WORK/PERSONAL LIFE BALANCE: A CONSTRUCT DEVELOPMENT STUDY

Gwenith Gwyn Fisher

A Dissertation

Submitted to the Graduate College of Bowling Green

State University in partial fulfillment of
the requirements for the degree of

DOCTOR OF PHILOSOPHY

August 2001

Committee:

Carlla S. Smith, Advisor

Vincent Kantorski,
Graduate Faculty Representative

Milton D. Hakel -

Dara Musher-Eizenman 11 9
 ii

ABSTRACT

Carlla S. Smith, Advisor

Many changes in the workplace and in employees’ lives have taken place over the

past couple decades and have led to an increase in the attention paid to the boundary

between work and non-work issues among employees. However, much of this research

has been limited by its emphasis on family in the domain of non-work. The present study

was undertaken to extend previous research by developing a broader construct:

work/personal life balance. Based on the conservation of resources theory and qualitative

research, work/life balance was defined as an occupational stressor regarding issues of

time, energy, goal accomplishment, and strain.

A 21-item scale was developed as part of this study to measure employees'

perceptions of work/life balance, and it was expected that this scale would be related to a

variety of stressors and strains. Data were collected from a heterogeneous sample of 540

managers employed in a variety of organizations across the U.S. Responses from 273 of

these managers were retained for use as a holdout sample for cross-validation purposes.

Across both samples, results indicated three dimensions of work/life balance: work

interference with personal life, personal life interference with work, and work/personal

life enhancement.

As predicted, this measure was significantly related to role overload, time

pressure, time management behavior, overall feelings of job stress, job satisfaction, life

satisfaction, organizational commitment, and turnover intentions in both samples. In

addition, results indicated initial evidence of discriminant validity between work/life

balance and work/family conflict. Overall, this study provides initial evidence of
 Ill

construct validity. Implications for theories of occupational stress as well as

organizational practice are discussed. Continued work/life balance scale development is

warranted.
 iv

“A balanced and skillful approach to life becomes an

important factor in conducting one’s everyday existence.
It is important in all aspects of life.”
--Dalai Lama
The Art of Happiness
 V

ACKNOWLEDGMENTS

I would like to thank the people who assisted me in achieving this educational

milestone. I am most grateful to my advisor. Dr. Carlla Smith, for her continued support,

dedication, open door, and confidence in me throughout this process. Many thanks to Dr.

Milt Hakel, not only for his feedback and service on this committee, but for much

wisdom and professional mentorship during my graduate school career. I would also like

to thank my other committee members, Drs. Dara Musher-Eizenman, and Vince

Kantorski for their time and helpful suggestions. Sincere appreciation is also extended to

Drs. Mike Zickar and Jeff Stanton for their contributions to this project.

This research would not have been possible without the support of the American

Management Association, for allowing me to survey a portion of their members and

expending the financial resources to make it all happen. Many thanks to Dr. Monica

Hemingway, Jody Hoffman, Julie Fuller, Maggie Laber. and Lilly Lin for assisting me

with the qualitative research conducted as part of this investigation. I also extend deep

gratitude to Dr. Steven Rogelberg for his mentoring earlier in my graduate career— for

stimulating my interest in quality of work life initiatives and survey measurement issues,

being an excellent statistics teacher, and always demonstrating enthusiasm for research.

I am extremely grateful for having had the opportunity to be a part of such a

successful graduate program at BG. Many thanks to the Penn State folks who gave me

knowledge and experience that opened windows of curiosity and doors of opportunity for

me— Drs. John Mathieu, Dave Day, Jim Farr, Tonia Heffner, and Jay Goodwin. I also

give thanks to Armand “Puck" Spoto for making those PSU days and many since then

more enjoyable with his friendship and great sense of humor.

 VI

I greatly appreciate the twenty-eight years of constant love, support, and

encouragement from my parents, Loring Fisher and Janet Dimotsis. Thanks, mom and

dad, for always believing in me and helping me to succeed in reaching my goals. I am

also grateful for the love and support of my other “family,” Mickey Duggan, Chrisy

Duggan, and Holly Mack, who have always been there for me. I am also thankful for the

friends who greatly improved the quality of my graduate school life: Emily, Stephanie,

Jen, Paul, Stefoni, Brenda, Ana, Michelle, Heath, Jody, and Laurie.

I appreciate having had the opportunity to work with Drs. Andrea Goldberg and

Ellie McCreery. They have been great mentors and sources of support, both personally

and professionally. 1 thank each of them for teaching me the importance of and

satisfaction in having a life beyond work, and for showing me that achieving a

work/personal life balance is possible.

I give many thanks to Savanna for her daily reminders to strive for this balance in

my own life. Most importantly, I am grateful to my partner, Devin, for his love, support,

and patience during this process, for many helpful suggestions on earlier drafts, and for

adding so much to my life. I look forward to charting the course together through the

remainder of life’s chapters.

 Vll

TABLE OF CONTENTS

INTRODUCTION.................................................................................................................1

CHAPTER I: BACKGROUND RESEARCH...................................................................... 3

Organizational Stress................................................................................................ 4

Definition of Stress....................................................................................... 4

Heuristic Models of Organizational Stress...................................................5

Theories Related to Work/Non-Work Conflict.........................................................6

Role Theory.................................................................................................. 6

Spillover Model ........................................................................................... 7

Congruence Model....................................................................................... 9

Compensation Model ................................................................................... 9

Segmentation Model .................................................................................. 10

Conservation of Resources Theory............................................................. 11

Summary......................................................................................................13

Empirical Research on Work/Non-Work Conflict ................................................ 14

Work/Family Conflict.................................................................................. 14

Measurement of Work/Family Conflict....................................................... 20

Previous Work/Life Balance Research........................................................ 22

Summary..................................................................................................... 24

CHAPTER II: THE PRESENT STUDY............................................................................ 26

Work/Personal Life Balance Definition.................................................................. 26

Work/Personal Life Balance Measurement............................................................ 30

Hypotheses.............................................................................................................. 32
 VIH

Hypothesis 1................................................................................................. 32

Hypothesis 2................................................................................................. 32

Hypothesis 3................................................................................................. 33

Hypothesis 4................................................................................................. 33

Hypothesis 5................................................................................................. 33

Hypothesis 6................................................................................................. 34

Hypothesis 7................................................................................................. 34

Hypothesis 8................................................................................................. 34

Hypothesis 9................................................................................................. 35

Hypothesis 10............................................................................................... 35

CHAPTER III: METHOD.................................................................................................. 37

Study One - Work/LifeBalance Scale Development.............................................. 37

Participants................................................................................................... 37

Procedure..................................................................................................... 38

Measures...................................................................................................... 38

Results.......................................................................................................... 39

Study Two................................................................................................................ 41

Participants and Procedure........................................................................ 41

Measures......................................................................................................46

Data Analyses.............................................................................................. 54
 IX

CHAPTER IV: RESULTS................................................................................................... 60

Initial Sample........................................................................................................... 60

Demographic Information........................................................................... 60

Scale Assessment.........................................................................................60

Summary...................................................................................................... 88

Validity Evidence......................................................................................... 88

Hypothesis Tests.......................................................................................... 91

Hypothesis 1.....................................................................................91

Hypothesis 2.....................................................................................92

Hypothesis 3.....................................................................................92

Hypothesis 4.....................................................................................92

Hypothesis 5.....................................................................................93

Hypothesis 6.....................................................................................93

Hypothesis 7.....................................................................................93

Hypothesis 8.....................................................................................93

Hypothesis 9.....................................................................................93

Hypothesis 10...................................................................................94

Cross-Validation Using Holdout Sample...............................................................106

Demographic Information.......................................................................... 106

Scale Assessment....................................................................................... 106

Scale Item Assessment Summary..............................................................119

Holdout Sample Scale ValidityEvidence................................................... 125

Hypothesis Tests........................................................................................ 125

 X

Hypothesis 1...................................................................................125

Hypothesis 2...................................................................................128

Hypothesis 3.................................................................................. 128

Hypothesis 4.................................................................................. 128

Hypothesis 5.................................................................................. 129

Hypothesis 6...................................................................................129

Hypothesis 7.................................................................................. 129

Hypothesis 8...................................................................................129

Hypothesis 9...................................................................................130

Hypothesis 10................................................................................ 130

Exploratory Analyses............................................................................................ 131

Demographic Differences......................................................................... 131

Incremental Validity of Enhancement...................................................... 137

CHAPTER V: DISCUSSION.......................................................................................... 146

Scale Item Generation............................................................................................ 146

Scale Item Assessment........................................................................................... 147

Work Interference with Personal Life....................................................... 148

Personal Life Interference with Work....................................................... 148

Work/Personal Life Enhancement............................................................ 149

Scale Validity........................................................................................................ 150

Convergent Validity................................................................................... 150

Discriminant Validity................................................................................. 154

Criterion-related Validity........................................................................... 155

 XI

Summary of Construct Validity Evidence................................................ 161

Theoretical Implications....................................................................................... 162

Comparison of Work/Life Balance to Work/Family Conflict..............................163

Practical Implications............................................................................................ 165

Limitations.............................................................................................................166

Overall Summary and Conclusions...................................................................... 168

REFERENCES.................................................................................................................. 170

APPENDIX A: Preliminary Study Interview Questions.................................................. 184

APPENDIX B: Work/Life Balance Scale Items in Study One

(Content Validation)............................................................................... 185

APPENDIX C: Questionnaire Cover Letter..................................................................... 186

APPENDIX D: Work/Life Balance Scale inStudy Two................................................... 187

APPENDIX E: Work/Family Conflict Scales.................................................................. 189

APPENDIX F: The Stress in General Scale..................................................................... 190

APPENDIX G: The Job in General Scale......................................................................... 191

APPENDIX H: Life Satisfaction Scale............................................................................. 192

APPENDIX I: Organizational Commitment Scale.......................................................... 193

APPENDIX J: Demographic Questions.......................................................................... 194

 xn

LIST OF TABLES
Table

1 Work/Life Balance Scale Items Retained after Content Validation Study.............. 40

2 Demographic Characteristics of the Total Study Sample (N=540).......................... 43

3 Demographic Characteristics of the Initial Sample................................................... 61

4 Work/Life Balance Scale Item Means and Standard Deviations - Initial Sample ....63

5 Correlation Matrix of Work/Life Balance Scale Items - Initial Sample................... 64

6 Pattern Matrix of Initial Work/Life Balance Items - Initial Sample.........................71

7 Component Correlation Matrix of Work/Life Balance Scale Items -

Initial Sample............................................................................................................. 72

8 Pattern Matrix of Work/Life Balance Items after Removing Item 13 -

Initial Sample............................................................................................................. 74

9 Component Correlation Matrix of Work/Life Balance Scale Items after

Removing Item 13 - Initial Sample........................................................................... 75

10 Component Matrix and Communalities Among Fourteen Work Interference

with Personal Life (WIPL) Items - Initial Sample.................................................... 78

11 Corrected Item-Total Correlations for Work/Life Balance Scale Items -

Initial Sample..............................................................................................................79

12 Correlations Between Each Work/Life Balance Scale Item and

One-Item Work/Life Balance Measure - Initial Sample........................................... 82

13 Fit Statistics for Confirmatory Factor Analysis Models - Initial Sample................ 86

14 Descriptive Statistics for all Study Variables - Initial Sample................................ 89

15 Correlation Matrix of Study Variables - Initial Sample.......................................... 90

16 Hierarchical Regression Analysis for Work/Family Conflict and Work/Life

Balance Variables Predicting Role Overload - Initial Sample (N = 267)................ 95

17 Hierarchical Regression Analysis for Work/Family Conflict and Work/Life

Balance Variables Predicting Overall Stress (Pressure) -
Initial Sample (N = 267).............................................................................................96
 XIII

18 Hierarchical Regression Analysis for Work/Family Conflict and Work/Life

Balance Variables Predicting Overall Stress (Threat) - Initial Sample (N = 267) ...97

19 Hierarchical Regression Analysis for Work/Family Conflict and Work/Life

Balance Variables Predicting Time Pressure - Initial Sample (N = 267)................ 98

20 Hierarchical Regression Analysis for Work/Family Conflict and Work/Life

Balance Variables Predicting Setting Goals and Priorities -
Initial Sample (N = 267)............................................................................................. 99

21 Hierarchical Regression Analysis for Work/Family Conflict and Work/Life

Balance Variables Predicting Planning and Scheduling -
Initial Sample (N = 267)........................................................................................... 100

22 Hierarchical Regression Analysis for Work/Family Conflict and Work/Life

Balance Variables Predicting Preference for Organization -
Initial Sample (N = 267)........................................................................................... 101

23 Hierarchical Regression Analysis for Work/Family Conflict and Work/Life

Balance Variables Predicting Overall Job Satisfaction -
Initial Sample (N = 267)........................................................................................... 102

24 Hierarchical Regression Analysis for Work/Family Conflict and Work/Life

Balance Variables Predicting Life Satisfaction -
Initial Sample (N = 267)........................................................................................... 103

25 Hierarchical Regression Analysis for Work/Family Conflict and Work/Life

Balance Variables Predicting Affective Organizational Commitment-
Initial Sample (N = 267)........................................................................................... 104

26 Hierarchical Regression Analysis for Work/Family Conflict and Work/Life

Balance Variables Predicting Turnover Intentions - Initial Sample (N = 267)..... 105

27 Demographic Characteristics of the Holdout Sample (N = 273)............................ 107

28 Work/Life Balance Scale Item Means and Standard Deviations -

Holdout Sample....................................................................................................... 109

29 Corrected Item-Total Correlations for Work/Life Balance Scale Items -

Holdout Sample........................................................................................................ 110

30 Correlations Between Each Work/Life Balance Scale Item and

One-Item Work/Life Balance Measure - Holdout Sample..................................... 112

31 IRT a Parameters for Each Work/Life Balance Scale Item - Total Study
Sample....................................................................................................................... 115
 XIV

32 Descriptive Statistics for all Study Variables - Holdout Sample............................126

33 Correlation Matrix of Study Variables - Holdout Sample...................................... 127

34 Hierarchical Regression Analysis for Work/Life Balance Variables

Predicting Setting Goals and Priorities - Total Study Sample (N=540).................139

35 Hierarchical Regression Analysis for Work/Life Balance Variables

Predicting Planning and Scheduling- Total Study Sample (N=540)...................... 140

36 Hierarchical Regression Analysis for Work/Life Balance Variables

Predicting Overall Job Stress (Threat) - Total Study Sample (N=540)..................141

37 Hierarchical Regression Analysis for Work/Life Balance Variables

Predicting Overall Job Satisfaction - Total Study Sample (N=540)....................... 142

38 Hierarchical Regression Analysis for Work/Life Balance Variables

Predicting Life Satisfaction - Total Study Sample (N=540)................................... 143

39 Hierarchical Regression Analysis for Work/Life Balance Variables

Predicting Affective Organizational Commitment - Total Study
Sample (N=540)....................................................................................................... 144

40 Hierarchical Regression Analysis for Work/Life Balance Variables

Predicting Turnover Intentions - Total Study Sample (N=540)............................. 145
 XV

LIST OF FIGURES
Figure

1 Scree Plot of Eigenvalues from Principal Components Analysis Among

Work/Life Balance Scale Items - Initial Sample...................................................... 69

2 Final Work/Life Balance Scale Confirmatory Factor Analysis Model -

Initial Sample............................................................................................................ 87

3 IRT Test Information Function for WIPL dimension - Total Study

Sample (N=540)....................................................................................................... 116

4 IRT Test Information Function for PLIW dimension - Total Study

Sample (N=540)....................................................................................................... 117

5 IRT Test Information Function for WPLE dimension - Total Study

Sample (N=540)....................................................................................................... 118

6 Age Differences in Work Interference with Personal Life - Total Study

Sample (N=540)....................................................................................................... 133

7 Age Differences in Personal Life Interference with Work - Total Study

Sample (N=540)....................................................................................................... 134

8 Age Differences in Work/Personal Life Enhancement - Total Study

Sample (N=540)....................................................................................................... 135

9 Work Interference with Personal Life by Annual Salary - Total Study

Sample (N=540)....................................................................................................... 138
 xvi

PREFACE

Because employees spend approximately half of their waking lives at work, I

decided to pursue graduate study in a field in which I could devote my career to applying

psychology to improve the quality of employees’ work lives. As part of my professional

development during graduate school, I opted to spend my third year on leave for a 12-

month internship as a marketing research consultant at IBM’s corporate headquarters in

Armonk, NY. During my tenure at IBM, I often experienced difficulty in finding time to

pursue my personal interests (e.g., spending time with friends, playing tennis, bicycling,

running, spending time with my dog, and traveling) and think about trying to get enough

sleep while try ing to perform well on a job that involved working at least 50 hours a

week, bringing work home, and traveling on occasion.

Even as a single, childless worker I encountered this struggle, and it was then that

I learned two important things. First, I found that having a balance between my work life

and my life away from work was important for my overall health and well-being. (In

retrospect. I find it hard to believe that it took years of part-time jobs to get through

college, long hours in graduate school, and almost six months of full-time employment

before I realized that, not only did I live, eat, & breathe my job, but that sacrificing work

accomplishments in order to have a satisfying personal life never really occurred to me. I

think many of my friends and family would have agreed I was highly involved in my

work— it is no wonder why I became so fascinated with the psychology of “work!” )

It was not long after this epiphany during the winter of 1997/1998 that I learned

that being able to exercise regularly actually helped me cope with stress on the job, and

gave me more energy to get through a long work day. Having more energy helped me to
 XVII

be more productive, which certainly was of benefit to my clients within the company.

Secondly, I realized that having a significant other, children, or having to care for aging

family members is not a prerequisite for having this struggle; the difficulty I experienced

was legitimate, despite others’ expectations that I would be available for lengthy travel

assignments overseas or could work overtime because I did not have anyone waiting for

me at home.

In light of my own anecdotal evidence for this phenomenon, I undertook this

study to take a more empirical approach to learning about work/life balance among other

employees. I am grateful for the experience that taught me a valuable lesson, and the

many people at IBM and elsewhere who helped get me to where I am today, striving for

my own work/personal life balance.

 INTRODUCTION

Workplace stress is affecting our national effectiveness and well-being. For

example, the estimated industry expense of stress-related disorders is $150 billion per

year, and more than 14% of insurance compensation claims are based on stress-related

disorders (Pelletier & Lutz, 1989; 1991). Stress at work is related to a number of

deleterious and costly individual problems (e.g., headaches, gastrointestinal disorders,

anxiety, hypertension, coronary heart disease, depression) and negative organizational

outcomes (e.g., job dissatisfaction, burnout, accidents, loss of productivity, absenteeism,

and turnover; Ganster & Schaubroeck, 1991; Gupta & Beehr, 1979; Ironson, 1992;

Murphy, Hurrell, Sauter, & Keita, 1995; Quick, Horn, & Quick; 1987, Sauter & Murphy,

1995). These consequences have sustained a continuing interest in assessing stress at

work in an effort to understand the etiology of workplace stress, to pinpoint sources of

stress, as well as to guide the use of stress reduction interventions.

One particular source of stress (i.e., a stressor) that has been the focus of much

empirical research is work/non-work conflict. Work/non-work conflict refers to the

challenge many of us face trying to juggle work responsibilities with aspects of one’s

personal life, such as maintaining friendships, exercising, spending time with a romantic

partner, and/or caring for one’s aging parents or young children. Previous work/non-

work conflict research has primarily focused on family as the salient aspect of life in the

non-work domain. For example, many studies have focused on developing and testing

models regarding the antecedents and outcomes of work/family conflict (e.g., Frone,

Russell, & Cooper, 1992; Frone, Yardley, & Markel, 1997; Grandey & Cropanzano,

1999).
 2

Although the study of work/non-work conflict to date has increased our

understanding of stress that results from work/family conflict, this research can be

criticized for having both conceptual and methodological limitations. For example, this

research is limited in scope because it fails to examine aspects of non-work life in

addition to family. Most studies only include employees who are married, have children,

and/or have elder care responsibilities. As a result, employees who are not married, do

not have children, or may not have the responsibility of caring for aging parents are

omitted from study, yet these employees may still perceive stress resulting from other

types of work/non-work conflict. The purpose of the present study is to address some of

the limitations of previous research on work/non-work conflict by developing a broader

construct: work/life balance. The goal of developing a broader construct for

conceptualizing work/non-work conflict is to deepen our understanding of stress that

results from the process of trying to juggle one’s work and personal life.

The remainder of this paper will describe the present study undertaken to develop

the construct of work/personal life balance (also referred to as work/life balance). The

beginning will summarize background research, including theories and previous

empirical studies examining work/family conflict. This background will be followed by

a description of the present study, the hypotheses to be tested in this study, and a

description of the methodology employed to test those hypotheses. Finally, the statistical

analyses that will be conducted to interpret the data will be described.

 3

CHAPTER I. BACKGROUND RESEARCH

Many changes in the workplace and in employees' lives have taken place over the

past couple decades and have led to an increase in the attention paid to work/non-work

conflict (Parcel, 1999). In the workplace, an increasing number of women are joining the

workforce, with the largest segment including single mothers (Lerner 1994; Zedeck,

1992). Employees are spending more of their time working, as evidenced by an

increasing number of hours that employees work, compared to the traditional 40-hour

work week (Casner-Lotto & Hickey, 1999; Figart & Golden. 2000). In addition, workers

often telecommute (i.e., work from home) or bring work home, thus blurring the

boundaries between work and home (Hill, Miller, Weiner, & Colihan, 1998; Shamir,

1992). Organizations have responded to these trends by implementing programs to

support employees' needs, such as on-site childcare (Rothausen, Gonzalez, Clarke, &

O'Dell, 1998) and flexible work arrangements (including flexibility in time and location

of work; Hill et al., 1998).

With regard to changes in employees’ personal lives, an increasing number of

women are delaying childbirth until their late thirties or early forties because of career

concerns (Ventura, Martin, Curtin, & Mathews, 1998). An increasing number of men

and women are adopting a more egalitarian approach to work and family issues, breaking

down traditional gender roles in work and family (Barnett & Rivers, 1996; Parcel, 1999).

According to the opens systems approach to the study of organizations, an

understanding of the relationship between an organization and its external environment is

paramount (Katz & Kahn, 1978). One’s work is one aspect of one’s life, and therefore

the examination of one’s work life should take into account the context in which that
 4

work takes place, including family and other personal concerns (Kopelman, Greenhaus,

Connolly, 1983; Schein, 1976). When one’s work life interferes with his/her life outside

of work or one’s personal life negatively impacts his/her work, a particular source of

stress (i.e., a stressor) referred to as work/non-work conflict may occur. In order to

understand how work/non-work conflict may be considered a stressor, the next section

provides a definition and overview of the organizational stress process. Following the

general description of organizational stress is a description of various theories in the

organizational psychology literature that relate specifically to work/non-work conflict.

Organizational Stress

Definition of Stress

The literature on organizational stress is replete with multiple definitions of the

term “stress.” For example, a stimulus definition focuses on stressors, such as events or

situations in the environment that “contribute to or cause a noxious or disruptive

experience” (Smith & Sulsky, in press). In the context of work stress, the stressor may be

an event or situation at work that leads to a disruptive experience at work or outside of

work. In addition, work stress is also concerned with non-work related events or

situations that lead to a noxious or disruptive experience at work.

A second definition of stress is a response definition, which is not concerned with

the stimulus or stressor, but rather with how an individual reacts to the stressor (Selye,

1965). This definition of stress is also referred to as the medical definition or approach.

Responses include immediate physiological reactions, such as changes in heartbeat, and

long-term or sometimes chronic changes called strains (Smith & Sulsky, in press).
 5

Examples of strains include long-term changes in behavior, attitudes, and health (e.g.,

absenteeism, job dissatisfaction, and cardiovascular disease).

A third definition of stress combines the first two approaches and is referred to as

a stimulus-response definition. This approach takes into account the interaction between

a stimulus (i.e., a stressor) and an individual's response to that stimulus. For the purpose

of the present study, stress is defined as “any circumstance (stressor) that places special

physical and/or psychological demands on an organism such that an unusual or out-of­

ordinary reaction or response occurs, which, if unabated, may result in long-term or

chronic undesirable outcomes (i.e., strains)” (Smith & Sulsky, in press).

Lazarus’ research on stress (Lazarus, 1966; Lazarus & Folkman, 1984)

underscores the role of cognitive appraisal in the stress process. Specifically, Lazarus

purports that a particular event or situation is not a stressor unless it is appraised as such

by an individual. As a result, some individuals may perceive a particular event as

stressful, although others may not.

Heuristic Models of Organizational Stress

A number of heuristic models of stress have been proposed (e.g., Beehr &

Newman, 1978; Edwards, 1992; Katz & Kahn, 1978). Although these models vary with

regard to specific elements, each acknowledges the dynamic nature of the stressor-strain

relationship such that stress is a process rather than a discrete event. Examples of these

models include the social environmental model of stress (French & Kahn. 1962; Katz &

Kahn, 1978), Beehr & Newman’s (1978) conceptual model of work stress, and

Edwards’s (1992) cybernetic model.

 6

In total, these heuristic models indicate that stress is a dynamic process that takes

into account characteristics of an individual (e.g., personality, ability, interests, etc.) as

well as characteristics of his/her work and non-work environment (e.g., role demands at

work, role demands outside of work) and may lead to a number of potential outcomes for

both the individual and the organization for which he/she works.

Such heuristic models are useful for understanding stress as a process in general.

However, these models are not specific enough to yield thorough understanding of the

stress process, especially with regard to understanding the impact of any particular

stressor. The remainder of this investigation focuses on the interface between

employees’ work and non-work lives as a potential stressor. In order to better understand

the stress process as it relates to work/non-work conflict, a number of specific theoretical

models will be explained.

Theories Related to Work/Non-Work Conflict

Previous research on work/non-work conflict has relied on a variety of theoretical

conceptualizations. Among the theories or models described in this research are role

theory, spillover theory, congruence, compensation, and the conservation of resources. A

description of each of these theories as they relate to work/non-work conflict research

follows.

Role Theory

In the late 1970’s, work/non-work conflict research began with the application of

role theory to understanding a specific type of work/non-work conflict: work-family

conflict (Kopelman et al., 1983; Quinn & Staines, 1979). Role theory was first

introduced in the 1950s by sociologists such as Parsons (1951) and Merton (1957),
 7

although Kahn, Wolfe, Quinn, Snoek, and Rosenthal (1964) were the first to discuss role

theory in the context of work organizations. Role theory is based on the premise that an

individual’s life is comprised of a number of roles, including roles at work as well as

roles outside of work. Role conflict occurs when two (or more) sets of pressures occur at

the same time such that compliance with one set makes it more difficult to comply with

the other set (Kahn et al., 1964). Objective role conflict is a product of the environment,

whereas subjective role conflict is based on an individual’s perception. In addition, Kahn

et al. (1964) suggest that interrole conflict will be greater if employees have multiple

roles. Multiple roles may lead to psychological conflict if each role cannot be adequately

fulfilled.

Role conflict is a type of role stress that is commonly studied in organizational

stress research (King & King, 1990; Smith & Sulsky, in press). This approach has been

related to work/non-work conflict research based on the notion that conflict experienced

between one’s work role and one’s role outside of work is a salient stressor. However,

reliance on role theory in work-family research has been criticized for a few reasons

(Grandey & Cropanzano, 1999; Zedeck & Mosier, 1990). First, Kahn et al.'s (1964) role

theory is primarily focused on roles in the context of the work environment and does not

address family or other non-work roles. Secondly, this theory fails to articulate variables

that may moderate the relationship between work and family stressors and stress

outcomes (Jackson & Schuler, 1985).

Spillover Model

In addition to the role theory approach to studying work and non-work aspects of

employees' lives, a number of other models appear widely throughout the work/non-work
 8

conflict literature. These include the spillover model (Staines, 1980), the congruence

model (Morf, 1989), the compensation model (Staines, 1980), and the segmentation

model (Evans & Bartolomé, 1984; Piotrkowski, 1978).

The spillover model purports that there is a correspondence between what occurs

at work and what occurs in one’s life outside of work (Staines, 1980). That is, what

happens at work can “spill over” and affect one’s non-work life, or vice versa in which

aspects of one’s personal life can spill over and affect his/her work life. For example,

this approach suggests that satisfaction with one’s work leads to satisfaction at home.

Spillover has been conceptualized in both a positive and a negative way (e.g.,

Grzywacz, 2000; Grzywacz & Marks, 2000). Positive spillover refers to enhancement

that results from one's work life affecting his/her family life or vice-versa. An example

of positive spillover was previously mentioned (i.e., satisfaction in one domain leads to

satisfaction in the other domain). Negative spillover, on the other hand, refers to conflict

between work and non-work, such that negative aspects of one’s work affect non-work in

a negative manner. For example, long hours and psychological stress at work could lead

to less time spent at home and psychological stress at home as well.

Most of the extant work-family research has relied upon a spillover explanation of

the relationships between aspects of work and family (e.g., Williams & Alliger, 1994;

Zedeck & Mosier, 1990). One of the problems with the spillover model, however, is this

notion of positive and negative spillover: One may experience both positive and negative

spillover at the same time (Gryzwacz & Marks, 2000). Research applying the spillover

theory to work-family conflict has yet to address this complication. For example, a job

with long hours or a great deal of travel may take away from time one spends doing non­
 9

work activities. However, the same job may also be financially rewarding and personally

satisfying to the employee, which can have a positive impact on one’s family.

Congruence Model

The congruence model described by Morf (1989) is similar to the spillover model

in that this model also postulates a positive relationship between work and non-work

roles. However, the congruence model attributes the cause for the relationship to a third

variable, whereas spillover is based on the notion that one domain directly influences

another (Morf, 1989; Staines, 1980; Zedeck & Mosier, 1990). In other words, this model

purports a positive relation between one's work and family life based on a third variable

underlying each domain (e.g., personality variables, genetic factors, and social or cultural

factors; Frone, et al., 1992; Morf, 1989; Staines, 1980).

One specific example of a variable that may lead to congruence is negative

affectivity. Negative affectivity is a dispositional trait that refers to one’s tendency to

interpret situations in a negative way (Watson, Clark, & Tellegen, 1988). For example,

an individual may seem dissatisfied with his/her job as well as with his/her personal life,

and this simultaneous dissatisfaction in both domains could be due to the individual

having a strong negative affective disposition that impacts his/her perceptions in both

domains.

Compensation Model

In contrast to the spillover and congruence models, another approach to

understanding the relations between one’s work life and his/her life outside of work is the

compensation model (Staines, 1980). This approach suggests that what an individual is

missing in one domain, she will seek to find (i.e., compensate for) in another domain.
 10

For example, an employee who finds her job to be boring may seek opportunities to do

something she finds exciting when not at work. In other words, the compensation model

predicts antithetical experiences in the work and family domain, or a negative

relationship between work and non-work experiences.

Segmentation Model

A fourth model articulating the relations between work and non-work variables is

the segmentation model (Evans & Bartolomé, 1984). According to this model, work and

non-work domains are separate from one another, and as a result, there is no relationship

between what an employee does at work and what he does when not at work. In other

words, this model suggests that employees can compartmentalize their lives such that

work is separated from non-work in time, space, and function (Zedeck, 1992). As a

result, one's work remains at work, and their non-work life does not affect work or is not

affected by work.

Although previous research has found empirical support for each of the models

just described (spillover, congruence, compensation, and segmentation; e.g., Frone et al.,

1992; Hart, 1999; Williams & Alliger, 1994), each model has been criticized for a

number of reasons (Grandey & Cropanzano, 1999; Zedeck, 1992; Zedeck & Mosier,

1990). One criticism is that these models are “atheoretical,” not specifying boundary

conditions regarding when relationships prescribed by the model should hold.

In addition, empirical research has failed to reconcile competing predictions made

by each of the models. For example, Zedeck and Mosier (1990) raised the question

regarding whether more than one model may be applicable to an individual and whether

spillover and compensation can occur simultaneously or in a sequential order (e.g., can
 11

one experience spillover for some factors, and seek to compensate for others?). Some

empirical support has been found for all four models, although among them, most

research has been based on the spillover model (Small & Riley, 1990; Zedeck & Mosier,

1990). Grandey & Cropanzano (1999) pointed out that these models are limited in that

they have not been integrated into a single, comprehensive theory to guide work/life

research.

Conservation of Resources Theory

The conservation of resources theory holds promise as an approach for guiding

research examining employees' work and non-work lives (Grandey & Cropanzano,

1999). In 1988, Hobfoll proposed the conservation of resources theory as a new theory

of stress. This theory is based on the notion that “people have an innate as well as

learned desire to conserve the quality and quantity of their resources and to limit any state

that may jeopardize the security of these resources” (p. 25).

According to this theory, stress is a reaction to an environment in which one is

threatened by a potential loss in resources, experiences an actual resource loss, or fails to

gain expected resources (Hobfoll 1988; 1989). “Resources” are supplies that can help or

support an individual, and can be conceptualized in four distinct categories, including

objects, personal characteristics, conditions, and energies. Objects are material

possessions, and examples include food, water, shelter, and a vehicle. Personal

characteristics are aspects of an individual, including traits, feelings, attitudes, or other

characteristics. Hardiness is a personal attribute characterizing one's psychological

resiliency that serves as a source of resistance to stress (Rush, Schoel, & Barnard, 1995).

Conditions refer to states that are considered desirable or that one seeks. Social support
 12

is an example of a condition that is sought and can buffer one against stress

(Viswesvaran, Sanchez, & Fisher, 1999). Various types of social support exist, including

support from one’s supervisor, coworker(s), or spouse. The fourth type of resource is

energies, which refers to the means for attaining such resources. Examples of energies

include time, knowledge, or physical energy. This latter category includes particularly

salient resources with regard to meeting the multiple demands of one’s work and non­

work life (Fisher & Hemingway, 2000).

The conservation of resources model takes into account that stress is a process

(Edwards, 1992; Jex & Beehr, 1991; Smith & Sulsky, in press). Specifically, it is an

experienced loss of resources, or the threat of a loss of resources, that initiates the stress

process. The cognitive evaluation of resources here is similar to the cognitive appraisal

process described by Lazarus (1966) and his colleagues (Lazarus & Folkman, 1984), such

that the perception of stress is in the eye of the beholder.

The conservation of resources theory can be applied to examining stress that

results from the process of balancing one's work and non-work life. As Grandey and

Cropanzano (1999) have indicated, the conservation of resources model is useful for

explaining the results of interrole stress. According to the conservation of resources

theory', interrole conflict between one's work role and his/her non-work role leads to

strains because one loses resources when try ing to perform both work and non-work

roles. This actual or potential loss of resources can lead to what Hobfoll (1989) terms a

“negative state of being,” including dissatisfaction, depression, anxiety, or physiological

strains.
 13

When comparing work and non-work roles. Edwards and Rothbard (2000) have

indicated that there is a negative relationship between one’s work and non-work

resources such that resources used in one domain limit their use in another domain. Time

is one example of a particular resource that is likely to be lost or threatened when

attempting to meet both work and non-work role demands. For example, the more time

one spends doing work, the less time is available for non-work activities. Edwards and

Rothbard (2000) point out that this negative relationship may be difficult to predict

because one’s non-work life includes multiple domains (e.g., family, community

activities, and personal hobbies/activities). In order to allocate more time to work, one

may spend less time sleeping or attending to one’s own, non-work activities.

How one allocates his/her resources may or may not be a conscious choice. For

example, one may choose to spend more time at work to complete a particular project in

time for a deadline. Or, one may be dealing with unforeseen non-work issues (e.g.,

family problems or the death of a close friend) that prevents one from focusing on work

even if he/she intends to devote attention to work.

Summary

A number of theories have been applied to the work/non-work research domain.

These include role theory, spillover, congruence, compensation, segmentation, and the

conservation of resources models. Role theory purports that an employee may

experience interrole conflict between his or her work and non-work roles as a stressor.

Although many work/non-work conflict studies have been conceptualized based on role

theory, this approach has been criticized for its focus on roles within the work

environment and failure to lead to specific predictions about role conflict. The spillover,
 14

congruence, compensation, and segmentation models have each served as a basis for

some previous research, but have been criticized for making competing predictions that

are not falsifiable. In addition, these models are more descriptive than prescriptive. In

other words, they describe relationships between work and non-work variables, but do

little to explain why we observe such behavior.

In contrast to other approaches, the conservation of resources model (Hobfoll,

1988; 1989) suggests that an individual experiences stress in the event of resource loss.

Although this model is a descriptive theory not unlike the spillover, segmentation, and

compensation models, this approach provides a useful framework for understanding

stress reactions that result from work/non-work conflict because time and energy are

salient, limited resources necessary for achieving one's work and non-work goals.

Empirical Research on Work/Non-work Conflict

Work/Eamily Conflict

Over the past two decades, an increased number of empirical studies have

examined work/non-work conflict (Allen, Herst, Bruck. & Sutton, 2000; Kossek &

Ozeki, 1998; Zedeck, 1992). The vast majority of this research has focused on “family”

in the domain of non-work, examining a specific type of work/non-work conflict:

work/family conflict. Work/family conflict occurs when work pressures and family

pressures occur at the same time such that compliance with pressures in one domain (e.g.,

work) makes it more difficult to comply with pressures in another domain (e.g., family).

Most work/family conflict research to date developed and tested antecedents and

outcomes of work/family conflict ( Frone et al., 1992; Frone, et al., 1997; Grandey &

Cropanzano. 1999; Greenhaus & Beutell, 1985; Kopelman et al., 1983).

 15

Kopelman et al. (1983) were among the first to systematically examine the

construct of work-family conflict. They began by defining three role conflict variables:

work conflict, family conflict, and interrole conflict. Work conflict is the extent to which

one experiences incompatible role pressures within the work domain. Similarly, family

conflict is the extent to which one experiences incompatible role pressures with regard to

family. Interrole conflict occurs when one experiences pressures of one role that are

discordant with pressures in another role. Kopelman et al. (1983) developed three scales

to measure each of these constructs and assessed the scale validity with two data samples.

Kopelman et al.’s (1983) results did provide evidence of construct validity for

three dimensions of conflict. However, the only outcomes they examined were job

satisfaction, family satisfaction, and life satisfaction. Specifically, results of their path

analyses indicated a negative relationship between work conflict and job satisfaction, and

a negative relationship between family conflict and family satisfaction. Work conflict

and family conflict were modestly correlated. Work conflict predicted interrole conflict in

both samples, although family conflict predicted interrole conflict in only one of two

samples.

A number of other researchers (e.g., Grandey & Cropanzano, 1999; Greenhaus &

Beutell, 1985; Greenhaus & Parasuraman, 1986; Frone et al., 1992) built upon Kopelman

et al.’s work by developing and testing additional models of work-family conflict. For

example, based on a review of previous literature, Greenhaus & Beutell (1985)

articulated three potential sources of work-family conflict: time based conflict, strain-

based conflict, and behavior-based conflict. T ime-based conflict occurs when time spent

in one role (based on the number of hours worked, inflexibility in work scheduling, and
 16

shiftwork) leads to difficulty in fulfilling another role. Strain-based conflict occurs when

strain such as tension, anxiety, irritability, etc. from one role make it difficult to perform

a second role. Behavior-based conflict takes place when behavior (e.g., expectations) in

one role lead to difficulty in meeting the requirements of another role. Greenhaus and

Beutell (1985) focused on potential sources of interrole conflict between work and non­

work only to the extent that aspects of the work role are likely to be directly related to

interrole conflict. In addition, they only focused on interrole conflict with family, not

other non-work roles (e.g., leisure). These researchers proposed a new model of work-

family conflict based on their review, though they did not empirically test the model.

Frone et al. (1992) extended work/family conflict research by empirically

demonstrating that the work-family interface is bi-directional, such that work interferes

with family, and family interferes with work. They found that job stressors (e.g., lack of

autonomy, work pressure, and role ambiguity) predicted work-to-family conflict and

family stressors (including both parental stressors and marital stressors) predicted family-

to-work conflict. In addition, family-to-work conflict was positively related to job

distress and depression, although work-to-family conflict did not predict depression or

family distress.

Using a time-lagged design, Grandey and Cropanzano (1999) developed and

tested a model of work/family conflict based on the conservation of resources model.

Their model identified aspects of employees’ work and home situations that led to work

and family role stress, and work-to-family conflict and family-to-work conflict. Based on

a survey among university professors, they found that age and tenure predicted work-to-

family conflict such that younger individuals report more work-to-family conflict. In
 17

addition, marital status, number of children at home, and gender explained 22% of the

variance in family role stress, and 26% of the variance in family-to-work conflict such

that those with more children, unmarried individuals, and women reported more family

role stress and family-to-work conflict.

With regard to outcomes of work/family conflict, Grandey and Cropanzano found

that work-to-family conflict predicted job distress, and family-to-work conflict predicted

family distress. Job distress, but not family distress, significantly predicted turnover

intentions, life distress, and poor physical health. Although Grandey and Cropanzano

found that having children at home is a significant predictor of work/family conflict,

other studies have found that the children's age is an important determinant regarding

one’s ability to meet work and family demands (Bedeian, Burke, & Moffett, 1988;

Voydanoff, 1988). For example, having younger children (pre-school-age) is more

closely related to work/family conflict than having older children.

Although empirically tested models of work/family conflict have purported a

number of antecedents, much more research has focused on outcomes related to

work/family conflict. Two recent meta-analyses integrated and summarized the findings

regarding work/family conflict outcomes. First, Kossek and Ozeki (1998) published a

meta-analysis of two of the most frequently studied outcomes: job satisfaction and life

satisfaction. Their study found that higher levels of work/family conflict are associated

with decreases in both job and life satisfaction. Secondly, Allen et al. (2000) reviewed

previous research regarding outcomes of work/family conflict, and included many more

outcome variables than Kossek and Ozeki (1998).

 18

Allen et al. (2000) categorized work/family conflict outcomes into three groups:

work-related, non-work-related, and stress-related outcomes. In terms of work-related

outcomes, higher levels of work/family conflict are associated with decreased job

satisfaction, decreased career satisfaction, decreased organizational commitment,

increased absenteeism and turnover intentions, and decreased job performance. Non­

work related outcomes of work/family conflict are marital dissatisfaction, life

dissatisfaction, leisure dissatisfaction, and family dissatisfaction (Allen et al., 2000;

Kossek & Ozeki, 1998). Stress-related outcomes include general psychological strain,

somatic and physical symptom complaints, depression, substance abuse, burnout, work-

related stress, and family-related stress (Adams, King & King, 1996; Allen et al., 2000;

Burke, 1988; Frone et al., 1992; Greenhaus & Beutell, 1985; Netemeyer. Boles, &

McMurrian, 1996; Thomas & Ganster, 1995).

A few studies to date have examined whether there are gender differences in

perceptions of work/family conflict. Duxbury & Higgins (1991) conducted an

exploratory study to examine gender differences in antecedents and outcomes of

work/family conflict. Specifically, they tested hypothesized gender differences in each

path of the model of work-family conflict articulated by Kopelman et al. (1983). In this

model, work conflict and family conflict are related constructs that each lead to interrole

conflict, and in turn lead to job satisfaction and family satisfaction, which each lead to

life satisfaction. Work conflict is the extent to which one experiences incompatible role

pressures within the work domain. Similarly, family conflict is the extent to which one

experiences incompatible role pressures with regard to family. Work-family conflict is a

type of interrole conflict that occurs when one experiences pressures of one role (e.g.,
work) that are discordant with pressures in another role (e.g., family; Kopelman et al.,

1983).

Duxbury & Higgins (1991) tested for differences between men and women in

perceptions of work involvement and work conflict, work involvement with work-family

conflict, family involvement and family conflict, family involvement with work-family

conflict, work expectation and work-family conflict, work expectations with work

conflict, family expectations with family conflict, each type of expectations with work-

family conflict, and so forth.

Their results were mixed, such that there were statistically significant differences

between men and women in 11 of 17 gender comparisons. Little support was found for

their hypotheses of gender differences in predictors of work and family conflict, and the

authors attributed the lack of differences to societal expectations and behavioral norms.

However, they found support for most of their hypotheses regarding predictors of work-

family conflict. As predicted, the relationship between work involvement and work-

family conflict was stronger for women than for men, and similarly the relationship

between family involvement and work-family conflict was stronger for men than for

women. With regard to predictors of quality of work life and quality of family life, all

but one gender hypothesis was supported. For example, the relationship between work

conflict and quality of work life, and the relationship between work-family conflict and

quality of family life was much stronger for men than for women. Among women, the

relationship between work-family conflict and quality of work life is stronger than for

men. However, there were no significant gender differences in the relationship between

family conflict and quality of family life.

 20

Although this study was prompted by the belief that gender roles are changing as

more women enter the workforce and more men assist with family responsibilities,

Duxbury and Higgins (1991) concluded that societal expectations regarding gender role

expectations have not changed as much as expected over the past few decades. However,

this study is now over a decade old, and it is possible that one might find slightly

different results as societal expectations and behavioral norms regarding gender roles

continue to change. For example, the trend toward increases in the number of on-site day

care programs (Rothausen et al., 1998) and use of flexible work arrangements (e.g.,

alternative work schedules and telecommuting) during the 1990s suggest that additional

research is warranted.

More recently. Eagle, Miles, & Icenogle (1997) sought to determine if there are

gender differences in the extent to which employees’ work demands are permitted to

intrude into the family role and vice-versa. These researchers found that work-to-family

conflict was more prevalent than family-to-work conflict, and that there were no

significant differences observed between men and women.

Although only a few studies have examined gender differences in perceptions of

the work-family interface, it seems that perceptions of work/family issues may be related

to gender. As a result, gender is an important variable to include and explore when

conducting research in this area.

Measurement of Work/Family Conflict

In addition to describing relations among work/family conflict and other

variables, a review of empirical research would not be complete without discussing the

measurement of the construct. Measures of work-family conflict have varied from study
 21

to study. For example, some have employed single item measures (e.g., asking the extent

to which work and family interfere with one another; Quinn & Staines, 1979). Other

work/family researchers (e.g.. Burke, 1988) have used lengthy scales. A number of

work/family conflict measures (e.g.. Kopelman et al., 1983; Quinn & Staines, 1979) have

not accounted for the bi-directionality of work/family conflict (i.e., work interferes w ith

family and family interferes with work; Netemeyer et al. 1996) More recently,

researchers (e.g., Carlson, Kacmar, & Williams, 2000; Netemeyer et al., 1996) have

developed scales of work/family conflict that take into account the bi-directional nature

of the construct (i.e., work-to-family conflict, and family-to-work conflict).

Many work/family conflict measures appear to have construct confusion. Some

work/family conflict scale items refer to “home life” or “things I want to do at home,”

whereas others include the words “family life,” “family responsibilities,” and “family or

spouse/partner.” As a result, respondents' perceptions of home may or may not relate

specifically to family. For example, “things I want to do at home” may be interpreted as

household chores or time for one’s self, rather than as specific to family. Secondly, it is

difficult to discern what, exactly, is meant by family. Does family refer to one’s spouse,

one’s children, or perhaps one’s parents? Use of the word “family” may pose problems

for survey respondents who must interpret survey questions, given increasing demands on

employees to deal with issues such as elder care (Hammer, Allen, & Grigsby, 1999;

Parcel, 1999), as well as an increase in the number of unwed couples cohabitating

(Sweet, 1989).

For example, employees may have household responsibilities/chores, actively

pursue personal interests/hobbies. participate in community activities, or participate in

 22

religious/spiritual or activities that interfere with work or with which work interferes. As

a result, employees may experience work/non-work conflict even without having family

responsibilities.

Without empirically investigating the extent to which other aspects of one’s

personal life interfere with work, we cannot say that we understand the complete

work/non-work interface. It is possible that researchers have underestimated the degree

of work/non-work conflict that employees face by only examining “family” in the non­

work domain. This also suggests that a more generic measure, one that assesses the

work/non-work interface more broadly, is warranted.

Previous Work/Life Balance Research

Recently a broader term has emerged in organizations and the popular press to

refer to work/non-work conflict. This new term is work/life balance, and holds promise

for a more inclusive approach to the study of work/non-work conflict compared to

work/family conflict.

The majority of journal articles to date addressing “work/life balance” appear in

trade journals (e.g., Casner-Lotto & Hickey, 1999; Messmer, 1999) rather than empirical

research journals. In a recent article in the trade journal. Business Credit. Messmer

(1999) highlighted work/life balance as a tool that organizations can use to attract quality

job candidates. Specifically, he recommended that employers accommodate a “healthy

work/life balance” (p. 55) because “employees are interested in making their careers

compatible with what is important in their personal lives” (p. 55). Shellenbarger (1997)

found that job candidates now ask questions about work/life balance in first round talks
 23

with employers, as opposed to years past when these questions were saved for the final

round of interviews, or never asked at all.

Casner-Lotto and Hickey (1999) indicated a recent change in employees' values

such that workers have an increased desire for a better quality of life. Specifically, they

found that in the past few years, employees have indicated that they want more time to do

volunteer work and pursue hobbies and leisure activities. In addition, these researchers

reported that workers desire more time to spend with family, and that fewer employees

seem willing to sacrifice their family or personal time in order to advance in their careers.

Jeff Hill and his colleagues (Hill et al., 1998) are among the only researchers to

publish their findings on work/life balance in an empirical research journal. Hill et al.

(1998) sought to examine the influence of a virtual office environment (i.e., telework, or

working from home) on aspects of work and work/life balance at IBM. They conducted a

quasi-experimental study in which they compared a group of employees who work in a

virtual office environment to a group of traditional office workers at IBM. Qualitative

results based on thematic content coding of written survey comments were somewhat

mixed. More employees indicated that working in a virtual office blurred the boundary

between home and work than the number of employees who indicated a virtual office was

supportive of achieving a work/life balance. Based on quantitative results, there was no

main effect in terms of a difference between the virtual office and traditional office

groups on work/life balance.

Although Hill et al. (1998) appear to take a broader approach to the study of

work/non-work conflict by employing the term “work/life balance'' in lieu of

work/family conflict or work/family balance, there is no evidence of construct validity

 24

for work/life balance. The specific measure of work/life balance used in the Hill et al.

(1998) was not identified. Other articles on work/life balance mention the term, but fail

to provide a conceptual definition or a specific measure of the construct (Casner-Lotto &

Hickey, 1999; Messmer, 1999).

Most research to date on work/non-work conflict appears to be deficient in that it

limits conceptualization of employees' non-work lives to family. Most participants in

work/family conflict research are married or have children, and as a result, employees

who are not married and do not have children are usually omitted. Although work/life

balance is a more encompassing term that appears to be growing in popularity when

referring to the work/non-work interface, research to date on work/life balance lacks

sufficient construct development. In other words, theoretical and empirical research

needs to be conducted to understand what work/life balance is, identify what other

variables are related to work/life balance, and what inferences can be drawn as a result of

knowing if an individual is considered to have achieved work/life balance. Therefore, the

purpose of the present research is to fill this gap by first defining work/life balance, and

subsequently developing and initiating validation of an appropriate work/life balance

measure.

Summary

The majority of work/non-work conflict research has focused on work/family

conflict. Previous empirical work/family conflict research developed and tested models

that have identified a number of work- and non-work related antecedents and outcomes

of work/family conflict. Antecedent variables related to work/family conflict include

work involvement, family involvement, amount of time spent at work, amount of time
 25

spent with family, role overload, work stressors, family stressors, age, marital status,

tenure, and number of children at home. Work-related outcomes include job

dissatisfaction, career dissatisfaction, decreased organizational commitment, increased

absenteeism and turnover intentions, and decreased job performance. Non-work related

outcomes of work/family conflict are marital dissatisfaction, life dissatisfaction, leisure

dissatisfaction, family dissatisfaction, and decreased family performance. Stress-related

outcomes include general psychological strain, somatic and physical symptom

complaints, depression, substance abuse, burnout, work-related stress, and family-related

stress. Measures of work/family conflict have varied from study to study, including

single- and multi-item measures. More recently researchers have developed multi­

dimensional scales that measure work-to-family conflict and family-to-work conflict.

Although previous work/non-work conflict research has increased our

understanding of what leads to and results from work/family conflict, our knowledge of

work/non-work conflict is limited to lamily, because research has focused on family in

the non-work domain. More recently, the term work/life balance has emerged in a few

journals and in the popular press. This term seems more inclusive of various activities in

one's personal life and is positively-worded (i.e., balance) rather than negatively-worded

(e.g., conflict).

Work/life balance research to date has not established a conceptual definition or

an appropriate measure of the construct. The purpose of the present study is to fill this

gap in the literature by developing the construct of work/life balance.

 26

CHAPTER II. THE PRESENT STUDY

The primary purpose of the present study is to define and measure the construct of

work/life balance. Work/life balance is a broader construct than work/family conflict

because it encompasses more than just family in the non-work domain. A broader

construct to examine work/non-work conflict will help to close the gap in the existing

literature and provide organizational researchers with a useful, valid measure of work/life

balance to facilitate future research in this area.

Developing any psychological construct is a process that involves defining the

domain of interest, identifying a way to assess the domain, and understanding how the

construct is related to other constructs (Cook & Campbell, 1976; Cronbach & Meehl,

1955). This process is undertaken to assess construct validity (i.e., understanding what is

being measured).

Work/Personal Life Balance Definition

The first step in a construct development study is to develop a conceptual

definition (Cook & Campbell, 1976; Kopelman et al., 1983). In order to effectively do

this, the author undertook an exploratory, preliminary study to better understand the term

“work/personal life balance” and what it means to employees.

A total of 77 employed individuals were interviewed face-to-face or on the phone

about their perceptions of work/life balance by one of three trained researchers. All of

these participants worked in a variety of occupations and types of organizations, and all

but two were employed full-time. The sample of participants also included both

individuals who were married and/or had children, as well as many who were single and

childless. Since previous work/non-work conflict research has typically focused on

 27

employees with immediate family (e.g., spouse and/or children), it was important to

include unmarried as well as childless workers in this study.

Employees interviewed in this study were asked a number of open-ended

questions pertaining to work/life balance (See Appendix A for a complete list of the

open-ended questions). For example, individuals were asked to indicate what the term,

“work/life balance,” means to them, and how they can tell if they have a balance between

their work and personal life or not.

Each interviewer wrote down each individual's responses. A panel of five

industrial organizational psychology graduate students and one practicing industrial-

organizational psychologist reviewed all of the interview responses, and using an

inductive approach, generated a list of content themes that emerged from the data. The

panel members subsequently discussed the results, agreed upon content coding

categories, and coded responses to questions pertaining to the definition of work/life

balance. Because panelists achieved acceptable levels of inter-rater reliability (i.e.,

consistency in coding responses), results were averaged across raters for interpretation.

Based on data obtained in this study, work/life balance was found to mean

different things to different employees. For example, many employees saw balance in

terms of how they spend or allocate their time, such as the amount of time they spend at

work versus doing things outside of work. Other employees said balance referred to their

ability to accomplish both their work and personal goals. In addition to time, many

participants stressed the importance of accomplishing work tasks (e.g., work-related

goals) and/or non-work tasks (e.g.. goals for accomplishing things not related to work).

For example, one respondent defined work/life balance as “efforts directed toward
 28

maintaining a happy and healthy life while pursuing a challenging and satisfying career.”

Other employees defined this balance as “successfully coordinating the time requirements

of my job with my responsibilities to my family - without negative consequence to

either,” “when you are able to leave your job at the office and come home to spend time

doing the things that you enjoy (e.g., hobbies, working out), find time to socialize, and

keep your life in order (e.g., have groceries in the fridge, bills paid, obligations met).”

This preliminary study suggested two dimensions of work/life balance. The first

dimension of work/life balance is time, such as how much time is spent at work,

compared to how much time is spent engaged in other activities. The second dimension

refers to behavior, such as work goal accomplishment, as work/life balance is based one’s

belief that he is able to accomplish what he would like at work and in his personal life.

Previous research on work/family conflict has identified three potential sources of

conflict: time, strain, and behavior (Greenhaus & Beutell, 1985; Carlson et al., 2000).

Time is a salient component in both the preliminary study as well as in previous

empirical research. Although previous work/family research defined behavior in terms of

expectations for how one should act in one role compared to another (Greenhaus &

Beutell, 1985), accomplishing one’s work and/or non-work goals appears to be a

behavior that was salient to employees in the qualitative work/life balance investigation.

As a result, there appears to be similarity between sources of work/non-work conflict

identified in previous research and findings among interviews with employees in the

preliminary study.

In addition to time and goal accomplishment, there are two additional issues that

seem relevant to defining work/life balance. One of these is strain, based on previous
 29

work/non-work conflict research that defined strain as the third source of interrole

conflict (i.e., in addition to time and behavior; Greenhaus & Beutell, 1985). Examples of

strains include anxiety, tension, having to miss important personal activities, and

difficulty focusing attention.

Secondly, the conservation of resources theory suggests that energy should be

considered when defining work/life balance. Although previous research has not

explicitly cited energy (or lack thereof) as a source of interrole conflict, the conservation

of resources model suggests that energy is a salient resource, and therefore a lack of

energy due to work or non-work demands may lead to an increase in perceived stress.

The rationale for including energy in the definition of work/life balance is consistent with

the notion of time; energy is a limited resource and relevant to employees being able to

accomplish work and/or non-work related goals.

In summary, consistent with the conservation of resources model and interviews

with employees about work/life balance, there are four components relevant to defining

work/life balance: time, behavior (e.g., goal accomplishment), strain, and energy. Each

of these components is congruent with the conservation of resources model (Hobfoll,

1988; 1989). Again, this model purports that stress is a reaction to an environment in

which one is threatened by a potential loss in resources, experiences an actual resource

loss, or fails to gain expected resources. According to the conservation of resources

model, time and energy are examples of salient and limited resources. In addition, strain

experienced as a direct result of a loss of time or energy may lead to a loss of resources or

prevent one from gaining resources.

 30

This model suggests that stress will result if individuals perceive that they lose

resources or devote resources to a task or activity without getting what they expect in

return. It follows that perceived stress is likely to result from a lack of work/life balance,

or that a lack of work/life balance can be considered an occupational stressor. More

specifically, when an individual expends a resource (e.g., time or energy) to accomplish

desired goals at work and in her life outside of work but does not reach both those goals,

she may perceive a lack of work/life balance. To the extent that she strives to regain or

attain balance, she is likely to expend rather than conserve resources. If she actually

loses time or energy, faces a potential loss of time or energy, or fails to gain what she

expects given the resources that she previously expended, then this stress may lead to any

of a number of negative outcomes (e.g., frustration, tension, anxiety, decreased job

satisfaction, etc.).

The cognitive appraisal approach to stress suggests that what constitutes a

“balance” for any one individual is in the eye of the beholder (Lazarus, 1966; Lazarus &

Folkman, 1984). For example, spending fifty hours per week at work and the remainder

of time participating in non-work activities may be considered a balance for one

employee, but not for another. The qualitative research finding that work/life balance

means different things to different employees further underscores this point.

Work/Personal Life Balance Measurement

The first step in developing a construct is to define the domain of interest. The

second step is to identify a way to assess this domain. As part of the present study, 31

self-report survey items were written incorporating time, energy, strain, and goal

accomplishment to measure work/life balance. These items were written based on a

 31

review of existing measures of work/family conflict and using responses gathered from

the preliminary study. Examples of items including time, energy, strain, and goal

accomplishment, respectively, include “I am happy with the amount of time I spend

doing activities not related to work,” “My personal life drains me of the energy I need to

do my job,” “I have to miss out on important personal activities because of my work,”

and “I am able to accomplish what I would like in both my personal and work lives.”

The present study involved two stages of scale development to assess the internal

consistency reliability and validity of these new items. In the first step, the items were

administered to and rated by a panel of subject matter experts in a pilot study to

determine the extent to which these items are consistent with the conceptual definition of

the work/life balance construct (i.e., content validity). Based on subject matter expert

ratings of each item, scale items that were perceived to be relevant to the work/life

balance construct were administered during the second step: a mail survey including the

work/life balance scale and other measures administered to a heterogeneous sample of

employees. Data obtained from this survey were used to assess validity for this new

measure, specifically by examining the psychometric properties of the scale and its

relations to other measures.

Based on the conservation of resources theory and results of the preliminary

study, I generated a number of hypotheses regarding the relations between work/life

balance and other constructs. These hypotheses were tested to assess the construct

validity of the work/life balance measure. Construct validity assesses the extent to which

the scale measures what it purports to measure (i.e., the extent to which the measure

assesses “work/life balance.”) Specifically, construct validity was established by

 32

examining convergent, discriminant, and criterion-related validity (Cronbach, 1971;

Cronbach & Meehl, 1955). Evidence for convergent validity was evaluated by

examining the extent to which work/life balance is empirically related to similar

constructs, such as work/family conflict and other stressors (e.g., time pressure and role

overload). Evidence for criterion-related validity was established by demonstrating the

extent to which work/life balance predicts a number of related outcomes or strains (e.g.,

feelings of overall job stress, job satisfaction, life satisfaction, organizational

commitment, and turnover intentions).

Hypotheses

Hypothesis 1. Work/life balance is strongly related to work/family conflict.

Assessment of convergent validity is an important step in the construct validation

process (Cook & Campbell, 1976; Guion. 1998). Convergent validity refers to finding

strong relationships among similar constructs. Because work/life balance and

work/family conflict are both defined as stressors that employees may face when trying

to manage their work and non-work responsibilities, measures of the two constructs

should be empirically related.

Hypothesis 2: Work/life balance is negatively related to role overload.

According to role theory, each role that an individual has includes a set of

behaviors to be performed (Kahn et al., 1964). The more roles one takes on or is

responsible for, the more role-related tasks or goals exist that need to be accomplished.

Role overload refers to too many behaviors that need to be performed as part of one’s

role(s). Work/life balance is the perception that one can accomplish both his work and

non-work goals, and these goals likely include the behavior in each role set. The more
 33

tasks one must complete as part of a role (role overload), the less likely he is to perceive a

balance between work and his/her personal life.

Hypothesis 3: Work/life balance is negatively related to perceptions of time

pressure.

Time pressure is a specific stressor in which individuals perceive that they have a

limited amount of time to accomplish a task or set of tasks. Based on the conservation of

resources model, employees who perceive more time pressure will perceive a lower

degree of work/life balance.

Hypothesis 4: Work/life balance is positively related to employees ' time

management behavior.

Employees who are better at managing their time will be better at allocating their

time in accordance with their priorities, and therefore more effective in terms of

conserv ing time as a limited resource. This is consistent with Adams and Jex’s (1999)

finding that time management behaviors were directly and indirectly related to

work/family conflict. Because work/life balance concerns one’s ability to accomplish

work and non-work goals, and effective time management behavior is positively related

to time and task accomplishment, work/life balance should be positively correlated with

employees’ ability to manage his/her time effectively.

Hypothesis 5: Work/life balance is negatively related to perceptions of overall job

stress.

Based on the conservation of resources theory, difficulty in achieving work/life

balance is purported to be a specific job stressor. As a result, lower perceptions of

work/life balance should correspond with greater feelings of overall job stress.
 34

Hypothesis 6: Work/life balance is positively related to job satisfaction.

Job satisfaction refers to the affective evaluation one has about his/her job

(Cranny, Smith, & Stone, 1992). Based on the definition of work/life balance, part of the

construct refers to accomplishing one’s goals at work. To the extent that one is able to do

so, and/or their job does not interfere with being able to accomplish non-work goals, an

employee is more likely to be satisfied with his/her job. Based on previous work/family

conflict research that has demonstrated a relationship between work/family conflict and

job satisfaction (e.g., Allen et al., 2000; Kossek & Ozeki, 1998), a similar relationship is

expected between work/life balance and job satisfaction.

Hypothesis 7: Work/life balance is positively related to overall life satisfaction.

Similar to job satisfaction, life satisfaction is defined as the summary affective

evaluation one holds about his/her life in general. Previous research on life satisfaction

considers this construct to be an indicator of overall quality of life (Duxbury & Higgins,

1991; Sekaran, 1985). Work/life balance literature has indicated that many employees

want both a fulfilling career and fulfilling personal life (Shellenbarger, 1997). Therefore,

for employees who desire both a career and a satisfying personal life, employees who are

able to achieve a work/life balance are more likely to be satisfied with their lives in

general.

Hypothesis 8: Work/life balance is positively related to organizational

commitment.

Previous research has indicated that the desire for both a fulfilling career and

fulfilling personal life is desired among employees (Shellenbarger, 1997). Organizational

commitment refers to an employee’s desire to join and remain with an organization

(Mathieu & Farr, 1991). Employees who can obtain the desired balance between work
 35

and their personal life will be more committed to the organization that employs them

compared to employees who do not perceive that they have a work/life balance. In

addition, previous work/family conflict research has found a negative relationship

between work/family conflict and organizational commitment.

Hypothesis 9: Work/life balance is negatively related to turnover intentions.

Intention to quit an organization refers to an employee’s plan to leave an

organization. According to Mobley, Horner, and Hollingsworth's (1978) model of

turnover, many steps are involved in an employee's decision to leave an organization and

follow through (i.e., actually carry out the behavior). An employee's intention to quit is

the last step in this process before actually leaving, and therefore a better indicator of

turnover than other variables (e.g., job satisfaction, organizational commitment). This

prediction is also consistent with the rationale for the expected relationship between

work/life balance and organizational commitment.

Hypothesis 10: There will be evidence of discriminant validity between work/life

balance and work/family conflict.

Work/life balance is a more general construct that work/family conflict because

personal life is more inclusive of non-work activities than work/family conflict, which

only focuses on the negative impact of work on family and vice-versa. To the extent that

other aspects of one’s personal life (i.e., salient activities or commitments in addition to

family) can affect one's work life or be affected by work, a more inclusive measure of the

interface between one's work and personal life will be empirically distinct from a more

narrow measure of work/non-work conflict (e.g., work/family conflict). Although

work/life balance and work/family conflict are expected to be empirically similar, I also
 36

hypothesize that work/life balance will account for some variance in other stressors and

strains that is not accounted for by work/family conflict.

 37

CHAPTER III. METHOD

The purpose of the present study was to develop and validate a measure of

work/life balance for use in organizational research. Work/life balance items were

written based on data gathered from interviews with employees, suggestions by a panel of

work family researchers, a review of existing measures of work/family conflict, and the

construct definition. First, these items were assessed with regard to content validity by a

panel of subject matter experts (SMEs). That is, these raters assessed the extent to which

each item seemed relevant to the construct of work/life balance, based on the conceptual

definition (i.e., a lack of work/personal life balance is a potential occupational stressor

involving time, energy, strain, and goal accomplishment). In addition, SMEs were given

an opportunity to generate additional items that they perceive as relevant to the construct.

Items that were judged by multiple raters as relevant to the construct domain were

incorporated in a larger survey administered to a heterogeneous sample of employees.

Data collected using the survey were used to assess the psychometric properties of the

scale using two samples: an initial sample, and a holdout sample. The hypotheses

presented earlier were tested to establish evidence of construct validity for the new'

work/life balance measure.

Study One - Work/Life Balance Scale Development

Participants

A total of fifteen (7 male and 8 female) industrial-organizational psychology

graduate students, faculty, and practitioners, all of which have experience developing

psychological measures, participated in this phase of the scale development process.

 38

Procedure

Using data collected as part of an exploratory, qualitative study, the

recommendations by a panel of work/family researchers (MacDermid et al., 2000), and a

review of existing measures of work/family conflict (e.g., Carlson et al., 2000;

Netemeyer et al., 1996), the author wrote a total of 31 work/life balance scale items.

Following the recommendations of previous researchers to examine both negative and

positive aspects of the work/personal life interface (e.g., Duxbury & Higgins, 1991;

MacDermid et al., 2000), these items sought to measure employees' perceptions of the

extent to which work interferes with personal life, personal life interferes with work, as

well as to what extent work enhances personal life, and personal life enhances work.

Items incorporated aspects of time, energy, strain, and behavior based on the conceptual

definition of work/life balance.

Each SME was given a paper and pencil survey listing the conceptual definition

of work/life balance, the 31 items, and a rating scale. Using the same procedure as

Netemeyer et al. (1996), SMEs rated each item based on the extent to which it is “2=Very

Representative,” “l=Somewhat Representative,” or “0=Not Representative” of the

work/life balance content domain. The SMEs were also given an opportunity to generate

additional work/life balance items.

Measures

Work/personal life balance scale. The preliminary work/life balance scale

included 31 items written to assess perceptions of work/life balance as they relate to time,

behavior, strain, and energy. A complete list of these items is presented in Appendix B.
 39

Item discrimination ability. Four items from Paulhus’s (1984; 1988) Balanced

Inventory of Desirable Responding were included among the work/life balance items to

verify the SMEs’ ability to identify and rate the new scale items appropriately. The

social desirability items included “I never cover up my mistakes,” “Sometimes I drive

faster than the speed limit,” “I dropped litter on the street at least once in my life,” and “I

sometimes tell lies if I have to.” Item rating responses from any raters who indicated that

at least one of these four items was “somewhat relevant” or “very relevant" to the

construct of work/life balance were removed from further analyses. In other words,

ratings were only retained from subject matter experts who were able to discriminate

between items purported to be relevant to work/life balance and those that were not.

Results

I discarded responses from three SMEs who identified at least one of the social

desirability items as somewhat or very representative of the work/life balance construct.

Items for which all of the twelve remaining raters indicated that the items were somewhat

or very representative were retained for use in the second study; 7 items (Item 1,3, 14,

15, 17, 25, and 31 as listed in Appendix B) were discarded. All items retained for

subsequent scale validation had mean item ratings of M = 2.6 or higher on the 3-point

item rating scale. Two raters suggested slight wording changes to two of the items. The

24 items that remained as a result of the content rating exercise were incorporated in

Study Two. A list of all items generated and retained for inclusion in Study Two appears

in Table 1.
 40

Table 1

Work/Life Balance Scale Items Retained after Content Validation Study

Item
1. I struggle with trying to juggle both my work and non-work responsibilities.
2. My personal life drains me of the energy I need to do my job.
3. I feel overwhelmed when I try to balance my work and personal life.
4. My work suffers because of everything going on in my personal life.
5. I am too tired to be effective at work because of things I have going on in my personal
life.
6. My job gives me energy to pursue activities outside of work that are important to me.
7. I have difficulty scheduling vacation time because of my workload.
8. I am unable to relax at home because I am preoccupied with my work.
9. I am happy with the amount of time I spend doing activities not related to work.
10. Because of my job, I am in a better mood at home.
11.1 often have to make difficult choices between my work and my personal life
12.1 have to put aspects of my personal life “on hold” because of my work.
13.1 am able to accomplish what I would like in both my personal and work lives.
14. I am in a better mood at work because of everything I have going for me in my
personal life.
15.1 often neglect my personal needs because of the demands of my work.
16. When I am at work. I worry about things I need to do outside of work.
17. I have difficulty getting my work done because I am preoccupied with personal
matters at work.
18. My personal life suffers because of my work.
19.1 have to miss out on important personal activities due to the amount of time I spend
doing work.
20. My personal life gives me the energy to do my job.
21.1 feel that I allocate appropriate amounts of time to both work and non-work activities.
22.1 make personal sacrifices to get work done.
23.1 come home from work too tired to do things I would like to do.
24. My job makes it difficult to maintain the kind of personal life I would like.
 41

Study Two

Participants and Procedure

An eight-page survey was sent via mail to 10,206 members of the American

Management Association (AMA) in February, 2001. The AMA is a large, national

professional association consisting of over 700,000 members nationwide that agreed to

sponsor this study to learn more about work/life balance perceptions among their

membership. Accompanying the survey was a cover letter written by the AMA Director

of Research (see Appendix C). No incentive, other than availability of survey results,

was offered to participants in exchange for completing the survey.

AMA members are employed in a variety of organizations, departments and

industries across the United States, and serve as a heterogeneous sample of U. S.

employees. Although the majority (i.e., 70%) of AMA members are male, females were

disproportionately sampled for inclusion in the present study. Of the 10,206 managers

sampled, 50% were female and 50% were male. An equal sample of males and females

was sought in the present study because previous work/non-work conflict research has

found some gender differences in paths of work/family conflict antecedents and

outcomes (Duxbury & Higgins, 1991). By oversampling females, 1 anticipated having

more females included in the final sample, thus providing enough statistical power to

detect gender differences if indeed there were any.

A total of 545 questionnaires were completed and returned for inclusion in this

next phase of scale development. 482 surveys were returned upon the initial survey

mailing (4% response rate). To increase the response rate, a second copy of the survey

was sent to 1,000 managers. This group of n=l ,000 was randomly selected among the
 42

group of non-respondents. A total of 63 managers returned the survey following the

second mailing, yielding a total sample size of N=545, and an overall survey response

rate of 5%. Due to missing data, 540 questionnaires could be used for analysis.

The sample of 540 usable surveys was divided approximately in half in order to

create a holdout sample for cross-validation of the work/life balance scale. As a result,

267 (49.4% of total sample) respondents were included in an initial sample, and the

remaining 273 respondents (50.6% of total sample) were included in the holdout sample.

The sample of respondents consisted of 55% males and 45% females, which is

statistically different from the expected frequencies based on the survey mailing to a

sample of 50% males and 50% females (x“ (1) = 5.81, p < .05); slightly more males

responded than females. On average, respondents were between 45 and 49 years old.

Forty-nine percent of respondents have been employed by the same organization for at

least 10 years; 30% employed for 5-9 years; and 22% for less than 5 years. On average,

managers reported working a total of approximately 52 hours per week (M = 52.3, SD =

5, median = 50). The majority of respondents were married (77%). In addition, the

majority of respondents (79%) had obtained at least a bachelor’s degree, and had been

working for the same organization for at least five years (78.1%). Fifty-five percent of

respondents indicated that they have at least one child living in their household at least

50% of the time; 17% reported having no children living with them, and 28% left this

item blank. Among those who answered the question, 76% of the sample had at least one

child living with them. A complete breakdown of sample demographic characteristics is

presented in Table 2.
 43

Table 2

Demographic Characteristics of the Total Study Sample (N=540)

Variable Categories Percentage (%'

Gender Male 55.2
Female 44.8

Age 18 to 24 years 0.2

25 to 29 years 0.2
30 to 34 years 3.3
35 to 39 years 10.7
40 to 44 years 16.7
45 to 49 years 19.6
50 to 54 years 19.6
55 to 59 years 18.3
60 to 64 years 8.9
65 years or older 2.4

Relationship Status Married, living together 77.0

Married living apart 1.2
Not married, but living with partner 5.6
Not married, not living with partner,
but involved in committed
relationship 5.6
Not married, and not involved in
committed relationship 10.7

Number of children
living in household at least
50% of the time 0 23.6
1 29.0
2 31.0
3 13.6
4 or more 2.8

Primary responsibility
for elder care Yes 12.0
No 88.0
 44

Table 2 (continued)

Demographic Characteristics of the Total Study Sample (N=540)

Variable Categories Percentage (%)

Total work hours per week Less than 40 2.6
40 to 49 31.8
50 to 59 39.5
60 to 69 18.9
70 to 79 5.9
More than 80 1.3

Highest Education Level

Obtained High school graduate or less 3.2
Some college 13.5
4-year college degree 4.3
Some post-graduate study 26.5
Masters Degree 14.4
Some post-masters study 27.7
Doctoral/professional degree 5.0

Job Tenure Less than 1 year 14.9

1 to 2 years 26.2
3 to 4 years 27.1
5 to 9 years 19.7
10 to 14 years 6.9
15 years or more 5.2

Organization Tenure Less than 1 year 1.9

1 to 2 years 7.1
3 to 4 years 13.0
5 to 9 years 29.5
10 to 14 years 19.5
15 years or more 29.1

Current Salary Under $35K 3.8

$35K to $49K 12.9
$50K to $74K 31.6
$75K to $99K 23.6
$100K to$149K 18.9
$150K or more 9.3
 45

Table 2 (continued)

Demographic Characteristics of the Total Study Sample (N=540)

Variable Categories Percentage (%)

Management Level Level 1 (CEO) 2.6
Level 2 (Report to CEO) 10.8
Level 3 24.7
Level 4 27.9
Level 5 19.2
Level 6 7.2
Level 7 3.4
Level 8 or lower 4.2

Primary Area of
Responsibility Communications 3.4
Finance 12.7
General and Administrative Services 8.9
General Management 15.8
Human Resources 16.0
Information Systems & Technology 7.0
Insurance & Risk Management 3.0
International 0.2
Manufacturing 9.3
Marketing 8.0
Purchasing 3.4
Research & Development 5.9
Sales 4.6
Transportation & Distribution 1.7

Position Description Strategic, policy-making 23.4

Selling and/or marketing 8.7
Product/service development or
delivery 25.1
Administrative or support functions 42.8
 46

With the exception of gender (as females were intentionally oversampled in the

present study), these demographic characteristics were representative of the AMA

membership population as a whole, as well as the sample of 10,206 to whom the survey

was mailed.

Measures

Work/life balance. The survey included a 24-item measure based on the

conceptual definition of work/life balance, a review of existing work/family conflict

measures, and data gathered in response to open-ended questions asked as part of a

preliminary, exploratory study. Work/life balance items included in this survey consisted

of the items retained from Study One; a complete list of these items is presented in

Appendix D.

Previous work/family researchers (e.g., MacDermid et al., 2000) have

recommended the use of a time-based stem so that all respondents have the same frame

of reference for responding to the items. In the survey, respondents were asked to

indicate the frequency with which they have felt a particular way during the last three

months using a five-point scale: “Not at all," “Rarely," “Sometimes,” “Often,” and

“Almost All the Time." Respondents had the option of choosing a “Not Applicable"

response if the item did not pertain to them. “Not Applicable" responses were coded as

missing values for analysis purposes.

In addition, a single, one-item measure of w'ork/life balance was included in the

survey to serve as an initial validity check for the work/life balance scale items.

Respondents indicated the extent to which they agreed or disagreed with the statement,

“In general I feel that I have an adequate balance between my work and personal life,”
 47

using a 5-point Likert-type rating scale ranging from “Strongly Agree” to “Strongly

Disagree.”

Work/family conflict. Work/family conflict was assessed using Netemeyer et

al.’s (1996) bi-directional measure of work-to-family conflict and family-to-work

conflict. The work-to-family conflict scale consisted of 5 items assessing the negative

impact of work on family. Respondents indicated the extent to which they agreed or

disagreed with each item using a 5-point Likert-type rating scale, ranging from “Strongly

Agree” to “Strongly Disagree.” The internal consistency reliability (coefficient alpha) for

this scale was a = .90 in both the initial and holdout sample. A work-to-family conflict

scale score was computed for each respondent by taking the mean across all 5 items.

The family-to-work conflict scale assessed the extent to which one’s family

responsibilities interfere with work, using a separate five-item scale and the same

response format as the work-to-family items. The coefficient alphas for family-to-work

conflict were a = .88 and a = .89 in the initial and holdout samples, respectively.

Family-to-work scale scores were computed for each respondent by taking the mean

across all 5 items. See Appendix E for the work-to-family conflict and family-to-work

conflict measures.

Role overload. Role overload refers to having too much to do and was assessed

using a 3-item scale, adapted from Beehr, Walsh, & Taylor (1976). These items include

“I am given enough time to do what is expected of me on my job,” “It often seems like 1

have too much work for one person to do,” and “There is always more work to be done.”

Participants were asked to respond to these items using a 5-point Likert scale, ranging

from “Strongly Agree” to “Strongly Disagree.” An overall scale score for each
 48

individual was computed by taking the average score across al, three items after reverse­

coding the negatively-worded item. Coefficient alphas were a = .70 for the initial

sample, and a = .67 for the holdout sample, respectively.

Overall job stress. Overall job stress was assessed using the 15-item Stress-in-

General (S1G) scale (Stanton, et al., in press; Fisher et al., 2000). This measure is an

overall rather than a facet measure of job stress, and a measure that is widely applicable

rather than tied to specific stressors or strains. The SIG assesses two dimensions of

overall job stress. The first factor, entitled "Pressure," contains 7 items, such as

“Demanding,” “Hectic,” and “Many Things Stressful." The second factor, entitled

“ Threat,” includes 8 items, such as “Nerve-Wracking, “Hassled," and “Overwhelming.”

Previous research (e.g., Stanton et al., in press) found that when the SIG scale is

scored as two separate dimensions, both factors were correlated with a one-item measure

of general stress as well as specific facets of stress as measured by the Job Stress Index

(Sandman, 1992; Sandman & Smith, 1988). In a subsequent study, Stanton et al. (in

press) examined the relation between the SIG and physiological measures, and found

evidence to support the general or overall nature of the Pressure subscale and its

connection to chronic strains. In a follow-up study, Fisher et al. (2000) demonstrated that

both dimensions were measurably distinct from negative affectivity. Responses to each

of the 15-items on the SIG measure were recorded as “Y” for “Yes,” “N” for “No” and

“?” for “Cannot decide.” Scoring for the SIG is as follows: “N” = 0, “?” = 1, “Y” = 2 for

the positively worded items, and “N”=2, “?”=1 and “Y”=0 for the negatively-worded

items. A separate sum for each dimension of the scale was calculated. Higher scores

indicate higher levels of stress.

 49

The SIG has demonstrated adequate internal consistency reliability for each

dimension in previous research. Coefficient alphas obtained for the Pressure and Threat

dimensions in the present study were consistent with previous findings (aprcssurc = .82, .78

and a,hreat = -82, .80 in the initial and holdout samples, respectively). See Appendix F for

the complete SIG measure.

Time management behavior. Time management behavior was assessed using

Macan, Shahani, Dipboye, and Phillips’ (1990) 33-item self-report inventory of time

management behaviors. This scale measures three dimensions of time management

behaviors, including “goal setting and prioritizing,” “planning and scheduling," and

“preference for organization” (Adams & Jex, 1997).

Goal setting and prioritizing is defined by Macan (1994) as “the setting of goals

concerning what the person wants or needs to accomplish and the prioritizing of tasks

necessary to achieve these goals.” (p. 391.) Examples of goal setting and prioritizing

items include “I set short-term goals for what 1 want to accomplish in a few' days or

weeks" and “I finish top priority tasks before going on to less important ones." She

defined the second dimension, planning and scheduling, as “the behaviors typically

associated with managing time, such as making lists, scheduling, and planning.” (p. 391).

Examples of items measuring this dimension of time management include “I schedule

activities at least a week in advance" and “1 make a list of things to do each day and

check off each task as it is completed.” The third aspect of time management is

preference for organization, which refers to “both a general organized approach to work

projects as well as maintenance of an organized work environment" (Macan, 1994, p.

391 ). A few of the preference for organization items are “At the end of the workday I
 50

leave a clear, we 11-organized work space,” and “I can find the things I need for my work

more easily when my work space is messy and disorganized than when it is neat and

organized.”

Previous research has demonstrated acceptable internal consistency reliability

(Adams & Jex, 1999; Macan, 1994; Macan et al., 1990) of each of these three

dimensions. Coefficient alphas obtained in the initial and holdout samples, respectively,

of the present study were a=.84 and a=.81 for setting goals and priorities, a = .73 and

a=,75 for planning and scheduling, and a =.72 and a = .76 for preference for

organization. With regard to validity, self-report ratings of time management behavior

using this scale have correlated highly with supervisors, coworkers, and friends’ reports

of time management behaviors (Macan, 1994). ,n addition, Macan’s (1994) test of a

process model of time management found that engaging in time management behaviors is

related to lower levels of tension and higher levels of job satisfaction.

Time pressure. Time pressure is a particular stressor defined as the extent to

which employees are pressured to complete their work within a fixed timeframe. Time

pressure was assessed using a 5-item scale developed for use in the present study in

which respondents indicated the extent to which they agreed of disagreed with each item

using a 5-point Likert-type rating scale. The five time pressure items were “I am often

pressured for time at work,” “There are too many things to be done at once,” “I often face

unwanted interruptions of my work,” “I do not have enough time to meet my

obligations,” and “I have too many responsibilities as part of my job." Participants

responded to each of the five time pressure items using a five-point scale ranging from
 51

“Strongly Agree” to “Strongly Disagree." Internal consistency reliability for these five

items was a = .84 in both the initial and holdout samples.

Job satisfaction. Employees’ overall level of job satisfaction was assessed using

the Job-in-General (JIG) scale. The JIG scale is a global rather than a facet measure of

job satisfaction. An overall rather than a facet measure was chosen for use in this study

because I was interested in how work/life balance relates to job satisfaction as a whole,

rather than how work/life balance relates to specific aspects of one's job.

This scale consists of eighteen items to which respondents indicate the whether

each word or phrase describes their current job situation. The response scale for the JIG

scale is identical to the SIG: “Y” for “Yes,” “N” for “No,” and “?” for “Cannot decide.”

Scoring for the JIG is as follows: “N” = 0, “?” = 1, “Y” = 3 for the positively worded

items, and “N”=3, “?”=1 and “Y”=0 for the negatively-worded items. An overall score is

calculated for each individual by summing one's responses across the eighteen items.

JIG scores may range of 0 to 54, with higher scores indicating higher levels of overall job

satisfaction. In order to reduce the impact of missing data on summed scale scores,

listwise deletion of cases was employed such that no job satisfaction score was calculated

for any individual who left more than three items blank (Balzer et al., 1997).

Coefficient alphas measuring the internal consistency of the job satisfaction items

in the present study were a = .92 for the initial sample and a = .90 for the holdout

sample, respectively. This is consistent with internal consistency reliability estimates

obtained in previous research, in which alpha has ranged from .91 to .95 (Ironson et al.,

1989). As evidence for construct validity for the JIG, this measure is highly correlated

with other measures of job satisfaction (r ranges from .67 to .80, median = .76; Ironson et
 52

al., 1989) and moderately correlated with overall job stress as measured by the SIG scale

(r = -. 15 to -.49; median = -.39). See Appendix G for the complete measure.

Life satisfaction. Life satisfaction is the degree to which individuals judge the

overall quality of their life favorably (Diener, Emmons, Larsen, & Griffin, 1985; Pavot &

Diener, 1993). Overall life satisfaction was assessed using the 5-item Satisfaction with

Life Scale (Diener et al., 1985). This measure has been used frequently in quality-of-life

research to assess individuals’ overall satisfaction with life, and has demonstrated

appropriate levels of reliability and validity (Hart, 1999; Pavot & Diener, 1993). For

example, the estimate of internal consistency reliability (coefficient alpha) obtained for

this measure in previous research was a = .88, and the measure was correlated r = .49

with job satisfaction, r = .56 with non-work satisfaction, r = -.34 with work hassles, and r

= -.31 with the personality trait, neuroticism.

Participants responded to the five items of life satisfaction using a 5-point Likert-

type scale, ranging from “Strongly Agree" to “Strongly Disagree." Life satisfaction

scores for each individual were computed by taking the arithmetic mean across the five

items. Higher scores indicate higher levels of life satisfaction. Coefficient alphas

(internal consistency) in the present study were a = .85 for both the initial and holdout

samples. See Appendix FI for the complete life satisfaction measure.

Organizational commitment. Affective organizational commitment refers to the

extent to which employees like and feel a part of the organization for which they work.

This dimension of organizational commitment was measured using a 6-item scale

developed by Mowday, Steers, and Porter (1979). Participants were asked to indicate the

extent to which they agree or disagree with each organizational commitment item using a
 53

5-point, Likert-type scale ranging from “Strongly Agree” to “Strongly Disagree.” An

organizational commitment score was computed for each individual by computing the

arithmetic mean of the six items. Higher scores indicate higher levels of organizational

commitment.

This measure of organizational commitment has been used extensively in

previous research, including a recent study by McClough, Rogelberg, Fisher, &

Bachiochi (1998). This measure has demonstrated acceptable reliability and validity. For

example, estimates of internal consistency reliability (coefficient alpha) range from a =

.80 to a = .86, and the measure is highly correlated with job satisfaction (r = .63) and

turnover intentions (r = -.62; McClough et al., 1998). Internal consistency reliability

estimates obtained in the present study (coefficient alpha) were a = .83 and a = .82 for

the initial and holdout samples, respectively. See Appendix I for the complete

organizational commitment measure.

Turnover intentions. Three items measuring employees’ intentions to quit were

included on the survey based on the Michigan Organizational Assessment Questionnaire

by Seashore, Lawler, Mirvis, and Cammann (1982). These items include “I often think

of quitting my job,” “1 will probably look for a new job within the next 12 months,” and

“I would like to leave my present job.” Participants were asked to respond to these items

using a 5-point Likert scale, ranging from “Strongly Agree” to “Strongly Disagree.” An

overall intention to quit scale score was computed by taking the average across

individuals’ responses to all three items. Previous research has used two of these three

items to measure intentions to quit, with a coefficient alpha of a = .91 (McClough et al.,

1998). Empirical research using the 2-item measure has found a negative relation with
 53

job satisfaction (r = -.56 to -.64) and organizational commitment (r = -.52 to -.62; Fisher

et al., 2000; McClough et al., 1998). To improve the internal scale properties (e.g.,

internal consistency reliability), a third item (“I would like to leave my present job.”) was

added to the scale. Internal consistency reliability estimates (coefficient alpha) obtained

in the present study were

a = .91 and a = .90 for the initial and holdout samples, respectively.

Demographic characteristics. For sample classification purposes and examination

of moderator variables, a number of items pertaining to demographic characteristics were

included in the survey. These include gender, age, marital/relationship status,

organizational tenure, job tenure, current salary, primary area of responsibility,

management level, average number of total work hours per week, average number of

hours spent working per week at home, number and ages of children, and responsibility

for elder care. See Appendix J for the exact demographic measures.

Data Analyses

Data analyses for this study were conducted in three phases. The first two phases

involved using data from the initial sample. The third phase involved using data from the

holdout sample. The first phase focused on item and scale analysis to gather empirical

evidence to assess the work/life balance scale. The second phase of data analysis was

conducted to assess construct validity evidence by testing the hypotheses made about the

relations between work/life balance and other constructs using data gathered in the initial

sample. The third phase employed a hold-out sample to cross-validate the scale

psychometric properties (e.g., factor structure, reliability, and validity of the scale) with

the results obtained during Phases One and Two.

 55

Holdout sample. Prior to conducting any statistical analyses, 273 respondents

(i.e., approximately half the sample) were randomly selected from the entire sample.

This group of respondents was set aside used as the holdout sample for cross-validation

purposes. The remaining 267 respondents were used in Phase One (scale item analysis)

and Phase Two (initial scale validity assessment). Consistent findings obtained in a

holdout sample would increase the generalizability of results by adjusting for random

sampling error.

The sample size for the holdout sample is considered more than sufficient for

conducting factor analyses and tests of the hypotheses. A minimum of 200 participants

and a participant-to-item ratio of 5 have been suggested for obtaining reliable factor

loading estimates in exploratory factor analysis (Cattell, 1978; Ford, MacCallum. & Tait,

1986; Tabachnick & Fidell, 1996). The participant-to-item ratio in each of these two

samples exceeded ten to one, and was considered sufficient for obtaining reliable

parameter estimates.

Phase one. For the first phase of the data analysis, item means and standard

deviations were reported. In addition, an exploratory' factor analysis using maximum

likelihood extraction and an oblique rotation of the factors was conducted. The

eigenvalues, rotated factor loadings, and item communalities were reported.

After completing the exploratory factor analysis, I computed the corrected item-

total correlations (which are the bivariate correlations of each scale item compared to the

composite of all scale items that is corrected for autocorrelation; Nunnally, 1978), and

Cronbach's coefficient alpha as a measure of internal consistency reliability. Based on

scale development recommendations by Stanton et al. (1999) and Clark & Watson
 56

(1995), I compared each scale item to external scale information as a preliminary

“validity check.” Specifically, I computed the bivariate correlations between each

work/life balance scale item and a one-item overall work/life balance item (i.e., “In

general I feel that I have an adequate balance between my work and personal life”).

Using the results obtained across all of those analyses, I tested the latent factor structure

of the scale using partial confirmatory factor analysis (PCFA; Bollen, 1989). In a partial

confirmatory factor analysis, the latent factor structure (i.e., which items are to load on

which factors) is specified a priori, but the item loadings are freed for estimation. This

step enabled me to determine the best-fitting model with regard to the number of latent

factors underlying the observed variables (i.e., the scale items). The results I obtained in

each of these steps to evaluate the work/life balance scale items are presented in the next

chapter.

Phase two. Examination of convergent and criterion-related validity evidence

obtained from the initial sample (N=267) provided initial evidence for construct validity

of work/life balance. Assessment of the correlations among work/life balance and

work/family conflict measures provided evidence of convergent validity. Strong,

statistically significant correlations (e.g., above r = .60) would provide evidence that the

two scales are measuring similar constructs. Additional evidence of convergent validity

was provided by comparing perceptions of work/life balance to other stressors, such as

time pressure and role overload. Criterion-related validity was evaluated by testing the

research hypotheses regarding the extent to which work/life balance was related to other,

conceptually related variables, including occupational strains, including feelings of

 57

overall job stress, life satisfaction, job satisfaction, organizational commitment, and

turnover intentions.

Finally, the work/life balance scale was compared to the work/family conflict

measure to determine whether there was an empirical difference between work/life

balance and work/family conflict in terms of the amount of variance shared with each

variable and related measures. This was accomplished by determining the extent to

which work/life balance accounted for a significant portion of unique variance in other

variables after controlling for variance in common with work/family conflict. This was

tested using hierarchical regression analysis in which work/family conflict was entered

into the regression equation as a predictor variable, and each outcome (e.g., life

satisfaction, job satisfaction, overall job stress) was used as a separate criterion or

dependent variable. Work/life balance was entered into the equation as a predictor in the

second step, after work/family conflict. A statistically significant F-ratio obtained by

comparing the difference or change between the multiple R-squared obtained from the

first step and the multiple R-squared found in the second step would indicate that

work/life balance accounts for variance in the criterion variable that work/family conflict

does not.

Phase three. The data analysis in the third phase was conducted to cross-validate

the scale development process undertaken in Phase One and the initial scale validity

evidence obtained in Phase Two. Internal consistency reliability estimates (Cronbach's

alpha), corrected item-total correlations, and correlations with a one-item overall measure

of work/life balance were computed for each of the work/life balance scale items using

data retained in the holdout sample.

 58

Next, I followed the recommendation of Reeve & Smith (2001) to use a

convergent evidence approach by conducting IRT item analyses in addition to the

classical test theory item analyses already conducted to gather additional information to

assess the psychometric properties of the work/life balance scale items. For example, I

computed the a parameter and test information function (TIF) for each scale item

obtained in an item response theory (IRT) analysis. I he a parameter measures how well

an item discriminates on the latent trait being assessed. The TIF provides evidence of the

sensitivity of the scale, which refers to how much information is provided by the scale at

all levels of the trait being measured. The purpose of examining the TIF is to determine

the extent to which the scale items measure the work/life balance content domain across

all levels of work/life balance. The IRT analysis requires a large sample size in order to

obtain reliable item response parameter estimates. Therefore, data from the initial and

holdout samples were combined for this analysis.

Data analyses conducted in Phase Two were repeated in Phase Three to further

assess the construct validity of the work/life balance measure using the holdout sample.

Consistent findings in Phase Three compared to results obtained in Phase Two would

serve as additional evidence for the validity of the scale.

In addition to evaluating the validity of the scale, a number of exploratory

analyses were conducted using demographic variables to determine if there were any

meaningful differences between groups of respondents in the initial sample based on

demographic differences. These data analyses involved computing t-tests and using

analysis of variance to compare mean differences in perceptions of work/life balance

across different demographic subgroups. The full sample (N=540), rather than the initial
 59

and holdout samples individually, was used for the exploratory analyses in order to

increase statistical power (i.e., the probability of finding significant effects if, in fact, they

exist).
 60

CHAPTER IV. RESULTS

Initial Sample

Demographic Information

The demographic characteristics of the initial sample are presented in Table 3.

Among these 267 respondents, 57% were male and 43% were female. The proportion of

males and females in this sample was not significantly different from the proportion

obtained in the overall sample (yw2 (1) = .40, p > .05). The majority of respondents were

between 35 and 54 years old. The majority of respondents were married (79%). In

addition, the majority of respondents (81%) had obtained at least a bachelor’s degree, and

had been working for the same organization for at least five years (82%).

Scale Assessment

Exploratory factor analysis. Item means and standard deviations for the initial

item pool of the work/life balance scale are presented in Table 4. Higher means indicate

that respondents report having experienced that situation more frequently. In most cases,

items with higher means are purported to indicate lower levels of work/life balance. The

inter-correlations among the scale items are given in Table 5.

An exploratory factor analysis (principal components analysis, PCA) was

conducted using the correlation matrix as input to assess the dimensionality among the 24

work/life balance scale items (Tabachnick & Fidell, 1996). The purpose of PCA is to

extract the maximum amount of variance with each component from the data, with the

assumption that the observed variables (i.e., items) are a linear combination of some
 61

Table 3

Demographic Characteristics of the Initial Sample (N=267)

Variable Categories Percentage (%)

Gender Male 56.9
Female 43.1

Age 18 to 24 years 0.4

25 to 29 years 4.5
30 to 34 years 10.5
35 to 39 years 15.0
40 to 44 years 21.7
45 to 49 years 18.0
50 to 54 years 19.1
55 to 59 years 9.4
60 to 64 years 1.5
65 years or older 0.4

Relationship Status Married, living together 78.0

Married living apart 0.8
Not married, but living with partner 4.2
Not married, not living with partner, 5.8
but involved in committed
relationship
Not married, and not involved in 11.2
committed relationship

Number of children 24.1

living in household at least
50% of the time 0
1 28.7
2 31.8
3 12.3
4 or more 3.0

Primary responsibility
for elder care Yes 9.5
No 90.5
 62

Table 3 (continued)

Demographic Characteristics of the Initial Sample (N=267)

Variable Categories Percentage (%)

Total work hours per week Less than 40 2.3
40 to 49 35.5
50 to 59 35.9
60 to 69 20.2
70 to 79 4.6
More than 80 1.5

Highest Education Level

Obtained High school graduate or less 2.6
Some college 12.0
4-year college degree 4.5
Some post-graduate study 25.6
Masters Degree 13.9
Some post-masters study 28.9
Doctoral/professional degree 6.4

Job Tenure Less than 1 year 15.0

1 to 2 years 23.3
3 to 4 years 28.2
5 to 9 years 21.1
10 to 14 years 7.1
15 years or more 5.3

Organization Tenure Less than 1 year 1.5

1 to 2 years 5.6
3 to 4 years 10.5
5 to 9 years 28.9
10 to 14 years 21.1
15 years or more 32.3
Table 4

Work/Life Balance Scale Item Means and Standard Deviations - Initial Sample

Item Mean SD
1. I struggle with trying to juggle both my work and non-work responsibilities. 3.32 1.00
2. My personal life drains me of the energy I need to do my job. 2.03 0.84
3. I feel overwhelmed when I try to balance my work and personal life. 2.60 0.93
4. My work suffers because of everything going on in my personal life. 1.88 0.77
5. I am too tired to be effective at work because of things I have going on in my personal life. 1.76 0.70
6. My job gives me energy to pursue activities outside of work that are important to me. 2.71 1.12
7. I have difficulty scheduling vacation time because of my workload. 2.94 1.16
8. I am unable to relax at home because I am preoccupied with my work. 2.73 1.00
9. I am happy with the amount of time I spend doing activities not related to work. 3.14 1.10
10. Because of my job, I am in a better mood at home. 2.77 1.04
11.1 often have to make difficult choices between my work and my personal life 2.70 0.94
12.1 have to put aspects of my personal life “on hold" because of my work. 2.59 0.99
13. I am able to accomplish what I would like in both my personal and work lives. 3.22 1.03
14. 1 am in a better mood at work because of every thing 1 have going for me in my personal life. 3.53 1.01
15.1 often neglect my personal needs because of the demands of my work. 2.84 1.02
16. When I am at work, I worry' about things I need to do outside of work. 2.55 0.86
17. I have difficulty getting my work done because I am preoccupied with personal matters at work. 1.93 0.70
18. My personal life suffers because of my work. 2.54 1.03
19.1 have to miss out on important personal activities due to the amount of time I spend doing work. 2.41 0.95
20. My personal life gives me the energy to do my job. 3.37 1.00
21.1 feel that I allocate appropriate amounts of time to both work and non-work activities. 3.30 1.17
22.1 make personal sacrifices to get work done. 3.14 0.91
23.1 come home from work too tired to do things I would like to do. 3.25 0.94
24. Mv iob makes it difficult to maintain the kind of personal life I would like. 2.82 1.16
O\
LU
Table 5

Correlation Matrix of Work/Life Balance Scale Items - Initial Sample

1 2 3 4 5 6 7 8 9 10 11 12 13
1 1.00
2 .37 1.00
3 .63 .40 1.00
4 .36 .58 .42 1.00
5 .29 .59 .36 .65 1.00
6 -.19 -.25 -.29 -.18 -.22 1.00
7 .35 .07 .34 .15 .07 -.16 1.00
8 .45 .24 .53 .23 .23 -.30 .56 1.00
9 -.46 -.25 -.46 -.18 -.26 .34 -.42 -.51 1.00
10 -.22 -.18 -.33 -.23 -.19 .53 -.27 -.35 .37 1.00
11 .56 .27 .50 .35 .21 -.19 .52 .53 -.43 -.28 1.00
12 .50 .24 .43 .28 .17 -.27 .54 .55 -.50 -.34 .64 1.00
13 -.46 -.38 -.50 -.38 -.34 .43 -.38 -.51 .61 .44 -.48 -.57 1.00
14 -.27 -.28 -.33 -.24 -.26 .34 -.19 -.31 .31 .37 -.28 -.27 .48

Note. Row and column labels refer to the numbering of Work/Life Balance scale items. (N=267)

2
Table 5 (continued)

Correlation Matrix of Work/Life Balance Scale Items - Initial Sample

I 2 3 4 5 6 7 8 9 10 H 12 13 14

15 .51 .22 .52 .24 .16 -.29 .53 .64 -.56 -.37 .55 .71 -.56 -.32
16 .28 .32 .26 .34 .34 -.27 .02 .10 -.18 -.18 .26 .25 -.26 -.19
17 .23 .44 .25 .50 .53 -.27 .05 .13 -.22 -.22 .28 .25 -.35 -.28
18 .53 .26 .54 .28 .21 -.31 .53 .59 -.55 -.38 .59 .72 -.59 -.23
19 .48 .23 .48 .24 .13 -.26 .51 .53 -.49 -.37 .64 .69 -.49 -.22
20 -.30 -.34 -.39 -.28 -.35 .39 -.24 -.32 .37 .40 -.34 -.35 .51 .71
21 -.47 -.21 -.51 -.26 -.24 .37 -.42 -.54 .62 .39 -.47 -.58 .65 .38
22 .39 .19 .40 .22 .13 -.15 .48 .53 -.45 -.31 .50 .65 -.50 -.28
23 .44 .33 .47 .28 .31 -.39 .37 .50 -.50 -.43 .41 .53 -.58 -.38
24 .50 .34 .51 .31 .27 -.38 .47 .57 -.57 -.44 .56 .73 -.60 -.37

Note. Row and column labels refer to the numbering of Work/Life Balance scale items. (N=267)

O'
Table 5 (continued)

Correlation Matrix of Work/Life Balance Scale Items - Initial Sample

15 16 17 18 19 20 21 22 23 24
15 1.00
16 .19 1.00
17 .19 .50 1.00
18 .74 .16 .23 1.00
19 .67 .16 .21 .76 1.00
20 -.35 -.18 -.37 -.35 -.29 1.00
21 -.59 -.26 -.28 -.64 -.55 .43 1.00
22 .65 .06 .11 .65 .59 -.25 -.50 1.00
23 .61 .29 .30 .57 .48 -.39 -.55 .48 1.00
24 .68 .26 .30 .71 .70 -.42 -.60 .58 .63 1.00

Note. Row and column labels refer to the numbering of Work/Life Balance scale items. (N=267)
 67

underlying construct (Ford et al., 1986). The first principal component is the linear

combination of observed variables that accounts for the most variance among the items.

The second component is computed from the residual correlations after forming the first

component, and is the linear combination of observed variables that accounts for the most

variance uncorrelated with the first component. The process of extracting the maximum

variance from the residual correlations that is uncorrelated with previously computed

components is repeated until accounting for all unique variance. Mathematically, if all

components are retained from a PC A, the observed correlation matrix can be reproduced

from the PCA. As a result, PCA is an exploratory factor analytic tool that is useful for

providing an empirical summary of the data. This method is slightly different from a

factor analysis (i.e., using a principal factor or maximum likelihood factor extraction

method), in which shared variance is estimated with communalities that attempt to

eliminate unique and error variance from the factors (Tabachnick & Fidell, 1996).

However, results obtained from a principal components analysis and principal factor

analysis are very similar and do not usually yield different results upon interpretation.

Prior to conducting the PCA, many of the work/life balance scale items were

recoded so that for all items, higher scores indicated higher levels of work/life balance.

Because the pattern of missing data among work/life balance scale items appeared to be

random rather than systematic, listwise deletion was performed to handle missing data.

Listwise deletion refers to omitting any cases with any missing values from the analysis,

and is a conservative approach to handling missing data. Data from N=242 respondents

were retained for use in the principal components analysis after listwise deletion. This
 68

sample size exceeds the minimum requirement of 200 cases or 5 respondents per item as

recommended by Tabachnick & Fidell (1996).

Two criteria were used to assess the number of components that describe the

correlations among the scale items: eigenvalues greater than one and the scree test (Cattell,

1978). The first step was to examine the eigenvalues obtained for each component.

Eigenvalues are related to the amount of variance accounted for by each component. A

rule of thumb is to retain components with eigenvalues greater than one (Tabachnick &

Fidell, 1996). There were three components with eigenvalues greater than one. These

eigenvalues were 10.4, 2.6, and 1.5, respectively.

The second step was to conduct a scree test. A scree test involves plotting

eigenvalues on the y-axis against the number of components on the x-axis, drawing a line

through each point on the graph, and determining where the line changes direction, or

levels off. The number of components to be retained for interpretation should be equal to

the number of components on the x-axis before the line changes direction. I he scree plot

of eigenvalues is presented in Figure 1. This plot suggests a three-component solution. As

a result, both criteria suggest that three components should be retained for interpretation.

As indicated earlier, the amount of variance that each observed variable (i.e., item)

contributes is one (Tabachnick & Fidell, 1996; Cattell, 1978). Because there were twenty-

four work/life balance scale items, the total variance is 24. Total variance accounted for by

each component includes shared (common) variance and unique variance. The three
 69

Figure 1

Scree Plot of Eigenvalues from Principal Components Analysis Among Work/Life

Balance Scale Items - Initial Sample

Component Number
 70

components that emerged from the PCA accounted for 60.6% of the total variance. The

first component accounted for 43.4%, the second accounted for 10.8%, and the third

component accounted for 6.4% of the total variance. Common or shared variance across

the set of items is computed by summing the communality estimates obtained for each

item, and is 14.55. The proportion of total variance accounted for by each component is

computed by dividing the common variance (sum of communality estimates, 14.55) by the

total variance (i.e., number of items, 24).

Because 1 predicted that the factors would correlate with one another, I performed

an oblique (direct oblimin) rotation of the components. An oblique rotation assumes that

the components are correlated. However, it is important to note that the components are

not necessarily correlated with one another when an oblique rotation is used. If results

indicate that the components are uncorrelated, then a more simple, orthogonal rotation can

be used (Tabachnick & Fidell, 1996).

Results of the PCA with the oblique rotation yielded three interpretable and

correlated components that converged in six iterations. The PCA pattern matrix (A) and

item communalities are presented in Table 6, and the component correlation matrix ((p) is

shown in Table 7. Because the pattern matrix was easily interpretable (only one item

cross-loaded on two components) and the three components were correlated with one

another (r range from .31 to .44), the oblique rotation seemed appropriate, and there was no

need to conduct a second principal component analysis using an orthogonal (varimax)

rotation.
 71

Table 6

Pattern Matrix of Initial Work/Life Balance Items - Initial Sample

Factor Loadings Communality

Item 1 2 3

18 .87 .02 -.01 .76

19 .87 -.02 -.06 .70
12 .85 .03 -.03 .71
15 .84 -.09 .07 .73
22 .80 -.11 -.02 .59
7 .76 -.20 .00 .52
11 .75 .18 -.12 .60
8 .74 -.04 .09 .61
24 .72 .09 .17 .71
1 .61 .32 -.13 .52
21 .59 .05 .29 .62
3 .56 .29 .04 .53
9 .55 .01 .29 .53
23 .52 .10 .29 .56
13 .47 .16 .40 .66
5 -.05 .81 .06 .66
4 .13 .80 -.12 .66
2 .05 .75 .01 .59
17 -.08 .70 .20 .58
16 -.02 .61 .05 .38
14 -.04 .06 .78 .62
20 .00 .17 .74 .66
6 .01 .03 .70 .51
10 .18 -.09 .66 .54

Note. Principal components extraction method with oblique rotation method. (N=242)
1 Refers to the work/life balance scale item number as listed in Table 4.
 72

Table 7

Component Correlation Matrix of Work/Life Balance Scale Items - Initial Sample

Component
I 2 3

1 1.00

2 .31 1.00

3 .44 .33 1.00

Note. Principal components extraction method with oblique rotation method. (N=242)
 73

The pattern matrix obtained in this analysis indicated a clear three-component

solution. Component loadings obtained based on an oblique rotation are measures of the

unique relationship between the component (i.e., latent variable) and the observed variable

(i.e., item; Tabachnick & Fidell, 1996). Higher loadings indicate stronger relationships,

and the higher the loading, the more the observed variable is a good indicator of the latent

factor. A minimum loading of .40 was interpreted as evidence that the item loaded on a

particular component.

The item communality estimates specified the proportion of variance accounted for

by the components. Communalities for items in the first component ranged from .52 to

.76. Communalities for items in the second component range from .38 to .66.

Communalities for items in the third component range from .51 to .66. These

communalities indicated that a large proportion of the variance is explained by these three

components.

The component pattern matrix was easy to interpret because the component

loadings for each item exceeded .40 on one component and were much less than .40 on

each of the remaining components. The only exception to this was Item 13 (i.e., “1 am able

to accomplish what I would like in both my personal and work lives"), which had a loading

of .47 on the first component and .40 on the third component. These results indicated that

this item cross-loaded on two components, which made it difficult to discern whether Item

13 was more similar to items on the first component or items on the third component.

Taking advantage of the exploratory' nature of principal components analysis, I

conducted a second PCA using an oblique rotation, but omitted Item 13 from the analysis.

The resulting component pattern matrix and communalities are presented in Table 8, and
 74

Table 8

Pattern Matrix of Work/Life Balance Items after Removing Item 13 - Initial Sample

Factor Loadings Communality

Item 1 2 3

18 .87 .02 -.02 .76

19 .86 -.02 -.05 .70
12 .85 .03 -.03 .71
15 .84 -.08 .07 .73
22 .80 -.11 -.02 .59
7 .75 -.20 .01 .52
11 .75 .18 -.12 .60
8 .74 -.03 .10 .61
24 .72 .09 .17 .71
1 .61 .32 -.13 .52
21 .60 .05 .28 .61
3 .56 .29 .04 .53
9 .56 .02 .27 .52
23 .53 .10 .29 .56
5 -.05 .81 .07 .66
4 .13 .80 -.12 .66
2 .06 .75 .01 .59
17 -.07 .70 .20 .58
16 -.02 .61 .06 .39
14 -.02 .07 .78 .63
20 .01 .17 .74 .67
6 .02 .03 .70 .52
10 .19 -.09 .66 .55

Note. Principal components extraction method with oblique rotation method. (N=242)
1 Refers to the work/life balance scale item number as listed in Table 4.
 75

Table 9

Component Correlation Matrix of Work/Life Balance Scale Items after Removing

Item 13 - Initial Sample

Component
I 2 3

1 1.00

2 .30 1.00

3 .43 .32 1.00

Note. Principal components extraction method with oblique rotation method. (N=242)
 76

the component correlation matrix appears in Table 9. These results were consistent with

the results obtained in the initial PCA, although none of the remaining items cross-loaded

on multiple components, and this factor solution was preferable to the initial results

obtained with all 24 items.

After removing the cross-loading item, fourteen scale items remained in the first

component. Examples of items on this component with high factor loadings included “My

personal life suffers because of my work,’* and “1 have to miss out on important personal

activities because of my work.” The content of these items reflected the extent to which

work interferes with personal life, and therefore 1 labeled the component worA: interference

with personal life (WIPL). The five items in the second component indicated the opposite

direction of work/personal life interference compared to items in the first component.

Examples of these items included “My work suffers because of everything going on in my

personal life,” and “I have difficulty getting my work done because 1 am preoccupied with

personal matters.” Specifically, these items depicted the extent to which one's personal

life interferes with work. As a result, 1 labeled this component personal life interference

with work (PLIW). The four items in the third component involved positive effects of

one's work on personal life or vice versa, the extent to which one's personal life enhances

work. Therefore, this I labeled the third component work/personal life enhancement

(WPLE). Lower levels of interference (WIPL and PLIW) were interpreted as higher levels

of work/life balance. Similarly, higher levels of work/personal life enhancement were

considered to be associated with higher levels of work/life balance.

Upon examination of the content of the fourteen items in the first component, some

items specifically indicated a directional relationship between work and personal life such
 77

that work interferes with personal life. For example, the item with the highest loading on

this factor was “My personal life suffers because of my work." However, this directional

influence was less apparent in a few of the items (e.g., “I struggle with trying to juggle my

work and non-work responsibilities.”) To explore whether these fourteen items reflected

more than one underlying dimension, I conducted an additional principal components

analysis among these fourteen items. Results yielded a clear one-component solution that

accounted for 58% of the variance among the items. The eigenvalue for this component

was 8.17, and the scree plot also indicated a one-component solution. The component

matrix and communalities for these fifteen items are presented in Table 10.

Additional item and scale analyses. After performing the exploratory factor

analysis to understand the underlying dimensions among the work/life balance scale items,

the next step was to compute additional item and scale statistical analyses to evaluate the

work/life balance scale items. First, 1 computed corrected item-total correlations among

work/life balance scale items on each dimension identified by the exploratory factor

analysis. Corrected item-total correlations measure the bivariate relationship between each

scale item and the total of all items in the scale and are corrected for autocorrelation

(Nunnally, 1978). In other words, an item-total correlation is a part-whole comparison of

how one particular item is related to the whole set of items.

Three sets of corrected item-total correlations (one for each of the three scale

dimensions) were computed and the results are presented in Table 11. Previous research

has suggested that corrected item-total correlations above r = .45 indicate that the
 78

Table 10

Component Matrix and Communalities Among Fourteen Work Interference with Personal

Life (WIPL) Items - Initial Sample

Item Component Matrix Communality

18 .87 .75

15 .85 .71

24 .83 .69

12 .83 .69

19 .81 .66

8 .77 .59

21 .76 .57

22 .75 .56

11 .74 .55

23 .71 .50

9 .71 .50

3 .68 .47

1 .68 .46

7 .67 .45

Note. Principal components extraction method (N=242)

 79

Table 11

Corrected Item-Total Correlations for Work/Life Balance Scale Items - Initial Sample

WIPL PLIW WPLE

Item ITC Item ITC Item ITC

18 .83 5 .69 6 .75

15 .81 4 .67 10 .73
12 .79 17 .63 14 .71
24 .79 2 .61 20 .68
19 .77 16 .44
8 .73
21 .71
11 .70
22 .70
9 .66
23 66
3 .64
1 .63
7 .62

Note. WIPL = Work Interference with Personal Life; PLIW = b Personal Life
Interference with Work; WPLE = Work/Personal Life Enhancement
 80

relationship between any one scale item and the whole scale is acceptable for scale

development (Ironson et al., 1989). All work/life balance items exceeded the r = .45

criterion, except Item 16 (on the PLIW dimension). The corrected item-total correlation

for Item 16 was r = .44, which was much lower than the corrected item-total correlations

for other items on the PLIW scale.

Next, I computed internal consistency reliability estimates (coefficient alpha) for

each of the three dimensions of work/life balance. Coefficient alpha is a statistic that

provides an estimate of internal consistency reliability. Specifically, coefficient alpha

measures the extent to which items are interrelated and have high communalities (Cortina,

1993). A coefficient alpha above a = .70 is considered to be a rule of thumb as an

acceptable level of internal consistency reliability for a scale. However, it is important to

note that the number of items in a scale affects the size of the alpha coefficient such that

scales with more items are more likely to the have higher coefficient alphas (Cortina,

1993). Therefore the level of coefficient alpha should be interpreted with the number of

scale items in mind. For example, a four-item scale w ith a coefficient alpha of .80 would

indicate more internal consistency among the items than a fifteen-item scale with the same

reliability estimate.

Among the work/life balance items, coefficient alphas obtained in the initial sample

were a = .94 for Work Interference with Personal Life (WIPL; 14 items), a = .81 for

Personal Life Interference with Work (PLIW; 5 items), and a = .77 for Work/Personal Life

Enhancement (WPLE; 4 items). The reliability analysis also indicated that removing Item

16 (“When I am at work, I worry about things I need to do outside of work") would

increase the reliability estimate from a = .81 to a = .83. The low corrected item-total
 81

correlation for Item 16 as well as the increase in scale reliability after removing Item 16

from the PLIW scale suggest that Item 16 should be removed from the PLIW scale prior to

performing additional analyses.

Next, I conducted an initial item validity investigation by correlating each work/life

balance scale item with a one-item measure of overall work/life balance. That is,

respondents rated a single item, “In general I feel that I have an adequate balance between

my work and personal life.” Table 12 presents the bivariate correlations between each of

the work/life balance scale items and the one-item measure. Correlations between items on

the WIPL dimension and overall work/life balance were moderate to high in magnitude.

Correlations between WIPL items were also moderately correlated with overall work/life

balance. However, personal life to work interference (PLIW) correlations with overall

work/life balance were much lower than any of the other items, although still significantly

related with r’s between .20 and .26. This suggested that work/life balance was more

similar to work interference with personal life than any other dimension.

Confirmatory factor analysis. Using the results of the exploratory factor analysis

obtained in the previous step, I tested a measurement model of the work/life balance scale

item factor structure using partial confirmatory' factor analysis (PCFA). In PCFA, the

latent factor structure is specified a priori, but the item loadings are freed for estimation

(Bollen, 1989). The purpose of this analysis was to apply a confirmatory approach to

testing the factor structure of the work/life balance scale and to obtain fit statistics

useful for assessing how well the model fit the data. However, due to the exploratory

nature of this study, it was premature to specify item path coefficients.

 82

Table 12

Correlations Between Each Work/Life Balance Scale Item and One-Item Work/Life

Balance Measure - Initial Sample

Item r Item r Item r

21 .69 4 -.26 20 .48
18 -.60 17 -.25 6 .40
3 -.56 2 -.22 14 .40
24 -.56 5 -.21 10 .39
15 -.55 16 -.20
19 -.55
12 -.52
9 .49
11 -.49
23 -.49
1 .48
8 -.48
22 -.44
7 -.41

Note. All correlations significant at p < .05.

 83

Results of the exploratory factor analysis suggested that a three-factor model would fit the

data best. I tested a three-factor model using a maximum likelihood procedure and the

covariance matrix as input using AMOS Graphics 4.0 software (Arbuckle, 1999).

Although the initial item pool contained 24 items, two items (13 and 16) were removed

based on results obtained from the exploratory factor analysis and additional item and scale

analyses. Therefore, the initial model 1 tested was a three-factor model containing a total of

22 items: 14 items on the WIPL dimension (Items 1,3, 7, 8, 9, 11, 12, 15, 18, 19, 21, 22,

23. and 24), 4 items on the PLIW dimension (Items 2, 4, 5, 17), and 4 items on the WPLE

dimension (Items 6, 10. 14, and 20). No items were allowed to cross-load on any other

dimension.

Four model fit statistics were examined to assess the fit of the model: the Tucker-

Lewis Index (TLI; Tucker & Lewis, 1973: Bentler & Bonett, 1980), the comparative fit

index (CFI; Bentler, 1990), the normed fit index (NFL McDonald and Marsh, 1990), and

the root mean square error of approximation (RMSEA; Steiger, 1990) based on the

recommendations of Hu and Bentler (1995), Medsker, Williams, and Holahan (1994), and

Hu & Bentler (1995). The TLI, CFI, and NFI are relative indices that are used to compare

alternative models using the same data. TLI, CFI, and NFI values below .90 indicate poor

model fit (Bentler & Bonett, 1980). Compared to widely used fit indices such as the NFI,

the TLI and CFI are robust across sample sizes (Hu & Bentler, 1995; Marsh. Balia, &

McDonald. 1988).

The fourth model fit index examined in this study was the RMSEA (Steiger, 1990.

The RMSEA is a fit index for comparing non-nested models that permits the relative

ordering of models from best fitting to worst fitting within a single data set, and represents
 84

a discrepancy per degree of freedom test (Maruyama, 1998). MacCallum (1998) and Hu

and Bentler (1995) recommend the use of the RMSEA in addition to other fit statistics

when evaluating model fit. According to Steiger, (1990) RMSEA values below .10

indicate good fit and values below .05 indicate very good fit. More recently, however,

Browne & Cudeck (1992) claimed that an RMSEA of .08 represents the upper bound of

good fit, and Hu and Bentler (1995) claimed that .06 indicates good fit.

In addition to these fit statistics, it is customary to report chi-square and degrees of

freedom for a PCFA model (Maruyama, 1998). However, the chi-square as a test of model

fit is not informative because the chi-square value is greatly affected by sample size. In

other words, the chi-square statistic is usually statistically significant because the sample

size required to conduct the analysis is typically large. Fit indices obtained from test of the

initial three-factor model are presented in Table 13.

Although the TLI, CFI, and NFI results obtained from the PCFA indicate good

model fit, the RMSEA was higher than desirable to indicate good model fit according to

Hu and Bentler (1995). To improve the fit of the model to the data, two steps were taken.

First, I removed Item 3 from the WIPL dimension. Because Items 1 and 3 were strongly

correlated (r = .63), examination of the modification indices produced by the PCFA

analysis indicated that the error terms for Items 1 and 3 were highly correlated, and the

item content was very similar (i.e., “I struggle with trying to juggle both my work and non­

work responsibilities" and “I feel overwhelmed when I try to balance my work and

personal life," respectively), it seemed appropriate to remove one of these items to improve

model fit. I chose item 3 for omission because the PCFA factor loading for Item 3 (.60)

was lower than the factor loading for Item 1 (.63). The fit indices for the three-factor
 85

model having removed Item 3 are presented in Table 13. These results indicate that the

model fit did improve by removing Item 3.

The second step taken to improve model fit was to remove the constraint that errors

among a pair of items are uncorrelated. In other words, the errors between Items 6 and

Item 10 were allowed to covary. These changes were made not only because the items

were empirically related, but also because they were semantically similar and therefore

theoretically related to one another. A third PCFA was conducted to test the fit of the

model to the data once this constraint had been removed. The fit indices for this model are

also presented in Table 13. Although allowing the errors from items 6 and 10 to

covary improved mode, fit, I removed the same constraint from items 9 and 21. The fit

indices found by testing this model appear in Table 13. With relative fit indices well

above .90 and an RMSEA of .06, these results suggest that this 21-item, three-factor model

with two pairs of errors allowed to covary fits the data well.

In order to rule out the alternative that a one- or two-factor mode, fit the data better

than the three-factor model, two additional PCFAs were conducted for comparative

purposes, and results obtained from these analyses are also presented in Table 13. As

expected, examination of these fit indices suggested that the three-factor model was

superior to the one-or two-factor model. A figure depicting the best-fitting model and

including the path coefficients (i.e., factor loadings) for each item is presented in Figure 2.
 86

Table 13

Fit Statistics for Confirmatory Factor Analysis Models - Initial Sample

Model Description Chi-Square df TLI CFI NFI RMSEA

1 Three-factor model, 514.08 206 .977 .981 .969 .075

22 items

2 Three-factor model, 426.21 186 .981 .984 .973 .070

21 items

3 Three-factor model, 386.29 185 .984 .987 .975 .064

21 items, one error
covariance

4 Three-factor model, 364.95 184 .985 .988 .977 .061

21 items, two error
covariances

5 Two-factor model, 21 559.69 186 .970 .976 .964 .087

items, two error
covariances

6 One-factor model, 21 868.79 187 .945 .956 .944 .117

items, two error
covariances
Note. TLI = Tucker-Lewis Index; CFI = Comparative Fit Index; NFI = Normed Fit Index;
RMSEA = Root Mean Square Error of Approximation
 87

Figure 2

Final Work/Life Balance Scale Confirmatory Factor Analysis Model - Initial Sample

Item 1 L
Item 7 L
Item 8—-

Item 9j
item 11 •

Item 12
WIPL
Item 15
Item 18
Item 19

Item 21
Item 22
Item 23

Item 24
-0.5:
Item 2

PLIW Item 4

Item 5

Item 17
Item 6~

Item 10
WPLE
Item 14

Item 20
 88

Summary

The initial pool of 24 work/life balance items was reduced to a final set of 21 items

based on the results of the item and scale analyses conducted among data from the initial

sample. The resulting work/life balance scale was multi-dimensional, with 13 items

measuring work interference with personal life (WIPL), 4 items assessing personal life

interference with work (PLIW), and four items assessing work/personal life enhancement

(WPLE). These dimensions demonstrated adequate internal consistency reliability,

although the convergent and criterion-related validity evidence regarding the relation

between each work/life balance dimension and other variables will be assessed based on

results obtained during Phase Two.

Validity Evidence

Scale descriptive statistics. Scale means and standard deviations for all study

variables are presented in Table 14. Scale mean distributions were normal for both WIPL

and WPLE, and somewhat positively skewed for PLIW. indicating that more managers

indicated low levels of personal life interference with work. I tested for mean differences

between the work/life balance scale dimensions using paired-samples t-tests. Results

indicated that managers perceived that their work interfered with their personal life (WIPL)

more than their personal life interfered with their work (PLIW), t (266) = 19.84, p < .01.

In addition, respondents’ mean levels of work/personal life enhancement were significantly

higher than their beliefs that their personal life interfered with their work, even after taking

into account the opposite direction of scaling for the two variables (t (266) = 20.84,

p < .01. There were no mean differences found between WIPL and WPLE.
Table 14

Descriptive Statistics for all Study Variables - Initial Sample

Standard
Scale Mean Deviation
1. WIPL - Work Interference with Personal Life 2.83 0.76
2. PLIW - Personal Life Interference with Work 1.90 0.61
3. WPLE - Work/Personal Life Enhancement 3.10 0.80
4. Overall Work/Personal Life Balance 3.32 1.05
5. WFC - Work-to-Family Conflict 2.84 0.93
6. FWC - Family-to-Work Conflict 2.03 0.74
7. Role Overload 3.61 0.78
8. Stress - Pressure 17.78 3.52
9. Stress - Threat 17.55 4.45
10. Time Pressure 3.45 0.76
11. Time Management Behavior (TMB)- Setting Goals and Priorities 3.64 0.58
12. Time Management Behavior (TMB)- Planning and Organizing 3.53 0.65
13. Time Management Behavior (TMB)- Preference for Organization 3.77 0.54
14. Job Satisfaction 41.90 10.91
15. Life Satisfaction 3.36 0.82
16. Organizational Commitment 3.80 0.74
17. Turnover Intentions 2.45 1.19

oo
o
Table 15

Correlation Matrix of Study Variables - Initial Sample

I 2 3 4 5 6 7 8 9 10 11 12 13 14 15 £6 12
1. WIPE .94
2. PLIW .39** .83
3. WPLE -.57** -.41** .77
4. Overall W/L Bal. -.67** -.29** .54** -
5. WFC .87** .35** -.53** -.60** .90
6. FWC .28** .58** -.29** -.27** .35** .88
7. Role Overload .45** .08 -.27** -.31** .48** .10 .70
8. Stress - Pressure .47** .12* -.30** -.26* ♦ .47** .11 .54** .82
9. Stress - Threat .52** .20** -.44** -.40** .49** .17** .50** .58** .82
10. Time Pressure .55** .22** -.34** -.35** .56** .18** .79* ♦ .61** .57** .84
11. TMB-Goals -.05 -.31** .29** .14* -.07 -.27** -.05 .00 -.20** -.09 .84
12. TMB-Planning -.09 -.24** .24** .12* -.08 -.18** -.02 .02 -.02 -.01 .62** .73
13. TMB-Org. -.41** -.37** .29* * .27** -.36** -.32** -.29** -.16** -.28** -.32** .36** .28** .72
14. Job Sat. -.31** -.28** .52** .31** -.32** -.13* -.14* -.17** -.43** -.23** .25** .15* .10 .91
15. Life Sat. -.55** -.42** .61** .63** -.53** -.35** -.24* -.24* -.42** -.32** .26** .19** .30** .49** .85
16. Org. Comm. -.16** -.21** .44** .22** -.21** -.16** -.10 -.05 -.31** -.14* .23** .12* .11 .57 .36** .83
17. Turnover .25** .21** -.38** -.28** .31** .19** .21** .13* .41** .23** -.14* -.07 -.14* -.61** -.39** -.64* • .90
Note. Coefficient alpha for each scale appears on diagonal. (N=267)
** p<.01
* p < .05

©
©
 91

After gathering initial construct validity evidence for the work/life balance scale

by examining the psychometric properties of the scale and examining the scale mean

distributions, the next step was to examine the extent to which the work/life balance scale

was related to other variables. I tested a series of hypotheses regarding the relations

among the work/life balance dimensions and other variables to assess the convergent and

criterion-related validity of the work/life balance scale. Specifically, 1 correlated each

dimension of work life balance with work-to-family conflict, family-to-work conflict,

role overload, time pressure, the pressure and threat subscales of overall feelings of job

stress, time management behavior (setting goals and priorities, planning and scheduling,

and preference for organization), overall job satisfaction, life satisfaction, organizational

commitment, and turnover intentions. The bivariate correlations obtained during this

analysis and interpreted to test the hypotheses are presented in Table 15.

Hypothesis tests

Each of the hypotheses tested in this study involved assessing the relations

between work/life balance and other variables. In order to test these hypotheses. Lower

levels of WIPL and PLIW (both directions of interference) were interpreted as higher

levels of work/life balance, and higher levels of WPLE (enhancement) corresponded to

higher levels of work/life balance.

Hypothesis 1. In order to assess the evidence of convergent validity of work/life

balance, I compared employees' perceptions of work/life balance to their ratings of work-

to-family conflict (WFC) and family-to-work conflict (FWC). As expected, work/life

balance and work/family conflict were strongly related to one another. In particular,

WIPL and WFC were very highly correlated, and PLIW was fairly strongly related to
 92

FWC. These results indicate evidence for convergent validity between work/life balance

and work/family conflict.

Hypothesis 2. Hypothesis 2 stated that work/life balance would be negatively

related to role overload. Hypothesis 2 was supported for two of the three work/life

balance dimensions: WIPL and WPLE. Contrary to what I predicted, no significant

relationship was found between personal life-to-work interference (PLIW) and role

overload.

Hypothesis 3. Hypothesis 3 stated that work/life balance would be negatively

related to perceptions of time pressure. Results indicated that WIPL was strongly related

to perceptions of time pressure. The other two dimensions of work/life balance were also

significantly related to time pressure in the expected direction, although the magnitude of

the correlations between time pressure and both PLIW and WPLE was not as high as the

r = .55 correlation between time pressure and WIPL.

I lypothesis 4. The fourth hypothesis predicted a positive relationship between

employees' perceptions of work/life balance and time management behavior. All three

dimensions of time management behavior (setting goals and priorities, time management

planning and scheduling, and preference for organization) were correlated with PLIW

and WPLE in the predicted direction. However, WIPL was only related to having a

preference for organization and was not related at all to setting goals and priorities or

planning and scheduling. Although results generally support Hypothesis 3, the lack of a

relationship between WIPL and setting goals and prioritizing was contrary to what I

expected.
 93

Hypothesis 5. Hypothesis 5 predicted a negative relationship between work/life

balance and feelings of overall job stress, or a positive relationship between stress and

interference. WIPL and WPLE were significantly related to both dimensions of overall

job stress. Pressure and Threat. However, PLIW was only related to the threat

dimension, not pressure. The magnitude of the obtained correlations was higher for

threat than pressure, suggesting a stronger relationship between work/life balance and

threat, compared to work/life balance and pressure

Hypothesis 6. As expected, work/life balance (WIPL. PLIW, and WPLE) was

positively related to overall job satisfaction. However, work/personal life enhancement

was much more strongly related to job satisfaction (r = .52) than either of the interference

dimensions, WIPL and PLIW (r = -.31 and -.28, respectively).

Hypothesis 7. Similar to the results for the previous hypothesis, all three

dimensions of work/life balance were related to perceptions of life satisfaction. All three

dimensions were correlated relatively strongly with life satisfaction, although consistent

with the results for overall job satisfaction, WPLE was more strongly related to life

satisfaction than either of the interference dimensions.

Hypothesis 8. Hypothesis 8 predicted that work/life balance would be positively

related to organizational commitment. PLIW was moderately related to affective

organizational commitment, and WPLE was strongly related to organizational

commitment. However, WIPL was only weakly related to commitment.

Hypothesis 9. Hypothesis 9 stated that employees’ perceptions of work/life

balance would be negatively related to their intentions to leave their work organization.

Although the distribution of turnover intentions was positively skewed (i.e., the majority
 94

of employees were not planning to leave their jobs), results were obtained in the

predicted direction for all three work/life balance dimensions. Work/personal life

enhancement was more strongly related to turnover intentions than the other two

dimensions.

Hypothesis 10. The next step was to discern whether there was any evidence for

discriminant validity between work/life balance and work/family conflict. That is, were

they each really measuring two, albeit related, but distinct constructs? To test the

hypothesis that indeed they are empirically distinct constructs (Hypothesis 10), I

conducted a series of hierarchical regression analyses to examine the incremental

variance in each of the criterion variables that work/life balance accounted for above and

beyond work/family conflict.

The results of this analysis are presented in Tables 16 to 26. R2 is the coefficient of

determination, which indicates the proportion of variance in the criterion variable that is

accounted for by the linear combination of the predictor variables. The change in R2

indicates the incremental variance accounted for by additional variables when they are

added to the regression equation. Results of these hierarchical regression analyses

indicated that work/life balance accounted for a significant proportion of the variance in

each of the criterion variables after accounting for work/family conflict, with the

exception of role overload and time pressure. This suggests that, although similar, the

work/life balance construct is measurably distinct from the work/family conflict

construct.
 95

Table 16

Hierarchical Regression Analysis for Work/Family Conflict and Work/Life Balance

Variables Predicting Role Overload - Initial Sample (N = 267)

Variable B SE B p

Step 1

Work-to-Family Conflict .43 .05 .51**

Family-to-Work Conflict -.08 .06 -.07

Step 2

Work-to-Family Conflict .32 .09 .37**

Family-to-Work Conflict -.01 .07 -.01
Work Interference with Personal Life .16 .12 .16
Personal Life Interference with Work -.15 .09 -.11

Work/Personal Life Enhancement -.03 .07 -.03

Note. R2 = .24 for Step 1; AR“ =.01 for Step 2 .

* p<.05
 96

Table 17

Hierarchical Regression Analysis for Work/Family Conflict and Work/Life Balance

Variables Predicting Overall Stress (Pressure) - Initial Sample (N = 267)

Variable B SEB P

Step 1

Work-to-Family Conflict 1.85 .22 .49**

Family-to-Work Conflict -.34 .28 -.07

Step 2
Work-to-Family Conflict .97 .44 .26**

Family-to-Work Conflict -.16 .33 -.03

Work Interference with Personal Life 1.10 .54 .24*

Personal Life Interference with Work -.42 .41 -.07

Work/Personal Life Enhancement -.28 .30 -.06

Note. R2 = .22 for Step 1; AR2 =.02 for Step 2 .
**p<.01
* p < .05
 97

Table 18

Hierarchical Regression Analysis for Work/Family Conflict and Work/Life Balance

Variables Predicting Overall Stress (Threat) - Initial Sample (N = 267)

Variable B SEB ß

Step 1

Work-to-Family Conflict 2.35 .28 .49**

Family-to-Work Conflict -.04 .35 -.01

Step 2
Work-to-Family Conflict .56 .53 .12

Family-to-Work Conflict .06 .40 .01

Work Interference with Personal Life 1.82 .66 .31**

Personal Life Interference with Work -.44 .50 -.06

Work/Personal Life Enhancement -1.22 .37 -.22**

vNote. ,772“=—.23
” . —R —
7TTfor . A ,77 — e.
Step 1 ; AR2 =.07 for Step 2 .
**p<.01
* P< 05
 98

Table 19

Hierarchical Regression Analysis for Work/Family Conflict and Work/Life Balance

Variables Predicting Time Pressure - Initial Sample (N = 267)

Variable B SEB P

Step 1

Work-to-Family Conflict .45 .04 .55**

Family-to-Work Conflict -.01 .06 -.01
Step 2
Work-to-Family Conflict .25 .09 .31**

Family-to-Work Conflict .00 .07 .00

Work Interference with Personal Life .26 .11 .26*

Personal Life Interference with Work -.01 .08 -.01

Work/Personal Life Enhancement -.03 .06 -.04

Note. R2 = .30 for Step 1 ; AR2 =.02 for Step 2 .
**p<.01
* p<.05
 99

Table 20

Hierarchical Regression Analysis for Work/Family Conflict and Work/Life Balance

Variables Predicting Setting Goals and Priorities - Initial Sample (N = 267)

Variable B SEB ß

Step 1

Work-to-Family Conflict .02 .04 .03

Family-to-Work Conflict -.22 .05 -.28**

Step 2
Work-to-Family Conflict .01 .07 .02

Family-to-Work Conflict -.10 .06 -.13

Work Interference with Personal Life .17 .09 .22

Personal Life Interference with Work -.20 .07 -.21**

Work/Personal Life Enhancement .22 .05 .30**

Note. R~ = .08 for Step 1; ARJ =. 10 for Step 2
**p<.01
* p<.05
 100

Table 21

Hierarchical Regression Analysis for Work/Family Conflict and Work/Life Balance

Variables Predicting Planning and Scheduling - Initial Sample (N = 267)

Variable B SEB ß

Step 1

Work-to-Family Conflict -.01 .05 -.02

Family-to-Work Conflict -.16 .06 -.18**
Step 2

Work-to-Family Conflict .06 .09 .09

Family-to-Work Conflict -.06 .07 -.07

Work Interference with Personal Life .03 .11 .03
Personal Life Interference with Work -.16 .08 -.15
Work/Personal Life Enhancement .18 .06 .23**
Note. R- = .04 for Step 1; AR~ =.06 for Step 2 .
**p<.01
* p<.05
 101

Table 22

Hierarchical Regression Analysis for Work/Famity Conflict and Work/Life Balance

Variables Predicting Preference for Organization - Initial Sample (N = 267)

Variable B SEB ß

Step 1

Work-to-Family Conflict -.16 .03 -.28**

Family-to-Work Conflict -.16 .04 -.22**

Step 2
Work-to-Family Conflict .03 .07 .05

Family-to-Work Conflict -.10 .05 -.14

Work Interference with Personal Life -.24 .08 -.34**

Personal Life Interference with Work -.15 .06 -.17**

Work/Personal Life Enhancement .01 .05 .02

■ ----777------ e..-.
, --- 777 --- 77
Note. R2 = . 17 for Step 1; AR2 =.06 for Step 2 .
vt

**p<.01
* p < .05
 102

Table 23

Hierarchical Regression Analysis for Work/Family Conflict and Work/Life Balance

Variables Predicting Overall Job Satisfaction - Initial Sample (N = 267)

Variable B SEB P

Step 1

Work-to-Family Conflict -3.75 .73 -.32**

Family-to-Work Conflict -.22 .93 -.01

Step 2
Work-to-Family Conflict -2.55 1.29 -.22*

Family-to-Work Conflict 1.90 1.00 .13

Work Interference with Personal Life 2.60 1.62 .18

Personal Life Interference with Work -2.72 1.24 -.15*

Work/Personal Life Enhancement 6.50 .90 .48**

Note. R2 = . 11 for Step 1; AR" =. 18 for Step 2 .
**p<.01
* p<.05
 103

Table 24

Hierarchical Regression Analysis for Work/Family Conflict and Work/Life Balance

Variables Predicting Life Satisfaction - Initial Sample (N = 267)

Variable B SEB ß

Step 1

Work-to-Family Conflict -.42 .05 -.47**

Family-to-Work Conflict -.21 .06 -.19

Step 2

Work-to-Family Conflict -.10 .08 -.11

Family-to-Work Conflict -.11 .06 -.10

Work Interference with Personal Life -.19 .11 -.18

Personal Life Interference with Work -.14 .08 -.11

Work/Personal Life Enhancement .38 .06 .37**

Note. RJ = .32 for Step 1; AR~ =. 14 for Step 2 .
**p< .01
* p<.05
 104

Table 25

Hierarchical Regression Analysis for Work/Family Conflict and Work/Life Balance

Variables Predicting Affective Organizational Commitment- Initial Sample (N = 267)

Variable B SEB P

Step 1

Work-to-Family Conflict -.14 .05 -.17**

Family-to-Work Conflict -.10 .06 -.10

Step 2

Work-to-Family Conflict -.17 .09 -.21

Family-to-Work Conflict .01 .07 .01

Work Interference with Personal Life .32 .12 .33**

Personal Life Interference with Work -.09 .09 -.07

Work/Personal Life Enhancement .45 .06 .48**

v.'j;:—772-----77FT—77“—. xr.2---- 7V7 777
Note. R~ = .05 for Step 1; AR~ =. 17 for Step 2 .
**g<.01
* g < .05
 105

Table 26

Hierarchical Regression Analysis for Work/Family Conflict and Work/Life Balance

Variables Predicting Turnover Intentions - Initial Sample (N = 267)

Variable B SEB ß

Step 1

Work-to-Family Conflict .35 .08 .27**

Family-to-Work Conflict .14 .10 .09

Step 2

Work-to-Family Conflict .36 .15 .28*

Family-to-Work Conflict .03 .12 .02

Work Interference with Personal Life -.33 .19 -.21

Personal Life Interference with Work .09 .15 .04

Work/Personal Life Enhancement -.48 .11 -.32**

Note. R2 = .10 for Step 1; ARJ =.07 for Step 2 .
**p<.01
* p<.05
 106

Cross-validation Using Holdout Sample

After completing the aforementioned analyses using the initial sample. I applied

the same statistical analyses to the holdout sample (N=273) as a method of cross-

validation to determine whether the findings were consistent in a second sample.

Demographic Information

The demographic characteristics of the holdout sample were similar to the initial

sample. Table 27 presents a summary of the holdout sample demographic characteristics.

Scale Assessment

Work/life balance scale item descriptive statistics are presented in Table 28.

Item-total correlations for each of the items were consistent with the results obtained

using the initial sample, and are presented in Table 29. Overall, the item total

correlations in the holdout sample were slightly lower on the WIPL and WPLE

dimensions and higher on the PLIW dimension compared to the corrected item-total

correlations obtained in the initial sample. Despite these general differences in trend, all

corrected item-total correlations in the holdout sample were above .45 (lowest was

r = .53), suggesting that each item was sufficiently related to the other items on the scale.

Internal consistency reliability estimates for each of the three dimensions of

work/life balance in the holdout sample indicated that the scale dimensions demonstrated

acceptable levels of reliability (ex's = .93, .76, and .76, respectively). The coefficient

alpha for the PLIW dimension was slightly lower in this sample compared to the initial

sample (a=.83 vs. a = .76; consistent with the trend of lower corrected item-total

correlations), but still indicated an acceptable level of internal consistency reliability.

 107

Table 27

Demographic Characteristics of the Holdout Sample (N = 273)

Variable Categories Percentage (%)

Gender Male 53.5
Female 46.5

Age 18 to 24 years .4
25 to 29 years 2.2
30 to 34 years 11.0
35 to 39 years 18.3
40 to 44 years 17.6
45 to 49 years 21.2
50 to 54 years 17.6
55 to 59 years 8.4
60 to 64 years 3.3
65 years or older 0

Relationship Status Married, living together 75.9

Married living apart 1.5
Not married, but living with partner 6.9
Not married, not living with partner,
but involved in committed
relationship 5.4
Not married, and not involved in
committed relationship 10.3

Number of children
living in household at least
50% of the time 0 23.1
1 29.2
2 30.3
3 14.9
4 or more 2.6

Primary' responsibility 14.5

for elder care Yes 85.5
No
 108

Table 27 (continued)

Demographic Characteristics of the Holdout Sample (N = 273)

Variable Categories Percentage (%)

Total work hours per week Less than 40 3.0
40 to 49 28.1
50 to 59 43.1
60 to 69 17.6
70 to 79 7.1
More than 80 1.1

Highest Education Level

Obtained High school graduate or less 3.7
Some college 19.0
4-year college degree 27.5
Some post-graduate study 14.9
Masters Degree 26.4
Some post-masters study 3.7
Doctoral/professional degree 4.8

Job Tenure Less than 1 year 14.7

1 to 2 years 29.0
3 to 4 years 26.1
5 to 9 years 18.4
10 to 14 years 6.6
15 years or more 5.1

Organization Tenure Less than 1 year 2.2

1 to 2 years 8.4
3 to 4 years 15.4
5 to 9 years 30.0
10 to 14 years 17.9
15 years or more 26.0
Table 28

Work/Life Balance Scale Item Means and Standard Deviations - Holdout Sample

Item Mean SD
1. I struggle with trying to juggle both my work and non-work responsibilities. 3.28 1.07
2. My personal life drains me of the energy I need to do my job. 2.03 0.79
4. My work suffers because of everything going on in my personal life. 1.85 0.70
5. Iam too tired to be effective at work because of things I have going on in my personal life. 1.82 0.71
6. My job gives me energy to pursue activities outside of work that are important to me. 2.74 1.07
7. I have difficulty scheduling vacation time because of my workload. 2.99 1.14
8. Iam unable to relax at home because I am preoccupied with my work. 2.77 1.03
9. I am happy with the amount of time I spend doing activities not related to work. 3.14 1.16
10. Because of my job, I am in a better mood at home. 2.76 1.00
11.1 often have to make difficult choices between my work and my personal life 2.79 0.93
12.1 have to put aspects of my personal life “on hold" because of my work. 2.73 1.03
14.1 am in a better mood at work because of everything I have going for me in my personal life. 3.64 0.95
15. I often neglect my personal needs because of the demands of my work. 2.88 1.02
17. I have difficulty getting my work done because 1 am preoccupied with personal matters at work. 1.94 0.72
18. My personal life suffers because of my work. 2.62 1.05
19. I have to miss out on important personal activities due to the amount of time I spend doing work. 2.44 0.95
20. My personal life gives me the energy to do my job. 3.39 1.03
21.1 feel that I allocate appropriate amounts of time to both work and non-work activities. 3.39 1.14
22.1 make personal sacrifices to get work done. 3.13 0.92
23.1 come home from work too tired to do things I would like to do. 3.21 1.03
24. My job makes it difficult to maintain the kind of personal life I would like. 2.77 1.17

©
©
 110

Table 29

Corrected Item-Total Correlations for Work/Life Balance Scale Items - Holdout Sample

WIPL1 PLIW2 WPLE3

Item ITC Item ITC Item ITC

18 .84 4 .72 14 .61

24 .82 5 .71 20 .54
15 .77 2 .60 6 .53
12 .76 17 .55 10 .53
19 .75
23 .73
21 .68
22 .66
11 .65
8 .65
9 .61
1 .58
7 .54

Note. WIPL = Work Interference with Personal Life’ PLIW = Personal Life
Interference with Work; WPLE = Work/Personal Life Enhancement
 Ill

Next, I computed the bivariate correlations between each scale item and the one-

item measure of overall work/life balance. These correlations are presented in Table 30.

In general, these correlations are similar to those obtained in the initial sample. However,

there are two differences worthy of noting. First, the correlations between PLIW items

and overall work/life balance were lower in the holdout sample compared to the initial

sample (r’s ranged from -.15 to -.18 in the holdout sample, whereas they ranged from -.20

to -.26 in the initial sample). This suggested that the personal life interference with work

items were less strongly related to perceptions of overall work/life balance in the second

sample compared to the first sample.

The factor structure of the holdout sample was assessed using partial confirmatory

factor analysis to test the fit of the PCFA model obtained with the initial sample to the

data in the holdout sample. The three-factor model with two errors allowed to covary fit

the data reasonably well (x2 (184) = 466.07, TLI = .977, CFI = .982, NFI = .971, and

RMSEA = .075. The RMSEA was below Browne and Cudeck's (1992) recommendation

of .08 as an upper bound of good fit, and overall, these results suggest that the 3-factor

model fit the data well.

Item Response Theory Analyses. The factor analysis, corrected item-total

correlations, and scale reliability estimates summarized in the previous section are

approaches to item and scale analysis based on classical test theory’. Classical test theory

(CTT) assumes that any obtained score on a variable of interest is equal to the sum of the

true score and measurement error (i.e., random error; Guion, 1998). Another approach to

scale assessment is an item analysis based on item response theory (IRT). Item response
 112

Table 30

Correlations Between Each Work/Life Balance Scale Item and One-Item Work/Life

Balance Measure - Holdout Sample

Item r Item r Item r

21 .66 2 -.18 10 .51
24 -.64 5 -.18 6 .48
18 -.63 4 -.17 14 .40
15 -.58 17 -.15 20 .36
23 -.58
19 -.56
12 -.54
9 .48
11 -.48
8 -.45
22 -.44
1 -.43
7 -.43

Note. All correlations are statistically significant at p < .05.

 113

theory links the probability of a particular item response (i.e., the likelihood of

responding to a particular item with “not at all," “rarely," “sometimes,” “often,” or

“almost all the time” ) to the characteristic or latent trait that is being assessed by the test

(e.g., work interference with personal life; Drasgow & Hulin, 1990). One advantage of

IRT over CTT is that IRT parameter estimates do not vary based on the distribution of the

trait (Hambleton, Swaminathan, & Rogers, 1991). In other words, CTT takes into

account the measure as a whole, whereas IRT permits researchers to use any individual

item in a measure as the unit of analysis (Drasgow & Hulin, 1990).

Recently, Reeve and Smith (2001) recommended a “convergent evidence

approach" to scale development that combines CTT and IRT approaches to seek

consistency in making conclusions about item psychometric properties. Following the

recommendations of Reeve and Smith (2001), I computed item response parameters by

applying Samejima's (1989) graded response model to the data using Multilog software

(Thissen, 1991). The graded response model has been recommended for use with

polytomous data (i.e., when there are more than two response options for an item; Zickar,

in press). Because this model assumes unidimensionality of the underlying trait, I

conducted three IRT analyses: one for each of the three work/life balance dimensions.

The first step in interpreting the IRT results was to examine the a parameter for

each of the initial work/life balance scale items. The a parameter measures item

discrimination, and a values of 1.00 or higher are considered acceptable levels of

discrimination ability (Zickar, personal communication; Reeve and Smith, 2001). IRT a

parameter estimates obtained in the present study and the standard error of measurement

for each a parameter estimate are reported in Table 31.

 114

In addition to the a parameter, the test information function (TIF) provides

evidence of the sensitivity of the scale, which refers to how much information is provided

by the scale at all levels of the trait being measured (e.g., WIPL, PLIW, or WPLE). The

purpose of examining the TIF is to determine the extent to which the scale items measure

the work/life balance content domain across all levels of work/life balance. In other

words, for the scale to be most useful and widely applicable, it is important for scale

items to measure work/life balance at low, moderate, and high levels of work/life

balance. The test information function is determined by plotting test information (0)

obtained for the test (i.e., each dimension of work/life balance) along the y-axis, and the

levels of theta (0), which range from -2.0 to +2.0 along the x-axis. Test information

functions for WIPL, PLIW, and WPLE are shown in Figures 3-5, respectively. These

figures are useful for identifying the theta levels (i.e., 0; levels of the latent trait) for

which the scale provides information or does not provide information.

The test information function for the first dimension, WIPL, indicated that this

13-item scale provided information about the latent trait at most levels, although it was

less sensitive toward higher levels of work interference with personal life. (This was

indicated by the line on the graph being lower for high levels of theta compared to low or

moderate levels.) In other words, this thirteen-item scale was better at measuring work

interference with personal life among respondents who had little interference or a

moderate amount compared to those who perceived that their work interferes with their

personal life a great deal.

With regard to personal life interference with work (PLIW) and work/personal

life enhancement (WPLE), each four-item scale was most sensitive at middle levels of the
 115

Table 31

IRT a Parameters for Each Work/Life Balance Scale Item - Total Study Sample

Work Interference Personal Life Interference Work/Personal Life

with Personal Life With Work Enhancement
Item a SE Item a SE Item A SE

1 1.24 .10 2 2.36 .22 6 1.02 .11

7 1.18 .12 4 3.36 .24 10 1.09 .11

8 1.56 .12 5 3.30 .29 14 3.20 .22

9 1.33 .13 17 1.88 .18 20 2.85 .19

11 1.59 .13

12 2.52 .20

15 2.41 .17

18 3.32 .24

19 2.28 .19

21 1.75 .13

22 1.78 .13

23 1.71 .15

24 2.70 .20

Note. N = 540
 116

Figure 3

IRT Test Information Function for WIPL dimension - Total Study Sample (N=540)
Information

-2.5 -2 -1.5 1 -0.5 0 0.5 1.5 2 2.5

Theta
 117

Figure 4

IRT Test Information Function for PLIW dimension - Total Study Sample (N=540)

Theta
 118

Figure 5

IRT Test Information Function for WPLE dimension - Total Study Sample (N=540)

Theta
 119

construct (thetas of-1.0 and +0.5) and much less sensitive at the high end (e.g., for

individuals with a high degree of personal life interference with work, etc.). For all three

scales, none was particularly useful for assessing high levels of each dimension, but

seemed to provide information at low and moderate levels of interference and

enhancement.

Scale Item Assessment Summary'

Twenty-four work/life balance items were administered to managers during the

second phase of scale development. Exploratory factor analysis results, corrected item-

total correlation coefficients, and estimates of scale reliability (coefficient alpha) in the

initial sample suggested that two of these items be removed from the scale: Items 13 and

16 (as listed in Appendix C). Item 13 (“I am able to accomplish what I would like in

both my personal and work lives”) cross-loaded on two factors in the initial exploratory

factor analysis, making it difficult to determine how to interpret the item. Item 16

(“When I am at work I worry about things I have to do outside of work”) was removed

because the corrected item-total correlations and coefficient alphas computed with and

without the item indicated that Item 16 was not strongly related to the other items on the

same dimension, and the scale demonstrated higher reliability without Item 16. A third

scale item. Item 3 (“I feel overwhelmed when I try to balance my work and personal

life”), was removed from the scale based on the results obtained from a partial

confirmatory factor analysis because it was redundant with Item 1.

Removal of Items 3, 13, and 16 improved the measurement properties of the scale

and resulted in a 21-item scale assessing three dimensions: work interference with

personal life (WIPL), personal life interference with work (PLIW), and work/personal
 120

life enhancement (WPLE). A partial confirmatory factor analysis of the scale yielded a

21-item, three-factor model that indicated good fit to the data across both samples.

Work interference with personal life. WIPL was assessed with 13 items, and

demonstrated very high internal consistency reliability across both samples (a = .94 and

a = .93 for the initial and holdout samples, respectively). The coefficient alphas obtained

for this scale were high due to similarity in item content as well as a large number of

items measuring this dimension.

The content of most of the items on this scale specifically assessed the extent to

which work has a negative impact on one’s personal life. The item with the highest

factor loading and best ability to discriminate on the latent construct was Item 18 (“My

personal life suffers because of my work”). Results also indicated that Items 12, 15, and

19 had excellent measurement properties. An explanation for the high correlations

among the items and high reliability is that there is semantic similarity in item content,

although the items did not seem to overlap completely with one another.

Two of the items on the WIPL scale (Items 9 and 21) were worded positively,

such that higher scores on these items indicated a lack of interference. These items were

recoded to be consistent with the other 11 items. Lower levels of work interference with

personal life were interpreted as indicative of higher levels of work/life balance. This

interpretation is consistent with the finding that WIPL items were strongly correlated

with a one-item measure of overall work/life balance employed as an initial “validity

check” among the items.

In addition to computing the scale item statistics based on classical test theory

(e.g., factor analysis, corrected item-total correlations, and reliability), I conducted item
 121

response theory analyses to gather additional psychometric information about the scale.

First, these IRT results described which items were more useful for discriminating among

individuals with high and low levels of WIPL. All thirteen items discriminated

adequately, as evidenced by a parameters above 1.0. Among the 13 items, Item 18

discriminated best, followed by Items 24, 12, 15, and 19, respectively. These results

converged with findings from the classical test theory analyses such that the items with

the highest a parameters also had the highest factor loadings and corrected item-total

correlations.

With regard to item content, nine of the WIPL items (Items 7, 8, 12, 15, 18, 19,

22. 23, and 24) implied a directional effect of work on personal life, such that work

affects or spills over into one's personal life. The remaining four items on this dimension

(Items 1,9, 11, and 21) did not imply this same directional effect, but were still more

strongly related to the items on this dimension than to items on either of the other two

dimensions in both samples. This finding suggested that these items were more

indicative of the extent to w hich work affects one's personal life than vice-versa (i.e., the

extent to which one’s personal life affects work).

The construct of work/life balance was defined as incorporating time, energy,

strain, and goal accomplishment. Review of the content of these items suggested that

four of the WIPL items (Items 7, 9, 19, and 21) incorporated issues of time, four items

(Items 8, 15, 18, and 24) described strain experienced in one’s personal life due to work,

one item (Item 23) pertained to energy, and the remaining items referred to

accomplishing non-work goals or meeting non-w'ork demands. As a result, all four

 122

components of the work/life balance definition were included in items on the WIPL

dimension.

Personal life interference with work. The second dimension of the work/life

balance scale included four items that assessed the extent to which one’s personal life

interferes with one’s work (PLIW). Internal consistency reliability coefficients for this

scale obtained in both samples (a = .83 and a = .76) indicated that the scale had

acceptable levels of reliability. The lower alphas obtained for this dimension compared

to the WIPL scale could be attributed to less overlap in item content and fewer items in

the scale (i.e., 4 items, compared to 13; Cortina, 1993).

All four of the PLIW items demonstrated acceptable levels of item discrimination

based on results of the IRT analysis. Among the four items. Items 4 (“My work suffers

because of everything going on in my personal life’’ and 5 (“I am too tired to be effective

at work because of things I have going on in my personal life’’) were the most

discriminating items. With regard to scale sensitivity, the PLIW scale provided more

information for individuals with low or moderate levels of PLIW than for high levels of

PLIW. In other words, these four items indicated very little about individuals whose

personal life interferes a great deal with their work. Additional items would have to be

written and assessed before being able to gather information about individuals with high

PLIW.

Compared to WIPL, items on the PLIW scale assessed the opposite direction of

the work/personal interface: the extent to which aspects of one’s personal life has a

negative impact on work. Examples of the items that loaded highly on this dimension

were Item 5 and Item 2 (“My personal life drains me of the energy I need to do my job.”).
 123

With regard to the four components of work/life balance, Items 2 and 5 dealt with having

energy for work, Item 4 referred to strain experienced as a result of one's personal life

interfering with work, and Item 16 pertained to goal accomplishment (e.g., getting work

done). None of the items on this dimension incorporated the other issue identified as a

salient aspect of work/life balance: time. Future scale development efforts should be

directed toward writing new items to incorporate personal life interference with work as

it relates to time. An example of this type of item would be “My personal life takes away

time that 1 would otherwise spend doing work."

The four items on this scale demonstrated high reliability and seemed to assess

personal life interference with work. However, unlike items on the WIPL scale, none of

the PLIW items correlated very highly with the one-item measure of overall work/life

balance. This finding suggested that perceptions of overall work/life balance were more

strongly related to the effect of work on personal life than the effect of personal life on

work.

Work/personal life enhancement. The third work/life balance scale dimension

assessed the extent to which one's personal life is enhanced by work or vice versa. The

four items on this scale were more highly related with one another than with items on

either of the other two dimensions, suggesting that measurement of the work/personal life

interface involved more than just directional differences in which work affects personal

life and vice versa. These four items demonstrated acceptable levels of reliability across

two samples of data (a = .77 and a = .76). The internal consistency reliability estimates

for this dimension were lower in the first sample compared to WIPL and PLIW, and

about the same as PLIW in the second sample. The corrected item total correlations
 124

obtained for these four items were high, ranging from .68 to .75, suggesting that each

item was strongly related to the other items.

Two of the WPLE scale items (Items 14 and 20) seemed to discriminate much

better than Items 6 and 10. Item 14 was the most discriminating item, followed by Item

20, and then Items 10 and 6. The test item function for the WPLE scale was similar to

the TIF obtained for the PLIW scale; these four items as a whole provided more

information at low and moderate levels of the construct than at high levels. In other

words, these items assessed the extent to which work and personal life enhance one

another for those who reported low and moderate levels of enhancement, but is not useful

for identifying or describing individuals who perceive their personal life adds a great deal

to their work life or vice versa.

In terms of item content, two of the WPLE items assessed the extent to which

work enhances personal life (Items 6 and 10), and two items referred to the positive

impact of one’s personal life on work. Unlike the previous two scale dimensions, in

which items were related with one another based on the direction of influence across the

work/personal life interface, these four items were related to one another regardless of the

direction of enhancement. These results can be explained in two ways. First, there were

not enough items to be able to detect this different response pattern. In other words, more

items assessing the positive contributions of work to one's personal life and vice versa

may have resulted in finding two separate dimensions. Secondly, the direction of

influence of one domain on another (work on personal life, or personal life on work) may

not matter when it comes to enhancing or contributing positively to one’s life.

 125

Validity Evidence

Results obtained in the initial sample (N = 267) supported the predictions made

regarding the relations among work/life balance and other variables. Further evidence for

validity of the scale is presented in this section, which describes the results obtained in

the holdout sample (N=273).

Scale descriptive statistics. Scale means and standard deviations computed for all

study variables in the holdout sample are presented in Table 32. The mean distributions

and pattern of work/life balance scale means compared to one another were the same as

in the initial sample.

Hypothesis Tests

Having completed the assessment of the psychometric properties of the work/life

balance scale, the next step was to examine the relations between each of the work/life

balance scale dimensions and the other variables measured in this study to test the

hypotheses in a second sample. Scale means and standard deviations for all study

variables are presented in Table 33, and the correlation matrix appears in Table 34. I

conducted a series of independent sample t-tests to examine whether any of the scale

means differed between the initial sample and the holdout sample. There were no

significant mean differences between the two samples (p < .05).

Hypothesis 1. With regard to convergent validity of the work/life balance scale.

Hypothesis 1 predicted that work/life balance would be related to measures of

work/family conflict. Consistent with the initial sample, correlations obtained from the

holdout sample indicated a very strong relationship between WIPL and WFC, and a fairly

strong relationship between PLIW and FWC. WIPL was related to both WFC and FWC.
Table 32

Descriptive Statistics for all Study Variables - Holdout Sample

Standard
Scale Mean Deviation
1. WIPL - Work Interference with Personal Life 2.85 0.75
2. PLIW - Personal Life Interference with Work 1.91 0.59
3. WPLE - Work/Personal Life Enhancement 3.13 0.78
4. Overall Work/Personal Life Balance 3.38 1.06
5. WFC - Work-to-Family Conflict 2.78 0.91
6. FWC - Family-to-Work Conflict 2.01 0.71
7. Role Overload 3.62 0.77
8. Stress - Pressure 17.98 3.17
9. Stress - Threat 17.92 4.39
10. Time Pressure 3.42 0.79
11. Time Management Behavior (TMB)- Setting Goals and Priorities 3.64 0.58
12. Time Management Behavior (TMB)- Planning and Organizing 3.55 0.67
13. Time Management Behavior (TMB)- Preference for Organization 3.75 0.61
14. Job Satisfaction 42.38 10.14
15. Life Satisfaction 3.31 0.85
16. Organizational Commitment 3.81 0.71
17. Turnover Intentions 2.46 1.16
K)
O
Table 33

Correlation Matrix of Study Variables - Holdout Sample

1 2 3 4 5 6 7 8 9 IP 11 12 13 ]4 J5 IP 12

1. WIPL .93
2. PLIW .31** .76
3. WPLE -.64** -.40** .76
4. Overall W/L Bal. -.70** -.21»* .57** --
5. WFC .84** .31** -.54** -.69** .90
6. FWC .22** .51** -.30** -.11 .27** .89
7. Role Overload .55** .16** -.34** -.44** .51** .02 .67
8. Stress - Pressure .51** .17** -.36** -.43** .45** .07 .53** .78
9. Stress - Threat .56** .19** -.45** -.53** .52** .10 .53** .63** .80
10. Time Pressure .62** .30** -.45** -.47** .58** .19** .74** .56** .64** .84
11. TMB-Goals -.20** -.29** .35** .24** -.20** -.26** -.07 -.05 -.17** -.15* .81
12. TMB - Planning -.16** -.18** .24** .19** -.16** -.16** -.06 -.10 -.17** - 15* .63 *♦ .75
13. TMB-Org. -.44** -.38** .38** .31** -.38** -.31** -.35** -.34** -.41»* -.50** .39** .35** .76
14. Job Sat. -.41** -.21** .45** .43** -.42** -.13* -.29** -.17** -.45** -.35** .21** .12* .18** .90
15. Life Sat. -.56** -.34** .63** .64** -.55** -.26** -.32** -.27** -.43** -.45** .34** .25** .31** .52** .86
16. Org. Comm. -.22** -.30** .32** .26** -.27** -.11 -.16** -.11 -.20** -.25** .23** .14* .11 .54** .37** .80
17. Turnover .34** .26** -.37** -.38** .36** .10 .29** .24** .40** .36** -.19** -.09 -.24** -.60** -.39** -.70** .91
Note. Coefficient alpha for each scale appears on diagonal. (N=273)
** p<.01
* p < .05

t-j
"si
 128

Hypothesis 2. Hypothesis 2 stated that work/life balance would be negatively

related to role overload. In the initial sample, Hypothesis 2 was supported for two of the

work/life balance dimensions: WIPL and WPLE, but not PLIW. The same pattern of

results was found in the holdout sample, such that no significant relationship was found

between personal life-to-work interference (PLIW) and role overload.

Hypothesis 3. Hypothesis 3 stated that work/life balance would be negatively

related to perceptions of time pressure. Results in the second sample indicated that all

three dimensions of work/life balance were related to perceptions of time pressure. The

relationship between time pressure and WIPL was strong, as indicated by a high

correlation of r = .62. These results were consistent with the correlations obtained in the

first sample. In other words, individuals with more interference and less enhancement

across the work/personal interface experienced more time pressure those with less

interference or more enhancement.

Hypothesis 4. The fourth hypothesis predicted a positive relationship between

employees’ perceptions of work/life balance and time management behavior. Similar to

the initial sample, all three dimensions of time management behavior (setting goals and

priorities, planning and scheduling, and preference for organization) were significantly

correlated with WPLE in the predicted direction, and two dimensions (setting goals and

priorities and preference for organization) were related to PLIW. In addition, WIPL was

also related to preference for organization. These results were consistent with findings

from the first sample, except that the correlation between PLIW and planning and

scheduling was weak (r = -. 18) in the second sample.

 129

Hypothesis 5. Hypothesis 5 predicted a negative relationship between work/life

balance and feelings of overall job stress, or a positive relationship between stress and

interference. Correlations among all three dimensions of work/life balance and stress

were statistically significant. However, although the general pattern of results obtained in

the second sample was consistent with findings from the first sample, the correlations

between PLIW and stress were low (r = . 17 and r = . 19), and therefore not indicative of a

very strong relationship between personal life interference with work and feelings of

overall work stress.

Hypothesis 6. The sixth hypothesis predicted that work/life balance would be

positively related to overall job satisfaction. The results obtained in the second sample

were consistent with the first sample, indicating that WIPL, PLIW, and WPLE were

significantly related to job satisfaction.

Hypothesis 7. Similar to job satisfaction, all three dimensions of work/life

balance were related to perceptions of life satisfaction. The same pattern of correlations

was obtained in the second sample, such that WPLE was more strongly related to life

satisfaction than either work interference with personal life or personal life interference

with work.

Hypothesis 8. Hypothesis 8 predicted a positive relationship between work/life

balance and organizational commitment. In the second sample, all three work/life

balance dimensions were related to affective organizational commitment. These results

indicated more support for Hypothesis 7 than I obtained in the initial sample because the

observed correlation between WIPL and organizational commitment in the holdout

 130

sample (r = -.22) was larger than the correlation between the same two variables in the

initial sample (r = -. 16).

Hypothesis 9. Hypothesis 9 stated that employees’ perceptions of work/life

balance would be negatively related to their intentions to leave their work organization.

Similar to results obtained in the initial sample, all three dimensions of work/life balance

were significantly related to employees' turnover intentions, and work/personal life

enhancement was more strongly related to turnover intentions than either of the other

work/life balance dimensions.

Hypothesis 10. Next I tested Hypothesis 10. which stated that work/life balance

would be empirically distinct from WFC. To accomplish this, I repeated the hierarchical

regression analyses conducted earlier to test for incremental validity of work/life balance

over work/family conflict.

Although results in the initial sample failed to indicate that work/life balance

accounted for incremental variance over work/family conflict in role overload and time

pressure, results in the second sample did find this effect (AR2 = .05, F (2,3) = 6.69,

p < .01 for role overload, and AR2 = .07, F (2,3) = 10.15, p < .01 for time pressure). For

the pressure dimension of overall work stress and time pressure, results in the holdout

sample indicated an even larger effect size compared to the initial sample (AR2 = .06,

F (2,3) = 7.14, p < .01 for overall stress in terms of pressure, compared to AR2 = .02 for

pressure in the prior sample). Although the incremental validity of work/life balance

with regard to job satisfaction was significant in the second sample (AR = .07,
 131

F (2,3) = 8.42, p < .01), the effect size of .07 was smaller than the effect size of .18

obtained in the initial sample. The effect sizes obtained for other criterion variables in

the holdout sample were comparable to those found in the initial sample.

In summary, results of these analyses were consistent with prior findings such that

work/life balance (WIPL, PLIW, and WPLE) accounted for significant variance in each

of the other variables beyond work/family conflict. Hence, the results of these

hierarchical regression analyses indicate support for Hypothesis 10 by demonstrating

that, although similar, work/life balance and work/family conflict are empirically distinct

constructs.

Exploratory Analyses

Demographic differences

In addition to testing the hypotheses in this study, a series of additional data

analyses were conducted to determine whether the managers who participated in this

study differed in perceptions of work/life balance based on demographic characteristics.

The full sample (N=540) was used for these analyses in order to have a large enough

sample size in each of the subgroups, which was necessary to increase statistical power.

Individual characteristics. Males did not significantly differ from females in

perceptions of work interference with personal life, personal life interference with work,

or work/personal life enhancement; nor were there any differences based on education

level, p > .05. However, managers' perceptions did differ with respect to age on all three

dimensions of work/life balance (F (3,536) = 4.12, p < .01 for WIPL; F (3, 536) = 4.68,

p < .01 for PLIW, and F (3, 536) = 4.96, p < .01 for WPLE). Specifically, managers age

sixty or above reported significantly less work interference with personal life compared
 132

to managers between the ages of 50 and 59 or younger. Personal life interference with

work appeared to decrease with age such that mean levels of PLIW were lower among

older managers (60 and older) compared to managers younger than 50 years of age.

Similarly, work/personal life enhancement increased with age; managers above age 60

reported significantly higher levels of work/personal life enhancement than younger

managers. Age differences in perceptions of work/life balance are shown in

Figures 6 to 8.

The total number of hours employees worked per week was significantly related

to work interference with personal life (r = .25, g < .01), but was unrelated to either of the

other two dimensions of work/life balance. Although the correlation between average

total work hours per week was significantly related to WIPL, the magnitude of the

correlation may have been attenuated by the low variability in work hours; more than half

the sample (54%) reported working an average of 45 to 55 hours per week, and 21%

indicated they work 50 hours per week.

The sixty-three managers who reported having elder care responsibilities (i.e.,

primary responsibility for caring for one or more of their parents, in-laws, or other

relatives) had significantly lower levels of work/personal life enhancement than the

managers without elder care responsibilities. There were no differences on other

dimensions of work/life balance with regard to elder care responsibilities.

 133

Figure 6

Age Differences in Work Interference with Personal Life - Total Study Sample (N=540)
Mean of WIP

Age
 134

Figure 7

Age Differences in Personal Life Interference with Work - Total Study Sample (N=540)
Mean of PLIW

Age
 135

Figure 8

Age Differences in Work/Personal Life Enhancement - Total Study Sample (N=540)

Age
 136

In terms of the number of children who live at home at least 50% of the time,

those who have at least one child reported higher levels of WPLE compared to employees

without children at home (t(388) = -2.16, p < .05). Consistent with other findings, no

effects were found for WIPL or PLIW, p > .05. No significant differences were found

regarding the relationship between work/life balance and the age of children, p > .05.

Another individual difference variable included as part of this study involved

asking respondents whether they perceived their personal life as a priority over work,

work as a priority over personal life, or both work and personal life as equally important.

Fourteen percent indicated that work was their priority, 29% indicated that personal life

was their priority, and the majority (i.e., 55%) indicated that they were both equally

important. A comparison of mean differences with regard to priority indicated that those

who prioritize their work over their personal life experience significantly more work

interference with their personal life (t (234) = -8.54, p < .05), less personal life

interference with work (t (234) = 2.32, p < .05), and less work/personal life enhancement

(t (234) = 3.47, p < .05) compared to those who indicated that their personal life is a

priority over work.

Organizational variables. Next, I tested whether there were any differences in

perceptions of work/life balance based on demographic variables related to managers'

jobs and organizations. No differences were found regarding length of time in current

position, length of tenure within one's organization, or management level within the

organizational hierarchy, or type of department (e.g., marketing, HR, finance, etc.),

p > .05. However, 1 found a clear difference with regard to salary level. Specifically,

managers who earn more money reported significantly more interference of their work
 137

with their personal life (F (5,523) = 4.01, p < .01. A graph depicting mean levels of

WIPL for each salary level is presented in Figure 9. There were no differences in mean

levels of PLIW or WPLE with regard to salary level, p > .05.

Incremental Validity of Enhancement

In addition to testing for demographic differences in perceptions of work/life

balance, I conducted additional analyses to determine the extent to which the

work/personal life enhancement dimension accounted for unique variance in the criterion

measures after accounting for work/personal life interference (i.e., WIPL and PLIW).

Using hierarchical regression analysis to gather this information, I found that

work/personal life enhancement was a significant predictor of other variables beyond

interference. Results for variables in which WPLE accounted for a significant portion of

variance beyond WIPL and PLIW are presented in Tables 34 to 40.

Figure 9

Work Interference with Personal Life by Annual Salary - Total Study Sample (N=540)

00
 139

Table 34

Hierarchical Regression Analysis for Work/Life Balance Variables Predicting Setting

Goals and Priorities - Total Study Sample (N=540)

Variable B SEB P

Step 1

Work Interference with Personal Life -.02 .03 -.02

Personal Life Interference with Work -.29 .04 -.30**

Step 2
Work Interference with Personal Life .11 .04 .14**

Personal Life Interference with Work -.22 .04 -.23**

Work/Personal Life Enhancement .23 .04 .31**

v,
Note. mm —---,1 ;. AR2---
. -- 'R~ = .09 for AStep =.06 for oï
mx
Step 2 .
**g<.01
* g<.05
 140

Table 35

Hierarchical Regression Analysis for Work/Life Balance Variables Predicting Planning

and Scheduling- Total Study Sample (N=540)

Variable B SEB ß

Step 1

Work Interference with Personal Life -.05 .04 -.06

Personal Life Interference with Work -.21 .05 -.19**

Step 2
Work Interference with Personal Life .04 .05 .05

Personal Life Interference with Work -.16 .05 -.14**

Work/Personal Life Enhancement .18 .05 .21**

TV.----- 77Z----- ----------- 77---- ;—7772 o-, r 0,
Note, R2 = .05 for Step 1 ; AR2 =.03 for Step 2 .
**p< .01
* p<.05
 141

Table 36

Hierarchical Regression Analysis for Work/Life Balance Variables Predicting Overall

Job Stress (Threat) - Total Study Sample (N=540)

Variable B SEB ß

Step 1

Work Interference with Personal Life 3.14 .23 .53**

Personal Life Interference with Work .04 .29 .01

Step 2
Work Interference with Personal Life 2.52 .27 .43**

Personal Life Interference with Work -.28 .29 -.04

Work/Personal Life Enhancement -1.12 .26 -.20**

7-- 772---- , A r.2----------------- 7777--- Ö7
Note. R" = .29 for Step 1 ; AR2 =.02 for Step 2 .
**p<.01
* p < .05
 142

Table 37

Hierarchical Regression Analysis for Work/Life Balance Variables Predicting Overall

Job Satisfaction - Total Study Sample (N=540)

Variable B SEB P

Step 1

Work Interference with Personal Life -4.30 .60 -.31**

Personal Life Interference with Work -2.39 .75 -.14**

Step 2
Work Interference with Personal Life -1.26 .67 -.09

Personal Life Interference with Work -.81 .73 -.05

Work/Personal Life Enhancement 5.49 .66 .41**

v, ■ —rTZ-— , .n ’--- T77T —FT
Note. R2 = . 14 for Step 1 ; AR~ =. 10 for Step 2 .
**p<.01
* p<.05
 143

Table 38

Hierarchical Regression Analysis for Work/Life Balance Variables Predicting Life

Satisfaction - Total Study Sample (N=540)

Variable B SEB ß

Step 1

Work Interference with Personal Life -.53 .04 -.48**

Personal Life Interference with Work -.30 .05 -.21**

Step 2
Work Interference with Personal Life -.30 .05 -.27**

Personal Life Interference with Work -.18 .05 -.13**

Work/Personal Life Enhancement .42 .04 .40**

x, :----- 772----- r/.--------------- .—7772—777"7~77
Note. R2 = .35 for Step 1 ; AR2 =. 10 for Step 2 .
**g< .01
* g<.05
 144

Table 39

Hierarchical Regression Analysis for Work/Life Balance Variables Predicting Affective

Organizational Commitment - Total Study Sample (N=540)

Variable B SEB ß

Step 1

Work Interference with Personal Life -.11 .04 -.12*

Personal Life Interference with Work -.25 .05 -.21**

Step 2
Work Interference with Personal Life .08 .05 .08

Personal Life Interference with Work -.15 .05 -.13**

Work/Personal Life Enhancement .35 .05 .38**

VT ~— 772--- nn r- O. , ~7772 77777—Ö7
Note. R2 = .08 for Step 1 ; AR2 = 09 for Step 2 .
**p<.01
* p < .05
 145

Table 40

Hierarchical Regression Analysis for Work/Life Balance Variables Predicting Turnover

Intentions - Total Study Sample (N=540)

Variable B SEB P

Step 1

Work Interference with Personal Life .37 .07 .24**

Personal Life Interference with Work .29 .09 .15**

Step 2
Work Interference with Personal Life .13 .08 .08

Personal Life Interference with Work .17 .09 .09

Work/Personal Life Enhancement -.43 .08 -.29**

v," — Ï72 --- 1 , --- KTT — KT
Note. R2 = .11 for Step 1; AR2 =.05 for Step 2 .
**p<.01
* p<.05
 146

CHAPTER V. DISCUSSION

The purpose of the present study was to extend previous work/non-work conflict

research by developing a more general construct - work/personal life balance. I defined

the construct and developed a 31-item scale and reduced it to 21 items through empirical

construct validation. Using two samples of data obtained from a heterogeneous

population of managers, responses to these items fit a three-factor model in both samples.

The three dimensions that emerged were work interference with personal life (WIPL),

personal life interference with work (PLIW), and work/personal life enhancement

(WPLE). In this final chapter, I discuss the present study with regard to scale item

generation, scale item assessment, and initial validity evidence for scale. Next, I compare

work/life balance to work/family conflict, and describe theoretical and practical

implications of the results. Finally, I present the limitations of this study and overall

conclusions.

Scale Item Generation

I developed a conceptual definition of work/personal life balance based on one-

on-one interviews with a diverse group of employees, a review of previous empirical

literature, and the conservation of resources theory (Hobfoll, 1988; 1989). This

definition indicated that having a lack of work/personal life balance is a potential

occupational stressor. Four potential sources of imbalance were identified: time (e.g., the

amount of time spent at work compared to time spent pursuing non-work activities),

behavior (e.g., accomplishing one's work and non-work goals), strain (e.g., tension or

anxiety resulting from the process of trying to meet work and non-work demands), and
 147

energy (e.g., having energy available to pursue non-work activities after a full-day’s

work).

Items included in the work/life balance measure were derived from the conceptual

definition, comments obtained from employees during one-on-one interviews conducted

as part of the construct definition process, existing measures of work/family conflict (e.g.,

Carlson et al., 2000; Netemeyer et al., 1996), and recommendations made by work/family

researchers (e.g., MacDermid et al., 2000). A total of 31 items were written and

administered to a panel of subject matter experts for content validation. Panelist ratings

of item relevance to the work/life balance construct were useful for reducing the item

pool to a smaller group of appropriate items; seven items were discarded from the initial

item pool because they did not seem to adequately assess the construct domain.

Scale Item Assessment

The twenty-four work/life balance scale items remaining after the initial content

validity assessment were administered to a large, heterogeneous sample of managers as

part of a larger survey. Responses gathered from managers were randomly divided into

two groups; the second group was retained as a holdout sample for cross-validation

purposes.

Three of the twenty-four work/life balance items were removed from the scale

due to poor psychometric properties. Doing so improved the measurement characteristics

of the scale and resulted in a 21-item measure with acceptable psychometric properties

assessing three dimensions: work interference with personal life (WIPL), personal life

interference with work (PLIW), and work/personal life enhancement (WPLE).

 148

Work Interference with Personal Life

WIPL was assessed with 13 items, and demonstrated very high internal

consistency reliability across both samples. The content of most of the items on this scale

specifically assessed the extent to which work has a negative impact on one's personal

life. Content examination of the WIPL items suggested all four components of the

work/life balance definition (e.g. time, energy, goal accomplishment, and strain) were

included in items on the WIPL dimension.

Personal Life Interference with Work

The second dimension of the work/life balance scale included four items that

assessed the extent to which one's personal life interferes with work (PLIW). Internal

consistency reliability coefficients obtained in both samples indicated that the scale had

acceptable levels of reliability.

Compared to WIPL, items on the PLIW scale assessed the opposite direction of

the work/personal interface: the extent to which aspects of one's personal life have a

negative impact on work. Although these four items assessed work/life balance in terms

of goal accomplishment, energy, and strain, none of the items on this dimension

incorporated the other issue identified as a salient aspect of work/life balance: time.

Future scale development efforts should be directed toward writing new items to

incorporate personal life interference with work as it relates to time. Examples of this

type of item would be “My personal life takes away time that I would otherwise spend

doing work,” or “The activities important to me in my personal life interfere with my

work time.”
 149

Work/Personal Life Enhancement

The third work/life balance scale dimension assessed the extent to which one's

personal life is enhanced by work or vice versa. These four items demonstrated

acceptable levels of reliability across two samples of data (a = .77 and a = .76).

Previous empirical research on work/non-work conflict has typically focused on

the negative impact of work on family and vice versa; only recently have other

researchers (e.g., Gryzwacz, 2000; Gryzwacz & Marks, 2000) quantitatively assessed

positive spillover with regard to work and family. The present study extended beyond

this work by examining negative as well as positive influences in more than just family in

the non-work domain. Future research should examine whether there arc directional

differences regarding the positive influence of work on personal life and vice-versa.

Results obtained in both samples regarding the factor structure of the scales

indicated that the three-factor model, with WIPL, PLIW, and WPLE as three separate

dimensions, fit the data well. Although this factor structure was cross-validated in the

holdout sample among a heterogeneous sample of employees across a variety of

organizations, the participants in this sample were primarily white-collar workers in

managerial jobs. Future research should examine the extent to which this factor structure

is invariant in other samples, such as among employees in different occupations.

Item response theory findings indicated that each item in the 21-item scale

discriminated well, although only the WIPL dimension was able to provide information

at high levels of the underlying construct. New items should be developed to increase the

sensitivity of the scale at higher levels of personal life interference with work and

work/personal life enhancement. Increasing the sensitivity of the scale across all levels
 150

of all three dimensions would serve to make the scale more useful for developing models

of work/life balance to understand the relationship work/life balance in the stress process

and to evaluate the effectiveness of work/life balance initiatives in organizations.

Scale Validity

The three subscales of the work/life balance scale were significantly related to one

another. Specifically, WIPE and WPLE were more highly correlated than WIPL and

PLIW, and PLIW and WPLE. Individuals who experienced less work interference with

personal life or personal life interference with work were more likely to perceive that one

domain enhanced the other. These correlations were not sufficiently large enough to

suggest that each dimension was in fact measuring the same thing (r’s ranged from .31 to

.64 across both samples); the empirical evidence suggested that the three scales seemed to

measure distinct dimensions of work/personal life balance.

Convergent Validity

Evidence of convergent validity was assessed in two ways. First, I examined the

correlation between the three dimensions of work/life balance and a one-item overall

measure of the construct (WEB). All three dimensions were significantly related to the

overall measure. However, the strength of the relationship varied significantly across the

three dimensions. WIPL was strongly related to overall work/life balance, suggesting

that they are measuring very similar constructs. The correlation between WPLE and the

one-item measure was also fairly strong, though not as high as the correlation between

WIPL and WLB. Although PLIW was also significantly related to WLB, the relationship

was less strong in comparison to WIPL and WPLE.

 151

These findings suggested that managers’ general perceptions of whether or not

they have a balance between their work and personal life are closely aligned with whether

they perceive their work as conflicting with their personal life and whether their work and

personal lives have positive effects on one another. The extent to which one’s personal

life interferes with work was related to overall perceptions of balance, but not nearly to

the same extent as the other two dimensions. This implies an asymmetry in the

importance of work versus personal life, which is consistent with the finding that more

managers place a higher priority on personal life than work.

Work/family conflict. The second approach to convergent validity assessment

involved comparing correlations of the new scale to measures of other, conceptually-

similar occupational stressors, including work-to-family conflict, family-to-work conflict,

role overload, and time pressure. Because two of the work/life balance scale dimensions

implied directional effects from one-domain to another (i.e., WIPL and PLIW) similar to

WFC and FWC, the respective scales were compared in two ways. First I compared

WIPL to WFC, and then compared PLIW to FWC. The correlations between WIPL and

WFC were very high in both samples (r = .87 and r =.84), which indicated very high

correlations between these two constructs. In other words, managers’ perceptions of

WIPL were very similar to their experience of work-to-family conflict. The correlation

between PLIW and FWC was also fairly high (above .50 in both samples), indicating that

the extent to which one's personal life interfered with work was similar to whether they

perceived family-to-work conflict.

Role overload. According to role theory, the more roles one takes on or is

responsible for, the more role-related tasks or goals exist that need to be accomplished
 152

(Kahn et al., 1964). Role overload refers to having too many behaviors to perform as part

of one’s role. Work/life balance was negatively related to role overload as expected for

both WIPL and WPLE. These results suggest that the more one is expected to do as part

of their job, the more likely their work would interfere with their personal life, and the

less their work would have a positive effect on one’s personal life.

Unlike WIPL and WPLE, however, PLIW and role overload were unrelated. The

role overload measure employed in the present study was specific to managers' work

role, and did not assess the extent to which managers' perceived they had too much to do

in any non-work roles. Therefore, having too much to do at work does not seem to be

related to whether an employee's personal life interferes with work. However, having

too much to do outside of work (i.e., non-work role overload) may be more closely

related to the extent to which one's personal life interferes with work. Future research

should incorporate a more general role overload measure, or a separate measure

pertaining to non-work role overload (e.g., having too much to do at home.)

In addition, managers who responded to this survey did not report a great deal of

personal life interference with work (i.e., M = 1.90, SD = .60 across both samples).

Perhaps PLIW would be more strongly related to work role overload in samples with

more variability in perceptions of PLIW. Future research should be conducted to explore

PLIW in additional samples.

Time pressure. Time pressure is a second specific stressor in which individuals

perceive that they have a limited amount of time to accomplish a task or set of tasks.

Work/life balance was negatively related to time pressure as expected for all three

dimensions of work/life balance. The conservation of resources theory claims that

 153

individuals will experience stress if they are threatened by a potential loss in resources,

experience an actual loss, or fail to gain expected resources (Hobfoll, 1988; 1989).

Taking time into account as a valued and limited resource, it follows from this theory that

one will experience stress if they lose time or fear they might lose time. Because

work/life balance involves time (e.g., the amount of time one spends at work compared to

pursuing non-work activities), it follows that employees with more of a balance between

work and personal life will experience less pressure for time. In other words, the results

seem consistent with the conservation of resources theory.

These results regarding time pressure are consistent with the findings for role

overload, except that time pressure was related to all three dimensions of work/life

balance whereas role overload was only related to two. It was not surprising to find

similar results for time pressure and role overload because these constructs are

conceptually similar (e.g., time pressure is related to having too much to do), and

measures of these two constructs were highly correlated (r = .74 and .79 in the two

samples, respectively). However, given the similarity between the two constructs, then

why was PLIW significantly related to time pressure but not role overload? There are

three possible explanations for these results. First, the time pressure scale was more

internally consistent than role overload (a = .84 for time pressure, compared to a = .70 or

.67 for role overload). According to classical test theory, less internal consistency

reliability indicates a greater degree of measurement error, which attenuates correlations

between variables (Guion, 1998). It follows that the higher the internal consistency of a

scale, the more the correlation between two measures estimates the true relationship

between those variables. In terms of the present studv, the low reliability for role
 154

overload may have attenuated the correlation between role overload and work/life

balance. Had the role overload measure been more reliable, it is possible that PLIW and

role overload would have been more highly correlated, as found for time pressure.

A second explanation pertains to the content of the survey items. Two of the five

time pressure items were not specific to work, whereas all of the role overload items

asked respondents about their work role(s). It could be that respondents took time

pressure in their personal lives into account when responding to these survey items,

resulting in more similar responses between time pressure and PLIW compared to role

overload and PLIW. Additional research using a verbal protocol analysis approach, in

which respondents explain their thought process while responding to survey items, might

provide more insight into how respondents interpret these items.

Finally, a significant relationship between time pressure and work/life balance

could exist because time pressure is more conceptually similar to work/life balance than

role overload. To the extent that time may be a more salient component of the work/life

balance construct than goal accomplishment (i.e., accomplishing the tasks set forth as

part of a role), then time pressure may be more closely related to work/life balance.

Taken together, the pattern of the correlations obtained in both the initial and

holdout samples indicates strong evidence of convergent validity for the new work/life

balance scale. Work/life balance was strongly related to measure of other occupational

stressors, including work/family conflict, role overload, and time pressure.

Discriminant Validity

Evidence for discriminant validity was established by finding lower correlations

between dimensions of work/life balance and other study variables (e.g., overall work
 155

stress, job satisfaction, etc.) than were found between specific work/life balance

dimensions and very similar constructs (e.g., overall work/life balance, work-to-family

conflict, and family-to-work conflict). For example, the general pattern of correlations

obtained across both samples in the present study indicated that the dimensions of

work/life balance were more strongly related to perceptions of overall work/life balance,

and work/family conflict (as in the very high correlations between WIPL and WFC, and

strong relationship between PLIW and FWC) than to most other variables which were not

as conceptually similar. These results indicate that, although work/personal life balance

was related to other variables in this study, it was empirically distinguished from other,

theoretically less similar constructs.

Criterion-related Validity

Overall job stress. Based on the conservation of resources theory, difficulty in

achieving work/life balance was predicted to be a specific job-related stressor. Results in

this study indicated that work/life balance was negatively related to perceptions of overall

job stress. Support for this hypothesis was based on the strong positive correlation found

between work interference with personal life (WIPL) and stress, as well as the strong

negative correlation between work/personal life enhancement (WPLE) and stress.

However, perceptions of personal life interference with work (PLIW) were not as

strongly related to stress.

One explanation for why PLIW was not related to perceptions of stress is

consistent with what was previously described for role overload. Specifically, overall job

stress refers to the amount of strain (e.g., feeling tense, overwhelmed, etc.) one feels on
 156

the job. It seems that work-related stressors rather than personal-life related stressors are

more salient with regard to feeling stress on the job.

Findings among correlations between work/life balance and each dimension of

stress indicated a slightly stronger relationship between work/life balance and pressure

compared to work/life balance and stress. These results may be explained by what is

actually being measured by each dimension of overall work stress. For example,

researchers who developed the overall stress scale employed in the present study (e.g.,

Stanton et al., in press; Fisher et al., 2000; Fisher, Stanton, Grauer, & Smith, 2001) have

suggested that although both dimensions of stress (pressure and threat) assess overall

levels of work stress, pressure assesses more moderate and perhaps benign levels of

strain, whereas threat is indicative of more severe perceptions of strain. Overall,

managers in the present study reported rather high mean levels of pressure, and more

moderate levels of threat. The distribution of overall stress responses was negatively

skewed for pressure, and more normally distributed for threat. As a result, the highly

skewed distribution of pressure may have attenuated the observed correlation between

work/life balance and pressure compared to threat.

Time management behavior. The hypothesis concerning the positive relationship

expected between managers’ time management behavior and their perceptions of

work/life balance was generally supported. In other words, the more these employees

performed a variety of time management behaviors, the more they perceived that they

achieved a balance between their work and personal lives. These results were consistent

with Adams & Jex’s (1999) finding that time management behavior predicted work-to-

family conflict and family-to-work conflict. This relationship was stronger for two
 157

dimensions, WIPL and WPLE, than for WPLE. In particular, preference for organization

was the only dimension of time management behavior that was significantly related to all

three aspects of work/life balance.

The stronger relationship with regard to organization may have been detected

because it is the organizational aspect of time management that helps employees with

being able to manage the demands in both their work and personal lives. In terms of the

conservation of resources theory, efforts toward being we 11-organized may be one way

employees can effectively manage their resources. Conservation of resources theory

suggests that limiting the expenditure of resources (e.g., energy or time) would increase

one’s perception of work/life balance, and therefore reduce or eliminate strains.

Evidence to support this explanation was suggested by the finding across both samples

that preference for organization was negatively related to overall job stress (both pressure

and threat).

Among the work/life balance results pertaining to time management behavior,

WIPL was not at all related to either setting goals and priorities, or planning and

scheduling. This unexpected finding could be due to the fact that the time management

measure asked respondents about their time management behaviors in the context of

work. As a result, this measure may have been more likely to explain variance in how

one’s personal life interferes with work (PLIW) rather than in how one’s work interferes

with personal life, or when personal life has positive effects on work or vice versa

(WPLE). Aspects of one's personal life are more likely to enhance one’s work life if

they focus on setting goals and priorities and do a lot to plan and schedule their time and

tasks at work. For example, having to attend to one's personal needs (e.g., going to a
 158

doctor or dentist appointment) would be more likely to interfere with work if one has a

busy or inflexible schedule, or is very focused on accomplishing important work goals.

Job satisfaction. As predicted, results indicated that work/life balance was related

to job satisfaction for all three dimensions of work/life balance. According to the

conceptual definition of work/life balance, part of achieving this balance involves being

able to accomplish one's goals at work. To the extent that an employee is able to

accomplish her work goals without it interfering with her personal life, she may be more

satisfied with her job. Similarly, one may enjoy her job more if her personal life is not

interfering with accomplishing her work.

Among the three dimensions of work/life balance, work/personal life

enhancement was more strongly related to job satisfaction than WIPL and PLIW. This

finding may be explained by the notion that the positive effects of one's job on his

personal life, or vice versa lead one to develop more favorable impressions of his job.

Employees with jobs that do not have a negative impact on their personal life, but make

positive contributions to their personal life (e.g., by putting them in a better mood or

giving them more energy at home) are likely to be more satisfied with their job in general

compared to employees who do not have the same experience with their work. The

conservation of resources theory indicates that one may experience strain in a situation

where he or she expects to gain resources but does not, or loses resources. Actually

gaining resources may serve to not only inhibit the development of strains, but lead to

beneficial outcomes (e.g., job satisfaction).

Another explanation for the positive relationship is consistent with the congruence

model (Frone et al., 1992; Morf, 1989; Staines, 1980). That is, a third variable (e.g., an
 159

underlying personality trait) may account for similar perceptions regarding work/life

balance and job satisfaction. For example, negative affectivity is one trait that has been

found to account for similarities in perceptions of stressors (e.g., lack of work/life

balance) and strains (e.g., job dissatisfaction; Chen & Spector, 1991)

Life satisfaction. Similar to job satisfaction, managers’ ratings of work/life

balance were highly related to their overall satisfaction with life. As with job

satisfaction, these findings could be explained by perceptions of work/life balance

influencing one’s level of life satisfaction, life satisfaction influencing perceptions of

work/life balance, or perhaps a third variable accounting for both variables.

Organizational commitment. Affective commitment refers to employees’ overall

feelings toward their work organization. As predicted, work/life balance was positively

associated with organizational commitment. This finding is consistent with work/family

conflict research that found that employees who experienced more work-to-family

conflict and family-to-work conflict were less committed to their work organization

(Allen et al., 2000).

The correlations between work/life balance variables and organizational

commitment obtained in the present study may have been slightly attenuated due to low

variability in managers' ratings of organizational commitment. In particular, the majority

of managers who participated in this study had been working for the same organization

for at least five years, and the distribution of mean organizational commitment ratings

was negatively skewed such that most respondents were highly committed to their work

organization. Perhaps results would differ in a sample in which there is more variability

in employees' commitment to their work organization.

 160

Further examination of the relationship between work/life balance and

organizational commitment in the present study indicated a partially mediating effect of

job satisfaction. In other words, results of a hierarchical regression analysis showed that

the majority of the variance in organizational commitment was related to job satisfaction

rather than work/life balance itself; work/life balance variables (WIPL, PLIW, and

WPLE combined) only accounted for a small portion of the variance in organizational

commitment after controlling for job satisfaction. That is, job satisfaction accounted for

31% of the variance in organizational commitment (R2 = .31), and work/life balance

explained an additional 3% (AR2 = .03, F (3, 523) = 8.88, p < .01.). This finding is not

surprising, as the organizational psychology literature is replete with studies that have

documented the strong empirical relationship between job satisfaction and organizational

commitment (e.g., Mathieu, 1991; Mathieu & Farr, 1991; Ostroff, 1992). It is worth

reminding the reader, however, that work/life balance was highly related to job

satisfaction. As a result, it seems that work/life balance is both directly and indirectly

(via its relationship with job satisfaction) related to organizational commitment. Future

research should employ longitudinal research designs and analytic tools available, such as

structural equation modeling (SEM; Maruyama, 1998), to test models of work/life

balance outcomes.

Turnover. Similar to organizational commitment and as predicted, work/life

balance was also related to managers' intentions to quit their present job. Specifically,

managers with more positive perceptions of work/life balance were had lower turnover

intentions. Although these results supported the research hypothesis, the same partial

mediating effect of job satisfaction was found with regard to turnover. Specifically, in
 161

the total study sample, work/life balance did account for unique variability in turnover

beyond job satisfaction (R2 = .37 for job satisfaction and turnover, AR2 = .014,

F (3, 523) = p < .01), as well as beyond both job satisfaction and organizational

commitment (R2 = .53 for job satisfaction and organizational commitment predicting

turnover, AR2= .01, F (3, 523) = p < .01). However, the effect sizes (the AR2 values

obtained in these hierarchical regression analyses) were small, suggesting that work/life

balance predicted little variance in turnover after taking these organizational variables

into account.

Summary of Construct Validity Evidence

Construct validity regarding work/personal life balance was established in four

ways across both studies in this research investigation. First, the item content ratings

made by the subject matter experts during the early stages of scale development (Study

One) provided initial evidence for content validity of the overall scale. Subsequent

examination of the item content once the factor structure had been identified indicated

that two of the three scale dimensions (PLIW and WPLE) did not seem to measure the

work/life balance content domain as well as they could. Nonetheless, empirical results

obtained in two samples in Study Two demonstrated strong initial evidence for

convergent, discriminant, and criterion-related validity of the 21-item work/life balance

measure.

Although not every dimension of work/life balance was related to all dimensions

of each criterion, the findings discussed here indicated strong support for the criterion-

related validity of the work/life balance construct. Given the exploratory nature of this

research and early development of this construct, the results presented here suggest that
 162

work/life balance holds promise for increasing our understanding of the intersection

between employees' work and personal lives, and hence for more clearly understanding

the stress process.

Altogether, the relationships between work/life balance and other individual and

organizational variables were consistent with previous findings regarding the

relationships between work/non-work stressors and strains (Allen et al., 2000; Kossek &

Ozeki, 1998). Findings that work/life balance was significantly related to strains (e.g.,

job satisfaction, overall job stress, etc.) provided empirical evidence that having a lack of

work/life balance is an occupational stressor that is related, directly and indirectly, to a

number of individual and organizational outcomes, such as life satisfaction, job

satisfaction, organizational commitment, and turnover intentions.

Theoretical Implications

The purpose of the present study was not to test Hobfoll’s (1988; 1989)

conservation of resources theory, but rather to use the theory to guide development of the

work/personal life balance construct. This theory, combined with a review of previous

research and qualitative information gathered as part of the present investigation,

identified four potential sources of imbalance between employees’ work and personal

lives: time, goal accomplishment, energy, and strain. The significant correlations

between work/life balance and other stressor and strain measures were consistent with

predictions made by Hobfoll’s (1988; 1989) theory'. Significant correlations between

work/life balance and other variables in the expected direction indicated that not being

able to conserve one’s resources was associated with higher perceptions of other stressors

and strains. To the extent that time, energy, goal accomplishment, and lack of strain are
 163

perceived as equally salient resources, the number and content of the items on each of the

work/life balance sub-scales should reflect that. As a result, future scale development

work is warranted.

Heuristic models of organizational stress have suggested that stress is a process in

which characteristics of an individual as well as characteristics of one’s work and non­

work environment may lead to a number of potential outcomes for both the individual

and the organization for which she works. This study demonstrated that a lack of

work/life balance is one possible stressor that is related to other stressors (e.g., role

overload and time pressure) as well as strains (e.g., life dissatisfaction, job dissatisfaction,

low organizational commitment, and high turnover intentions). Future research is needed

to develop more specific models regarding the role of work/life balance in the stress

process by identifying antecedents and other outcome variables.

Comparison of Work/Life Balance to Work/Family Conflict

The purpose of the present study was to develop a new construct to extend

previous work/family conflict research both conceptually and empirically. However, one

remaining question is whether the work/life balance scale developed in the present study

demonstrated any utility over a widely-used measure of work/family conflict. With

regard to content validity, the work/life balance scale items were more inclusive of the

non-work domain than Netemeyer et al.'s (1996) measure of work/family conflict.

Specifically, the items in the work/life balance scale included the term “personal life” so

that respondents could interpret the item in the context of what was salient to their own

respective personal lives. This more general approach to item wording is desirable for
 164

being able to assess more of the non-work content domain than just family, as family may

not be relevant to all employees.

Secondly, unlike work/family measures that only assess tension or interference

(e.g., work-to-family conflict and family-to-work conflict; e.g., Netemeyer et al., 1996),

the scale developed as part of the present study also measured positive spillover (i.e.,

enhancement) between work and personal life. Findings in the present study indicated

that the enhancement dimension of work/life balance accounted for additional variance in

some of the criterion variables after accounting for interference, and therefore it is

important to consider the positive effects of one domain on the other. Findings from this

study indicated that both positive and negative work and personal life issues are relevant

to conceptualizing “balance.” In other words, balance does not simply refer to having a

lack of interference, but also includes the extent to which one’s personal life benefits

from work and vice-versa. Reliance on measures that only assess conflict or interference

seems to be limiting if the purpose is to make accurate predictions based on individuals

total perceptions of the boundary between work and non-work.

Empirically, the WIPL (work interference with personal life) sub-scale of

work/life balance was highly correlated with work-to-family conflict (i.e., r’s above .80

in both samples). Not surprisingly, then, was the pattern of relationships with other

variables similar for WIPL and WFC. However, the PLIW scale (personal life

interference with work) was superior to the family-to-work conflict measure in terms of

its ability to predict other variables. For example, PLIW was at least moderately related

to time pressure, time management behavior (preference for organization), job

satisfaction, organizational commitment, and turnover intentions, whereas FWC was not.
 165

Furthermore, PLIW was more strongly related to overall stress and life satisfaction than

FWC. These results suggest that there may be more to one's personal life than just

family that can interfere or conflict with work, and the FWC measure was criterion

deficient in comparison to the 4-item PLIW scale. Taken together, the more broad item

content, inclusion of positive as well as negative aspects of work and personal life, and

superior predictive validity of the work/life balance scale suggest that the new scale is

superior to the measure of work/family conflict employed in the present study.

Practical Implications

In addition to contributing to our knowledge of occupational stressors and strains,

the development of a new construct to measure work/life balance has a number of

practical implications. First, many organizations have been developing work/life

programs without regard to explaining what is meant by the term, “work/life balance.”

(Hill et al., 1998) The construct definition developed as part of this study can assist

organizations in articulating what they mean when marketing their ability to help

employees attain work/personal life balance (Messmer, 1999).

Secondly, the scale developed as part of the present study holds promise as a

useful tool for organizations interested in assessing work/life balance perceptions among

employees. To date, no other measure of work/life balance has been developed and

validated. The more inclusive wording of personal life compared to family offers

organizations the opportunity to measure the interface between work and personal life

among employees without regard to employees' marital status and family size. This scale

could also be used to evaluate the effectiveness of work/life balance programs. The

multiple scale dimensions afford an opportunity to assess interference and enhancement

 166

separately. Although work/life balance programs designed to reduce work/personal life

interference could be effective for reducing strains, this study provides initial evidence to

suggest that programs designed to enhance employees’ personal lives via work will be

particularly useful for improving employee attitudes and retaining valued employees.

To increase the practical utility of the work/life balance scale, the work

interference with personal life (WIPL) scale could be reduced in length while retaining its

psychometric properties with regard to reliability, content validity, and predictive

validity. Reducing the scale could ease the survey administration process and may have a

positive impact on response rate.

Limitations

One limitation of the present study was the use of a single sample design using a

holdout sample for cross-validation of the scale. This single sample methodology has

received some criticism (e.g., Murphy, 1983). Although the use of a holdout sample

assists a researcher with adjusting for random sampling error, this approach does not

compensate for violations of sampling assumptions. In other words, any anomalous

results specific to a single sample would likely be evident in both the initial and holdout

sample. Therefore, future research should be conducted to continue the scale validation

process using additional samples.

Although the sample used in the present study was heterogeneous, (i.e.,

respondents were employed by organizations that differ in size, industry, and

geographical location), all participants were managers, few were under the age of 40,

most were married with children, and most had been with the same organization for at

least 5 or 10 years. Consequently, their perceptions may differ from non-managers,

 167

employees in other occupations, younger employees, unmarried and/or childless workers,

or those with less experience with a particular organization. Samples obtained for future

scale development research should be conducted by sampling from different populations.

All data collected from managers in this study relied on their self-reports of the

variables assessed in this study. The use of self-report questionnaires has been a source

of debate in organizational psychology literature (e.g., Spector, 1994; Schmitt, 1994;

Howard, 1994), and has been criticized for resulting in artificially inflated correlations

among measures of psychological constructs. Although method bias due to the use of a

cross-sectional, self-report research design may have artificially inflated the magnitude of

the observed correlations, this effect would not alter the direction of the observed

relations between variables.

The constructs included in the present study were most appropriately measured by

asking employees to report their own attitudes and perceptions (Schmitt, 1994). For

example, work/life balance was conceptualized as an occupational stressor, and according

to the cognitive appraisal definition of stress (Lazarus, 1966; Lazarus & Folkman, 1984),

perceptions of stress are in the eye of the beholder. Therefore, the use of self-report

questionnaires was the most appropriate methodology. The same rationale was used for

using self-report methodology to gather information about other stressors (e.g., time

pressure and role overload).

Because this study was exploratory and early in the development of work/life

balance research, this methodology has been suggested as appropriate for gathering

information to assess initial construct validity (e.g., Schmitt, 1994). Future research to

gather further validity evidence should incorporate other methods of measurement (e.g.,
 168

coworkers’ or family members’ perceptions) and employ longitudinal research designs to

rule out the explanation that cross-sectional, self-report research methodology accounted

for any of the results obtained in this study.

Overall Summary and Conclusions

Despite these limitations, the present investigation extended previous research on

work/non-work conflict by defining the construct of work/personal life balance and

developing a measure of the construct for assessing employees’ level of work/life

balance. Item and scale analyses based on both classical test theory and item response

theory across two samples of data demonstrated initial evidence for appropriate scale

psychometric properties. In addition, data provided initial evidence for the convergent,

discriminant, and criterion-related validity of the construct.

Overall, work/life balance was defined as an occupational stressor based on lost

resources of time and energy, lack of goal accomplishment, and strain between work and

personal life role demands. This study identified three specific dimensions of work/life

balance: work interference with personal life, personal life interference with work, and

work/personal life enhancement. Empirically, a lack of work/life balance was related to a

host of strains, including higher perceptions of stress, job dissatisfaction, life

dissatisfaction, lack of organizational commitment, and intent to quit one’s job. Results

indicated support for measuring positive as well as negative aspects of the work/personal

life interface, as well as incremental validity beyond work/family conflict.

Clark and Watson (1995) commented that the work of validating a scale is never

complete. Given the early stage of development for the work/life balance construct,

much additional research is warranted to improve the work/life balance scale and
 169

continue understanding its role in the stress process. It is the author’s hope that continued

research on work/personal life balance will lead to efforts for improving the quality of

peoples’ work and personal lives.

 170

REFERENCES

Adams, G. A. & Jex, S. M. (1999). Relationships between time management,

control, work-family conflict, and strain. Journal of Occupational Health Psychology, 4,

72-77.

Adams, G. A. & Jex, S. M. (1997). Confirmatory factor analysis of the time

manafement behavior scale. Psychological Reports, 80, 225-226.

Adams, G. A., King, L. A., & King, D. W. (1996). Relationships of job and

family involvement, family social support, and work-family conflict with job and life

satisfaction. Journal of Applied Psychology, 81,411-420.

Allen, T. D., Herst, D. E. L., Bruck, C. S., & Sutton, M. (2000). Consequences

associated with work-to-family conflict: A review and agenda for future research.

Journal of Occupational Health Psychology, 5, 278-308.

Arbuckle, J. (1999). Amos Graphics software, version 4.0. Smallwaters Corp.

Balzer, W. K., Kihm, J. A., Smith, P. C., Irwin, J. L., Bachiochi, P. D., Robie, C.,

Sinar, E. F., & Parra, L. F. (1997). Users' manual for the Job Descriptive Index (JDI;

1997 Revision) and the Job In General scales. Bowling Green, OH: Bowling Green State

University.

Barnett, R. & Rivers (1996). She works, he works: How two-income families are

happy, healthy, and thriving. Cambridge, MA: Harvard University Press.

Bedeian, A. G., Burke, B. G., & Moffett, R. G. (1988). Outcomes of work-family

conflict among married male and female professionals. Journal of Management, 14,

475-491.
 171

Beehr, T. A. & Newman, J. E. (1978). Job stress, employee health, and

organizational effectiveness: A facet analysis, model, and literature review. Personnel

Psychology, 31, 665-699.

Beehr, T. A., Walsh. J. T., & Taylor, T. D. (1976). Relationship of stress to

individually and organizationally valued states: Higher order needs as a moderator.

Journal of Applied Psychology, 61,41 -47.

Bentler, P. M. (1990). Comparative fit indices in structural models.

Psychological Bulletin, 107, 238-246.

Bentler, P. M. & Bonett, D. G. (1980). Significance tests and goodness of fit in

the analysis of covariance structures. Psychological Bulletin, 88, 588-606.

Bohen, H. H. & Viveros-Long. A. (1981). Balancing jobs and family life: Do

flexible schedules really help? Philadelphia: Temple University Press.

Bollen, K. A. (1989). Structural equations with latent variables. New York: John

Wiley.

Browne, M. W. & Cudeck, R. (1992). Alternative ways of assessing model fit.

Sociological Methods and Research, 21,230-258.

Burke, R. J. (1988). Some antecedents and consequences of work-family conflict.

Journal of Social Behavior and Personality, 3, 287-302.

Carlson, D. S., Kacmar. K. M.. & Williams, L. J. (2000). Construction and initial

validation of a multidimensional measure of work-family conflict. Journal of Vocational

Behavior, 56, 249-276.

Casner-Lotto, J. & Hickey (1999). Holding a job, having a life - Making them

both possible. Employment Relations Today, 25, 37-47.

 172

Cattell, R. B. (1978). The scientific use of factor analysis in behavioral and life

sciences. New York: Plenum.

Chen, P. Y., & Spector, P. E. (1991). Negative affectivity as the underlying cause

of correlations between stressors and strains. Journal of Applied Psychology. 76, 398-

407.

Clark, L. A., & Watson, D. (1995). Constructing validity: Basic issues in

objective scale development. Psychological Assessment. 7. 309-319.

Cook, T. D., & Campbell, D. T. (1976). Quasi-experimentation : Design &

analysis issues for field settings. Boston : Houghton Mifflin.

Cortina, J. M. (1993). What is coefficient alpha? An examination of theory and

applications. Journal of Applied Psychology, 78, 98-104.

Cranny, C. J., Smith, P. C., & Stone, E. F. (1992). Job satisfaction: How people

feel about their jobs and how it affects their performance. New York: Lexington Books.

Cronbach, L. J. (1971). Test validation. In R. L. Thorndike (Ed.), Educational

Measurement (2nd ed.). Washington, D.C.: American Council on Education.

Cronbach, L. J. & Meehl, P. E. (1955). Construct validity in psychological tests.

Psychological Bulletin, 52, 281-302.

Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The satisfaction

with life scale. Journal of Personality Assessment, 49, 71-75.

Drasgow, F. & Hulin, C. L. (1990). Item response theory. In M. D. Dunnette &

L. M. Hough (Eds.), Handbook of industrial and organizational psychology, (2nd ed.).

Palo Alto, CA: Consulting Psychologists Press.

 173

Duxbury, L. E. & Higgins, C. A. (1991). Gender differences in work-family

conflict. Journal of Applied Psychology, 76, 60-74.

Eagle, B. W., Miles, E. W., & Icenogle, M. L. (1997). Interrole conflict and the

permeability of work and family domains: Are there gender differences? Journal of

Vocational Behavior, 50, 168-184.

Edwards, J. R. (1992). A cybernetic theory of stress, coping, and well-being in

organizations. Academy of Management Review, 17, 238-274.

Edwards, J. R. & Rothbard, N. P. (2000). Mechanisms linking work and family:

Clarifying the relationship between work and family constructs. Academy of

Management, 25, 178-199.

Evans, P. & Bartolomé, F. (1984). The changing pictures of the relationship

between career and family. Journal of Occupational Behavior. 5, 9-21.

Figart, D. M. & Golden, L. (2000). Introduction and overview: understanding

working time around the world. In L. Golden & D. M. Figart (Eds), Working time:

International trends, theory, and policy perspectives. New York: Routledge.

Fisher, G. G. & Hemingway, M. A. (2000). Is there more to life than work and

family? Symposium presented at the 15th Annual Conference for the Society of Industrial

and Organizational Psychology, New Orleans, LA, April 14-16.

Fisher, G. G., Stanton, J. M., & Grauer, E., & Smith, C. S. (2001). Further

examination of the pressure and threat dimensions of the stress-in-general scale.

Manuscript in progress.

Fisher, G. G., Stanton, J. M., Thoresen, P., Julian, A. L., Sinar, E. F., Aziz, S„

Balzer, W. K.., & Smith, P. C. (2000). The stress-in-general scale: Exploration and
 174

validation. Poster presented at the 15th Annual Conference for the Society of Industrial

and Organizational Psychology, New Orleans, LA, April 14-16.

Ford, J. K., MacCallum, R. C., & Tait, M. (1986). The application of exploratory

factor analysis in applied psychology: A critical review and analysis. Personnel

Psychology, 39, 291-314.

French, J. R. P., & Kahn, R. L. (1962). A programmatic approach to studying the

industrial environment and mental health. The Journal of Social Issues, 18, 1-47.

Frone, M. R., Russell, M., & Cooper, M. L. (1992). Antecedents and outcomes of

work-family conflict: Testing a model of the work-family interface. Journal of Applied

Psychology, 77, 65-78.

Frone, M. R., Yardley, J. K.. & Market K. S. (1997). Developing and testing an

integrative model of the work-family interface. Journal of Vocational Behavior, 50, 145-

167.

Ganster, D. C. & Schaubroeck. J. (1991). Work and stress and employee health.

Journal of Management, 17, 235-271.

Good, L. K., Sisler, G. F., & Gentry, J. W. (1988). Antecedents of turnover

intentions among retail management personnel. Journal of Retailing. 64, 295-314.

Grandey, A. A. & Cropanzano, R. (1999). The conservation of resources model

applied to work-family conflict and strain. Journal of Vocational Behavior, 54. 350-370.

Greenhaus, J. H. & Beutell, N. J. (1985) Sources of conflict between work and

family roles. Academy of Management Review, 10, 76-88.

 175

Greenhaus, J. H. & Parasuraman, R. (1986). A work/non-work interactive

perspective of stress and its consequences. Journal of Organizational Behavior

Management, 8, 37-60.

Gryzwacz, J. G. (2000). Work-family spillover and health during midlife: Is

managing conflict everything? American Journal of Health Promotion, 14, 236-243.

Grzywacz, J. G. & Marks, N. (2000). Reconceptualizing the work-family

interface: An ecological perspective on the correlates of positive and negative spillover

between work and family. Journal of Occupational Health Psychology. 5, 111-126.

Guion, R. (1998). Assessment, measurement, and prediction for personnel

decisions. Mahwah, NJ: Lawrence Earlbaum Associates.

Gupta, N., & Beehr, T. A. (1979). Job stress and employee behavior.

Organizational behavior and Human Performance, 23, 373-387.

Gutek, B. A., Searle. S., & Klepa, L. (1991). Rational versus gender role

expectations for work-family conflict. Journal of Applied Psychology, 76, 560-568.

Hambleton, R. K., Swaminathan, H., & Rogers, H. J. (1991). Fundamentals of

item response theory. Newbury Park, CA: Sage.

Hammer, L. B., Allen, E., & Grigsby, T. D. (1997). Work-family conflict in dual­

earner couples: Within-individual and crossover effects of work and family. Journal of

Vocational Behavior, 50, 185-203.

Hart, P. M. (1999). Predicting employee life satisfaction: A conherent model of

personality, work and nonwork experiences, and domain satisfactions. Journal of Applied

Psychology, 84, 564-584

 176

Hill, E. J., Miller, B.C., Weiner, S.P., & Colihan, J. (1998). Influences of the

virtual office on aspects of work and work/life balance. Personnel Psychology, 51,

667-683.

Hobfoll, S. E. (1988). The ecology of stress. New York: Hemisphere.

Hobfoll, S. E. (1989). Conservation of resources: A new attempt at

conceptualizing stress. American Psychologist, 44, 513-524.

Howard, G. S. (1994). Why do people say nasty things about self reports?

Journal of Organizational Behavior, 15, 399-404.

Hu, L. & Bentler, P. M. (1995). Evaluating model fit. In R. H. Hoyle (Ed.),

Structural equation modeling: Concepts, issues, and applications (pp. 158-176).

Thousand Oaks. CA: Sage.

Ironson, G. H. (1992). Job stress and health. In C. J. Cranny, P. C. Smith, & E. F.

Stone (Eds.), Job satisfaction: How people feel about their jobs and how it affects their

performance (pp. 219-239). New York: Lexington Books.

Ironson, G. H., Smith, P. C., Brannick, M. T., Gibson, W. M., & Paul, K. B.

(1989). Construction of a “Job-in-General"' scale: A comparison of global, composite,

and specific measures. Journal of Applied Psychology, 74, 193-200.

Jackson, S. E. & Schuler, R. S. (1985) A meta-analysis and conceptual critique of

research on role ambiguity and role conflict in work settings. Organizational Behavior

and Human Decision Processes, 36, 16-78.

Jex, S. M. & Beehr, T. A. (1991). Emerging theoretical and methodological issues

in the study of work-related stress. In G. R. Ferris, & K. M. Rowland (Eds.), Research in

personnel and human resources management. Greenwich, CT: JAI Press, Inc.
 177

Kahn, R. L., Wolfe, D. M., Quinn, R., Snoek, J. D., & Rosenthal, R.A. (1964).

Organizational Stress. New York: Wiley.

Katz & Kahn, R. L. (1978). The social psychology of organizations (2nd ed.).

New York: John Wiley & Sons.

King, L. A., & King, D. W. (1990). Role conflict and role ambiguity: A critical

assessment of construct validity. Psychological Bulletin, 107, 48-64.

Kopelman, R. E., Greenhaus, J. H.. & Connolly, T. F. (1983). A model of work,

family, and interrole conflict: A construct validation study. Organizational Behavior and

Human Decision Processes, 32, 198-215.

Kossek, E. & Ozeki, C. (1998). Work-family conflict, policies, and the job-life

satisfaction relationship: A review and directions for organizational behavior - human

resources research. Journal of Applied Psychology, 83, 139-149.

Lazarus, R. S. (1966). Psychological stress and the coping process. New York:

McGraw Hill.

Lazarus R., S. & Folkman, S. (1984). Stress, appraisal, and coping. New York:

Springer.

Lerner, J. V. (1994). Working women and their families. Thousand Oaks, CA:

Sage.

Macan, T. H. (1994). Time management: Test of a process model. Journal of

Applied Psychology, 79, 381-391.

Macan, T., Shahani, C., Dipboye, R. L., & Phillips, A. P. (1990). College

students’ time management: Correlations with academic performance and stress. Journal

of Educational Psychology, 82, 760-768.

 178

MacCallum, R. (1998). Commentary on quantitative methods in 1-0 research.

The Industrial-Organizational Psychologist, 35, 19-30.

MacDermid, S. M., Barnett, R., Crosby, F., Greenhaus, J., Koblenz, M.. Marks,

S., Perry-Jenkins, M.. Voydanoff, P., Wethington, E., & Sabbatini-Bunch, L. The

measurement of work/life tension: Recommendations of a virtual think tank. [On-line].

Available: http://www.bc.edu/bc_org/avp/csom/cwf/thinktanks/worklife tension vtt.pdf

Marsh, H. W., Balia, J. R., & McDonald, R. P. (1988). Goodness of fit indices in

confirmatory factor analysis: The effect of sample size. Psychological Bulletin. 103,

391-411.

Maruyama, G. M. (1998). Basics of structural equation modeling. Thousand

Oaks, CA: Sage.

Mathieu, J. E. (1991). A cross-level non-recursive model of the antecedents of

organizational commitment and satisfaction. Journal of Applied Psychology, 76.

607-618.

Mathieu, J. E., & Farr, J. L. (1991). Further evidence foe the discriminant

validity of measures of organizational commitment, job involvement, and job

satisfaction. Journal of Applied Psychology, 76, 127-133.

McClough, A. C., Rogelberg, S. G., Fisher, G. G., & Bachiochi, P. D. (1998).

Cynicism and the quality of an individual's contribution to an organizational diagnostic

survey. Organization Development Journal. 16, 31-41.

McDonald, R. P. & Marsh, H. W. (1990). Choosing a multivariate model: Non­

centrality and goodness of fit. Psychological Bulletin, 107, 247-255.

Medsker, G. J., Williams, L. J., & Holahan, P. J. (1994). A review of current

 179

practices for evaluating causal models in organizational behavior and human resources

management research. Journal of Management, 20, 439-464.

Merton, R. K. (1957). Social theory' and social structure. Glencoe, IL: Free Press

of Glencoe.

Messmer, M. (1999). Attracting quality job candidates. Business Credit, 101, 55-

57.

Mobley, W.H., Homer, S. O., & Hollingsworth, A. T. (1978). An evaluation of

the precursors of hospital employee turnover. Journal of Applied Psychology, 63, 408-

414.

Morf, M. (1989). The work/life dichotomy: prospects for reintegrating people

and jobs. Westport, CT : Quorum Books.

Mowday, R. T., Steers, R. M., & Porter, L. W. (1979). The measurement of

organizational commitment. Journal of Vocational Behavior. 14, 224-247.

Murphy, K. R. (1983). Fooling yourself with cross-validation: Single sample

designs. Personnel Psychology. 36, 111-118.

Murphy, L. R., Hurrell, J. J. Jr., Sauter, S. L., & Keita, G. P. (Eds.) (1995). Job

stress interventions. Washington, DC: American Psychological Association.

Netemeyer, R. G., Boles, J. S., & McMurrian, R. (1996). Development and

validation of work-family conflict and family-work conflict scales. Journal of Applied

Psychology, 81,400-410.

Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New' York: McGraw-Hill.

 180

OstrofT, C. (1992). The relationship between satsifaction, attitudes, and

performance: An organizational level analysis. Journal of Applied Psychology, 77, 963-

974.

Parcel, T. L. (1999). Work and family in the 21st century. Work and

Occupations, 26, 264-274.

Paulhus, D. L. (1984). Two-component model of socially desirable responding.

Journal of Personality and Social Psychology, 46, 598-609.

Paulhus, D. L. (1988). Assessing self-deception and impression management in

self-reports: The Balanced Inventory of Desirable Responding—Reference Manual.

Unpublished manuscript.

Parsons, T. (1951). The social system. Glencoe, IL: Free Press of Glencoe.

Pavot, W., & Diener, E. (1993). Review of the satisfaction with life scale.

Psychological Assessment, 2, 164-172.

Pelletier & Lutz (1989). Mindbody goes to work: A critical review of stress

management programs in the workplace. Advances. 6, 28-34.

Pelletier, K. R. & Lutz, R. (1991). Healthy people-healthy business: A critical

review of stress management programs in the workplace. In Weiss, S. M. & Fielding, J.

E. (Ed.), Health at work, (pp. 189-204). Hillsdale, NJ: Lawrence Erlbaum Associates.

Piotrkowski, C. S. (1978). Work and the family system: A naturalistic study of

working-class and lower-middle-class families. New York: Free Press.

Quick, J. D., Horn, R. S. & Quick, J. C. (1987). Health consequences of stress.

Journal of Organizational and Behavioral Management. 8, 19-36.

 181

Quinn, R.P. & Staines, G. L. (1979) The 1977 Quality of Employment Survey.

Ann Arbor: University of Michigan, Institute for Social Research.

Reeve, C. L. & Smith, C. S. (2001). Refining Lodahl and Kejner’s job

involvement scale with a convergent evidence approach: Applying multiple methods to

multiple samples. Organizational Research Methods, 4, 91-111.

Rothausen, T., Gonzalez, G. A., Clarke, N. E., & O'Dell, L. A. (1998). Family-

friendly backlash - Fact or fiction? The case of organizations' on-site childcare centers.

Personnel Psychology. 51,685-706.

Rush, M. C., Schoel, W. A., & Barnard, S. M. (1995). Psychological resiliency in

the public sector: “Hardiness” and pressure for change. Journal of Vocational Behavior.

46, 17-39.

Sandman, B. A. (1992). The measurement of job stress: Development of the Job

Stress Index. In C. J. Cranny, P. C. Smith, & E. F. Stone (Eds.), Job satisfaction: How

people feel about their jobs and how it affects their performance (pp. 241-254). New

York: Lexington Books.

Sandman, B.A., & Smith, P. C. (1988). The job stress index. Unpublished

manuscript.

Samejima, F. (1969). Estimation of latent ability using a response pattern of

graded scores. Psychometrika Monograph Supplement, 17 (4 Pt 2).

Sauter, S. L., & Murphy, L. R. (1995). Organizational risk factors for job stress.

Washington, D.C.: American Psychological Association.

 182

Schein, E. H. (1976). Career development: Theoretical and practical issues for

organizations. Career planning and development. Geneva: International Labour Office,

9-48.

Schmitt, N. (1994). Method bias: The importance of theory and measurement.

Journal of Organizational Behavior, 15, 393-398.

Seashore, S. E., Lawler, E. E., Mirvis, P., & Cammann, C. (1982). Observing

and measuring organizational change: A guide to field practice. New York: Wiley.

Sekaran, U. (1985). The paths to mental health: An exploratory study of husbands

and wives in dual-career families. Journal of Occupational Psychology, 58, 129-137.

Selye, H. (1956). The stress of life. New York: Mcgraw-Hill.

Shamir, B, (1992). Home: The perfect workplace? In S. Zedeck (Ed.), Work,

families, and organizations. San Francisco: Jossey-Bass.

Shellenbarger, S. (January 29, 1997). New job hunters ask recruiters, “Is there a

life after work?” The Wall Street Journal, BI.

Small, S. A. & Riley, D. (1990). Toward a multidimensional assessment of work

spillover into family life. Journal of Marriage and the Family, 52, 51-61.

Smith, C. S., & Sulsky, L. M. (In press). Work and stress. Pacific Grove, CA:

Wadsworth.

Spector, P. E. (1994). Using self-report questionnaires in OB research: a

comment on the use of a controversial method. Journal of Organizational Behavior, 15,

385-392.

Staines, G. L. (1980). Spillover versus compensation: A review of the literature

on the relationship between work and non-work. Human Relations, 33, 111-129.
 183

Stanton, J. M., Balzer, W. K., Smith, P. C., Parra, L. F„ & Ironson, G. (In press).

A global measure of work stress: The stress in general (SIG) scale. Educational and

Psychological Measurement.

Stanton, J. M., Sinar, E. F., Balzer, W. K., Julian, A. L., Thoresen, P., Aziz, S.,

Fisher, G. G., & Smith, P. C. (In press). Development of a compact measure of job

satisfaction: The abridged job descriptive index. Educational and Psychological

Measurement.

Steiger, J. II. (1990). Structural model evaluation and modification: An interval

estimation approach. Multivariate behavioral research, 25, 173-180.

Sweet, J. A. (1989). Changes in the life cycle composition of the United States

population and the demand for housing. CDE Working Paper, No. 89-14. Madison,

Wisconsin: University of Wisconsin, Center for Demography and Ecology.

Tabachnick, B. G. & Fidell, L. S. (1996). Using multivariate statistics (3rd ed.).

New York: Harper Collins.

Thissen, D. (1991). MULTILOG: Multiple, categorical item analysis and test

scoring using item response theory (Version 6). Chicago: Scientific Software.

Thomas, L. & Ganster, D. C. (1995). Impact of family-supportive work variables

on work/family conflict and strain: A control perspective. Journal of Applied

Psychology, 80, 6-15.

Tucker, L. R., & Lewis, C. (1973). A reliability coefficient for maximum

likelihood factor analysis. Psychometrika, 38, 1-10.

 184

Ventura, S. J., Martin, J. A., Curtin, S. C., & Mathews, T. J. (1998). Monthly

Vital Statistics Report, Vol. 46, No. 11, supp. Hyattsville, MD: National Center for

Health Statistics.

Viswesvaran, C., Sanchez, J. I., & Fisher, J. (1999). The role of social support in

the process of work stress: A meta-analysis. Journal of Vocational Behavior, 54, 314-

334.

Voydanoff, P. (1988). Work role characteristics, family structure demands, and

work/family conflict. Journal of Marriage and the Family. 50, 749-761.

Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of

brief measures of positive and negative affect: The PANAS Scales. Journal of

Personality and Social Psychology, 54, 1063-1070.

Williams, K. J. & Alliger, G. M. (1994). Role stressors, mood spillover, and

perceptions of work-family conflict in employed parents. Academy of Management

Journal, 37, 837-868.

Zedeck, S. (1992). Introduction: Exploring the domain of work and family

concerns. In S. Zedeck (Ed.), Work, families, and organizations. San Francisco: Jossey-

Bass.

Zedeck, S. & Mosier, K. L. (1990). Work in the family and employing

organization. American Psychologist, 45, 240-251.

Zickar, M.J. (In press). Modeling data with polytomous item response theory. In

F. Drasgow & N. Schmitt (Eds.), Measuring and analyzing behavior in organizations:

Advances in measurement and data analysis. San Francisco: Jossey-Bass.

 185

APPENDIX A

Preliminary Study Interview Questions

1. How do you define work/life balance?

2. What aspects of your life outside of work come to mind?

3. How do you spend your time outside of work?

4. How important are each of these activities to you?

5. What constitutes a “balance” for you between work and aspects of your life aside
from your primary job?

6. Do you feel that you currently have a balance? If so, how do you know? If not, what
indicates to you that you don’t have a balance?

7. Do you perceive that striking a balance between work and life outside of work is an
important goal to obtain? Why/ why not?

8. What (in particular) do you do to balance your work and life outside of work?

9. Is it easier to balance some roles than others (including work)? How is it easier/more
difficult? Why do you think this is?

10. What effect(s) does NOT having this balance have on your work? On your life
outside of work? Please explain.

11. What aspects of your job help/hinder you in achieving a balance between work and
life outside of work? Please explain.

12. Is the company you work for mindful of your having a life outside of work? If yes, in
what way?

13. Is your supervisor supportive of your life outside of work? Please explain.

14. Do you have any other comments that come to mind now that you have given some
thought to this issue of work/life balance and the roles that you have in your work and
life outside of work? What other issues seem relevant that weren’t asked about?
 186

APPENDIX B

Work/Life Balance Scale Items in Study One (Content Validation)

1. There is not enough time in the day to do everything I need to do at work and
at home.
2. I struggle with trying to juggle both my work and non-work responsibilities.
3. I spend more time at work than I would like.
4. My personal life drains me of the energy I need to do my job.
5. I fee, overwhelmed when I try to balance my work and personal life.
6. My work suffers because of everything going on in my personal life.
7. I am too tired to be effective at work because of things I have going on in my
personal life.
8. My job gives me energy to pursue activities outside of work that are important
to me.
9. I have difficulty scheduling vacation time because of my workload.
10. I am unable to relax at home because I am preoccupied with my work.
11.1 am happy with the amount of time I spend doing activities not related to
work.
12. Because of my job, I am in a better mood at home.
13.1 often have to make difficult choices between my work and my personal life.
14.1 would spend more time at work if it were not for my non-work
responsibilities.
15. My career interferes with my ability to have a satisfying personal life.
16. I have to put aspects of my personal life “on hold” because of my work.
17. Due to all the pressures at work, I am too drained to do other activities I enjoy.
18.1 am able to accomplish what I would like in both my personal and work lives.
19.1 am in a better mood at work because of everything I have going for me in
my personal life.
20.1 often neglect my personal needs because of the demands of my work.
21. When I am at work, I worry about things I need to do outside of work.
22. 1 have difficulty getting my work done because I am preoccupied with
personal matters at work.
23. My personal life suffers because of my work.
24.1 have to miss out on important personal activities due to the amount of time I
spend doing work.
25.1 am satisfied with my ability to get things done both at home and at work.
26. My personal life gives me the energy to do my job.
27.1 feel that I allocate appropriate amounts of time to both work and non-work
activities.
28.1 make personal sacrifices to get work done.
29.1 come home from work too tired to do things I would like to do.
30. My job makes it difficult to maintain the kind of personal life I would like.
31. My work allows me enough time to do the things that are important to me.
 187

APPENDIX C

Questionnaire Cover Letter

American Management
Association

February 2001

Dear Friend:

Pie American Management Association's Research Department is pleased to assist Bowling

Green State University in a new survey project on Work-Life Balance.

By completing and returning the enclosed questionnaire, you’ll realize these benefits:

• You’ll receive a summary of and analysis of key findings in advance of the

public release of survey results.
• You’ll gain access to a password-protected area of AMA’s research
website where you can view the complete databooks of not only this survey
but ALL AMA surveys conducted in the 2000-2001 season.
• You'll have the opportunity to purchase, at a deep discount, a CD-ROM of
the raw data files for your own query and tabulation.

Your confidentiality is assured; neither you nor your organization will be identified to any outside
party without your advance, written consent.

Won't you take the few minutes necessary to complete this questionnaire and return it to us in the
enclosed postage-paid envelope? We thank you very much for your consideration, and we look
forward to your participation. Please feel free to contact me at [email protected] if you
have any questions or concerns.

Sincerely,

Eric Rolfe Greenberg

Director of Management Studies
 188

APPENDIX D

Work/Life Balance Scale in Study Two

For each of the following statements, please indicate HOW FREQUENTLY YOU HAVE
FELT THIS WAY DURING THE PAST 3 MONTHS using the scale below. If the
statement is not applicable to you. please fill in “N/A” (Not Applicable) as your response
to that question. Please note that for the following group of statements:

“Work” refers to YOUR PRESENT JOB

Work Interference with Personal Life

Almost
Not at Rarely Sometimes Often All the Not
All Time Applicable

I. I struggle with trying to juggle both my 1 2 3 4 5 N/A

work and non-work responsibilities.
3. I feel overwhelmed when I try to balance 1 2 3 4 5 N/A
my work and personal life.
7. I have difficulty scheduling vacation time 1 2 3 4 5 N/A
because of my workload.
8. Iam unable to relax at home because I arr 1 2 3 4 5 N/A
preoccupied with my work.
9. I am happy with the amount of time I 1 2 3 4 5 N/A
spend doing activities not related to work.
II. 1 often have to make difficult choices 1 2 3 4 5 N/A
between my work and my personal life
12.1 have to put aspects of my personal life 1 2 3 4 5 N/A
“on hold” because of my work.
13.1 am able to accomplish what I would like 1 2 3 4 5 N/A
in both my personal and work lives.
15.1 often neglect my personal needs because 1 2 3 4 5 N/A
of the demands of my work.
18. My personal life suffers because of my 1 2 3 4 5 N/A
work.
19.1 have to miss out on important personal 1 2 3 4 5 N/A
activities due to the amount of time I
spend doing work.
21.1 feel that I allocate appropriate amounts 1 2 3 4 5 N/A
of time to both work and non-work
activities.
 189

22.1 make personal sacrifices to get work 1 2 3 4 5 N/A

done.
23.1 come home from work too tired to do 1 2 3 4 5 N/A
things I would like to do.
24. My job makes it difficult to maintain the 1 2 3 4 5 N/A
kind of personal life I would like.

Personal Life Interference with Work (PLIW)

2. My personal life drains me of the energy 1 2 3 4 5 N/A

I need to do my job.
4. My work suffers because of everything 1 2 3 4 5 N/A
going on in my personal life.
5. Iam too tired to be effective at work 1 2 3 4 5 N/A
because of things I have going on in my
personal life.
16. When I am at work, I worry about things 1 2 3 4 5 N/A
I need to do outside of work.
17.1 have difficulty getting my work done 1 2 3 4 5 N/A
because I am preoccupied with personal
matters at work.

Work/Personal Life Enhancement

6. My job gives me energy to pursue 1 2 3 4 5 N/A

activities outside of work that are
important to me.
10. Because of my job, I am in a better mood 1 2 3 4 5 N/A
at home.
14.1 am in a better mood at work because of 1 2 3 4 5 N/A
everything I have going for me in my
personal life.
20. My personal life gives me the energy to 1 2 3 4 5 N/A
do my job.
 190

APPENDIX E

Work/Family Conflict Scales

Please indicate the extent to which you agree or disagree with the following statements:

Work-to-Family Conflict
StrongUy Strongly
Disagi ee Disagree Neutral Agree Agree

The demands of my work interfere with my 1 2 3 4 5 N/A

home and family life.
The amount of time my job takes up makes 1 2 3 4 5 N/A
it difficult to fulfill my family
responsibilities.
Things I want to do at home do not get 1 2 3 4 5 N/A
done because of the demands my job puts
on me.
My job produces strain that makes it 1 2 3 4 5 N/A
difficult to fulfill family responsibilities.

Due to work-related duties, I have to make 1 2 3 4 5 N/A

changes to my plans for family activities.

Family-to-Work Conflict
My home life interferes with my 1 2 3 4 5 N/A
responsibilities at work, such as getting to
work on time, accomplishing daily tasks,
and working overtime.

The demands of my family or 1 2 3 4 5 N/A

spouse/partner interfere with work-related
activities.
I have to put off doing things at work 1 2 3 4 5 N/A
because of the demands on my time at
home.
Things I want to do at work don’t get done 1 2 3 4 5 N/A
because of the demands my family or
spouse/partner.

Family-related strain interferes with my 1 2 3 4 5 N/A

ability to perform job-related duties.
 191

APPENDIX F

The Stress in General Scale

Think about your present iob. How well does each of the following, words or phrases
describe it? Please check only ONE answer for each:

Y (for “Yes”)
N (for “No”)
? (for “Cannot decide)

Yes No ?
Demanding Y N ?
Pressured Y N ?
Hectic Y N ?
Calm Y N ?
Relaxed Y N ?
Many things stressful Y N ?
Pushed Y N ?
Irritating Y N ?
Under control Y N ?
Nerve-wracking Y N ?
Hassled Y N ?
Comfortable Y N ?
More stressful than I'd li Y N ?
Smooth running Y N ?
Overwhelming Y N ?
 192

APPENDIX G

The Job in General Scale

Think about your present job. How well does each of the following words or phrases
describe it? Please check only ONE answer for each:

Y (for “Yes”)
N (for “No”)
? (for “Cannot decide)

Yes No ?
Pleasant Y N ?
Bad Y N ?
Ideal Y N ?
Waste of time Y N ?
Good Y N ?
Undesirable Y N ?
Worthwhile Y N ?
Worse than most Y N ?
Acceptable Y N ?
Superior Y N ?
Better than most Y N ?
Disagreeable Y N ?
Makes me content Y N ?
Inadequate Y N ?
Excellent Y N ?
Rotten Y N ?
Enjoyable Y N ?
Poor Y N ?
 193

APPENDIX H

Life Satisfaction Scale

Strongly Strongly
Disagree Disagree Neutral Agree Agree

In most ways my life is close to ideal. 1 2 3 4 5 N/A

The conditions of my life are excellent. 1 2 3 4 5 N/A
If I could live my life over, I would 1 2 3 4 5 N/A
change almost nothing.
I am satisfied with my life. 1 2 3 4 5 N/A
So far I have gotten the important things I 1 2 3 4 5 N/A
want in life.
 194

APPENDIX I

Organizational Commitment Scale

Please indicate the extent to which you agree or disagree with the following statements
about the ORGAN1ZATION/COMPANYfor which you work:

Strongly Strongly
Disagree Disagree Neutral Agree Agree

1 2 3 4 5 N/A
I really care about the fate of my
organization.
1 2 3 4 5 N/A
I am proud to tell others what
organization I work for.
1 2 3 4 5 N/A
For me, my organization is the best of all
possible organizations for which to work.
1 2 3 4 5 N/A
If I were offered a job with similar pay
and benefits with another organization, I
would take it.
1 2 3 4 5 N/A
I am willing to put in a great deal of effort
beyond that normally expected in order to
help my organization be successful.
1 2 3 4 5 N/A
me in the way of job performance.
 195

APPENDIX J

Demographic Questions

1. On average, how many hours per week do you work on your job IN TOTAL
(including at work and at home)? ______hours per week

2. On average, how many hours per week do you work on your job AT HOME?
_____ hours per week

3. Please indicate what best describes your current marital/relationship status:

_____ Married, living together
_____Married, living separately
_____Not married, but living with a romantic partner
_____Not married, not living with a romantic partner, but currently involved in a
committed relationship
_____Not married, and not currently involved in a committed relationship

4. Please indicate the number of children who live with you in your household at least
50% of the time: _____

If one or more, how old are they?_______________________________________

5. Do you currently have primary responsibility for caring for one or more of your
parents/in-laws/other relatives? _____Yes _____ No

6. Your gender: ______ Male ______ Female

7. Your age:
_____ 18 to 24 40 to 44 60 to 64
_____ 25 to 29 45 to 49 65 or older
_____ 30 to 34 50 to54
_____ 35 to 39 55 to59

8. Your highest level of completed education:

_____ High school graduate or less
_____ Some college
_____ Junior college
_____ Senior (four-year) college degree
_____ Some post-graduate study
____ Masters degree
_____ Some post-masters study
_____ Doctoral degree or professional degree
 196

9. How long have you been with your current organization?

____ Less than 1 year _____5 to 9 years
_____1 to 2 years ____ 10 to 14 years
_____ 3 to 4 years _____15 years or more

10. How long have you held your current position?

_____ Less than 1 year _____ 5 to 9 years
_____ 1 to 2 years _____10 to 14 years
_____ 3 to 4 years _____ 15 years or more

11. Your current salary:

___ Under $35K _____$75K to $99K
_____ $35K to $49K _____$ 100K to $ 149K
_____$50K to $74K _____$ 150K or more

12. Your primary' area of responsibility (please check only one):

_____Communications
_____ Finance
_____General and Administrative Services
_____General Management
_____Human Resources
_____Information Systems & Technology
_____Insurance & Risk Management
_____ International
_____ Manufacturing
_____ Marketing
_____ Purchasing
_____ Research & Development
_____Sales
_____ Transportation & Distribution

13. Which of these descriptions best fits your current position?

_____Strategic, policy-making
_____Selling and/or marketing
_____Product/service development or delivery'
_____ Administrative or support functions

14. Please indicate the management level of our current position, in terms of its reporting
relationship to the Chief Executive Officer (CEO) of your organization. If you serve
as CEO, mark Level 1; if you report directly to the CEO, mark Level 2; if you are
further removed from the CEO, please mark the number indicating your management
level, where the CEO equals Level 1.

_____ Level 1 (CEO) _____Level 4 _____ Level 7

_____ Level 2 (Report to CEO) _____ Level 5 _____ Level 8 or lower
Level 3 Level 6