Influence of Diversity on Conflicts, Turnover, and Performance

Do Lim Choi & Yun Jeong Kim

(Chungnam National Univ.)
Introduction
Diversity issue has been treated as an important element in organizational
management in many advanced countries(Barsade, Ward et al. 2000; Pitts 2009). As
evidence, many papers on diversity were presented at 2012 American Society for
Public Administration Conference held in Las Vegas, NV. Initially, diversity meant
something to be related to demographic variables, affirmative action (AA), and equal
employment opportunity (EEO). In other word, the term was considered as something
equal to the participation of more colors and women. Therefore, the concept was
more normative. It means that more colors and women in organization are
meaningful.
Since 1980s, according as organizations have the diverse workforce, many
practitioners and scholars have had interests in workforce diversity(Wentling and
Palma-Rivas 1998; Mor-Barak 1999). The focus of diversity has been expanded from
simple demographic characteristics to belief, background, and function (Thomas Jr
1990; Slack 1997; D'Netto and Sohal 1999; Gardenswartz and Rowe 2010). Unlikely
in the past, they would like to reveal the relationship between diversity and
organizational performance(Pelled, Eisenhardt et al. 1999). It, however, is difficult to
find empirical studies that examine the relationship between diversity and
performance in the public sector, although there are some studies in the private
sector(Pitts 2005; Pitts 2009).
In the past decades, empirical studies have focused on direct relationship between
demographic characteristic variables and performance, but recently a group of
scholars tries to suggest more sophisticated model to explain diversity and
performance. For example, conflict variable emotional or task conflict can be
used as a moderate variable between diversity and performance(Pelled, Eisenhardt et
1

al. 1999).
The purpose of this study is to examine the relationships between diversity and
conflicts, and performance. All three variables are classified into two dimensions.
Diversity variable can be classified into indirectly work related and directly work
related. Conflict variable is divided into two categories: emotional and task. Finally,
performance is categorized into turnover and cognitive performance. This model is
very similar to that of Pelled and her colleagues(1999).
Theoretical Background and Hypotheses
Diversity, Emotional Conflict, and Task Conflict
Organizations are consisted with employees who have various demographic
characteristics. Employee composition in gender, age, race, education, tenure,
functional background, and so on differs according to organizations. The diversity in
demographic characteristics will influence emotional conflict. It can be explained by
the term categorization, which is a very useful way to simplify and make our
perceived organizations more predictable(Tajfel, Billig et al. 1971).
Employees in organizations are categorized by demographic characteristics and
categorization is a good means to explain people who belong to any group that shows
the same demographic characteristic. Once categorization is formed, people try to
describe the core characteristics of the group and try to compare with other groups.
Categorization can be formed by gender, age, tenure, education, originality, race,
functional background, and so on. This can be also explained by the model of
inclusion and exclusion of Mor-Barak(1999).
When categorizations interact, there is higher possibility that conflicts occur
because they have different tendency in both thinking and acting. They will share
different values, assumptions, and expectations according to categorization. Majority
group tends to have negative attitudes, discrimination and prejudice, stereotyping,
and racism to minority groups(Wentling and Palma-Rivas 1998). If diversity in
demographic variables increases, individuals in workgroup have more chances to
interact people who have different characteristics. It will cause any type of conflict in
2

the process of interaction. As discussed above, each categorization has its own
unique values, assumptions, and expectations. Therefore, the more diverse
categorization is the more conflicts happen.
According to the demographic characteristics, different conflict can be stimulated in
organization. Two types of conflict can be discussed: emotional and task
conflict(Pelled 1996; Pelled, Eisenhardt et al. 1999). First, emotional or affective
conflict means anger, resentment, and other negative feelings caused by the
interaction of different social categories in the same workgroup. Among diversity
variables, age, gender, originality, and race will contribute to the forming of emotional
conflict in process of interaction. For example, young generation will show different
way of thinking from old generation. Likewise, man and woman will put emphasis on
different values(Gilligan 1982). As a result, when they confront in the process of
work, negative feelings will be produced.
Hypothesis1: Diversity in gender, age, and originality will have stronger positive
associations with emotional conflict.
Second is task or substantive conflict, which is more related to work than to
emotion. Other demographic characteristics will bring more task conflict than
emotional conflict. Task conflict means conflicts occurring due to different ideas,
different ways to do work, or different work experience(Carrell and Mann 1995). For
example, functional background is more associated with task conflict. According to
employees, experiences that they have had will be different. There are some
employees who experience just a few functional departments and others who
experience various functional departments. They may have different belief structures
or way of thinking mechanism. When employees who have different functional
background get together to decide what to do, they may spend more time to reach an
agreement than those who have the same or similar functional background. Also,
education will have similar effect on conflict. Different education level among
employees will bring different ideas or suggestions in work. It will cause work related
conflict.
3

Hypothesis2: Diversity in education, tenure in organization, and functional

background will have stronger positive associations with task conflict.
Tenure in workgroup means the term that employees have worked in the current
workgroup. Employees who work for longer term in the same workgroup will form and
share similar values, beliefs, and expectations. They will more easily accept different
categories. As a result, emotional conflict will be weakened(O'Reilly, Caldwell et al.
1989). Likewise, the negative effect of diversity in education, tenure in organization,
and functional background on task conflict will be weakened, as tenure in the same
workgroup increases. Employees who have worked in the same workgroup have more
chances to learn from one another. Therefore, conflict relevant to task will decrease.
Hypothesis3: Group longevity will diminish the positive associations between
diversity variables and conflict in work groups.
Emotional and Task Conflict as Mediators: Links to Turnover
Turnover means rate that employees leave their organization. Without having other
alternatives, leaving current organization is not easy, because they need job to live or
support their families. In order to measure the rate of turnover, employees opinion,
whether they have any intention to leave or not, has been asked in many previous
empirical studies.
Emotional conflict caused by diversity will increase turnover. Because emotional
conflict is relevant to anger, being isolated, and hatred, employees who feel
emotional conflict in current organization will have turnover intention. In contrast to
emotional conflict, task conflict caused by diversity will not be significantly associated
with turnover intention. Task conflict is not based on category bias, prejudice, or
personal negative feeling.
Hypothesis4: Emotional conflict will be positively associated with turnover
intention.
4

Hypothesis5: Task conflict will not be a significant variable to influence turnover

intention.
Task and Emotional Conflict as Mediators: Links to Performance
Performance has been interpreted as a term to have various meanings. It may be
interchangeably used with productivity, efficiency, consequence, result, effectiveness,
output, outcome, and so on. Simply, performance implies consequences or results
that organizations or employees accomplish in a certain time period.
Performance can be measured in two ways. First, it can be measured by the
secondary data. Through real numbers, goal achievement, efficiency, output, and so
on can be evaluated. Second, it can be measured by employees cognitive judgment.
In other words, performance level can be asked to employees. For examples,
following questions can be employed to measure organizational performance. How
much productive is it? How much efficient is it?
Task conflict is able to be evaluated as a process seeking better alternatives or
making better decisions to achieve organizational goals. Therefore, generally
speaking, task conflict is necessary to improve organizational performance. However,
emotional conflict will not be a significant variable to explain organizational
performance in organization. Rather, it may play a role to decrease organizational
performance.
Hypothesis6: Task conflict will be positively associated with performance.
Hypothesis7: Emotional conflict will decrease organizational performance.
Methodology
Research Framework
Research framework will be briefly explained here, although it is done in theoretical
background in details. As in Figure 1, three dimensions are identified: independent,
mediator, and dependent variables. First, there are two types of diversity variables:
indirectly job related and directly job related variables. The former includes gender,
age, and originality, the latter includes functional background, tenure, and education.
5

Second, there are two types of conflict variables: emotional and task conflicts. Third,
turnover and cognitive performance are used as dependent variables.

Less Task Related Variables:

Gender, Age, Originality

Emotional
Conflict

Turnover

Task Conflict

Performance

Longevity

More Task Related Variables:

Functional Background.
Education, Tenure

<Figure 1> Research Framework

In the relationship between diversity and conflict, diversity variables that are not
directly related to task will be positively associated with emotional conflict. Diversity
variables directly related to task will be positively associated with task conflict. And
longevity will alleviate the effect of diversity variables on conflict variables. In the
relationship between conflict and performance, emotional conflict will be negatively
associated with employees turnover and task conflict will be positively associated
with cognitive performance. And emotional conflict will have negative effect to
performance and task conflict will also have negative effect to employees turnover.
Measurement and Research Methods
Initially, it was planned to ask direct values on diversity, but public servants felt
uncomfortable with it. They seemed to worry that somebody might know about who is
who. Also, they didnt know about exact information on colleagues age, tenure, and
educational institution that coworkers graduate. Many respondents in pilot study
didnt answer to questions such as final educational institution, their hometown, age,
and so on. So, instead of asking direct information or values, the degree of diversity
was asked. For example, what is the level of diversity in age? Each question has five
6

point Likert scale.

To test hypotheses suggested in this paper, a few quantitative analysis methods
were implemented. For basic analyses, frequency, mean, and correlation analysis
were implemented. And for hypothesis test, regression and path analysis were
employed.
Findings
Basic Statistics
This section reports basic information and statistics for further research such as
response rate, demographic characteristics, and Cronbach alpha.

Response Rate
A metropolitan city was selected for this research and respondents are those who
are working in the metropolitan city hall. Total 172 questionnaires were distributed in
March 2012 and 152 collected. It showed a response rate of 88.4%. All
questionnaires collected were used in final analysis.

Descriptive Statistics
Detailed descriptive statistics are reported in Table 1. As shown in Table 1, 68.4
percent of the respondents were male and 31.6 percent female. Most of the
employees were in the 40s (62.5%). Regarding education, 61.8 percent were
university graduates and 21.7 percent had graduate degrees. Of the respondents,
46.6 percent had worked for 11 to 20 years and 37.7 percent for 21 to 30 years. In
functional background, most of respondents (82.1%) had experienced less than 11
work types. Finally, most of respondents were from Daejeon(44.3%) and
Chungnam(30.9%).
<Table 1> Demographic Characteristics
Classification
Gender

Frequency(total no.: 152)*

Number
Percentage

Male
Female

104
48
7

68.4
31.6

Age

Education

Work years

Functional
background

Originality

20s
30s
40s
Older than 50
High school
Community college
University
Graduate school
1 to 10 years
11 to 20 years
21 to 30 years
More than 30 years
Less than 6
6 to 10
11 to 15
16 to 20
More than 20
Daejeon
Chungnam
Chungbuk

1
20
95
36
15
10
94
33
13
68
55
10
64
46
11
7
6
66
46
20

0.6
13.2
62.5
23.7
9.9
6.6
61.8
21.7
8.9
46.6
37.7
6.8
47.8
34.3
8.2
5.2
4.5
44.3
30.9
13.4

others
17
11.6
*: Each demographic characteristic may have a frequency missing which indicates
the number of respondents that did not mark that question.

Credibility and Correlation

Table 2shows means, standard deviation, Cronbach alph, and correlation
coefficients. Through this table, one can obtain basic information about the thoughts
of respondents for main variables and to a degree guess the results. A Cronbach
alpha was used to measure the unidimensionality of multiple questions. All variables
show reliability coefficients of 0.7 or higher which confirms the unidimensionality of
variables. Among four questions to measure emotional conflict, one question was
removed, because respondents didnt recognize it as the same dimension.
Regarding correlation, all variables are statistically significant with each other.
Emotional conflict variable is negatively correlated with task conflict and performance
and positively with turnover. Task conflict variable is negatively correlated with
turnover and positively with performance. Finally, turnover variable is negatively
correlated with performance.

<Table 2> Means, Standard Deviation, Credibility, and Correlation (n=152)

Variables

Emotional

Emotional

Correlation coefficient
Task
Turnover Performance

1.0

conflict

Mean

Std.

0.7

0.

0.7

0.

2
0.9

7
0.

0
0.7

8
0.

2.15

Task conflict

-0.57***

1.0

3.90

Turnover

0.27***

-0.19**

1.0

Performance

-0.49***

0.71***

-0.22***

2.96
1.0

3.77

**: significant at 0.05 ***: significant at 0.01

Respondents recognize that task conflict is higher than emotional conflict. Almost
half of respondents have thought of turnover and they have positive mind toward their
performance.
Hypothesis Test

Less work-related variables and emotional conflict

Hypothesis1: Diversity in gender, age, and originality will have stronger positive
associations with emotional conflict.
As shown in table 3, gender and originality diversity variables were associated with
emotional conflict, but in the opposite direction. Gender diversity was a significant
variable to increase emotional conflict, but originality diversity was negatively
associated with emotional conflict. Maybe, the samples of this study are
inappropriate to measure the effect of originality diversity, because most of
respondents are from Daejeon(44.3%) and Chungchung Province(44.3%). Further
research will be necessary in the future. And age diversity was not a variable to
influence emotional conflict. Thus, hypothesis 1 would be partially accepted.
<Table 3> Relationships between less work-related variables and emotional conflict
Independent variables

Dependent variable: Emotional conflict

9

Beta
Standard error
Gender diversity
.16730
.30538
Age diversity
-.06301
.07082
Originality diversity
-.23185
.06653
F Value: 5.20
R-Square: 0.0977
Adj R-Square: 0.0789
**: significant at 0.05 ***: significant at 0.01

t
2.04**
-.78
-2.86***

More work-related variables and task conflict

Hypothesis2: Diversity in education, tenure in organization, and functional
background will have stronger positive associations with task conflict.
As shown in table 4, functional background diversity and tenure diversity in public
service were statistically associated with task conflict. Both variables are positively
associated. It means that task conflict increases when employees have diverse
functional background experience. Also, task conflict increases when employees have
diverse tenure. And educational diversity was not a variable to affect task conflict.
Thus, hypothesis 2 would be partially accepted.
<Table 4> Relationships between more work-related variables and task conflict
Independent variables

Dependent variable: Task conflict

Beta
Standard error
t
Educational diversity
.01933
.36497
.24
Functional background diversity
.20855
.07079
2.51**
Tenure diversity in public service
.18106
.08127
2.15**
F Value: 5.28
R-Square: 0.1004
Adj R-Square: 0.0814
**: significant at 0.05 ***: significant at 0.01

Effect on group longevity between diversity variables and conflicts

Hypothesis3: Group longevity will diminish the positive associations between
diversity variables and conflicts in work groups.
Group longevity was not a significant variable to mediate the relationships between
diversity variables and conflicts in work groups. In another word, there was no
difference in the relationships between diversity variables and conflicts according to
group longevity.

Relationship between conflicts and turnover

10

Hypothesis4: Emotional conflict will be positively associated with turnover

intention.
Hypothesis5: Task conflict will not be a significant variable to influence turnover
intention.
As shown in table 5, emotional conflict was positively associated with turnover
intention. It means that employees have intention to leave current organization when
they feel more emotional conflict. Thus, hypothesis 4 would be accepted. Task
conflict was not a significant variable to influence turnover intention. Thus, hypothesis
5 would also be accepted.
<Table 5> Relationships between emotional conflict and performance
Independent variables

Dependent variable: Turnover

Beta
Standard error
t
Emotional conflict
.23721
.12691
2.44**
Task conflict
-.05803
.12233
-0.60
F Value: 5.81
R-Square: 0.0751
Adj R-Square: 0.0622
**: significant at 0.05

Relationship between conflicts and performance

Hypothesis6: Task conflict will be positively associated with performance.
Hypothesis7: Emotional conflict will decrease organizational performance.
As shown in table 6, both emotional and task conflict were statistically associated
with performance variable, but in the opposite direction. Like in hypothesis 6, the
higher task conflict is, the greater organizational performance is. On the contrary, the
higher emotional conflict is, the lower organizational performance is. Thus,
hypothesis 6 and 7 would be accepted.
<Table 6> Relationships between emotional conflict and performance
Independent variables

Dependent variable: Performance

Beta
Standard error
t
Emotional conflict
-.16174
.07293
-2.29**
Task conflict
.61428
.07040
8.71***
F Value: 5.81
R-Square: 0.0751
Adj R-Square: 0.0622
**: significant at 0.05 ***: significant at 0.01
11

Implications and Conclusion

In two aspects, the usefulness of this study can be explained: theoretical and
practical. First is a theoretical aspect. This research is the primary study in Korea,
especially in the public sector. This will be a starting point to discuss about the effect
of diversity variables on various result variables.
Second is a practical aspect. First, organizations should start to consider gender
composition as a variable to affect conflict, satisfaction, and performance. It doesnt
mean that there is the best composition in gender. Because gender composition
causes emotional conflict, organizations or leaders ought to make more efforts to
reduce the level of conflict. Second, emotional conflict is a variable to explain the
opinion toward turnover. Organizations should pay more attention to workgroups that
have a few females with most males. Third, task conflict is helpful to increase
organizational performance, so organizations should try to maintain the appropriate
level of task conflict in workgroups considering various function experience and
tenure in public service. Fourth, emotional conflict is harmful for organizational
performance.

12

References
Barsade, S. G., A. J. Ward, et al. (2000). "To Your Heart's Content: A Model of Affective Diversity in
Top Management Teams." Administrative Science Quarterly 45(4): 802-836.
Carrell, M. R. and E. E. Mann (1995). "Defining Workforce Diversity in Public Sector Organizations."
Public Personnel Management 24(1): 99-112.
D'Netto, B. and A. S. Sohal (1999). "Human resource practices and workforce diversity: an
empirical assessment." International journal of Manpower. MCB university Press 20(8):
530-547.
Gardenswartz, L. and A. Rowe (2010). Managing Diversity: A Complete Desk Reference & Planning
Guide. SHRM, Society For Human Resource.
Gilligan, C. (1982). "In a Different Voice: Psychological Theory and Women's Development." Harvard
University Press.
Mor-Barak, M. E. (1999). "Beyond Affirmative Action: Toward a Model of Diversity and Organizational
Inclusion." Administration in Social Work 23(3-4): 47-68.
O'Reilly, C. A., D. F. Caldwell, et al. (1989). "Work Group Demography, Social Integration, and
Turnover." Administrative Science Quarterly 34(1): 21-37.
Pelled, L. H. (1996). "Demographic Diversity, Conflict, and Work Group Outcomes: An Intervening
Process Theory." Organization Science 7(6): 615-631.
Pelled, L. H., K. M. Eisenhardt, et al. (1999). "Exploring the Black Box: An Analysis of Work Group
Diversity, Conflict, and Performance." Administrative Science Quarterly 44(1): 1-28.
Pitts, D. W. (2005). "Diversity, Representation, and Performance: Evidence about Race and Ethnicity
in Public Organizations." Journal of Public Administration Research and Theory 15(4): 615631.
Pitts, D. W. (2009). "Diversity Management, Job Satisfaction, and Performance: Evidence from U.S.
Federal Agencies." Public Administration Review 69(2): 328-338.
Slack, J. D. (1997). "FROM AFFIRMATIVE ACTION TO FULL SPECTRUM DIVERSITY IN THE AMERICAN
WORKPLACE: SHIFTING THE ORGANIZATIONAL PARADIGM." Review of Public Personnel
Administration. 17(4): 75-87.
Tajfel, H., M. G. Billig, et al. (1971). "Social categorization and intergroup behavior." European
Journal of Social Psychology 1: 149-177.
Thomas Jr, R. R. (1990). "From Affirmative Action to Affirming Diversity." Harvard Business Review
68(2): 107-117.
Wentling, R. M. and N. Palma-Rivas (1998). "Current status and future trends of diversity initiatives
in the workplace: Diversity experts' perspective." Human Resource Development Quarterly
9(3): 235-253.

13