NF

.
4;. I
5‘. A...
$1

{3.53.0

7
(3 d .
. has.
5 . 3‘. A.
3." 8.”...ku
\‘rv ’11}! ‘ I; I
8.. in?

shun.

5:”... I
~ I
g

. I .
Mum. 4..le
v. I t:
t 2.340.. t

‘Wme!

3:.
It I’—

 

 

. 512‘ .
. . ; f c
J 25:} . Tu .

Michigan State

{Mﬂé UUHAHY
University

20%

 

This is to certify that the
dissertation entitled

THE EFFECT OF COWORKER SUPPORT ON A
WORKER’S STRESS:

THE MEDIATING EFFECTS OF PERCEIVED JOB
CHARACTERISTICS

presented by

KYOKO KATO

has been accepted towards fulﬁllment
of the requirements for the

Ph. D. degree in Labor and Industrial Relations

 

 

2832;;

 

Major Professt Signature
Fabhwwa 13’, Zea—2

Date

MSU is an affirmative-action. equal-opportunity employer

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:/ProjIAcc&Pres/ClRC/DateDue‘Indd

 

THE EFFECT OF COWORKER SUPPORT ON A WORKER’S STRESS:
THE MEDIATING EFFECTS OF PERCEIVED JOB CHARACTERISTICS

By

Kyoko Kato

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY

School of Labor and Industrial Relations

2008

ABSTRACT

THE EFFECT OF COWORKER SUPPORT ON A WORKER’S STRESS:
THE MEDIATING EFFECTS OF PERCEIVED JOB CHARACTERISTICS

By
Kyoko Kato

Increased demands on workers due to competitive pressures have resulted in high
levels of workplace stress (Green, 2001; Hoel, Sparks, & Cooper, 2002; Landsbergis,
Cahill, & Schnall, I999; NIOSH, I999) . This can cause work-family conﬂict, stress-
related illness, and absenteeism, resulting in reduced productivity, organizational
performance and competitiveness, as well as increased costs for compensation and legal
challenges (Thomas & Ganster, 1995; Wells, I982). Social support from the work
environment has been considered one of the critical resources that can reduce the
negative effect of workplace stressors; however research has failed to identify the
mechanism through which this occurs (Carlson & Perrewe, 1999). The purpose of this
dissertation is therefore to begin to explain how social support, more speciﬁcally
coworker support, inﬂuences an individual’s stress level by introducing perceived job
characteristics as mediators.

The proposed model considers perceptions of job characteristics as socially
constructed through coworker support; these perceptions, in turn, become critical
determinants of an individual’s stress. This dissertation explores the possible antecedents
of coworker support, that is, demographic similarity in terms of age, gender, and ethnic
background. It is assumed that such similarity increases a coworker’s sense of
identiﬁcation and elicits more cooperative behavior. The proposed model is developed

based on an integration of two seminal theories: the Social Information Processing (SIP)

approach developed by Salancik and Pfeffer ( l 978) and the Demand-Control model of
job characteristics developed by Karasek (1979).

Using nationally representative data collected by the Families and Work Institute,
this dissertation examines the mediating effect of perceived job characteristics, predicting
that the direct effect of coworker support on stress is supplemented by effects of
individuals’ perceptions of two job characteristics: job demands and job control.

The empirical results generally support the notion that job characteristics play a
mediating role for the relationship between coworker support and stress, but they did not
conﬁrm a good theoretical ﬁt of the proposed model. Recommendations for ﬁiture
research, theoretical contributions, practical implications, and overall conclusions are

discussed.

Copyright by
KYOKO KATO
2008

To my husband and parents

ACKNOWLEDGEMENTS

Completing a doctoral study is a long journey that never occurs in isolation. First
and foremost, I want to express my sincere appreciation to my husband, Eiji, who has
given me the strength to pursue the journey of the degree. He has not only been a loving
and supportive husband, but also been an academic mentor to help me get through the
process. He has demonstrated the discipline of academic work, which kept my
motivation high when this dissertation project seemed to stall. I would also like to thank
my parents and my husband’s parents for their consistent love, trust and support
throughout this long road. I would like to dedicate this dissertation to my husband and
my parents as an indication of my heartfelt acknowledgement.

My advisor and committee chair, Dr. Peter Berg, was essential to my success in
completing this research study. He was the person who encouraged me to study at the Ph.
D. level and inspired my interest in academic research. I feel very fortunate to have had
his guidance throughout the process. I also would like to thank my committee members,
Dr. Rich Block, Dr. Mark Roehling, and Dr. Linn Van Dyne, whose constant
encouragement, faith in my abilities, and constructive feedback that helped me become
more conﬁdent in my identity as a researcher. Through your role modeling, I commit
myself to be a mentor to others in their professional development process.

I would also like to thank colleagues and friends who helped to shape my
professional development and higher education practice. Although there are too many to
list, I must acknowledge my friend, Yo Nagai, for his offering support to check my
progress weekly, critique my writing, and encourage me keeping up my work. I also

thank my colleagues and friends, Kaumudi Misra and Russ Ormiston, for their support

vi

and encouragement. They deﬁnitely made my academic life more enjoyable by
constantly cheering me up. Lastly, but certainly not least, I must also acknowledge the
supportive colleagues, faculty members, and staff within the School of Labor and
Industrial Relations. The completion of such a demanding project could not have been
accomplished without the assistance of those individuals who served as critical sources of

support and guidance.

vii

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................... x
LIST OF FIGURES ........................................................................................................... xi
CHAPTER I: INTRODUCTION ........................................................................................ 1
CHAPTER 2: LITERATURE REVIEW ........................................................................... I 1
Stress .............................................................................................................................. l 1
Job Characteristics — The Demand-Control Model ........................................................ 13
Theoretical background .............................................................................................. I 3

The model ................................................................................................................... 14
Social Support ................................................................................................................ 18
Social Information Processing (SIP) Approach ............................................................. 23
The concept ................................................................................................................. 23
Demographic diversity/similarity ............................................................................... 26
CHAPTER 3: MODEL DEVELOPMENT ....................................................................... 30
CHAPTER 4: HYPOTHESES ........................................................................................... 37
Coworker support and stress .......................................................................................... 38
Coworker support and Job Characteristics ..................................................................... 39
The effects of Job Characteristics on stress ................................................................... 4]
Demographic similarity as an antecedents of coworker support .................................... 43
CHAPTER 5: METHODS ................................................................................................. 44
Sample ............................................................................................................................ 44
Development of measures .............................................................................................. 44
Measures ......................................................................................................................... 46
Stress ........................................................................................................................... 46

C oworker support ....................................................................................................... 46

Job characteristics ...................................................................................................... 47
Demographic similarity .............................................................................................. 48
Control Variables ....................................................................................................... 48
Analytic Strategies ......................................................................................................... 5]
Multiple Regression for the mediating effects ............................................................ 52
Structural Equation Modeling (SEM) ........................................................................ 53

C onﬁrmatory Factor Analysis (C FA )and Path Analysis ............................................ 54
Model Fit .................................................................................................................... 55

F it Indices‘ for C FA ..................................................................................................... 57

viii

CHAPTER 6: RESULTS ................................................................................................... 59

Descriptive Statistics ...................................................................................................... 59
Correlation Matrix .......................................................................................................... 60
Results of the Multiple Regression Analyses ................................................................. 69
Job characteristics as Mediators ................................................................................ 69
Demographic Similarity ............................................................................................. 77
Results of the Structural Equation Modeling ................................................................. 78
Model Identiﬁcation ................................................................................................... 78
Fit Indices ................................................................................................................... 79
CHAPTER 7: DISCUSSION ............................................................................................. 82
Discussion of the Results ............................................................................................... 82
Reason for the Poor Fit .............................................................................................. 84
Future Research .............................................................................................................. 87
Contribution of the Study ............................................................................................... 88
Implications .................................................................................................................... 90
Conclusion ...................................................................................................................... 92
APPENDICES ................................................................................................................... 93
APPENDIX A
VARIABLE DESCRIPTIONS AND METHODS OF CONSTRUCTION ...................... 94
APPENDIX B

RESULTS OF THE ROTATED PRINCIPAL COMPONENT FACTOR ANALYSIS ..98

REFERENCES ................................................................................................................ 100

LIST OF TABLES

Table 1: Descriptive Statistics of All Variables for the Analyses ..................................... 59
Table 2: Correlation Matrix ............................................................................................... 62
Table 3: Regression Results of Stress on Coworker Support ............................................ 71

Table 4: Regression Results of three Job Characteristics (mediators) on Coworker

Support ............................................................................................................................... 73
Table 5: Regression Results of Stress on Coworker Support, Job Characteristics, and both
Coworker Support and Job Characteristics ....................................................................... 74
Table 6: Results of the Sobel Test for Mediating Effects of Job Characteristics .............. 76
Table 7: Regression Results of Demographic Similarity ................................................... 78

LIST OF FIGURES

Figure l: The Proposed Model ...................................................................................... 4, 30
Figure 2: Karasek’s Demand-Control Model and Proposed Outcomes ............................. 15
Figure 3: The Theoretical Paths ........................................................................................ 37
Figure 4: The First/Final Model of Job Characteristics ..................................................... 58
Figure 5: The Alternative Model of Job Characteristics .................................................... 58
Figure 6: Mediating Paths .................................................................................................. 76
Figure 7: The Results of the SEM analyses for the Proposed Model ................................ 81

xi

CHAPTER 1: INTRODUCTION

According to a report by the International Labour Organization (ILO),
approximately 30 percent of the workforce in developed countries suffer from work
related stress, and the ﬁgure appears to be increasing (Hoel, Sparks, & Cooper, 2002).
This increase of stress in the workplace can be attributed to rapid changes in the current
work environment, such as the increase of global competition, the introduction of new
forms of work organization, and changes in the demographic composition of the
workforce (Barker, 1993; Kirkman & Rosen, I999; Liden, Wayne, & Bradway, 1997;
NIOSH, 2002). Global competition has contributed to downsizing and restructuring,
which increases pressures on people at work (Hoel et al., 2002). High performance
workplaces are associated with more worker discretion and involvement in decision
making (Appelbaum, Bailey, Berg, & Kalleberg, 2000), as this work environment tends
to result in work intensiﬁcation (Francis Green, 2001, 2002; F. Green & McIntosh, 2001;
Landsbergis, Cahill, & Schnall, I999; NIOSH, 1999; C. S. Smith, Tisak, Hahn, &
Schmieder, I997; Zeytinoglu, Denton, Davies, Baumann, & Boos, 2007). In addition,
demographic changes within the workforce, such as increases in female participation,
immigrant workers and non-standard work arrangements, also heighten the vulnerability
of particular workers.

Previous research has indicated that stress is associated with negative
consequences for individuals, such as (I) the quality of the working experience (e.g., job
satisfaction, self-esteem, and commitment), (2) general well-being (e.g., life satisfaction
and happiness), (3) physical health outcomes (e.g., heart disease), and (4) problematic

behavior (e.g.. alcoholism and drug abuse) (Cary L. Cooper, Dewe, & O'Driscoll, 200];

Wells, 1982). Research also shows that having highly stressed workers can be costly for
organizations, since these stressed workers are more likely to be involved in legal
challenges and compensation for stress-related illness and accidents (L. T. Thomas &
Ganster, 1995). Due to such negative outcomes for both individual workers and their
organizations, investigating the factors that affect the level of an individual’s stress is an
important area of stress research.

Social support has emerged as one of the critical resources that can affect
individuals’ stress levels (Dawn S Carlson & Perrewe, 1999; L. T. Thomas & Ganster,
1995). Since the 19805, a substantial amount of research on the beneﬁcial effects of
social support has appeared in the social science literature (Ducharme & Martin, 2000;
House, l981b; Kaufmann & Beehr, 1986; Parasuraman, Greenhaus, & Granrose, 1992).
Social support is conceptualized as a ﬂow of communication between people involving
emotional concern, caring, and informational/instrumental help (Daniels & Guppy, I994;
Seers, McGee, Serey, & Graen, 1983). This theoretical concept rests on the assertion that
social relations in the workplace may make a key contribution to employees’ job
satisfaction, productivity, and well-being (Dawn 8. Carlson, 1999; Ducharrne & Martin,
2000).

A number of recent studies have investigated the relationship between a
supportive social environment and stress (Terry A. Beehr, Jex, Stacy, & Murray, 2000;
Dawn S. Carlson, Kacmar, & Williams, 2000; Ducharrne & Martin, 2000; Parasuraman
et al., 1992; Searle, Bright, & Bochner, 2001). Researchers have suggested that there are
two dominant hypotheses that describe the effect of social support (Cohen & Wills, I985;

Daniels & Guppy, 1994;.lex, I998). The ﬁrst hypothesis, called a direct/main effect of

social support on the outcome, proposes that a high level of social support may directly
reduce levels of stress (.lex, I998). The second hypothesis, known as the
moderating/buffering hypothesis, suggests that social support interacts with stressors such
that social support moderates/buffers the effect of stressors upon the outcomes. In other
words, the moderating/buffering hypothesis proposes that the strength of the relationship
between stressors and the outcomes depends on the level of social support. It means that
when social support is high, the relationship between stressors and their outcomes is
relatively weak. In contrast, if social support is low, the relationship between stressors
and their outcomes becomes stronger.

While many studies have found a signiﬁcant direct effect of social support on
various outcomes including stress, there is considerable debate and dispute over the
moderating/buffering effect of social support given inconsistent empirical ﬁndings of
such an effect (Parasuraman et al., 1992; Yang & Carayon, 1995; Carlson & Perrewe,
1999). In fact, many studies have been plagued by methodological problems such as
measurement unreliability and low statistical power (Cohen & Wills, 1985; Ganster et al.,
1986; Beehr et al., 2000). A more conceptual problem is, however, the lack of theoretical
underpinnings for the moderating/buffering effect of social support. A theoretical
rationale that explains the effect of social support on a stressor-stress relationship remains
an unanswered question (Carlson & Perrewe, 1999). Researchers have suggested that
discussions of the direct and moderating/buffering effects of social support are oﬁen
empirically-driven and their hypotheses are theoretically weak in explaining the

relationship of social support to its outcomes (Cary L. Cooper et al., 2001).

Instead of using a moderating/buffering hypothesis of social support, this
dissertation proposes a different mechanism; it explains the relationship of social support
to an individual’s stress by introducing perceived job characteristics as mediators. In
other words, this dissertation tests the model shown in Figure 1 below. in which the effect
of social support on stress is mediated by the effect of an individual’s perceptions of two
job characteristics: job demands and job control. Job demands are referred to as the
amount of workload that individuals perceive regarding their job (Robert A Karasek,
1979). Job control, on the other hand, represents the potential control that workers have
over their tasks during a working day (Karasek, 1979). In this proposed model, job
control consists of two dimensions: autonomy and skill discretion. Autonomy is the
authority that workers possess to make decisions on the job, and skill discretion is the

breadth of skills used by the worker on the job (R. Karasek & Theorell, 1990).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Perceived Job
Demographic Social Support Characteristics
Similarity
Job Demands
Gender /1 Support N } individual’s
Ethnic Job Control Stress
background \A - Autonomy
- Skill ,
Discretion

 

 

 

 

 

 

 

 

 

 

 

 

 

\ 1W4
V

Social Information Processing (SIP) I Demand — Control (D-C) model ]

 

 

 

 

Figure 1: The Proposed Model

The uniqueness of the proposed model can be described in two ways. First, rather
than seeing social support as a moderator/buffer for the relationship between stressor and
stress, the proposed model considers social support as an antecedent that affects the level
of stressors, which are job characteristics in this model. Second, rather than emphasizing
individuals’ predispositions, the model focuses on social construction of information that
may determine individuals’ perceptions of job characteristics. In short, the model
proposes that such perceptions are constructed by social cues, such as social support,
which individuals receive from their work environment. The perceived job
characteristics in turn affect the level of an individual’s stress.

As shown in the ﬁgure above, the notion that an individual’s perception is socially
constructed comes from the Social Information Processing (SIP) approach developed by
Salancik and Pfeffer (1978); this approach suggests that individuals’ perceptions are
inﬂuenced by the social environment or by the network of social and informational
relationships. The theoretical relationship between perceived job characteristics and an
individual’s stress, on the other hand, is explained by Karasek’s (I979) Demand-Control
(D-C) model of job characteristics. By integrating two well-developed and well-tested
theories, this dissertation provides a theoretical foundation for examining the impact of
social support on an individual’s stress.

Social support studies have been dominated by the examination of supervisor
support, but they have paid far less attention to the potential effect of another type of
social support in the workplace, namely coworker support (Ducharme & Martin, 2000).
As articulated in Figure I, this dissertation focuses only on coworker support. The

signiﬁcance of coworkers’ inﬂuence on individuals, due especially to the frequency of

interactions, has been continuously increasing because of the globalization of business,
the introduction of new work practices, and demographic changes of the workforce,
(Barker, 1993; Kirkman & Rosen, I999; Liden et al., 1997). For instance, the
globalization of business requires ﬂatter organizational structures for the purpose of
ﬂexibility; in such workplaces, employees are frequently asked to cooperate with their
colleagues across boundaries, such as functional areas and divisions, to secure
cooperation from individuals over whom they have no hierarchical control (M. Williams,
2001 ).

In addition, the introduction of a self-directed team places more collective
responsibilities on individual workers for the processes and outcomes of their work,
including situations where scheduling and production goals are determined by consensus
among team members (Parker, Wall, & Cordery, 2001). Such team-based work requires
more interaction among coworkers than supervisors in order to coordinate tasks smoothly.

Furthermore, workforce diversity, due to changes in the labor force such as the
change of family composition (e.g., single parents), the increase of minorities and
females in the labor force, and an aging society, represents a great challenge for workers
(Hodson, 2005). For example, they have to make an extra effort to communicate with
others to ﬁll the gap between cultural/ethnic backgrounds and generations. In addition,
many organizations use non-standard employment arrangements (e.g., hiring part-time,
temporary, and contract workers) to cope with environmental turbulence (Ang, Van Dyne,
& Begley, 2003; Davis-Blake, Broschak, & George. 2003; Kalleberg, Reskin, & Hudson,
2000). The use of non-standard workers has created a blended workforce, where workers

in both standard and non-standard arrangements work side-by-side doing the same tasks,

sometimes even on the same work teams (Davis-Blake et al., 2003; Pearce, 1993; Rogers,
2000; V. Smith, 2001). Under these circumstances, more frequent communication across
diverse work groups is necessary for cooperative work (Hodson, 2005).

If workers are required to interact more with their coworkers, such interactions
play a signiﬁcant role in affecting their psychological and physical states. In this regard,
positive interactions, such as social support from coworkers, become a critical issue in
the workplace. Investigating the effect of coworkers is therefore an increasingly
important component of workplace studies (Hodson, 2005). This dissertation
differentiates coworker support from that of supervisors in order to articulate their
speciﬁc effects on an individual’s stress level.

In addition. this dissertation explores a possible antecedent of social support:
demographic similarity (see Figure 1 above). Demographic similarity in this model
refers to similarity in age, gender, and ethnic background of coworkers. Previous studies
have found that demographic similarity has a signiﬁcant effect on the level of social
support by increasing workers’ sense of identiﬁcation and by eliciting more cooperative
behavior (Hodson, 2005). In short, this dissertation not only focuses on the relationship
among coworker support, job characteristics, and stress, but also focuses on the
relationship of coworker support with its antecedent, demographic similarity. Identifying
an antecedent of social support may strengthen the theoretical implications of the
proposed model.

This dissertation uses the 2002 National Study of Changing Workforce (N SCW),
a survey conducted by the Families and Work Institute. This data set consists of a

representative sample of 3,500 employed adults. Using the method suggested by Baron

and Kenny (I986 ), this dissertation primarily examines the mediating effect of job
characteristics on the relationship between coworker support and an individual’s stress
level. In addition, Structural Equation Modeling (SEM) is used to investigate the
theoretical ﬁt of the proposed model, including demographic similarity, coworker support,
job characteristics and stress.

This dissertation has several conceptual and methodological strengths. First, this
dissertation attempts to explain ‘how’ and ‘why’ social support, more speciﬁcally
coworker support, reduces individuals’ stress levels, which has not yet been explained
(Carlson & Perrewe, 1999). By introducingjob characteristics as mediators, the linkage
between social support and stress can be explained by two existing theories: the Social
Information Processing approach and the Demand-Control model. The integration of
these two theories may provide a solution for the debate over the effect of social support
on various psychological outcomes.

Second, this dissertation challenges some criticisms of job characteristics studies
that have failed to take into account factors that constrain individuals’ perception of job
characteristics (Cappelli & Sherer, I991; Clegg, 1984; Morgeson & Campion, 2003;
Parker et al., 2001 ). By introducing social support as a contextual inﬂuence, this
dissertation proposes that individuals’ perception of job characteristics is socially
constructed by such inﬂuence. Although research interest in job characteristics appears
to have declining in recent years, this dissertation revitalizes job characteristics research
by broadening the job design/characteristics paradigm (Morgeson & Campion, 2003).

Third, this dissertation provides practical implications, especially for coworker

relationships in a workplace. For example. empirical ﬁndings from this dissertation may

support organizational attempts to promote worker integration on a social level, such as
company-sponsored parties and other social activities. Such activities may be even more
beneﬁcial when combined with efforts aimed at building effective work teams that
enhance workers’ opportunities to provide on-the-job assistance, advice, and information
to their coworkers (Ducharme & Martin, 2000). Furthermore, arranging a
structure/system in the workplace that promotes interactions among coworkers may
facilitate workers’ access to positive aspects of practical support. Such a system may be
particularly beneﬁcial in a team setting when work demands exceed individuals’ abilities
to fulﬁll their job requirements. In other words, these systematic efforts to promote both
instrumental and affective networks among coworkers will enhance workers’ affect and
may consequently improve job performance.

This dissertation provides another critical implication. Traditionally, studies of
job characteristics have been analyzed using the ‘need-satisfaction paradigm,’ which
emphasizes individuals’ predispositions for explaining the effect of job characteristics
(Morgeson & Campion, 2003). Consequently, these studies have looked to potentially
expensive and time-consuming practices such as job enlargement or enrichment and
improve an individual’s perception of job characteristics (Piccolo & Colquitt, 2006).
However, the proposed model suggests that, in addition to predispositions, social
construction of information provided by coworkers could also determine perceptions of
job characteristics (R. W. Grifﬁn, Bateman, Wayne, & Head, 1987; Piccolo & Colquitt,
2006). Job design, therefore, could be improved by attending to the demographic

composition of work environments (e.g., age, gender, and ethnicity), training for

interpersonal skills, or promoting a cultural change within an organization (Parker et al.,
2001).

A ﬁnal advantage of this dissertation is that it uses a large and nationally
representative dataset; this may resolve a critical issue commonly identiﬁed in both social
support and job characteristics studies, namely, the issue of inconsistent ﬁndings (D.
Ganster, M. R. Fusilier, & B. T. Mayes, 1986). Previous studies generally agree that
inconsistent ﬁndings result from confounding factors such as size and heterogeneity of
samples, socioeconomic status, health-related behavior, individual differences (e.g., age,
sex and constitution), and duration of exposure (de Jonge & Kompier, 1997; de Jonge,
van Breukelen, Landeweerd, & Nijhuis, 1999; de Rijk, Le Blane, Schaufeli, & de Jonge,
1998; van der Doef & Maes, 1999). The dataset used for this dissertation is large enough
to provide statistically powerful tests of models of social support and job characteristics.

This dissertation will proceed as follows. Chapter 2 reviews previous research
and studies in detail to develop a theoretical model. Chapter 3 discusses the proposed
model, and descrbes the hypotheses tested in Chapter 4. The research method, including
a discussion of the sample, scales, and analytical strategies, is covered in Chapter 5.
Chapter 6 discusses the analyses and results, and Chapter 7 concludes this dissertation
with a discussion of the major ﬁndings and their implications for research and practice.
Discussions of study limitations, suggestions for future research, and the primary

contributions of this dissertation are also included in Chapter 7.

CHAPTER 2: LITERATURE REVIEW

m

When describing stress, researchers often break down the concept of stress into
two components: stressors and strains. Stressors are external events or conditions, such
as excessive noise and temperature in the workplace and high and repetitive task
demands, which require adaptive response from individuals (Goldberger & Breznitz,
I993; .Iex, I998). The National Institute of Occupational Safety and Health (NIOSH) has
identiﬁed several important psychological risk factors in the workplace (Jex, I998):
workload and work pace, work schedule, role stressors, career security factors,
interpersonal factors, and job content (characteristics). Stress researchers have also paid
special attention to such factors. In a more recent study, Cooper et al. (2001) added
‘work-home interface’ as an emerging stressor to the above work-related factors.

Workload is one of the most signiﬁcant and salient stressors consistently
discussed in stress research. Workload is the amount of work that workers have to
perform and usually consists of ‘perceived’ and ‘actual/objective’ workload (Jex, 1998;
Cooper et al., 2001). Stress researchers have been particularly interested in perceived
workload since an individual’s perception of his/her environment is critical for the
experience of strain. Perceived workload is further divided into ‘quantitative’ and
‘qualitative’ workload. Quantitative workload is deﬁned as the sheer amount of work
required and the time frame in which work most be completed (Jex, 1998; Cooper et al.,
2001). Qualitative overload occurs, on the other hand, when individuals believe they do
not have the necessary skills or capabilities to conduct their tasks (Cooper et al., 2001).

In addition, perceived workload is often considered conceptually proximal to ‘role

11

overload,’ deﬁned as the stressor that occurs when people perceive the demands of work
as excessive (Jex, 1998; Jones, Flynn, & Kelloway, I995).

Strains are a multitude of negative ways that individuals may respond when faced
with stressors (Goldberger & Breznitz, I993; .lex, 1998). Note that individuals may
respond to stressors in a positive or a neutral way. In this case, the response would not be
regarded as strain (Jex, 2998). Strains are generally classiﬁed as psychological, physical,
or behavioral. Psychological strains usually include anxiety, depression and job
dissatisfaction. Physical strains are symptoms ranging from headaches to coronary heart
disease. Behavioral strains, especially in a workplace, include poor productivity,
turnover, and absenteeism.

Stress is deﬁned from one of the three perspectives: (I) the stimulus-based view;
(2) the resource-based view; (3) the transactional view (Cary L. Cooper, Dewe, &
O'Driscoll, 2000; Cary L. Cooper et al., 2001; Goldberger & Breznitz, 1993; 1998; Kasl,
I983). The stimulus-based view focuses only on stressors for describing stress (e.g.
working conditions), while the resource-based view sees stress as strains, such as
psychological reactions to these working conditions (Fumham, 2005). Until recently,
psychological stress studies have been dominated by these two perspectives.

Despite their dominance, such ‘one-component’ perspectives (e.g., focusing only
on stressors or strains) do not adequately predict whether a particular individual is under
stress. since what is stressful for one individual may not be stressful for another (Cary L.
Cooper et al., 200 I ). Also, these one-component perspectives explain little about the

perceptual and cognitive processes of stress. Since stress involves both a stimulus and a

12

response in relation to one another, the relational nature of stress should be included in
any deﬁnition (Lazarus, 1990).

To deal with such criticisms, a third perspective was developed which
incorporated both stimulus and response elements as well as possible intervening factors
such as personality differences (Fumham, 2005). This perspective, which emphasizes the
dynamics of relationships between stressors and strains, is called the transactional view
(Cary L. Cooper et al., 2001; S. Fisher, 1989; Wells, 1982). This view states that stress
involves disturbances in the relationship between individuals (responses) and their
environment (stimulus) (Cary L. Cooper et al., 2001). In this view, stress does not reside
solely in the individual or solely in the environment but in the conjunction between the
two (Edwards & Cooper, 1990; Tansey, Mizelle, Ferrin, Tschopp, & Frain, 2004; Van
Harrison, 1978; Xie, 1996). In other words, the stress process can be considered a
continual transaction between external demands and constraints, external supplies and
supports, personal resources and internal needs or values to be balanced (Caplan, I987;

Daniels & Guppy. I994).

Perceived Job Characteristics - The Demand-Control Model
Theoretical background

Karasek’s (1979) Demand-Control (D-C) model is one of the most recognized
and inﬂuential job characteristics models. The D-C model was founded upon two
distinctive academic research streams: psychological strain and job satisfaction. Karasek
(1979) found that research on psychological strain tended to focus on job demands (e.g.,

workload and role overload) as its antecedent but not on autonomy. while research on job

13

satisfaction tended to focus on autonomy as its antecedent and ignoredjob demands
(Karasek, 1979). The study concluded that job characteristics, job demands and
autonomy, must be analyzed at the same time, since omitting one could result in

misinterpretation and/or inconsistent ﬁndings (van der Doef & Maes, I999).

The model

The Demand-Control model therefore consists of two psychosocial job
characteristics: perceived job demands and job control (see Figure 2 below). According
to Karasek and Theorell (1990), job demands are psychological stressors present in the
work environment. Examples of the psychological demands of work include
requirements for working hard, having a great deal to do, not having enough time, and
having conﬂicting demands (Xie, 1996). In contrast, perceived job control represents the
potential control that workers have over their tasks in the workplace, and is made up of
two distinctive constructs: autonomy and skill discretion. Autonomy is the authority that
workers possess to make decisions on the job, while skill discretion is the breadth of
skills used by the worker on thejob (Cary L. Cooper et al., 2000; Robert A. Karasek,
Triantis, & Chaudhry, 1982). In a strict sense, perceivedjob control is not a stressor but
the lack of job control or autonomy can be a stressor (Susan E. Jackson, 1983). While the
D-C model does not refer to lack of control as a stressor, some researchers have proposed
and found the direct effect of ‘Iack of control’ on psychological strain (Ten'y A. Beehr,

Glaser, Canali. & Wallway, 2001).

I4

 

 

 

 

 

 

 

 

 

 

 

Job Demands _ Well-being
- St
Job Control - Dergizssion
- Autonomy . - Job satisfaction
- Skill discretion - Emotional
Dob Demands x Job Control exhaustion, etc.

 

 

Figure 2: Karasek’s Demand-Control Model and Proposed Outcomes

For more than 20 years, many studies have applied the Demand-Control (D-C)
model and its hypotheses to a broad array of outcomes, particularly psychological well-
being: stress, exhaustion, depression,job/Iife satisfaction, work-family interference,
anxiety, and burnout (e.g., Butler, Grzywacz, Bass, & Linney, 2005; de Jonge &
Schaufeli, 1998; de Rijk et al., 1998; Janssen, Peeters, de Jonge, Houkes, & Tummers,
2004; van Vegchel, de Jonge, & Landsbergis, 2005). In contrast, only a limited set of
studies have examined behavioral outcomes such as the number of sick days, job
performance, and organizational citizenship behavior (de Jonge, Reuvers, Houtman,
Bongers, & Kompier, 2000; Jex, I998; Noblet, McWilliams, Tea, & Rodwell, 2006).

The uniqueness of the D-C model lies in its examination of the interaction effects
of job demands and job control. Karasek (1979) proposes that although excessive work
demands may be associated with higher levels of psychological strain, the impact of these
demands may be offset by the perception that one has control over important aspects of
the work environment. In order to explain the interaction, Karasek (1979) developed a 2-
by-Z matrix in which there are high and low job demands as well as high and lowjob
control. The four possible outcomes of the interactions are labeled as follows: low-strain

(high control-low demands), passive (low control-low demands), active (high control-

high demands) and high-strain (low control-high demands). According to Karasek
(1979), stress is more likely to occur in ‘high-strain’ jobs that are simultaneously high in
job demands and low injob control. In other words, the source of work stress is notjob
demands, but the high level of job demands in combination with a lack of job control.
This interaction effect has been attracted by considerable theoretical and
methodological criticism, however. This is primarily because there is no consensus in
conceptualizing and operationalizing the interaction effects of job demands and job
control (Xie, 1996; de Rijk et al., 1998; de Jonge et al., 1999; Beehr et al., 2001; Wong et
al., 2007; de Jonge & Kompier, 1997; van der Doef, 1999). Originally, Karasek(1979)

proposed that the interaction effect could be expressed by the following formula:

Interaction effect = job demands —job control

However, such conceptualization and operationalization of the interaction effect is

different from the more prevalent way of conceptualizing and operationalizing interaction

effects, which is expressed by the following formula (Edward & Cooper, 1990; Ganster

& Gusilier, I989; Beehr et al., 2001):

Interaction effect = job demands x job control

This is called a ‘multiplicative’ interaction (Cappelli & Sherer, I991; Clegg,

1984; Morgeson & Campion, 2003; Parker et al., 2001). Studies that conceptualize

interactions in this way consider perceived job control as a buffering factor reﬂecting
individuals’ perceptions of their speciﬁc work environment (Cooper et al., 2001).

Although the ‘multiplicative’ interaction is conceptually distinct from the ‘relative
excess’ interaction, studies, including Karasek’s later pieces (R. A. Karasek, 1989;
Robert A. Karasek et al., 1982), began to use these two formulas interchangeably for
examining the interaction effects of job demands and job control. As a result, some
studies found evidence of interactions (de Jonge et al., 1999; Robert A Karasek, 1979),
while others found no interaction effects (de Rijk et al., 1998; Fletcher & Jones, 1993).
Yet others have found that the empirical results changed depending on how the
interaction effects are statistically interpreted (Edwards & Cooper, 1990; van der Doef &
Maes, 1999; van Vegchel et al., 2005). Having a sound theoretical rationale is essential
for examining the Demand-Control interaction (Terry A. Beehr et al., 2001).

At the same time, however, it is worth mentioning that the majority of prior
studies have found evidence of the direct/main effects of job demands and job control on
various psychological outcomes. In other words, job demands and job control are still
signiﬁcant factors that independently affect individuals’ psychological states such as
stress and depression, when considering lack of autonomy/skill discretion as a stressor.
As mentioned previously, some researchers have proposed and found the main effects of
both job demands and job control on stress (Terry A. Beehr et al., 2001; Hendrix, Ovalle,
& Troxler, 1985; Susan E. Jackson, 1983).

Other previous studies have identiﬁed posssible sources of inconsistent empirical
ﬁndings. 3 major shortcoming of the D-C model (Terry A. Beehr et al., 2001; de Jonge &

Kompier, 1997: de Jonge & Schaufeli, 1998; de Rijk et al., 1998; Theorell & Karasek,

I996; van der Doef & Maes, 1999). For example, such ﬁndings may result from the
problematic nature of the variables of job demands and control. Researchers have
warned against the use of a global index of perceived job control, which consists of
autonomy and skill discretion, since it may mask the impact of some forms of control
(Cary L. Cooper et al., 2001; de Jonge & Kompier, 1997). In fact, some studies have
empirically tested the dimensionality of job control and concluded that studies using the
D-C model should not use the global job control measurement (Carayon & Zijlstra, 1999;

C. S. Smith et al., 1997).

Social Support

Stress researchers have been interested in exploring possible contextual factors
that affect stressor-strain relationships (Cooper et al., 2001). Social support is one of
these factors and is considered a resource or coping mechanism that may reduce an
individual’s stress level. Social support in the workplace is a helpful social interaction
available on ajob from both coworkers and supervisors (Dawn S. Carlson, 1999). More
speciﬁcally, social support is an interpersonal transaction that involves emotional concern,
instrumental aid, information, or appraisal at work (Terry A. Beehr et al., 2000; Dawn S.
Carlson et al., 2000; Ducharme & Martin, 2000; Etzion, 1984; C. D. Fisher, 1985;
LaRocco, House, & French, 1980; Parasuraman et al., 1992; Schaubroeck, Cotton, &
Jennings, 1989; Searle et al., 2001; Seers et al., 1983; Wells, 1982).

There has been substantial research on the relationship between social support and
psychological well-being and physical outcomes (Dawn S Carlson & Perrewe, I999;

Kaufmann & Beehr, 1986). In particular. considerable research has examined the impact

18

of social support on stress (Beehr, I995; Jex, 1998). In fact, Johnson (1988) presented an
expanded ‘Demand-Control-Support (D-C-S)’ model that introduces the notion of ‘iso-
train,’ in which jobs with high demands, low control and low social support are more
likely to induce heart disease, fatigue, and other health problems (Karasek et al., 1982).
They argue thatjob-related social support may facilitate successful coping with high
strain jobs, by preventing or buffering the potential harmful effects of jobs (Karasek &
Theorell, 1990).

The types, sources, and mechanisms of social support have received signiﬁcant
attention within the research on the subject (Ducharme & Martin, 2000). In terms of
types of social support, for example, research has proposed various taxonomies that
suggest different dimensions of social support: instrumental, active, and material support
(Cobb, I979); affect, afﬁrmation, and aid support (Kahn & Antonucci, I980); tangible,
informational, and emotional support (Cohen & Wills, 1985; House, 1983). Among these
taxonomies, two types of social support have been studied the most: instrumental and
emotional support (Beehr et al., 2000). Instrumental support refers to the different types
of tangible help that others may provide, such as material assistance in response to
speciﬁc needs (Ducharme & Martin, 2000). Emotional support, on the other hand, refers
to the things, such as sympathy, that provide the receivers with feelings of being accepted
and cared for (Beehr et al., 2000).

Research has been interested in sources of social support, though in a limited
sense. For instance, previous studies have been sensitive to the distinction between the
two sources of social support (Dawn S. Carlson, 1999); one from a work domain, such as

supervisors and coworkers; the other from a non-work domain, such as family and

I9

spouses (Dawn S Carlson & Perrewe, 1999; Wallace, 2005). However, these studies
have paid less attention to comparing and contrasting sources of social support within a
work domain. Such insensitivity becomes apparent when operationalizing social support
at work. For example, Etzion’s (1984) social support indicator consists of ten items
asking about the quality of the relationship with all of supervisors, coworkers, and
subordinates. Karasek et al. (1982) initially differentiated supervisor support from
coworker support but later combined them into one social support measure due only to a
methodological reason (high factor loading). As a result, discussions about the
differences between supervisor and coworker support have rarely occurred in social
support studies (Ducharme & Martin, 2000). Similar to other studies, these examples
demonstrate how research has typically combined supervisor and coworker support
despite the fact that they may be distinct constructs. As such, it is important to understand
these differences in the context of social support studies (Ducharme et al., 2000).

Some researchers have warned that coworker support may not work in the same
way that supervisor support works (Cohen & Wills, 1985). For example, House (1983)
states that informal and nonprofessional support, such as that from family, friends, and
coworkers, should be more focused on as primary sources, since they are much more
commonly mentioned when people are asked to name actual sources of support. Though
few, there are empirical studies that compared the effect of supervisor support with that
of coworker support (House, McMichael, Wells, Kaplan, & Landerman, 1978; Robert A.
Karasek et al., 1982; LaRocco et al., 1980), but these ﬁndings were not fully incorporated

into later studies.

In terms of the mechanism of social support, there are two dominant hypotheses
to explain the effect of social support on an individual’s stress level (Jex, 1998; Cohen &
Wills, I985; Beehr, I985). The ﬁrst hypothesis states that social support may work
directly to reduce psychological strains and increase well-being. This is called the
‘direct/main’ effect of social support (Cohen & Wills, I985). The underlying assumption
of the hypothesis is that social support may provide positive affect, a sense of
predictability and stability in one’s life situation, and recognition of self-worth. These, in
turn, are posited to prevent negative experiences that lead to psychological and physical
disorders (R. Karasek & Theorell, I990). The second hypothesis is suggestive of social
support as a moderator/buffer to reduce the positive relationship between stressors and
the outcomes (Terry A. Beehr et al., 2000; Ducharme & Martin, 2000; D. C. Ganster, M.
R. F usilier, & B. T. Mayes, I986; Jex, I998; Kaufmann & Beehr, 1986; Parasuraman et
al., 1992). In other words, the strength of the relationship between stressors and the
outcomes depends on the level of social support; when social support is high, the
relationship between stressors and their outcomes is relatively weak. In contrast, if social
support is low. the relationship between stressors and the outcomes becomes stronger.

While many studies have found evidence of the direct effect of social support,
there is considerable debate over the moderating/buffering effect of social support (Terry
A. Beehr et al., 2000; LaRocco et al., 1980). The notion of the moderating/buffering
effect seems intuitively appealing, but the research on social support as a moderating
variable is far from conclusive (Carlson & Perrewe, 1999). Some studies have found
evidence for this hypothesis (Cohen & Wills, 1985; Daniels & Guppy, I994; Etzion,

I984; Kaufmann & Beehr, I986; Seers et al., 1983), while others failed to ﬁnd such

evidence regarding the moderating/buffering effects of social support (Terry A. Beehr et
al., 2000; Ducharme & Martin, 2000; C. D. Fisher, 1985; D. C. Ganster etal., 1986;
LaRocco et al., 1980; Seers et al., 1983).

My examination of the dataset prior to this dissertation analysis showed mixed
results in terms of the moderating/buffering effect of social support on the relationship
between job characteristics (job demands and job control) and an individual’s stress.
Using multiplicative interaction terms and the procedure suggested by Baron and Kenny
(I986), the empirical results showed that coworker support moderated the relationship
between job demands and stress, but not the relationship between job control (autonomy
and skill discretion) and stress.

The methodological and conceptual rationales for the inconsistent ﬁndings
regarding the moderating/buffering effects of social support have been discussed in the
literature. For methodological reasons, Ganster et al. (1986) noticed that the evidence of
the moderating/buffering effect might depend on the source of support, the people who
receive such support, and the stressors/strains chosen for a study. Other studies have
pointed out measurement unreliability, low statistical power, and speciﬁcation of social
support (Dawn S. Carlson, 1999; C. D. Fisher, 1985; L. T. Thomas & Ganster, 1995).

Like the Demand-Control interaction, however, a fundamental problem of the
moderating/buffering effect of social support is a lack of theoretical rationale (Dawn S.
Carlson, 1999; Cary L. Cooper et al., 2001; Jex, 1998). For example, Cooper (2001)
observed the problem of using a moderator/buffer approach to analyze the stressor-strain
relationship:

"(J)ob stress research has investigated a very large array of
moderator variables. sometimes with little theoretical

Ix)
IQ

rationale for their inclusion in a study’s design, resulting in

inconsistent (and frequently ambiguous) ﬁndings about the

role of these variables. We suggest that it is important now

to move beyond the simple identiﬁcation of potential

moderator variables to more comprehensive theories that

attempt to explain the mechanisms by which all relevant

factors interact. Furthermore, even when moderators are

selected on a theoretical basis, empirical ﬁndings often

simply demonstrate a moderator effect rather than

explicating the role that the moderator plays in the stress

process (p. 147).”

Similar to the Demand-Control interaction, the above arguments demonstrate the

imperative nature of developing a solid theoretical foundation when testing the effect of
social support on psychological outcomes. This dissertation, therefore, uses the Social

Information Processing (SIP) approach to explain how social support affects the

relationship between job characteristics and social support.

Social Information Processing (SIP) Approach
The concept

Studies of job characteristics have often been criticized for failing to reﬂect on
factors that inﬂuence and constrain choices of job design (Adams, 1965). In fact, most
studies of job attitudes and task design have been discussed in the ‘need-satisfaction
paradigm,’ which emphasizes individuals’ predisposition for explaining behaviors. There
have been not enough attempts to explore factors, other than dispositional ones, that
alternatively explain work attitudes and behaviors (Cappelli & Sherer, I991; Rousseau,
I978; Salancik & Pfeffer, 1978). Such factors are called ‘contextual’ factors and the term
context refers to the set of surroundings that help to illuminate events (Rousseau, 1978, p.

522). In an organizational setting, contextual factors are characteristics of organizations,

individuals and their role in an organization, and other environmental factors including
social information at work (Salancik & Pfeffer, 1978). As Blau and Katerberg (1982)
suggested, contextual factors are important determinants, since people often spend a large
amount of time understanding and interpreting information from their social context to
determine their attitude and behavior.

Based on these arguments, Salancik and Pfeffer (1978) developed a theoretical
model, called the Social Information Processing (SIP) approach, to explain how
individuals develop their attitude and behavior toward job characteristics. They propose
thatjob characteristics, such as conditions of a workplace, are not a given but rather are
constructed by individuals’ observing and interpreting information from their social
context. According to Salancik and Pfeffer (1978), there are two paths to explain the
effects of social information on individuals’ attitudes and behavior. First, information
from the social context and past actions directly shape individuals’ attitude and behavior
(Salancik & Pfeffer, 1978). Alternatively, social context focuses individuals’ attention on
certain information by making it more salient and provides expectations regarding
individuals’ attitudes and the logical consequences of such attitudes (Salancik & Pfeffer,
1978).

A number of studies have examined the effects of social information on job
characteristics and outcomes, though predominantly in laboratory and ﬁeld settings
where a manipulation of social cues is possible. For example, O’Reilly and Caldwell
(1979) found that major perceptions of task characteristics were informational inﬂuences
in the forms of social cues about the tasks in addition to objective realities. Grifﬁn

(1 983) used social cues from supervisors and found that both the objective task changes

24

and the informational cues from supervisors signiﬁcantly inﬂuenced research
participants’ perceptions of task attributes and their affective responses. The
experimental study of Grifﬁng et al. (1997) reported how research participants perceived
various social cues to determine their attitudes and behavior toward their assigned jobs.
They provided participants with three forms of manipulated social cues: written
descriptions of the job, written evaluations by others who had done these tasks, and
verbal statements from the experimenters. They found that social information as well as
the objective elements of tasks affected perceptions of job characteristics. Field studies
examined an even broader range of social cues. For instance, Zalesny and Farace (1986)
used the agency newsletter for manipulating social cues. They found that socially-
transmitted information focused a person’s attention to certain aspects of the work/social
environment of their organization. They also found that workers with socially-
transmitted information tended to have more similar perceptions than those without such
information. Thomas (I 986) expanded the sources of social cues to family members,
customers (patients) and other people working in the same hospital to examine how
nurses used those sources to gather information. He found that, in addition to supervisors
and coworkers, family members and people who are not immediate supervisors or
coworkers were signiﬁcant sources of social cues affecting nurses’ perception of task
characteristics.

Coworkers are an important source of social cues (Oldham, Kulik, Stepina, &
Ambrose, 1986). In fact, a review of the SIP literature demonstrates that the most
widely-used sources of social cues in empirical research have been coworkers and

supervisors (J. Thomas & Griffin, 1983). Oldham et al. (1979) found that workers

25

determine their attitudes toward their job by comparing themselves with their coworkers.
If they feel that they have fewer opportunities in the workplace, they are less internally
motivated. In contrast, if they feel they have more opportunities compared to others, they
are more likely to be internally motivated and perform well. White and Mitchell (1979)
found that participants receiving positive social cues from coworkers were more satisﬁed
and more productive than those receiving negative social cues from coworkers. These
ﬁndings indicate that individuals indeed receive various social cues from their
environment in terms of sources (e.g., supervisors, coworkers, and family members) and
forms (e.g., verbal or written). However, social cues that have been used for previous
studies are mostly arbitrarily provided by experimenters, primarily because most of these
studies have been conducted in a laboratory setting.

As introduced above, social support is conceptualized as a ﬂow of communication
between people involving emotional concern, caring, and informational as well as
instrumental help (Daniels & Guppy, I994; Seers et al., 1983). From this deﬁnitional
standpoint, social support could be a social cue that inﬂuences an individual’s perception
of job characteristics. However, previous studies have not explicitly used social support
as one of social cues. This dissertation therefore takes the SIP standpoint and sees social
support as a social cue, proposing that social support affects individuals’ perceptions of

job characteristics.

Demographic diversinz/similariw

So far, this chapter has covered social support and its dependent and mediator

variables (stress and job characteristics, respectively). This subsection reviews a possible

26

antecedent of coworker support, namely demographic similarity (see Figure I).
Demography research assumes that there are compositional or demographic effects
caused by demographic distribution which are more than the sum of the effects of the
individual-level variants (Charles A. O'Reilly, Caldwell, & Barnett, 1989; Jeffrey Pfeffer,
1983; Jeffrey Pfeffer & Moore, 1980; Zenger & Lawrence, 1989). A number of studies
have found that demographic variety in organizations has substantial effects on
organizational outcomes, such as innovation, performance, executive succession, and
turnover (Susan E. Jackson, Joshi, & Erhardt, 2003; Frances J. Milliken & Martins,
1996; K. Y. Williams & O'Reilly III, 1998).

Demography research has been called diversity research, in which the term
diversity covers not only demographic compositions but also other elements found in an
organization, such as the length of service, educational background, and functional
background (Harrison, Price, & Bell, 1998; Jehn, Northcraft, & Neale, 1999; Frances J.
Milliken & Martins, 1996; van Knippenberg & Schippersand, 2007). Milliken and
Martins (1996) called these emerging elements of diversity ‘underlying diversity’ (e.g.,
personality, culture, skills, and tenure), as distinct from ‘observable diversity’ (e.g.,
demographic characteristics). Others use the term ‘deep-level’ for such underlying
elements, versus ‘surface-Ievel’ diversity (Harrison et al., 1998). Yet others use terms
such as ‘value,’ ‘social category,’ and ‘informational’ diversity (Jehn et al., 1999).

While the majority of diversity research has been concerned with the proportion
of a certain demographic group, some research has argued that compositional property,
measured by the variance in demography within the unit, cannot adequately capture the

full impact of potential demographic effects (Chattopadhyay, Tluchowska, & George,

27

2004; Tsui, Egan, & O'Reilly, 1992). As a result, it has been proposed that the effects of
individuals’ similarity to their work group or to their leaders as predictors of individual
outcomes (Tsui & O'Reilly, 1989), which is called ‘relational demography.’ The key
feature of relational demography is ‘comparative’ similarity or dissimilarity in given
demographic attributes, such as age and gender, of supervisors, subordinates, or
colleagues. In other words, workers compare their own demographic characteristics with
those of other members of the unit. The level of perceived similarity in turn inﬂuences
workers’ identiﬁcation with the unit (Tsui et al., 1992).

The mechanisms of demographic effects from either the traditional or relational
view can be explained by a seminal framework called the similarity-attraction paradigm,
which proposes that similarity of attributes such as attitudes and values increases
interpersonal attraction and liking (D. E. Byme, 1971). In this paradigm, individuals
classify themselves and others based on categories that are salient such as age, gender,
ethnicity, or status (van Knippenberg & Schippersand, 2007). They then tend to favor
more, trust more, and be more willing to cooperate with group members than people
outside the group (K. Y. Williams & O'Reilly III, 1998). Also, those individuals may
share common life experience and values, since they may ﬁnd that those with similar
background interact more readiliy and are more attracted to each other (van Knippenberg
& Schippersand, 2007).

A primary purpose of diversity research is to examine how diversity affects group
process (e.g., cognitive processing, communication problems, and cohesiveness) and
outcomes (e.g., performance, creativity, and problem-solving), as well as affective

outcomes such as commitment and job satisfaction. Diversity studies, however, have

tended to focus more on performance outcomes than either the process or the affective
outcomes. According to Jackson et a1. (2003), only eight percent of the studies that they
reviewed examined affective outcomes at the team level. Inconsistent empirical ﬁndings
are a primary reason for the decline of such studies; diversity studies of affective
outcomes (e.g.,. such as satisfaction) have sometimes positive, sometimes negative, and
sometimes not signiﬁcant ﬁndings depending on the particular dimension of diversity
considered (Rico, Molleman, Sanchez-Manzanares, & Van der Vegt, 2007). Surprisingly,
no study has examined the effect of demographic diversity on the level of social support

that individuals perceive.

Reviewing the literature reveals that investigating the relationship among stress,
social support, job characteristics, and demographic similarity is not a new idea. In fact,
each paired set of concept has a research history: social support and stress, job
characteristics and stress, and job characteristics and social support. However, the
problem commonly identiﬁed among these concepts is the lack of a sound theoretical
argument to explain the relationships. This dissertation therefore implements the Social
Information Processing (SIP) approach, including the similarity-attraction paradigm, and
the Demand-Control model of job characteristics to arrange the three concepts in a way
that makes theoretical sense. In other words, this dissertation aims at developing and
testing a cohesive theoretical argument regarding the causal linkages among stress, social
support, and job characteristics. The next chapter proposes the model and discusses each

concept in detail.

29

CHAPTER 3: MODEL DEVELOPMENT
Figure 1 represents the graphical representation of the proposed model. As the
ﬁgure shows, the model consists of four components: an individual’s stress, job
characteristics, social (coworker) support, and demographic similarity. These four

components and the theoretical linkage of these components are discussed in detail.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Perceived Job
Demographic Social Support Characteristics
Similarity
J Job Demands
Age Coworker // \ An
Gender /I' Support } individual’s
Ethnic \ Job Control Stress
background . N - Autonomy
- Skill 1
Discretion
\ J \ J
V V
Social Information Processing (SIP) Demand — Control (D-C) model

 

 

 

 

 

 

Figure l: The Proposed Model

An individual’s stress in this model is deﬁned as a subjective feeling in response
to stressors in the environment (Wells, 1982). This deﬁnition takes the transactional
perspective, which considers stress as a continual transaction between external demands
and constraints (S. Fisher, 1989). As described in the literature review, the term
transaction implies that stress is neither in the environmental nor in the person, but
reﬂects the conjunction of a person with an environment (Lazarus, I990). The term

transaction also implies a process, meaning that the stressor-strain relationship is not

30

static but is constantly changing as a result of the continual interplay between the person
and the environment. Generally, researchers agree with the notion that term stress should
denote the overall process incorporating stressors, strains, and the process. When
operationalized, however, this notion contains a number of limitations in its ability to
explain the stress process (Cooper et al., 2001).

Instead of searching for a single satisfactory measure that explains the stress
process, this dissertation follows the suggestion by Lazarus (1990), in which a series of
measures are developed to capture various aspects of the stress process or the stressor-
stress relationship. In addition, the dependent variable of the model is called ‘stress’
instead of ‘strain’ throughout this dissertation, though it is a ‘strain’ in a context of a
stressor-strain relationship. The process of stressor-strain relationship will be discussed
later in the context of the ‘Demand-Control’ model.

Social support is an important resource, or coping mechanism, that can reduce the
negative effect of stressors (Daniels & Guppy, 1994). Previous studies deﬁne social support
as an interpersonal transaction that involves emotional concern, instrumental aid, or
appraisal (Dawn S Carlson & Perrewe, 1999; House, I981a; Kaufmann & Beehr, 1986;
Wallace, 2005). As this deﬁnition indicates, social support contains both instrumental
and emotional support. Instrumental support refers to providing material or practical
assistance in response to the individual’s speciﬁc needs, such as information, ﬁnancial
aid, and help with one’s tasks (House, I981a). Emotional support, on the other hand,
involves providing individuals with the feeling of being cared for, and includes empathy,
affection, understanding, and trust (House, 1981a; Wallace, 2005). In addition to

instrumental and emotional support, this proposed model includes another dimension of

31

social support: a sense of belonging that workers perceive from their coworkers.
Researchers have argued that feelings of being a part of the group are also an important
invisible support that workers receive from their coworkers (Berkman & Syme, 1979;
Cohen & Wills, 1985; House, Robbins, & Metzner, 1982). Social isolation, or a lack of a
sense of belonging, is the most serious phenomenon for individual workers that affects
individuals’ health outcomes because it decreases the availability of social support from
others in a workplace (Salancik & Pfeffer, 1978). Based on the above argument, social
support in this proposed model is therefore deﬁned as an interpersonal transaction that
involves emotional concern, instrumental help, ﬂow of information, a sense of belonging
to the group, and the feeling of being cared for.

This dissertation focuses only on a single source of social support in the
workplace: coworker support. Although previous studies have paid less attention to
comparing and contrasting sources of social support within a work domain, this
dissertation proposes that differences exist between supervisors and workers especially in
terms of the relationships, such as ‘supervisor-subordinate (worker)’ and ‘worker-
worker’ relationships. First, there is a structural difference between the two relationships.
While relationships among workers are considered horizontal, relationships between
workers and their supervisors are considered vertical. In a horizontal relationship,
interactions more likely take place on one-to-many basis, while in a vertical relationship,
interactions are more likely done on one-on-one basis. This structural difference also
includes power differences between workers and their supervisors. Furthermore, there is
a difference in physical and psychological distance; in most cases supervisors are more

physically and psychologically distant from workers than their colleagues are. Such

physical and psychological closeness affects not only the numbers of interactions, but
also the signiﬁcance of the impact that social support has on individuals. Based on the
above argument, this dissertation considers coworker support as distinct from supervisor
support.

There are two theoretical paths for explaining the effect of coworker support on
an individual’s stress. As introduced in the previous chapter, coworker support may
directly reduce strain by providing positive affect and a sense of predictability as well as
preventing negative experiences (Fisher, 1985; Karasek & Theorell, 1990; Carlson &
Perrewe, 1999). Alternatively, coworker support may affect an individual’s perceptions
of job characteristics, which in turn determine the level of stress. In other words, in the
presence of environmental stressors, the perceived job characteristics in this case are
either not perceived, or are objectively reduced through the instrumental aid of the
supporter. In this proposed model, coworker support is considered a social cue that helps
construct an individual’s perception of the job characteristics. Again, this argument is
based on the Social Information Processing (SIP) approach, proposing that individuals’
perceptions are inﬂuenced by the social environment or by the network of social and
informational relationships (Salancik & Pfeffer, 1978).

Job characteristics in this proposed model consist of perceived job demands and
job control. Note thatjob characteristics in this proposed model are ‘perceived’ job
characteristics, not ‘objective’ characteristics; hereinafter the term job characteristics in
this dissertation indicates perceived job characteristics. Jex (l 998) lists job demands as
one of the primary sources of stressors in the workplace. Job demands are referred to as

the amount of workload that individuals perceive regarding theirjob (Robert A Karasek,

33

1979). In other words, psychological job demands are the perception of ‘how hard you
work’ (R. Karasek & Theorell, 1990); In this regard, it does not include dimensions such
as ‘actual’ workload (Jex, I998), ‘physical’ workload (R. Karasek & Theorell, 1990), and
‘qualitative’ workload (Fletcher & Jones, 1993; Landsbergis, I988). However,job
demands in this model may overlap with other common research perspectives associated
with job-related stress, such as Kahn’s (1964) role overload.

Job control represents the potential control that workers have over their tasks
during a working day (Karasek, 1979). In this proposed model, job control consists of
two dimensions: autonomy and skill discretion. Autonomy is the authority that workers
possess to make decisions on the job, and skill discretion is the breadth of skills used by
the worker on the job (R. Karasek & Theorell, 1990). Although some studies combine
these two dimensions into a single construct of job control (R. Karasek & Theorell, 1990),
this dissertation takes the position that autonomy and skill discretion are closely related
but conceptually distinctive (Carayon & Zijlstra, 1999; C. S. Smith et al., 1997), and
should not be combined into a composite job control measurement (Edwards & Cooper,
1990; Tansey et al., 2004; Van Harrison, 1978; Xie, 1996).

In this model, job demands and job control are the important determinants of the
level of an individual’s stress. Job demands themselves have potentially negative effects
on the physical and psychological state of individuals, while a person’s ability to control
the environment has the potential to lead to positive consequences. As mentioned in the
literature review, the two determinants are developed based on the Person-Environment
ﬁt (P-E ﬁt) approach of stress, which articulates that a strain develops when there is a

misﬁt/discrepancy between the demands of thejob and the abilities of a person to meet

34

those demands (Edwards & Cooper, 1990). The misﬁt/discrepancy is often called an
‘interaction effect’ of job demands and job control, and is hypothesized to generate
deleterious psychological, physiological and behavioral outcomes, including stress (van
der Doef & Maes, 1999; van Vegchel et al., 2005).

The proposed model, however, does not introduce any forms of interaction effects
of job demands and job control (e.g., relative excess, multiplication, and
ratio/proportional — see the previous chapter). In other words, the
interactionaI/transactional nature of an environment and a person in this model is not
conceptualized by the interaction term of job demands (an environment) and job control
(a person), primarily because no form of the interaction effect could adequately express
the ‘continuous’ transaction of stressor-strain relationships. Since the stress measure of
this dissertation captures that aspect of stress process, the proposed model focuses simply
on the effect of two determinants of an individual’s stress.

Finally, demographic similarity in this model is deﬁned as the compositional
distribution of team members across any personal attribute that potentially leads to the
perception of being different/similar from each other (Tsui & O'Reilly, 1989). Note that
the term team in the above deﬁnition does not necessarily imply teams in a strict sense;
teams in this model include occasions in which coworkers simply work with others
regardless of whether they are formally considered team members. According to Pfeffer
(I983), demography is the composition of a unit in terms of basic attributes such as age,
sex, educational level, length of service or residence, and race. In this dissertation,

demographic similarity is assessed by similarity in age, gender, and ethnic background.

35

The relationship between demographic similarity and coworker support is
explained by the similarity-attraction paradigm, which can be considered a branch of the
Social Information Processing (SIP) approach. In this paradigm, individuals who have
similar demographics (e.g., age, gender, and ethnic background) are more likely to share
common life experience and values. People may ﬁnd that those with similar backgrounds
interact more easily, and may be more attracted to each other (Reuben M Baron & Kenny,
1986). This similarity increases group members’ sense of identiﬁcation and integration
with the group (Hodson, 2005). Consequently, it elicits more cooperative behavior
toward coworkers. Note that the proposed model focuses only on the effect of
demographic similarity on coworker support. In other words, demographic similarity is
treated as an antecedent of coworker support, instead of a predictor of stress.

In sum, the model proposes that the three components discussed - demographic
similarity, coworker support, and job characteristics - play a critical role in explaining an
individual’s stress level. Coworker support helps construct an individual’s perception of
job characteristics, which in turn inﬂuences the level of an individual’s stress. In addition,

demographic similarity promotes the social support of coworkers in the workplace.

36

CHAPTER 4: HYPOTHESES

Based on the proposed model introduced in the previous chapter, this dissertation
examines the relationship between social support and an individual’s stress (Path A) by
introducing a mediating effect of job characteristics. For establishing the mediating
effect, variations in social support should account for variations in an individual’s stress
(Path A), which is a fundamental assumption of the model. Based on this assumption,
Baron and Kenny (1986) suggest the following three conditions for mediating effect.
First, in Figure 3 below, variations in coworker support account for variations in job
characteristics (Path B). Second, variations in job characteristics account for an
individual’s stress (Path C). Third, when job characteristics are given, the previously
signiﬁcant relationship between coworker support and stress (Path A) becomes less/not

signiﬁcant.

   

 

    
 

Perceived Job
Social Support Characteristics

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Demographic Coworker JOb Demands 5. An
Similarity @’ Support 3 C individual’s
\ .lOb Control Stress
Na - Autonomy
- Skill ‘
Discretion

 

 

 

 

 

 

Figure 3: The Theoretical Paths
The hypotheses below regarding the mediating effects of job characteristics are
developed as follows: (I) a discussion of the relationship between coworker support and

stress (Path A): (2) a discussion of the relationship between coworker support and job

37

characteristics (Path B); and (3) a discussion of the relationship between job
characteristics and stress (Path C). Once the mediating effects of job characteristics are
hypothesized, a discussion of demographic Similarity is introduced to determine its

relationship with coworker support (Path D).

Coworker Support and Stress

In a Simple relationship between coworker support and stress, coworker support is
an antecedent and is considered to have a direct/main effect on stress (C. D. Fisher, 1985).
In this regard, coworker support is predicted to directly reduce tension because it may
serve to lower the experienced strain and other negative outcomes (C. D. Fisher, 1985;
LaRocco et al., 1980; Parasuraman et al., 1992; Schaubroeck et al., 1989; Seers et al.,
1983; Wells, 1982). For instance, it provides individuals with a social network that may
bring positive experiences and a set of stable and socially-rewarding roles in their work
environment. Such experiences and roles lead to positive affect, a sense of predictability
as well as stability in life situations, and the recognition of self-worth (Cohen & Wills,
1985). Consequently, coworker support helps reduce an individual’s stress level. As
introduced in the literature review, empirical ﬁndings have also indicated the direct effect
of social support (Cohen & Wills, 1985; C. D. Fisher, 1985;Jex, 1998). In this
dissertation, the direct effect of coworker is therefore predicted to have a positive effect

on an individual’s stress.

38

Coworker Support mid Job Characteristics

In this proposed model, the relationship between coworker support and job
characteristics is explained by the Social Information Processing (SIP) approach, in
which coworker support is considered as an antecedent of job characteristics. Some
empirical ﬁndings in previous studies underpin the argument that support in the
workplace could be viewed as an antecedent of various outcomes (Bateman, Grifﬁn, &
Rubinstein, 1987; Charles A. O'Reilly & Caldwell, 1979; Schnake & Dumler, 1987). For
example, Fisher (1989) found that social support served as an antecedent of an
individual’s perception of unmet expectations, which in turn affects job satisfaction.
Schaubroeck et al. (1989) also examined coworker support as an antecedent of job
satisfaction, which in turn affected turnover intention. Other studies found that social
support affected perceived stressors, such as work time demands, work
ambiguity/conﬂict, and control over tasks (C. D. Fisher, 1985).

In the context of the SIP approach, coworker support is a social cue for
individuals that helps Shape their perception of job characteristics. Researchers indeed
have acknowledged that coworkers are an important source of social cues (Cohen &
Wills, 1985). Grifﬁn and colleagues (1983; R. W. Grifﬁn et al., 1987) tested and found
that individuals’ perceptions of job characteristics were constructed based on their
processing information from coworkers. White and Mitchell (1979) noted that
participants receiving positive social cues from coworkers were more satisﬁed and more
productive than those receiving negative social cues from coworkers. Since coworker
support is a positive cue for individuals, it may help develop a positive work environment,

and consequently, workers with high social support may be less likely to perceive job

39

demands as stressors in their work environment. Fisher (1 985) concluded that social
support possessed a protective function that instills an ongoing sense of emotional
concern prior to the actual stressful event. In the presence of social support, individuals
may not perceive, or may perceive to a lesser extent, stressors such as job demands. In
other words, social support affects the appraisal of the environment where the perceived
stressor may be weaker when people perceive the environment as less threatening
(Oldham et al., 1986). In this regard, coworker support becomes a necessary resource to
redeﬁne the potential for harm posed by one’s work situation.

Social support may also work in a more practical way. For example, Thomas and
Ganster (1995) found that supportive supervisors affect workers’ level of control over
theirjob. Such supervisor support includes accommodating a worker’s ﬂexible schedule
or allowing one to bring a child to work on a snow day, which provides more
freedom/ﬂexibility for workers to conduct their tasks (P. R. Jackson, Wall, Martin, &
Davids, 1993). Although coworkers usually do not have such authority in a workplace,
they could provide instrumental support (e.g., trading work shifts to mitigate the impact
of an absence) that makes individuals feel more at ease in such an environment.

In the case of job control (autonomy and skill discretion), coworker support may
allow individuals to perceive an increased a sense of predictability that strengthens their
perceived ability to cope with imposed demands. In addition, workers may learn how to
exercise Skills and autonomy by sharing information and observing their coworkers
(Tansey et al., 2004; Wallace, 2005). Thus, coworker support may enhance and promote

the perception of control over ajob by providing instrumental help to clarify ambiguous

40

task responsibilities, organize work processes, and ﬁnd ways to cope with work overload

(Terry A. Beehr et al., 2001).

The Effects of Job Characteristics on Stress

 

In the D-C model, or nearly any theory of occupational stress, job demands are
considered to be stressors, which bring negative outcomes such as work stress,
exhaustion, depression, and high-blood pressure (C. L. Cooper, 1986; Hackman &
Oldham, 1980; Jex, 1998; R. Karasek & Theorell, 1990). In contrast, autonomy and Skill
discretion are mostly considered as having a beneﬁcial effect on stress since they provide
individuals with opportunities to change the nature of the environment (Daniels & Guppy,
I994). Autonomy is the most important and the most extensively examined concept in
job characteristics studies. According to Hackman and Oldham (1980), individuals tend
to believe that their work outcomes depend substantially on their own efforts, initiatives,
and decisions when autonomy is given. AS autonomy increases, workers are more likely
to feel responsible for their jobs, and consequently “they are more willing to accept
personal accountability for the outcomes of their work,” (Hackman & Oldham, 1980; p.
77). In addition, autonomy may reduce stress in practice by letting individuals design
their work processes that ﬁt their needs/schedules. For instance, Batt and Valcour (2003)
suggest that autonomy in decision making translates into the employee’s ability to control
decisions, such as when, where, and how to integrate work and family responsibilities.

Compared to autonomy, the term skills in the context of job control takes various
forms and measurements, including skill discretion (Dodd & Ganster, I996), skill level

(Frances J Milliken, Martins, & Morgan, 1998), skill variety (Voydanoff, 2004), and

41

learning opportunities (D. Byme, Clore, & Worchel, 1966; Frances J. Milliken & Martins,
1996). However, these constructs share a common assumption that people “from
newborn to mature adults” (Hackman & Oldham, 1980; p. 78) always look for
opportunities to utilize their ability and to gain a sense of efﬁcacy. When people have a
chance to develop and utilize their skills at work, such as problem-solving, taking
initiative, and active Ieaming, it encourages self-conﬁdence and a proactive stance, which
may lead to redued stress levels. The above discussions lead to the following mediating

hypothesis:

HI .' Job characteristics (job demands, autonomy and skill
discretion) mediate the relationship between coworker support and
an individual’s stress. The following hypotheses should hold in
order to conﬁrm the mediating effects of job characteristics:

H I a: Coworker support is negatively associated with stress.

H1 b: Coworker support is negatively associated with job
demands.

H] c: Coworker support is positively associated with
autonomy.

H1 d: Coworker support is positively associated with skill
discretion.

HIe: When job characteristics (job demands, autonomy
and skill discretion) are controlled, the relationship between
coworker support and stress becomes statistically
insigniﬁcant.

Demographic Similarity and Coworker Support

The relationship between demographic similarity and coworker support is
explained by the similarity-attraction framework. As introduced in the literature review,
this framework states that people with Similar backgrounds are more likely to share
common life experiences and values, and are thus more likely to be attracted to each
other (Hodson, 2005). They also ﬁnd that such similarity makes it easier to interact and
to communicate with each other (Williams & O’Reilly, 1998). In fact, Zenger and
Lawrence (1989) found that age and tenure homogeneity are positively associated with
the frequency of technical communication in teams. The similarity also increases group
members’ sense of identiﬁcation and integration with the group (Bond, Galinsky, Prottas,
& Thompson, 2003). Through communications and interactions, workers may ﬁnd it
easier to share their feelings with those who have similar backgrounds, which in turn
fosters trust and cooperation among them (O’Reilly et al., 1993). The above discussion
leads to the following hypothesis:

H2: Demographic similarity such as age, gender, and ethnic
background, is positively associated with coworker support.

43

CHAPTER 5: METHOD

Sam le

The present study uses 2002 data from the National Study of the Changing
Workforce, a survey conducted by the Families and Work Institute. Using random-digit
dialing, a telephone survey interviewed a nationally representative sample of 3,504 adults
who met the following criteria: (I) worked at a paid job or operated an income-producing
business, (2) aged 18 years or older, (3) were in the civilian labor force, (4) resided in the
continental United States, and (5) lived in a non-institutional residence with a telephone.
The estimated response rate was 52.0 percent of eligible households contacted. Self-
employed workers were excluded from the analyses, since it was not clear whether they
had coworkers. This resulted in a sub-sample that includes 2,819 wage and salary

workers out of the total 3,504.

Development of Measures

Since existing data were used for the analyses, measures were constructed from
items asked of the survey respondents. The survey author, the Families and Work
Institute, does not cite all the sources used in developing these items, but an effort has
been made to determine how similar these items are to others used in previous studies. A
complete list of measures, including all items, scales, and alphas, can be found in
Appendix A. As the appendix indicates, there are eight total variables developed for the
analyses: one dependent variable (an individual’s stress), one independent variable

(coworker support), three mediator variables (job demands, autonomy, and skill

44

discretion), and three antecedents of an independent variable (age similarity, gender
similarity, and ethnic similairty).

To develop above measures, the survey instrument was examined for items
related to the constructs of interest. For variables that are constructed from multiple
items, appropriate preliminary scales were created and factor analyses were conducted.
Factor analysis is a technique for analyzing interrelationships among a large number of
items to explain these items in terms of their common underlying dimensions, called
factors (Hair et al., 1998 from Kan). Because the source of dependent, independent, and
mediator variables is the same, all the items used for composite variables were included
in the explanatory factor analyses (EF A) to see if these composite variables are
conceptually distinctive. The results indicate that there are six factors with eigenvalues
over 1.0, which accounted for 57.52 percent of the total variance

The initial pre-rotated matrix indicates that items are most likely to be interrelated
mainly because of the similarity of what was being measured; they are mostly
individuals’ perceptions of the work environment, such as coworker support, job
characteristics and stress levels. Because the constructs were expected to be interrelated,
a varimax rotation was used to simplify the interpretation. A varimax rotation method is
an orthogonal rotation method that is suitable for minimizing the number of variables
with high loadings.

The results of the rotated principal component factor analysis are shown in
Appendix B. As shown in the appendix, all factor loadings were above 0.55 and the
difference between the primary loading and the cross loading for all the items was less

than 2.00, indicating that these six factors are fairly distinctive. Following the factor

45

analyses, alphas were calculated for each composite (multi-item) variable. The ﬁnal

variables are described in the next sections.

Measures
Stress

Stress is assessed with four items from the l6-item global measure of perceived
stress developed by Cohen et al. (1983). The four items assess the degree to which
situations in one’s life are appraised as stressful. According to Cohen et al. (1983), this
measurement shows adequate reliability and is correlated to psychological and
physiological symptoms. A sample item from the survey includes the question, “How
often have you felt nervous and stressed?” There are two additional items that exist in
the survey that were from the global measure of stress, but they were dropped due to low
factor loadings. A higher score indicates a higher level of an individual’s stress.

Cronbach’s alpha for stress is 0.61.

C oworker Support

Coworker support is measured by three items developed for the survey. Although
the survey author does not provide information about how these items were developed,
they are consistent with the type of measurements used in previous studies to tap
respondents’ perceptions of the likelihood of receiving both on-the-job and off-the-job
support (Ducharme & Martin, 2000). As mentioned in Chapter 3, the variable of
coworker support combines one item from instrumental support and an other item from

emotional support, which are, respectively: ‘I have support from coworkers that I need to

46

do a good job’ and ‘I have support from coworkers that helps me to manage my work and
personal or family life.’ In addition, one more item is included that asks respondents if
they feel that they are a part of the group that they work with. As discussed previously, a
sense of belonging is an important part of coworker support that may change the
signiﬁcance of social support itself (Way, 2002; Zacharatos, Barling, & Iverson, 2005).
In fact, some researchers have included similar items into their coworker support
measurement (Cary L. Cooper et al., 2000). Cronbach’s alpha for coworker support is
0.75. All three items are reverse-coded to create a single measure in which a higher score

indicates higher levels of coworker support.

Job characteristics

Job demands, autonomy and skill discretion are measured by items developed for
the survey. They are, however, consistent with the items developed by Karasek (1979).
The job demands construct consists of four items, asking the level of individuals’
workload that they perceive when performing their tasks. Initially, two additional items
were included that asked respondents whether they have to work fast and hard. They
were, however, extracted due to low factor loading. Since one of the four items has a
different response scale (four-point scale ranging from strongly agree (one) to strongly
disagree (four)) from others (ﬁve-point scale ranging from very often (one) to —never
(ﬁve)), all the items are standardized. Cronbach’s alpha forjob demands is 0.77. A
higher score indicates higher level of job demands. Autonomy is assessed by four items
with a four-point response scale ranging from strongly agree (one) to strongly disagree

(four). All items are reverse-coded and Cronbach’s alpha of autonomy is 0.71. A higher

47

score indicates higher autonomy. Finally, skill discretion is measured with four items,
asking about respondents’ opportunity to learn and utilize skills, knowledge, and
creativity. All items are reverse-coded, indicating that a higher score means more

discretion in relationships. Cronbach’s alpha for the variable is 0.71.

Demographic similarity

Demographic similarity is measured by three single-item demographic
characteristics: age, gender, and ethnic background. A sample item; ‘About what
percentage of your coworkers are of people from your racial, ethnic, or national
background?’ Instead of answering the actual percentage, the survey respondents chose a
category that is closest to their actual percentage (e.g. I -100%, 2 -75-99%, 3-50 -74%, 4-
25 - 49%, 5-1 -25%, and 6 - 0%). The problem with the scale, however, is that it is
neither a ratio nor an interval scale. Therefore, the three items are corded as l = equal to

or more than 50% of coworkers, and 0 = less than 50% of coworkers.

Control variables

In addition to the above eight variables, control variables regarding working
conditions are included for the analyses: job tenure, work hours per week, job insecurity,
and occupational types (Census 3-digit occupational codes). Although they are not the
primary focus of the analyses, they could affect the results of the analyses by inﬂuencing
independent, mediator and dependent variables. For example, a number of studies have
found a signiﬁcant correlation between the overall number of hours worked and various

indices of health and well-being. AlSo, Theorell and Karasek (1996) suggest that

48

occupational differences play a signiﬁcant role in determining the level of job control.
According to Hanisch (I999), job insecurity may be one of the single most salient
sources of stress for workers in contemporary workplaces.

In the analyses, job tenure consists of a single item asking respondents about their
length of service in the current line of work. The work hours variable also consists of a
single item asking hours of work per week. .lob insecurity consists of a single item
asking the extent to which respondents are likely to lose their currentjob. Finally, the
respondents’ occupation is measured using a single item asking what their occupation is,
with responses coded on the basis of 2000US Bureau of Census Occupational Codes.
The coded responses are then classiﬁed into seven categories of work: executives,
administrative, and managerial; professional; technical; sales; administrative support;
service; and production, operation, or repair. For the analyses, a dummy variable was
created to represent each occupational category in which 1 indicates ‘applicable’ and 0
indicates ‘all other cases.’

In addition, supervisor support is included as a control variable to avoid the
confounding effect of this variable on the relationship between coworker support and
stress. Supervisor support consists of nine items asking respondents if their supervisors
are supportive in both instrumental and affective ways. Although some of them are
slightly modiﬁed, the nine items are consistent with the type of measurement used in
previous studies to examine supervisor support (Dawn 8 Carlson & Perrewe, 1999). To
be consistent with coworker support, four instrumental support items and ﬁve emotional

support items are combined to create a single variable. Cronbach’s alpha of the variable

49

is 0.90. Items are coded such that a higher score indicates higher levels of supervisor
support.

Variables related to demographic characteristics are controlled, since they are also
potential factors that may affect the results of the analyses. In fact, researchers have
suggested that age and gender play a signiﬁcant role in determining the level of stress
(Broman, Hamilton, Hoffman, & Mavaddat, 1995). Michelle and Gary (2001) found
racial differences in burnout among managerial people. Broman et al. (1995) found that
race, gender, and educational level interacted and affected the level of stress and drinking
problems of blue-collar workers. Other studies have also found that marital status and
family Situations also affect Signiﬁcantly individuals’ psychological strain (Gutek,
Repetti, & Silver, 1988). Although whether non-work situations can be positive or
negative factors remains unresolved, they do have a signiﬁcant impact on an individual’s
stress (David P. MacKinnon, 2000). Based on these arguments, the following variables
are included: gender, age, race, educational levels, marital status, having/not having
children, and having/not having elderly people at home to take care of. Gender is
recorded as a dichotomous variable indicating l = male and 0 = female. Age is also a
single item asking respondents their age. Race consists of ﬁve categories: White, African
American, Native American, Asian, and other. The category White is omitted for the
regression analyses. Education has six categories: Less than high school diploma, high
school diploma or GED, some college, associate degree, bachelor degree, graduate or
professional schools; the category of high school diploma is omitted for the regression
analyses. Respondents are asked about their marital status: married, remarried, living

with spouse, single and never married; divorced. widowed, and separated; the category of

50

single is omitted for the regression analyses. Finally, two items ask respondents if (1)
they are the parent or guardian of any child of any age including, own, stepchildren,
adopted, foster, grandchildren, and (2) they have provided special attention or care for a
person with more than 65 years old within the past year. They are dichotomously coded

such that l = yes and 0 = no.

Analytic Strategies

Two statistical methods are employed for the analyses of the proposed model:
multiple regression for the mediating effects and structural equation modeling (SEM).
Multiple regression is an apprOpriate method for analyzing the relationship between a
Single dependent variable and several independent variables. The purpose of using
multiple regression is to identify and evaluate the mediating effect of job characteristics.
Structural equation modeling, on the other hand, allow for the development and testing of
theories concerning observed and unobserved variables and their relationships. The
purpose of using SEM analyses is therefore to examine and evaluate the overall ﬁt of the
proposed model in Figure l. The major difference between multiple regression and SEM
is that SEM provides a mechanism for explicitly taking measurement error in the
observed variables into account (both dependent and independent variables), though it
does not provide a clear and straightforward relationship between the proposed variables

(David P. MacKinnon & Dwyer, I993).

51

Multiple regression for the mediating eﬂects

Baron and Kenny (1986) state that the following conditions must hold to establish
mediation. First, the independent variable (coworker support) must affect the mediators
(job characteristics). Second, the independent variable (coworker support) must affect
the dependent variable (stress). Third, the mediators (job characteristics) must affect the
dependent variable (stress). If a partial mediating effect of job characteristics exists, the
independent variable (coworker support) should be less signiﬁcant than it was in a simple
coworker-stress relationship without mediators. In addition, if job characteristics are
perfect mediators, coworker support shows no signiﬁcance. In a more mathmatical way,

MacKinnon et al., (1993) introduced the following equations;

Formula 1: Y0 = Bl‘l'l‘Xp'l'el;
Formula 2: YM = Bz+aXp+e2;

Formula 3: Y0 = B3+r’Xp+BXM+€3;

where Y0 is the dependent variable (stress), Xp is the independent variable (coworker
support), and XM is the mediator (job characteristics). [31, [32, and B3 denote the intercept
for Formula 1, Formula 2, and Formula 3, e1, e2, and e3 denote unexplained variability.
The value of the mediated effect of job characteristics equals the difference in the
independent variable coefﬁcients (r-r’) in Formula 1 and Formula 3. If the independent
variable coefﬁcient (r’) is zero when the mediator is included in the model (see Formula
3), then the effect of the independent variable is entirely mediated by the mediating

variable.

52

Structural Equation Modeling

In this dissertation, structural equation modeling (SEM) is conducted using
LISREL 8.80. As mentioned above, this analytical strategy is best suited for examining
whether or not a proposed model explains the relationship between the independent
variable (coworker support) and the dependent variable (stress) by way of mediators (job
characteristics). Two methodological steps are taken for testing the proposed theoretical
model; they are conﬁrmatory factor analysis (CF A) and path analysis. Although SEM
analyses cover both steps, the term SEM refers to path analysis throughout this
dissertation.

Before describing both steps, note that four control variables that have multiple
categories were condensed into dummy variables; they are race, marital status, education,
and occupations. For example, originally there were ﬁve dummy variables for race:
White, African American, Native American, Asian, and others. In the condensed version,
however, race was expressed by a dummy variable (whether the survey participants were
White or not). In the same way, marital status, previously consisting of three dummy
variables, was dummy variable (are the survey participants married or not). In the case of
education level, a dummy variable was created (did the survey participants receive a
bachelor degree/higher degree or not). Finally, seven occupational dummy variables
were compressed into a single dummy variable asking if the participants were
executives/professionals or not. The purpose of simplifying these control variables is to
reduce the degree of freedom of the proposed model, Since having many parameters may
reduce its testability. In Appendix A, these variables are labeled as follows: race 2,

marital status 2, education 2, and occupation 2.

53

Confirmatory Factor Analysis ( C FA ) and Path Analysis

Conﬁrmatory factor analysis (CF A) is used for examining patterns of
interrelationships among several constructs (Raykov & Marcoulides, 2006). While
explanatory factor analysis (EF A) seeks to uncover the underlying structure of a
relatively large set of variables, CF A is used for determining if the number of factors and
the loadings of measured variables on them conﬁrm what is expected. The difference
between EFA and CPA is that EFA does not necessarily require prior theory, while CFA
is conducted based on pro-established theory. In other words, the number of factors in
the model is hypothesized before conducting the analysis in the case of CFA. In this
regard, CFA is the most common method to test the construct validity of survey items.

Path analysis model is used for testing two or more causal relationships among
observable variables (Ku, 2006). Although some researchers do not consider path
analysis models SEM analysis, this dissertation takes the perspective that path analysis
Shares a common underlying idea of model ﬁtting and testing like any other SEM model
(Hu & Bentler, 1995). This dissertation therefore uses the term SEM to indicate path
analysis hereinafter. The goal of conducting SEM is to ﬁnd the most parsimonious
model, which is usually explained by the term ‘ﬁt.’ This term represents the process that
compares the population covariance matrix, 2, to the restricted covariance matrix, 2 (6),
implied by the proposed model (Hu & Bentler, 1995). In other words, for this statistical
analysis, the null hypothesis (H0) is that the population covariance matrix, 2, is equal to

the covariance matrix, 2 (9). based on the proposed model.

Model Fit

In this dissertation, the following three indices are used as methods of evaluating
model ﬁt for both CF A and path analysis: the chi-square (35’), the root-mean-square-error-
of—approximation (RMSEA), and the comparative ﬁt index (CF 1). Although there are
other indices for evaluating model ﬁt, the analyses utilize the above three indices
primarily because they are most commonly used by organizational researchers. Since
each index has both strengths and weaknesses, it is better to use multiple indices for
examining ﬁt in a comprehensive way.

The chi-square (x2) is a test for evaluating the closeness between the sample
covariance matrix, S, and the covariance matrix of the proposed model 2 (6). A large
chi-square, relative to its degrees of freedom, indicates a ‘poor ﬁt’ of the proposed model,
while a small value indicates that the proposed model describes the data well. The
problem of using the chi-square for the proposed model is, however, that the chi-square is
not sensitive to the distribution of the observations when the sample size is large (Ku,
2006). As a result, in the proposed model the null hypothesis is more likely to be rejected.
The remaining two indices, RMSEA and CF], are considered measures that are least
affected by sample size (Browne & Cudeck, 1993).

Root-mean-square-error-of-approximation (RMSEA) is a measure of the
discrepancy between the sample, S, and the model-implied matrix per degree of freedom.

The formula is as follows:

 

2
RMSEA = Jx‘/df.(~ — 1)

where xyz is the chi-square for the tested model, dﬁ is the degree of freedom for the model,

and N is the sample size (Fan, Thompson, & Wang, 1999; Zacharatos et al., 2005).

55

When additional parameters are included in the model, they reduce the value of chi-
square and the RMSEA will decrease to a minimum value of 0 accordingly. However,
because RMSEA considers the degree of freedom (df), its value can increase as
additional model parameters increase. According to Browne et al. (I 993), a value of the
RMSEA of about 0.05 or less would indicate a close ﬁt of the model, in relation to the
degrees of freedom; this evaluation is based on the assumption that an exact ﬁt with the
RMSEA equals 0.00. A value of about 0.08 or less for the RMSEA would indicate a
reasonable error of approximation. Finally, if the value is equal to or greater than 0.1, the
model would not be considered as a good ﬁt.

Finally, Comparative Fit Index (CF 1) compares the covariance matrix of the
proposed model with the one based on the null hypothesis, meaning the covariance

matrix of zeroes. The value of CFI is calculated by the following formula:

2 __ df
CFI = X: r/
j 1721—de

It represents the percent of lack of ﬁt which is accounted for by going from the null

 

model to the proposed model. CFI ranges from 0 to l, with a value greater than 0.9

indicating an acceptable ﬁt (David P. MacKinnon & Dwyer, I993).

56

F it Indices for CFA

In this dissertation, CF A was performed to conﬁrm the theoretical distinctiveness
of the three job characteristics variables. The analysis started with the model in which
twelve items (from V1 to V12) are loaded into three conceptual factors —job demands,
autonomy and skill discretion — based on the Demand-Control model of job
characteristics. Figure 4 shows the results from the initial analysis. The numbers to the
left of the variables (from V1 to V12) are errors while the numbers to the right of the
variables are t—values of each path coefﬁcient. As the ﬁgure shows, all the t-values of the
coefficients are signiﬁcant. In terms of the ﬁt indices, the three indices (chi-square,
RMSEA, and CF I - not shown in the ﬁgure), showed the following results: chi-square =
509.87, RMSEA = 0.065 and CFI = 0.95. These results conﬁrm that the model is
acceptable.

Since, however, the above results indicated that autonomy and Skill discretion
were signiﬁcantly related (t = 35.11 in Figure 4), the alternative model was also
examined in which autonomy and skill discretion were combined into one single factor,
job control (see Figure 5 below). The indices for the alternative model deteriorated
signiﬁcantly: Chi-square = 1201.37, RMSEA = 0.098, and CFI = 0.90. The results
conﬁrmed that autonomy and skill discretion were conceptually distinctive variables,
though they were correlated.

The results of the SEM analyses are presented in the next chapter.

57

32.51
21.12
26.70

28.35
32.25
29.81
28.04
25.91
23.44
32.45
26.35
32.89

32.49
21.00
26.56

28.32
35.06
32.71
32.27

29.94

29.88
34.09
29.77
34.68

V

ﬂ

V2

<
w

<
4:.

<
Ur

V6

V7

<
00

<
o

V

—

0

S

V12

Vl

<<
um

<<
LII-b

<
as

\0

V7

<
00

<

10

<

11
V12

27.29
40.46
36.04

34.30
25.56
29.91
32.21
34.46
37.19
25.84
34.57
24.79

Figure 4: The First and the Final Model of Job Characteristics

Job Demands \

-0.
Job Control ‘/

Figure 5: The Alternative Model of Job Characteristics

27.24
40.37
36.03

34.22
19.63
27.68
28.77

33.45

33.55
23.53
33.73
21.27

58

Job Demands

   
 

  

-3.78

71

Descriptive Statistics

CHAPTER 6: RESULTS

The means, standard deviations, and other descriptive statistics of all variables are

shown in Table 1 below.

Table 1: Descriptive Statistics of All Variables for the Analyses

 

 

N Min Max Mean SD

Stress 2,788 1.00 5.00 2.343 0.777
Std Job Demands 2,777 -l .45 1.55 0.000 0.765
Autonomy 2,777 1.00 4.00 2.949 0.775
Skill Discretion 2,797 1.00 4.00 3.279 0.655
Coworker Support 2,773 1.00 4.00 3.429 0.631
Ethnic Similarity 2,764 0.00 1.00 0.764 0.425
Gender Similarity 2,787 0.00 1.00 0.817 0.387
Age Similarity 2,752 0.00 1.00 0.655 0.476
Gender (male=1) 2,810 0.00 1.00 0.513 0.500
Age 2,785 18.00 99.00 40.947 12.917
White 2.775 0.00 1.00 0.791 0.407
African American 2,775 0.00 1.00 0.104 0.305
Native American 2,775 0.00 1.00 0.010 0.101
Asian American 2,775 0.00 1.00 0.015 0.123
Other races 2,775 0.00 1.00 0.079 0.270
> Highschool 2,810 0.00 1.00 0.110 0.312
Highschool grad 2,810 0.00 1.00 0.308 0.462
Some college 2,810 0.00 1.00 0.206 0.405
Associate degree 2,810 0.00 1.00 0.090 0.287
Bachelor degree 2,810 0.00 1.00 0.195 0.397
Grad or professional degree 2,810 0.00 1.00 0.090 0.287
Married & Partner 2,802 0.00 1.00 0.653 0.476
Single 2,802 0.00 1.00 0.187 0.390
Divorced/Widowed/Separated 2,802 0.00 1.00 0.160 0.367
Have children 2,808 0.00 1.00 0.677 0.468
Have elder people 2,804 0.00 1.00 0.351 0.477
Supervisor Support 2,348 1.00 4.00 3.377 0.632
Job tenure 2,810 0.00 63.00 7.562 8.347
Hours per week 2,795 1.00 l 10.00 44.871 13.214
Job insecurity 2,758 1.00 4.00 2.008 0.947
OCC exectives 2,810 0.00 1.00 0.136 0.343
OCC professional 2,810 0.00 1.00 0.192 0.394
OCC tech 2,810 0.00 1.00 0.040 0.195
OCC sales 2,810 0.00 1.00 0.089 0.285
OCC admin 2,810 0.00 1.00 0.144 0.352
OCC service 2.810 0.00 1.00 0.128 0.334
OCC prod/oper/repair 2,810 0.00 1.00 0.266 0.442
Valid N (listwise) 2,145

 

59

The mean of the composite scale measuring the degree of an individual’s stress is
2.34 on a scale of one to ﬁve, indicating that overall individual workers do not feel
heavily stressed. The mean of the composite scale measuring the degree of coworker
support is 3.43 on a scale of one to four, indicating that, overall, individual workers
receive a fairly good level of coworker support in their workplace. In terms of job
characteristics, the job demands variable is standardized such that its mean is 0.00. The
mean of the composite scale of autonomy is 2.95 on a scale of one to four, while the
composite scale of skill discretion is 3.28; it indicates that respondents in the sample hold
a relatively higher degree of Skill discretion, while they Show moderate degrees of

autonomy.

Correlation Matrix

Examination of the zero—order correlations below (Table 2) indicates that stress is
signiﬁcantly correlated with job demands, autonomy, skill discretion, and coworker
support. The job demands variable is positively associated with stress, while autonomy,
Skill discretion, and coworker support are negatively associated with stress. These
directions of the correlations are consistent with the hypotheses of the model; job
demands were predicted to increase an individual’s level of stress, whilejob control,
autonomy and skill discretion, were predicted to decrease an individual’s level of stress.
No similarity variable (ethnic, age and gender) is found to be signiﬁcantly correlated with
stress. In terms of demographic control variables, all marital status categories (e.g.
married, single, and divorced) Show a signiﬁcant correlation with stress, but in a different

way; while being married or having a partner shows a negative correlation with stress,

60

being Single or divorced demonstrates a positive correlation with stress. Interestingly
enough, having children or elder people at home to take care are not variables that are
signiﬁcantly correlated with stress.

In terms of work—related control variables, ﬁndings are generally consistent with
previous research. For example, supervisor support and job tenure are negatively
correlated with stress, while job insecurity is positively correlated with stress. Also,
employment in some occupations, such as professional vocations, shows a negative
correlation with stress, while employment in service occupations holds a signiﬁcantly
positive correlation with stress. This ﬁnding is consistent across three job characteristics,
implying that certain occupations may affect the relationship between stress and job
characteristics more than others.

Finally, coworker support shows Signiﬁcant correlations with stress and all three
job characteristics proposed in the model. Coworker support is negatively correlated
with stress and job demands, and positively correlated with autonomy and skill discretion.
Also, coworker support demonstrates a Signiﬁcant positive relation with ethnic similarity,
but not with gender and age similarity. Interestingly, coworker support is negatively
correlated with gender, meaning that being male is correlated with decreasing coworker
support. Coworker support shows no signiﬁcant correlation with having children and
older people at home to take care of. However, coworker support does show positive
correlation with supervisor support, conﬁrming the importance of controlling supervisor
support for the regression and SEM analyses. In addition, longer work hours and job

insecurity Show negative correlations with coworker support.

61

Table 2: Correlation Matrix

 

 

I 2 3 4 5

1 Stress I

2 Std Job Demands .240("”") l

3 Autonomy -.207(**) -0.039 I

4 Skill Discretion -.l7l(**) .091(**) .431(**) l

5 Coworker Support —.252(**) -.l80(**) .283("‘*) .382(**) I

6 Ethnic Similarity -0.040 -.055(*) 0.035 0.039 .072(**)

7 Gender Similarity -0.015 -0.001 0.009 0.006 0.025

8 Age Similarity 0.003 -0.003 0.029 0.034 0.022

9 Gender (male=1) -.108(**) -0.028 .076(**) 0.016 -.047(*)
10 Age -.l35(**) 0.009 .083(**) .068(**) 0.013
11 White -0.014 0.030 .080(**) 0.038 .077(**)
12 African American 0.024 -0.017 -.087(**) -.086(**) -.085(**)
13 Native American -0.020 -0.033 0.006 0.029 -0.036
14 Asian American -0.008 0.027 0.000 0.005 -0.027
15 Other races 0.003 -0.027 -0.025 0.029 0.005
16 > Highschool .121(**) -0.020 -.151("‘*) -.060(**) 0.016
17 Highschool grad .046(*) -.073(**) -.096(**) -.146(**) 0.001
18 Some college -0.009 -.046(*) 0.029 -.055(*) -0.040
19 Associate degree -0.040 0.035 0.036 .061(**) 0.037
20 Bachelor degree -.043(*) .070(**) .083(**) .108(**) -0.009
21 Grad or profdegree -.080(**) .068(**) .113(**) .157("”") 0.013
22 Married & Partner -.087(**) 0.042 0.042 .114(**) 0.024
23 Single .046(*) -.077(**) -.054(*) -.106(**) -0.002
24 Divor/Widow/Separat .064(**) 0.027 0.003 -0.036 -0.030
25 Have children 0.036 .062(**) 0.030 .1 l3(**) 0.006
26 Have elder people 0.012 .090(**) 0.033 0.018 0.001
27 Supervisor Support -.277(**) -.301(**) .296(**) .341(**) .560(**)
28 Job tenure -.089(**) .065(**) .102(**) .105(**) 0.029
29 Hours per week 0.005 .264(**) .087(**) .153(**) -.058(**)
30 Job insecurity .215(**) .l36(**) -.150(**) -.173(**) -.212(**)
31 Exectives -0.023 .080(**) .l82(**) .073(**) 0.009
32 Professional -.059(**) .086(**) .085(**) .265(**) .051(*)
33 Tech -0.013 0.032 0.036 0.027 -.044(*)
34 Sales 0.027 -0.003 0.032 -.056(**) -0.030
35 Administrative 0.001 -.078(**) -0.041 -.108(**) 0.019
36 Service .049(*) -.053(*) —.119(**) -.096(**) 0.031
37 Prod/oper/repair 0.023 -.057(**) -.l34(**) -.121(**) -.053(*)

 

** = Correlation is signiﬁcant at the 0.01 level (2-tailed).
*= Correlation is signiﬁcant at the 0.05 level (2-tailed).
Lstwise N=2l45

62

Table 2: Correlation Matrix (cont.)

 

 

6 7 8 9 10

1 Stress

2 Std Job Demands

3 Autonomy

4 Skill Discretion

5 Coworker Support

6 Ethnic Similarity I

7 Gender Similarity .125(**) 1

8 Age Similarity .084(**) .069(**) l

9 Gender (male=l) -.045(*) -0.020 -0.014 1
10 Age .057(**) -0.010 -.080(**) -.086(**) l
11 White .388(**) 0.003 .044(*) -0.015 .109(**)
12 African American -.268(**) -0.015 -0.032 -0.015 -.062(**)
13 Native American -.057(**) 0.040 -.061(**) 0.004 0.001
14 Asian American -.l68(**) 0.010 -0.007 0.006 -0.025
15 Other races -.I86(**) -0.004 -0.007 0.036 -.082(**)
16 > Highschool -0.012 -0.003 -0.037 .095(**) -.070(**)
17 Highschool grad 0.010 .075(**) 0.025 0.014 -0.015
18 Some college -0.035 -0.002 0.032 -0.013 -.047("')
19 Associate degree 0.001 -0.031 0.009 -0.004 0.017
20 Bachelor degree 0.024 -0.020 -0.032 -.078(*"‘) 0.005
21 Grad or prof degree 0.011 -.055(*) -0.012 0.013 .l37(*"‘)
22 Married & Partner .088(**) 0.014 0.036 .059(**) .188(**)
23 Single -.086(**) -0.024 -0.018 .055(*) -.430(**)
24 Divor/Widow/Separat -0.023 0.008 -0.028 -.I36("”") .214(**)
25 Have children .060(**) .059(**) .070(**) -.066(**) .314(**)
26 Have elder people 0.011 0.016 -0.004 -0.004 .111(**)
27 Supervisor Support .087(**) -o.oor 0.023 -.044(*) 0.019
28 Job tenure .051(*) 0.038 0.010 .046(*) .467("”")
29 Hours per week -0.031 0.008 0.018 .247(**) 0.014
30 Job insecurity -.099(**) -0.029 -0.035 0.035 -.076(**)
31 Exectives -0.015 -0.039 -0.006 -0.037 .092(**)
32 Professional .058(**) -0.007 -0.013 -.113(**) 0.036
33 Tech -0.034 0.010 0.009 -0.014 0.026
34 Sales 0.040 .044(*) 0.013 -.055(*) -0.025
35 Administrative 00“ -.083(**) -0.002 -.214(*"‘) 0.017
36 Service -.046(*) 0.008 0.040 -0.014 -.097(**)
37 Prod/oper/repair -0.002 .064(**) -0.024 .356(**) -0.040

 

** = Correlation is signiﬁcant at the 0.01 level (2-tailed).
*= Correlation is signiﬁcant at the 0.05 level (2-tailed).

Lstwise N=2145

63

Table 2: Correlation Matrix (cont.)

 

 

ll 12 13 14 15 16

1 Stress

2 Std Job Demands

3 Autonomy

4 Skill Discretion

5 Coworker Support

6 Ethnic Similarity

7 Gender Similarity

8 Age Similarity

9 Gender (male=l)
10 Age
11 White 1
12 African American -.678(**) 1
13 Native American -.l7l(**) -0.027 1
14 Asian American -.248(**) -0.039 -0.010 I
15 Other races -.573(**) -.091(**) -0.023 -0.033 I
16 > Highschool -.101(**) .104(**) -0.027 -0.039 .061(**) 1
I7 Highschool grad 0.010 -0.025 0.005 -.045(*) 0.032 -.210(**)
18 Some college -.O45(*) .056(**) 0.039 0.007 -0.010 -.169(**)
19 Associate degree 0.036 -0.016 -0.027 -0.029 -0.014 -.106(**)
20 Bachelor degree .047(*) -.053(*) 0.012 0.030 -0.029 -.163(**)
21 Grad or prof degree .049(*) -.056(**) -0.026 .089(**) -.043(*) -.102(“‘*)
22 Married & Partner .087(**) -.l41(**) 0.035 0.019 0.009 0.008
23 Single -.091(**) .112(**) -0.039 0.017 0.014 -0.012
24 Divor/Widow/Separat -0.017 .065(**) -0.004 -.042(*) -0.027 0.002
25 Have children -0.036 .054(*) 0.021 -0.025 -0.003 0.021
26 Have elder people 0.038 -0.022 -0.015 -.047(*) -0.006 0.025
27 Supervisor Support 0.034 -0.040 0.002 -0.018 0.002 0.002
28 Job tenure .074(**) -.060(**) 0.029 -0.017 -.046(*) -.126(**)
29 Hours per week -0.029 0.018 0.007 0.011 0.016 -0.037
30 Job insecurity -.075(**) .087(**) -0.035 -0.007 0.028 .1 14(**)
31 Exectives 0.008 -0.008 0.017 -0.015 -0.002 -.084(**)
32 Professional .076(**) -.075(**) -0.006 .046(*) -.049("‘) -.144(**)
33 Tech 0.007 -0.014 -0.017 0.002 0.010 -0.024
34 Sales 0.009 -0.014 -0.026 0.017 0.003 -0.018
35 Administrative -0.026 .045(*) 0.018 0.004 -0.019 -.119(**)
36 Service -0.038 .050(*) -0.012 -0.007 0.007 .125(**)
37 Prod/oper/repair -0.031 0.008 0.012 -0.039 .052(*) .226(**)

 

*"‘ = Correlation is signiﬁcant at the 0.01 level (2-tailed).
*= Correlation is signiﬁcant at the 0.05 level (2-tailed).
Lstwise N=2 l 45

64

Table 2: Correlation Matrix (cont.)

 

 

l7 18 19 20 21

1 Stress

2 Std Job Demands

3 Autonomy

4 Skill Discretion

5 Coworker Support

6 Ethnic Similarity

7 Gender Similarity

8 Age Similarity

9 Gender (male=1)
10 Age
11 White
12 African American
13 Native American
14 Asian American
15 Other races
16 > Highschool
17 Highschool grad 1
18 Some college -.343(**) 1
19 Associate degree -.214(**) -.l72("‘*) 1
20 Bachelor degree -.330(**) -.266(**) -.l66(**) I
21 Grad or prof degree -.206(**) -.l65(**) -.103(**) -.159(**) l
22 Married & Partner -0.008 -.062(**) 0.017 0.035 0.026
23 Single 0.030 .044(*) -.052(*) -0.014 -0.026
24 Divor/Widow/Separat -0.021 0.034 0.033 -0.031 -0.007
25 Have children -0.001 0.006 0.029 -0.034 -0.009
26 Have elder people -.07l(**) 0.014 0.013 0.001 .053(*)
27 Supervisor Support 0.007 -0.038 0.006 0.018 0.009
28 Job tenure -0.002 -0.014 0.030 0.011 .106(**)
29 Hours per week -0.040 -.053(*) 0.012 .046(*) .101(**)
30 Job insecurity -0.020 0.027 0.010 -.049(*) -.064(**)
31 Exectives -.1 l6(**) -0.005 0.012 .146(**) .060(**)
32 Professional -.245(**) -.148(**) 0.031 .270(**) .339(**)
33 Tech -.075(**) 0.026 .104(**) 0.016 -0.023
34 Sales .068(**) -0.016 -0.019 -0.013 -0.029
35 Administrative .048("') .120(**) 0.000 -0.027 -.091(**)
36 Service .089(**) 0.024 -0.041 -.112(**) -.104(**)
37 Prod/oper/repair .202(**) 0.026 -.051(*) -.256(**) -.177(**)

 

** = Correlation is signiﬁcant at the 0.01 level (2-tailed).
*= Correlation is signiﬁcant at the 0.05 level (2-tailed).
Lstwise N=214S

65

Table 2: Correlation Matrix (cont.)

 

1 Stress
2 Std Job Demands
3 Autonomy
4 Skill Discretion
5 Coworker Support
6 Ethnic Similarity
7 Gender Similarity
8 Age Similarity
9 Gender (male=1)
10 Age
11 White
12 African American
13 Native American
14 Asian American
15 Other races
16 > Highschool
17 Highschool grad
18 Some college
19 Associate degree
20 Bachelor degree
21 Grad or prof degree
22 Married & Partner
23 Single
24 Divor/Widow/Separat
25 Have children
26 Have elder people
27 Supervisor Support
28 Job tenure
29 Hours per week
30 Job insecurity
31 Exectives
32 Professional
33 Tech
34 Sales
35 Administrative
36 Service

37 Prod/oper/repair

22

1

-.662(“)
-.601(**)

.354(**)
0.036
0.003

.135(**)

.084(**)

-.07l(**)

0.035
.0640")
0.01 l

-.058(**)

-.048(*)

-.068(**)

0.041

23

I

-2010")
-.520(**)

-.049(*)
0.023

-.213(**)
-.088(**)
.068(**)

-.044(*)
-0.041
0.005
.049(*)
0.017

.070(**)

-0.030

24

1
.0930")
0.005
-0.028
.052(*)
-o.015
0.021
0.0m
-o.o4o
-0019
0.023
.0450)
0.014
-o.021

25

I
0.019
-0.003

.144(**)
.078(**)

-0.042
0.024
0.014

-0.012

-0.009

-.046(*)

-0.026

0.036

26

I
-0.041

.072(* *)

0.004
0.036
0.013
0.014
0.019
-0.018
-0.005
-0.021

0.003

** = Correlation is signiﬁcant at the 0.01 level (2-tailed).
*= Correlation is signiﬁcant at the 0.05 level (2-tailed).
Lstwise N=2145

66

Table 2: Correlation Matrix (cont.)

 

1 Stress
2 Std Job Demands
3 Autonomy
4 Skill Discretion
5 Coworker Support
6 Ethnic Similarity
7 Gender Similarity
8 Age Similarity
9 Gender (male=l)
10 Age
11 White
12 African American
13 Native American
14 Asian American
15 Other races
16 > Highschool
17 Highschool grad
18 Some college
19 Associate degree
20 Bachelor degree
21 Grad or prof degree
22 Married & Partner
23 Single
24 Divor/Widow/Separat
25 Have children
26 Have elder people
27 Supervisor Support
28 Job tenure
29 Hours per week
30 Job insecurity
31 Exectives
32 Professional
33 Tech
34 Sales
35 Administrative
36 Service

37 Prod/oper/repair

27

I

-0.002
-.087(*"‘)
-.255(**)
0.015
0.033
-0.020
0.036
0.022
0.005

-.078(**)

28

1
.096(**)
-.I 15(**)
.060(**)
.0890”)
.044(*)
-.085(**)
0.000
-.079(**)
-o.034

29

1
-0.002
0760*)
.058(**)
-o.o22
-.o70(**)
-.108(**)
-.108(**)
.104(**)

30

1

-0.039
-.047(*)
-0.037
0.000
-0.006
.053(*)
.05 8(* *)

31

I
-.205(**)
-.os4(**)
-.128(**)
-.165(**)
-.151(**)
-.240(**)

** = Correlation is signiﬁcant at the 0.01 level (2-tailed).
*= Correlation is signiﬁcant at the 0.05 level (2-tailed).
Lstwise N=2145

67

Table 2: Correlation Matrix (cont.)

 

32
1 Stress
2 Std Job Demands
3 Autonomy
4 Skill Discretion
5 Coworker Support
6 Ethnic Similarity
7 Gender Similarity
8 Age Similarity
9 Gender (male=l)
10 Age
11 White
12 African American
13 Native American
14 Asian American
15 Other races
16 > Highschool
l7 Highschool grad
18 Some college
19 Associate degree
20 Bachelor degree
21 Grad or prof degree
22 Married & Partner
23 Single
24 Divor/Widow/Separat
25 Have children
26 Have elder people
27 Supervisor Support
28 Job tenure
29 Hours per week
30 Job insecurity
31 Exectives
32 Professional 1
33 Tech -.104(**)
34 Sales -.159(**)
35 Administrative -.204(**)
36 Service -.187(**)

37 Prod/oper/repair -.298(**)

33

l
-.065(**)
-.083(**)
-.076(**)
-.122(**)

** = Correlation is signiﬁcant at the 0.01 level (2-tailed).
*= Correlation is signiﬁcant at the 0.05 level (2-tailed).

Lstwise N=2145

68

34

l
-.128(**)
-.1 l7(**)
-.186(**)

35

i
-.150(**)
-.239(**)

36

1
-.219(**)

37

Result of the Multiple Regression Mses
Job Characteristics as Mediators

As discussed in the previous chapter, multiple regression analysis is the most
appropriate method for analyzing the relationship between a single dependent variable
and several independent variables. Tables 3-5 below Show the results of the three steps
taken for detecting the mediating effects of job characteristics on the relationship between
coworker support and stress. The three steps correspond to the three conditions described
in the previous chapter, which must be held to establish mediation (David. P. MacKinnon,
1994). First, the independent variable (coworker support) must be statistically signiﬁcant
for the dependent variable (stress); second, the independent variable must be statistically
signiﬁcant for the mediators variables (job demands, autonomy and skill discretion); third
and ﬁnally, the independent variable Should decrease in magnitude when three mediators
are present in the same regression equation.

Results of the multiple regression analyses are organized in the following way.
Table 3 provides the results of the ﬁrst regression equation, in which stress is regressed
on coworker support. This equation also corresponds to a sub—hypothesis, Hypothesis Ia,
proposing that coworker support is negatively associated with stress. Table 4 Shows the
results of the second equation, in which job characteristics are regressed on coworker
support, in order to examine the mediating effect of job characteristics. Since the
proposed model of this dissertation includes three job characteristics, there are three
equations for each mediator variable. These equations also correspond to three sub-
hypotheses, Hypothesis 1b, 1c, and 1d, stating, respectively, that coworker support is

negatively associated with job demands, while coworker support is positively associated

69

with autonomy/skill discretion. Column 3 in Table 5 provides the results of the third
equation, in which stress is regressed not only on coworker support but also on the three
job characteristics. This equation is for examining the third condition, and corresponds to
Hypothesis 1e, proposing that the relationship between coworker support and stress
becomes statistically insigniﬁcant when job characteristics are controlled.

Note that all tables (Tables 3-5) exhibiting regression results include an additional
set of columns named ‘condensed,’ in which coefﬁcients and standard errors of variables
for the SEM analyses are listed. As described in the previous chapter, some control
variables are condensed to a single dummy variable for the SEM analyses in order to
identify the proposed model. The purpose of presenting the results with the condensed
control variables is to show that the results do not differ between the two versions (e.g.,
original and condensed), which is important for comparing the results from the multiple
regressions and the SEM analysis.

Table 3 shows the result of the ﬁrst equation. The coefﬁcient of coworker
support is -.1 81 (-.l77 for the condensed version) with signiﬁcance level less than 0.01
(0.01 for the condensed version), which indicates that the coworker support variable has a
negative and statistically signiﬁcant association with stress. This result indicates that
having higher support from coworkers is associated with lower individual stress levels.
Accordingly, the result supports Hypothesis 1a that coworker support is signiﬁcantly

associated with an individual’s stress.

70

Table 3: Regression Results of Stress on Coworker Support

 

 

 

DV= stress DV = stress
Std. Std.
Variables (original) Coeff Error Variables (condensed Coelf Error
Gender(male=l) -0225 0.034 ** Gender (male=l) -0_2l4 0.032 *“
Age -0.009 0.002 ** Age 0009 0.001 **
African American -013] 0.053 * Dummy White 0,094 0.039 *
Native American -0.191 0.171
Asian American -0.048 0.127
Other races -0.087 0.059
> Highschool 0,255 0_058 ** Dummy > Bachelor -0.091 0.039 *
Some college -0.095 0.043 *
Associate degree -0.l36 0.057 "'
Bachelor degree -0.1 17 0.049 *
Grad or profes degree -0.158 0.066 *
Single 0.120 0.050 "' Dummy Mar & Par 01 19 0.035 **
Divor/Widow/Sep 0.138 0.044 **
Have children 0162 0,039 ** Have children 0.166 0.037 **
Have elder people 0.025 0.032 Have elder people 0,027 0.032
Supervisor Support .0196 0,029 ** Supervisor Support -O.19l 0.029 "*
Job tenure 0.001 0002 Job tenure 0.000 0.002
Hours per week 0,002 0,001 Hours per week 0.001 0.001
Job insecurity 0,110 0,017 ** Job security 0.119 0.017 *"‘
OCC exectives 0.025 0.058 Dummy Ocup -0.031 0.037
OCC professional 0013 0.056 (1 = manager)
OCC tech -0.024 0.084
OCC sales 0.027 0.064
OCC service 0.018 0.060
OCC prod/oper/repair -0016 0,054
Coworker Support -018] 0.029 ** Coworker Support -0.l77 0.030 *"
(Constant) 3.726 0.148 ** (Constant) 3.699 0.140 **
R-Squared 0. I66 R-squared 0.153
N 2,184 N 2,175

 

Table 4 Shows the results of the second equation in which the three mediator
variables (job characteristics) are regressed on the independent variable (coworker

support). The ﬁrst pair of columns represents the results of regressing job demands on

71

coworker support. As the results indicate, the coefﬁcient on coworker support is -.065 (-
.068 for the condensed) at a Signiﬁcance level of 0.026 (0.020 for the condensed),
indicating that having higher support from coworkers is more likely to be associated with
a reduced level of perception of job demands. The second set of columns presents the
results of regressing autonomy on coworker support. Contrary to job demands, coworker
support has a positive association with autonomy (coefﬁcient is .107 at a signiﬁcance
level of 0.000 for the original, and coefﬁcient is .091 at a signiﬁcance level of 0.002 for
the condensed). The results indicate that having coworker support is more likely to
increase the perception of autonomy. Similar results are identiﬁed for Skill discretion; as
shown in the last set of columns in Table 4, the coefﬁcient of coworker support on skill
discretion is .231 (.237 for the condensed) at a signiﬁcance level less than 0.001 (less
than 0.001 for the condensed). This indicates that having coworker support is more likely
to increase positive perceptions of an individual’s skill discretion. The above ﬁndings
support Hypothesis lb, 1c, and 1d, and therefore meet the second condition that the

independent variable must be Statistically signiﬁcant with the mediator variables.

Table 4: Regression Results of Three Job Characteristics (Mediators)
on Coworker Support

 

DV= Job demands

DV=Autonomy

DV=Skill discretion

 

 

 

Original Condensed Original Condensed Original Condensed
Coworker -0.065 * -0.068 * 0.107 ** 0.091 ** 0.231 ** 0.237 **
(0.029) (0.029) (0.029) (0.029) (0.022) (0.022)
(Constant) 0.129 ** 0.033 0.629 ** 0.583 ** 0.882 ** 0.940 **
(0.145) (0.138) (0.142) (0.137) (0.109) (0.104)
R-squared 0.204 0.200 0.268 0.236 0.369 0.351
N 2,196 2,187 2,196 2,187 2,196 2,187
Standard Error in parenthesis.

Control variables are not shown in the table.

* < 0.05, ** < 0.01

Table 5 consists of three columns for both original and condensed regressions

using stress as the dependent variable. The ﬁrst columns (Column 1 of the original and

the condensed) Show the coefﬁcient of coworker support when it is used as a solitary

control variables. The second columns represent results from the regression in which

stress is regressed on job characteristics and other control variables, but without coworker

support. The third columns are the results of the regressions with all variables: controls,

job characteristics, and coworker support.

73

Table 5: Regression Results of Stress on Coworker Support,
Job Characteristics, and
Both Coworker Support and Job Characteristics

 

 

 

Original (DV=stress) Condensed (DV=stress)
Column 1 Column 2 Column 3 Column 1 Column 2 Column 3
Job Demands - 0.187 ** 0.181 ** - 0.192 ** 0.187 **
(0.022) (0.022) (0.022) (0.022)
Skill Discretion - -0.064 " -0.031 - -0.068 "' -0.037
(0.029) (0.029) (0.028) (0.029)
Autonomy - -007 ** -0.06 ** - -0.08*"' -0.073 **
(0.023) (0.023) (0.022) (0.022)
Coworker Support -0.l81 *” - -0.153 " -0.177 " - -0.l45 "
(0.029) (0.030) (0.030) (0.030)
R-squared 0.166 0.185 0.194 0.153 0.176 0.184
N 2,184 2,175

 

Standard Error in parenthesis.
Control variables are not shown in the table.
* < 0.05, ** < 0.01

For meeting the third condition required for a mediating relationship, ﬁrst of all, job
characteristics must be signiﬁcantly associated with stress. In both the original and
condensed regressions, the coefﬁcients of all job characteristics Show statistical
signiﬁcance (see Column 2 of the original and the condensed): coefﬁcient = .187 (.192
for condensed) at a signiﬁcance level of less than 0.001 (less than 0.001 for condensed)
forjob demands; coefﬁcient = -.064 (-.068 for the condensed) at a signiﬁcance level of
0.025 (0.016 for the condensed) for skill discretion; coefﬁcient = -.070 (-.080 for the
condensed) at a signiﬁcance level of 0.002 (less than 0.001 for the condensed) for
autonomy.

When including these job characteristics, the coefﬁcients of coworker support

decrease in value from -.181 (-.l77 for the condensed) in Column 1 to -.153 (-.l45 for the

74

condensed) in Column 3. As introduced in the previous chapter, the mediating effect can

be expressed in the following formula (MacKinnon, 2000):

Formula 4: Mediated effect = r - r’

where r represents the coefﬁcient of coworker support in F omula 1 (see Methods chapter)
where nojob characteristics are present in the equation (-.l81 for the original and -.177
for the condensed) and 1" represents the coefﬁcient of coworker support in Fomula 3
where job characteristics variables are present in the equation (-.153 for the original and -
.145 for the condensed). The mediating effect of three job characteristics is therefore -
0.181 — (-0.153) = -0.028 (for the condensed: -.17l-(-.l45) = 0.032). Since coworker
support still shows statistical signiﬁcance (see Column 3 in Table 3) even with the
presence of job characteristics, job characteristics partially, but not fully, mediate the
relationship between coworker support and stress.

There is a post-hoe test of an approximate signiﬁcance of a mediating effect,
called the Sobel test (Baron & Kenny, 1986). As presented in Figure 6 below, the path
from the independent variable to the mediator is denoted as a and its standard error is 50;
the path from the mediator to the dependent variable is denoted as b and its standard error

IS 8],.

75

 

    

Mediator

 

Figure 6: Mediating Paths

The approximate Signiﬁcance is calculated based on the following formula:
61 * b

\/(b2 * so2 + a2 * sb2 + sa2 * sbz)

2 value =

 

Table 6 shows the results of the Sobel Test for eachjob characteristic variable. Job
demands and autonomy Show a statistically signiﬁcant mediating effect on the
relationship between coworker support and stress, while skill discretion does not

demonstrate a statistically signiﬁcant mediating effect.

Table 6: Results of the Sobel Test for the Signiﬁcance
of the Mediating Effects of Job Characteristics

 

 

Job Demands Autonomy Skill Discretion

a -0.065 a 0.107 a 0.231

sa 0.029 sa 0.029 sa 0.022

b 0.181 b -0.060 b -0.031

sb 0.022 sb 0.023 sb 0.029
Z = -2.l48 ** Z = -2.080 * Z = -1.059

 

* <0.05, ** < 0.01

76

In short, the results described above lead to the conclusion that job characteristics
partially mediate the relationship between coworker support and stress which, in part,
supports Hypothesis 1 (overall). More speciﬁcally,job demands and autonomy in the
proposed model partially mediate the relationship between coworker support and an
individual’s stress, but Skill discretion does not mediate the relationship. Note that the
variance-inﬂation factors (VIF) diagnostics (available from the author) for each
regression analysis indicate that there are some levels of collinearity among the
independent and mediator variables, but they are not serious enough to eliminate

variables from the entire analyses.

Demographic similarity

Table 7 shows the results of the regression estimates in which coworker support is
regressed on three demographic similarity variables: ethnic, gender and age similarity.
Note that no other control variable is included in the regression equation. The coefﬁcient
of ethnic similarity is .126 with signiﬁcance level at less than 0.01. Gender similarity
fails to be statistically signiﬁcant, however (coefﬁcient = .037 at a signiﬁcance level of
0.240). Age Similarity demonstrates marginal statistical signiﬁcance (coefﬁcient = 0.050
at a signiﬁcance level of 0.051). In addition, the R-squared value suggests that less than
1 percent of variance in coworker support can be explained by the three demographic
similarity variables. In short, these results only partially support Hypothesis 2 that
demographic similarity is signiﬁcantly associated with coworker support; only ethnic

similarity shows a signiﬁcant association with coworker support.

77

Table 7: Regression Results of Coworker Support
on Demographic Similarity

 

DV= Coworker Support

 

Coef Std. Error

 

Ethnic Similarity 0.126 0.029 **
Gender Similarity 0.037 0.031
Age Similarity 0.050 0.026
(Constant) 3.268 0.037 **
R-squard 0.009

N 2,693

 

Control variables are not included in the equation.
* > 0.05, ** > 0.01

Results of Structural Eggation Modeling
Model Identiﬁcation

Before conducting SEM analyses, the proposed model has to satisfy a requirement
for model identiﬁcation. To check the necessary conditions for model identiﬁcation, a
calculation is performed to count the number of independent parameters in the model and
subtract them from the number of non-redundant elements in the sample covariance

matrix. The product of the calculation represents the ‘degree of freedom.’

Degree of freedom = p (p + I) / 2 — (number of model parameters)

78

where p is the number of observed variables to which the model is ﬁtted (Raykov &

Marcoulides, 2006). The model parameters in the proposed model are calculated based
on (I) the variance and covariance of the exogenous latent factors, measurement errors,
and disturbances, (2) the direct effects on the indicators from factors (e.g., factor loading).
and (3) the direct effects on latent endogenous factors from other factors (e.g., path

coefﬁcients). The model parameter is therefore:

ll9+5+7+15=l46

The degree of freedom of the proposed model is:

Degree of freedom (df) = (20 * 21) / 2 — 146 = 63

Since the degree of freedom is nonnegative, the proposed model meets the requirement

for identiﬁcation called the “t-rule” (Raykov & Marcoulides, 2006).

Fit Indices

AS mentioned in the previous chapter, the purpose of using structural equation
modeling (SEM) is to examine the overall ﬁt of the proposed model. Figure 6 shows the
t-value of each path coefﬁcient of the model. In fact, the predicted outcomes of the
relationships among demographic similarity, coworker support, job characteristics, and
stress are quite similar to the predicted outcomes obtained from the multiple regression

estimates. For example. the t-values of the path coefﬁcients for gender and age Similarity

79

are not Signiﬁcant (t = .75 for gender, t = .54 for age). Coworker support demonstrates
both a direct effect and an indirect effect through job characteristics (t = -4.97 for the
direct effect of coworker support). Also, the relationship between skill discretion and
stress is non-signiﬁcant (t = -.161) when coworker support is controlled. AS shown in
Figure 6, gender and age similarity did not Show a Signiﬁcant relationship with coworker
support. Similar to the results from the multiple regression estimates, several control
variables, including race, having elderly people to care for at home, job tenure, work
hours, and occupation also did not Show signiﬁcant relationship with stress. The
three ﬁt indices, the chi-square, the comparative ﬁt index (CF I), and the root-mean-
square error of approximation (RMSEA), are used to examine the theoretical ﬁt of the
model. The ﬁt indices however showed a signiﬁcantly poor ﬁt of the proposed model
(Chi-square = 1854.91 , CF I = 0.72, and RMSEA = 0.12), despite the similar results
regarding the Signiﬁcance of the path coefﬁcients with those of regression analyses.

Reasons for the unsatisfactory ﬁt of the proposed model are discussed in the next chapter.

80

 

32.56 3.29
32.56 0.75 32.56
32.56 —m 0.54
_,,,,
.5.

32.56 A -5.84

ge 13.69
32.56 189

2.
32.56 Edu -2.52 AU’Onom 1930356
32.56 -3.61
' ' 32.5

. 9. 0 0 Skill Disc
32.56 Children 4.58
32.56 Eldery 0.45 373
32.56 -4.96 16]
32.56 —0.13
32-56 —-’_ -0-44 32.56
32.56 —>_ 5.48
32.56 —>_ -0.81

Figure 7: The results of the SEM analyses for the proposed Model

81

CHAPTER 7: DISCUSSION

Discussion of the Results

This dissertation proposed a theoretical model that was developed based on two
well-known theories, the Social Information Processing (SIP) approach and the Demand-
Control (D-C) model, to examine the relationship among demographic similarity,
coworker support, job characteristics, and an individual’s stress. In the proposed model,
coworker support was considered a social cue that individual workers receive from their
work environment. The effect of coworker support was then explained by the SIP
approach, which proposes that individuals’ perceptions are constructed by the social
environment or by the network of social and informational relationships. Such socially-
constructed job characteristics in turn are predicted to affect the level of an individual’s
stress. The effects of job characteristics on stress were explained by the D-C model,
which proposed that job demands and job control (consisting of autonomy and skill
discretion) were critical determinants of individuals’ stress levels. Finally, the model
proposed possible antecedents of coworker support: demographic Similarity in terms of
age, gender, and ethnic background. The relationship between demographic similarity
and coworker support was explained by the similarity-attraction paradigm, which was
treated as part of the SIP approach. By integrating the D-C model and the SIP approach,
the proposed model tested the main hypothesis (Hypothesis 1) thatjob characteristics
mediate the relationship between coworker support and an individual’s stress.

AS predicted in a set of sub-hypotheses in Chapter 5 (e.g., Hypothesis la through
1d), coworker support showed a signiﬁcant association with stress as well as three job

characteristics: job demands. autonomy and skill discretion (see Table 3 and 4 in Chapter

6). These results may support the notion that coworker support is one of the social cues
that affect individuals’ perceptions of job characteristics.

The coefﬁcient of coworker support decreased in value when three job
characteristics were present in the same regression equation (see Table 5 in Chapter 6),
but the strength of its direct effect on stress remained statistically signiﬁcant. This result
partially supports the Hypothesis 1e, which proposed that coworker support would lose
its statistical signiﬁcance when job characteristics were present in the same equation.
This result also indicates that job characteristics partially, but not fully, mediate the
relationship between coworker support and stress. The statistical signiﬁcance of the
partial mediating effects were conﬁrmed by the results from the Sobel test regarding job
demands and autonomy.

All three job characteristics showed a signiﬁcant association with stress when
coworker support was not present in the equation (see Column 2 in Table 5). However,
skill discretion lost its signiﬁcance when coworker support was present in the regression
equation. This indicates that the variance in skill discretion may be shared in large part
with coworker support, though the conceptual distinctiveness between coworker support
and skill discretion was methodologically conﬁrmed by factor analyses. In addition, the
variance inﬂation factor of the regression analyses did not indicate serious
multicollinearity for these variables. However,job demands, autonomy and skill
discretion showed high correlations with each other (see Table 2 in Chapter 6), which
may lead to instability of the estimation regarding the relationship among variables.

Based on the above arguments, the ﬁndings from the multiple regression analyses

partially support the main hypothesis (Hypothesis 1), stating thatjob characteristics

83

mediate the relationship between coworker support and an individual’s stress. The term
partially indicates the lack of statistical signiﬁcance of skill discretion as a mediator on
the relationship between coworker support and stress.

Hypothesis 2 states that demographic similarity affects the level of coworker
support. The result from the multiple regression analysis, however, only supported this
hypothesis for ethnic background, which seems contradictory to previous ﬁndings. Such
an insigniﬁcant relationship between age/gender similarity and coworker support may
result from the nature of the measurement. In fact, scales for the similarity items are
difﬁcult to use directly for a quantitative purpose, since they do not have equal intervals,
even though each scale is assigned to a range of percentages (see Appendix A for detail).
Consequently, these three items were treated as dummy variables in the multiple
regression equations, which lost considerable variance among survey participants.
Variables with better scales would have helped obtain better results.

Although the results from the multiple regression analyses generally support the
main hypothesis (Hypothesis 1), the structural equation modeling (SEM) did not support
the model from a ‘theoretical ﬁt’ standpoint. The overall ﬁt indices, such as chi-square,
RMSEA, and CFI, indicated a poor ﬁt. The following sub-section discusses possible
reasons for the discrepancies between the results from the multiple regression analyses

and that of the SEM analyses, as well as reasons for the poor ﬁt of the model.

Reasons for the Poor F it
AS described in the method chapter, the major difference between multiple
regression analysis and SEM analysis is that SEM analysis provides a mechanism for

taking measurement error into account for both dependent and independent variables

84

(Raykov & Marcoulides, 2006). In contrast, multiple regression analyses do not consider
potential measurement error of independent variables. Since the independent and
mediator variables of the proposed model are perceptions of support and job
characteristics, they are highly likely to contain measurement error.

In addition, as shown in the correlation matrix (Table 2 in Chapter 6), coworker
support, job characteristics, and stress are highly correlated variables despite being
theoretically distinctive. These high correlations may result from the cross-sectional
nature of the data set. In fact, all the information to develop dependent, independent, and
mediator variables were collected at a single point in time from a single source. Using
the same source for both dependent and independent variables may posses some potential
ﬂaws, such as the risk of method variance (Jenkins, Nadler, Lawler, & Cammann, 1975).
Common method variance is a form of systematic error contamination due to the method
of measurement rather than the trait (Charles A. O'Reilly, Parlette, & Bloom, 1980). The
primary problem associated with using common methods is the overlap of the method
constructs, which is especially problematic when respondents associate the dimensions of
one variable while answering another. Such measurement error may make the regression
results different from that of the SEM analyses.

The reliability of measures used in this dissertation may have contributed to the
less-than-satisfactory SEM results. For instance, coworker support, job demands,
autonomy, and skill discretion are acceptable, but not highly reliable variables (alphas
between 0.70 and 0.80). In terms of the dependent variable (stress), the reliability is even
lower than other variables (alpha = 0.61). Since the dataset is not designed for this

particular study, these composite variables do not necessarily consist of all items that are

85

suggested in the previous studies, which is part of the weakness of using a secondary
dataset.

In addition, this dissertation did not differentiate emotional support from
instrumental support of coworkers, but emotional and instrumental support may affect
individuals’ stress differently. For example, emotional social support may contribute
more to developing a sense of stability in one’s life situation, which would affect job
demands more than autonomy and Skill discretion. On the other hand, instrumental
support, such as practical advice and survival tips in the workplace, may contribute more
to developing a sense of predictability, which would affect autonomy and skill discretion
more than job demands. If this is true, the combined measure of social support may
affect speciﬁcation error of the SEM analysis.

As mentioned in the method chapter, the dependent variable of the proposed
model consists of items called ‘global measurement of stress,’ which asks the overall
stress level of individuals in their life as a whole. Using this variable for the analyses
Should be advantageous for this dissertation from a ‘common-method variance’
standpoint, Since, unlike more work-related outcomes (e.g., occupational stress), the
survey participants would be less likely to associate theirjob characteristics perceptions
with their stress level. Also, they would be less likely to associate their evaluation of
coworker support with stress. However, this global stress measure may not be explained
by workplace factors such as coworker support and job characteristics. Additional work-
related variables such as occupational stress, job satisfaction, and work-family conﬂict

would have been better explained by the proposed independent and mediator variables.

86

Future Research

Several issues are pointed out for future research. First, this dissertation
examined a US sample. US society is usually considered an individualistic society where
priority is placed on self-reliance and personal well-being (Hofstede, 1980). People in
such a society focus more on individual goals and personal attributes. People in
collectivistic societies, on the other hand, give more priority to group goals welfare.
They also put more emphasis on group identities, harmony, and cooperation (Hofstede,
1980). When comparing people in individualistic societies with collectivistic societies,
the magnitude of the coworker effect may be more signiﬁcant in collectivistic societies.
In other words, social cues, such as support form coworkers, may have more impact on
individuals’ perception of job characteristics in a collectivistic workplace than in an
individualistic workplace. Future research on coworker support Should include factors
such as cultural differences (e.g., individualistic vs. collectivistic) that may affect
individuals’ perception of their job characteristics, which may consequently inﬂuence
their stress levels.

Second, this dissertation focused on coworker support and intentionally excluded
supervisor support, but it did not address theoretical differences nor did it test the
difference empirically. There may be explanations to support the difference between
coworker and supervisor support. For example, workers may use coworker support as a
reference point to build their perceptions of job characteristics. In other words, workers
interpret their work environment by observing their colleagues. In such cases, coworker
support is an antecedent of job characteristics. In contrast, workers may see supervisor

support as encouragement to adjust their perceptual association of job characteristics with

87

stress. In this case, social support from supervisors is not an antecedent, but a moderator
that changes the strength of a stressor-strain relationship. Future research is necessary to
ﬁnd a theoretical rationale to explain the difference between supervisor and coworker
support in terms of how individuals understand and interpret such support.

Finally, the dataset used in this dissertation consists of a nationally representative
sample of working adults. Using a nationally representative sample has strength in
allowing the generalizing the empirical ﬁndings. However, large heterogeneous samples
may include the confounding effects of socio-economic status (Cappelli & Sherer, I991;
Clegg, 1984; de Rijk et al., 1998; Morgeson & Campion, 2003; Parker et al., 2001).
Analyses for speciﬁc groups of people, in addition to analyses with a whole sample, may
overcome the weakness of using large representative samples. The analyses for speciﬁc
groups of people may also help reﬁne the effect of job demands, autonomy and skill
discretion. Future studies should therefore investigate theoretical foundations for

grouping people within a representative sample.

Contributions of the Study

Although the overall ﬁt indices of the SEM analyses showed unsatisfactory
results, the ﬁndings from the regression analyses at least support the notion that coworker
support may affect perceptions of job characteristics, which consequently affect
individuals’ stress level. Instead of focusing on the moderating/buffering effect of social
support, this dissertation proposed that social support was an antecedent of stressors (e.g.,
job characteristics). In other words, social support is a social cue from the environment,

which can be explained by the Social Information Processing (SIP) approach. The

88

empirical results generally conﬁrmed the role of coworker support as an antecedent of
job characteristics, as well as the role of job characteristics as mediators. In this regard,
the proposed model explained ‘conceptually and methodologically’ the effect of
coworker support on an individual’s stress. This ﬁnding provides a solution for the
considerable debate over how social support affects an individual’s stress.

As mentioned in the previous chapters, studies of job characteristics have often
been criticized for failing to reﬂect on factors that inﬂuence and constrain individuals’
perceptions of job characteristics (Cappelli & Sherer, 1991; Rousseau, I978). Salancik
and Peffer (1978) argued that there have been few attempts to explore social/situational
(rather than dispositional) factors as an alternative explanation for work attitudes and
behaviors. This dissertation introduced social support from coworkers as a social cue,
and the results showed that this does affect individuals’ perceptions of job characteristics.
In recent years, research interest in job characteristics appears to be declining (Morgeson
& Campion, 2003). However, by expanding and broadening the focus of ajob
characteristics paradigm, this dissertation provides a way of revitalizing the research on
job characteristics. In fact, some researchers have tried to consider ajob characteristics
theory in a larger context, and obtained evidence of a potential mediating effect of job
characteristics. Piccolo and Coloquitt (2006) found that transformational leadership was
associated with the way individuals View their jobs, which in turn affected their
performance and organizational citizenship behavior (OCB). By integrating ajob
characteristics theory with other well-established theory, this dissertation opens up a new

domain for the potential mediating effect of job characteristics.

89

Implications

These results have several implications. First, individual workers in a
contemporary workplace are required to have more interactions with their colleagues due
primarily to structural and demographic changes in the workplace discussed in the
previous chapter. While such interactions with coworkers may be a source of support,
they may also become a potential source of stress (Piccolo & Colquitt, 2006). In other
words, whether they are a source of stress or a source of social support depends on
whether or not the environment facilitates positive interactions among coworkers. For
example, organizations can develop a buddy system for on-the-job training in which
senior workers are paired with new workers to teach knowledge and skills, as well as
informal information (e.g., tips for dealing with a particular boss).

Management can also utilize existing systems such as Kaizen/Teian meetings for
the purpose of promoting a supportive environment. Originally, such meetings were
developed as a way to make suggestions for increasing productivity in a plant or an
organization. However, these meetings can also be used for improving individuals’ job
skills and knowledge. In addition, speciﬁc off-the-job training, such as training for
improving interpersonal skills, is an effective way to create positive interactions among
workers. These implications can also be applicable to team building within an
organization. Team members usually have more interactions than those not in a team
setting. Particularly in a self-directed team, individual members may be Signiﬁcantly
inﬂuenced by their team members since they are the dominant source of information for
accomplishing their tasks. Under the circumstances, arranging a better structure/system

that promotes interactions may facilitate access to practical as well as emotional support

90

from other team members. Such systematic efforts to promote both instrumental and
affective networks among team members will enhance workers’ affect and may
consequently improve job performance.

Traditionally, studies of job characteristics have been analyzed by the ‘need-
satisfaction paradigm,’ which emphasizes individuals’ predisposition for explaining the
effect of job characteristics (Morgeson & Campion, 2003). Consequently, these studies
have looked to potentially expensive and time-consuming practices such as job
enlargement or enrichment to promote and improve individuals’ perception of job
characteristics (Piccolo & Colquitt, 2006). However, the empirical results of this
dissertation conﬁrmed the notion that not only predisposition, but also social construction
of information provided by coworkers could determine the perceptions of job.
characteristics (R. W. Grifﬁn et al., 1987; Piccolo & Colquitt, 2006). Job design,
therefore, could be improved by paying attention to the ethnic composition of workers,
running training for interpersonal skills, or promoting cultural change within an
organization (Parker et al., 2001).

Note, however, that the results of the analyses do not Speak to a causal connection
among coworker support, job characteristics, and stress given the nature of the data set
used for the analyses. Therefore, the implications presented in this section rely on the
assumption that the association among coworker support, job characteristic, and stress

does represent a causal relationship.

91

Conclusion

The purpose of this dissertation was to explain the effect of coworker support on
an individual’s stress in both theoretical and methodological ways. By introducing job
characteristics as mediators, the proposed model attempted to explain how coworker
support affects an individual’s stress. Although the empirical results from the analyses
did not conﬁrm a theoretical ﬁt of the proposed model, they generally supported each of
the theoretical relationships: demographic similarity and coworker support, coworker
support and job characteristics, and job characteristics and stress. In other words, the
results indicated that individuals’ perceptions of job characteristics were partially
constructed by coworker support. Job characteristics in turn affected an individual’s
stress level. In addition, ethnic Similarity may affect the level of coworker support.
These empirical ﬁndings suggest that creating supportive interactions among workers
may not only reduce individuals’ stress levels, but also have positive consequences for
workers in terms of developing perceptions of job characteristics.

As mentioned previously, having a less stressed workforce is also beneﬁcial for
organizations. Stressed workers can lead to occupational accidents, legal challenges, and
compensation for stress-related illness. In a situation where workers have more
opportunities to work with their colleagues, maintaining positive social relations among
workers will thus be the most critical and cost-effective way to create a win-win situation

for both management and workers.

92

Appendices

93

APPENDIX A: Variable Descriptions and Methods of Construction

 

 

Variable Name Descriptions and Method of Construction
Dependent Variable
Stress This variable consists of four questions. It is deﬁned as the
sum of the four items divided by 4. (Reverse coded: Alpha =
0.61.)

* QPW3: How often have you felt nervous and stressed?
* QPW4: How often have you felt that you were unable to
control the important things in your life?
* QPW5: How often have you felt conﬁdent about your
ability to handle your personal problems?
* QPW6: How often have you felt that things were going
your way?
(1 = very often, 2 = often, 3 = sometimes, 4 = rarely, 5 =
never)
Independent Variables
C oworker Support
This variable consists of three questions. It is deﬁned as the
sum of the three questions divided by 3. (Reverse coded:
Alpha = 0.76.)
* QWC27: I feel I am really a part of the group of people 1
work with.
* QWC28a: I have the support from coworkers that I need
to do a good job.
* QWC28b: I have support from coworkers that helps me
to manage my work and personal or family life.
(I = strongly agree, 2 = somewhat agree, 3 = somewhat
disagree,4 = strongly disagree)
Mediator Variables
Job Demands
This variable consists of ﬁve questions. It is deﬁned as the
sum of the four standardized items divided by 4. (Reverse
coded: Alpha = 0.77.)

* QWC 12: I never seem to have enough time to get
everything done on my job.

(1 = strongly agree, 2 = somewhat agree, 3 = somewhat
disagree,4 = strongly disagree)

* QWC53: Thinking about your main job, tell me how
often you have felt overwhelmed by how much you had
to do at work in the last 3 months?

* QWC54: And how often in the past 3 months, have you
been asked by your supervisor or manager to do
excessive amounts of work?

"' QWC56: During a typical workweek, how often do you
have to work on too many tasks at the same time?

(1 = very often, 2 = often, 3 = sometimes, 4 = rarely, 5 =
never)

94

Job Control
Autonomy This variable consists of four questions. It is deﬁned as the
sum of the four questions divided by 4. (Reverse coded:
Alpha = 0.71.)

* QWC I: l have the freedom to decide what I do on my
job.

* QWC4: It is basically my own responsibility to decide
how my job gets done.

* QWC7: I have a lot of way about what happens on my
job.

* QWC9: I decide when I take breaks.

(I = very true, 2 = somewhat true, 3 = a little true, 4 =
not at all true)
Skill Discretion This variable consists of three questions. It is deﬁned as the
sum of the four questions divided by 4. (Reverse coded:
Alpha = 0.71.)

* QWC3: My job requires that I keep learning new things.

* QWCI I: My job requires that I be creative.

* QWCI3: My job lets me use my skills and abilities.

(I = strongly agree, 2 = somewhat agree, 3 = somewhat
disagree, 4 = strongly disagree)

* QWC47: How true is the following statement about your
job: 1 have the opportunity to develop my own special
abilities.

(I = very true, 2 = somewhat true, 3 = a little true, 4 =
not at all true)
Antecedent of coworker support
Demograhic Similariﬁ
There are three single-item variables regarding similarity
(ethnic background, age, and gender). They are recoded to 1-
0 variable, indicating that 1 = more than 50% are similar / 0
= less than 50% are similar.

Ethnic similarity * QWC43: About what percentage of your coworkers are
of people ftom your racial, ethnic, or national

Gender similarity * QWC43a: About what percentage of your coworkers are
people of your sex?

Age similarity * QWC43b: About what percentage of your coworkers are

within 10 years of your age — either younger or older?
(1-100% of coworkers / 2-75 through 99% / 3-50 through
74% / 4-25 through 49% / 5-Less than 25% but more

__ i__than 0 / 6-0%)

 

 

Control Variables
Working environment

Job experience QEB38: How long have you worked for this employer/ been
involved in this line of work?
Work hours QEB38: How long have you worked for this employer/ been

involved in this line of work?

95

Job insecurity QEB37r: How likely is it that during the next couple of years
you will lose your present job and have to look for ajob with
another employer? (1 = very likely, 2 = somewhat likely, 3
= not too likely, 4 = not al all likely: Reverse coded)

Occupations Creating 7 dummy categories based on the 3 digit codes
from US bureau of the census (occupat / oc80)
Executive, administrative, and managerial occupations
(003-03 7)
Professional specialty occupations (042-199)
Technicians and related support occupations (203-235)
Sales occupations (243-285)
** Administrative support occupations, including clerical
(303-3 89)
Service occupations (403-469)
Production, operators, and repair occupations (473-889)
0ccupations2 Creating a dummy; 1 indicates executives and professional
(003-199), 0 indicates other (203-899)
Supervisor Support This variable consists of nine questions. It is deﬁned as the
sum of the nine questions divided by 9. (Reverse coded:
Alpha = 0.90.)
"‘ QSUP6: My supervisor or manager keeps me informed
of the things I need to know to do my job well.
* QSUP7: My supervisor or manager has expectations of
my performance on the job that are realistic.
* QSUP8: My supervisor or manager recognizes when I do
a good job.
* QSUP9: My supervisor or manager is supportive when I
have a work problem.
* QSUPIO: My supervisor or manager is fair and doesn't
show favoritism in responding to employees' personal or
* QSUPI 1: My supervisor or manager accommodates me
when I have family or personal business to take are of --
for example, medical appointments, meeting with child's
teacher, etc.

* QSU‘PIZ: My supervisor or manager is understanding
when I talk about personal or family issues that affect my
work.

* QSUP13: I feel comfortable bringing up personal or
family issues with my supervisor.

* QSUPI4: My supervisor or manager really cares about
the effects that work demands have on my personal and
family life.

(I = strongly agree, 2 = somewhat agree, 3 = somewhat
disagree,4 = strongly disagree)
Demogmphic characteristics

Gender QSC8: Gender (1 = male, 0 = female)
Age QPDI: May I ask how old you are?
Race QPD4: Race (white, black, A-Indian, Asian, other/mixed)

96

**

Race2
Education

**

E ducation2
Marital status

**

Marital status2

Having children at
home

Having elderly people
at home

White

African American

Native American

Asian

Other races
1 indicates white, 0 indicates all other.
RED6: Highest level of schooling completed

Less than high school diploma

High school diploma or GED

Some colleges

Associate degree

Bachelor degree

Graduate or professional schools
1 indicates more than bachelor degree, 0 indicates all other.
QEB38: How long have you worked for this employer/ been

Married for a ﬁrst time, remarried, and living with

someone as a couple.

Single and never married

Divorced, widowed, and separated
1 indicates married, 0 indicates all other (single, divorced,
separated, widowed).
QEN 1: Are you the parent or guardian of any child of any
age? Please include your own, stepchildren, adopted, foster,
grandchildren, or other for whom you act as a parent ( l =
yes, 0= no)

QECI: Within the past year have you provided special
attention or care for a relative or in-law >65 -- helping w/
things that were difﬁcult or impossible for them to do
themselves? (I = yes, 0 = no)

 

97

 

 

 

 

 

83- Name Se... ea; Roe 3.3 2255.. pee 2.3... be a: us as 8.. a: $630
33. 986. a memo nae Sod cases on _ a5 8:32 no» a: u. 630
Sea 35o g ass- mm _ .o 235 use see mass. 83 _ as cases no.4 o2 N$25
8:285 .55 .4
ES. Bod Red. 3.2. oz; coed .385 one _ 5:3 oeeoe _ ”830
Sod- 2 3 £2 E 23.. g 3 So» be S 28%; :23 gene as up 22 e 22 _ ”530
~85 ace Sod wool... e85 $3 .88
2 S- :3 82 mi See. em; so» be :6 oe _ see; page a eeeooe 2: use _ “630
»Eo=3=< .m
33 $3- a; weed ..Ieﬂ See- ape: pea 2: a was baa es
co 3:95 8 05: so» on cote Bo: £83.58? Eoib m MESQ ”$030
mooo o_o.o Nooo Goo a E _.o- $.83 oo 358:8 033088 ow 2 comm—8E .8 ..oﬂigg So»
»n coo—me soon so» eta: .3288 m “men 2: E cote Bo: 92. ”$030
83. :3 ES- See- mg 32. $588 m as. us a 5.83 a ee 8 e2 so» 52: so; 3 85.2263
zoo 03: so» Soc 3o: 2: :2 do» :38 So» Soon wet—SE. ”$030
83 ”Nos. was 83. I51... 33. so»
»E :o econ wasp—Ego How 8 2:: cmzoco gm: 9 Eoom .85: _ ”N630
2.52:0»— aoo .N
ﬂel... 33- £3. mood- See 83. as; so» meow so; owes as ea 3% use come 25: HEEG
g owed Sod- $3. 83. age. $6285 .2628 so»
298: 9 bEnm So» Eons Econ—80 zoo so» 023 Ste >5: ”man—O
Wei... e23- 33 Rod. mm; :55- Se: so» e 0&5 288$
05 35:8 8 032: 225 :o» 85 :oo so» 32 cote 3o: ”van—O
Neel... NN _ .o- wood- o as- some BS. seamen es. 395: go so» use 5% Bo: ”ego
82% .—

e m .. m a _ use:

 

£9223. ..Soem 3259.80 .363...— couﬁom 2: .3 $.33— “m 5:73:53.

98

 

 

 

 

 

mo_.o- mid ooﬁo om ..o oooo- mono do: »=an use .9883 »E :o 96; 8.38% x53
85 Source 05 Sons 330 »=e2 commence co coﬂtoozm »2 ”3osz

33. at: .86 we _ .o :56. a .5283 be

53> mosmﬂ €an .o .2088 a: wEwEB oEetﬂEoo .ooo _ um 7:50

83. was secs 2 3 was- $3 .083 be some as can see a 88.8
Sega is _ 5:3 wﬁncﬁﬂoocs m_ cowacaE co .oﬂtoozm »2 ”NEDmO

mm _ .o- wmos nae ee _ .o e85- Neel... ea 8:82 8:8 sE mesooe £88588
30:58 63:85 So .. me 2a 8.8 8 3.2.6.5 .9533 co »=E£

3m; _ 5:3 2: 8835888 commons: co c8339; »2 H. EDmO

owoo- om_.o mnoo mm _ .o o. _.or ad .mooo: »=E£ co Ecoﬂoo .moo»o_oEo 9 wﬁocoomm: E
EmEcoZo Beam remooo was So ﬂ commas: ._o coﬂtoozm »2 ”oEDmO

was- 3:6 was 085 owes- «El... .5288
age- was ewes woos sod- g .8» seem a 8 _ 8:3 8888. page a 82:83 a: ”228
see. we; mos w :3 83- as amass. ea as 8.. es 8

oocmEcotom »E Co 228835 was homage co aoﬂioozm »_>_ KmDmO

e2; 83 83 53. £3. Hole .83 8.25 8 e as: 2 Bee
_ meE 2: oo BEQE oE 88x cowmcmE co .8389; »_>_ Hon—Dmo

tooozm Sarcasm .o

ewes- lis 83 em _ .o Need- a _ 2 .6: 38s a .828 ea sea
»E owacaE 2 2: 8.0; :2: 38.838 :5: tooosm 3a: _ “anOBO

89o- as... am _ .o was $3. a _ no .8»
33. E... was See 83. 2 no .83 sea _ 2888 deem use :8 e baa ea _ as _ uRos/o
l r - - i tooosm ..ofokoUl ..m
we? 82 En: ego Rod. 33 82:8 38% :8 be 8.28 e 38:88

2: gm; _ 52.50» Sosa EoEBSm wEBoz£ 2: mm 0?: Bo: US$30

99

References

100

References:

Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.), Advances in
experimental social psychology (pp. v.). New York,: Academic Press.

Ang, S., Van Dyne, L., & Begley, T. M. (2003). he employment relationships of foreign
workers versus local employees: A ﬁeld study of organizational justice, job
satisfaction, and OCB. Journal of Organizational Behavior, 24(5), 561-5 83.

Appelbaum, E., Bailey, T., Berg, P. B., & Kalleberg, A. L. (2000). Manufacturing
advantage .' why high-performance work systems pay of. Ithaca: Cornell
University Press.

Barker, J. R. (1993). Tightening the iron cage: Concertive control in self-managing
teams. Administrative Science Quarterly, 38(3), 408.

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(6), 1 173.

Baron, R. M., & Kenny, D. A. (I 986 ). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic and statistical considerations.
Journal of Personality and Social Psychology, 51(6), 1 173-1 182.

Bateman, T. S., Grifﬁn, R. W., & Rubinstein, D. (1987). Social Information Processing
and Group-Induced Shifts in Responses to Task Design. Group & Organization
Management, 12(1), 88-108.

Batt, R.. & Valcour, P. M. (2003). Human Resources Practices as Predictors of Work-
Family Outcomes and Employee Turnover. Industrial Relations, 42(2), 189-220.

Beehr, T. A., Glaser, K. M., Canali, K. G., & Wallway, D. A. (2001). Back to basics: Re-
examination of Demand-Control Theory of occupational stress. Work & Stress,
15(2), 115-130.

Beehr, T. A., Jex, S. M., Stacy, B. A., & Murray, M. A. (2000). Impact of Context on Job
Stress Coping Work Stressors and Coworker Support as Predictors of Individual
Strain and Job Performance. Journal of Organizational Behavior, 21(4), 391-405.

Berkman, L. F., & Syme, S. L. (1979). Social networks, host resistance, and mortality: A
nine-year follow-up study of Alameda County residents. American Journal of
Epidemiology, 109(1 86-204).

Blau, G. J ., & Katerberg, R. ( 1982). Toward Enhancing Research with the Social

Information Processing Approach to Job Design. The Academy of Management
Review, 7(4), 543-550.

101

Bond, J. T., Galinsky, E., Prottas, D., & Thompson, C. (2003). Highlights of the 2002
National Study of the Changing WorkfbrceUnpublished manuscript, New York,
NY.

Broman, C. L., Hamilton, V. L., Hoffman, W. S., & Mavaddat, R. (1995). Race, gender,

and the response to stress: Autoworkers' vulnerability to long-term unemployment.
American Journal of Community Psychology, 23(6), 813-832.

Browne, M. W., & Cudeck, R. (1993). Alternative Ways of Assessing Model Fit. In K. A.
Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 320 p.).
Newbury Park: Sage Publications.

Butler, A. B., Grzywacz, J. G., Bass, B. L., & Linney, K. D. (2005). Extending the
demands-control model: A daily diary study of job characteristics, work-family
conﬂict and work-family facilitation. Journal of Occupational and
Organizational Psychology, 78, 155.

Byme, D., Clore, G. L. J., & Worchel, P. (1966). Effect of economic similarity-
dissimilarity on interpersonal attraction. Journal of Personality and Social
Psychology, 4(2), 220-224.

Byme, D. E. (1971). The attraction paradigm. New York,: Academic Press.

Caplan, R. D. (1987). Person-environment ﬁt theory and organizations: Commensurate
dimensions, time perspectives, and mechanisms. Journal of Vocational Behavior,
31(3), 248-267

Cappelli, P., & Sherer, P. D. (Eds.). (1991). The missing role of context in OB: The need
for a mesa-level approach (Vol. 13). Greenwich, CT: JAI Press.

Carayon, P., & Zijlstra, F. (1999). Relationship between job control, work pressure and
strain: studies in the USA and in The Netherlands. Work & Stress, 13(1), 32-48.

Carlson, D. S. (1999). Personality and Role Variables as Predictors of Three Forms of
Work-Family Conﬂict. Journal of Vocational Behavior, 55(2), 236-253.

Carlson, D. S., Kacmar, K. M., & Williams, L. J. (2000). Construction and Initial
Validation of a Multidimensional Measure of Work-Family Conﬂict“ l , *2.
Journal of Vocational Behavior, 56(2), 249-276.

Carlson, D. S., & Perrewe, P. L. (1999). The role of social support in the stressor-strain
relationship: An examination of work-family conﬂict. Journal of Management,

25(4), 513.

Chattopadhyay, P., Tluchowska, M., & George, E. (2004). Identifying the Ingroup: A
Closer Look at the Inﬂuence of Demographic Dissimilarity on Employee Social

102

Identity. Academy of Management. The Academy of Management Review, 29(2),
180.

Clegg, C. W. (1984). The Derivation of Job Designs. Journal of Occupational Behaviour,
5(2), 131-146.

Cobb, S. (1979). Social Support and Health through the Life Course. In M. W. Riley
(Ed.), A ging ﬁ‘om Birth to Death: Interdisciplinary Perspective. Washington, D.
C.: American Association for the Advancement of Socience.

Cohen, S., Kamarck, T., & Merrnelstein, R. (1983). A Global Measure of Perceived
Stress. Journal of Health and Social Behavior, 24 (4), 385-396.

Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering hypothesis.
Psychological Bulletin, 98(2), 310-357.

Cooper, C. L. (1986). Job distress: recent research and the emerging role of the clinical
occupational psychologist. Bulletin of the British Psychological Society, 39, 325-
331.

Cooper, C. L., Dewe, P., & O'Driscoll, M. P. (2000). Organizational stress .' a review
and critique of theory, research, and applications. Thousand Oaks, Calif.: Sage.

Cooper, C. L., Dewe, P., & O'Driscoll, M. P. (2001). Organizational stress .' a review
and critique of theory, research, and applications. Thousand Oaks, Calif: Sage.

Daniels, K., & Guppy, A. (1994). Occupational Stress, Social Support, Job Control, and
Psychological Well-Being Human Relations, 4 7(12), 1523-1544.

Davis-Blake, A., Broschak, J., & George, E. (2003). Happy Together? How Using
Nonstandard Workers Affects Exit, Voice, and Loyalty Among Standard
Employees. Academy of Management Journal, 46(4), 475-485.

de Jonge, J., & Kompier, M. A. J. (1997). A Critical Examination of the Demand-
Control-Support Model from a Work Psychological Perspective. International
Journal of Stress Management, 4(4), 235.

de Jonge, J., Reuvers, M. M. E. N., Houtman, I. L. D., Bongers, P. M., & Kompier, M. A.
(2000). Linear and Nonlinear Relations Between Psychosocial Job Characteristics,
Subjective Outcomes, and Sickness Absence: Baseline Results From SMASH.
Journal of Occupational Health Psychology, 5(2), 256-268.

de Jonge, J ., & Schaufeli, W. B. (l 998). Job characteristics and employee well-being: a

test of Warr’s Vitamin Model in health care workers using structural equation
modeling. Journal of Organizational Behavior (1986-1998), 19(4), 387.

103

de Jonge, J., van Breukelen, G. J. P., Landeweerd, J. A., & Nijhuis, F. J. N. (1999).
Comparing group and individual level assessments of job characteristics in testing

the job demand-control model: A multilevel approach. Human Relations, 52(1),
95.

de Rijk, A. E., Le Blanc, P. M., Schaufeli, W. B., & de Jonge, J. (1998). Active coping
and need for control as moderators of the job demand-control model: Effects on
burnout. Journal of Occupational and Organizational Psychology, 71, l.

Dodd, N. G., & Ganster, D. C. (1996). The interactive effects of variety, autonomy, and

feedback on attitudes and performance. Journal of Organizational Behavior,
1 7(4), 329.

Ducharme, L. J., & Martin, J. K. (2000). Unrewarding Work, Coworker Support, and Job
Satisfaction Work and Occupations, 27(2), 223-243.

Edwards, J. R., & Cooper, C. L. (1990). The Person-Environment Fit Approach to Stress:

Recurring Problems and Some Suggested Solutions. Journal of Organizational
Behavior, 11(4), 293.

Etzion, D. (1984). Moderating effect of social support on the stress-bumout relationship.
Journal of Applied Psychology, 69(4), 615-622.

Fan, X., Thompson, B., & Wang, L. (1999). Effects of sample size, estimation method,
and modelspeciﬁcation on structural equation modeling ﬁt indexes. Structural
Equation Modeling, 6, 56-83.

Fisher, C. D. (1985). Social Support and Adjustment to Work: A Longitudinal Study
Journal of Management, 11(3), 39-53.

Fisher, S. (1989). Chapter 10: Stress, Control, Worry Prescriptions and the Implications
for Health at Work: A Psychological Model. In S. L. Sauter, J. J. Hurrell & C. L.
Cooper (Eds), Job Control and Worker Health (Vol. Wiley Series on Studies in
Occupational Stress). New York, NY: John Wiley & Sons.

Fletcher, B. C., & Jones, F. (I 993). A refutation of Karasek's demand - discretion model
of occupational stress with a range of dependent measures. Journal of
Organizational Behavior, 14(4), 319.

Fumham, A. (2005). The Psychology Of Behaviour At Work: The Individual in the
Organisation Psychology Press.

Ganster, D., F usilier, M. R., & Mayes, B. T. (1986). Role of social support in the
experience of stress at work. Journal of Applied Psychology, 71(1), 102-110.

104

Ganster, D. C., Fusilier, M. R., & Mayes, B. T. (1986). Role of social support in the
experience of stress at work. Journal of Applied Psychology, 71(1), 102-1 10.

Goldberger, L., & Breznitz, S. (1993). Chapter 1: Stress Research at a Crossroads. In L.
Goldberger & S. Breznitz (Eds), Handbook of stress .' theoretical and clinical
aspects (2nd ed., pp. xxv, 819 p.). New York: Free Press.

Green, F. (2001). It's been a hard day's night: The concentration and intensiﬁcation of
work in late twentieth-century Britain. British Journal of Industrial Relations,
39(1), 53.

Green, F. (2002). Work Intensiﬁcation, Discretion and the Decline in Well-Being at Work.
Paper presented at the the Conference on Work Intensiﬁcation, Paris.

Green, F ., & McIntosh, S. (2001). The Intensiﬁcation of Work in Europe. Labor
Economics, 8(2), 291-308.

Grifﬁn, M. A. (1997). Interaction between individuals and situations: Using HLM

procedures to estimate reciprocal relationships. Journal of Management, 23(6),
759.

Grifﬁn, R. W. (1983). Objective and Social Sources of Information in Task Redesign: A
Field Experiment. Administrative Science Quarterly, 28(2), 184-200.

Grifﬁn, R. W., Bateman, T. S., Wayne, S. J., & Head, T. C. (1987). Objective and Social
Factors as Determinants of Task Perceptions and Responses: An Integrated

Perspective and Empirical Investigation. The Academy of Management Journal,
30(3), 501-523.

Gutek, B., Repetti, R. L., & Silver, D. L. (1988). Nonwork Roles and Stress at Work. In

C. L. Cooper & R. Payne (Eds.), Causes, coping, and consequences of stress at
work (pp. xii, 418 p.). Chichester ; New York: Wiley.

Hackman, J. R., & Oldham, G. R. (1980). Work redesign. Reading, Mass.: Addison-
Wesley.

Hanisch, K. A. (1999). Job Loss and Unemployment Research from 1994 to 1998: A
Review and Recommendations for Research and Intervention. Journal of
Vocational Behavior, 55(2), 188-220.

Harrison, D. A., Price, K. H., & Bell, M. P. (1998). Beyond Relational Demography:

Time and the Effects of Surface- and Deep-Level Diversity on Work Group
Cohesion. Academy of Management Journal, 41(1), 96-107.

105

Hendrix, W. H., Ovalle, N. K., & Troxler, R. G. (1985). Behavioral and physiological
consequences of stress and its antecedent factors. Journal of Applied Psychology,
70(1), 188-201.

Hodson, R. (2005). Management Behaviour as Social Capital: A Systematic Analysis of
Organizational Ethnographies. British Journal of Industrial Relations 43(1), 41-
65.

Hoel, H., Sparks, K., & Cooper, C. L. (2002). The cost ofviolence/stress at work and the
beneﬁts of a violence/stress-ﬁee working environment. (accessed 10 Feb 2003).
Geneva: Report commissioned by the International Labour Organisation (ILO).

Hofstede, G. H. (1980). Culture ’s consequences, international diﬂerences in work-related
values. Beverly Hills, Calif.: Sage Publications.

House, I. S. (l 98 l a). Work stress and social support. Reading, Mass: Addison-Wesley
Pub. Co.

House, J. S. (l981b). Work stress and social support (Vol. 4). Reading, Mass: Addison-
Wesley.

House, J. S., McMichael, A. J., Wells, J. A., Kaplan, B. N., & Landerman, L. R. (1978).
Occupational Stress, Social Support, and Health. Paper presented at the Reducing
Occuaptional Stress: Proceedings of a Conference.

House, J. S., Robbins, C., & Metzner, H. L. (1982). The association of social relationship
and activities with mortality: Prospective evidence ﬁ'om the Tecumseh
Community Health Study. American Journal of Epidemiology, 116, 123-140.

Hu, L.-T., & Bentler, P. M. (1995). Evaluating Model Fit. In R. H. Hoyle (Ed.),

Structural equation modeling: concepts, issues, and applications (pp. xxii, 289
p.). Thousand Oaks: Sage Publications.

Jackson, P. R., Wall, T. D., Martin, R., & Davids, K. (1993). New measures of job
control, cognitive demand, and production responsibility. Journal of Applied
Psychology, 78(5), 753.

Jackson, S. E. (1983). Participation in decision making as a strategy for reducing job-
related strain. Journal of Applied Psychology, 68(1), 3-19.

Jackson, S. E., Joshi, A., & Erhardt, N. L. (2003). Recent Research on Team and
Organizational Diversity: SWOT Analysis and Implications Journal of
Management, 29(6), 801 -830.

Janssen, P. P. M., Peeters, M. C. W., de Jonge, J., Houkes, I., & Tummers, G. E. R.
(2004). Speciﬁc relationships between job demands, job resources and

106

psychological outcomes and the mediating role of negative work-home
interference. Journal of Vocational Behavior, 65(3), 411.

Jehn, K. A., Northcraﬁ, G. B., & Neale, M. A. (1999). Why Differences Make a
Difference: A Field Study of Diversity, Conﬂict, and Performance in Workgroups.
Administrative Science Quarterly, 44(4), 741-763.

Jenkins, G. D., Nadler, D. A., Lawler, E. E., & Cammann, C. (1975). Standardized
observations: An approach to measuring the nature of jobs. Journal of Applied
Psychology, 60(2), 171-181.

Jex, S. M. (1998). Stress and job performance .' theory, research, and implications for
managerial practice Thousand Oaks, CA: Sage Publications.

Johnson, J. V., & Hall, E. M. (1988). Job strain, work place social support, and
cardiovascular disease: a cross-sectional study of a random sample of the Swedish
working population. American Journal of Public Health, 78(10), 1336-1342.

Jones, B., Flynn, D. M., & Kelloway, E. K. (1995). Perception of support from the
organization in relation to work stress, satisfaction, and commitment. In S. L.
Sauter & L. R. Murphy (Eds.), Organizational risk factors for job stress (pp. xii,
400 p.). Washington, DC: American Psychological Association.

Kahn, R. L. (1964). Organizational stress: studies in role conﬂict and ambiguity. New
York,: Wiley.

Kahn, R. L., & Antonucci, T. (1980). Convoys over the Life Course: Attachment, Roles,
and Social Support. In B. B. Baltes & O. Brim (Eds), Life-Span Development and
Behavior (Vol. 3). Boston: Lexington Press.

Kalleberg, A. L., Reskin, B. F., & Hudson, K. (2000). Bad job in America: Standard and
nonstandard employment relations and job quality in the United States. American
Sociological Review, 65(April).

Karasek, R., & Theorell, T. (1990). Healthy work: stress, productivity, and the
reconstruction of working life. New York: Basic Books.

Karasek, R. A. (1979). Job Demands, Job Decision Latitude, and Mental Strain:
Implications for Job Redesign. Administrative Science Quarterly, 24(2), 285.

Karasek, R. A. (1989). Control in the workplace and its health-related aspect. In S. L.

Sauter, J. J. Hurrell & C. L. Cooper (Eds), Job control and worker health (pp. xx,
31 1 p.). Chichester ; New York: Wiley.

107

Karasek, R. A., Triantis, K. P., & Chaudhry, S. S. (1982). Coworker and Supervisor
Support as Moderators of Associations between Task Characteristics and Mental
Strain. Journal of Occupational Behaviour, 3 (2), 181-200.

Kasl, S. V. (1983). Chapter 4: Pursuing the link between stressful life experiences and
disease: A time for reappraisal. In C. L. Cooper (Ed.), Stress Research: Issues for
the Eighties (pp. 79-102). New York, NY: John Wiley & Sons.

Kaufmann, G. M., & Beehr, T. A. (1986). Interactions Between Job Stressors and Social
Support: Some Counterintuitive Results. Journal of Applied Psychology, 71(3),
522-526.

Kirkman, B. L., & Rosen, B. (1999). Beyond self-management: Antecedents and
consequences of team empowerment. Academy of Management Journal, 42(1), 58.

Ku, C.-H. (2006). The eﬂects of organizational factors and scheduling on fatigue, health
and social well-being of locomotive crews. Unpublished Ph.D., The University of
Wisconsin - Madison, United States -- Wisconsin.

Landsbergis, P. A. (1988). Occupational Stress Among Health Care Workers: A Test of
the Job Demands-Control Model. Journal of Organizational Behavior, 9(3), 217.

Landsbergis, P. A., Cahill, J., & Schnall, P. (1999). The impact of lean production and
related new systems of work organization on worker health. Journal of
Occupational Health Psychology, 4(2), 108-130

LaRocco, J. M., House, I. S., & French, J. R. P. J. (1980). Social Support, Occupational
Stress, and Health. Journal of Health and Social Behavior, 21(3), 202-218.

Lazarus, R. S. (1990). Theory-Based Stress Measurement. Psychological Inquiry, 1(1), 3-
13.

Liden, R. C., Wayne, S. J ., & Bradway, L. K. (1997). Task Interdependence as a
Moderator of the Relation Between Group Control and Performance (Vol. 50, pp.
169-181).

MacKinnon, D. P. (1994). Analysis of mediating variables in prevention and intervention
research. In A. Cazares & L. A. Beatty (Eds.), Scientific methods in prevention
research (Vol. 139, pp. 127-153): US Govt. Print. Ofﬁce.

MacKinnon, D. P. (2000). Technical Assistance Report: Mediation Analysis: Research in
Prevention Laboratory.

MacKinnon, D. P., & Dwyer, J. H. (1993). Estimating Mediated Effects in Prevention
Studies. Evaluation Research, 1 7(2), 144-158.

108

Michelle, P. S., & Gary, R. B. (2001). An exploratory analysis of racial factors in staff
burnout among assertive community treatment workers. Community Mental
Health Journal, 37(5), 393-404.

Milliken, F. J ., & Martins, L. L. (1996). Searching for Common Threads: Understanding
the Multiple Effects of Diversity in Organizational Groups. The Academy of
Management Review, 21(2), 402-433.

Milliken, F. J ., Martins, L. L., & Morgan, H. (1998). Explaining organizational
responsiveness to work-family issues: The role of human resource excutives as
issue interpreters. Academy of Management Journal, 41(5), 580.

Morgeson, F. P., & Campion, M. A. (2003). Work Design. In I. B. Weiner, D. K.
F reedheim, J. A. Schinka, W. F. Velicer & D. R. Ilgen (Eds.), Handbook of
psychology. New York: Wiley.

NIOSH. (1999). Stress. Cincinnati, OH: National Institute for Occupational Safety and
Health.

NIOSH. (2002). The changing organization of work and the safety and health of working
people. Cincinnati, OH: National Institute for Occupational Safety and Health.

Noblet, A. J ., McWilliams, J., Tea, S. T. T., & Rodwell, J. J. (2006). Work characteristics
and employee outcomes in local government. The International Journal of
Human Resource Management, 1 7(10), 1804-.

O'Reilly, C. A., & Caldwell, D. F. (1979). Informational inﬂuence as a determinant of
perceived task characteristics and job satisfaction. Journal of Applied Psychology,
64(2), 157-165.

O'Reilly, C. A., Caldwell, D. F ., & Barnett, W. P. (1989). Work Group Demography,
Social Integration, and Turnover. Administrative Science Quarterly, 34(1), 21-37.

O'Reilly, C. A., Parlette, G. N., & Bloom, J. R. (1980). Perceptual measures of task
characteristics: The biasing effects of differing frames of reference and job
attitudes. Academy of Management Journal (pm-1986), 23(000001), 1 18.

Oldham, G. R., Kulik, C. T., Stepina, L. P., & Ambrose, M. L. (1986). Relations between
Situational Factors and the Comparative Referents Used by Employees. The
Academy of Management Journal, 29(3), 599-608.

Parasuraman, S., Greenhaus, J. H., & Granrose, C. S. (1992). Role Stressors, Social
Support, and Well-Being Among Two-Career Couples. Journal of Organizational
Behavior, 13(4), 339-356.

109

Parker, S. K., Wall, T. D., & Cordery, J. L. (2001). Future work design research and
practice: Towards an elaborated model of work design Journal of Occupational
and Organizational Psychology, 74(4), 413-440.

Pearce, J. L. (1993). Toward an Organizational Behavior of Contract Laborers: Their
Psychological Involvement and Effects on Employee Co-Workers. Academy of
Management Journal, 36(5), 1082-1096.

Pfeffer, J. (1983). Organizational Demography. In L. L. Cummings & B. M. Staw (Eds.),
Research in Organizational Behavior (Vol. 5, pp. 299-357): JAI Press Inc.

Pfeffer, J ., & Moore, W. L. (1980). Average Tenure of Academic Department Heads: The
Effects of Paradigm, Size, and Departmental Demography. Administrative
Science Quarterly, 25(3), 387-406.

Piccolo, R. F., & Colquitt, J. A. (2006). Transformational Leadership and Job Behaviors:
The Mediating Role of Core Job Characteristics. Academy of Management
Journal, 49(2), 327-340.

Raykov, T., & Marcoulides, G. A. (2006). A ﬁrst course in structural equation modeling
(2nd ed.). Mahwah, NJ: Lawrence Erlbaum Associates, Publishers.

Rico, R., Molleman, E., Sanchez-Manzanares, M., & Van der Vegt, G. S. (2007). The
effects of diversity faultlines and team task autonomy on decision quality and
social integration Journal of Management, 33(1), 111-132.

Rogers, J. K. (2000). Temps : the many faces of the changing workplace. Ithaca, N.Y.:
Cornell University Press.

Rousseau, D. M. (1978). Characteristics of Departments, Positions, and Individuals:
Contexts for Attitudes and Behavior. Administrative Science Quarterly, 23(4),
521-540.

Salancik, G. R., & Pfeffer, J. (1978). A Social Information Processing Approach to Job
Attitudes and Task Design. Administrative Science Quarterly, 23(2), 224-253.

Schaubroeck, J., Cotton, J. L., & Jennings, K. R. (1989). Antecedents and Consequences
of Role Stress: A Covariance Structure Analysis. Journal of Organizational

Behavior, 10(1), 35-58.

Schnake, M. E., & Dumler, M. P. (1987). The Social Information Processing Model of
Task Design: Conflicting Cues and Individual Differences (Vol. 12, pp. 221-240).

Searle, B., Bright, J. E. H., & Bochner, S. (2001). Helping people to sort it out: The role
of social support in the Job Strain Model Work & Stress, 15(4), 328 - 346.

110

Seers, A., McGee, G. W., Serey, T. T., & Graen, G. B. (1983). The Interaction of Job
Stress and Social Support: A Strong Inference Investigation. Academy of
Management Journal, 26(2), June 1983.

Smith, C. S., Tisak, J., Hahn, S. E., & Schmieder, R. A. (1997). The measurement ofjob
control. Journal of Organizational Behavior, 18(3), 225.

Smith, V. (2001). Teamwork vs. Tempwork: Managers and the dualizm of workplace
restructuring. In D. B. Cornﬁeld, K. E. Campbell & H. J. McCammon (Eds.),
Working in restructured workplaces .' challenges and new directions for the
sociology of work (pp. xxii, 484 p.). Thousand Oaks, Ca1if.: Sage Publications.

Tansey, T. N., Mizelle, N., Ferrin, J. M., Tschopp, M. K., & F rain, M. (2004). Work-
Related Stress and the Demand-ControI-Support Framework: Implications for the
P x E Fit Model. Journal of Rehabilitation, 70(3), 34.

Theorell, T., & Karasek, R. (1996). Current Issues Relating to Psychosocial Job Strain

and Cardiovascular Disease Research. Journal of Occupational Health
Psychology, 1(1), 9-26.

Thomas, J ., & Grifﬁn, R. (1983). The Social Information Processing Model of Task
Design: A Review of the Literature. The Academy of Management Review, 8(4),
672-682.

Thomas, J. G. (1986). Sources of Social Information: A Longitudinal Analysis Human
Relations, 39(9), 855-870.

Thomas, L. T., & Ganster, D. C. (1995). Impact of family-supportive work variables on
work-family conﬂict and strain: A control perspective. Journal of Applied
Psychology, 80(1), 6.

Tsui, A. S., Egan, T. D., & O'Reilly, C. A. I. (1992). Being Different: Relational
Demography and Organizational Attachment. Administrative Science Quarterly.
37(4), 549-579.

Tsui, A. S., & O'Reilly, C. A. I. (1989). Beyond Simple Demographic Effects: The
Importance of Relational Demography in Superior-Subordinate Dyads. The
Academy of Management Journal, 32(2), 402-423.

van der Doef, M., & Maes, S. (1999). The Job Demand-Control (-Support) Model and
psychological well-being: a review of 20 years of empirical research. Work &

Stress, 13(2), 87 - 114

Van Harrison, R. (1978). Person-environment ﬁt and job stress. In C. L. Cooper & R.
Payne (Eds.), Stress at work (pp. xii, 293 p.). Chichester ; New York: Wiley.

111

van Knippenberg. D., & Schippersand, M. C. (2007). Work Group Diversity. Annual
Review of Psychology, 58, 515-541.

van Vegchel, N., de Jonge, J ., & Landsbergis, P. A. (2005). Occupational stress in
(inter)action: the interplay between job demands and job resources. Journal of
Organizational Behavior, 26(5), 535.

Voydanoff, P. (2004). The Effects of Work Demands and Resources on Work-to-Family
Conﬂict and Facilitation. Journal of Marriage and Family, 66(2), 398.

Wallace, J. E. (2005). Job Stress, Depression and Work-to-Family Conﬂict: A Test of the
Strain and Buffer Hypotheses. Relations Industrielles, 60(3), 510.

Way, S. A. (2002). High Performance Work Systems and Intermediate Indicators of Firm
Performance Within the US Small Business Sector. Journal Of Management,
28(6), 765-785.

Wells, 1. A. (1982). Objective Job Conditions, Social Support and Perceived Stress
Among Blue Collar Workers. Journal of Occupational Behaviour, 3(1), 79-94.

White, S. E., & Mitchell, T. R. (1979). Job enrichment versus social cues: A comparison
and competitive test. Journal of Applied Psychology, 64(1), 1-9.

Williams, K. Y., & O'Reilly III, C. A. (1998). Demography and Diversity in
Organizations: A review of 40 years of research. In L. L. Cummings & B. M.
Staw (Eds.), Research in Organizational Behavior (Vol. 20, pp. 77-140): JAI
Press Inc.

Williams, M. (2001). In whom we trust: Group membership as an effective context for
trust development. Academy of Management. The Academy of Management
Review, 26(3), 377.

Xie, J. L. (1996). Karasek's model in the People's Republic of China: Effects of job
demands, control, and individual differences. Academy of Management Journal,
39(6), 1594.

Zacharatos, A., Barling, J ., & Iverson, R. D. (2005). High-Performance Work Systems
and Occupational Safety. Journal of Applied Psychology, 90(1), 77-93.

Zalesny, M. D., & Farace, R. V. (1986). A Field Study of Social Information Processing
Mean Differences and Variance Differences (Vol. 13, pp. 268-290).

Zenger, T. R., & Lawrence, B. S. (1989). Organizational Demography: The Differential
Effects of Age and Tenure Distributions on Technical Communication. Academy
of Management Journal, 32(2), 353-376.

112

Zeytinoglu, 1. U., Denton, M., Davies, S., Baumann, A., & Boos, L. (2007). Associations
between Work Intensiﬁcation, Stress and Job Satisfaction: The Case of Nurses in
Ontario. Relations Industrielles, 62(2), 201.

113

    

  

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

11111111111111111111111111111111111111

3 1 93 02956 5763

 

1