffman, L. 1978. The group problem-solving process. In L. Berkowitz (ed.), Group Processes: 101-

114. New York: Academic press.

Hogg, M. 1996. Social identity, self-categorization, and the small group. In E. Witte & J. Davis (Eds).

Understanding group behavior, Vol. 2: Small group processes and interpersonal relations. (pp. 227-

253). Hillsdale, NJ.

Hraba, J., Radloff, T., & Gray-Ray, P. 1999. A comparison of Black and White social distance.

Journal of Social Psychology, 139 (4): 536-539.

Jackson, S. 1992. Team composition in organizations. In S. Worchel, W. Wood, and J. Simpson

(eds.), Group Process and Productivity: 1-12. London: Sage.

James, L. R., Demaree, R. G., & Wolf, G., 1984. Estimating within-group interrater reliability with and

without response bias. Journal of Applied Psychology, 69: 85-98.

Janis, I.L. 1972. Victims of Groupthink. Boston: Houghton-Mifflin.

Jehn, K.A. 1997. A qualitative analysis of conflict types and dimensions in organizational groups.

Administrative Science Quarterly, 42: 520-557.

Jehn, K.A. 1995. A multimethod examination of the benefits and detriments of intragroup conflict.

Administrative Science Quarterly, 40, 256-282.

Jehn, K.A., Northcraft, G. & Neale, M. 1999. Why Differences Make a Difference: A Field Study of

Diversity, Conflict, and Performance in Workgroups. Administrative Science Quarterly, 44: 741-763.

Jobson, J.D. 1992. Applied Multivariate Data Analysis. Volume II: Categorical and Multivariate

Methods, New York: Springer – Verlag.

Kessler, T., Mummendey, A. 2001. Is there any scapegoat around? Determinants of intergroup

conflicts at different categorization levels. Journal of Personality and Social Psychology, 81 (6): 1090-

1102.

Kimble, C. E. & Musgrove, J. I. 1988. Dominance in arguing mixed-sex dyads: Visual dominance

patterns, talking time, and speech loudness. Journal of Research in Psychology, 22: 1-16.

Kirchmeyer, C. & Cohen, A. 1992. Multicultural groups: Their performance and reactions with

constructive conflict. Group & Organization Management, 17 (2): 153-170.

24

Komorita, S.S. & Kravitz, D.A. 1983. Coalition formation: A social psychological approach. In Paulus

(Ed.), Basic group processes: 1979-203. New York: Springer-Verlag.

Kosnik, R. 1990. Effects of board demography and directors’ incentives on corporate greenmail

decisions. Academy of Management Journal, 33: 129-150.

Lau, D. & Murnighan, J.K. 1998. Demographic diversity and faultlines: The compositional dynamics of

organizational groups. Academy of Management Review, 23 (2): 325-340.

Mack, R. and Snyder, R.S. 1957. The Analysis of Social Conflict – Toward an Overview and

Synthesis. Journal of Conflict Resolution, 1, 212-248.

Mackie, D. M., Devos, T., & Smith, E. R. 2000. Intergroup emotions: Explaining offensive action

tendencies in an intergroup context. Journal of Personality and Social Psychology, 79(4): 602-616.

Morrison, D. G. 1967. Measurement problems in cluster analysis. Management Science, 13 (2): B-

775-B-780.

Mullick, R., & Hraba, J. 2001. Ethnic attitudes in Pakistan. International Journal of Intercultural

Relations, 25 (2): 165-179.

Murnighan, J.K. 1978. Models of coalition behavior: Game theoretic, social psychological and political

perspectives. Psychological Bulletin, 85: 1130-1153.

Murnighan, J.K. & Brass, 1991. Intraorganizational Coalitions. Research on Negotiation in

Organizations, 3. Handbook of Negotiation Research, 283-306

Nemeth, C., 1986. Differential contributions of majority and minority influence processes. Psychological

Review, 93: 10-20.

O’Reilly, C. & Flatt, S. 1989. Executive team demography: Organizational innovation and firm

performance. Working paper, School of Business, University of California at Berkeley.

Pelled, L., Eisenhard, K. & Xin, K. 1999. Exploring the Black Box: An Analysis of Work Group

Diversity, Conflict, and Performance. Administrative Science Quarterly, 44: 1-28.

Plaks, J. E. & Higgins, E. T. 2000. Pragmatic use of stereotyping in teamwork: Social loafing and

compensation as a function of inferred partner-situation fit. Journal of Personality and Social

Psychology, 79(6): 962-974.

Polzer, J.T., Mannix, E.A., & Neale, M.A. 1998. Interest alignment and coalitions in multiparty

negotiation. Academy of Management Journal, 41 (1): 42-54.

25

Pool, J. 1976. Coalition formation in small groups with incomplete communication networks. Journal of

Personality and Social Psychology, 34 (1): 82-91.

Richard, O. 2000. Racial Diversity, Business Strategy, and Firm Performance: A Resource-Based

View. Academy of Management Journal, 43 (2): 164-177.

Scheider, B. 1983. An interactionist perspective on organizational effectiveness. In L. L. Cummings &

B. M. Staw (Eds.), Research on organizational behavior, Vol. 5, pp. 1-31. Greenwich, CT: JAI Press.

Sharma, S. 1996. Applied Multivariate Techniques, New York: John Wiley & Sons.

Sidanius, S., & Pratto, F. 1999. Social dominance: An intergroup theory of social hierarchy and

oppression. New York: Cambridge University Press.

Simons, T., Pelled, L.H., & Smith, K.A. 1999. Making use of difference: Diversity, debate, and

decision comprehensiveness in top management teams. Academy of Management Journal, 42 (6): 662-

673.

Steiner, I.D. 1972. Group Process and Productivity. San Diego: Academic Press.

Stevenson, W.B., Pearce, J.L. & Porter, L.W. 1985. The concept of “coalition” in organization theory

and research. Academy of Management Review, 10 (2): 256-268.

Tajfel, H. & Turner, J.C. 1986. The social identity theory of intergroup behavior. In Worchel & Austin

(Eds.), Psychology of intergroup relations (2nd ed., pp. 7-24). Chicago: Nelson-Hall.

Teachman, J.D. 1980. Analysis of population diversity. Sociological Methods and Research, 8: 341-

362.

Thatcher, S.M., Jehn, KA., and Zanutto, E., 2000. Cracks in diversity research: The effects of faultlines

on conflict and performance. In press.

Tsui, A. S., Egan, T. D., & O’Reilly, C. A. 1992. Being different: Relational demography and

organizational attachment. Administrative Science Quarterly, 37 (4): 549-577.

Turner, J. C., Hogg, M. A., Oaks, P. J., Reicher, S.D., & Wetherell, M.S. 1987. Rediscovering the

social group: A self-categorization theory. Oxford, England: Blackwell.

Tziner, A. & Eden, D. 1985. Effects of crew composition on crew performance: Does the whole equal

the sum of its parts? Journal of Applied Psychology, 70: 85-93.

Wagner, W. G., Pfeffer, J., & O’Reilly, C. A. 1984. Organizational demography and turnover in top

management groups. Administrative Science Quarterly, 29 (1): 74-92.

26

Williams, K. & O’Reilly, C. 1998. Demography and Diversity in Organizations: A Review of 40 Years

of Research. Research in Organizational Behavior, 20: 77p-140.

Wittenbaum, G. & Stasser, G. 1996. Management of information in small groups. In Nye and Brower

(Eds.), What’s social about social cognition? Social cognition research in small groups (pp. 3-28).

Thousand Oaks, CA: Sage.

Worchel, S., Lind, E.A. and Kaufman, K.H. 1975. Evaluations of group products as a function of

expectations of group longevity, outcome of competition, and publicity of evaluations. Journal of

Personality and Social Psychology, 31:1089-1097.

27

Figure 1. Research Model

Group faultlines: -Strength -Distance

Conflict: -Relationship -Task -Process

Group Performance: -Group bonus -Group stock -Group perf. rtg -Satisfaction

Group Identity

28

Appendix 1. Examples of the archival textual data

Leadership Portfolios extracts

Group A:

Develop the leadership and administrative skills of the current group of Team Leaders to enable them to handle a wider range of operational and client issues. 1. Increase spending and time for training. 2. Customer Service Manager and Account Manager will need to provide additonal team and one to one coaching. Start in second quarter of 1999. Completion of process workflows will provide the management team with an effective training tool. 1.Subject to availability of the XXXXXXX1. Start between fourth quarter of 1999 and first quarter of 2000. Improve distribution process for XXXXXX thereby helping them to gain efficiencies and cost savings while generating incremental revenue for XXXX. 1.Increase in resources for appropriate Addressing Systems and Software. 2.We will need to source for a Business Services Consultant when critical mass is achieved. Start between fourth quarter of 1999 and first quarter of 2000. Contract Review - Perform complete analysis of all XXXX sites to right size account and reduce expenses by 5% versus current billings. 1. Open XXXX XXXX position must be filled by 7/19/99. 2. Possible reduction in annual revenue of $XXXK per annum. 1. XX, XXX & XXX sites must be completed by 8/31/99. 2. XX, XXX, XXX & XXX to be completed by 9/30/99. Improve distribution process for XXXX thereby helping them gain efficiencies and cost savings while generating incremental revenue for XXXX. 1. Increase in resources for appropriate XXXX Systems and Software. 2. We will need to source a Business Services Consultant when critical mass is achieved. 4Q 1999 & 1Q 2000. Continue to grow the business at our current XXXX locations and add at least one addition location (i.e. XXXX). 1. Appropriatie resources are available. 2. RVP to communicate with XXX VP. XXXXX of 1999.

Group B

Increase skill mix of workforce. Field effected by training immediate and ongoing throughout 99. Strenghten mgmt capability Time in field an personal dev time needed for loc. Mngr ongoing In house support development training needed immediate and ongoing The speed of change in the amount and variety of information that we are required to be experts on. Have regular and frequent training forums to stay abreast of these challenges. Monthly technical conference calls. Weekly mgt staff meetings to keep all informed of information changes and requirements. Decrease in operating budgets. Re-evaluate territories be vacated by XXX. Meet with territory teams to restructure territory to reduce staff by off loading non technical work. change of skill mix of workforce A+ certify all XXXs by 6/2000 Make advanced training in networking and programming technolgy available to qualified XXXs. (currently in process) Develop challenging strategies to interest XXXs in further development and use of technical abilities. (currently in process) Changing focus and complexity of systems business. change of skill mix of

1 Certain words and phrases have been removed from the documents to protect the confidentiality of the company.

29

workforce Implement a strategy that allows service and sales to work together to provide necessary solutions in a sensible and effective manner.

30

Table 1. Faultline Strength and Faultline Distance Scores Calculated for the Four Groups.

Member Age Race Gender Tenure Function Education Fau Strength

Fau Distance

Team 1 0.8057 2.9334 1 65 1 1 26 3 5 2 37 1 1 2 3 7 3 50 1 0 26 3 4 4 36 1 1 4 3 7 5 46 1 0 1 3 7

Team 2 0.8304 2.0265 1 61 2 1 6 1 7 2 34 1 0 10 1 5 3 45 1 0 4 1 5 4 47 2 1 9 1 7 5 37 1 0 1 1 5

Team 3 0.3589 1.9091 1 42 1 0 18 2 5 2 55 1 1 3 2 2 3 40 1 0 11 2 7 4 41 1 1 16 3 7 5 49 1 0 29 4 4

Team 4 0.3572 1.2746 1 48 1 1 7 2 4 2 50 1 1 20 2 5 3 40 2 0 1 2 5 4 42 3 1 9 2 5 5 35 4 1 13 2 5

31

Table 2. Means, Standard Deviations, and Zero-Order Correlations Among Variables.

Variables M SD Correlations

1 2 3 4 5 6 7 8 9 10

1.Faultline Strength .49 .12

2.Faultline Distance 1.88 .46 .487*

3.Relationship Conflict 1.62 1.42 -.129* -

.168**

4.Task Conflict 2.63 1.81 -.026 -

.208**

.018

5.Process Conflict 2.72 1.55 -.051 .001 -.064 .282**

6.Bonuses 30218.4

0

35675.3

4

-.052 .008 .060 .223** .370**

7.Stocks 1461.88 2058.03 -

.160**

-.048 -.045 .094 .292** .837**

8.Performance Ratings 3.88 .38 .123* -.064 -.078 .008 .101 .356** .122*

9.Satisfaction (negative) 3.33 2.09 -.089 -

.177**

.455** .431** .119* .347** .219** .077

10.Satisfaction

(positive)

3.95 1.98 .073 .078 .257** .171** .309** .512** .379** .042 .313**

11.Team Identity 3.89 2.01 .032 -.026 .359** .338** .518** .559** .415** .082 .599** .682**

*p<.05; **p<.01 (two-tailed); N = 518.

Table 3. Hierarchical Regression Analyses Predicting Conflict (N = 518)

Relationship Conflict Task Conflict Process Conflict Step 1: Controls Group Size Adjusted R2 F

.037 .002 .428

.072 .002 1.609

-.057 .002 1.024

Step 2: Main Effects Faultline Strength (FauS) Faultline Distance (FauD) Change in R2 F change R2 Adjusted R2 F

-.020 -.190** .037 5.883** .038 .029 4.069**

.197** -.380*** .085 14.409*** .090 .081 10.188***

-.114 .092* .009 1.368 .012 .003 1.254

32

Table 4. Regression Analyses Predicting Performance and Satisfaction.

Bonuses Stock Options Performance Ratings

Satisfaction (negative)

Satisfaction (positive)

Step 1: Controls Group Size Adjusted R2 F

.144 .017 6.519*

.045 .002 .642

-.009 .000 .025

.159** .022 8.054**

.212*** .042 14.538***

Step 2: Main Effects Faultline Strength (FauS) Faultline Distance (FauD) Change in R2 F change R2 Adjusted R2 F

-.006 -.040 .002 .273 .022 .013 2.345

-.178* .038 .025 3.916* .027 .017 2.829*

.248*** -.220** .044 7.048** .044 .035 4.708**

.120* -.331*** .069 11.651*** .094 .085 10.637***

.161* -.088 .017 2.732 .061 .052 6.722***

*p<.05; **p<.01; ***p<.001.

Table 5. Hierarchical Regression Analyses (N = 518)

Relationship Conflict Task Conflict Process Conflict Step 1: Controls Group Size Adjusted R2 F

.037 .002 .428

.072 .002 1.609

-.057 .002 1.024

Step 2: Main Effects Faultline Strength (FauS) Faultline Distance (FauD) Team Identity (TI) Change in R2 F change R2 Adjusted R2 F

-.020 -.129* .355*** .158 19.249*** .159 .149 14.563***

.151* -.328*** .303*** .172 21.466*** .178 .167 16.581***

-.200** .189** .561*** .311 146.399*** .314 .305 35.168***

Step 3. Interactions FauS X TI FauD X TI Change in R2 F change R2 Adjusted R2 F

-.176 -.491 .017 3.152* .177 .160 10.895***

-.678* .656* .016 3.065* .194 .178 12.224***

-.008 .376 .006 1.449 .321 .307 23.996***

*p<.05; **p<.01; ***p<.001.

Documents

A Field Study of Group Faultlines, Team Identity, Conflict, and Performance in Organizational Groups