:81 .33
. 2%..”5;
Q

r...

 

«swam,»

.
f. V. ..... ..
h? . T. . ., ..
onwflf.m«.w. A

£35....“
: 14:?)

31 ..
2:31.
I.

3.4..)er
.r. .

:
A..:-.:5......
.. pittf
o1; hunts}:

3-5 :93. 3.1:!
..I J91, 59... U7. 1:...
A 1783!». .21:

 

a 09.:
nnnfefléa.»

Llaanni

r

2:51 at

, » .rhfihuht. x ...

‘txsfi!
31 a}:
.03).:
1.151. .

gfifiifiwfig

 

fl... 1:93.
n. P:
I

 

 

llllllllllllllllllllllllllllllllllllllllllll’lllllllllllll

LIBRARY 3 1293 01787 976
Michigan State
Universlty

 

 

 

This is to certify that the

dissertation entitled

Interrelationships Among
Adult Attachment Style, Work Stress,
Social Support, and Indexes of Strain

presented by

Lisa L. Schirmer

has been accepted towards fulfillment
of the requirements for

Ph.D. Counseling Psychology

 

 

degree in

flame, 17%“?

Major professor

 

MSUiJ an Affirmative Action/Equal Opportunity Institution 0-12771

PLACE IN RETURN Box to remove this checkout from your record.

To AVOID FINES return on
MAY BE RECALLED with earlier

or before date due.
due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11” W14

INTERRELATIONSHIPS AMONG
ADULT ATTACHMENT STYLE, WORK STRESS,
SOCIAL SUPPORT, AND INDEXES OF STRAIN
By

Lisa L. Schirmer

A DISSERTATION

Submitted to
Michigan State University
in partial fiilfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Counseling, Educational Psychology, and Special Education

1 999

ABSTRACT

THE INTERRELATIONSHIPS AMONG ADULT ATTACHMENT STYLE, WORK
STRESS, SOCIAL SUPPORT, AND INDEXES OF STRAIN

By

Lisa L. Schirmer

The present study explored the contributions of work stress, social support, and
adult attachment styles to job satisfaction and symptomatic distress within an adult
worker sample. Participants were 117 Michigan Sate University employees (61% staff;
39% faculty) who completed both categorical and continuous measures of adult
attachment style along with self-report measures of the other variables under study. Due
to their low frequencies in this sample, participants with preoccupied and fearful
attachment styles were combined to form a single “anxious” attachment style group. This
group was compared to workers evidencing secure and dismissive attachment styles.
Results indicated that secure workers reported significantly less work stress and
symptomatic distress than did anxiously attached workers. Relative to their anxious
peers, secure workers also reported significantly higher levels of supervisor support. No
' main effects for attachment style on job satisfaction were observed. Dimensional
measures of adult attachment orientation significantly enhanced the prediction of
symptoms (but not job satisfaction) even after work stress and supervisor support levels
were controlled. Implications of the findings to prior inconsistencies in the literature
regarding relationships among work stress, social support, and indexes of work strain are

discussed.

Copyright by
LISA LOUISE SCHIRMER
1999

I dedicate this dissertation to my parents,
Lucille H. Schirmer and Clarence R. Schirmer (1918-1992).

iv

ACKNOWLEDGMENTS

I could not have completed this dissertation without a great deal of both
instrumental and emotional support. Academically, the primary source of this support
was my advisor and dissertation committee chairperson, Dr. Frederick Lopez, who
continually provided me with his expert knowledge and encouragement. I would also
like to thank my dissertation committee members: Drs. Kevin Ford, Linda Forrest, and
Don Hamachek for their excellent feedback and advise.

My gratitude goes out to my many friends who provided me with encouragement
and emotional support during the nine long years of my doctoral program. In addition, I
thank the following people for their technical/instrumental support: Bn’an Guerin, for
statistical consultation; Leslie Agnes-Pons Guerin, for data entry, survey collating, and
tea; Ted \Vrlliamson, for computer support; and Anne Williamson, for survey collating.
For editorial assistance I relied upon Patti Jones, Pam Comstock, and Mary S. Powell.

I was also fortunate have three colleagues in my program, Jennie Leskela, Ellen
Narusis Behrens, and Mary Gilbert who provided me support through the tempests of
graduate school and beyond. Storms such as those experienced in graduate school are
diflicult to manage without the security of a safe harbor. I am blessed to have such a
haven in my family. My Father, Clarence Schirmer, always told me that I could achieve
whatever I set my mind to. My mother, Lucille Schirmer; my brother, Richard Schirmer;
and my sister-in-law, Florence Walraven never gave up on me and were always there
when I needed them. Finally, my love and gratitude goes to Michael S. Powell for his

patience, understanding, and support throughout our life together.

TABLE OF CONTENTS

LIST OF TABLES ...........................................................................................
LIST OF FIGURES .........................................................................................

CHAPTER I

INTRODUCTION ...........................................................................................
Attachment Style and the Development of Social Support 1., ......................
Attachment Theory and Vocational Behavior ............................................
Summary and Problem Statement .............................................................

CHAPTER 11
REVIEW OF LITERATURE ...........................................................................
Occupational Stress ..................................................................................
“(Social Support and the Work Stress-Strain Relationship ...........
‘/ Support for the Buffering Hypothesis .......................................
The Absence of Buffering .........................................................
Instances of “Reverse Buffering” and “Mixed Buffering” .........
Summary ..................................................................................
Attachment Theory as a Framework for Conceptualizing the
Interrelationships Among Work Stress, Social Support and
Worker Strain ....................................................................................
Attachment Theory: Key Concepts and Assumptions ...............
Contemporary Attachment Theory: Theoretical and
Empirical Foundations ........................................................
The Classification and Measurement of Adult
Attachment Styles ...................................................
Adult Attachment Related Distinctions in Affect
Regulation and Interpersonal Behavior ....................
Adult Attachment and Work-Related Behavior ..............
Summary and Problem Restatement .........................................
Definitions ...........................................................................................
" Stressors ...................................................................................
Strains ......................................................................................
. Social Support ..........................................................................
AduIt'AttaCIiment Styles ...........................................................
Hypotheses ..........................................................................................
Relations Among Work Stress, Social Support, and Strain ........
Relationships of Attachment Style Differences to Work
Stress, Supervisor Support, and Indexes of Strain ...............

CHAPTER III
METHODOLOGY ..........................................................................................
Participants .............................................................................................

Procedures ..............................................................................................

ix
xi

\IOUIv—t

10
13
15
16
17
21

21
23

26

26

3O
33
35
37
37
37
37
38
39
39

4O

42

42
4S

Instruments ............................................................................................. 46

Demographic and Background Information Form ..................... 47
Attachment Style ...................................................................... 47
Relationship Questionnaire ........................................... 47
Experiences in Close Relationships ............................... 48
Social Support .......................................................................... 49
The Work Stress Inventory ....................................................... 51
The Job Satisfaction Scale ........................................................ 52
The Hopkins Symptom Checklist ............................................. 52
Research Hypotheses .............................................................................. 53
Data Analysis .......................................................................................... 54
CHAPTER IV
RESULTS ....................................................................................................... 56
Treatment of Missing Data ...................................................................... 56
Descriptive Statistics ............................................................................... 57
Recomposition of RQ Classification ....................................................... 58
Correlational Findings ............................................................................ 6O
Supervisor Support as a Moderator of the Work Stress-Strain
Relationship ...................................................................................... 63
Relationship of Attachment to Work Stress, Supervisor Support
and Indexes of Strain ......................................................................... 64
Contributions of Work Stress, Social Support, and Attachment
Orientations to indexes of Work-Related Strain ................................. 69
Post Hoc Analyses .................................................................................. 69
Summary of Findings .............................................................................. 75
CHAPTER V
DISCUSSION ................................................................................................. 82
The Relationships of Work Stress, Supervisor Support, and Strains ......... 82
Findings Regarding the Buffering Effect of Social Support ..................... 83
The Relationship of Adult Attachment to Work Stress, Job
Satisfaction Symptomatic Distress and Social Support ...................... 86
Limitations ............................................................................................. 9O
Implications ............................................................................................ 92
Implications for Theory ............................................................ 92
Implications for Practice ........................................................... 95
Recommendations for Future Research ..................................... 96
Conclusions ............................................................................................ 98
REFERENCES ................................................................................................ 100

vii

APPENDICES ................................................................................................ 108

Appendix A. Prenotification Postcard ................................................... 109
Appendix B. Request for Participation Cover Letter ............................. 110
Appendix C. Participation Consent Form .............................................. 112
Appendix D. Demographic and Background Information ...................... 113
Appendix E. RQ ................................................................................... 114
Appendix F. ECR ................................................................................. 115
Appendix G. Social Support Scale ........................................................ 118
Appendix H. WSI ................................................................................. 119
Appendix I. J SS .................................................................................. 125
Appendix J HSCL .............................................................................. 126
Appendix K Thank You/Reminder Postcard ........................................ 129
Appendix L Cover Letter to Fourth Mailing ........................................ 130

viii

Table 1.
Table 2.

Table 3.

Table 4.
Table 5.
Table 6.
Table 7.

Table 8.

Table 9.

Table 10.

Table 11.

Table 12.

Table 13.

LIST OF TABLES

Hazan and Shaver Attachment Style Inventory ...............................

Sample Demographic Information ..................................................

Means and Standard Deviations of ECR Scores for

RQ Categories ................................................................................
Descriptive Statistics for Continuously Scored Measures ...............
Frequencies and Percentages of Attachment Styles ........................
Frequencies and Percentages of Recoded Attachment Styles ..........

Intercorrelation of Variables ...........................................................

Hierarchical Regression Analyses for Work Stress (W S),
Supervisor Support (SS), and their Interaction in

Predicting Job Satisfaction and Symptomatic Distress ....................

Means and SD of Attachment Groups on Work Stress,

Symptomatic Distress, Job Satisfaction, and Supervisor Support

Hierarchical Regression Analyses for Work Stress (W S),
Supervisor Support (SS), Attachment Dimensions and their
Interaction in Predicting Symptomatic Distress and Job

Satisfaction ....................................................................................

Hierarchical Regression Analyses for Supervisor Support
and Attachment Dimensions and their Interaction in

Predicting Job Satisfaction .............................................................

Hierarchical Regression Analyses for Organizational Stress
Intensity (OSI), Supervisor Support (SS) and their
Interaction in Predicting Job Satisfaction and

Symptomatic Distress .....................................................................

Hierarchical Regression Analyses for OSI, Supervisor
Support (SS), Attachment Dimensions and their Interaction

in Predicting Job Satisfaction and Symptomatic Distress ................

ix

27

43

50

59

61

61

62

65

68

70

72

76

77

Table 14.

Table 15.

Hierarchical Regression Analyses for Organizational Stress

Frequency (OSF), Supervisor Support (SS), Attachment

Dimensions and their Interaction in Predicting Job

Satisfaction .................................................................................... 78

Hierarchical Regression Analyses for Organizational Stress

Frequency (OSF), Supervisor Support (SS), Attachment

Dimensions and their Interaction in Predicting Symptomatic

Distress .......................................................................................... 79

LIST OF FIGURES

Figure 1. The Interaction of Avoidance and Supervisor
Support in Predicting Job Satisfaction ............................................ 73

Figure 2. The Interaction of Anxiety and Supervisor
Support in Predicting Job Satisfaction ............................................ 74

Figure 3. The Interaction of Supervisor Support and
Organizational Stress Intensity in Predicting
Job Satisfaction .............................................................................. 80

CHAPTER I

INTRODUCTION

Occupational stress is a major problem affecting the health and well being of
millions of Americans each year. According to the 1985 National Health Interview
Survey, approximately 11 million workers reported that they were experiencing health-
endangering levels of "mental stress" at work (Shilling & Brackbill, 1987). The
psychological consequences of work related stress are costly for the employer as well as
the individual. In 1985 the insurance industry reported that "claims for gradual mental
stress alone account for about 11% of all claims for occupational disease" (1985 report of
the National Council on Compensation Insurance cited in Sauter, Murphy, & Hurrell,
1990). Furthermore, in a recent symposium, Joseph Hurrell, Associate Director of the
National Institute for Occupational Safety and Health (NIOSH) reported that claims for
job related stress disorders are increasing and are considered the most disabling
occupational illness in terms of lost time (Hurrell, 1998). The exact cost of losses in
productivity and employment, as well as in total expenditures for medical services,
resulting from occupational stress has been difficult to calculate. However, Sauter et al.
(1990) have projected that the expense of psychological disorders related to work stress
costs in the tens of billions of dollars annually in the United States alone. Clearly,
gaining a greater understanding of these disorders and their treatment would benefit both

individual workers and industry.

The term 'stress' is often used to describe both the precipitant and the result of
experiencing difficulty at work. For the purpose of this study, Fenlason and Beehr's
(1994) distinction of stressors and strains will be adopted. Stressors are those events or
demands that contribute to the experience of work related stress. Strains are the physical
or emotional manifestations of work related stress. Stress will be reserved as a term to
describe our general field of interest.

Reducing the effects of work related stressors has concerned psychologists,
employers and the general public for at least two decades. The Michigan Person-
Environment Fit (P-E fit) model is one of the most dominant etiological models of how
occupational stress develops (Gore, 1987). This theory posits that a worker’s experience
of stressors and strain are a result of an imbalance between that individual’s
needs/characteristics and the environmental provisions on the job (Caplan, Cobb, French,
Harrison, & Pinneau, 1975). In this model, the relationship between stressors and strain
is moderated by individual difference variables such as defenses, coping predispositions,
genetic factors, social background characteristics, and unmet needs. Therefore, according
to the P-E fit model, not all workers will experience identical stressors in the same way.

Researchers have investigated numerous variables that may moderate the stressor-
strain relationship. One popular notion is that by increasing a worker's social support
system he/she may avoid many of the negative consequences of working with stressors.
However, the exact method by which social support alleviates work related distress has
been greatly debated among psychologists and researchers.

Researchers have proposed that social support reduces work related distress in

three different ways. Fenlason and Beehr (1994) suggested the following summary of

this body of literature: social support reduces distress by (a) acting directly on strains
(main effect), (b) acting directly on the stressors (main effect), and (c) “interacting with
stressors so that the relation between stressors and strain is stronger for persons with low
levels of support than for those with high levels of support” (p. 158). The idea that social
support can directly reduce strains has been demonstrated consistently in the empirical
literature on job stress (Blau, 1981; LaRocco & Jones, 1978; Leiter, 1991). However,
mixed results have been found regarding the main effect of social support on stressors.
Beehr (1985) speculated that the effect of social support on stressors may be more
pronounced if the support is provided by supervisors rather than by peers or family
members. This hypothesis has been partially supported in subsequent research (F enlason
& Beehr, 1994).

Research investigating the 'buifering effect' of social support on the relationship
between stressors and strains has produced inconsistent findings. Whereas some authors
(Gore, 1978; LaRocco, House, & French, 1980; Kobasa & Puccetti, 1983; Haines,
Hurlbert, & Zimmer, 1991) have found support for the hypothesis that social support
moderates the relationship between stress and strains, other researchers have found no
support for the buffering hypothesis (Blau, 1981; Ganster, Fusiler, & Mayes, 1986;
LaRocco & Jones, 1978). Other studies have found support for the buffering hypothesis,
but not in the expected direction. In these studies the presence of social support appears
to increase rather than decrease the effects of stressors on strain (Beehr, 1976; Kaufinann
& Beehr, 1986). This result has been labeled ‘reverse buffering’. Still other authors have
reported both buffering and reverse buffering results (Chisholm, Kasl, & Mueller, 1986;

F enlason and Beehr, 1994).

These inconsistent findings have been the subject of a great deal of speculation.
One hypothesis, first proposed by Beehr, King, and King (1990), and later by Fenlason
and Beehr (1994), is that the content of the supportive communication influences the
effect of social support on the relationship between stressors and strains. In a study of
173 professional secretaries, F enlason and Beehr (1994) found that positive work related
conversation was most helpful in reducing worker strains and that non-work related
conversation is second best at producing this effect. However, they stated that negative
work-related conversation does not appear to help workers reduce their experience of
strains.

Another factor that may contribute to discrepant findings in the stress buffering
literature is that researchers have not been consistent in their use of outcome measures.
After reviewing this body of work, Beehr (1985) asserted, “it appears that the strains that
are more health oriented (psychological or physiological health) and less attitude oriented
(e. g., job satisfaction) are more likely to be affected by social support interacting with
stressors” (p.387). While this theory has not been tested in a systematic manner, in
accordance with the P-E Fit Model, it seems reasonable to hypothesize that individuals
may be affected differently by occupational stressors and that social support may be
helpful in relieving some indexes of strain but not others. .

Finally, inadequacies in the conceptualization and measurement of social support
may have contributed to inconsistent findings in the stress buffering literature. In a
review of this literature, Coyne and DeLongis (1986) stated that the concept of social
support becomes “systematically misleading when it is accepted in place of a more

elaborated understanding of the complexities of people’s involvement with others”

(p. 458). They asserted that it is important to research how an individual’s personal
characteristics affect the way that he/she finds, develops, nurtures, and terminates
relationships. Sarason, Sarason, and Shearin (1986) contended that much social support
research has been based on the erroneous assumption that social support is an
environmental provision. They argued that it is important to consider the contribution that
peOple make to their own social support levels and to include personality variables in the
design of research on social support. Three studies conducted by these authors revealed
evidence that social support may indeed possess trait-like characteristics. They found that
self-reports of social support availability and satisfaction are stable for up to 3 years and
appear consistent with retrospective indexes of early patterns of social contact.
Individual attachment style is one personality variable that may provide an explanation
for this consistency in social support patterns over time.
Attachment Style and the Development of Social Support

The concept of “attachment styles” has been derived from Attachment theory
(Bowlby 1969, 1982; Ainsworth, 1978, 1982). Attachment theory posits that the
outcome of children’s early attempts at proximity seeking contributes to their subsequent
expectations of their own competence and lovability, and of the accessibility and
responsiveness of significant others in their social environment. These cognitive
expectancies are referred to as the person’s “internal working model” of self and others.
Lopez (1995) stated that Bowlby “considered them ‘working’ models because they (a)
organized internal appraisals and interpersonal behaviors along pathways that were
adaptive in the persons earlier development, and (b) thereby shaped the person’s later

social experience in schema-consistent ways.” (p. 399). These working models develop

into “attachment styles”, or patterns of relating to self and others, which are especially
activated when the individual is under stress. Contemporary attachment theory has been
extended to provide a framework for understanding adult adjustment. Current research
supports the existence of four distinct adult attachment styles (Bartholomew, 1990;
Bartholomew & Horowitz, 1991).

Recent research has revealed that an individual’s adult attachment style influences
his/her affect regulation, perception of social support, development of network
orientation (beliefs, attitudes, and expectations regarding the utility of social support),
and satisfaction with support that is provided (Kobak & Sceery, 1988; Priel & Shamai,
1995; Wallace & Vaux, 1993). This body of literature suggests that by understanding an
individual’s adult attachment style we may be better able to predict the usefulness of
social support in buffering the stressor-strain relationship. Furthermore, there has been
some evidence that the construct of adult attachment styles may be helpful in
understanding an individual’s perceptions and behaviors in occupational settings.
Attachment Theog and Vocational Behavior

Although most research on adult attachment styles has focused on romantic
relationships, several authors have begun to explore the role of adult attachment style in
vocational behavior (Blustein, Schultheiss, & Prezioso, 1995; Hardy & Barkham, 1994;
Hazan & Shaver, 1990). Hazan and Shaver (1990) provided theoretical and empirical
support for attachment style differences in how individuals approach their work. They
found that secure workers reported greater work adjustment than those individuals’s
endorsing insecure attachment styles. In addition, Hardy and Barkham (1994) studied a

clinical population who where reporting occupational stress and depression. They found

attachment style differences in worker’s anxiety regarding their job performance, work
relationships, and job satisfaction.

Although these findings lend support to the hypothesis that an individual’s adult
attachment style influences work-related attitudes and behaviors, there are several
important weaknesses in this literature. First, sampling techniques that limit the
generalizability of findings have been frequently used. Second, the methods of
measurement used in these studies were also problematic. Hazan and Shaver’s study
used exploratory and unstandardized measures of work behavior. They also relied solely
on a single item forced choice measure to indicate attachment style. Hardy and Barkham
relied upon a new measure developed for their study to assess attachment style.

Finally, those studies which directly attempted to measure the constructs
underlying attachment did so using a three-category model of attachment (secure,
anxious/ambivalent, and avoidant) based on the work of Mary Ainsworth (Ainsworth, et
al., 1978). Recently, Bartholomew and Horowitz (1991) have developed a newer model
to describe attachment style in adults which identifies four different attachment
categories: secure, preoccupied, dismissive, and fearful.

Summagy and Problem Statement

The literature examining the influence of social support on the relationship
between work stressors and strains has produced conflicting results. It has been
hypothesized that this inconsistency may be due to personality differences that may
moderate the effect of social support on the stressor-strain relationship. Adult
attachment style is an individual difference variable that may both directly affect work-

related strains and also moderate stressor-strain relationships. It has been linked to

network orientation which describes the beliefs, attitudes, and expectations that people
hold regarding the usefulness of social support in providing assistance to deal with life
problems. It has also been demonstrated that an individual’s attachment style is related to
his/her perception and satisfaction with social support, as well as his/her ability to
regulate affect. Therefore, an individual’s adult attachment style may predict how the
provision of social support may influence the stressor-strain relationship. To enhance the
current research base, future studies need to incorporate the following: 1) the recently
developed four-category attachment typology; 2) well established measures of all
constructs being examined; 3) measures of both psychological health and work appraisal
oriented strains (e.g., job satisfaction); and 4) a sampling strategy which is not limited to
one class or type of occupation.

The purpose of this study is to examine the contributions of work-related
stressors, social support, and adult attachment styles to job satisfaction and psychological
functioning within an adult worker sample. Special consideration will be given to
whether stressor-strain and social support-strain relationships are moderated by an adult
worker’s attachment style. In addition, the direct contribution of attachment style in
reducing strains will be explored.

Should support be found for the moderating effect of attachment style on the
relationship between work stress and social support, this study would assist in resolving a
long-standing controversy in the literature on stress management and social support.
Furthermore, this evidence would contribute to the growing body of literature indicating
that workers’ attachment styles may influence their work-related attitudes and behaviors.

Support for this hypothesis would also underscore the desirability of considering adult

attachment styles as a component in the design of stress management plans for workers.
For example, although some workers may benefit from group oriented stress management
programs which encourage supportive exchanges among members, others may receive
greater assistance from completely didactic sessions. In fact, individuals who may
develop greater strains when social support is delivered may benefit from information
regarding how to say ‘no’ when such support is proffered.

Given the great deal of time that adults devote to work and the detrimental
consequences of experiencing stress in work settings, it is important that we be able to
manage work stress in the most effective manner possible. Taking significant individual
differences into account when designing stress management plans may be profitable for

both workers and industry.

CHAPTER 11

REVIEW OF THE LITERATURE

This chapter reviews three lines of research relevant to the proposed study: 1)
studies which examine interrelationships among work stressors, social support and
strains; 2) studies which investigate how variation in adult attachment style is related to
individual differences in the perception and expectations of others, as well as in
relationship behavior; and 3) studies which propose that attachment theory can be useful
in understanding work-related attitudes and behaviors.

Occupational Stress

Many negative consequences of work stress have been discussed in the medical,
psychological and risk management literature. House, Wells, Landerman, McMichael,
and Kaplan (1979) demonstrated the wide variety of strains that workers may experience
when under stress in a study of 1,809 white, male, blue-collar workers. In this study, the
authors examined the relationships of perceived stress to reports of ill health and to five
medical conditions. Controlling for confounding variables (e.g., exposure to noxious
chemical agents), they found that perceived stress was positively related to self-reported
physical symptoms such as angina, ulcers, and medical evidence of hypertension. They
also found that perceived stress was positively related to neurotic symptoms.

Numerous psychological strains have been related to job stress in the literature

including depression, anxiety, irritability, somatic complaints, and job dissatisfaction

10

(Fenlason & Beehr, 1994). Caplan et al. (1975) found evidence of a strong job stress-job
dissatisfaction relationship in their study of 2,010 male workers fi'om 23 occupations.
Basing their predictions on the Michigan Person-Environment Fit (P-E fit) model of job
stress, the authors posited that negative psychological and physical consequences
(strains) result from a discrepancy between participants’ work related needs/desires and
their perceptions of how well these needs/desires were met in their work environment.
While obtaining support for the relationship between job stress and job dissatisfaction,
they did not find a significant direct relationship between job stress and physiological
strains (e. g., blood pressure, pulse rate, cholesterol) or on psychological strains such as
depression, irritation, and somatic complaints.

Although the idea persists that occupational stress contributes to the negative
physical and emotional experiences of workers, there are many inconsistent findings in
this literature. Numerous limitations in this research base may be responsible for these
discrepant results. As recently as 1987, Kahn acknowledged that researchers in this area
have not agreed on the use of the term stress, nor its definition. He cited Selye’s 1976
use of the term stressor to describe an external stimulus that evokes the “nonspecific
response of the body to any demand made upon it” which he called the stress response as
one definition (Kahn, 1987, p. 312). Other authors have used the term stress to define
external Stimuli and strain as its effect. Kahn advocated that the term stressors be used to
refer to external stimuli; the term strain be designated as the product of experiencing,
stressors; and the term stress be reserved for a general description of this area of interest.
These distinctions appear to be gaining popularity in more recent research (Fenlason &

Beehr, 1994).

ll

Barone (1988) criticized the occupational stress literature for failing to be
grounded in a sophisticated assessment of work stressors. He reported that many studies
have relied upon relatively few questions, and sometimes single items or the presence of
indicators such as ‘turn over’, to assess work related stressors. In a similar vein, he was
critical of studies using scales designed by Caplan et al.(1975) to assess stressors. Barone
noted that these scales were developed using an exclusively male participant pool,
consisting primarily of blue-collar workers. In response to these criticisms, Barone
developed and validated the Work Stress Inventory via four studies involving a diverse
group of over 1300 workers (Barone,1988).

Another criticism of the occupational stress literature is that it has been difficult to
make comparisons across studies due to inconsistent Operational definitions of strains.
Some research have explored behavioral strains such as poor job performance, others
have obtained physiological measures (e. g., blood pressure, pulse rate, cholesterol), and
still others have relied upon self-reports of psychological distress symptoms (e. g. somatic
complaints, anxiety, depression, and job satisfaction). Again, these studies often rely
heavily on the instrumentation developed by Caplan et al. (1975) on male blue-collar
workers. These measures contain relatively few items per identified strain.

Restricted sampling has also created problems in the generalizability of much of
the occupational stress literature. Many studies have been confined to one sex and/or to
specific occupations. Furthermore, failure to control for potential covariates such as
‘length of service’ has been characteristic of these studies (Blau, 1981). Investigating
such relationships, Blau (1981) found a strong negative relationship between length of

service and job strains (ineffective performance and job dissatisfaction).

12

Finally, consistent with the P-E fit model of work stress, individual difference
variables may be responsible for discrepant findings among studies exploring the
relationship between occupational stressors and strains. The search for potential
moderators of this relationship has focused primarily on the interrelationships of social
support, stressors, and strains.

Social Support and the Work Stress-Strain Relationship

Perhaps inspired by earlier leadership studies, which revealed that there was a
relationship between perceived supervisor support and subordinate job satisfaction,
Caplan et a1. (1975) investigated the effect of social support on the job stress-strain
relationship (Blau, 1981). The popular belief that social support ameliorates some of the
negative effects of work stress can be traced back to this study. The authors of this
investigation concluded:

A final type of environmental change which may be used to promote

psychological well-being is the provision of social support fi'om one’s supervisor

and from one’s co-workers. . .Our findings indicate that social support can reduce

job dissatisfaction and depression. (Caplan et al., 1975, p. 210)

Since the publication of this study, a great deal of research has investigated the
relationship among these variables.

Social supporhhas been defined in different ways in the literature. House (1981)
summarized thegdifis’tinctions in social support made by previous researchers into four _
groups: emotional support, appraisal support, information support, and instrumental

support. However, Beehr (1985) reported that there is little empirical evidence upon

which these distinctions were based. He suggested a division of social support into two

13

4» _.—. 1.4...” N... ,..__.-

general types: emotional and instrumentaLsupport. Emotional support inc des

-vw/l/

@hipcludes delivering tangible assistance to another worker in the form of physical ..

assistance, adage, knowledge, or the Inaterial goods needed to accomplish his/her job.
More recently, researchers have combined scales of emotional and instrumental support
into a single measure. This strategy is supported by findings that these scales are highly
intercorrelated when the support is provided by a single source, such as from supervisors
(Caplan et al., 1975; Fenlason & Beehr, 1994; Kaufmann & Beehr, 1986).

Research exploring the interrelationships of work stressors, social support, and
strains has proceeded from a predominantly atheoretical basis. Some research has
focused on the main effect of social support on work stressors and strains while others
have explored social support as a buffer of the effect of stressors on strains.

Fenlason and Beehr (1994) asserted that, “The idea that social support can
directly reduce strains is consistent with most of the empirical literature on job stress” (p.
158). They summarized the findings of numerous authors who have demonstrated that
social support is negatively correlated with various types of strains including job
dissatisfaction, life dissatisfaction, somatic complaints, depression, and burnout (Blau,
1981; Ganster et al., 1986; LaRocco & Jones, 1978; Leiter, 1991). However, there has
been little research and mixed findings regarding the role of social support in directly
reducing work stressors. Cohen and Vlfrllis (1985) hypothesized that social support may
act upon a worker’s appraisal of a stressor making him or her perceive a situation as less

threatening. Other researchers have found negative correlations between supervisor

support and role stressors, time pressure, and role ambiguity (Beehr, 1976; Blau, 1980;

14

Caplan et al., 1975). However, LaRocco and Jones (1978) did not find support for a main
effect of leader or co-worker social support on work stressors (i.e., perceived conflict and
ambiguity of organizational goals).

The notion that social support may buffer the effects of work stressors has been
widely studied over the 20 years since the publication of Caplan et al.’s (1975) landmark
study; however, this literature has yielded inconsistent support for buffering hypotheses.
The following section will organize these studies into three groups: (a) those which
yielded support for the “buffering hypothesis”, (b) those which found no support for the
buffering hypothesis, and (c) those which demonstrated instances of “reverse buffering”
and “mixed buffering”.

Support for the Buffering Hypothesis

Gore (1978) examined social support as a bufier of the health consequences of
unemployment. She found that when unemployed workers were also unsupported they
reported significantly greater negative changes in their health status (blood pressure,
cholesterol level, and illness symptoms) than did their supported peers. House and Wells
(1978) found that supervisor support moderated the relationship between a stressor
labeled “role conflict” and neurosis. Support from the worker’s spouse also appeared to
have a buffering effect with regard to several job stressors and strains. LaRocco et al.
(1980) studied men from 23 occupations and found that support bufl‘ered the effects of
stressors on depression, irritation, anxiety, and on somatic complaints. No evidence was
found for the buffering effect of social support on job related strains such as job
dissatisfaction and boredom. Co-worker support was found to be more effective at

buffering workers from strains than was support from either supervisors or from family.

15

Although acknowledging the popularity of the above findings, Haines et al.
(1991) criticized the stress buffering research on primarily methodological grounds.
They observed that previous research frequently analyzed multiple bivariate relationships
rather than using a more appropriate multiple regression procedure, thereby increasing
the risk of Type 1 error in these studies. The authors were also critical of the sampling
procedures used in previous studies, noting that only male participants, representing a
few industries and occupations, were sampled. This restricted sampling strategy severely
limits the generalizability of the findings of these Studies. Haines et al. (1991) attempted
to address these issues in a study of 685 workers (16 years and older) drawn from a
national sample. In particular, their sampling pool consisted of both men and women
fiom a variety of occupations. Additionally, the authors used multiple regression, where
appropriate, to analyze their data. Their findings suggested that workers from different
age, sex, education and income groups may be differentially exposed to stressors and that
these factors should be examined as potential confounding variables in future research.
They also found a significant relationship between work support and several indexes of
work strain. While arguing that “support for the buffer hypothesis is more tenuous than it
appears”, the authors found that work support did moderate the relationship between
stressors and strains Specifically caused by workload and conflict (Haines et al., 1991, p.
226). They found no differences in this effect across occupational groups.
The Absence of Buffering

Although social support has emerged as the primary moderator hypothesized to
mitigate the effects of stress in the workplace, not all research has yielded support for this

“buffering” expectation. LaRocco and Jones (197 8), while supporting the hypothesis that

16

there is a main effect of social support on strains, found no evidence for the buffering
hypotheses in a survey of 3,725 Navy enlisted personnel. Blau (1981) investigated the
buffering hypothesis in study testing the person-environmental fit model of job stress. He
found only partial support for this model. Specifically, he concluded that although work
stress and social support had independent main effects on job dissatisfaction, “no type of
social support (supervisor, co-worker, or off-job) acted as a buffer between any job stress-
job strain relationships” (Blau, 1981, p. 299). Interestingly, he did find, and dismiss, two
“buffering” interactions in this study. He conjectured that these findings were merely
evidence of Type I error in his statistical analysis (i.e., due to the large number of
regression equations employed).

Consistent with the above findings, Ganster et al. (1986) found modest support for
the direct effect of social support in lowering strain, but no support for any buffering
effect of social support on the relationship between stressors and strains. These authors
also acknowledged the problematic nature of computing a large number of regression
equations to investigate the potential buffering effects of different types of social support
on various stressors and strains. They also concluded that the buffering hypothesis may
receive more support in studies concerning stressful life events rather than specific
studies of work stress. The authors speculated that this may be due to the fact that many
life stressors (e. g., death of a spouse or change in residence) directly affect access to
social support whereas work stressors do not.

Instances of “Reverse Bufl‘eriaglaad “Mixed Bufferin ”
Some of the most intriguing findings in the buffering literature have been those

studies which reported a ‘reverse buffering effect’. In these studies, increased social

17

support appeared to exacerbate the effects of stressors on strains. Beehr (1976) found
that coworker support appeared to increase the effects of role ambiguity on worker job
dissatisfaction. This author hypothesized that workers experiencing this stressor may
communicate with each other in a manner which reduces self-blame, instead assigning
blame for role ambiguity on the job itself. He stated that this type of communication may
produce greater overall job dissatisfaction.

In a subsequent study of nurses, Kaufmann and Beehr (1986) found that both
work and non-work related support had reverse buffering effects on the relationship
between an index of work stressors (future job ambiguity and role overload) and
psychosomatic strain. There was also a reverse buffering effect of social support in the
relationship between work stressors and absenteeism. These authors offered three
possible explanations for their findings. First, they extended Beehr’s (1976) argument
that the content of supportive communication may either help workers see that their
situation is not as difficult as it appears, or it may convince them that they are suffering
ten'ibly. Second, they speculated that the source of support may not be independent of
the source of stress (e.g., a supervisor who is causing stress among workers approaches
an individual to give support but actually increases the workers stress level by singling
him/her out). Third, they offered the possibility that frequent stressors leading to greater

strains may cause workers to seek increased levels of support, as opposed to social

 

support interacting with stressors to produce more strain.
In addition to incidents of reverse buffering in this literature, there have also been

9,

reports of “mixed buffering . In these studies, social support was found to reduce the

effect of some strains from certain work stressors, but at the same time, increase the

18

effect of other strains from these work stressors. Chisholm, Kasl, and Mueller (1986)
found mixed buffering results in their study of nuclear worker responses to the Three
Mile Island (TMI) accident. Again, these authors found evidence for a main effect of
social support on strains such as job satisfaction, greater optimism about their future
employment as nuclear workers, and fewer psychosomatic symptoms. Despite these
positive main effects, social support did not always moderate the stressor-strain
relationship in a helpful manner. According to these authors, the direction of the
moderating effect depended upon two factors:

(a) the level of stress being considered and (b) the types of variables being

examined. Social support consistently affects stress/strain relationships at low

levels of stress but not at high stress levels. In contrast, stress/health outcome
relationships are buffered at high stress levels but not at low levels (Chisholm,

Kasl, & Mueller, 1986, p. 191).

In this study, worker’s role classification as either supervisory or non-supervisory
appeared to have an impact on whether buffering or reverse buffering was observed.
Specifically, supervisors at TM] appeared to be adversely effected by supervisor support
when under greater levels of stress. These workers reported greater job dissatisfaction,
lower optimism regarding their job firture, and less occupational self-esteem after the
TMI incident when they were given high amounts of supervisor support. Non-
supervisory staff experienced reverse bufi‘ering of co-worker support on job satisfaction
and perceptions of job firture. “In contrast, all three of the significant effects of
supervisor support on job strains of non-supervisory employees at the two plants support

the buffer hypothesis” (Chisholm et al., 1986, p. 189).

19

Chisholm et al. concluded that social support generally produces positive main
effects on strain; however, the interaction of stressors and social support does not always
yield beneficial results (lower strain). The authors concluded that the reason for
differences in the buffering effect between supervisory and non-supervisory workers was
not clear from their current study. Finally, they pointed to literature suggesting that it is
unlikely that workers use only one strategy to cope with the complexity of organizational
institutions. The authors stated that workers may use a combination of social support and
“defensive copin ” to deal with stress in their work settings.

Fenlason and Beehr (1994) also obtained “mixed buffering” results. They
explored whether the content of supportive communication differentially affected the
bufl‘ering effect of social support on strains in a sample of 351 female secretaries. These
authors identified the following three types of supportive communication: 1) positive
work-related conversation (e. g., congratulating one another on a job well done); 2)
negative work-related conversation (e.g,. commiserating regarding the problems with
working in a given environment); and 3) non-work related conversation (e. g., discussing
non-work interests). When including the content of communication measure, they found
that stressed workers who received positive job-related communication from their
supervisors and family/friends reported less strains. However, greater co-worker
negative job-related communication (interacting with the stressor “underutilization of
skills”) was associated with increased strain (reverse buffering). Increased job related
communications with family and friends was also associated with greater strains. The
authors speculated that positive communication results in buffering whereas negative

communication (especially as related to work) is more likely to contribute to reverse

20

buffering effects.

In addition to the above analysis, the authors also examined the effects of social
support using a general support index. They hypothesized that, although support from all
sources (i.e., supervisors, co-workers and family/fiiends) would be related to strain,
supervisor support would be more highly related to strain than the other sources of
support. They speculated that co-worker support would be the second most strongly
related to strain. The authors found this relationship with regard to the general support
index, but not with the contents of communication scale. The only interaction observed
using a general support index, rather than their contents of communication scale, revealed
that the experience of role conflict/overload appeared to be greater under conditions of
higher coworker instrumental support indicating a ‘reverse buffering effect’.

Serum

Research regarding the interrelationships of work stressors, social support, and
strains has generated inconsistent findings. Future research should focus on theory-based
individual difference variables which are presumed to moderate these interrelationships.
Clearer definitions of all of these constructs, along with improved sampling,
instrumentation, exploration and control of potential covariates, and analysis of distinct
types of strains (i.e. job-satisfaction and symptomatic distress) would greatly enhance the
quality of this line of occupational stress research.

Alpaalrment Theory Mmework for Conceptualizing the Interrelationshi_p§
Among Work Stress. Social Support, and Worker Sm
An explanation for the inconsistencies found in the bufi‘ering research may be

derived through understanding individual differences in affect regulation and the

21

perception of social support. Attachment theory provides a conceptual framework for
making such distinctions and may explain why increased social support appears to assist
some workers in coping with stress, whereas disposing others to experience greater
strains.

The origins of attachment theory can be found in the work of John Bowlby and
Mary Salter Ainsworth (Ainsworth & Bowlby, 1991; Bretherton, 1992). Trained as a
psychoanalyst, Bowlby initially sought empirical support of the tenets of object relations
theory (Holmes, 1993). However, his exploration outside of the realm of analytic theory,
into ethology (the scientific study of animals) and cybemetics, was the catalyst for
Bowlby’s elaboration of Attachment Theory. This developmental theory emphasizes the
importance of close, enduring emotional bonds (or “attachments”) in the formation of the
human character. Bowlby posited that a child’s actual experiences in the relationship
with his/her primary caregiver were instrumental in determining his/her later personality
development and relationship behavior. Ainsworth, a co-worker of Bowlby’s in the early
1950’s, would later provide empirical evidence supporting the existence of distinct
attachment styles in humans. The classification and measurement of attachment styles,
the continuity of attachment classification from infancy through childhood, and the
lasting influence of attachment styles throughout the lifespan have been the focus of
recent research investigations. The following review will provide the reader with an
overview of the theoretical foundations of attachment theory, the current status of the
classification and measurement of adult attachment styles, and a sample of research
findings regarding the application of attachment theory to understanding an individual’s

reaction to stress, perception of others, and adult interpersonal behavior.

22

Attachment Theory: Key Concepts and Assumptions

Bowlby considered attachment as a basic behavioral system, distinct from mating
and feeding, with its own evolutionary purpose. He defined attachment behavior as
behavior with the goal of achieving proximity to a caregiver “whose evolutionary
function is protection of the infant from danger” (Bretherton, 1992, p. 766). The child
and his/her primary caregiver develop a “goal-corrected partnership” that preserves a
specific proximity or “set goal”. Infants display “attachment behaviors” that help
maintain this set goal of proximity. These behaviors include verbal protests (e. g., crying
or calling out) and also following or clinging to the primary caregiver. The parental
contribution to this attachment partnership is to maintain a child within safe boundaries in
the environment (i.e., a range where the child may explore but can return to safety if
threatened).

Bowlby believed that actual experiences in an infant’s life lead him/her to develop
internal representations of the elements in the world around them. These representations
are considered “working models” of the self, others, and the environment. Two types of
experiences are crucial to the development of secure attachments to others; these are the
experience of a secure base and the opportunity to explore the environment. The ‘secure
base’ is a concept introduced by Ainsworth and embraced by Bowlby in the early
nineteen-eighties (Bowlby, 1988). This term describes a condition in which children may
explore their environment secure in the knowledge that when they return to their primary
caregiver(s) they will be welcomed, comforted, reassured, and that their basic physical
and emotional needs will be met. Children may safely engage their curiosity and test the

limits of their abilities under these circumstances. Holmes (1993) summarized the efi‘ect

23

of this phenomenon as follows: “We can endure tough seas if we are sure of a safe
haven” (p. 70).

Furthermore, Bowlby asserted that when a child’s needs for attention, comfort, and
protection are not met in infancy, lasting affective and behavioral problems can result.
He proposed that children who lack a secure base would be more prone to depression,
less resilient to stress, and would experience difficulty in subsequent intimate
relationships (Bowlby, 1988). In addition, he believed that the absence of a secure
working model inhibited the child’s exploration of his or her surroundings, leading to
continued insecurity regarding self, others, and the environment.

Mary Salter Ainsworth has been credited with advancing attachment theory
through her development of an empirical methodology for identifying infant attachment
styles (Bretherton, 1992). This method, called The Strange Situation, consists of a
specific sequence of events, orchestrated in a laboratory setting, with the intention of
activating the attachment system of the child through the stress of separation fi'om his/her
mother. Observers recorded and rated the mother-child interaction, as well as the child’s
behavior after a series of episodes involving separation and reunion.

The Strange Situation proved to be very usefirl in distinguishing patterns of
attachment in the mother-child dyads. Ainsworth identified three distinct patterns of
attachment which she labeled: secure, insecure/ambivalent, and insecure/avoidant.
Children who were labeled “secure” in their attachment protested when their mothers left
the room and sought contact when she returned. They were described as easily consoled
and able to resume exploratory behavior. Their mothers were described as sensitive and

responsive to their infant’s needs. The insecure/ambivalent infants protested when their

24

mothers left and were alternately clingy and angry upon her return. Their mothers were
observed to be inconsistently responsive to their infant’s needs. Insecure/ambivalent
children were distinguished fiom another group (insecure/avoidant) who also
demonstrated insecure behavior, but who appeared unaffected by the separation from
their mothers in the laboratory and ignored them when they were reunited. The mothers
of insecure/avoidant children were observed to be insensitive to their infant’s signals and
were especially rejecting when their child sought physical contact.

Ainsworth’s pioneering studies were greatly influenced by her correspondence
with Bowlby as he developed the first volume of his classic trilogy, Attachment and Loss
(Ainsworth & Bowlby, 1991). In these volumes, Bowlby (1969/1982, 1973, 1980)
proposed that our childhood attachment styles, while not fixed, are relatively stable over
time and continue to influence how we view ourselves and others. He believed that this
is achieved through a cybernetic process in which attachment style influences an
“individual’s perceptions, information-processing, and interpersonal behavior in ways
that produce schema-consistent experiences” (Lopez, 1995, p. 402). As an individual
grows into adulthood, others who occupy emotionally significant roles in the person’s life
and who are perceived as more powerful and/or wiser (e. g., supervisors) may become the
objects of attachment-related perceptions and behaviors (Bowlby, 1979; Lopez, 1997).
Although the frequency and intensity of attachment behaviors diminish in adulthood, they
tend to be “especially evident when a person is distressed, ill, or afraid “ (Bowlby, 1979,
p. 129). However, this ‘continuity hypothesis’ has not been critically tested as there have
been no longitudinal studies of attachment styles from childhood through adulthood

(Hendrick & Hendrick, 1994).

25

Contemporm Attachment Theory: Theoretical and Empirical Fourydations

A major advance in contemporary attachment theory has been its application to
understanding the behavior and relationships of adults. This body of research has been
building since 1987 when Hazan and Shaver developed the first self-report measure of
adult attachment styles. Since that time adult attachment researchers have focused on: a)
improving the classification and measurement of adult attachment styles; b) the extension
of attachment theory to understanding the affective and interpersonal relationships of
adults; and c) the exploration of the impact of adult attachment style of work-related
attitudes and behaviors.

The classification and measurement of adult atpaphment styles.

In their pioneering study, Hazan and Shaver (1987) translated Ainsworth’s
classification of childhood attachment into language appropriate to describe three
different patterns of adult functioning (secure, avoidant, and anxious/ambivalent) and
asked participants to decide which description best fit their feelings in close relationships
(See Table 1). In their subsequent research, these authors found support for the idea that
Ainsworth’s attachment style categories could apply to adult romantic relationships.
They also found that differences in adult attachment were related, in a theoretically
consistent manner, to retrospective self-reports of childhood relationships with parents.

Although this pioneering work provided the impetus for a great deal of research
into adult attachment, numerous authors criticized Hazan and Shaver’s use of a single
item, categorical measure to assess attachment style in adults. This criticism lead to the
development of continuously scaled measures of adult attachment style (Collins & Reed,

1990; Simpson, 1990). These measures were developed through decomposing Hazan and

26

Table 1

Hazan and Shaver Attachment Stfle Inventory

 

Which of the following best describe your feelings?

Secure
I find it relatively easy to get close to others and am comfortable depending on
them and having them depend on me. I don’t often worry about being abandoned
or about someone getting too close to me.

Avoidant
I am somewhat uncomfortable being close to others; I find it difficult to trust them
completely, difficult to allow myself to depend on them. I am nervous when
anyone gets too close, and often, love partners want me to be more intimate than I
feel comfortable being.

Anxious/Ambivalent
I find that others are reluctant to get as close as I would like. I often worry that
my partner doesn’t really love me or won’t want to stay with me. I want to merge
completely with another person, and this desire sometimes scares people away.

 

27

Shaver’s three descriptive paragraphs into separate items. Respondents rated themselves
on each of these items. A factor analysis of these responses provided for the
identification of two- (Simpson, 1990) or three- (Collins & Reed, 1990) dimensions
thought to underlie adult attachment styles. After reviewing this research, Hazan and
Shaver (1993) asserted that there are two dimensions underlying their measure: the first
dimension represents the extend to which a participant expresses comfort with
interpersonal closeness and dependence on others; the second dimension reflects the
degree of anxiety or tension that the participant reports regarding separation and distance
in romantic relationships.

More recently, Bartholomew (1990) proposed a four-group classification of adult
attachment style (secure, preoccupied, avoidant, and fearful). This typology is grounded
in Bowlby’s theory that individuals have internalized working models of self and others
based on their childhood experiences. Contained within a working model of self is the
individual’s belief regarding his/her worthiness of receiving support and love. The
individual’s working model of others holds his/her perception of the availability of the
attachment figure to meet his/her needs. In this classification system, secure individuals
are said to have positive views of self and of others; while insecure adults hold negative
beliefs about either self, others, or both. Preoccupied individuals correspond closely to
Hazan and Shaver’s anxious/ambivalent description. They hold a negative view of self
and a positive view of others. Both the dismissive and the fearful groups report wariness
regarding intimacy; however, the difference between these two groups may be found in
their distinct motivations for avoidance in relationships. Although dismissive individuals

hold a negative view of others, they report a positive view of self. Their avoidance of

28

intimacy may be due to their unwillingness to compromise their independence for
closeness with others. Fearful individuals hold both a negative view of others and
themselves which may cause them to believe that they are unworthy of closeness. Unlike
their dismissive counterparts, these individuals also fear interpersonal rejection.

Bartholomew and Horowitz (1991) developed a categorical measure of these four
styles based on the design of Hazan and Shaver’s Attachment Style Inventory. These
authors obtained empirical support for the four-category typology of adult attachment.
Grifl'rn and Bartholomew (1994) assessed the validity of the self and other dimensions of
the four-category model and found that these dimensions of attachment had construct,
discriminant, and predictive validity.

In a study comparing the three- and four- category models of attachment, the
same two dimensions were found to underlie both models (Brennen, Shaver, & Tobey,
1991). The results of this study also indicated that some of Hazan and Shaver’s
participants may have been misclassified based on their need to conform to the three
choices which they were offered. The authors found that most participants who classified
themselves as secure did so on both measures; however a minority of the respondents
who classified themselves as secure on Hazan and Shaver’s measure, endorsed the
dismissive item on Bartholomew’s scale. Anxious/ambivalent participants classified
themselves as fearful or preoccupied on Bartholomew’s measure, and avoidant
participants distributed themselves into the dismissive and fearful categories. These
findings support the use of a four-category typology of adult attachment style as it allows
for finer distinctions along the two dimensions reflecting an individual’s view of

him/herself and others.

29

In a recent review, Lopez (1995) discussed numerous conceptual and
methodological limitations in adult attachment research including a critique of these
measurement-related difficulties. In addition, he revealed that there have been
inconsistent findings regarding sex differences in adult attachment classification which
may be due, in part, to the use of both three- and four- group classification models in
across study comparisons. One study employing the four-group scheme found women
overrepresented among individuals classified as fearfully attached and males
overrepresented among persons classified with dismissive attachment styles. In other
studies employing the four group taxonomy, no sex differences were noted across three
different independent samples (Lopez, 1995). Cultural differences in the relative
distribution of attachment styles have also been found using The Strange Situation. In
response to these criticisms, Lopez stated that firture research should carefully examine
the potential moderating effects of gender, sex-role orientation, and cultural differences
on the manner in which “attachment behavior is expressed and cooperatively managed in
close relationships” (LOpez, 1995, p. 408).

Adult flchment related distinctions in affect regplation and interpersonal
mm

Research on adults has produced a body of literature that supports numerous
distinctions among individuals with different attachment styles. Researchers have found
that adults with secure attachment styles report more relationship satisfaction, higher
levels of trust in relationships, greater use of appropriate self-disclosure, more
constructive approaches to conflict resolution, and greater collaboration in problem

solving than their insecure peers (Camelley, Pietromonaco, & Jaffe, 1994;

30

Lopez, Gover, Leskela, Sauer, Schirmer, & Wyssmann, 1996; Mikulincer & Nachshon,
1991; Pistole, 1989; Simpson, 1990). Adult attachment style differences in affect
regulation and the development and perception of social support have also been observed.
These findings can be helpfirl in theorizing about how individuals of different attachment
styles may react to work related stress, and whether or not social support would be usefirl
to them in reducing strains.

Kobak and Sceery (1988) found that secure college students endorsed fewer
symptoms of distress on a self-report measure than did insecure respondents. This study
also revealed that peers rated secure participants as more ego-resilient, less hostile, and
less anxious than insecure participants. Attachment style differences in affect regulation
were further supported through the work of Priel and Shamai (1995) who found that
secure respondents reported less anxiety and depression than insecure respondents.

Researchers have also begun to investigate the relationship between patterns of
attachment to others and the development and perception of social support. Several
authors have shown that secure participants tend to perceive more support in times of
distress than do their insecure peers (Florian, Mikulincer, & Bucholtz, 1995; Kobak &
Sceery, 1988; Preil & Shamai, 1995). Preil and Shamai (1995) also found that secure
participants reported more satisfaction with social support than did insecure participants.
In addition, researchers have studied support seeking and support giving in couples
during an anxiety provoking situation (Simpson, Rholes, & Nelligan, 1992).

Simpson et al. (1992) designed an experiment to study the relationship of adult
attachment to support seeking and giving in heterosexual dating couples. During this

experiment the female member of each couple was placed under stress. The researchers

31

observed the extent to which these women sought and accepted social support fiom their
male partners. In addition they recorded the extent to which male participants offered
support to their partners. Participants were classified on two continuous scales, believe to
represent the underlying dimensions of attachment style: a) Anxious versus Nonanxious;
and b) Secure versus Avoidant. No significant differences in support seeking and giving
were revealed with regard to classification on the Anxious versus Nonanxious scale.
However, with regard to support seeking by female partners and support giving by male
partners, the authors found that more secure females sought greater support as their level
of anxiety increased, and more secure men tended to ofl‘er greater support as their
partner’s level of anxiety increased. In contrast, more avoidant females sought less
support as their anxiety increased and more avoidant males gave less support as their
partner’s anxiety increased. Interestingly, at lower levels of anxiety, more avoidant
females sought greater support, and more avoidant males delivered more support than
their more secure peers. This finding was explained in terms of the avoidant participants
conflicting desire for, and yet fear of, proximity. The authors speculated that, “increases
in perceived threat or distress sharply accelerate the onset of fear of proximity, resulting
in decreased proximity seeking and giving” within the more avoidant groups (Simpson et
al., 1992, p. 443). With regard to their reactions to support, both avoidant and secure
participants appeared calmed by their partner’s support.

Wallace and Vaux (1993) focused their research on understanding personality
characteristics that influence an individual’s ability to develop social support networks.
They reasoned that negative help-seeking experiences, especially in the formation of

early attachment styles, would play a critical role in the development of a negative belief

32

system regarding obtaining help from others (i.e., result in a negative network
orientation). In their research on the relationship between network orientation and
attachment style, these authors found that individuals who reported insecure attachment
styles (i.e., anxious/ambivalent or avoidant) were more likely to “endorse beliefs and
expectations reflecting the risk, costs, and futility of seeking help from network
members” (Wallace & Vaux, 1993, p. 362).

Adult mchmerL and work-related behavior.

In a recent paper, Blustein et al. (1995) addressed the application of attachment
theory to career development and organizational behavior. These authors asserted that
our understanding of work-related behavior in adulthood would be enhanced by
incorporating the role of relationships into career development theory. After a review of
relevant literature, they proposed that workers with secure attachment styles would be
more likely to experience adaptive relationships at work than insecure workers and would
report higher levels of job satisfaction.

In a groundbreaking study, Hazan and Shaver (1990) proposed that, in several
respects, adult work is functionally similar to Bowlby’s concept of ‘exploration’ in
childhood. These authors asserted that, for adults, work provides a source of actual and
perceived competence just as play and exploration provide these opportunities for
children. Furthermore, they stated that, “the balance between attachment and exploration
associated with healthy functioning early in life is, in important respects, similar to the
love/work balance that makes healthy functioning in adulthood” (Hazan & Shaver, 1990,

p.270)

33

Hazan and Shaver (1990) recruited members for their study using a questionnaire
published in a Sunday magazine supplement. Respondents were, on average, graduates
of college whose household income was between $30,000 to $40,000 per year. Their
results yielded tentative support for several theoretically derived hypotheses. For
example, they found that secure participants “approach their work with the confidence
associated with secure attachments”(I-Iazan and Shaver, 1990, p. 278). In this respect,
secure participants appeared to value and enjoy their work while still placing their
relationships with others as the primary focus of their lives. Anxious/ambivalent workers
reported that they often feared rejection for poor performance on the job and that their
preoccupation with issues related to love gets in the way of their productivity at work. Of
the three attachment style groups, avoidant workers reported the least satisfaction with
their jobs; firrthermore, their response patterns suggested that they used work to avoid
social interactions. The results of this study are considered preliminary as the
instrumentation used by the authors consisted of an unstandardized questionnaire of work
attitudes and of a single item, forced choice measure to classify attachment style.
Furthermore, their sampling technique limited the generalizability of these findings to a
population that cannot be considered as representative of the larger population of adult
workers.

In an effort to support and extend the findings of Hazan and Shaver, Hardy and
Barkham (1994) studied a clinical population who where reporting occupational stress
and depression. Their sample consisted of 219 participants who were referred to a clinic
for distressed white collar workers. For inclusion in the study, participants were

employed and were determined to be clinically depressed. They found that workers

34

classified as more anxious/ambivalent were fearful about their work performance and
their relationships on the job. In contrast, the authors found that workers who received
higher scores on the avoidant scale reported greater job dissatisfaction, more conflict with
co-workers, and greater difficulties in their social relationships outside of work than did
workers who scored lower on this scale. The sampling procedure used in this study limits
the generalizability of the findings to populations of workers admitting distress. The
authors in this study did improve upon the methodology of Hazan and Shaver by using
more standardized measures; however, to measure attachment style they developed their
own scale that should be considered exploratory.
Summw and Problem Restatement

There have been inconsistent findings in the occupational Stress literature
regarding the effect of social support on the relationship between work stressors and
strains. Problems regarding sampling techniques, instrumentation, and data analysis may
be contributing to these discrepant findings. In addition, important individual difference
variables, which have yet to be adequately studied in this context, may (a) be
significantly related to perceptions of stress, strain, and social support, and (b) moderate
the stressor-strain relationships. Attachment theory provides a useful framework for
hypothesizing about how individuals differ in their affect regulation and social behavior
when under stress, their perceptions of social support, and their work-related attitudes and
behavior.

Contemporary research on adult attachment has revealed adult attachment style
differences in affect regulation and on indexes of adaptive relationship behavior. Given

these findings, it seems reasonable to hypothesize that adult attachment style may be

35

directly related to indexes of symptomatic distress. In addition, recent studies regarding
the contribution of adult attachment style to work behavior suggest that, by understanding
a worker’s attachment style, we may better predict job satisfaction. Furthermore, since
attachment behavior is activated during times of stress, it seems reasonable to
hypothesize that attachment style would moderate the relationship between work stress
and indexes of both symptomatic distress and job satisfaction.

Hardy and Barkham’s findings would also suggest that attachment style may
influence the quality of an individual’s work relationships. Since attachment theory
predicts that attachment behavior in adults is likely to be demonstrated with regard to
preferred individuals perceived as more powerful/or wiser, it is assumed that supervisory
relationships will be more likely to be influenced by attachment style differences than co-
worker relationships. In addition, adult attachment has been demonstrated to influence
individuals perceptions and beliefs regarding the safety of seeking help fiom others in
times of distress; therefore, under stressfirl conditions, attachment style may moderate the
relationship between supervisor social support and symptomatic distress by clarifying
which individuals are likely to seek and use social support when stressed. It is also
predicted that attachment style will moderate the relationship between supervisor social
support and job satisfaction under stressful work conditions.

Definitions
Stressors
In this study, work stress is conceptualized as the worker’s perception that

existing work demands, or “stressors”, exceed his/her capabilities (Barone et al., 1988).

36

“Work stress” is an interactional construct that involves an appraisal regarding the
intensity and frequency of this discrepancy by workers.

“Stressors” are defined as the specific job-related contributors to work stress
(F enlason & Beehr, 1992). Work stressors for an employee may include a perceived lack
of one or more of the following: a) information, b) input into decisions, c) autonomy, (1)
clear communication, e) support, and f) recognition from supervisors (Barone, 1988).
Other work stressors involve job risks and may include conflicting and excessive job
demands, the need for emergency responding, extended work without reprieve, and other
threats of harm to self and/or others (Barone, 1988).
Sam

For the purpose of this Study, “strains” are the psychological outcomes that are a
result of experiencing work stressors (F enlason & Beehr, 1992). In this study, two types
of strains will be examined: a) self-reported job dissatisfaction, and b) self-reported
symptomatic distress. Job dissatisfaction is defined as the extent to which an employee
indicates lack of overall satisfaction with his/her current job. Symptomatic distress is
defined as the intensity of discomfort experienced by participants with regard to
symptoms commonly observed among mental health outpatients.
Social Support

In this study, social support will be defined as emotionayor instrumental support
given to the participant by one of the following three sources: supervisors, co-workers,
and others (spouse, friends, and relatives). Emotional support includes statements of
caring or listening to one’s concerns. Instrumental support refers to providing tangible

assistance that is necessary to complete a task. This type of support may include

37

supplying advice, physical assistance, or the necessary materials for a worker to do
his/her job.
Adult Attachment Styles

In this study, Bartholomew's (1990) four-group taxonomy of adult attachment
styles will be used. The four adult attachment styles are defined as follows:

Secure adults report that they are comfortable with connection with others. They

 

can also tolerate separation from others. They generally have a positive view others and
of themselves. Secure adults report positive family memories from their childhoods.
They describe their parents as being available and responsive to their needs.

Preoccupifidults focus intensely on relationships. They are, in general, more
comfortable with connection than separation, although they express strong fears of
abandonment. They have developed a negative view of themselves while viewing others
in positive terms. For preoccupied adults, obsessiveness and jealousy often interfere with
their relationships with others. Preoccupied adults report negative family memories from
childhood. They describe developmental histories marked by inconsistent parenting, lack
of parental support, and role reversal in their relationships with their parents.

Dismissive adults generally dismiss the need for connection with others and they
tend to be counter-dependent in their relationships. Individuals with this style of
attachment view themselves in positive terms while maintaining a negative view of
others. Dismissive adults expect their partners to be unavailable and unresponsive to
their needs. These expectations results in a general deactivation of attachment proximity-

seeking behaviors.

38

Learfll adults are generally afraid of connection with others, express
abandonment fears, and demonstrate a social avoidance. Individuals with this style of
attachment view themselves and others in negative terms. F earfirl individuals doubt their
ability to be loved and expect their partners to be rejecting. They may demonstrate an
erratic combination of avoidant and anxious proximity-seeking behaviors.

Hypotheses

There have been inconsistent findings regarding the role of social support in
moderating the relationship between work stressors and strains. Attachment theory
provides a conceptual model for reasoning about these discrepancies. Specifically,
individuals with different attachment styles may perceive and react to social support in
different ways, thereby moderating the effect of social support on the stressor-strain
relationship. This investigation will address these relationships as well as other gaps in
the adult attachment and social support buffering literature. The following hypotheses
will be set forth in this study:
1. Relations among work stress, social support, and strain

Hypothesis 1. Work stress will be significantly correlated with indexes of strain.

1a. Stress will be significantly correlated (negatively) with job
satisfaction.

1b. Stress will be significantly correlated (positively) with symptomatic
distress

Hypothesis 2. Supervisor support will be significantly correlated with indexes
of strain.

2a. Supervisor support will be significantly correlated (positively) with
job satisfaction.

2b. Supervisor support will be significantly correlated (negatively) with
symptomatic distress.

39

Hypothesis 3. The source of social support will be significantly related to job
dissatisfaction.

3a. Supervisor support will be more highly related to job satisfaction than
will support from family/friends.

Hypothesis 4. Supervisor support will significantly moderate the relation of work
stress to indexes of strain. Specifically, under high stress conditions,
workers with high supervisor support will report significantly less strain
than will workers with low support.

4a. Under high stress conditions, workers with high levels of supervisor
support will report significantly greater job satisfaction than will
workers with low levels of supervisor support.

4b. Under high stress conditions, workers with high levels of supervisor
support will report significantly less symptomatic distress than will
workers with low levels of support.

II. Relationships of attachment style differences to work stress, supervisor support,
and indexes of strain.

Hypothesis 5. Workers’ attachment styles will be significantly related to their
levels of work stress.

5a. Secure workers will report significantly less work stress than will
fearful workers.

Hypothesis 6. Workers’ attachment styles will be significantly related to their
perceptions of supervisor support.

6a. Secure workers will report higher levels Of supervisor support than
will fearful workers.

Hypothesis 7. Workers’ attachment styles will be significantly related to their
indexes of strain.

7a. Secure workers will report significantly higher job satisfaction than
will fearful workers.

7b. Secure workers will report significantly less symptomatic
distress than will fearful workers.

40

Hypothesis 8. Controlling for the main effects of work stress and supervisor
support, adult workers’ attachment orientations will significantly interact
with their level of supervisor support to predict indexes of strain.
Specifically, workers reporting lower anxiety and higher avoidance will
exhibit significantly lower levels of strain under low support conditions
than will workers reporting higher anxiety and lower avoidance.

8a. Under conditions of low supervisor support, individuals
acknowledging high anxiety will report less job satisfaction and
higher symptomatic distress than workers expressing less anxiety.

8b. Under conditions of low supervisor support, individuals
acknowledging high avoidance will report higher job satisfaction
and lower symptomatic distress than workers expressing less
avoidance.

41

CHAPTER IH
METHODOLOGY

Participants

A sample of 250 Michigan State University (MSU) employees was recruited to
participate in this study through a weighted sampling procedure. The two subsamples in
this study were university support staff and faculty/academic staff. University support
stafl‘ is comprised of clerical-technical personnel, maintenance and skilled trades laborers,
campus police, and operating engineers. The faculty/academic staff subsample is
composed of professors, coaches, administrators, extension personnel, and library staff.
Participants were randomly selected from their respective populations producing a total
sample which was representative of MSU employees. A total response rate of 50% was
achieved (N = 125). Surveys were dropped fi'om the study if they contained a significant
degree of incomplete information (e. g. blank measures), leaving117 valid surveys which
were included in the data analysis. This sample size exceeded the number of responses
needed (N = 94) to detect a medium effect at an alpha level of .05, with eight predictors
and a power of .80 (Cohen, 1992).

The final sample was comprised of 61% Staff members (n = 71) and 39%
Faculty/Administrators including those who categorized their occupational group as
“Other” (n = 46). In the overall population of MSU employees, 58% were classified as

Staff and 42% were classified as Faculty/Administrative staff. Table 2 provides detailed

42

Table 2

1 Dem a hic Inf ti n

 

 

 

 

 

 

 

 

 

Variables Population Sample
N = 8574 n = 117
Total % Total %
Occupational Group
Staff 4936 58% 71 61%
Faculty, Administration, Other 3638 42% 46 39%
Sex
Female 43 77 51% 70 60%
Male 4197 49% 47 40%
Age Mean =
44 years
Ethnicity
African-American 578 7% 5 4%
Asian-American 399 5% 4 3%
Hispanic/Latino(a) 270 3% 2 2%
Multiracial 0 0%
Other 3 3%
Caucasian/White 103 88%
Other, Including Caucasian 7327 85%
Marital Status
Single 14 12%
Committed Partnership 10 9%
Married 80 68%
Divorce 12 10%
Widowed 1 1%
Highest Level of Education
High School/GED 10 9%
Some College 21 18%
Associates Degree 9 8%
Bachelors Degree 1298 15% 31 27%
Masters Degree 1086 13% 15 13%
Doctorate Degree 2784 32% 25 21%
Other 6 5%
Length of Service at Present Job Median Median = 5-10 yrs.
= 5-9
years

 

 

Yearly Income

 

 

 

 

Median =

$3 0,000-$40,000
Mode =

$20,000 to $29,999

 

43

 

Table 2

Sample Demoggaphic Information - Continpad

 

 

 

 

 

 

 

 

Variables Population Sample
N = 8574 n = 117
Total % Total
Are Supervised by Others
Yes 112 96%
No 5 4%
Supervise Others
Yes 66 57%
No 47 40%
Missing Data 04 03%

 

 

 

44

 

demographic information about MSU workers and the sample that was used for this
study. The sample population had more women (60%) than men (40%) represented,
although the MSU worker population is more evenly divided with 51% women and 49%
men. African-American and Asian-American workers were also slightly under
represented in this sample. The sample was similar to the general worker population in
terms of average length of service at MSU (5-10 year vs. 5-9 years respectively). No
information was available regarding the average age, marital status, or the average yearly
income of MSU employees. In our sample the average age of respondents was 44, 68%
were married, and their average income was between $30,000 and $39,999 (the modal
income was between $20,000 and $29,999). Ninety-six percent of the sample reported
that they are supervised by others and 57% reported that they were supervisors as well.
Procedures

To minimize the risk of an order effect, the order of the survey instruments was
varied, creating 6 versions of the survey. However, all packets had the background and
demographic form as the last instrument, as recommended by Dillman (1991). Each
participant was randomly assigned one version of the survey. The survey packet was
comprised of a demographic and background information form, two measures of adult
attachment style, a social support measure, an inventory of work stress, a job satisfaction
scale, and a scale of symptomatic distress.

To increase the likelihood of survey response, a pre-notification postcard was sent
to participants approximately two weeks prior to sending the actual survey packet
(Weather, Furlong, & Solozano, 1993). This letter provided an introduction to the study,

notification of when the participant could expect the survey packet to anive, and a

45

request for participation in the study (See Appendix A). As incentive for participating in
this study, subjects were notified that individuals returning completed surveys would be
entered into a drawing for a $100.00 prize.

The survey packet consisted of a) a personalized cover letter explaining the
purpose of the study, guaranteeing confidentiality and requesting participation in the
study (See Appendix B), b) an informed consent form (See Appendix C), c) a
demographic form (see Appendix D), d) one of six versions of the survey, and e) a
stamped return envelope.

Two follow up contacts were made with participants. The first follow up contact
was a postcard sent to each participant one week after the survey had been sent (See
Appendix K). The postcard served as a reminder to those participants who had not
completed the survey, and to thank those participants who had already returned them.
The second follow-up mailing was sent to the non-respondents three weeks following the
initial mailing. The content of this mailing was a) a letter explaining that the participant’s
completed survey had not yet been received and reviewing the information from the
original cover letter (See Appendix L), b) a second copy of the same version of the
survey, and c) another stamped return envelope.

Instruments

The survey packet included a demographic and background information form;
two measures of adult attachment (the Relationship Questionnaire and Brennan et al.’s
Experiences in Close Relationships); one measure of work stress (Work Stress

Inventory); one measure of social support (Caplan et al., 1975); one measure of job

46

satisfaction (Job Satisfaction Scale); and one global measure of psychological adjustment
(The Hopkins Symptom Checklist).

Demographic and Background Infomation Form. Participants were asked to
provide the following demographic information: sex, age, level of education, ethnicity,
and marital status. This form also inquired about the participant's job classification
(faculty, staff, or administrative personnel) and length of time that he/she has been at
his/her present job. Length of services was reported on a continuous scale with 1 = less
than 1 year and 7 = more than 20 years on the same job. In addition, subjects were asked
to indicate whether or not they supervise others and if they have a supervisor themselves
(See Appendix D).

_A_tt_achment Styde. Two measures of adult attachment, one categorical and the
other continuous, were included in this survey.

_Re_lationship Ouestionaaire. (RQ; Bartholomew & Horowitz, 1991). The
RQ is a categorical measure designed to classify a respondant's attachment style based on
his/her response to a single item (See Appendix E). Participants are asked to indicate
which one of four descriptive paragraphs best portray their feelings about closeness and
intimacy in romantic relationships. The four paragraphs respectively represent secure,
dismissive, preoccupied, and fearfirl attachment styles. This measure has demonstrated
moderate stability (test-retest correlations of .49 to .71) over an eight month period
(Scharfe & Bartholomew, 1994a). Over a period of two years, this instrument has
demonstrated test-retest reliability of .30 to .67 (Scharfe & Bartholomew, 1994b).
Bartholomew and Horowitz (1991) provided convergent validity for this instrument with

both friend and self reports of respondent’s self-concept and sociability. A recent study

47

(Mickelson, Kessler, & Shaver, 1997) using a nationally representative sample reported
the following attachment style distributions among 35 — 44 year olds: 9% endorsed
anxious attachment styles, 59% classified themselves as secure, 28% reported avoidant
attachment styles, and 4% of respondents were unclassifiable. Among 44-54 year olds in
the same sample 8% endorsed preoccupied attachment styles as characteristic of them,
64% classified themselves as secure, 23% reported avoidant attachment styles, and 5% of
respondents were unclassifiable. These findings are consistent with the distribution of
attachment scores in other studies using older adults (Diehl, Elnick, Bourbeau &
Labouvie-Vref, 1998; Klohen & Bera, 1998). For exploratory purposes, a second part
was added to this measure which required participants to rate how characteristic each
paragraph was of them with 1 = “not at all” and 7 = “extremely” (Behrens, 1998). This
rating scale can also be used to assign an attachment category in the event that a
respondent endorses more than one paragraph as most descriptive of them or neglects to
choose a category at all (Davila, Burge, & Hammen, 1997).

Experiences in Close Relationships. (ECR; Brennan, Clark, & Shaver,
1996). This 36-item inventory is a continuously scaled measure of adult attachment (See
Appendix F). Respondents were asked to indicate the extent to which each item
describes how they have typically felt in romantic relationships. Responses were scored
using a 7-point Likert scale (1 = Disagree Strongly and 7 = Agree Strongly). This
instrument provides scores on two subscales that were derived through a comprehensive
factor analysis of multiple self-report indexes of adult attachment orientations. The
avoidance subscale measures the respondents’ reported level of comfort with

interpersonal intimacy and dependency. A high score on this subscale indicates greater

48

avoidance in relationships. The amaty subscale measures the reported level of worry
and tension that the respondent experiences in close relationships. A high score on this
subscale indicates greater anxiety in relationships. Cronbach alpha reliability coefficients
of .94 an .91 for the avoidance and anxiety subscales respectively, have been previously
obtained (Brennan et al., 1996). In this study Cronbach alphas of .93 and .91 were
obtained for the avoidance and anxiety subscales, respectively.

To explore the correspondence between the categorical and continuous measures
of adult attachment, a one-way AN OVA and follow up between group comparisons using
Scheffe’s procedure were conducted. Table 3 presents a summary of the means and
standard deviations on each of the ECR scales for secure, dismissive, preoccupied and
fearfully categorized participants. These results indicate that the categorical and
continuous measures are generally correspondent providing support for the concurrent
validity of these measures. Secure individuals scored significantly lower on the
avoidance dimension of adult attachment than either dismissive or fearfirl respondents;
yet secure workers were not significantly different from preoccupied individuals on this
dimension, I: (3,111) = 15.20, p < .001. Regarding the anxiety dimension, secure and
dismissive respondents reported significantly lower anxiety than either preoccupied or
fearfirl individuals, F (3,111) = 13.55, p < .001.

Spcial Suppprt was assessed by a set of three scales first developed by Caplan et
al. (1975; See Appendix G). These scales measured respondents’ perception of the level
of emotional and instrumental support that they receive from the following three sources:

their supervisor, co-workers, and family/friends. These scales were scored

49

Table 3

Means and Standard Deviations on ECR Scores for R9 Categories

 

 

RQ Categog Avoidance Anxiety
Secure
Mean 38.93 47.22
SD 14.23 15.52
Dismissive
Mean 53.94 47.51
SD 19.44 15.88
Preoccupied
Mean 54.83 74.67
SD 16.61 17.91
Fearful
Mean 67 .87 71.27
SD 16.75 19.02

 

50

using a 4-point Likert scale with 4 = very much and 1= not at all. A score of 0 was also
included which indicated that the respondent has no such person available to them.
Cross-sectional estimates of reliability of social support from supervisors, co-workers,
and friends and family were, .83, .73, and .81, respectively (Caplan et al., 1975). This
measure, and adaptations of it, have been used extensively by previous investigators
attempting to understand the relationship between stress and social support (F enlason &
Beehr, 1994). In this study a = .88 for supervisor support, or = .74 for coworker support,
and or = .82 for support from others.

The Work Stress Inventory (W SI; Barone, Caddy, Katell, Roselione, & Hamilton,
1988). This 40-item inventory required respondents to rate both the intensity and
frequency of stress at work (See Appendix H). Ratings of stress were made on a 5-point
Likert scale with 0 = none (intensity) or never (frequency) and 4 = very much (intensity)
and “daily” (fi'equency). A Composite Score was also derived by multiplying scores for
intensity by frequency score (Ix F). The following two scales were derived from this
inventory: Organizational Stress and Job Risk. Test-retest reliability on these scales (for
intensity, fi'equency, and composite, respectively) are high, with levels .88, .83, and .84
on Scale 1 (Organizational Stress) and .90, .91 and .90 on Scale 2 (Job Risk) being
reported in the literature. The frequency and composite indexes on the Organizational
Stress Scale have been shown to be moderately correlated in the expected direction with
work satisfaction, anxiety, and organizational commitment (Barone et al., 1988). The
intensity of organizational stress alone was not significantly correlated with any of these
outcomes. For the purpose of this study, composite scores on the Organizational Stress

Scale were used to measure work stress across a wide range of stressors such as job

51

overload, underload, role conflict, role ambiguity, non-participation, and interpersonal
conflicts at work. In this study an alpha level of .89 was obtained for the composite score
on the Organizational Stress Scale.

The Job Satisfaction Scale (J SS; Pond & Geyer, 1987: Lent, 1992) is a modified
version of the general job-satisfaction scale developed by Quinn and Sheppard (1974). It
was originally adapted to written form by Pond and Geyer (1987) and refined by Lent
(1992). This six item measure asked the respondents to rate their level of satisfaction
with their present job (See Appendix I). According to this instrument’s developers, these
items measure “facet-free job satisfaction”, which reflect a worker’s affective reaction to
their job without referring to any specific elements of their work. Items are rated on a 5-
point scale, with l = indicating complete dissatisfaction and 5 = indicating complete
satisfaction. Responses to each item were totaled to form an overall index of
participants’ job satisfaction, with high scores indicating a greater level of satisfaction.
Lent (1992) reported an alpha coefficient of .93 on this measure, which is consistent with
Pond and Geyer’s earlier report of an alpha coefficient of .90 for this scale. In this study,
an alpha coefficient of .89 was obtained for this measure.

The Hopkins Symptom Checklist (HSCL; Derogatis, Lipman, Rickels, Uhlenhuth,
& Covi, 1974) is a 58-item self report measure of psychological symptoms often reported
by outpatients (See Appendix J). Respondents rate their level of distress, during the
preceding seven days, regarding these symptoms. Items are rated on a 4-point Likert
scale with 1 = no distress and 4 = extreme distress. Results of factor analysis of this
instrument revealed the following five underlying symptom dimensions: somatization,

obsessive-compulsive, interpersonal sensitivity, anxiety and depression. Alpha

52

coefficients of these factors range from .84 (anxiety) to .87 (somatization and obsessive-
compulsive). Test-retest reliability ranged from .75 (anxiety) to .84 (obsessive-
compulsive) in evaluations completed one week apart. Derogatis et al. (1974) reported
that the HSCL has demonstrated sensitivity to low levels of symptoms and to changes in
emotional status among non-psychiatric outpatients.

In this study, intercorrelations among subscales ranged from r = .51, p < .01
(somatization and interpersonal sensitivity) to r = .74, p < .01 (anxious and obsessive-
compulsive). Research has consistently revealed high intercorrelations among the
subscales on the HSCL; therefore, subscale ratings were summed to obtain a total score
that assessed the overall level of psychological distress reported by the respondent
(Folkman, Lazarus, Gruen, & DeLongis, 1986). An alpha of .96 was obtained on the
total score indicating the participants’ overall level of psychological distress.

Research Hypotheses

As in previous research, the results of data analysis are expected to confirm a
relationship between work stress and indexes of strain (job satisfaction and symptomatic
distress). Social support is also expected to be related to these indexes of strain.
Furthermore, it is predicted that social support will moderate the relationship between
work stress and indexes of strain. The results of data analysis are also expected to reveal
that adult attachment style is significantly related to a worker’s level of stress, perception
of support, and to indexes of work strain. This research will also attempt to explain
previously discrepant findings regarding the buffering effect of social support on strain
under conditions of high stress through the exploration of adult attachment style as

moderating variable.

53

Data Analysis

The Statistics Package for Social Sciences (SPSS, Version 7.5.1) was used to
conduct data analyses. Descriptive sample statistics (means, standard deviations, and
ranges) were calculated and examined for all variables in the study. An intercorrelation
matrix was calculated for the demographic variables and the key measures of interest.

Correlation analysis was used to examine the relationship between work stress, as
measured by the WSI (Organizational Stress Composite Score), and the following two
variables: 1) job satisfaction, as measured by the J SS, and 2) symptomatic distress, as
measured by the HSCL (Hypotheses 1a and 1b). Correlation analysis was also used to
examine the relationship between overall social support, as measured by the Caplan et al.
(1975) social support scale, and the following two indexes of work strain: 1) job
satisfaction (J SS), and symptomatic distress (HSCL). (Hypotheses 2a and 2b).

A Fisher r to z transformation, and a subsequent z-test of the transformed r values,
was used to examine the between group differences in worker job satisfaction among
various types of social support (i.e, supervisor support, co-worker support, and
family/friend support). (Hypothesis 3a).

Multiple regression was used to examine the relationship(s) of social support and
work stress to indexes of strain and thereby test hypotheses 4a and 4b. To provide for a
more sensitive test to the interaction effects, raw scores were converted to z-scores, as
recommended by Holmbeck (1997). Two separate regression models examined the
effects of a) work stress (W SI) and supervisor support (Caplan et al, 1975) on job
satisfaction (J SS); and b) work stress (W SI) and supervisor support (Caplan et al, 1975)

on symptomatic distress (HSCL). Specifically, work stress was entered first into the

54

regression equation, followed by the social support variable, and the last step was to enter
the interaction of supervisor support and stress.

To test Hypotheses 5a, 6a, 7a, and 7b, a one-way MANOVA was conducted. In
this analysis, the independent variable was attachment style and the dependent variables
were work stress, supervisor support, symptomatic distress, and job satisfaction. As the
multivariate F was significant, follow-up univariate tests of attachment style on each
dependent variable were conducted. Planned contrasts were used to examine between-
group differences.

Once raw scores on predictor variables were converted to z-scores, hierarchical
regression was used to test Hypotheses 8a and 8b. Specifically, the main effects of work
stress and supervisor support were controlled in step 1 of these analyses. The main
effects of the adult attachment dimensions of avoidance and anxiety on symptomatic
distress and job satisfaction were tested in step 2. Finally, step 3 tested the two way
interactions of work stress and avoidance and anxiety; the two way interactions of
supervisor support and avoidance and anxiety; the three way interactions of work stress,
supervisor support and avoidance; and the three way interactions of work stress,

supervisor support and anxiety on symptomatic distress and job satisfaction.

55

CHAPTER IV

RESULTS

This chapter details the results of data analyses. The treatment of missing data and
a summary of descriptive statistics for the sample are presented in the initial sections of
this chapter. Next, the correlational findings are presented. Following this are the results
pertaining to social support as a moderator of the work stress and strain relationship, as
well as those relevant to the interrelationships among attachment style, work stress, social
support and indexes of strain. Finally, post hoc analyses suggested by the previous
findings are presented.
I_re_atment of MissingDat_a

All variables were examined for possible data entry errors and missing values
prior to data analysis. Mean substitution was used to replace 13 missing values on the
WSI and 4 missing values on the HSCL. Since a “neutr ” rating was included in the
ECR, this value was substituted for 4 missing data points on the ECR.

On the RQ, a number of participants did not respond to the continuously
measured portion of this instrument. Therefore, this portion of the RQ was only used to
derive the attachment style for those respondents who did not check one of the categories
of attachment style as most descriptive of them (i.e., the item rated as most highly
descriptive of the respondent was used to assign their attachment style). For two

participants, a “most descriptive” attachment style could not be assigned because the

56

participant gave no response to either part of the RQ or gave equivalent ratings on the
continuous portion of this instrument. The data for these individuals were automatically
deleted from further analyses using the RQ. Participants providing incomplete data on
the J SS (n=1) or the various subscales of the SSI (n=2 supervisor support; n=4 coworker
support; n=5 other support) were also dropped from analyses using those subscales, due
to the small number of items on each of these measures.

Also, prior to conducting further analysis, the data were examined for outliers by
plotting all instruments against the attachment style categories. One outlier with regard
to WSI scores was detected and deleted fiom firrther analysis using this measure.
Descriptive Statistics

Table 4 contains the mean, standard deviation, skewness, and range of the
continuously scored variables in this study. The descriptive statistics for work stress are
most similar to those found by Barone, Caddy, Katell, Roselinoe, and Hamilton (1998)
for police officers, hospital nurses and women in general. The HSCL mean and standard
deviation were similar to those found by Behrens (1998) in a study employing the HSCL
as a global measure of symptomatic distress.

Supervisor support in this sample was higher than recently reported in the
literature. Using the same measure of social support as in this study, Fenlason and Beehr
(1994) separated support into instrumental and emotional types and reported means of
3.64 and 3.54, respectively. The findings of this Study, regarding social support separated
into the same categories, revealed a mean of 5.53 for instrumental support and 5.65 for

emotional support.

57

Workers in this study also reported somewhat higher job satisfaction than in a
recent study using the J SS. Lent (1992) reported an item mean on the J SS of 3.39 in her
study of 168 adult workers. When the J SS data from this study was transformed, an item
mean of 3.79 was obtained.

Regarding the measurement of adult attachment dimensions using the ECR,

K. A. Brennan (January 22, 1999, personal communication) reported item means for the
anxiety and avoidance scales as 3.46 and 2.93, respectively. To obtain item means, after
re-coding the appropriate items and summing across scale items, the total is divided by
the number of items (18) on each scale. Once the data from this study were transformed
in the above manner, the item mean for the anxiety scale was 2.88 and the avoidance
scale was 2.67.

Table 5 describes the distribution of participants’ attachment style classification.
Consistent with recent studies including post-college age respondents, few preoccupied
participants were identified in this study (5%). Among 45-54 year olds in a nationally
representative sample, 8% endorsed preoccupied attachment styles as characteristic of
them, 64% classified themselves as secure, 23% reported avoidant attachment styles, and
5% of respondents were unclassifiable (Mickelson, Kessler, & Shaver, 1997).
Furthermore, at age 52 only 5% of women, participating in a longitudinal study of
personality characteristics and future plans, identified themselves as preoccupied
(Klohnen & Bera, 1998).

Recomposition of R0 Classification
Theoretically, both preoccupied and fearful individuals have negative self models

and are disposed toward high levels of attachment-related anxiety. These individuals are

58

Table 4

Descriptive Statistics for Continuously Scored Maasures

 

 

Variable Name M S_D S_K Ra__r_rga
Work Stress Inventory 82.01 47.29 .90 2 - 248
Hopkins Symptom 85.02 20.32 1.126 58 - 158
Checklist

Job Satisfaction Survey 18.93 4.41 -.830 7 - 25
Supervisor Support 11.21 3.67 -.540 0 — 16
Coworker Support 12.11 2.44 -.584 4 - 16
Support from Others 13.82 2.47 -1.23 5 — l6
Avoidance 48.11 19.07 .47 18 - 98
Anxiety 51.90 18.56 .651 19 - 100

 

59

also thought to share the characteristics of hypervigilance and hyperarousal in regard to
threatening or stressfirl events (Lopez, 1995). Since a relatively small number of
respondents endorsed preoccupied and fearfirl styles in this study, these categories were
collapsed to form an “anxious” category to facilitate data analyses. The frequencies and
percentages of attachment styles produced by this recoding of the RQ can be found in
Table 6.
Correlational Findings

Table 7 presents an intercorrelation matrix for the demographic variables and key
measures of interest in this study. No significant correlations were observed among
measures of work strain and the key demographic variables “length of service” and “sex”.
However, “length of service” was moderately correlated with “avoidance” (r = .26,
p < .05) indicating that workers reporting longer tenure in their present jobs expressed
higher levels of attachment-related avoidance.

Examination of the correlational data provides support for hypotheses la and 1b.
Work stress was significantly negatively correlated with job satisfaction ( r = -.44,
p < .01) indicating that workers who reported higher work stress acknowledged lower job
satisfaction. Also, work stress was significantly positively correlated with symptomatic
distress with r = .65, p < .01, demonstrating that workers reporting high levels of work
stress experienced higher levels of symptomatic distress.

Supervisor support was significantly positively correlated with job satisfaction
(r = .49, p < .01). This confirms hypothesis 2a in this study. In addition, supervisor

support was significantly negatively correlated with symptomatic distress (r = -.31,

60

Table 5

Fre_quencies and Percentages of Attachment Sgles

 

 

 

Attachment Style N %
Secure 59 5 1 .3
Dismissive 35 30.4
Preoccupied 6 5.2
Fearful 15 13.0
Total 1 1 5 99.9%

 

Note. Percentage does not total 100.00% due to rounding error.

Table 6

Frequean and Percentages of Recoded Attachment Styles

 

 

 

Attachment Style N %
Secure 59 5 1 .3
Dismissive 35 30.4
Anxious 21 1 8.3
Total 1 1 5 100%

 

61

Hugo q

5830:0520: 0m $322

_. >mo --
5. max .5 I-

u. macoemos -. _m .5... I-

a. r832. an... .3 -8...
maniac

M. £9: -55... no.2. .: JNM --
madam

0. 3588.8 .00 Amos... -. 3 .55... .5 --

 

4. >33 1... -8 .8 -8 8...... we...
m. 960232 .3 )8 JG: .8 33...... :5 now :-
mauve...
0. 6020184 Joo A: 3?. A: -53.... JOA now .51. :-
mauve:
_o. 0:82 .5 .8 be -5»... ANA... -05.... AS... .5 .51.. :-
mauve...
Z. Ema—r -.: .om wow .5 .31. but. .3... 33...... Jugs... -51.... --
5. Hmm .3 a: -.8 .S 33...... .3 A: be: .3: .3 n51. :-
inanimmo: a wmmamoea a :5 PS _a<o_ 5-3563. .. 0026530: mm 36.58:. a =5 98 _98. 5-653.

Ooh—3m” max 2 u mags—o. 5 u 7.35“ masomao: C n Em: worse—Emu. a N 98888 Coma.» q u 0283
Enema. e». wagon C n _amw as: _ <8: G u o<2 no <83. Eme- n mow—“Sm $5683 9.85:3 amw u .56 mmammmosos 9.28.

62

p < .01), confirming that those individuals reporting higher levels of supervisor support
reported less symptomatic distress (Hypothesis 2b).

A Fisher r to z transformation was conducted in order to examine group
differences in the relationship between different sources of social support and worker job
satisfaction. There were no significant differences found between the effect of supervisor
support as compared to coworkers support on job satisfaction (2 =.63, p = .26). However,
supervisor support was more highly related to job satisfaction than was support from
family/friends (z = 3.31, p <.Ol), confirming hypothesis 3a. Co-worker support was also
found to be significantly more related to job satisfaction than support from family and
friends (2 = 2.70, p < .01). These data indicate that on-the-job support is more highly
related to job satisfaction than is support from family and fiiends.

Supervisor Support as a Moderator of the WflStress-Strgin Relationship

Prior to examining the respective unique and interactive contributions of work
stress and supervisor support to the two indexes of work-related strain, the data were
examined regarding the necessary assumptions related to regression. The only violation
of these assumptions was with regard to normality. A histogram of the WSI exposed a
positive skew, whereas histograms of the J SS and the social support variable revealed
negatively skewed distributions. However, regression analysis is robust against
violations of normality when sample sizes are large as in this study (Berry & Feldman,
1995). The assumptions of pairwise linearity and homoscedasticity were satisfied by
examining bivariate scatterplots of the variables. Muticollinearity was evaluated through

the collinearity diagnostics of SPSS. All tolerance proportions were found to be above

63

0.1, indicating that the independent variables are not subject to the problem of
multicollinearity.

Table 8 presents the results of hierarchical regression analyses testing the direct
and interactive contributions of work stress and supervisor support to the two indexes of
strain. Both supervisor support and work stress demonstrated main effects on the
prediction of job satisfaction with work stress accounting for 20% of the variation on this
measure, and with supervisor support explaining an additional 10% of the variance. The
interaction of supervisor support and work stress did not incrementally enhance the
prediction of job satisfaction. VVrth regard to the prediction of symptomatic distress, only
work stress made a significant contribution; neither supervisor support nor its interaction
with work stress significantly improved the prediction of symptomatic distress.
Therefore, contrary to hypotheses 4a and 4b, when the main effects of work stress and
supervisor support were controlled, the interaction of work stress and supervisor support
did not significantly enhance the prediction of either job satisfaction or symptomatic

distress.

Relationships of Adult Attachment Styles to Work Stress, Supervisor Supmrt, and

 

Indexe; of Strapip

Significant relationships were hypothesized between adult attachment style and
work stress (hypothesis 5a) and perceptions of supervisor support (hypothesis 6a). It was
also hypothesized that worker’s attachment style would be related to job satisfaction
(hypothesis 7a) and symptomatic distress (hypothesis 7b). A MANOVA was used as an

initial test of these hypotheses.

Table 8

Hierarchical Reggession Anplyses for Work Stress (WS), Supervisor Support (SS), and
their Interaction in Predicting Job Satisfaction and Symptomatic Distress

 

Job Satisfaction

 

Unstandardized SE B Standardized t p

 

 

 

[3 beta
Step 1 Work Stress (WS) -2.0 .375 -.451 -5.32 .000
Step 2 Supervisor 1.62 .405 .358 3.99 .000
Support (SS)
Step 3 WS X SS .37 .347 .090 1.08 .283

 

Note. R2 = .20 for Step 1; A R2 = .10 for Step 2 (p < .001); A R2 = .01 for Step 3 (ns).

Symptomatic Distress

 

Unstandardized SE B Standardized t p

 

 

 

[3 beta
Step 1 Work Stress (WS) 13.02 1.45 .647 8.99 .000
Step 2 Supervisor .13 1.66 .006 .08 .939
Support (SS)
Step 3 WS X SS 1.04 1.43 .055 .72 .470

 

Note. R2 = .42 for Step 1; A R2 = .00 for Step 2 (ns); A R2 = .00 for Step 3 (ns).

65

All assumptions pertinent to MANOVA were tested. No outliers were found
using the Mahalanobis distance with p < .001. As stated previously, one outlier was
discovered using a boxplot of work stress and attachment style. This data point was
removed from analyses.

Histograms were used to test the normality hypothesis. A histogram of the HSCL
exposed a moderate positive skew and a histogram of the supervisor support variable
revealed a moderate negative skew. However, since MAN OVA analyses are robust to
violations of normality when sample sizes are large and when the violation is created by
skew rather than outliers, these data were not transformed (Tabachnick & Fidell, 1996).
It is also noteworthy that the HSCL positive skew indicates that the sample reported few
psychological symptoms of distress and that the negative skew of supervisor support
ratings indicates that relatively high supervisor support was reported by these
respondents. All other variables were found to have fairly normal distributions. Using
bivariate scatterplots of all combinations of the dependent variables, the assumption of
pairwise linearity was satisfied. Multicollinearity was evaluated through the collinearity
diagnostics of SPSS. All conditioning indexes were satisfactory. The homogeneity of
the variance-covariance matrices were assessed with Box’s M. The results of this test
produced strong evidence for the homogeneity of these variances, E (20, 15511) = .765, p
= .759. This outcome means that the variance-covariance matrices can be combined to
create a single estimate of error without distorting alpha levels (Tabachnick & Fidell,
1996). Since Box’s M is considered a “notoriously sensitive test of homogeneity of
variance-covariance matrices”; it is reasonably certain that the significance tests were

robust, despite unequal cell sizes in MANOVA (Tabachnick & F idell, 1996, p. 382).

66

A oneway MAN OVA, with attachment style as the predictor, was performed on
four dependent variables: symptomatic distress, job satisfaction, supervisor support and
work stress. An alpha level of .05 was selected. Table 9 presents the means and standard
deviations of the three attachment style groups on the measures of work stress,
symptomatic distress, job satisfaction and supervisor support.

A significant multivariate F for workers attachment style was observed, Wilks’

F (8,210) = 3.62, p < .001. Follow up univariate F tests revealed a significant
relationships between attachment style and symptomatic distress [F (2,108) = 12.10,

p < .001], work stress [F (2,108) = 5.70, p < .01], and supervisor support, [F (2,108) =
3.42, p < .05]. However, contrary to hypothesis 7a , no significant relationship was found
between attachment style and job satisfaction, F (2,108) = .76, p = .472.

Due to the combination of preoccupied and fearful attachment style groups,
secure participants were compared to the “anxious” attachment group in planned
contrasts used to test hypotheses 5a, 6a, 7a, and 7b. Given the discrepancy in cell sizes,
Levene tests for equality of variance were executed. The results of these tests indicated
that the null hypothesis could not be rejected; therefore; the variances for each group
were assumed to be equal (Norusis, 1997).

Univariate analysis of variance was used to test differences between secure and
anxious attachment style groups on work stress, supervisor support, job satisfaction and
symptomatic distress. Contrast coefficients of 1, 0, -1 were assigned to the secure,
dismissive, and anxious groups respective. Results of these ANOVA tests revealed that
secure workers reported significantly less work stress (p < .01), more supervisor support

(p < .05), and less symptomatic distress (p < .01) than did anxious workers, thereby

67

Table 9

Means and SD of Attachment Groups on Work Stress, Symptomatic Distresp. Job

Satisfaction, and Supervisor Support

 

 

Variable Name Mean S_D
Work Stress Inventory
Secure 76.20 43.51
Dismissive 74. 89 38.18
Anxious 113.61 59.24
Hopkins Symptom Checklist
Secure 80.37 17.25
Dismissive 82.46 18.73
Anxious 103.14 22.43
Job Satisfaction Survey
Secure 19.29 4.35
Dismissive 18.83 3.91
Anxious 17.86 5.44
Supervisor Support
Secure 12.09 3.50
Dismissive 10.26 3.65
Anxious 10.05 3.67

 

68

supporting hypotheses 5a, 6a, and 7b, respectively. No significant between group
attachment style differences were discovered in the rate of job satisfaction between
secure and anxious groups, disconfirming hypothesis 7a.
Contributions of Work Stress, Supervisor Supppg, and Adult Attachment Orientations to
Indexes of Work-Related Strai_n

Hierarchical regression analyses employing the dimensional indexes of adult
attachment was used to gain a better understanding of the relationship between adult
attachment, work stress, social support, job satisfaction and symptomatic distress. The
results presented in Table 10 demonstrate that, when work stress and supervisor support
were controlled, both adult attachment dimensions significantly enhanced the prediction
of symptoms, but not job satisfaction. Adult attachment indexes did not significantly
interact with work stress or supervisor support to predict job satisfaction or symptoms.
However, it should be noted that more avoidant workers reported greater job satisfaction
than less avoidant workers, and that the interaction of supervisor support with each adult
attachment index appears to have some relationship to job satisfaction.
Post H_oc mm

Since attachment style related behaviors are thought to be especially activated in
times of distress, removing the influence of stress in our initial regression equations (by
controlling for work stress in the first step) may have prevented the detection of
attachment style contributions to job satisfaction. To explore this possibility, a post-hoc
regression of job satisfaction that included only supervisor support and the two
attachment indexes was conducted. In this regression (see Table 11), attachment indexes

marginally enhanced the prediction of job satisfaction (R2 change = .04, p < .06), after

69

Table 10

Hierarchical Regression Analyses for Work Stress (W8 1, Supervisor Supmrt (SS 1,
Attachment Dimensions and their Interaction in Predicting Symptomatic Distress and Job

 

 

 

 

 

 

 

 

 

 

Satisfaction
Symptomatic Distress
Unstandardized SE [3 Standardized t p
[3 beta
Step Work Stress (W S) 13.08 1.64 .650 8.00 .000
1 Supervisor . l3 1 .66 .006 .76 .940
Support (SS)
Step Avoidance 3.76 1.47 .187 2.55 .012
2 Anxiety 3.01 1.54 .150 1.96 .053
Step WS X Avoidance —1.13 1.91 -.055 -.59 .556
3 WS X Anxiety 2.10 1.73 .117 1.21 .229
SS X Avoidance -4.40 2.25 -.002 -.02 .984
SS X Anxiety .56 1.83 .029 -.31 .758
WS X SS X Avoidance .64 1.78 .035 .36 .720
WS X SS X Anxiety .23 1.48 -.016 .15 .879
Note. R2 = .42 for Step 1; A R2 = .07 for Step 2 (p < .001); A R2 = .01 for Step 3 (ns).
mm
Unstandardized SE B Standardized t p
(3 beta
Step Work Stress (W S) -1.26 .397 -.285 -3. 18 .002
1 Supervisor 1 .62 .405 .3 58 3 .99 .000
Support (SS)
Step Avoidance .73 .375 .165 1.95 .054
2 Anxiety -.46 .391 -.103 -1.17 .245
Step WS X Avoidance -.31 .469 -.07 -.66 .509
3 WS X Anxiety .57 .426 .14 1.33 .187
SS X Avoidance -1.25 .555 -2.26 -2.26 .026
SS X Anxiety 1.02 .450 2.24 2.27 .025
WS X SS X Avoidance .30 .438 .07 .68 .495
WS X SS X Anxiety -.37 .364 -.12 -1.01 .317
Note. R2 = .30 for Step 1; A R2 = .03 for Step 2 (ns); A R2 = .06 for Step 3 (ns).

70

supervisor support was controlled. In addition, attachment indexes significantly
interacted with supervisor support levels to predict satisfaction (R2 change = .05, p < .03).
Workers reporting higher avoidance reported significantly greater job satisfaction under
low support conditions (Figure 1). Figure 2 describes the relationship between supervisor
support, anxiety and job satisfaction. Those individuals who reported lower anxiety
endorsed significantly greater job satisfaction under high support conditions than did
those respondents who reported higher anxiety under the same high support conditions.

Several post hoc analyses were conducted to investigate possible explanations for
unexpected findings (e. g., the lack of a moderating effect of supervisor support on the
work stress-strain relationship). Post hoc analyses also further explored the relationship
of workers’ adult attachment styles to these variables.

In the original hypotheses, work stress was conceptualized as an interactional
construct involving a worker’s appraisal regarding the frequency that work demands
exceed his/her capability and the intensity of distress that is created by this discrepancy.
To measure this construct, the composite score on the Organizational Stress Scale of the
Work Stress Inventory was used as the work stress variable. This score is created by
multiplying respondents’ ratings of the intensity (I) of stress that would be caused if they
experienced various problems at work, by a rating of the frequency (F) which these
events are actually experienced on their current job. This is a much more global measure
of work stress than those used in the studies that have found a buffering effect of social
support on the relationship between work stress and strains. Also note that a “0” rating
on either the intensity or frequency scale cancels out the rating of the other in the

composite score. Given the absence of an expected “buffering effect” of social support in

71

 

Table 11

Hierarchical Regiession Analyses for Supervisor Supmrt and Attachment Dimensions

and their Interaction in Predicting Job Satisfaction

Mam

 

 

 

 

Unstandardized SE B Standardized t p
(3 beta

Step Supervisor Support 2.17 .366 .49 5.93 .000
1 (SS)

Step Avoidance .78 .3 85 . 18 2.04 .044
2 Anxiety -.76 .381 -.17 -1.99 .049
Step SS X Avoidance -1.20 .485 -2.48 -2.48 .015
3 SS X Anxiety .87 .385 2.26 2.26 .026

 

Note. R2 = .24 for Step 1; A R2 = .04 for Step 2 (p < .06); A R2 = .05 for Step 3 (p < .03).

72

F igare 1. The Interaction of Avoidance and Supervisor Support in Predicting Job

Satisfaction.

 

22

21I

20‘

191

 

 

 

 

c:

a

0 18:

c9.

.9 .

‘8 Avondance

m

.0 171

o

-- L0

2 15 Hi
Lo Hi

Support

73

Figure 2. The Interaction of Anxiety and Supervisor Support in Predicting Job

Satisfaction.

Mean Job Satisfaction

16

 

 

 

 

 

74

Hi

Anxiety

Lo

Hi

our original findings, this series of post hoc tests was designed to separate respondents’
ratings of organizational stress intensity (081) and organizational stress frequency (OSF)
to determine if supervisor support interacted differently with either of these ratings to
predict job satisfaction and symptomatic distress.

Results of these regression analyses indicated that OSI interacted with supervisor
support to significantly increase the prediction of job satisfaction (see Table 12) The
nature of this interaction effect was explored and summarized in Figure 3. Those
individuals who reported high stress intensity and high support reported greater job
satisfaction than those reporting lower stress intensity under the same high support
conditions. However, under low support conditions, those reporting lower stress intensity
endorsed higher job satisfaction than those reporting higher stress intensity.
Hierarchical regression was also used to explore the possible interrelationships of OSI,
supervisor support and adult attachment indexes on job satisfaction and symptomatic
distress. Results revealed no new interaction effects; however, avoidance scores and the
interaction of 081 and supervisor support scores were significantly predictive of job
satisfaction (see Table 13). Furthermore, OSI, avoidance, and anxiety scores were
predictive of symptomatic distress. When the OSF score was substituted as the work
stress variable, OSF, supervisor support and avoidance added to the prediction of job
satisfaction (see Table 14), whereas OSF and anxiety were predictive of symptomatic
distress (see Table 15).
Samar; of Findinga

Workers reporting high levels of work stress also reported higher levels of

symptomatic distress and lower job satisfaction than workers reporting less work stress.

75

Table 12

Hierarchical Regression Analyses for Organizational Stress Intensit OSI Su ervisor

Support (SS) and their Interaction in Predictig Job Satisfactipn and Symptomatic

 

 

 

 

 

Distress
Job Satisfaction
Unstandardized SE B Standardized t p
B beta
Step Organizational Stress -.418 .414 -.095 -l.01 .315
1 Intensity (081)
Step Supervisor Support 2.16 .372 .485 5.80 .000
2 (SS)
Step OSIXSS .90 .331 .226 2.70 .008
3

 

Note. R2 = .01 for Step 1; A R2 = .23 for Step 2 (p < .001); A R2 = .05 for Step 3
(p< .01).

Symptomatic Distress

 

Unstandardized SE B Standardized t p

 

 

 

B beta
Step Organizational Stress 6.35 1.81 .313 3.51 .001
1 Intensity (081)
Step Supervisor Support -5.52 1.77 -.271 -3.11 .002
2 (SS)
Step 081 X SS -1.32 1.63 -.072 -.81 .422
3

 

Note. R2 = .01 for Step 1; A R2 = .07 for Step 2 (p < .01); A R2 = .05 for Step 3 (ns ).

76

Table 13

 

Hierarchical Regession Analyses for OSI, Supervisor Suppprt (SS), Attachment
Dimensions 33! their Interaction in Predictirg Job Satisfaction and Symptomatic Distress

191mm

 

 

 

 

Unstandardized SE B Standardized t p
B Beta
Step Organizational Stress
1 Intensity (081) -.37 .359 -.084 -1.00 .321
Supervisor Support (SS) 2.15 .362 .484 5.94 .000
051 X SS .90 .331 .226 2.70 .008
Step Avoidance .80 .381 .182 2.10 .038
2 Anxiety -.71 .387 -. 162 -1.83 .069
Step 081 X Avoidance -.33 .437 -.079 -.76 .446
3 OSI X Anxiety .43 .426 .101 1.00 .319
SS X Avoidance -1.14 .536 -.218 -2. 12 .037
SS X Anxiety .44 .506 .105 .86 .391
081 X SS X Avoidance .16 .503 .035 .33 .745
081 X SS X Anxiety 6.45 .442 .018 .15 .884

 

Note. R2 = .29 for Step 1; A R2 = .04 for Step 2 (ns, p = .063); A R2 = .04 for Step 3 (ns).

Symptomatic Distress

 

 

 

 

Unstandardized SE B Standardized t p
B Beta
Step Organizational Stress
1 Intensity (081) 5.50 1.77 .272 3.12 .002
Supervisor Support (SS) -5.52 1.77 -.271 -3.11 .002
Step Avoidance 4.16 1.75 .205 2.38 .019
2 Anxiety 5.26 1.77 .260 2.97 .004
Step 081 X Avoidance 1.37 1.92 .070 .71 .447
3 081 X Anxiety .61 1.91 .031 .32 .752
SS X Avoidance 1.58 2.42 .066 .65 .514
SS X Anxiety -2.03 1.95 -.106 -1.04 .300
081 X SS X Avoidance -.80 2.31 -.037 -.35 .730
081 X SS X Anxiety -1.68 1.95 -.103 -.86 .391

 

Note. R2 = .17 for Step 1; AR2 = .15 for Step 2 (p < .001); AR: = .03 for Step 3 (ns).

77

Table 14

Hierarchical Regression Analyses for Or anizational Stress Fre uenc OSF Su ervisor

 

Support (SS), Attachment Dimensions and their Interaction in Predicting Job Satisfaction

Job Satisfaction

 

Unstandardized SE B Standardized t p

 

 

 

B beta
Step Organizational Stress
1 Frequency (OSF) -1.32 .424 -.298 -3.12 .002
Supervisor Support (SS) 1.47 .427 .330 3.45 .001
Step Avoidance .84 .376 .189 2.22 .028
2 Anxiety -.54 .379 -.122 -l.42 .160
Step OSF X Avoidance -.36 .511 -.08 -.70 .488
3 OSF X Anxiety .74 .477 .19 1.56 .123
SS X Avoidance -1.11 .555 -.21 -2.01 .047
SS X Anxiety 1.13 .481 .27 2.35 .021
OSF X SS X Avoidance .15 .461 .04 .32 .750
OSF X SS X Anxiety -.46 364 -.15 -1.26 .210

 

111.03.. R2= .30 for Step 1; AR2 =

.03 for Step 2 (ns); A R2 = .06 for Step 3 (ns).

78

Table 15

Hierarchical Reggession Analyses for Organizational Stress Frequency (OSF). Supervisor

Support (SS), Attachment Dimensiorflnd their Interaction in Predicting Symptomatic

 

 

 

 

 

 

Distress
Symptomatic Distress
Unstandardized SE B Standardized t p
B beta
Step Organizational Stress
1 Frequency (OSF) 13.16 1.77 .647 7.46 .000
Supervisor Support (SS) .53 1.78 .026 .30 .767
Step Avoidance 2.79 1.52 .137 1.84 .069
2 Anxiety 4.51 1.53 .223 2.95 .004
Step OSF X Avoidance -l .82 2.12 -.086 -.86 .393
3 OSF X Anxiety 3.48 1.98 .195 1.76 .082
SS X Avoidance -1.20 2.30 -.050 -.52 .603
SS X Anxiety 1.27 1.99 .066 .64 .526
OSF X SS X Avoidance .60 1.91 .033 .31 .755
OSF X SS X Anxiety -.29 1.51 -.021 -.19 .847
Note. R2 = .40 for Step 1; A R2 = 08 for Step 2 (p < .001); A R2 = .02 for Step 3 (ns).

79

Figure 3. The Interaction of Supervisor Support and Organizational Stress Intensity in

Predicting Job Satisfaction.

 

23
22 a
21 I
20 ‘

19T

 

 

 

 

:
-8
8
f2 18 I
a Stress Intensity
_o __
.2 17 I LO
g 16 Hi

Lo Hi

Support

80

When the effects of social support on job satisfaction were explored, on-the-job support
(by supervisors and co-workers) was more strongly related to job satisfaction than was
support from family and fiiends. Those workers who described greater supervisor
support also reported lower levels of symptomatic distress than did workers reporting
lower supervisor support. No evidence was initially found for the buffering effect of
supervisor support on the work stress-strain relationship; however when work stress was
redefined as the perception of the intensity of potential work stressors, supervisor support
did moderate the relationship between stress and job satisfaction (see Figure 3).

Adult attachment style was found to be significantly related to work stress,
symptomatic distress and supervisor support. Tests of planned contrasts between
attachment groups revealed that secure workers reported significantly less work stress
and less symptomatic distress than did anxious workers. Secure workers also reported
more supervisor support than did anxious workers. No significant differences were found
between secure and anxious attachment styles groups with regard to job satisfaction.
However, post hoc analyses showed that the interaction of attachment style dimensions
and supervisor support did significantly affect the prediction of job satisfaction, even

after the main effect of supervisor support was controlled (See Figures 1 and 2).

81

CHAPTER V

DISCUSSION

This study sought to examine the contributions of work-related stressors, social
support and indexes of adult attachment to job satisfaction and psychological firnctioning
within an adult worker sample. Special consideration was given to exploring adult
attachment style as a moderator of the stressor-strain and social support-strain
relationships. In addition, the direct contribution of attachment styles to indexes of work-
related strains was explored.

In this chapter, conclusions regarding the interrelationships of work stress, social
support, and strains will be discussed initially. Second, the importance of the findings of
this study in understanding the buffering effect will be considered. Third, conclusions
regarding the relationships of adult attachment orientation to work stress, job satisfaction,
symptomatic distress and social support are presented. Fourth, the limitations of this
study are elaborated. Finally, the implications of the results of the present study to
counseling theory, practice, and research are discussed.

The Relationships of Wofirk Stress, Supervisor Supmrt, and Strains

Consistent with previous findings, results revealed that workers reporting high

levels of work stress also reported higher levels of symptomatic distress and lower job

satisfaction than did workers reporting less work stress. In addition, workers reporting

82

high levels of supervisor support reported higher job satisfaction and less symptomatic
distress than did workers acknowledging low levels of supervisor support.

Similar to the findings of Fenlason and Beehr (1994), support received at the
workplace (i.e., from supervisors and co-workers) was found to be more highly related to
job satisfaction than was support from family and friends. Furthermore, as expected in
this study, supervisor support remained predictive of job satisfaction even after work
stress was controlled.

Ms Regarding the BufferingEffect of Social Support

Contrary to expectations, when controlling for the main effects of work stress and
supervisor support, the interaction of supervisor support and work stress did not
significantly enhanced the prediction of symptomatic distress or of job satisfaction.
Therefore, no evidence was initially found for the buffering effect of social support on the
relationship between work stress and strain.

These results may indicate that work stress is a much more eflicient predictor of
symptomatic distress, and that it may subsume the contribution of supervisor support to
this measure. Alternatively, it is possible that the negatively skewed distribution of
supervisor support ratings in our sample (i.e., most participants reported high levels of
support) may have diminished the ability of this measure to predict unique effects
regarding symptomatic distress when stress levels were controlled.

Due to the unexpected nature of the above findings regarding job satisfaction,
post hoc testing explored alternative operationalizations of work stress in terms of the
respondent’s perception of the intensity of various stressors versus the frequency of actual

stressful events experienced on the job. Work stress was originally conceptualized as a

83

product of worker’s ratings of the intensity of potential stressors by the frequency that
these stressors occurred (Ix F). In the revised models, the intensity and frequency scores
were used separately to explore the unique contribution of each of these elements of work
stress to the prediction of strains.

In these post hoc tests, occupational stress intensity (081) alone was not found to
be significantly predictive of job satisfaction; however, supervisor support was found to
moderate the relationship between worker’s perceptions of the intensity of potential
stressors and job satisfaction. Those who perceived stressors as more intense reported
less job satisfaction under conditions of low supervisor support than those who perceived
these stressors as less intense. However, those who perceived stressors as more intense
and who received high supervisor support endorsed more job satisfaction than those who
perceived stressors as less intense under the same high support conditions. These
findings indicate that workers who perceive stressors intensely respond very well to high
levels of supervisor support and perceive their jobs as more satisfying than workers who
perceive stressors as less intense.

Unlike occupational stress intensity scores, occupational stress frequency (OSF)
scores were predictive of job satisfaction but did not significantly interact with supervisor
support to enhance the prediction job satisfaction. Therefore, it appears that, to
effectively help workers manage their stress on the job, supervisors need to be sensitive
not just to the amount or frequency of work stressors but more importantly their worker’s
perception of these stressors.

These findings are consistent with the theory that individual

differences/personality variables affect the way that an individual may create or use

84

social support (Sarason et al., 1996). One advantage of this study over previous
investigations into the buffering effect was the use of a psychometrically sound
instrument to measure work stress capable of partialing out individuals’ perceptual
differences regarding stressors fi'om the frequency of actual stressors. Since past research
has not made this distinction, workers’ negative affectivity (captured here in stressor
intensity ratings) may have differentially affected the relationship of social support and
stress to job satisfaction, and thus may have contributed to some of the inconsistent
findings regarding the buffering effect in this literature.

Both occupational stress intensity and fi'equency were predictive of psychological

 

distress in our sample; however, no interaction effects with supervisor support were
detected. Therefore, the breakdown of work stress into the above two components did
not enhance our understanding of the relationship between work stress and symptomatic
distress. This finding is contrary to Beehr’s (1985) assertion that psychological distress
appears to be more amenable to the moderating effect of social support on the stressor-
strain relationship.

Negative affectivity, defined as a stable interpersonal trait encompassing such
characteristics as trait anxiety, depression, and low self esteem has been explored by
previous researchers interested in the relationships between work stress and strain
(Decker & Borgen ,1993; Schonfeld, 1996). Decker and Borgen (1993) explored various
dimensions of work appraisal and the contribution of negative affectivity to work stress,
strain, coping and job satisfaction. These authors found that negative affectivity has
some effect on self-reports of psychological and physical strain; however it does not alter

the relationship between occupational stress and these strains. Furthermore, negative

85

affectivity had no significant effect on measures of job satisfaction. The findings of the
present study seem to be consistent with Decker and Borgen’s (1993) findings in that the
081 (if considered an indicator of negative afi‘ectivity or the tendency to catastrophize),
did not, by itself, predict job satisfaction but was predictive of symptomatic distress.
Also, the 081 did significantly interact with social support in the prediction of job
satisfaction.

Although, Shonfeld (1996) concluded that negative affectivity does not overly
distort the relationship of selected self-report measures of work stressors and job
satisfaction, the findings of this study suggest otherwise. Furthermore, Shonfeld chose to
assess negative affectivity indirectly through the use of elements of the Center for
Epidemiologic Studies Depression scale and a psychophysiologic scale. This
conceptualization of negative affectivity would not have captured the type of work-
related catastrophization revealed by the OSI in this study.

Previous studies investigating the buffering effect of social support on the work
stress-work strain relationship have found discrepant results which could be due, in part,
to the contamination of their work stress measures (to a greater or lesser degree) by work-
related negative affectivity. If this speculation is accurate, those studies using measures
that are susceptible to the bias of negative affectivity may have been more likely to
demonstrate a buffering effect.

The Relationship of Adult Attachment to Worlgtress. Job Satisfactiop, Symptomatic
Distress and Supervisor Supmrt
The results of this study did confirm the hypotheses that adult attachment style

would be significantly related to indexes of work stress and symptomatic distress.

86

Specifically, workers with higher levels of adult attachment anxiety reported significantly
more work stress and symptomatic distress than did secure workers. In addition,
dimensional measures of adult attachment orientation significantly enhanced the
prediction of symptoms even after work stress and supervisor support levels were
controlled. Although work stress was clearly the most efl'rcient predictor of symptomatic
distress (accounting for 42% of the variance), individuals with higher levels of avoidance
and those with higher anxiety reported more symptomatic distress than individuals who
scored lower on these dimensions of adult attachment (accounting for an additional 7 %
of the variance on this strain index).

Considering these results, together with previous findings, it is unclear whether or
not these individuals actually experience more stressful events at work or if they merely
perceive their work environment as more stressful. In his comprehensive review of the
literature regarding adult attachment style differences in affect regulation, Fuendeling
(1998) concluded that studies have consistently found that insecurely attached individuals
tend to make higher appraisals of threat to a variety of stimuli. Both Hazan and Shaver
(1990) and Hardy and Barkham (1994) reported that anxious/ambivalent workers were
fearful of rejection for poor job performance. Furthermore, Hardy and Barkham (1994)
found that these workers reported fears regarding their work relationships. It is possible
that, due to these fears, individuals with insecure attachment styles may actually instigate
or exacerbate stressful events and interpersonal encounters in their work environment,
thereby increasing their overall work-related distress.

In an innovative series of laboratory experiments, Mikulincer (1998) found

support for the idea that the default coping strategies for avoidantly attached adults and

87

anxious-ambivalently attached adults are very dissimilar. For avoidant adults,
disengagement from stress related cues and the pursuit of self reliance are the primary
coping strategies; whereas, for anxious-ambivalent adults, hypervigilance regarding
distress-related cues and attempts to secure the care and assistance of others are the
default strategies. These coping strategies are believed to occur below the level of
conscious awareness. However, Mikulincer found that these strategies do bias the way
individuals process information about the self, with avoidant persons predisposed toward
a positive view of self and anxious-ambivalent persons predisposed to a negative view of
self under stressfirl conditions. Therefore, one possible result of work stress on anxiously
attached individuals may be a hypersensitivity to criticism of work performance which, if
perceived in this fashion by others, may cause supervisors to withdraw from them
resulting in firrther distress to the worker who is already disposed to perceive him or
herself in negative terms.

Additionally, previous research has indicated that individuals with insecure
attachment styles demonstrate more anxiety and less competent affect regulation
strategies than their secure peers (Bartholomew & Horowitz, 1991; Fuendeling, 1998;
Kobak & Sceery, 1988). These characteristics may contribute to anxiously attached
individuals reporting higher levels of symptomatic distress when under stressfirl
conditions at work, as was found in this study.

The relationships among adult attachment style, supervisor support, and job
satisfaction appear to be complex. Although no significant differences were found in job
satisfaction among the three adult attachment style groups, the results of this study

indicated a significant difference in the perception of supervisor support by anxious and

88

secure workers. Moreover, dimensional measures of adult attachment orientation did not
enhance the prediction of job satisfaction after work stress and supervisor support were
concurrently controlled. However, after controlling for work stress and supervisor
support, the interactions of supervisor support with both the avoidance and anxiety
dimensions of adult attachment, respectively, significantly enhanced the prediction of job
satisfaction.

These findings may be due, in part, to the fact that this study’s sample size was
modest and that the distribution of supervisor support scores was skewed in a favorable
direction. Also, the fact that work stress alone explained a substantial amount of job
satisfaction variance may have made it especially difficult to discern any unique
contributions of adult attachment orientations to the prediction of this work strain index.
Indeed, when post-hoc regressions were repeated without inclusion of work stress scores,
both adult attachment indexes were significantly predictive of job satisfaction. These
findings are also consistent with attachment theory which posits that attachment-related
cognitions and behaviors are especially activated during periods of stress; therefore, when
work stress was initially controlled in the regression model, this relevant source of
variability was removed, making it more difficult to detect attachment style-related
contributions to job satisfaction.

These post-hoe analyses also revealed that the interactions of supervisor support
and attachment dimensions were significantly predictive of job satisfaction when work
stress was not part of the model. Under low support conditions, workers acknowledging
greater adult attachment avoidance expressed greater job satisfaction than did their peers

reporting lower avoidance. Moreover, workers endorsing higher levels of adult

89

attachment anxiety reported significantly lower job satisfaction under conditions of high
supervisor support than did workers reporting lower anxiety.

These findings lend support to the ideas set forth by Fuendeling who asserted that
what distinguishes avoidant from anxious individuals is their tendency to “make self-
enhancing secondary appraisals of their ability to cope with threats” (F uendeling, 1998,
p. 304). The results of the present study also are consistent with those of Mikulincer
(1998) who found evidence for the primacy of self-reliant coping among avoidantly
attached individuals and also demonstrated the central importance of winning the support
and compassion of others to anxiously-ambivalently attached adults. Therefore, the
tendency to use self-reliant coping strategies may allow avoidant workers to preserve
higher job satisfaction under low support conditions whereas less avoidant and more
anxious workers may rely on external resources (in this case, supervisor support) as a
preferred element of their coping strategy, thereby becoming more vulnerable to job
dissatisfaction when needed support is unavailable.

Limitations

To fully appreciate these findings, the limitations of this study should be
considered. First, data-gathering exclusively relied on self report measures; therefore, the
data solely reflect the respondents’ perceptions of their thoughts, feelings, behaviors, and
of the world around them. Even the reports of respondents regarding the frequency of
stressors at work is subject to distortion.

Second, although this study improved sampling over previously reviewed studies
with regard to the variety of occupations studied, sampling from one employer may have

contributed to less diversity in other important areas such as levels of supervisor support

90

and job satisfaction. This method of sampling thus limits the generalizability of these
findings to workers on other large state university campuses. Although the university
employees in this sample had different immediate supervisors and worked in different
settings, they still shared the same institutional employer offering standard benefits and
employee protection policies. These shared features may have contributed to less
variable job satisfaction than if the sample was drawn from multiple employers. Reliance
on a convenience, non-random, sample may have also introduced unknown bias. For
example, workers who responded to this mail survey may have been more satisfied with
their jobs than were non-respondents. Furthermore, large university settings are less
vulnerable to stressors such as corporate take over rumors or threats of massive layoffs
that have affected workers in for-profit industries in recent years. Theoretically, these
stressors could affect the overall level of work-related stress in a sample which, in turn,
could then affect findings regarding the buffering effect of social support on the stress-
strain relationships.

Third, there were few participants who endorsed a preoccupied adult attachment
style, necessitating their combination with fearful workers to form an “anxious”
attachment style group. The results of this action is the potential loss of valuable
information regarding attachment style variation on the other constructs under study. For
example, although adults with fearfirl attachment styles, as well as those with
preoccupied attachment styles, both have concerns about interpersonal rejection and are
theoretically predisposed to hypervigilance, they are likely to differ with regard to their
orientation toward support seeking. Fearfirl adults are likely to be apprehensive about

forming close relationships due to fears of rejection, whereas preoccupied adults are more

91

likely to pursue closeness under stressful conditions. Unfortunately, this study could not
explore these theoretical differences due to the aggregation of these two attachment style
categories.

Finally, the choice of instrumentation for this study may have contributed to
inconsistencies in findings from previous research on the affects of work stress and social
support on strains. The WSI is a measure of the intensity and fi’equency of work stresses
but does not break down into minute subscales with regard to different types of stress.
Previous studies using measures of work stress which breaks down into these subscales
(with relatively few items) have found the most evidence for the buffering effect of social
support. However, the WSI has some unique characteristics (i.e., the ability to
breakdown scores into Intensity (I) and Frequency (F) scales, while at the same time
using the product of these scales as a system for checking the potential for negative
affectivity to unduly influence overall composite scores), which allowed the researcher to
explore work-related negative affectivity as a potential reason for previous inconsistent
findings regarding the buffering hypothesis.

Implications

Despite the above mentioned limitations, the results of this study do provide
meaningful implications for attachment theory and counseling practice. These results
also suggest firture directions for research on attachment theory and the management of
work stress.

Implications for theog.

The results of this study advance attachment theory by demonstrating that

attachment style differences impact the lives of individuals outside of their parental and

92

romantic relationships. Furthermore, these findings generally support the conclusions of
several authors who have explored the role of attachment style in vocational behavior
(Blustein et al., 1995; Hardy & Barkham, 1994; Hazan & Shaver, 1990; Lopez, 1997).
Hazan and Shaver (1990) convincingly argued that work for adults functions similarly to
exploratory behavior in childhood and is fostered by security in attachments. They found
that secure workers approach their work with the confidence consistent with their
attachment style. They also found that when adults feel insecure in their attachments,
their concerns and fears may get in the way of their job satisfaction and their ability to
cope with work-related demands and relationships. These findings were generally
supported by the work of Hardy and Barkham (1994) as well. However, both of these
studies were considered to be preliminary in nature due to sampling techniques employed
and the use of exploratory measures. The present study sought to address these
limitations through improved sampling and the use of established instrumentation, as well
as to elaborate on the results of preceding studies.

Consistent with previous findings, job satisfaction was significantly related to adult
attachment orientation with higher avoidance and lower anxiety associated with greater
contentment. In addition, the results of this study extends the work of previous authors
by demonstrating that workers endorsing anxious attachment styles reported significantly
higher levels of work stress than did secure or dismissive workers.

Results also demonstrated that adult attachment styles were significantly related to
symptomatic distress among a sample of adult workers. These findings add to the body
of research suggesting the contribution of attachment behavior to affect regulation

(Bartholomew & Horowitz, 1991; Fuendeling, 1998; Kobak & Sceery, 1988). Although

93

previous studies exploring adult attachment and affect regulation have primarily sampled
from college age populations, this study provides evidence that these findings are not
exclusive to young adults. Consistent with the findings of Bartholomew and Horowitz
(1991) and Kobak and Sceery (1998), this study found that anxiously attached workers
reported greater symptomatic distress than either of their secure or dismissive
counterparts. Interestingly, when examining attachment-related orientations, this study
revealed that workers endorsing higher anxiety an_d those acknowledging higher
avoidance reported greater distress in terms of psychological symptoms, even after work
stress was controlled. This finding is somewhat surprising in that avoidance is usually
associated with a repression of feelings and with less willingness to report signs of
“personal weakness” or distress (Fuendeling, 1998). However, there may be differences
in the manner in which fearful-avoidant versus dismissive-avoidant individuals manage
their affect, differences which could not be detected by the instrumentation in this study.
For example fearfirlly attached individuals (who theoretically would endorse both high
anxiety and high avoidance on dimensional measures of attachment) may be less inclined
toward self-inflation and repression of affect as ways to regulate affect (Mikulincer,
1998)

Although this study did not completely resolve the controversy regarding the
buffering effect of social support on the stress-strain relationship, the findings did
demonstrate that the effect of supervisor support on job satisfaction depends upon the
worker’s perception of the intensity of potential work stressors. These results revealed
that workers who are disposed to perceive work stressors more intensely are responsive

to high levels of supervisor support and may depend upon this support more heavily than

94

other workers to ensure their job satisfaction. These findings suggest that individual
differences are important to consider when building theoretical models regarding the
response of workers to job stress. The findings of previous studies which did not
consider occupational stress intensity and frequency separately may have been biased, to
a greater or lesser degree, by work-related negative affectivity resulting in discrepant
findings regarding the buffering effect. Future studies should explore this hypothesis H
further by including a measure of negative affectivity such as the Negative Emotionality
(NEM) scale of Tellegen’s (1982) Multidimensional Personality Questionnaire.

Implications for counseling practice.

 

These findings support the idea that supervisors may play a critical role in
workers’ experience of job-related stress and job satisfaction. Although, in general,
workers with high supervisor support reported more job satisfaction, it is workers who
perceive stressors more intensely who appear to garner the most benefit from supervisor
support. In fact, these supported workers reported higher levels of job satisfaction
relative to their similarly supported peers who see potential stressors as less intense.
These findings also suggest that in order to maintain a highly satisfied workforce,
supervisors and employers should continue to invest in developing their understanding of
what constitutes high quality supervisor support and how to determine if workers are
particularly in need of greater support.

Furthermore, gaining knowledge of worker’s individual differences would be
beneficial for employee assistance program (BAP) counselors in developing tailored
stress management plans for workers presenting with psychological distress and job

dissatisfaction related to work stress. Beyond being helped to cope with their distress,

95

workers who perceive stressors very intensely might be coached in ways of assertively
expressing their need for supervisor support so that they can continue to maintain (or
develop) a positive feeling about their jobs.

In addition, understanding a worker’s attachment orientation may aid counselors
in assisting adults who enter counseling with work related interpersonal concerns.
Counselors may provide a secure base for workers who have anxious attachment Ft
orientations to explore their concerns and develop more effective coping strategies. For
example, an anxiously attached worker who may be hypervigilant to work stressors,

might benefit from “coaching” regarding detachment fi'om stressors in the work

 

environment or assertiveness training for clarifying and appropriately addressing
interpersonal concerns.

Recommendations for fixture reseath.

It is recommended that future research address the limitations of this study in
terms of sampling, design, and instrumentation. Sampling procedures could be improved
by using more than one employer, while still retaining diversity of occupations in the
sample. Furthermore, since preoccupied and fearfirl attachment styles are less common
in older adult samples, more participants (and especially younger workers) may need to
be recruited to study the unique contributions of these adult attachment style difl’erences
to the understanding of relationships among work stress, supervisor support, job
satisfaction, and symptomatic distress.

Future research exploring these constructs would also benefit by using diverse
methods of data gathering. One option might be to study small groups of supervisors and

workers in objectively high stress conditions such as reorganization, during a budgetary

96

crisis, or under strict deadlines. These groups could be assessed through qualitative
observation and by interview instruments, as well as through self-report measures. This
strategy would provide a framework for more clearly understanding individual
differences in perception of stress intensity and how it relates to attachment style
activation. Furthermore, using interview strategies could help illuminate interpersonal
dynamics around issues of trust, fear of abandonment and shame that may be activated
under stressfirl conditions at work. By employing this multi-level strategy, the
researcher(s) could also gain a greater understanding of which supervisory behaviors are
perceived as the most helpful to workers in coping with their stress. These clarifications,
in turn, could help supervisors better tailor their support to workers who are coping with
stressfirl work conditions or with fears regarding potential stressors.

Future research using the above data gathering strategy would also provide a
setting to study how different attachment style pairings (e. g., secure supervisor/avoidant
worker or preoccupied supervisor/fearfirl worker) affects the request, delivery, and
acceptance of social support under stressful condition. Studying groups of colleagues
would also enable researchers to further investigate how adult attachment style affects
self-presentation and identity negotiations in the work environment. For example,
studying work units might help us answer questions such as, “If avoidant individuals do
indeed inflate their self-reliance presentation how is this perceived by others?”; and
“How does the presentation of self-reliance affect support that is offered to workers?”.

Finally, the results of this study support the idea that the perception, and perhaps
the anticipation, of stress is crucial to understanding how a worker responds to supervisor

support. Future research regarding the buffering effect of social support should further

97

 

explore this finding. Furthermore, since anticipatory distress may reflect elements of
neuroticism, an instrument designed to measure this construct should be incorporated in
future research.
Conclusion

The results of this study indicate that the perception of work stress intensity and
adult attachment orientation are important to consider when investigating the relationship f
between work stress and strains. The findings also suggest that negative affectivity may
play a crucial role in understanding discrepant results in the stress buffering literature.

The present study has provided support for adult attachment style distinctions in

 

1‘."-

the perception or experience of work stress and symptomatic distress. The findings
confirm that adult attachment orientations significantly predict worker symptoms even
after levels of work stress and supervisor support are controlled. Furthermore, there is
preliminary evidence to suggest that adult workers’ attachment orientations may interact
with their perceptions of supervisory support in predicting current levels of job
satisfaction. These finding substantiate Bowlby’s assertion:

For not only young children, it is now clear, but human beings of all ages are

found to be at their happiest and to be able to deploy their talents to best

advantage when they are confident that, standing behind them, there are one or

more trusted persons who will come to their aid should difficulties arise (Bowlby,

1973,p.359)

Given the central role of work in adult life and the negative consequences of
experiencing stress in the workplace, work stress must be managed in the most effective

manner possible. Taken together, the findings of this study support continued inquiry

98

regarding attachment related issues in the workplace as a means of developing and

implementing more effective interventions for reducing worker strain.

99

 

REFERENCES

100

 

3'-

REFERENCES

Ainsworth, M. D. S., Blehar, M. C., & Waters, S. (1978). Patterns of attachment,
Hillsdale, NJ: Erlbaum.

Ainsworth, M. D. S. (1982). Attachment beyond infancy. American Psychologist. 44.
709-716.

Ainsworth, M. D. S., & Bowlby, J. (1991). An ethological approach to personality
development. American Psychologistp46. 333-341.

Bartholomew, K. (1990). Avoidance of intimacy: An attachment perspective. Journal of
SociaLand Personal Relationships, 7, 141-178.

Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among young adults:

A test of a four-category model. Journal of Personality and Social Psychology,
61, 226-244.

Baron, D. F., Caddy, G. R., Katell, A. D., Roselione, F. B., & Hamilton, R. A. (1988).
The work stress inventory: Organizational stress and job risk. Educational and
Psychological Measurement, 48, 141-154.

Beehr, T. A. (1976). Perceived situational moderators of the relationship between
subjective role ambiguity and role strain. Journal of Applied Psychology, 61,
35-40.

Beehr, T. A. (1985). The role of social support in coping with organizational stress. In
T. A. Beehr & R. S. Bhagat. (Eds), Human stress aad cognition in

organizations: An integrated perspective (pp. 375-398). New York: VVrley.

Beehr, T. A., King, L. A., & King, D. W. (1990). Social support and occupational stress:
Talking to supervisiors. Journal of Vocational Behavior 36, 61-81.

 

Behrens, E. N. (1998). Ralations of concordant and discordant parentaadult mchment
styles to adult psychological and relationship adjustment. Unpublished doctoral
dissertation, Michigan State University, East Lansing.

Berry, W. D., & Feldman, S. (1985). Multiple regression in practice. London, UK: Sage
Publications.

Blau, G. (1981). An empirical investigation of job stress, service length, and job strain.
Organizational Behavior and Humaa Performance, 22, 279-302.

101

 

Blustein, D. L., Schultheiss, D. P., & Prezioso, M. S. (1995). Attachment theory
and career development: Current status and future directions. The Counseling
Psychologist. 23. 416-432.

Bowlby, J. (1969) Attachment and loss. vol. 1: Attachment. New York: Basic Books.

Bowlby, J. (1982). Attachment an_d loss. vol. 1: Attachment 2“d ed. New York: Basic
Books.

 

Bowlby, J, (1973). Attachment and losgavol. 2: Separation: Anxiety and anger. New
York: Basic Books. P

Bowlby, J. (1979/1992). The making and breaking of affectional bonds. London:
Tavistock/Routledge.

Bowlby, J. (1980). Att_achmen_t_ and losa, vol. 3: Loss. New York: Basic Books. :5

 

Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human E
development. New York: Basic Books. '

Brennan, K. A., Clark,C. L., & Shaver, P. R. (1996, August). Development of anew
multi-item measure of adult romantic attachment: A preliminary report. Poster

presented at the International Society for the Study of Personal Relationships,
Banff, Alberta, Canada.

Brennan, K. A., Shaver, P. R., & Tobey, A. E. (1991). Attachment styles, gender, and
parental problem-drinking. Journal of Social and PersorraL Relationships, 8,
451-466.

Bretherton, I. (1992). The origins of attachment theory: John Bowlby and Mary
Ainsworth. Developmental Psychology, 2. 8, 759-775.

Caplan, R. D., Cobb, 8., French, J. R. P, Jr., Harrison, R. U., & Pinneau, S. R., Jr. (1975).
Job demands and worker health, US. Department of Health, Education, and
Welfare Publication No. 175-160. Y.S. Government Printing Office. The Institute
for Social Research, Washington, DC.

Camelley, K. B., Pietromonaco, P. R., & Jaffe, K. (1994). Depression, working models
of others, and relationship functioning. Journal of Personality and Social

Psychology, 66, 127-140.

Chisholm, R. F., Kasl, S. F., & Mueller, L. (1986). The effects of social support on
nuclear worker responses to the Three Mile Island accident. Journal of

Occupation Behavior, 7, 179-193.

Cohen, J. (1992). A power primer. Psychological Bulletin. 112, 155-159.

102

Cohen, S. & “Allis, T. A. (1985). Stress, social support, and the buffering hypothesis.
Psychological Bulletin. 98. 310—357.

Collins, N. L., & Read, S. J. (1990). Adult attachment, working models, and relationship
quality in dating couples. Journal of Personality and Social Psychology, 58,
644-664.

Coyne, J. C., & DeLongis, A. (1986). Going beyond social support: The role of social

relationships in adaptation. Journal of Consulting and Clinical Psychology, 54,
454-460.

Davila, J ., Burge, D., & Hammen, C. (1997). Why does attachment style change? n
Journal of Personality and Socjal Psychology, 73, 826-838.

Decker, P. J ., & Borgen, F. H. (1993). Dimensions of work appraisal: Stress, strain,
cOping, job satisfaction, and negative affectivity. Journal of Counseling

Psychology, 40, 470-478. J

 

Derogatis, L. R., Lipman, R. S., Rickels, K., Uhlenhuth, E. H., & Covi, L. (1974). The
Hopkins Symptom Checklist (HSCL): A self-report symptom inventory.
Behavioral Science. 1_9, 1-15.

Dillman, D. A. (1991). The design and administration of mail surveys. Anngl Review
of Sociology, 17, 225-249.

Fenlason, K. 1., & Beehr, T. A. (1994). Social support and occupational stress: Efl‘ects
of talking to others. Journal of Organizational Bahavior. 15. 157-175.

Florian, V., Mikulincer, M., & Bucholtz, I. (1995). Effects of adult attachment style on
the perception and search for social support. flip Joumaflf Psychology, 129,
665-676.

Folkman, S., Lazarus, R. S., Gruen, R. J., & DeLongis, A. (1986). Appraisal, coping,
health status, and psychological symptoms. Journal of Personality and Social
Psychology, 50, 571-579.

Fuendeling, J. M. (1998). Affect regulation as a stylistic process within adult attachment.
Journal of Social and Personal Relationships, 15, 291-322.

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

Gore, 8. (1978). The effect of social support in moderating the health consequences
of unemployment. Journal of Health and Social Behavior. 19. 157-165.

103

 

Gore, S. (1987). Perspectives on social support and research on stress moderating
processes. In J. M. Ivancevich & D. C. Ganster (Eds), Job stress: From
theory to suggestion. NY: The Haworth Press.

Griffin, D., & Bartholomew, K. (1994). Models of the self and other: Fundamental
dimensions underlying measures of adult attachment. Journal of Personality and
Social Psychology, 67, 430-445.

Haines, V. A., Hurlbert, J. S., & Zimmer, C. (1991). Occupational stress, social support
and the buffer hypothesis. Work and Occupations, 18, 212-235.

Hardy, G. E., & Barkham, M. (1994). The relationship between interpersonal attachment
styles and work difficulties. Human Relations, 47, 263-281.

Hazen, C., & Shaver, P. R. (1990). Love and work: An attachment-theoretical
perspective. Journal of Personality and Social Psychology, 59, 270-280.

Holmbeck, G. N. (1997). Toward terminological, conceptual, and statistical clarity in the
study of mediators and moderators: Examples from the child-clinical and
pediatric psychology literature. Journal of Consulting and Clinical Psychology,
6;, 599-610.

Holmes, J. (1993). John Bowlby and attachment theory. London: Routledge.

 

House, J. S. (1981). Work stress and social support. Reading, MA: Addison-Wesley.

House, J. S., & Wells, J. S. (1978). Occupational stress, social support, and health. In A.
McLean, G. Black & M. Colligan (Eds), Reducing occupational stress:
Proceedings of a conference (NIOSH Publication No. 78-140, pp. 8-29).
Washington, DC: US. Department of Health, Education, and Welfare

House, J. 8., Wells, J. A., Landerman, L. R., McMichael, A. J., & Kaplan, B. H. (1979).
Occupational stress and health among factory workers. Journal of Health and
Social Belavior. 20. 139-160.

Hurrell, J. J., Jr. (1998). Occupational health psychology in the United States: Past and
present. In H. E. Roberts-Fox (Chair), Building the future of occupational
health psychology. Symposium conducted at the American Psychological
Association 106 Annual Convention, San Francisco, CA.

 

Kaufman, G. M., & Beehr, T. A. (1986). Interactions between job stress and social

support: Some counterintuitive results. Journjal of Applied Psychology,ll_,
522-526.

104

 

 

 

Klohnen, E. C., & Bera, S. (1998). Behavioral and experiential patterns of avoidantly
and securely attached women across adulthood: A 3 l-year longitudinal
perspective. Journal of Personality and Social Psychology, 74, 211-223.

Kobak, R. R., & Sceery, A. (1988). Attachment in late adolescence: Working models,
affect regulation, and representations of self and others. Child Development, 59,
135-146.

Kobasa, S. E., & Puccetti, M. C. (1983). Personality and social resources in stress
resistance. Personality and Social Psychology, 45, 839-850.

LaRocco, J. M., & Jones, A. P. (197 8). Coworker and leader support as moderators of
stress-strain relationshps in work situations. Journal of Applied Psychology, 63,
629-634.

LaRocco, J. M., House, J. S., & French, In, J. R. P. (1980). Social support, occupational
stress, and health. Journal of Health and Social Behavior. 21. 202-218.

 

Leiter, M. P. (1991). Coping patterns and predictors of burnout: The firnction of control
and escapist coping patterns. Ila: Journal of Occupational Behavior. 12. 123-144.

Lent, E. B. (1992). The predictive ability of congruence and career self-efficacy in
adult workers: A study of job satisfaction (Doctoral dissertation, Michigan State
University, 1992). Dissertation Abstracts International 53 (6-B) 3195.

 

Lopez, F. G. (1995). Contemporary attachment theory: An introduction with
implications for counseling psychology. The Counseling Psychologist, 23,
395-415.

Lopez, F. G. (1997). Student-professor relationship styles, childhood attachment bonds
and current academic orientations. Journal of Social and Personal Relationshipa,
_15, 271-282.

Lopez, F. G., Gover, M. R., Leskela, J., Sauer, E., Schirmer, L., & Wyssmann, J. (1996).
Attachment styles, guilt, shame, and collaborative problem-solving
orientations. Personal Relationships, 4, 187-199.

Lopez, F. G., Melendez, M., Sauer, E.M., Berger, E., & Wyssman, J. (1998). Internal
working models, self reported problems, and help seeking attitudes among college
students. Journal of Counseling Psychology, 45, 79-83.

 

Mickelson, K. D., Kessler, R. C., & Shaver, P. R. (1997). Adult attachment in a
nationally representative sample. Journal of Personality and SocialPsycholggy,
13, 1092-1106.

105

ll

Mikulincer, M. (1998). Adult attachment style and affect regulation: Strategic
variations in self-appraisals. Journal of Personality said Social Psychology, 75,
420-435.

Mikulincer, M, Kessler, R. C., & Shaver, P. R. (1997). Adult attachment in a nationally

representative sample. Journal of Personality and Social Psychology, 73,
1092-1106.

Mikulincer, M., & Nachshon, O. (1991). Attachment styles and patterns of self-
disclosure. Jourrral of Personality and Social Psychology, 61, 321-331.

Norusis, M. J. (1997). SPSS 7.5: Guide to data analysis. Upper Saddle River, NJ:
Prentice Hall.

Ognibene, T. O., & Collins, N. L. (1998). Adult attachment styles, perceived social

support and coping strategies. Journal of Social and Personal Relationships, 15,
323-345.

Pistole, M. C. (1989). Attachment in adult romantic relationships: Style of conflict
resolution and relationship satisfaction. Journal of Social and Personal

Relationships, 6, 505-510.

Pond, S. B., & Geyer, P. D. (1987). Employee age as a moderator of the relations
between perceived work alternatives and job satisfaction. Journal of Applied
Psychology, 7_2_, 552-557.

Priel, B., & Shamai, D. (1995). Attachment style and perceived social support: Effects
on affect regulation. Personality and IndividuiDifferencea. 19. 235-241.

Quinn, R. P., & Shepard, L. (1974). The 1973-1974 quality of employment survey:
Descriptive statistics. Ann Arbor, MI: Institute for Social Research, Survey
Research Center.

Sarason, I. G., Sarason, B. R., & Shearin, E. N. (1986). Social support as an individual
difference variable: Its stability, origins, and relational aspects. Journal of
Personality and Social Psychology, 50, 845-855.

Sauter, S. L., Murphy, L.R., & Hurrell, J. J., Jr. (1990). Prevention of work-related
psychological disorders: A national strategy proposed by the National Institute
for Occupational Safety and Health (NIOSH). American Psychologist, 45,
1 146-1 158.

Scharfe, E., & Bartholomew, K. (1994a). Reliability and stability of adult attachment
patterns. Personal Relationships, 1, 23-43.

106

 

 

Scharfe, E., & Bartholomew, K. (1994b). Stability of adult attachment representations:
A two year follow-up. Paper presented at the Annual Convention of the
American Psychological Association, Los Angeles, California.

Schonfeld, I. S. (1996). Relation of negative affectivity to self-reports of job stressors
and psychological outcomes. Journal of Occupational Health Psychology, 1,
397-412.

Shilling, S. S., & Brackbill, R. M. (1987). Occupational health and safety risks and
potential health consequences perceived by US. workers, 1985. Public Health
Reports. 102. 36-46.

Simpson, J. A. (1990). Influence of attachment styles on romantic relationships. Journal '-
of Personality and Social Psychology, 59, 971-980.

 

 

Simpson, J. A., Rholes, W. S., & Nelligan (1992). Support seeking and support giving .
within couples in an anxiety-provoking situation: The role of attachement styles. ,
Journal of Personality and Socjal Psychology, 62, 434-446.

Tabachnick, B. G., & Fidell, L. S. (1996). Using multivariate statistics (3rd ed.).
New York, NY: HarperCollins.

Tellegen, A. (1982). Brief manual for the differential personality questionnaire.
Unpublished manuscript, University of Minnesota, Minneapolis.

Wallace, J. L., & Vaux, A. (1993). Social support network orientation: The role of
adult attachment. Journal of Social and Clinical Psychology, 12, 354-365.

Weathers, P. L., Furlong, M. J., & Solorzano, D. (1993). Mail survey research in
counseling psychology: Current practice and suggested guidelines. Journal of
Counseling Psychology 40, 238-244.

107

APPENDICES

108

 

 

APPENDIX A

Prenotification Postcard

 

 

Dear (Name inserted here):

I am writing to ask your cooperation in a study regarding the reaction of workers
to stress in the workplace. You are one of a small number of MSU employees
selected at random as potential participants in this study. A research packet,
including 6 brief questionnaires, will be mailed to you in about 2 weeks along with a
stamped return envelope. All survey responses will be kept completely confidential.
It is important that we understand how workers react to stress and how we may keep
them from the potential negative consequences of these experiences. Your
participation in this study would be greatly appreciated.

Sincerely,
Lisa L. Schirmer, M.Ed.

Doctoral Candidate
MSU Counseling Psychology Program

P.S. As an added incentive, when you return a completed survey, you will be entered
into a drawing for a $100.00 cash prize.

 

109

 

 

 

APPENDIX B

Work Stress Project
PO. Box 510021
Livonia, MI 48185

November 12, 1997

 

Name

Address a
City, MI Zip l '
Dear (Name inserted here):

I am writing once again to ask for your participation in a study regarding worker’s ,. -
reaction to stress in the workplace. This research is being conducted by Lisa L. Schirmer, B
a doctoral candidate in the Department of Counseling, Educational Psychology, and -

Special Education at Michigan State University (MSU), under the supervision of Dr.
Frederick Lopez. Your name was drawn at random from a listing of MSU employees as
a potential participant in this study.

The purpose of this research is to learn more about how workers are affected by stress
from their jobs or at their workplace. We are also interested in exploring factors that may
help protect workers from developing emotional/health problems due to stress. Enclosed
you will find a survey packet containing questionnaires which solicit information
regarding your general background; how you have been feeling recently; the level and
frequency of stress that you experience at work; and, information about your work and
personal relationships.

Please know that all survey responses will be kept completely confidential. Code
numbers have been assigned to your name and you will not be asked to place your name
on the survey packet. The primary researcher, Lisa Schirmer, will be the only person that
has access to the list of code numbers and names. This list will be used only to keep
track of surveys that have been returned and will be destroyed after the study is
completed. I expect that it will take you about 20-30 minutes to complete the survey
packet. I do not anticipate that your participation will result in any physical or emotional
risk to you. Your participation is completely voluntary and you may withdraw from the
study at anytime without penalty to you. As a benefit for your participation in this study,
your code number will be entered into a drawing for a $100.00 cash prize when you
return a completed survey packet.

I would be happy to answer any questions that you may have about this project. You can
write to me at the above address, or call me collect at (313) 525-9236. You may receive a
. summary of the results of this study by writing “copy of results requested” on the back of
your return envelope and printing your name and address below it. Please do not put this

110

 

 

information on the survey itself. Thank you, in advance, for your participation. Your
cooperation in this study is greatly appreciated.

Sincerely,
Lisa L. Schirmer, M.Ed.

Doctoral Candidate
MSU Counseling Psychology Program

111

 

 

\‘ag- _ m

APPENDIX C

Participation Consent Form

Thank you for agreeing to participate in my study of worker’s response to stress in
the workplace. You will find a research code number in the top right hand corner
of the first page of the survey packet. Please copy your research code number
below and also enter today’s date in the appropriate spot.

By placing your research number below, and returning a completed survey packet,
you will be indicating that you understand:

1) the nature of this project and the nature of your participation;

2) that your participation is voluntary and that you can temrinate your
participation at any time without penalty;

3) that there will be no risk to you by completing this survey;

4) that it will take approximately 20-30 minutes to complete the survey;

5) that the results of this survey will be treated with strict confidence and
that the participants will remain anonymous in any report of research
findings.

 

By returning this form, along with your completed survey, you will also be entered
into a drawing for a $100.00 cash prize in the form of a money order. If you are
the participant randomly selected to receive this prize it will be mailed to you on or
before December 22, 1997.

 

 

I have read the above and agree to participate in the study, being conducted by Lisa
Schirmer, of worker’s response to stress in the workplace. Please enter me in the
drawing for a $100.00 cash prize!

 

 

YOUR RESEARCH CODE NUIVIBER TODAY’S DATE

112

APPENDIX D

Demographic and Background Information

Thank you for participating in this study. These final questions ask about your
background. Please circle the number to the right of the information which best describes

 

you, or place your response in the blank space provided ( _) .
Your sex: Your Occupational Group:
Female 1 Administration 1
Male 2 Faculty 2
Staff 3
Your current age: Other (please explain) 4

 

Your ethnic/racial background:

 

 

 

African-American 1 How long have you been at your
Asian-American 2 present job?
Caucasian/White 3 Under 1 year 1
Hispanic/Latino(a) 4 1 to 2 years 2
Multiracial 5 2 to 5 years 3
Other (please describe) 6 5 to 10 years 4
10 to 15 years 5
15 to 20 years 6
Your Marital Status: Over 20 years 7
Single 1
Committed Partnership 2 Do you have a supervisor at work?
Married 3 Yes 1
Divorced 4 No 2
\Vrdowed 5
Do you supervise others at work?
Please indicate the level of education Yes 1
which you have received: No 2
High School/GED 1
Some College 2 Please indicate your yearly income from
Associates Degree 3 your current job:
Bachelors Degree 4 $0 to $10,999 1
Masters Degree 5 $11,000 to $19,999 2
Doctorate Degree 6 $20,000 to $29,999 3
Other (please explain) 7 $30,000 to $39,999 4
$40,000 to $49,999 5
$50,000 to $59,999 6
$60,000 to $69,999 7
$70,000 and over 8

113

 

APPENDIX E

RQ

Directions: Please read each of the descriptive paragraphs below and place a checkka
(V) next to the ONE paragraph that best describes how you feel about close
relationships. After this, using the scale provided, choose a number from 1 to 7 to rate
how characteristic each paragraph is to your typical relationships.

Scale:
Not at all Slightly Moderately Very Extremely
1 2 3 4 5 6 7

 

 

Most descriptive Rating
of me (1 to 7)

(Check one) (Rank all)

1. It is easy for me to become emotionally
close to others. I am comfortable
depending on others and having others
depend on me. I don’t worry about being
alone or having others not accept me -) 1.

 

2. I am comfortable without close emotional
relationships. It is very important for me to feel
independent and self-sufficient, and
I prefer not to depend on others or
have others depend on me -) 2.

 

3. I want to be completely emotionally intimate
with others, but I often find that others are
reluctant to get as close as I would like. I am
uncomforable being without close relationships,
but I sometimes worry that others don’t value
me as much as I value them -) 3.

 

4. I am uncomfortable getting close to others.
I want emotionally close relationships, but
I find it difficult to trust others completely,
or to depend on them. I worry that I will
be hurt if I allow myself to become too close
to others. -) 4.

 

114

 

 

 

APPENDIX F

ECR

Instructions: The following statements concern how you feel in romantic relationships.

We are interested in how you generally experience relationships, not just in what is

happening in a current relationship. Respond to each statement by indicating how much

you agree or disagree with it.

Disagree
Strongly
1. I prefer not to show a partner how I feel deep 1 2
down.
2. I worry about being abandoned. I 2
3. I am very comfortable being close to romantic 1 2
partners.
4. I worry a lot about my relationships. 1 2

5. Just when my partner starts to get close to me, 1 2
I find myself pulling away.

6. I wony that romantic partners won’t care 1 2
about me as much as I care about them.

7. I am uncomfortable when a romantic partner 1 2
wants to be close.

8. I worry a fair amount about losing my partner. 1 2

9. I don’t feel comfortable opening up to l 2
romantic partners.

10. I often wish that my partner’s feelings for me 1 2
were as strong as my feelings for him/her.

11. I want to get close to my partner, but I keep 1 2
pulling back.

115

Neutral/

Mixed
4 5
4 5
4 5
4 5
4 5
4 5
4 5
4 5
4 5
4 5
4 5

Agree
Strongly

7

-a.1--

 

M.,}:- .L

1 ..
Q

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

I often want to merge completely with
romantic partners, and this sometimes scares
them away.

I am nervous when partners get too close to
me.

I worry about being alone.

I feel comfortable sharing my private
thoughts and feelings with my partner.

My desire to be very close sometimes scares
people way.

I try to avoid getting too close to my partner.

I need a lot of reassurance that I am loved by
my partner.

I find it relatively easy to get close to my
partner.

Sometimes I feel that I force my partners to
show more feeling, more commitment.

I find it difficult to allow myself to depend on
romantic partners.

I do not often worry about being abandoned.

I prefer not to be too close to romantic
partners.

If I can’t get my partner to show interest in
me, I get upset or angry.

I tell my partner just about everything.

I find that my partner(s) don’t want to get as
close as I would like.

116

Disagree
Strongly

1

2

3

Neutral!

Mixed

4

5

Agree
Strongly
6 7
6 7
6 7
6 7
6 7
6 7
6 7
6 7
6 7
6 7
6 7
6 7
6 7
6 7
6 7

 

 

 

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

I usually discuss my problems and concerns
with my partner.

When I am not involved in a relationship, I
feel somewhat anxious and insecure.

I feel comfortable depending on romantic
partners.

I get frustrated when my partner is not around
as much as I would like.

I don’t mind asking romantic partners for
comfort, advise or help.

I get fi'ustrated if romantic partners are not
available when I need them.

It helps to turn to my romantic partner in
times of need.

When romantic partners disapprove of me, I
feel really bad about myself.

I turn to my partner for many things,
including comfort and reassurance.

I resent it when my partner spends time away
from me.

117

Disagree Neutral/
Strongly Mixed

12345

Agree
Strongly
6 7
6 7
6 7
6 7
6 7
6 7
6 7
6 7
6 7
6 7

 

 

 

APPENDIX G

Social Support Scale

Using the scale adjacent to each of the items below, circle the number which best
represents how much each of these individuals give you support in the following ways:

Don’t
Very Some- A Not Have Any
Much what Little At All Such Person

 

1. How much does each of these people go
out of their way to do things to make
your work 1113: easier for you?

 

 

A. Your immediate supervisor (boss)... 4 3 2 1 0
B. Other people at work... 4 3 2 1 0
C Your spouse, friends and relatives... 4 3 2 1 0
2. How m i_s_ i_t t_o fl with each of the

following people?
A. Your immediate supervisor... 4 3 2 1 0
B. Other people at work... 4 3 2 1 0
C. Your Spouse, friends and relatives... 4 3 2 1 0
3. How much can each of these people

be relied o_n when things get tough

at work?
A. Your immediate supervisor (boss)... 4 3 2 1 0
B. Other people at work... 4 3 2 1 O
C. Your spouse, friends and relatives... 4 3 2 1 0
4. How much is each of the following

people willing rp listen to your

personal problems?
A. Your immediate supervisor (boss)... 4 3 2 1 0
B. Other people at work... 4 3 2 1 0
C. Your spouse, friends and relatives... 4 3 2 1 0

118

APPENDIX H

WSI

Stress, for the purpose of this inventory, is defined as feelings of emotional strain,
pressure, discomfort, uneasiness, and/or tension.

INSTRUCTIONS:

You are to evaluate your current job for the amount and frequency
of stress experienced. For each item, use the following scale to
indicate the amount of stress that i_s_ experienced or would be if it
were to occur.

 

None

A little
Moderate
Much
Very much

:55"pr

Then, use the following scale to indicate how often it occurs,
stressful or not:

Never

Rarely (annually)

Sometimes (at least monthly)
Often (at least weekly)

Daily

95”.“???

Be sure to circle an answer for amount and frequency for each
item.

 

119

 

EXAMPLE:

If you feel that “Not knowing what superiors expect of you,” is very stressful, whether
or not you currently experience this problem on the job, you would rate the amount of
stress for item #1 as a “4”, as shown below. In the next step, you rate the frequency that
you experience this type of stress on your current job. If you never experience this type
of stress, you would rate it a “0”, as shown below.

AMOUNT OF FREQUENCY
STRESS

 

HODIN
NEIldO
KIIVCI

EINON
SILLY]
EIAEIN
A'IEDIVH

SEWILEIINOS

EILV‘tIEICIOW
HOIIW AHEIA
‘d

 

 

 

 

 

 

 

 

 

 

1. Not knowing what superiors
expect of you.

 

 

O
u—
N
1.»
G9
@
-_o
N
w
.h-

 

120

 

AMOUNT OF FREQUENCY

 

 

 

 

 

 

 

 

 

 

 

 

STRESS
a r g 5 a a a o v
a 3 v n a g a a E
1T1 E :1: E 52 1-1 '<
El g E
1. Not knowing what superiors
expectofyou. 0 1 2 3 4 0 1 2 3 4
1. Having to respond on an
“emergency basis”. 0 1 2 3 4 0 l 2 3 4

 

2. Disagreeing with superiors.

 

4. Not knowing how much
authority you have. 0 1 2 3 4 0 1 2 3 4

 

 

5. Being injured as a result of the
mistakes of others. 0 1 2 3 4 0 1 2 3 4

 

6. Having to deal with injury or
deathaspartofyourjob. 0 l 2 3 4 0 1 2 3 4

 

7. Having to make decisions that
will dramatically affect other
pe0ples’ live.

 

8. Finding that rewards are not
based on performance (e. g.,
promotions, raises).

 

9. Having to deal with several
pressing problems at once. 0 1 2 3 4 0 1 2 3 4

 

10. Working in a “high crime area.”

 

 

 

121

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AMOUNT OF FREQUENCY
STRESS
W O U
E E Q -<: § E F11
g E W E: '-1 z '<
3:] o a
m m
11. Not knowing what superiors
thinkofyou. 0 2 3 4 0 1 2 3 4
12. Not having the opportunity to
participate in decision making. 0 2 3 4 0 l 2 3 4
13. Having conflicting job
responsibilities. 0 2 3 4 0 1 2 3 4
14. Working without adequate safety
standards. 0 2 3 4 0 1 2 3 4
15. Having inadequate personnel or
equipment to respond in an 0 2 3 4 0 1 2 3 4
emergency situation.
16. Feeling there is no clear chain of
command. 0 2 3 4 012 34
17. Having periods of inactivity
Separated by perrods of 0 2 3 4 0 1 2 3 4
emergency response.
18. Having to physically restrain
others. 0 2 3 4 012 34
19. Potential for being injured on
thejob. 0 23401234
20. Being held responsible for too
many different activities. 0 2 3 4 0 1 2 3 4

 

 

 

122

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AMOUNT OF FREQUENCY
STRESS
Z 5 z W O U
0 o E a 5 O a E
E 3 E’- 9 -< § E
['11 a E W a z '<
E] o
a: m
21. Knowing that your error may
harm another person. 0 1 2 3 4 0 2 3 4
22. Failing to receive recognition of
achievement by supervisors. 0 1 2 3 4 0 2 3 4
23. Having to do things on the job
that are agarnst your better 0 1 2 3 4 0 2 3 4
judgement.
24. Never knowing when a
potentrally dangerous event 0 1 2 3 4 0 2 3 4
mrght occur.
25. Feeling that your work ability is
underrated. 0 l 2 3 4 0 2 3 4
26. Not being permitted to make
decisions on your own. 0 1 2 3 4 0 2 3 4
27. Working for long periods of time
without rest. 0 1 2 3 4 0 2 3 4
28. Performing duties that are
potentially dangerous to others. 0 1 2 3 4 0 2 3 4
29. Receiving criticism from
superiors. 0 1 2 3 4 0 2 3 4
30. Recervrng conflrctrng requests. 0 1 2 3 4 0 2 3 4

 

 

 

123

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AMOUNT OF FREQUENCY
STRESS
2 E: Z W O U
o —1 8 E) g g E O "’1 E
% a “1 5“ '< w r' E a '<
E? E < a
at g V,
31. Finding a lack of assistance or
support from superiors. 0 1 2 3 4 O 1 2 3 4
32. Working in excess of eight
hoursperday. O 1 2 3 4 0 1 2 3 4
33. Working with dangerous
materials. 0 l 2 3 4 O 1 2 3 4
34. Having ideas considerably
different from those of your 0 1 2 3 4 0 1 2 3 4
superiors.
35. Doing another person’s job in
addition to yours. 0 1 2 3 4 0 1 2 3 4
36. Having to maintain prolonged
Vigilance to protect the safety of 0 1 2 3 4 O 1 2 3 4
others.
37. Potential for being the victim of
acrimewhile onthe job. 0 1 2 3 4 O 1 2 3 4
38. Being held responsible for
mistakes made by co-workers. 0 1 2 3 4 0 1 2 3 4
39. Working while fatigued or tired. 0 1 2 3 4 O 1 2 3 4
40. Working under inconsistent
policies and guidelines. 0 1 2 3 4 0 1 2 3 4

 

 

 

124

APPENXIX I

JSS

Directions: Circle one number on each of the following scales to represent your answer
to these questions:

1. If you had to decide all over again whether to take the job you now have, what would
you decide?

1 2 3 4 5
definitely not take job definitely take job

2. If a qualified friend asked if he/she should apply for a job like yours with your
employer, what would you recommend?
1 2 3 4 5
recommend not at all recommend strongly
3. How does this job compare to your ideal job?
1 2 3 4 5

very far from ideal very close to ideal

4. How does your job measure up to the sort of job you wanted when you took it?

1 2 3 4 5
not at all like what just what I wanted
I wanted

5. All things considered, how satisfied are you with your current job?

1 2 3 4 5
not at all satisfied completely satisfied

125

APPENDIX J

HSCL

Using the scale below please indicate: “How have you felt during the past seven days
including today?” Please rate yourself on the following symptoms using the four-point

distress scale to the right of each item.
:5 is vs fax
,5 . . s
a at” «a

1. Headaches 1 2 3 4
2. Nervousness or shakiness inside 1 2 3 4
3. Being unable to get rid of bad thoughts or ideas 1 2 3 4
4. Faintness or dizziness 1 2 3 4
5. Loss of sexual interest or pleasure I 2 3 4
6. Feeling critical of others 1 2 3 4
7. Bad dreams 1 2 3 4
8. Difliculty in speaking when you are excited I 2 3 4
9. Trouble remembering things 1 2 3 4
10. Worried about sloppiness or carelessness l 2 3 4
11. Feeling easily annoyed or irritated 1 2 3 4
12. Pains in the heart or chest 1 2 3 4
13. Itching 1 2 3 4
14. Feeling low in energy or slowed down 1 2 3 4
15. Thoughts of ending your life 1 2 3 4
16. Sweating l 2 3 4
l7. Trembling l 2 3 4
18. Feeling confused I 2 3 4
19. Poor appetite 1 2 3 4
20. Crying easily 1 2 3 4
21. Feeling shy or uneasy with the opposite sex 1 2 3 4
22. A feeling of being trapped or caught 1 2 3 4

126

23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.

37.

38.

39.
4o.
41.
42.
43.
44.
45.

46.
47.
48.
49.

Suddenly scared for no reason

Temper outburst you could not control
Constipation

Blaming yourself for things

Pains in the lower part of your back
Feeling blocked or stymied in getting things done
Feeling lonely

Feeling blue

Worrying or stewing about things
Feeling no interest in things

Feeling fearful

Your feelings being easily hurt

Having to ask others what you should do

Feeling others do not understand you or are
unsympathetic

Feeling that people are unfiiendly or
dislike you

Having to do things very slowly in order
to be sure you are doing them right

Heart pounding or racing
Nausea or upset stomach
Feeling inferior to others
Soreness of your muscles
Loose bowel movements
Difficulty in falling asleep or staying asleep

Having to check and double check what
you do

Difficulty making decisions
Wanting to be alone
Trouble getting your breath
Hot or cold spells

127

§

_ay—uHH—ay—AH—ny—a

p—nu—Ip—dp—a

ud—ap‘udud—n...‘

W sf f

NNNNNNNNNNNNN

NNNNNNN

NNNNN

wwwwwwwwwwwww

wwwwwww

wwwww

AA-hAA-h-bAA-bA-b-b

AAAA-h-h-h

AbbA-h

50

51.
52.
S3.
54.
55.
56.
57.
58.

. Having to avoid certain places or activities
because they frighten you

Your mind going blank

Numbness or tingling in parts of your body
A lump in your throat

Feeling hopeless about the future

Trouble concentrating

Weakness in parts of your body

Feeling tense or keyed up

Heavy feelings in your arms or legs

128

@V'B'

1 2
1 2
1 2
1 2
l 2
l 2
1 2
l 2
1 2

3

wwwwwwww

§$$ a
state”

A-h-h-A-h-h-h-h

 

APPENDIX K

ThankYou/Reminder Postcard

 

 

 

    
   

Dear (Name inserted here):
Thank you for agreeing to participate in the Work Stress Project. Ifyou

haven not already done so, please complete and return the Work Stress Project

Survey ’97. Upon receipt of your completed survey, you will be entered in a

drawing for a $100.00 cash prize. Thanks again for your participation!

Sincerely,

Lisa L. Schirmer, M.Ed.
Doctoral Candidate
MSU Counseling Psychology Program

 

129

APPENXIX L

Work Stress Project
P. O. Box 510021
Livonia, MI 48151

December 3, 1997

(NAME)

(ADDRESS)

(CITY, STATE ZIP)

Dear (Insert Name):

Approximately three weeks ago I wrote to you requesting your participation in a study
regarding the reaction of employees to stress in the workplace. As of this writing, I have
not received your completed survey. If you have not already completed the survey, I ask
that you take the time to do so now. I have enclosed another copy of the survey for your
convenience. Upon receipt of your completed survey packet, I will enter your survey
code number into a drawing for a $100.00 cash prize.

Your response to this survey is important to the success of this study. You were one of
250 MSU employees selected at random to represent all university workers. In order for
the results of this study to truly represent the community, it is essential that each person
in the sample return his or her completed survey. The purpose of this research is to learn
more about how employees are affected by stress from their jobs or at their workplace. I
am also interested in learning more about factors that may help protect workers from
developing emotional/health problems due to stress.

Remember that all survey responses will be kept completely confidential. Code numbers
have been assigned to your name and you will not be asked to place your name on the
survey packet. I will be the only person that has access to the list of code numbers and
names. This list will be used only to keep track of responses that have been received and
will be destroyed after the study is completed. I would be happy to answer any questions
that you may have about this project. You can write to me at the above address, or call
me collect at (313) 525-9236. You may receive a summary of the results of this study by
writing “copy of results requested” on the back of your return envelope and printing your
name and address below it. Please do not put this information on the survey itself.

Thank you in advance for your participation.

 

Sincerely,
Lisa L. Schirmer

Doctoral Candidate
MSU Counseling Psychology Program

130

"‘11111111111111“