O I'l.g «Am; ‘ a "row“ Wfifig {Nun-D1113. 3.2.5 :. .r k. .1... (.5. 1|.) . . #3:: b... Jun tun...) .5 t 3., .ii. , . at... .. ..1' I 1' ‘ tit]. )ufl menu? , III... 7 .u. . L3 .. n9), ‘1» .f 12.5... .5 in 0.. I: .1. "MK... z}. 3.13 “I?“ l. 4...", 3.31.3. 1:30 ~51. .l. I... .8” ii . . ‘3 . i: a“: Atria .x. I? . n: avii évw.u..u.5!. A“..uwd..v.....3 31.393 . k A)! 2": So 3 J. :3 Emuiah , ! 31.1.1..- .2. 3.21:1... . I... 121.51 In». - It. 1...: .. .. .0... an! . incur: a) ukfln... , : 33-.....332, v -' r915 . . . E 1.1. Ihnfinqh. .. "xvi“... .- .3 t (.9: ashrsflflvfltmv II} II. D .uJunN. 3 s .K t...‘ finer...) in”... I -|.l. 5 5.3.9.. .. new"... IT {mu}? .. fin)! .H...r... «fiannvanFB 4.2.2... l: 1.3: .I..lv:v.l.I.-I.IJVJ1.I£ V. .:.-..zu...n.vmnwui:.uvv. v: z. .iiuhfiash . Ln ,XR...Iuv.u2ua&«Venl . do. .A finish, till. .. 3.1.3.5"? 1!. )1! I .p 1. .13....317. In: 2.2.1.... b ‘ ‘i i Kit! I "um-“9.. it . . L, A..." . THESIS IIIIMIIIHIIIUIJHMUIIN!lllelHHllllJlIHUIIHI 1293 01588 0580 LIBRARY Michigan State University This is to certify that the dissertation entitled THE ROLE OF STRESS IN EMPLOYEE PREFERENCES FOR FAMILY-FRIENDLY BENEFITS: TESTING AN INTEGRATED MODEL presented by Beverly Jeanne De Marr has been accepted towards fulfillment of the requirements for PhD degree in Industrial Relations WW KM Major professor Date September 25, 1996 MSU is an Affirmatiw Action/Equal Opportunity Insulation 042771 PLACE N RETURN BOX to romovo thlo chockout from your rocord. TO AVOID FINES rotum on or bolero doto duo. DATE DUE DATE DUE GATE DUE THE ROLE OF STRESS IN EMPLOYEE PREFERENCES FOR FAMILY-FRIENDLY BENEFITS: TESTING AN INTEGRATED MODEL BY Beverly Jeanne DeMarr A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY School of Labor and Industrial Relations 1996 ABSTRACT THE ROLE OF STRESS IN EMPLOYEE PREFERENCES FOR FAMILY-FRIENDLY BENEFITS: TESTING AN INTEGRATED MODEL BY Beverly Jeanne DeMarr Research in the area of work—family conflict has increased steadily as the number of women working outside of the home continues to rise and men begin to take on more family and dependent care responsibilities. At the same time increasing numbers of employers are offering various programs to help employees cope with. both child and elder care responsibilities, yet little is known about what kinds of programs are actually valued by employees. While such programs are implicitly assumed to ameliorate stress between work and family, both the dependent care and work-family conflict literatures remain separate and have yet to be integrated. Especially in the area of dependent care, there is a general ladk of empirical research, and studies typically focus on employees with either child or elder care responsibilities, not both. In.an effort to begin to bridge the gap between these two literatures this study investigated the relationship between employees' work and non-work characteristics, and attitudes toward employer-sponsored family-friendly benefits and perceptions of work productivity. The general research question that this study seeks to answer is whether the relationship is direct, or fully- or partially-mediated by the amount of stress experienced. To answer this question no- mediation, full-mediation, and partial-mediation models were developed based on a review of the literature. The models were tested using a Lisrel VII path analysis on responses from 5273 employees of a large financial services organization. The results show that for all of the dependent variables, the partial-mediation model provides the best fit providing support for spillover theory, as well as Greenhaus and Beutell's (1985) contention that there is a positive reciprocal relationship between family stress and work stress. The results highlight the importance of measuring the direct effects of work and.non-work characteristics, as well as their indirect effects through stress. This study also provides insight for practitioners in the development of programs to help employees balance their work and personal lives. Copyright by Beverly Jeanne DeMarr 1996 To my children, Melita May and Eric Steven Cioe. ACKNOWLEDGMENTS I would like to thank the following people whose support made this dissertation, and thus the degree, possible. Dr. Ellen Ernst Kossek, who in addition to serving as the chair of my dissertation committee, has supported me throughout my doctoral education and provided many opportunities to engage in research. The members of my dissertation committee: Drs. Richard N. Block, J. Kevin Ford, Daniel R. Ilgen, and Michael L. Moore for their helpful suggestions and support throughout this Process. The Families and Work Institute for providing me with access to the data used in this dissertation, and especially Dabbie Schwartz and Cali Williams for their efforts on my behalf. Dr. Randy Fotiu of the MSU Computer Center for reviewing my Lisrel runs to ensure they were correct. Dr. Sandy Gleason for telling me on my very first visit to MSU that completing a PhD is often a matter of persistence. Truer words have never been spoken. My children, Melita and Eric Cioe, for understanding when I had to work on the dissertation and allowing me work with vi very few interruptions. My parents, Hazel and Arthur DeMarr, for their continued support in everything I do. My dearest friends, Julie and John Bamfield for their encouragement throughout this process and generous doses of comic relief. Finally, I would like to thank my very good friend Dr. Marie McKendall, who having been through this process herself, taught me that writing a dissertation is a depressing process and that to actually finish one has to learn to "work depressed" . vii TABLE OF CONTENTS LIST OF TABLES xi LIST OF FIGURES xii CHAPTER ONE: INTRODUCTION AND OVERVIEW 1 Importance of the Topic 1 Research Strategy 5 Key Assumptions and Limitations 7 Contributions 9 Outline of the Dissertation 10 CHAPTER TWO: LITERATURE REVIEW 13 The Relationship Between Work and Non-Work 13 Work-Family Literature and Theories 16 Forms of Work-Family Conflict 17 Processes Linking Work and Family 21 Gender Differences in Work-Family Conflict 26 Sources of Work-Family Conflict 31 Dependent Care Literature and Theories 33 Dependent Care Problems and Programs 34 Dependent Care Preferences and Needs 40 CHAPTER THREE: MODEL AND HYPOTHESES 44 Models of the Role of Stress in Employee Attitudes 44 Family Characteristics 48 viii Child Care Characteristics 63 Elder Care Characteristics 68 Work Characteristics 79 Supervisor Support 83 Organizational Work-Family Culture 87 Dependent Care Stress, Work Stress, and Psychological Distress 92 Attitude Toward Benefits 96 Perceived Work Productivity 99 CHAPTER FOUR: METHOD 106 The Organization Under Study 106 The Sample 106 Data Collection Procedure 109 Operationalization of Variables 110 Data Analysis Strategy 1I7 CHAPTER FIVE: RESULTS 120 Overall Pit of the Model 120 Results of the Causal Model's Hypotheses 129 Family Characteristics 130 Child Care Characteristics 136 Elder Care Characteristics 137 Work Characteristics 139 Supervisor Support 140 Organizational Work-Family Culture 141 Dependent Care Stress, Work Stress, and Psychological Distress 141 Attitude Toward Benefits 142 ix Perceived Work Productivity CHAPTER SIX: DISCUSSION Support for Spillover Theory The Importance of Direct and Indirect Effects Limitations of the Study Implications for Employers Implications for Theory and Suggestions for Future Research LIST OF REFERENCES APPENDIX A: SCALE ITEMS 143 144 144 148 157 160 162 167 179 LIST OF TABLES Table 1 - Means, standard deviations, and intercorrelations of all variables Table 2. Results of Factor Analysis of Work Stress, Psychological Distress, and Dependent Care Stress Scale Items Table 3 - Goodness of fit measures for all models Table 4. Direct Versus Indirect Effects of all Independent Variables xi 121 125 127 149 LIST OF FIGURES Figure 1a - No-Mediation Segmentation Model Figure 1b - Full-Mediation Spillover Model Figure 1c - Partial-Mediation Spillover Model Figure 2. Partial-Mediation Spillover Model with Child Care Benefits as the dependent variable Figure 3. Partial-Mediation Spillover Model with Elder Care Benefits as the dependent variable Figure 4. Partial-Mediation Spillover Model with Work Environment Flexibility Training as the dependent variable Figure 5. Partial-Mediation Spillover Model with Perceived Work Productivity as the dependent variable xii 45 46 47 131 132 133 134 CHAPTER ONE INTRODUCTION AND OVERVIEW This chapter provides an introduction.and overview of the entire dissertation. The introduction and overview include: a) why it is important to investigate work-family stress and dependent care issues, b) the research strategy, c) the key assumptions and limitations of this research, d) contributions of this research, and e) an outline of the subsequent chapters in the dissertation. WW Changing workforce demographics, in particular the dramatic rise in the number of working mothers and the reduction in the number of traditional two-parent families (U.S. Bureau of the Census, 1994), have resulted in a great deal of interest in work-life and dependent care issues over the past decade. The media frequently reports on the child care crisis in this country, as well as issues that affect the elderly. Public policy makers debate the breakdown of the family, and with the passage of the Family Medical Leave Act in 1993 the United States finally has legislation requiring employers to allow employees to take unpaid parental leaves. Businesses are getting more involved in employees’ personal lives (Friedman, 1990; Zedeck & Mosier, 1990) and many offer a variety of dependent care assistance programs. While employers today are more likely to offer such programs, there is still a great deal yet to be learned about the dynamics 2 involved in the interaction and balancing of work and family roles (Zedeck & Mosier, 1990), and what types of benefits are most desirable. The early research.in the relationship of‘work.and family was based on a segmentation.approach and focused primarily on the incompatibility of work and family responsibilities and gender differences (cf. Burke, Weir, & Duwors, 1979, 1980a, 1980b,- Jones & Butler, 1980,- Locksley, 1980). The basic assumptions were that for men unemployment had detrimental effects on the family; while foeromen.employment was presumed to have a negative impact on the family, and especially the children (Voydanoff, 1988a). These assumptions reflected societal beliefs and stereotypes of the 19503 and 1960s. The stereotypical family included two parents. The father worked during the day in an office, and where family was not to intrude, and the mother spent her days at home cooking, cleaning, and tending to family matters. Thus, work and family were segmented with men and women filling distinct, separate roles that were determined by gender. These assumptions were implicitly, if not explicitly, carried forward into both research and practice, where to some extent, they are still alive and well today. This is evidenced by the preponderance of work-family and dependent care research done on populations that are predominately female (cf. Aryee, 1992,- Kossek & Nichol, 1992; Rosin & Korabik, 1990; Williams, Suls, Alliger, Learner, & Wan, 1991) 3 Over the last few decades our society has seen considerable change. People are marrying later, having fewer children, are more likely to divorce, and are living longer and with more serious medical conditions. In 1970, 71.7% of people in the United States were married, while by 1993 this number had dropped to 61.2%, with only 55% of households headed by a married person whose spouse was present (U.S. Bureau of the Census, 1994). The trend is the same even when there are minor children in the household. In 1970, 87% of family groups with children under age 18 were two-parent families, while by 1993 that number had dropped to 70% (U.S. Bureau of the Census, 1994). Not only has there been.a decline in the number of two-parent families, but among two-parent families more women are working outside of the home to help support the family. Between 1975 and 1993 the labor force participation rate for wives with husbands present and children under age 18 increased from 44.9% to 67.5% (U.S. Bureau of the Census, 1994). The rates of change are similar regardless of the age of the children. In 1992, 54% of women who had a child under one year of age were in the labor force compared to only 38% in 1980 (U.S. Bureau of the Census, 1994). The net effects of these changes are that more children today are in some type of daycare arrangement, and more parents are having to cope with stresses between work and family. At the same time with advances in medical technology, 4 life expectancies have increased. A child born in 1970 has a life expectancy of 70.8 years, while a child born in 1990 has a life expectancy of 75.4 years, a trend which is projected to continue well into the let century (U.S. Bureau of the Census, 1994). The trend has had an impact on those providing assistance for elder dependents, which historically have been women.who were not employed outside of the home. The fact that more women are working outside of the home, combined with increased life expectancies, means that there are fewer traditional caregivers for more elderly persons. While many believe the myth that children today do not care for their elders as they once did, the fact is they do, regardless of other work and family responsibilities (Brody, 1985). Thus, elder care responsibilities serve as another contributor to work-family stress, and are yet another facet of dependent care. Although work-family stress and dependent care are clearly intertwined, research in these areas is not. The dependent care research has tended to be directed more toward human. resource jpractitioners and. lacks solid. theoretical models. This literature typically focuses on the types of dependent care assistance programs offered by organizations or utilization rates of specific programs. While theoretical models explaining certain aspects of how an individual experiences work-family stress have been developed, a great deal remains to be done. The models typically do not address S a variety of predictors and outcomes of work-family stress, and most of the studies are conducted.with.homogeneous, white, middle class subjects. Strickland (1992) asserts that the scope of work-family research needs to be broadened relative to udnority, single-parent, working-class, chronically unemployed, and truly upper class families, and maintains that greater depth is also .needed. Because these streams of research have evolved in large part independently of one another, there is no integrated model and nothing that links the two empirically. B§§§§I£h_§fi£§£§31 A fundamental goal of this dissertation is to begin to integrate the more theoretical work-family literature and the more practitioner-oriented dependent care literature by developing and testing three versions of an integrated model that considers the effects of individual, work and organizational characteristics (n1 employees' jperceived productivity and attitudes toward benefits. The work-family literature (of. Greenhaus & Beutell, 1985,- Lambert, 1991,- Lobel, 1991) has tended to take a more psychological approach focusing on the individual level of analysis. Most of the models that have been developed and tested in this area focus on the causes of work-family stress, which is often referred to as work-family conflict, and often either stop there, or link conflict with some form of psychological distress or measures of job or marital satisfaction. The implicit 6 assumption is that work-family conflict is bad, but in large part only affects the individual. Further, 'most of the research ignores the influence of the work group and the organizational culture as it relates to work and family. The dependent care literature tends to be more Human Resource focused, as today more employers are offering some type of dependent care assistance programs. In general, the programs start as initiatives to help employees cope with child care responsibilities and over time are expanded to include eldercare. Many of these programs have been adopted without a formal needs assessment (Kossek, 1990). Even when employers have done a needs assessment they often stop after implementation and do not evaluate the effectiveness of the program to see if it truly serves the needs of their employees. Not surprisingly, the utilization rates of these programs tend to be quite low. There have Ibeen. a few studies that have looked. at absenteeism and some that have looked at utilization rates, however, little or no research has looked at whether employees see various benefits as attractive (an exception is Kossek, 1990). Similarly, studies have generally focused on either child or elder care responsibilities, not both. Most studies in this area are based on relatively small, homogeneous samples. They typically do not use rigorous data analysis techniques, but instead report simple descriptive statistics (e.g., what percent of employees prefer which option, or use 7 a particular program). With the exception of Goff, Mount, and Jamison (1990), and Kossek and her colleagues (Kossek, 1990; Kossek, DeMarr, Backman, & Kollar, 1993; Kossek 8: Nichol, 1992) most of this research is not based on theoretical models. It is believed that this dissertation will fill some of these gaps by developing and testing three versions of a model that considers the impact of various work and non—work characteristics on perceptions of productivity and attitudes toward famdly-friendly benefits. The independent variables include various family, child care, and elder care characteristics, the amount of control one has over his/her schedule, perceptions of career penalties, supervisor support, and organizational work-family culture. The model will be tested using a Lisrel VII path analysis to determine whether the independent variables have a direct effect, or are fully or partially mediated by work and family stress. The sample includes a large, heterogeneous group of respondents. K A ' L' ' i This dissertation is based on a resource allocation view that in today’s complex world individuals face demands from.a variety of sources, the major of which are work and family. For success, both typically require time, energy, and emotional commitment, all of which are fixed resources. As fixed resources the individual must allocate them between the two domains. When there are not enough resources to satisfy 8 both work and family roles the individual experiences role conflict. It is assumed that ultimately the type and level of stress that one experiences is influenced by a complex set of individual, dependent care, ‘work, and. organizational characteristics. In other words, there is no single predictor or level of predictors of work and family stress. Consistent with the work of Kopelman, Greenhaus, and Connolly (1983), Greenhaus and Beutell (1985), Gutek, Searle and Klepa (1991). Frone, Russell, and Cooper (1992), and Edwards and Rothbard (1995), it is also believed that there is a reciprocal relationship between work stress and family stress and together they result in psychological distress. Finally, it is assumed that the amount of psychological distress a person experiences affects his/her perceptions of productivity in the workplace and attitudes toward employer assistance in coping with his/her dependent care responsibilities. For example, a person experiencing a high level of psychological distress may be more prone to illnesses that would result in absence, or s/he may be distracted while at work resulting in lower work productivity. Further a person who is having difficulty coping with his/her dependent care responsibilities may be more likely to place a higher value on employer-sponsored.dependent care assistancejprograme than an individual who is already c0ping effectively with his/her responsibilities. 9 Work and family stress and dependent care are relatively new areas of study, and as such this work should be considered exploratory, especially with respect to the extent that the model deals with employee views toward family-friendly benefits. In that area existing theory on which to base a model is very scarce. Further, unlike some other areas of study, there are no well established standand measures and scales. While this study is quite broad in scope, there are also limitations as to how much can be accomplished in any one study. As a practical matter, this study does not consider all sources of demands on an individual (e.g., clubs, social organizations, friends, non-dependent relatives). Similarly, there may be other family or work-related characteristics not captured.by this study (e.g., the mental or physical health of the individual). 0 ri i While as with any study there are limitations, this study is also expected to make a number of both theoretical and practical contributions. First, the model will be used to test a variety of variables that influence family and work stress and psychological distress which.has rarely'been.done. This is important to assess the relative impact of each variable and begin to sort out which are the most important predictors. While all of the variables may not prove to be significant, assuming that other’plausible alternative explanations can.be 10 ruled out, this too*would.provide guidance for future research and theory development. This study also provides a link between the work-family and dependent care literatures which has yet to be done. By using a large heterogeneous dataset, this study also overcomes some of the limitations in generalizability in much of the prior research. The study makes a number of practical contributions for both employers and employees as well. Studying the influences on employee attitudes toward family-friendly benefits will provide insight for employers who wish to develop proactive programs to help employees balance their work and personal lives. To the extent that attitudes toward family-friendly benefits can be predicted, the study will help employers get the largest return on their investment in dependent care programs. To the extent that employers utilize the information provided by this study to develop policies and programs to help employees cope with dependent care responsibilities, employees will be better able to balance work and family responsibilities. Outline of the Dissertatign This dissertation consists of six chapters. Chapter one, thus far, has provided a broad overview of the goals, scope, and potential contributions of this dissertation. Chapter two includes a review of the existing work-family and dependent care literatures, focusing on current 11 alternative theories. The purpose of this review is to provide a general background and highlight the need for more theoretical work to integrate the two areas of study. Chapter three presents the model and related hypotheses on which this study is based. The model is based on a combination of the work by Kopelman, et a1 (1983), Higgins, Duxbury, and Irving (1992) , and Greenhaus and Parasuraman (1986), which has been expanded to include a variety of specific antecedents. The model considers the relationship between a number of work, family, and dependent care characteristics, stress, and employee views toward family- friendly benefits and perceived work productivity. It is offered as an early step in the development of a more theoretical body of literature on the dependent care issues faced by employed parents and those providing care to elder dependents. Chapter four presents the methods of investigation. It includes the organization under study, the subjects of the study, the data collection procedure, the Operationalization of the variables, and the data analysis strategy. Chapter five contains the results of the data analysis for the model and the hypotheses. Descriptive statistics are also provided. Chapter six includes a discussion of the results of the analyses focusing on the study’s support for spillover theory and the importance of measuring both the direct and indirect 12 effects of individual variables. Implications for theory and practice, as well as limitations of the study and suggestions for future research are also presented. 12 effects of individual variables. Implications for theory and practice, as well as limitations of the study and suggestions for future research are also presented. Chapter 2 Review of the Literature This chapter will review the literature and theories related to work-family conflict and dependent care after providing a more general overview of the relationship between the work and non-work domains. T ' ' w W rk N -W r Research on the relationship between the work and non- work aspects of people’s lives has evolved over time from the role conflict literature. While various authors have used differing terminology to describe what is essentially the same construct (Higgins, et a1, 1992), the most commonly used term thus far is "work-family conflict" (for a summary of studies that use different terms interchangeably see Greenhaus and Beutell, 1985, p. 79). Formally, Greenhaus and Beutell (1985, p 77) define work-family conflict as "a form of interrole conflict in which the role pressures from the work and family domains are mmtually incompatible in some respect" whereby participation in one role is made more difficult by virtue of participation in the other. The earliest view of the relationship between work and home was that they are segmented and independent (Lambert, 1990). Indeed, "work" and "family“ have typically'been studied separately by scholars in the various branches of psychology and sociology. as well as economics, organizational behavior, 13 14 and human relations. In large part the early research was lacking in theoretical frameworks for understanding the phenomenon. More recently the focus has been broadened from "work-family" to "work-life" to include other aspects of individuals’ personal lives which recognizes the diversity of demands on individuals today. While 'much. of the existing research 'uses the term "conflict", there are often.at least implicit links to stress. Several researchers have drawn from stress research (cf. Frone, Russell, & Cooper, 1992; Greenhaus & Parasuraman, 1986; Higgins, et al, 1992; Kopelman, et a1, 1983) bringing theoretical rigor to the study of work-family issues. Greenhaus (1989) suggested that viewing work-family research from a stress perspective is useful because it allows researchers to draw from an established paradigm. More explicit links to stress are made by Edwards and Rothbard (1995) in their recently proposed cybernetic model of stress, coping, and well-being within and between work and family domains. Thus, it appears there is a shift to more integrated approaches for studying the relationship between work and nonwork domains. While researchers have more recently begun to develop theories explaining' ‘work-life stress and study the effectiveness of various dependent care programs, significant gaps in the literature still remain. Lambert (1990) argues that a fuller understanding of the processes linking work and 15 family life is necessary to adequately evaluate the effectiveness of the family supportive policies currently being implemented by many U. S. firms, as well as to identify additional strategies for helping workers find satisfaction in both their work and personal roles. The existing literature typically falls into one of two general categories. The first type of research, which is more theoretical in nature, focuses on explaining how work-family conflict operates from a psychological standpoint. This type of research may study levels of work-family conflict, often with respect to gender, butgenerally does not consider what can be done to ameliorate it. The second type of research, which is more practitioner oriented, tends to focus on specific dependent care programs, usually onsite day care facilities or parental leave programs, offered by employers. Most of this type of research looks at either child or elder care, not both. It is also usually limited to a fairly narrow range of occupational levels, and hence doesn’t reflect the diversity in income levels, work schedules, and household configurations that is often found in large organizations. While this type of research offers some specific guidance for employers, it lacks a more general framework for viewing the broader picture. Both areas of research have made significant contributions to the body of knowledge surrounding the integration of work and nonwork, however, this is still a 16 relatively new area of study and there is much yet to be learned. To date there have not been any major empirical studies that link the two domains. In addition, there is a lack of research that includes the effects of having both elder and child care responsibilities. This is a serious void in the research given the changing demographics noted earlier and the growing trend for employers to offer benefit programs to help employees manage their elder care responsibilities. Typically such programs are modeled after programs to help employees manage their child care responsibilities, yet without research the efficacy of these programs remains unknown. In the following sections I will review the existing literature and theories in the areas of both work-family conflict and dependent care, beginning with the existing research on work-family conflict. Werk—Femily Liteteture end Theeries There are a number of frameworks and theories that have been used over the last decade to help further the understanding of various aspects of work-family conflict. Each makes a contribution to the body of knowledge on the relationship between work and non-work domains. Some have considered the forms of conflict (cf. Greenhaus & Beutell, 1985; Gutek, et al, 1991), while others have focused on the processes that link work and family (cf. Lambert, 1990). Yet other research has sought explanations for differences between men and women in work-family conflict (cf. Lambert, 1991; 17 Lobel, 1991). Finally, researchers have also considered the sources of work-family conflict, specifically work interference with family and family interference with work (cf. Frone, et al, 1992, Greenhaus & Beutell, 1985, Gutek, et al, 1991). Each of these frameworks and theories will be reviewed in the following pages. WW One of the earliest models of work-family conflict was developed by Greenhaus and Beutell (1985) who maintained that work-family conflict has three major forms: time-based conflict, strain-based conflict, and.behavior-based.conflict. Time-based.conflict is based on the premise that time spent on activities within one role cannot be devoted to activities within another role. This ‘may result from the physical constraint that a person can not be in two places at one time, or the psychological constraint that if a person is preoccupied with one role they may not be able to adequately fill the needs of a second role even though they are physically present . Pressures resulting in time-based conflict may arise from the work domain (e.g., long hours, inflexible work schedules, and shiftwork) or the family domain (e.g., young children, spousal employment and large families). Strain-based conflict exists when strain in one role affects one’s performance in another role (Greenhaus & Beutell, 1985). The roles are incompatible in that the strain produced by one role makes it difficult to comply with the 18 demands of another role. Pressures in the work domain that may result in strain-based conflict include role conflict, role ambiguity, and boundary-spanning activities. From the family domain, family conflict and low levels of spousal support are believed to contribute to strain-based conflict. Behavior- based conflict results when specific patterns of role behaviors are incompatible with expectations regarding behavior in another role (Greenhaus & Beutell, 1985). For example, the work domain may hold expectations for a manager to be objective and maintain confidentiality, while the family domain may hold expectations for openness and warmth. Greenhaus and Beutell's (1985) ‘model proposes that conflict between two roles can be produced by any role characteristic that affects a person's time involvement, strain, or behavior within a role. The model also posits that work-family conflict is intensified via role pressures when work and family roles are salient to a person's self-concept. As individuals become more motivated and ego-involved in a role they are more likely to spend more of their time and energy on that role at the expense of the other. Finally, the model holds that work-family conflict will be strongest when there are strong negative sanctions for noncompliance with role demands. The rationale is that the absence of strong negative sanctions for noncompliance reduces the pressure to comply with role demands. A takeoff on the time component of Greenhaus and 19 Beutell's (1985) framework is what Gutek, et al (1991) describe as the rational view. The rational view holds that the amount of conflict that an individual experiences rises in proportion to the amount of time spent in both work and family domains (Keith & Schafer, 1984; Staines, Pleck, Shepard, & O'Connor, 1978) . Specifically, the rational view predicts that as adults increase the amount of time spent in paid employment relative to family work they will experience more work interference with family (WIF) than family interference with work (FIW; Gutek, et a1, 1991) . Thus, the amount of WIF rises with the number of hours spent in work activities, while the amount of FIW increases with the number of hours spent in family activities (Gutek, et a1, 1991) . Since women on average spend more time in family work than men, and men spend more time in paid employment than women (Pleck, 1985), the rational view predicts that women will experience more FIW than men, while men will experience more WIF than women (Gutek, et a1, 1991). Although the rational view is quite simplistic, it does seem quite logical that the less free time one has available, the more conflict will be experienced. The rational view does not, however, fully account for the gender differences that are often found in this type of research, nor does it account for any individual differences in the type of work being performed or a person's perceived ability to cope with stress. Gender differences are specifically addressed by the gender role 20 framework. The gender role framework for understanding work-family conflict posits that gender both directly influences perceived work family conflict, and moderates the relationship between time spent in paid and family work, and perceived work-family conflict (Gutek, et a1, 1991). Central to this framework are the traditional gender roles, with men being primarily focused on work and women focused primarily on family matters. The gender role framework holds that gender role expectations may distort the rational view such that the level of conflict men and women report will depart from the rational view in a manner consistent with gender role expectations (Gutek, et al, 1991). According to this view additional hours spent in one’s own sex role domain (e.g., more housework for women or more time spent on paid employment for men) are perceived to be less of an imposition and create less conflict for the role holder than additional time spent in the other sex role domain. Specifically, the gender role framework.predicts that men will be more sensitive to the amount of time spent in family work, while women will be more sensitive to the amount of time spent in paid employment. Further, the gender role framework predicts that with hours of paid employment held constant, women will report more WIF and men more FIW (Gutek, et a1, 1991), which is exactly the opposite of the rational view's prediction. In a study of psychologists and managers, Gutek, et al, 21 (1991) found some support for both the rational and gender role explanations for work-family conflict. The results, which provided support for the rational perspective, showed that there was a fairly high correspondence between hours spent in a particular domain and conflict originating in that domain, while hours spent in one domain were not associated with conflict originating from the other domain. The gender role perspective received some support as well in that people did seem.to interpret their perceptions of conflict in accordance with traditional gender' role expectations. There *was no support for the gender role prediction that women are oblivious to the amount of family demands placed on them.and men are equally oblivious to work demands. Gutek and her colleagues (1991) concluded that researchers need to look at an individual’s perceptions as well as the actual amount of time one spends in work and family activities. Pr L' i W k n F mi . Another group of competing frameworks that are sometimes used to explain the processes that link work and family life are segmentation, compensation, spillover, and accommodation (Lambert, 1990). The earliest of these frameworks is segmentation, which. holds that work and. home lives are independent and do not affect one another. Thus, an individual's subjective reactions, whether they are positive or negative, to objective conditions in one domain do not affect outcomes in the other domain. This view is consistent 22 with the traditional two-parent family structure where the husband is the sole provider and the wife has primary responsibility for home and family matters. The presence of an at-home spouse to tend to non-work.matters allows an employed spouse to focus solely on work. Thus, in the days when the majority' of families *were "traditional", segmentation is likely' to Ihave occurred. naturallyu Today, however; *most families do not consist of two-parents with a stay-at-home mom and segmentation is unlikely to occur naturally (Lambert, 1990) . Today, segmentation is more likely to occur because workers "actively attempt to separate work and family life in order to deal with work-related stresses" (Piotrkowski, 1979, p. 98). The compensation model holds that individuals may compensate for a lack of satisfaction in one domain by trying to find. more satisfaction in the other (Lambert, 1990; Staines, 1980). The result of trying to find more satisfaction in a particular domain often leads to a higher level of involvement. This theory has been used primarily to explain why some individuals who are engaged in unsatisfying and uninvolving work often become more active in nonwork activities such as clubs, groups, or their children’s activities. With respect to working class men it has been said that they "look to their homes as havens", and "look to their families as sources of satisfaction lacking in the occupational sphere" (Piotrkowski, 1979, p. 98) . The 23 compensation model may also be used to explain why some individuals become more involved in their work when they experience family problems, such as divorce. In this case work may become an escape from the problems at home. Consistent with the compensation model is the utilitarian approach to role investment that holds the more often an activity is rewarded, and the more valuable the reward of an activity is to a person, the more likely a person is to engage in the activity (Homans, 1976). A primary difference is that the compensation model focuses on role investment to counterbalance a deficiency in the other role, while the utilitarian view focuses on role investment to maximize net rewards. In our culture, women have traditionally received greater recognition, and hence rewards, for household and family activities, while for men rewards have been primarily associated with work related activities. Thus, the traditional sex-role stereotypes have been reinforced. With the utilitarian view competition between work and family roles is viewed as inevitable, and a particular role gains acceptance only at the expense of the other. Thus, with this view a person’s sense of balance increases only as an individual’s role investments become more unequal (Lobel, 1991) . The corollary is that the highest degree of work-family conflict occurs when pressures to participate in both roles are equal (Greenhaus & Beutell, 1985). The spillover framework (Staines, 1980; Zedeck & Mosier, 24 1990) holds that the effects of each domain, both.positive and negative, carry over to the other. Research has suggested that individuals transfer the attitudes, behaviors, emotions, and skills established at work to the family environment (Belsky, Perry-Jenkins, &. Crouter, 1985; Crouter, 1984; Kelly, & Voydanoff, 1985; Piotrkowski, 1979), and vice versa (Belsky, et a1, 1985; Crouter, 1984). The spillover framework has been used for much of the existing research on work-family issues. This is particularly true for research.on.both elder and child care where the emphasis is placed on the spillover from.family roles and obligations to the caregiver's employment (cf. Kossek, 1990; Kossek, et a1, 1993). Spillover can be classified as either direct or indirect. Direct spillover occurs when objective conditions in one domain have a direct impact on the other regardless of how the person subjectively experiences the conditions (Lambert, 1990). An.example of this would be the parent who must stay at home with his sick child or the employee who misses her child’s ball game because she has to work. Indirect spillover occurs when an individual' 3 subjective reactions to objective conditions in one domain affect the other domain (Lambert, 1990). An example of this would be a working parent or caregiver who is anxious about his or her dependent(s) while at work, or the parent who is preoccupied with work while at his/her child's school play. Certain job and family characteristics may operate directly, indirectly, 25 or both (Lambert, 1990). However, the impact of indirect spillover may be more difficult to ascertain since employees seldom mention family responsibilities in general for fear their employer will not understand (Winfield, 1987), and may not see their employer as a source of support (Creedon, 1988). While the processes of segmentation, compensation, and spillover have frequently been viewed as competing, independent theories, Lambert (1990) argues that these processes may occur simultaneously and proposes that the accommodation model may also help to explain the processes linking work and family. The process of accommodation is characterized by individuals limiting their involvement in one domain so they may better accommodate the demands of the other (Lambert, 1990). An example of this is a woman who opts for part time work to allow her to be available to the children before and/or after school. The.accommodation model holds that a high involvement in one domain leads to low involvement in the other, which is the reverse of the compensation model. This may be in part due to the fact that the accommodation ‘model was conceived.on.the basis of women’s experiences, while compensation was initially based on men's experiences. Because of our traditional gender role stereotypes many employer sponsored work-family programs have been implicitly, if not explicitly, designed to help women accommodate their family responsibilities. While this approach certainly may provide some degree of assistance to employed caregivers, it does not 26 take into consideration men's increasing level of participation in family responsibilities and the diversity of family structures in today's society. As such employers may not be getting the maximum return on their investment in family-friendly benefit programs. W Employers often hope that by offering dependent care assistance programs and other family-friendly benefit programs, employees will experience higher levels of job involvement, intrinsic motivation, and job satisfaction. In an attempt to better understand the relationship between such outcomes and job and family characteristics, Lambert (1991) compared the expectation and value hypotheses . Both hypotheses attempt to explain why women generally appear to be more satisfied than men under similar work situations. The expectation hypothesis postulates that while women and men are equally attracted to certain job features, women have lower expectations of the workplace than men. Mottaz (1986) argues that these differences reflect adjustments to the workplace, in other words, women are more likely to turn to social satisfactions when other workplace satisfactions are lacking. The value hypothesis holds that women and men are attracted by different job features and that satisfaction depends on whether or not one receives what one values. Men and women are believed to value different job features because of differences in gender-role socialization (Lambert, 1991). 27 Traditionally in our culture women have been socialized to value relationships while men have been socialized to place a higher value on achievement. Lambert (1991) in a study that used data from the 1977 Quality of Employment Survey (QES), found that men and women reported similar levels of job satisfaction and job involvement even though women receive fewer intrinsic and extrinsic rewards than ‘men. This seeme to support the expectation hypothesis proposition that because women have lower expectations they are able to obtain levels of satisfaction similar to men. This result may be due to the fact that women's jobs may be less stressful than men’s and that women find coworker and supervisory relationships more supportive than men do (Lambert, 1991). When stressful job conditions and social rewards were controlled there were no significant differences between the levels of job satisfaction of men and women which did not support the expectation hypothesis. She did, however, find that women report greater intrinsic motivation even after controlling for intrinsic and extrinsic job characteristics, which is consistent with the expectation hypothesis. Lambert (1991) found only moderate support for the value hypothesis as well. In short, there were more similarities in the relationships between job characteristics and work responses than differences. The study found.that regardless of gender, skill variety' and. task. significance ‘promote job 28 satisfaction, intrinsic motivation, and job involvement, while stressful work is negatively related to job satisfaction and intrinsic motivation, but positively related to job involvement. The opportunity for promotion was significant in explaining job satisfaction for both men and women, however, the effect was significantly larger for men than women. Support was found for the value hypothesis’s assertion that women value the social aspects of work, and may be more sensitive to social rewards than men. Supportive coworker relationships were positively related to women's job satisfaction and supportive supervision was positively related to women’s intrinsic motivation. For men, job autonomy was positively related to intrinsic motivation. Limited support was found for Pleck's (1977) contention that the boundaries between work and family are asymmetrically permeable for men and women, although there was not support for the contention that women's work responses are more sensitive to family responsibilities than men’s. The study found that characteristics of the spouse’s employment did influence work responses of both men and women although in different ways. The number of hours a spouse spends in paid employment was related to job involvement for both men and women. Wives' income was related to both men's intrinsic motivation and job involvement, while husband's job security was related to women’s job satisfaction. The study found, however, that parenting responsibilities do not help explain 29 why some workers are more or less involved, motivated and satisfied in their jobs than others. This may be due to the fact that this study only considered the presence of parenting responsibilities as opposed to the type of responsibilities which can vary dramatically from family to family. Another explanation for the nonsignificance of parenting responsibilities may be found in Cooke and Rousseau’s (1984) finding that children are a source of both stress and reward. The value of such rewards is a focus of the social identity approach to work and family role investment. The social identity' approach. holds that individuals identify themselves as members of various social groups, and the more salient a particular identity is for a person the more they will invest in that role (Lobel, 1991) . Group identification results from an individual's perceptions, and unlike the utilitarian view'discussed.earlier, does.not strive to maximize net rewards. The results of an individual's identification with a group include: selection of identities congruent with the salient social identity; loyalty to the group despite negative attributes; conformity to group norms and attribution of typical characteristics to oneself; and reinforcement of the group's prestige, values, and practices (Ashforth & Mael, 1989). Individuals typically identify with several groups (e.g., employee, parent, child, alumni, civic organization member, etc.), and the enactment of a particular identity is triggered by situational cues (Lobel, 1991). For 30 example, a phone call from one’s superior would trigger the employee role, while a call from a one's child care provider concerning a sick child would trigger the parent role. The social identity view posits that conflict between work and family roles can.be minimized in one of two ways. The first is by keeping identities either physically, psychologically, or temporally separate (Allen, Wilder, & Atkinson, 1983; Lobel, 1991). This is consistent with the notion of intentional segmentation discussed earlier where an individual actively strives to keep their family and work lives separate to minimize conflict. Unfortunately, complete segmentation may not always be possible. A second way that an individual can minimize conflict according to the social identity view is to maintain consistent personal values across identities (Allen, et al, 1983; Lobel, 1991). For example, an individual that works for an organization that values people, openness, and concern for others, and whose family has those same values, would not be faced with enacting roles based on different values in each setting. Thus, the social identity approach maintains that balance between work and family roles is possible, and increases as the overlap in underlying values increases (Lobel, 1991). Lobel and St. Clair (1992) found support for social identity theory in their study that tested the effects of career identity salience, as well as gender and family responsibilities, on work effort and merit increases. Their 31 study found that the direct effect of career identity salience was both positive and significant. With respect to the effect of career identity salience on merit increases, the findings suggested that merit increases may be allocated based on congruence to gender role stereotypes. More specifically, family-oriented women with preschoolers received higher merit increases than family-oriented men with preschoolers, and career-oriented women with preschoolers received lower merit increases than career—oriented men with preschoolers. This study also tested predictions based on both human capital theory which holds that home and family responsibilities lead to less effort at work, and gender discrimination which asserts that women do not have equal access to opportunities and rewards. Neither of those theories received support. I W - Much of the existing research on work-family conflict does not explicitly make the distinction between work interference with family (WIF) and family interference with work (FIW) noted earlier. Researchers from different disciplines have, however, tended to implicitly focus on one or the other. Sociologists and developmental psychologists generally focus on the effect of work on family life (cf. .Aldous, 1969; Piotrkowski, Rapaport & Rapaport, 1987). On the other hand, organizational behaviorists tend to focus on the effects of family responsibilities on work (cf. Cooke & Rousseau, 1984; Lobel 8: St. Clair, 1992) . Researchers are now 32 beginning to recognize that the conflict relationship between work and family is bidirectional (cf. Frone, et al, 1992; Greenhaus & Beutell, 1985; Gutek, et al, 1991), and are making a distinction between FIW and WIF. Research.has shown that WIF and FIW do in fact operate differently, are clearly separable, that the amount of time spent in paid employment and the amount of time spent in family work are relatively independent of one another, and that the amount of time spent in one domain was not associated with conflict originating from the other domain (Gutek, et al, 1991). Frone, et a1 (1992) expanded on the notion of a difference between WIF and FIW with the development of a model that views the relationship between the two to be a positive reciprocal one. Their argument is that if one’s work (or family) related responsibilities and problems begin to interfere with the accomplishment of one's family (work) obligations, the unfulfilled family (work) obligations may begin to interfere with one's functioning at work (home). Thus, there can.be a snowball effect with respect to the total level of work-family conflict. The results of their study supported.the argument that the relationship between work and family is bidirectional or reciprocal (Frone, et al, 1992). Further, the results replicated the findings of others that job stressors and job involvement were positively related to the frequency of WIF conflict, and conversely, that family stressors and family involvement were positively related to 33 the frequency of FIW conflict. While research has generally provided support for a reciprocal relationship, a study by O’Driscoll, Ilgen, and Hildreth (1992) found that the relationship between time pressures in one domain and interference with activities in the other domain was asymmetrical. While some researchers have stated that FIW has been understudied or even that it has not been studied at all (Frone, et al, 1992) , I would argue that a major component of FIW has received attention in a number of studies under the guise of dependent care. More specifically, the impact of having child or elder care responsibilities on an employed caregiver’s paid employment. It is with that in mind that I shift the focus to the dependent care literature. D n i r Unlike the work-family literature, the dependent care literature is still in the early stages of development and tends to be fairly narrow in focus. Much of this literature is oriented toward practitioners and deals with specific employer-sponsored dependent care programs and their effectiveness in helping employees meet their dependent care obligations. In general, there are few models and little theory in this literature. The few models that have been developed usually focus on demographic characteristics as either direct or indirect predictors of work related outcomes such as absenteeism (cf. Goff, et al, 1990; Kossek, 1990; 34 Kossek, et al, 1993; Kossek & Nichol, 1992). A few studies have sought to evaluate employer-sponsored dependent care programs in terms of perceived fairness (Grover, 1991), work- family conflict (Goff, et al, 1990), performance (Kossek & Nichol, 1992), absenteeism (Goff, et a1, 1990; Kossek & Nichol, 1992), and employees’ assessment of the program's effectiveness (Kossek, et a1, 1993). More often what is published are accounts of what various organizations are doing to help their employees cope with. dependent care responsibilities (cf. Denton, Love, & Slate, 1990; Friedman, 1990; Mattis, 1990; Zedeck & Mosier, 1990), or what employees need in the area of employer-sponsored dependent care assistance programs (cf. Anastas, Gibeau, & Larson, 1990; Kossek, 1990; Miller, Stead, & Pereira, 1991; Scharlach & Boyd, 1989; Sizemore & Jones, 1990). Each of these areas will be reviewed in the following pages. D nd P m Kossek (1990) developed a model that holds that an employed caregiver’s demographic characteristics such as gender, household employment configuration (e.g., single parent, dual-career, or traditional family), use of familial care, and dependent care profile, can be used to predict problems with dependent care. Problems with dependent care, including the quality, cost and availability of care, subsequently predict an employee's attitude toward managing work and dependent care responsibilities (e.g., the level of 35 stress experienced at work as a result of family responsibilities). Attitudes toward managing work and dependent care responsibilities in turn predict absence due to dependent care responsibilities. The more problems with dependent care that an.employee experiences, the less positive their attitudes, and the more likely they will be absent. The general framework has received empirical support in both the context of child care (Kossek, 1990; Kossek & Nichol, 1992) and eldercare (Kossek, et al, 1993). In one study (Kossek, 1990) that tested this model as it relates to child care responsibilities found that demographic characteristics were significant predictors of problems with dependent care. In particular, those who used total nonfamilial care experienced the most problems with care, followed by those who used a combination of familial and non familial care. Those who relied solely on familial care were found to have the fewest problems with care. The study also found that women and those who experienced problems with dependent care were significantly more likely to hold negative attitudes toward managing work and family responsibilities. Further, those who held more negative attitudes (but not necessarily women) were more likely to be absent due to dependent care responsibilities. A subsequent study (Kossek & Nichol, 1992) found that the greater the number of hours that an employee uses off-site nonfamilial care or the lower the degree to which an employee 36 could rely on family members to care for a sick child, the greater the problems with care. As in the pmevious study, having problems with care was associated with negative attitudes toward managing work and child care responsibilities. Unlike the previous study, this study used supervisor's perceptions of child care related absenteeism. Significant relationships were found between gender and the extent to which an employee can rely on family members to care for a sick child and supervisor's perceptions. In short, supervisors perceived that women and those without familial support were absent more due to child care responsibilities. The relationship between an employee’s attitude toward managing work and child care responsibilities and supervisor' s 'perceptions of child.care related absence was not significant, however, the relationship between a supervisor’s perceptions of child care related absence and the employee's performance rating was. In other words, the more a supervisor perceived that an employee was absent due to child care responsibilities, the more likely they were to assess the employee’s performance as lower. One study has also used the same general model in the context of eldercare (Kossek, et al, 1993). In that study the employed caregiver's demographic characteristics included: household employment configuration, form of adult dependent care (e.g., familial, paid companion/nurse, adult care facility, or nursing home), living arrangements (e.g., another 37 town, the elder’s own residence locally, a special care facility, or living with the employee), and gender. Parallel to the child care studies, this study found that the more employees experienced problems with elder care arrangements the less favorable their attitudes toward managing work and eldercare responsibilities, and that absence was significantly related to both problems with eldercare arrangements and less favorable attitudes. Specifically, employees using a paid companion or nurse in the home, or employees whose parents are living with them are significantly more likely to hold negative attitudes toward managing work and eldercare responsibilities. Goff, et al (1990) developed a similar model that held that an employee’s use of an on-site day care center, the number of children under 5 years old, providing care for an ill child, having primary responsibility for child care, satisfaction with care arrangements, and the level of supervisor support directly influenced work-family conflict. Work-family conflict, as well as use of an on-site center, supervisor support, and pre-treatment absenteeism were subsequently believed to provide a direct influence on absenteeism. Due to sample size limitations they were not able to test the complete model simultaneously, and instead split the model to examine the predictors of work-family conflict and absenteeism separately. They found that supportive supervision and satisfaction with child care arrangements were 38 significantly related to lower work-family conflict, and lower work-family conflict was related to lower absenteeism. Surprisingly, they' found. that supportive supervision.*was weakly related to higher absenteeism. The models that address dependent care issues tend to be fairly narrow in scope, however, other models that have been developed tend to focus either on the more general concept of work-family conflict which was reviewed earlier or specific work related factors. This is the case with Brett and Yogev’s (1988) model that considers the effect of a married couple’s characteristics (i.e., family stage, number of children, the use of paid help, and financial power) along with his and her work and nonwork factors on the level of his and her work restructuring. Not surprisingly the study found that both.men and women are restructuring work to accommodate family, although women restructure ‘more than men do, and that restructuring is systematically related to conditions of work and nonwork. Grover (1991) studied perceptions of the fairness of a hypothetical parental leave policy and attitudes toward parental leave takers. Specifically, he considered the effects of one's attitudes toward women, gender, whether or not one has children or is of childbearing age, and the likelihood of having children and taking a parental leave. The study, which supported social justice theory, found that individuals who would either benefit directly from parental leave or who were 39 similar to those who would benefit perceived the parental leave policy as more fair, and held more positive attitudes toward leave takers than those who would not benefit and were not similar to leave takers (Grover, 1991) . While these types of studies provide very limited insight on the relationship between work and family, they do not offer a comprehensive theory. One exception is a recently developed model that considers the effect of individual and organizational contextual influences on the work and family orientations, and the work and family outcomes of employed caregivers (Kossek & DeMarr, 1996). That model holds that the characteristics of the employed caregiver (i.e., household employment configuration, gender, job characteristics, and access to familial care) and the characteristics of the dependent (s) (i.e. , relationship to the caregiver, living arrangements, time spent on care, and limitations in activities of daily living) determine an individual's orientation toward the management of the boundaries between work and family, and influence the employee’s perception of the attractiveness of family supportive policies. This perception is also influenced by variables indicative of the work context (e.g. , formal family supportive policies, organizational climate regarding work and family integration, and peer and supervisor social support). How attractive an employee views various policies influences his/her decision to use a benefit. An individual's orientation 40 toward work.and family boundary“management (i.e., whether s/he prefers to segment or integrate work and family) and an employee’s use of a benefit determine his/her perceived level of work-family conflict which in turn influences both psychological and behavioral work and family outcomes. Psychological outcomes include things such as organizational commitment, turnover intent, burnout, life and job satisfaction, and depression. Behavioral outcomes include: turnover, absenteeism, work quality, willingness to accept work assignments, substance abuse, and withdrawal. While this model has yet to be tested it does hold promise as a link between the work—family conflict and dependent care literatures. D n r Pr f n N The remaining studies in the area of dependent care have primarily considered employees’ preferences and need for dependent care assistance, and the effectiveness of specific programs. Overall these studies have shown that employees’ needs are diverse, multifaceted, and changing (Kossek, 1990) and suggest that demographic characteristics may also play a role in employee preferences. Women are more likely to prefer job sharing or part time work (Kossek, 1990; Kossek, et al, 1993), while men are more likely to prefer referral services (Kossek, et al, 1993). Overall employees often rank sick care as the most desirable form of employer sponsored assistance (Kossek, 1990; Kossek, et al, 1993). One study also found that 41 on- or near-site child care is highly ranked as a employer-sponsored form of care, especially among single parents (Kossek, 1990). Other studies have found that onsite day care is unrelated to performance (Kossek & Nichol, 1992). and does not reduce work-family conflict (Goff, et a1, 1990) or absenteeism (Goff, et al, 1990, Kossek & Nichol, 1992). Kossek and Nichol (1992) concluded that child care benefits, such as onsite day care, provide the most help in attracting and retaining employees. They did, however, note the presence of a "frustration effect" for people on a waiting list for onsite care that involved the lowering of their perceptions of the fairness and attractiveness of employer sponsored child care. In the area of eldercare, research tends to focus predominantly on the impact of caregiving on the caregiver's well-being and employment (cf. Barnes, Given, & Given, 1992; Brody, Kleban, Johnsen, Hoffman, & Schoonover, 1987; George & Gwyther, 1986; Pratt, Schmall, & Wright; 1987; Scharlach & Boyd, 1989; Stone, Cafferata, & Sangl, 1987), the types of programs offered by employers (cf. Sullivan & Gilmore, 1991), and employee preferences for employer-sponsored assistance programs (cf. Anastas, et al, 1990; Scharlach & Boyd, 1989; Sizemore & Jones, 1990). Specifically, Scharlach and Boyd (1989) found that a 'majority of caregivers experienced emotional, physical, and financial strain and some degree of conflict between. work. and family as a result of their 42 caregiving responsibilities. Similarly, Brody, et al (1987) found that among daughters who were caring for an elderly parent, those who either had quit, considered quitting, or reduced their hours of work to provide care had the most impaired elders and experienced the most lifestyle disruptions and caregiving strain. George and Gwyther (1986) found that caregivers have substantially more stress symptoms and considerably lower levels of affect balance and life satisfaction, and are less able to pursue social activities at preferred levels. Pratt, et al (1987), focusing on the ethical concerns of family caregivers to dementia patients, reported that caregivers were concerned with the ethical dilemmas arising from conflicts between caregiving and other commitments to family, career, or personal well-being. Evidence of competing demands that resulted in the caregivers having to alter their work schedules in some fashion was also found by Stone, et al (1987). Based on the existing research it appears that providing care for an elder dependent is likely to have an impact on employees. Unfortunately, the current research merely scratches the surface. This is also true in the study of employee needs and preferences for employer—sponsored eldercare assistance programs. Studies typically find the programs most favored by employees providing eldercare are flexible hours, family illness hours, and information or referral services (cf. 43 Anastas, et al, 1990; Scharlach & Boyd, 1989). Kossek, et al, (1993) found.that employees ranked sick care assistance (i.e., improved leave policies for care of elder dependents), a companion program (i.e., home visitors organized by the employer), and eldercare referral assistance to be the top three options for employer-sponsored eldercare assistance programe. Sizemore and Jones (1990) found that the topics for informational seminars most desired by employees were: legal considerations, coping with mental illness in the older relative, changes as a relative grows older, and family decision making. While these studies identify some general needs, they do not address what types of employees are more likely to prefer certain types of programs. Overall the dependent care research, both child and elder, is more applied than theoretical in nature. While practitioners may find it helpful to some extent, a more holistic approach is needed to fully capture the intricacies of the work-family nexus and bridge the gap between the work— family conflict and dependent care literatures. This study attempts to fill gaps and overcome: deficiencies in the existing work-family and dependent care literatures by developing and testing competing models of the relationship between an employee's work and non-work characteristics, and his/her attitudes toward employer-sponsored. benefits and perceptions of work productivity. CHAPTER THREE MODEL AND HYPOTHESES As noted in the introduction, the work-family and dependent care literatures have not been integrated, and a comprehensive model that includes both predictors and outcomes of work and family stress has yet to be developed and tested, leaving gaps in the existing literature. The general research question that this dissertation seeks to answer is whether the relationship between an employee's work and non-work characteristics, and his/her attitudes toward employer- sponsored benefits and perceptions of negative work productivity is direct, or fully or partially mediated by the amount of stress experienced. MODELS OF THE ROLE OF STRESS IN EMPLOYEE ATTITUDES Three competing models have been developed and will be tested to determine which model provides the best fit. The first, the no-mediation model (Figure 1a) shows only direct effects from variables in each domain to domain specific outcome variables. In other words, family and dependent care characteristics will influence attitudes toward.benefits, but not perceptions of work productivity, while work characteristics will influence perceptions of work productivity but not attitudes toward benefits. Implicit in this model is the assumption that the work and family domains are segmented and do not influence one another. The second, 44 Ix >ufl>fiuuooonm xnoz no mcoflummoumm mufiumcwm oumzos mosufiuua wummwamulmwa ._ HMGOL.” ”MN HGMWHO 4. .uIQmH Gm w mdwawmuqmw wmfiuamcmm Hmmumu Houucoo maoomcom nlll oeumfiu on Mon xnoz mumo umoam no umou mxmmu no nonfidz mumo so ucmmm wEflB .IIII wocwofimmu-oo mumoam mo Hmnfinz mumo oafico no umou mumo \3 cofluUMumflumm cmnoawco me qu852 u H.n mEoocfi oHocmmoox cofiumusmflucou oaocmmsom uwocmo .Hmooz COADMAomznoz .MH whomflm 46 >ua>fluosooum xuoz wummmamUImwa qumwwmmwdmde aqmqmmmw 4‘ mqwammmmqw mmmuum xuoz A mmeuamcmo nmmnmu mo mcofluomoumm A monumcmm oumgos manuauua mmmuumao HmowmoHocoxmm \/ Houuooo masomcom "m a» as u mhmao xu mumo umoam mo umoo mxmmb mo nonssz mumo so ucmmm mEaB wocmoammu-oo 1< AMIIII mmmuumA. mnmoam no uwQEsz mumo ucwocmmmo \ 7 mumo oafico mo umou AVIIIIIII mumo \3 sofiuomHmwumm cmuoafino uo Hmn852 mEoocfl odocmmoom cofiumuomfiucoo oaonmmsom nmocmw .Hmooz cesumacmz-aasm .hH whamam 47 mammqfiso mwa .1 HMCOH MMNflGmmHO L\ WMMdmmflW mamammwmmw mmwuum EH03 ATIIIIIIIIII mmfiuamcmm ummumu I: Houucoo masowcum IIIL >ufi>auosoonm "mofiumauwuomu co xuoz xuoz mo macauomonmmAllllllllll wwwuuwflo wumo “moan no umoo mufiumcmm HMOfimoHocoxwm wxmmu uo HmnESZ oumzoh mosufiuu< wumu co ucmmm wEaB .IIIJ . _ wucmoammuuoo \7\7\r mmmuummlllllllllln muwoaw no Honssz mumu "wofiumfinmuomumcu on u up Hm ucmocmmmo a mnmo oflfico mo uwou wumo \3 :ofluomwmfiumm smuoflfico uo Hwnfioz "wofiumflnmuomumnu mum odes mEoocfl oHocmmsom cofiumuomfiucoo oHocwmsom “cocoa .Hmooz :oflumfiomz-HMAuHmm .UH whomflm 48 the full mediation model (Figure 1b) shows only indirect effects on the outcome variables through the stressvariables. Thus, stress is the cross-over point between the work and family domains. The third, the partial-mediation model (Figure 1c) shows an employee’s work and non-work characteristics having' both. direct and. indirect effects on the outcome variables. Both the full— and partial-mediation models are grounded in spillover theory which holds that the work and family domains influence one another. In addition to considering whether a noamediation, full- mediation, or partial mediation model best explains the relationship between the variables, this dissertation will also evaluate the impact of the individual variables in the models. Thus, each variable included in the models is discussed below and competing hypotheses are presented for each of the three models. W: QBBQBE Gender has long been one of the most studied variables in both the work-family and dependent care literatures. Typically, research has shown that women spend more time on family responsibilities (Galinsky, 1989; Pleck, 1985), and experience more role conflict (Wiersma, 1990) and work-family conflict (Glass & Camarigg, 1992) than men. Demands of the home have been found to affect women more strongly than men, regardless of their employment status (Broman, 1991). 49 Similarly, spillovers of distress and fatigue from work to family, and from family to work have been reported to be stronger for women than for men (Williams & Alliger, 1994). Stress, however, may manifest itself in different ways for men and women. Jick and Mitz (1985) found that women are more likely to experience psychological and emotional stress, while men are more likely to suffer from physical illness. Indeed, women have been found to be less likely to hold favorable attitudes toward managing work and family responsibilities (Kossek, 1990; Kossek & Nichol, 1992) , however, merely being female did not directly'predict absence (Kossek, 1990). It is interesting to note, however, that being female was negatively related to supervisor perceptions of child care related absenteeism (Kossek. & INichol, 1992). This suggests that traditional notions that women are less reliable employees due to family obligations are alive and. well despite research to the contrary. Wolf & Soldo (1994) found that among employed married women, caring for an elderly parent is not associated.with any reduction in hours of work. In other words, married women who accept parental care responsibilities do not make compensating reductions in their hours of paid work. Finley (1989) found that while females are more involved in elder caregiving than males, males do feel responsible to take care of their elderly parents but they do not actually fulfill this responsibility 50 to the extent that females do. It is also important to note that differences in levels of education.and working status did not account for those differences (Finley, 1989). One study (Anastas, et al, 1990) that used a rather broad definition of caregiving found that although 48% of elder care providers were male, women provided nearly twice as many hours of care per week as men and reported more strain. Women are significantly more likely to restructure their work to accommodate family needs than men (Brett & Yogev, 1988). Specifically, flexible work arrangements are far more likely to be utilized by female than male employees (Mattis, 1990). Greenhaus and.Parasuraman (1994) argue that wives’ work accommodations tend to be ongoing, whereas husbands' accommodations represent "special arrangements" to acute family stresses. When men do make work accommodations they may not be as obvious as women's. It has been reported that many men do take time off from work after the birth of a child, but they do so by piecing together other forms of leave (e.g., vacation, personal leave, sick leave) that they see as more acceptable (Pleck, 1989, cited in Rodgers & Rodgers, 1989). This suggests that impression management tactics may color employer perceptions of the true impact of family and dependent care responsibilities on their employees. In a study of gender differences in work-family conflict Duxbury & Higgins (1991) found some support both for Pleck's (1977) theory of asymmetrically permeable role boundaries for 51 men and women, as well as Hall’s (1972) concept of simultaneous (for women) versus sequential (for men) role demands. Specifically, the relationship between work involvement and.work-family conflict was stronger among women and the relationship between family involvement and work- family conflict was stronger for men. This suggests that high levels of involvement in nontraditional roles is problematic for men and women in dual career families. Significant gender differences were also found in the relationship between work conflict and family conflict in that men were more likely to allow work conflict to spill over into the home environment (Duxbury & Higgins, 1991). Significant differences were also found in the relationship between family expectations and family conflict, leading Duxbury .and Higgins (1991) to conclude that work and family roles are not mutually supportive for women and that women have less control given the same demands as they relate to family. Conversely, there were no gender differences in the relationship between work expectations and work conflict, which suggests that men and women in managerial or professional jobs are more similar (Duxbury & Higgins, 1991). Similarly, there were no gender differences between work involvement and work conflict, and family involvement and family conflict (Duxbury & Higgins, 1991) . In other words, women were not more likely to have a higher relationship between family involvement and family conflict than men, and 52 men.were not more likely to have a higher relationship between work involvement and work conflict than women. For the most part, the gender differences found by Duxbury and Higgins (1991) were attributed to societal expectations and behavioral norms. While traditional gender role stereotypes have in large part been supported by research, the results of some studies suggest that the differences may not be as great as once thought. Gutek, et a1 (1991) found no support for the hypothesis that women would report more FIW than men and men would report more WIF than.woment As part of a larger study of the relationship between. work and family stressors and psychological distress, Frone, Russell, and Cooper (1991) conducted a set of exploratory moderated regression analyses to examine whether gender influenced the magnitude of the relationship and found that none of the stressor by gender interactions were statistically significant. Similarly, Rice, Frone, and McFarlin (1992) found that gender did not moderate the relationship between work-family conflict and family satisfaction, job satisfaction, and leisure satisfaction. The conflicting results with respect to gender differences may be a result of the types of studies done and/or the assumptions made. Much.of the existing research has only studied women (cf. Brody, et al, 1987; Katz & Piotrkowski, 1983; Rosin & Korabik, 1990; Schwartzberg & Dytell, 1988; Williams, et al, 1991; Wolf & Soldo, 1994) or 53 organizations that have a high proportion of women such as hospitals (cf. Kossek & Nichol, 1992). The conflicting results may also be a result of the way variables are defined and/or stereotypes that people in our society hold, whether conscious or subconscious. This is especially true in the case of elder care research where the statistics that show a larger proportion of women proving care may be due to definitions of care that focus on personal care (Anastas, et al, 1990) or a man’s view that providing assistance for an elderly person is simply "helping out", as opposed to "eldercare". Some research also suggests men are and will be doing more with respect to family responsibilities. While women historically have been.more likely to have a larger burden of direct care even if they do work, there appears to be a value shift in our culture toward greater family involvement by men (Pleck, 1985) and a shift toward more egalitarian attitudes and a corresponding movement away from beliefs in traditional roles for men and women (Deaux, 1985). It has also been noted that as children get older fathers often tend to become increasingly involved in family responsibilities (Galinsky, 1991). In two studies at DuPont, Rodgers and Rodgers (1989) found that men’s reports of certain family-related problems nearly doubled from 1985 to 1988. Indeed, the more that men take on the responsibility for balancing work and family life the more likely they are to report higher levels of depression and stress (Burden & Googins, 1986). 54 Although values are beginning to shift and men are beginning to take on a larger proportion of family responsibilities, it is believed that at this point women still shoulder the majority of dependent care responsibilities. Thus, the following hypotheses are offered for the no-mediation (NM), full-mediation (FM), and partial- mediation (PM) models respectively. H 1-NM: Gender will have only a direct effect on attitudes toward family-friendly benefits. Women will have more favorable attitudes toward family-friendly benefits than men. H 1-FM: Gender will have only indirect effects through stress. Women will experience more dependent care stress than men. H 1-PM: Gender will have both direct and indirect effects on attitudes toward famdly-friendLy benefits and perceptions of negative work productivity. H h nfi r i n There has been limited research on the effect of household employment configuration on work-family stress and dependent care issues. This is in.spite of calls for a greater focus from both researchers and organizations on family structure (Schneer & Reitman, 1993). The research that does consider household employment configuration tends to focus specifically on the types and levels of work-family conflict in dual-career or dual-income families (cf. Brett & Yogev, 1988; Falkenberg & Monachello, 1990; Greenhaus, 1988; Higgins, et al, 1992) or considers the effect of spousal support (cf. Granrose, Parasuraman, & Greenhaus, 1992; Greenhaus & 55 Parasuraman, 1994; Schwartzberg & Dytell, 1988). With few exceptions (cf. Higgins & Ebeury, 1992; Strickland, 1992) researchers typically don’t consider differences among and within dual career and traditional (two-parents with a nonworking spouse) households. This also true for single parent and single person households. With the exception of Schneer and Reitman (1993), studies that examine differences involving alternative family structures are almost nonexistent. This is surprising given the considerable media and public attention that has been devoted to the decline in the number of two-parent households in our society. Research on dual career or dual income families has generally shown that when wives are employed, the percentage of housework and child care done by the husbands increases not due to his doing more work, rather it is a result of the wife spending less time in housework than before and less time than her non—employed counterparts (Pleck, 1985) . This is supported by Barnett and.Baruch’s (1987) finding that fathers spent more proportional interaction time, performed more child-care tasks alone, and.did more feminine home chores when their wives were employed, however, with total interaction time as the dependent variable the only significant predictor was the number of hours the wife worked. Aryee (1992) also found the greater the hours worked by the spouse, the more job-parent conflict. In general there are two opposing views on the impact of 56 dual career status on work-family conflict. The first is that it generally has a negative impact. Higgins and Duxbury (1992) in a study comparing dual career and traditional family men, found the relationships between work conflict and family conflict, and work-conflict and work-family conflict were significantly stronger for men in dual-career families. Wife’s employment has also been found to be significantly negatively related to husband's job satisfaction and quality of life (Parasuraman, Greenhaus, Rabinowitz, Bedeian, & Mossholder, 1989), as well as quality of work life (Higgins & Duxbury, 1992). Parasuraman, et al (1989) found that husbands of employed women were also less satisfied with child care arrangements than were husbands of housewives. In contrast, Goff, et al (1990) found that employees whose spouse cared for the children did not experience less work-family conflict and absenteeism than those whose spouse was employed and child care occurred outside the home. A second view posits that advantages associated with having a working spouse results in a weaker relationship between family conflict and work-family conflict (Higgins & Duxbury, 1992). One of the advantages is an increased family income which affords both spouses more career autonomy and independence since the responsibility for providing for the family does not rest entirely on one person. Rosin (1990) concluded that wives' income allowed husbands the option to leave unsatisfactory employment, reject excessive demands for 57 travel or overtime, refuse disruptive relocations, participate in high risk startup ventures, or start their own businesses. Another advantage that dual career status may afford both partners is that the more they become involved in nontraditional roles the greater their opportunity for gratification and the better able they are to cope. Lambert (1991) found limited support for the contention that family responsibilities play a major role in determining the work responses of both men and women. The study found that characteristics of the spouse's employment did influence work responses of both men and women although in different ways. The number of hours a spouse spends in paid employment was positively related to job involvement for men, however it was negatively related to job involvement for women. In contrast, the presence or absence of a husband was not a significant predictor of role strain among employed black women (Katz & Piotrkowski, 1983). It may be that the relationship between spousal employment and work-family conflict is moderated by the level of support one receives from his/her spouse. Spousal support can be either emotional (i.e., moral support) or physical (i.e., help with housework or dependent care). Emotional support is especially important in dual- career marriages because stressful work events tend to carry over into family life through "negative emotional spillover" (Greenhaus &.Beutell, 1985). Studies that have considered the effects of spousal support have found positive effects on 58 family satisfaction (Parasuraman, Greenhaus, & Granrose, 1992) in that the more spousal support, the less job-spouse and job- parent conflict (Aryee, 1992) . Lack of spousal support is significantly correlated with lower self-esteem among employed women (Schwartzberg & Dytell, 1988). Falkenberg and Monachello (1990), acknowledging differences within dual-earner households, hypothesized that spouses in.dual-career households will exhibit similar levels of work-family involvement and that ‘men in dual-career households will provide more support to their wives than men in dual-income households. Strickland (1992) argues that instead of a simple dichotomy (e.g. , housewife vs. career wife), wives’ achievement roles actually comprise a continuum from lesser to greater direct achievement. It follows that there would also be a range in the need for organizational supports, child care and housekeeping needs, marital power structures, and degrees of role conflict and overload (Strickland, 1992). While there appear to be differences within two-earner families related to the amount of spousal support, this study instead focuses on the differences between family structures, as it is believed that these differences are more substantial. Unfortunately, there is a dearth of research on the effects of alternative family structures work-family conflict and dependent care. One small exploratory study found that single parents and dual career parents were more likely to experience 59 problems with their eldercare arrangements, hold less favorable attitudes toward managing work and eldercare, and be absent more than two-parent households with only one spouse working or households without child dependents (Kossek, et al, 1993). While the results of that study were not highly significant, they do suggest that a variety of family structures should be studied. Schneer and Reitman (1993) have proposed a three— dimensional family structure topology with dimensions for marital status, parental status, and the employment status of the spouse. Combinations of the dimensions allow for six different family structures: (1) single, no children; (2) single, children; (3) married, no children, one income; (4) married, no children, dual-income; (5) married, children, one income; (6) married, children, dual-income. Using their topology, Schneer and Reitman (1993) found that family structure was related to both income and career satisfaction for both men and women. Specifically, they found that men in post-traditional families are less rewarded than men in traditional families, but this was not true for women (Schneer & Reitman, 1993) . This suggests that men are still expected to fit the traditional managerial model. If this is true, a logical extension in the area of work-family conflict is that dual-earner households will experience more work-family conflict than traditional households. This is consistent with the contention that members of two-earner families, one-parent 60 families, and families with young children are likely to experience work-family conflict and job tension (Kelly & Voydanoff, 1985; Voydanoff, 1988b; Voydanoff & Kelly, 1984). Although there has not been significant empirical research on the impact of single parent status on work-family conflict, it is expected that single parents will experience the highest levels of work-family conflict. This is due to having to shoulder the burden of household responsibilities and dependent care without the benefits of physical or emotional spousal support discussed earlier. Indeed, Kossek (1990) concluded that single parents appear to be a special employee group that is high on child care assistance needs. Clearly, a major shortcoming in the work-family and dependent care literatures is the lack of attention given to household configuration. Based on the preceding review, the following hypotheses are offered: H 2-NM: Household configuration will have direct effects on attitudes toward family-friendly benefits. Single parents, and those in dual earner families will have more favorable attitudes toward benefits than those with a nonworking spouse or single persons. H 2-FM: Household configuration will have only indirect effects through stress. Single parents will experience the most dependent care stress, followed by those in a dual—earner families. Persons with a nonworking spouse and single persons will experience the least dependent care stress. H 2-PM: Household configuration will have both direct and indirect effects on attitudes toward family- friendly benefits and perceptions of work productivity. 61 lBQQmfi Income is another understudied variable as it relates to work-family conflict, even though an adequate income appears to be integral to a fulfilling family life (Piotrkowski & Katz, 1982; Pleck, 1985). The level of disposable income available in a household will clearly be affected by the amount of money spent on dependent care. Low-income families spend a much greater proportion of their budgets on child care than wealthier families do, often as mmch as they pay for housing (Hofferth, 1988). In 1992, 14.5% of the people in the United States were below the poverty level, which for a family of four was $14,335 (U.S. Bureau of the Census, 1994). According to Ellen Galinsky of the Families and Work Institute, those who pay a higher proportion of their family income for child care have more conflict (Solomon, 1994). Conceptually this can.be extended to additional expenses associated. with. eldercare. One study found that 54% of respondents with eldercare responsibilities reported some degree of financial strain associated with their caregiving responsibilities (Scharlach & Boyd, 1989). Barnes, et al (1992) found that currently employed caregiver daughters had more personal and material resources in terms of money, than did other groups of caregiving daughters. Similarly, Brody, et al, (1987) found that women who had.quit work.due to eldercare responsibilities had lower family incomes and received the least amount of paid help. Stone, et al (1987) also found that 62 eldercaregivers who shared care responsibilities with paid helpers reported higher incomes than those who did not. Higher household incomes also allow for additional flexibility'in.dealing with.dependent care issues and strains, which should help the caregivers to better cope with their dependent care responsibilities. For example, an individual who is experiencing considerable stress related to their elder or child care responsibilities can hire a babysitter or nurse/companion to allow them some time away, provided they can afford it. Higher income may also reflect flexibility in scheduling where and when the work gets done. In other words, individuals in higher income jobs are less likely to have to "punch a clock", and more likely to have work that can.be'made up by taking work home or coming in early or staying late than lower' income jobs. Indeed. after' using 'management 'versus nonmanagement status to tap the concept of flexibility, Kossek (1990) concluded that using salary data might be a better indicator' of the level and importance of an employee’s position, the probable freedom to juggle one’s schedule for care, and the amount of resources the employee has with which to purchase quality care. Based on that rationale, the following hypotheses are offered. H 3-NM: Household income will have direct effects on attitudes toward family-friendly benefits. The lower an individual's household income the more favorable his/her attitudes toward employer- sponsored benefits. 63 H 3-FM: Household. income ‘will have indirect effects on attitudes toward family-friendly benefits through stress. The higher an individual’s household income the less dependent care stress. H 3-PM: Household income will have both direct and indirect effects on. attitudes toward family-friendly benefits and perceived work productivity. WW: r r 14 In 1993, 35% of all households in the United States had children under age 18 (U.S. Bureau of the Census, 1994). Research has shown that parents experience conflict between work.and family more often than.other workers (Pleck, Staines, & Lang, 1980), and the greater the parental demands the more job-parent conflict (Aryee, 1992). Larger families are likely to place more demands on a person's time than small families (Cartwright, 1978; Keith & Schafer, 1980), and result in more strain (Katz & Piotrkowski, 1983). These findings are likely associated with the additional time required by additional children. This is consistent with Barnett and Baruch’s (1987) finding that fathers' individual and.proportional interaction time with the children was greater when there were more children in the family. One study found that each child under age six reduced weekly hours of work by about four and one half, while children age 6-12 reduced hours of work by about three per week (Wolf & Soldo, 1994). Logically, the more children a person.has, the more visits to the doctor and.dentist, and the 64 more school and social events to attend or for which to provide transportation. Having more children also increases the likelihood of having to make multiple dependent care arrangements which has been hypothesized to have a negative influence on the problems with child care arrangements experienced by a parent (Kossek, 1990) . The greater the likelihood of having to deal with multiple child care arrangements, the greater the likelihood of subsequent breakdowns in those arrangements. There is, however, some level of disagreement in the literature on the relationship of family size to work-family conflict. Goff, et al (1990) found the number of children under age five was not significantly related to the level of work-family conflict. Similarly, Lobel and St. Clair (1992) found that neither increasing numbers of children nor the presence of preschoolers had a significant direct effect on work effort. The presence and number of children in the household may have other effects as well. Kirchmeyer (1992) found that time spent in parenting was significantly positively related to both job satisfaction and organizational commitment. Conversely, Lambert (1991) found the effects of the number of children, presence of children, and the age of the youngest child in the household to be nonsignificant in. relation. to job involvement, job satisfaction, and intrinsic motivation. While it appears there are some potential work-related benefits of having children, 65 based on the literature I argue that when considering the relationship between number of children under age 14 in the household, and the level of stress and one's views toward benefits, that the relationship will be positive. H 4-NM: The number of children under age 14 will have a direct effect on attitudes toward family-friendly benefits. The more children one has the ‘more favorable one’s attitudes toward benefits. H 4-FM: The number of children under age 14 will have only indirect effects through stress. The more children for whom one is responsible the higher the dependent care stress. H 4-PM: The number of children under age 14 will have both direct and indirect effects on attitudes toward family-friendly benefits and perceptions of negative work productivity. Setiefeetien with gete While there is not a tremendous amount of research on a working parent’s satisfaction with child care arrangements, the research.that has been.doneehas produced fairly consistent results. In short, a person who is less satisfied with their child care arrangements can reasonably expect to have it affect them at work. Rosin and Korabik (1990) reported that women MBAs who had left paid employment cited difficulty finding good.quality child care as a:beason for leaving. Goff, et a1 (1990) found that employees who were more satisfied.with the quality of their child's care experienced less work-family conflict. Problems with child care arrangaments have been found to be significantly negatively related to holding favorable 66 attitudes toward managing work and child care responsibility (Kossek & Nichol, 1992) . Problems with child care arrangements may be more common with younger children who are not yet in school all day. While school age children might need before and after school care depending on their age, the fact is that schools provide adult supervision for children for a large part of the day. Rosin and.Korabik (1990) found.women.MBAs who had left paid employment and those working part time were most likely to be married, with multiple preschoolers, and husbands in the top income bracket. The heavy demands associated with the rearing of preschool children were the most important determinants of the career decisions made by these women (Rosin & Korabik, 1990). Based on the literature the following hypotheses are offered. H S-NM: The degree to which an employee is satisfied with child care arrangements will be directly related to his/her attitudes toward family-friendly benefits. Individuals who are less satisfied with their child care arrangements will have more favorable attitudes toward benefits. H 5-FM: Satisfaction.with child care arrangements will have only indirect effects through stress. Those who are less satisfied will experience more dependent care stress and have more favorable attitudes toward benefits and.perceive themselves as less productive due to their personal responsibilities. H 5-PM: Satisfaction with child care arrangements will have both direct and indirect effects on attitudes toward family-friendly benefits and perceptions of negative work productivity. 67 il r The cost of child care is likely to vary dramatically depending on the source of care (e.g., relative vs. formal daycare facility). as well as the geographic region of the country. As was noted in the discussion of the income variable, those who pay a higher proportion of their family income for child care have more conflict (Solomon, 1994). Although it might be argued that the more one spends on child care the higher the quality of the care, it could also be argued that low or no cost child care (i.e., familial care) is of higher quality. Since research has not specifically addressed the issue of the relationship between the cost of child care, the amount of stress one experiences, and subsequently the degree to which an employee needs or values employer-sponsored dependent care benefits, the following hypotheses are offered based on the premise that the higher the cost of child care the more of a burden it becomes. H 6-NM: The cost of child care will be directly related to attitudes toward family-friendly' benefits. Employees who incur higher costs for child care will have more favorable attitudes toward.benefits. H 6-FM: The cost of child care will have an indirect effect on attitudes toward benefits through stress. Those who incur higher costs for child care will experience ‘more dependent care stress and subsequently have more favorable attitudes toward benefits and.perceive themselves as less productive due to their personal responsibilities. H 6—PM: The cost of child care will have both direct and indirect effects on attitudes toward family- friendly benefits and perceptions of negative work productivity. 68 l r h r eristi s: Nu r f El Unlike the child care research that shows both positive and negative employment-related outcomes associated with minor dependents, the research on eldercare overwhelmingly finds only negative outcomes. When caring for an infant or child, the future holds promise of a gradual reduction in dependency, while caring for an impaired older person typically presages continuing or increasing dependence (Brody, 1985) . It has also been said that eldercare needs may be more complex than child care (Friedman, 1990) and that for many women, parent care is not a single time-limited episode in the life course (Brody, 1985). For some time now there has been a common myth that children in the U.S. are more self-centered today and do not provide the same level of care for their elders that they once did. The fact is that nowadays adult children provide more care and more difficult care to more parents over much longer periods of time than they did in the "good old days" (Brody, 1985). Surprisinglyy only 22% of people over 85 are in.nursing homes (Winfield, 1987), and it has been estimated that for every elderly person in a nursing home there are at least two others with an equivalent level of disability that are not institutionalized (Brody, 1985). Approximately 80% of their care is provided by members of their family (U.S. Senate Special Committee on Aging, 1985-86). The negative consequences of caregiving for the 69 caregivers themselves include: the emotional stresses of coming to terms with the changing role and capability of the older person; the restrictions on time and freedom; economic burdens, including loss or curtailment of employment; and detrimental effects on the caregiver’s marital, family, and social relationships (Horowitz, 1985). In a study of spouses and adult children with and without siblings, all caregiver groups reported a negative impact on health, reported receiving less affective support, and expressed feelings of abandonment over time (Barnes, et al, 1992). A.study by Pratt, et a1 (1987) reported that 29% of caregivers identified ethical dilemmas that arose from conflicts between caregiving and obligations to family, career, or personal well-being. George and Gwyther (1986) found that caregivers average nearly three times as many stress symptoms as noncaregivers from community-based samples, and are less able to pursue social activities as preferred levels. In addition to their caregiving responsibilities, many caregivers are employed outside of the home. Overall, employees who are assisting an elder dependent are likely to experience increased stress and related health problems, less leisure time, greater work-family conflict, and more frequent instances of coming in late and leaving early (Galinsky, 1991). Scharlach and Boyd (1989) found that 23% of nearly 2000 respondents of an employee survey were assisting an elderly person. Caregivers were more likely to experience 7O interference between their jobs and family responsibilities and were more likely to miss work than the noncaregivers. Specifically, 80% reported some degree of emotional strain, 60% cited physical strain, 54% identified financial strain, and 40% reported interference with their work activities as a result of their caregiving responsibilities (Scharlach & Boyd, 1989) . Similarly, in a study among working women, 58% reported that parent care made them miss work, 47% reported work interruptions, 18% said it made them lose pay, 17% indicated it robbed them of the energy to do their work well, 15% had limited their job choices, and 17% said caregiving made them wish they did not work (Brody, et al, 1987). Employees with responsibility for the oldest old (age 85 and older) must also contend with the length, diversity, and intensity of their needs (Friedman, 1986). The 1982 National Long-Term Care Survey by the U.S. Department of Health and Human Services reported that 31% of all caregivers were employed outside of the home (Stone, et al, 1987). Whilewmany may assume that working caregivers spend less time on eldercare, studies have indicated that employment status is not related to the overall amount of help provided to elderly persons (Brody, 1981). The amount of time spent on eldercare varies considerably. Surveys have reported that on average employees who provide eldercare assistance spend fram as little as six to ten hours per week (Azarnoff and Scharlach, 1988) to as much as four hours per day, seven days 71 a week (Stone, et al, 1987; U.S. Department of Labor Women’s Bureau, 1986). Scharlach (1987) found that employed daughters did not provide significantly less contact and assistance than those daughters who were not employed. Among caregivers in the 1982 national survey, 9% had.quit work to provide care, and of those who were working outside of the home, 20% were conflicted, 21% worked fewer hours, 29% rearranged their schedule, and 19% took time off without pay (Stone, et al, 1987). Bunting (1989) proposed that caregivers facing increased requirements for care of an older family member will, when choosing between this dependent care and care of themselves, perform fewer self-care actions. Despite the aging population, the number of working caregivers, and the potential impact on the business community, relatively little empirical research has been done on the effects on employees of caring for elder dependents. Not only has there been little interest in this area in the academic community, there has also been little interest shown by the business community with relatively few U.S. employers having formally addressed the issue of employees’ needs concerning eldercare. A survey of personnel executives revealed that 70% believed that some percentage of their employees were affected by eldercare, yet most of these employers had done little or nothing to help employees who have eldercare responsibilities (Magnus, 1988). Overall, companies have identified excessive stress and physical 72 complaints, and a decrease in productivity and quality of work as problems for employees providing eldercare (Friedman, 1986). More specifically, employers have cited: absenteeism, tardiness, visible signs of stress, excessive phone calls, unavailability for overtime work, requests for reduced hours, turnover, health problems, decreased quality of work and increased work accidents as problems related to eldercare responsibilities (Bureau of National Affairs, 1989). The problems that have been cited by employers thus far may only be the tip of the iceberg. Since most of the growth in the female workforce involves comparatively younger women whose parents are not yet old enough to require daily assistance, the workplace has probably not yet felt the full effects of elder care problems (Rodgers & Rodgers, 1989). It is also likely that over time employees will be providing assistance for multiple elderly dependents. As with.the number of children for whom one is responsible, the number of elders for whom one is providing care is likely to influence the amount of time and the logistical complexity for the caregiver. Thus, it is believed that the relationship between the number of elder dependents and the level of family-stress will be positive. H 7-NM: The number of elders for whom one is providing care will have a direct effect on attitudes toward family-friendly benefits. The more elders for whom one is providing care the more favorable one's attitudes toward benefits. H 7-FM: The number of elders for whom one is providing care will have only indirect effects through stress. The 73 more elders for whom one is providing care the higher the dependent care stress. H 7-PM: The number of elders for whom one is providing care will have .both. direct and indirect effects on attitudes toward family-friendly benefits and perceptions of negative work productivity. Ce-Residedee With Elder Dependents Unlike the number of elders for whom one is providing care, research.has considered the role of living arrangements, and in particular, living with an elder dependent. Overall, the research has shown that shared households are a strong predictor of strain (Brody, 1985). This is not to say that other living arrangements do not produce stress. For example, it has been reported that anxieties about final separation usually occur when a parent enters a nursing home (Shaw, 1987). However, caregivers who reside with their dependents are most likely to use psychotropic drugs, report the highest level of stress symptoms, significantly lower household incomes, and the lowest levels of affect, life satisfaction, and participation in and satisfaction with their social participation (George & Gwyther, 1986). This is likely a result of having to deal with the situation on a continuous basis. In other words, the caregivers can not simply go home for a break to get away from the caregiving situation. Shared households may be more common than one may think. One study found that of adult children providing care for an elderly parent, 80% lived with their elderly parents (Barnes, et al, 1992). It has been suggested in the caregiver 74 literature that shared households become a more common phenomenon as the health of the older person deteriorates (Noelker and Poulshock, 1982; Troll, 1971). This is supported by the finding of Stone, et al (1987) of a positive relationship between the level of disability among elders and shared living arrangements. Assuming that shared households reflect poorer health of the elderly person and increased assistance required, combined with the more continuous nature of the care, it is expected that caregivers who live with the elderly person will experience a greater impact on their paid employment than those with other living arrangements. H 8-NM: Living with an elder dependent will have a direct effect on attitudes toward family-friendly benefits. Those who live with an elder dependent will hold more favorable attitudes toward benefits. H 8-FM: Living with an elder dependent will have only indirect effects through stress. Those who live with an elder dependent will experience higher levels of dependent care stress than those who do not share a household. H 8-PM: Living with an elder dependent will have both direct and indirect effects on attitudes toward family-friendly benefits and perceptions of negative work productivity. Time Spent en Care The amount of time involved in providing care for an elder dependent can vary considerably. One caregiver may spend a few hours each week helping with household or financial matters while another may spend several hours each day providing meals and personal care. As noted earlier the average time spent providing eldercare assistance can vary 75 considerably (Azarnoff and Scharlach, 1988; Stone, et al, 1987; U.S. Department of Labor Women’s Bureau, 1986). In one study of employed caregivers, the median amount of assistance given was six hours per week, with 6% of the respondents providing 40 hours per week.or more (Scharlach.&.Boyd, 1989). Another study of a group of married daughters providing care to their mothers found that over 50% had helped during only two or fewer days during the week prior to the survey, yet about 36% provided help all seven days of the pre-survey week indicating tremendous variability in the actual intensity of caregiving (Wolf & Soldo, 1994). Thus, it seems that even when living arrangements is included as a variable, that the time spent providing care to an elderly dependent must also be assessed to determine the impact of caregiving. Further, caregivers may be "on call" making adjusting the volume or timing of paid employment more difficult (Wolf & Soldo, 1994). Barnes and her colleagues (1992) reported that caregiver daughters who had terminated employment indicated that caregiving activities had a greater impact on their schedule than did daughters who had never been employed. Thus, the following hypotheses are offered. H 9-NM: The amount of time a caregiver spends providing care for an elder dependent will have a direct effect on attitudes toward family-friendly benefits. The more time spent on caregiving the more favorable one’s attitudes toward benefits. H 9—FM: The amount of time a caregiver spends providing care for an elder dependent will have only indirect effects through stress. The more time spent providing care the higher the dependent care 76 stress. H 9-PM: The amount of time a caregiver spends providing care for an elder dependent will have both direct and indirect effects on attitudes toward family- friendly benefits and perceptions of negative work productivity. Numh§r_9£_lasks The most frequently cited type of eldercare assistance is help with shopping and transportation (Stone, et al, 1987). however, there is also a great deal of variation in the type and number of tasks with which a caregiver may be assisting. A caregiver may be assisting with activities of daily living (ADLs) such as eating, dressing, bathing, toileting, transference, and mobility, as well as instrumental activities of daily living (IADLs) such as shopping, housecleaning, or transportation. A national study of caregivers found 80% helped with household tasks, 67% provided assistance with one or more personal hygiene functions, 46% assisted with indoor mobility, 50% administered medication, and 50% helped with financial matters (Stone, et al, 1987). Clearly, the 'more tasks with. which. a caregiver is providing assistance the more time required, however, it is also believed that the more tasks the greater the perception of burden and the more stress experienced. George and Gwyther (1986) found the more serious the patient’s symptoms, the lower the self-rated health, the higher the level of stress symptoms, and the less time the caregiver spends relaxing. Another study found that daughters who had terminated 77 employment were more involved in ADL tasks than were employed daughters (Barnes, et al, 1992). Similarly, Wolf and Soldo (1994) report that the number of ADL limitations reduce a woman's propensity to be employed, but do not deter her from providing parent care. Thus, the following is offered. H 10-NM: The number of tasks with which a caregiver is providing assistance will have a direct effect on attitudes toward family—friendly benefits. The more tasks for which one is providing assistance the more favorable one’s attitudes toward benefits. H 10-FM: The number of tasks with which a caregiver is providing assistance will have only indirect effects through stress. The more tasks the greater the dependent care stress. H lo-PM: The number of tasks with which a caregiver is providing assistance will have both direct and indirect effects on attitudes toward family- friendly benefits and perceptions of negative work productivity. stt df Elder gate While the caregiver's income is sometimes included as a variable in elder care research (cf. Barnes, et al, 1992; Stone, et al, 1987; Wolf & Soldo, 1994), most studies have not considered the additional expense incurred by caregivers. These costs can vary dramatically and can be substantial. One study reported that the amount of financial assistance provided by employed caregivers ranged from.$20 to $3,000 per month, with a median of $200 (Scharlach & Boyd, 1989). While the perception may be that the cost of eldercare is shouldered by either the elderly persons themselves or the government, 45% of the respondents in a study of employed caregivers were 78 providing direct financial assistance to their elder dependents (Scharlach & Boyd, 1989). Providing financial assistance to elder dependents may result in financial strain on the caregivers. Scharlach and Boyd (1989) found that over half of the caregivers reported some degree of financial strain associated with their caregiving responsibilities. In a study of the ethical concerns of family caregivers to dementia patients, 13% discussed inequities in health care financing (Pratt, et al, 1987). It is believed that the greater the cost involved, the more stress experienced, and subsequently the more employees will need and value employer-sponsored dependent care benefits. This is supported by George and Gwyther’s (1986) finding that a caregiver's household income and perceived economic status were related to the need for more social support. Thus, the following hypotheses are offered based on the premise that the higher the cost of elder care the more of a burden it becomes. H 11-NM: The cost of elder care will be directly related to attitudes toward family-friendly' benefits. Employees who incur higher costs for elder care will have more favorable attitudes toward benefits. H 11-FM: The cost of elder care will have an indirect effect on attitudes toward benefits through stress. Those who incur higher costs for elder care will experience more dependent care stress and subsequently have more favorable attitudes toward benefits and.perceive themselves as less productive due to their personal responsibilities. H 11-PM: The cost of elder care will have both direct and indirect effects on attitudes toward family- friendly benefits and perceptions of negative work 79 productivity. W h ri i W In addition to the effects of individual and dependent care characteristics, it is believed that characteristics of the work situation also influence an employee’s perceptions of stress. The first of which is characteristics of the job, specifically the amount of control one has over his/her work schedule. Work schedule inflexibility may create structural interference between work roles and family roles (Pleck, et al, 1980; Quinn & Staines, 1978). Lack of flexibility in one's work schedule induces work- family conflict in that it limits the time or freedom one has to attend to family matters (Greenhaus & Beutell, 1985; Greenhaus, Parasuraman, Granrose, Rabinowitz, & Beutell, 1989; Keith & Schafer, 1980; Pleck, et al, 1980). In developing a model of social support provided. by two-earner couples, Granrose, et al (1992) hypothesized that schedule inflexibility would be negatively related to both instrumental and emotional support provided by wives and husbands. Glass and Camarigg (1992) found a significant negative relationship between flexibility and ease of work, and work-family conflict. It has been suggested that the greatest stress may be caused by the rigidity of the work day and one's inability to handle emergencies, or to schedule needed appointments 80 (Friedman, 1990). In 1991, only 15.1% of all workers had schedules that allowed them to vary the time they begin and end their work day (U.S. Bureau of the Census, 1994). What this does not reveal is how much flexibility one actually has in starting or ending his/her work day. In a study by Rodgers & Rodgers (1989), one-third to one-half of parents said they do not have the workplace flexibility to attend teacher conferences and important school events. In a study of employed black women, Katz and Piotrkowski (1983) found that autonomy was a much stronger predictor of family role strain than.number of hours worked. Similarly, Aryee (1992) found the more task autonomy the less job-spouse, job-parent, and job- homemaker conflict. Feelings of personal control, with respect to having the ability to choose when and how to complete an activity and having personal control over the activity, have been associated with higher levels of calmness and lower levels of distress (Williams & Alliger, 1994) . Friedman (1990) reported that of all the benefit options or policy changes a company might contemplate, employees seem to overwhelmingly prefer ways to make their work schedules more flexible (i.e., large windows of start and stop times). In.the area.of schedule control, the following hypotheses are offered. H 12-NM: The amount of control one has over his/her schedule will have a direct effect on perceptions of work productivity. The more control one already has the less likely one is to perceive his/her productivity to be lower due to personal responsibilities. 81 H 12-FM: The amount of control one has over his/her schedule will have only indirect effects through stress. Having more control will be related to lower levels of work stress. H 12-PM: The amount of control one has over his/her schedule will have .both. direct and indirect effects on attitudes toward family—friendly benefits and perceptions of negative work productivity. W Greenhaus and Beutell (1985) proposed that work-family conflict would be strongest when there are negative sanctions for noncompliance with role expectations. The basic premise is that if strong negative sanctions are absent, there will be less pressure to comply with role demands (Gross, Mason, & McEachern, 1958). For'exampleg if employees perceive that they are likely to "get in trouble" with their boss for calling home to check on their children after school they have two basic choices - either make the call or not. If they make the call they are likely to do it in a covert manner to avoid trouble with the.boss, which may leave them feeling guilty for not being up front. If they opt not to make the call they are likely to worry more about their children. Regardless of the choice they make their level of work stress is likely to be higher than if they could simply make a quick call without concern for any negative impact on their employment situation. While the Greenhaus and Beutell (1985) article is generally considered to be a seminal piece in the work-family conflict literature and is widely cited by others, research testing that particular proposition is scant. That may be due to a 82 reluctance on the part of organizations (and the researchers studying them) to leave themselves open to the possibility that employees may perceive the organization to be unsupportive. The lack of research does not mean that employees do not believe that career' penalties are real. One study that evaluated the effectiveness of flexible spending accounts and leave of absence program found a majority (75%) of the respondents felt that there was a subtle pressure not to take a leave of absence, and 40% felt that those who had taken a leave were less likely to get promoted than an equally qualified person who had not taken a leave (Kossek, et al, 1993). Thus, it should come as no surprise that only about 9% of the respondents had actually taken a leave under the program (Kossek, et al, 1993). This suggests that employee perceptions of career penalties do influence their behavior. While they did not specifically consider employee perceptions of negative sanctions, Lobel and St. Clair (1992) did investigate the effects of family responsibilities and career identity salience on merit increases. They found that family-oriented women with preschoolers received higher merit increases than family-oriented men with preschoolers, and career-oriented women with preschoolers received lower merit increases than career-oriented men with preschoolers. This lead them. to conclude that ‘merit pay increases may' be allocated on the basis of conformity to gender role 83 stereotypes (Lobel & St. Clair, 1992). In a similar vein, Powell and Mainiero (1992) suggested that the relationship between success in career and success in relationships with others may be positive when organizations place minimum constraints on women's opportunities for career success and offer programs that help working mothers to handle their family responsibilities, and negative when organizations place considerable constraints on women’s opportunities and offer no such programs. H 13-NM: Employee perceptions of career penalties will have a direct effect on perceptions of work productivity. The more one perceives there to be career penalties the more likely one is to perceive his/her productivity to be lower due to personal responsibilities. H 13-FM: Employee perceptions of career penalties will have only indirect effects through stress. Perceptions of more penalties will be related to higher levels of work stress. H 13-PM: Employee perceptions of career penalties will have both direct and indirect effects on attitudes toward family-friendly benefits and perceptions of negative work productivity. W Greenhaus and Parasuraman (1994) identify three roles of social support in the stress process. The first is referred to as a "health sustaining" function which reflects a positive main effect of support on well-being. The second, identified as a "stress prevention" function, exhibits a negative main effect of support on the stressor in that social support reduces the environmenta1.pressures that produce the stress in I'm".- _._._x _ _ 84 the first place. The third, called the "buffering" function, reflects a moderating effect of support on the relationship between stress and well-being in that social support attenuates the negative relationship between stress and well- being, thereby protecting the individual from the severe consequences of stress. Research has found a negative association between. perceived burden and social support (Zarit, Reever, & Bach-Peterson, 1980). More specifically as it relates to this study, a more negative work setting and less social support are significantly related to work-family conflict (Burke, 1988) in that low levels of leader support and interaction facilitation appear to produce work-family conflict (Jones & Butler, 1980). It has been said that all the policies and programs in the world don’t mean much to an employee who has to deal with an unsupportive boss (Rodgers & Rodgers, 1989) . Friedman (1986) maintains that midlevel management is often unaware that the company has identified a particular problem and is interested in.a more compassionate response from.supervisors. Thus, a boss may often be unsupportive because of mixed signals from above (Rodgers & Rodgers, 1989). Even without mixed signals supervisors must simply be aware of formal programs. With respect to programs established by organizations to provide assistance to employees with elder dependents, companies say that without that knowledge, managers might be reluctant to give employees the release time 85 to participate in seminars, or the flexibility to take personal leave time, even though it is sanctioned by the company (Friedman, 1986). Rodgers and Rodgers (1989) maintain that given the lack of government regulation dealing with work-family issues, we leave decisions about flexibility and the organization of work to individual companies, which means that the decisions of first—line managers in large part create our national family policy. Thus, even when there is no formal organization policy, work restructuring can and is being done informally, by agreement between employee and manager (Hall, 1989). Anastas and her colleagues (1990) found that many survey respondents received informal support in the workplace, with 54% reporting that their immediate supervisors knew of their caregiving responsibilities. Conversely, another study found that only 26% of respondents felt their supervisor supported a flexible spending account program, and only 25% perceived supervisor support for a leave of absence program (Kossek, et al, 1993). Supportive coworker relationships have been found to be positively related to women's job satisfaction, and supportive supervision positively related to women’ s intrinsic motivation (Lambert, 1991). Galinsky (1988) maintains the relationship with the supervisor is a particularly important and powerful predictor of work-family problems. Taken to an extreme, a poor relationship with one's supervisor can ultimately lead to 86 turnover. Rosin and Korabik (1990) found the working relationship with supervisor was in fact a significant reason for women to leave employment. While women tend to receive more co-worker support than men, both men and women have been found to receive similar levels of support from their supervisors (Greenhaus & Parasuraman, 1994). Such work support has significant main effects on the job satisfaction. of both. men and. women (Parasuraman, et al, 1992). While on the surface it appears that supervisor support is highly desirable for employees, there may be other considerations. One study found that while supportive supervision was related to less work-family conflict, it was also related to increased absenteeism at a marginally significant level (Goff, et al, 1990). Further, not all studies have come to the same conclusions concerning the effects of supervisor support. Frone, et al (1991) found that social support did not buffer the stressor-distress relationship. Similarly, Kossek and Nichol (1992) found the relationship between.the amount of supervisor support for work-family conflict was not a significant predictor of an employee's attitude toward managing work and child care responsibilities. Despite these conflicting results, it is believed that overall there is an inverse relationship between supervisor support and work- family conflict. Thus, the following hypotheses are offered. H 14-NM: Employee perceptions of supervisor support will have a direct effect on perceptions of work 87 productivity. The more one perceives his/her supervisor to be supportive the less likely one is to perceive his/her productivity to be lower due to personal responsibilities. H 14-FM: Employee perceptions of supervisor support will have only indirect effects through stress. Having a mere supportive supervisor will be related to lower levels of work stress. H 14-PM: Employee perceptions of supervisor support will have both direct and indirect effects on attitudes toward family-friendly benefits and perceptions of negative work productivity. E . . J W l-E '] 2 1 Just as a supervisor can be supportive about work-family issues, an organization’s culture can also be supportive. Galinsky (1991) identifies three common stages in the evolution of corporate responses to help employees balance their work and famdly lives. The first is termed "overcoming resistance", a stage in which organizations recognize child care as a legitimate business issue. The second stage is identified as "developing an integrated approach" where the focus on child care becomes broadened to include eldercare. The third stage is "mainstreaming work and family" where work- family concerns are'viewed.as just another facet of diversity. In this final stage companies work.to assure that the programs they have implemented are, in fact, utilized, and the culture change efforts begun are effective in permitting flexibility while retaining or enhancing productivity (Galinsky, 1991). Organizations today vary considerably on how they see and deal with work-family issues. While some organizations are 88 willing to help but are uncertain how to respond to work- family issues, other organizations are unwilling to address work-family issues (Hall & Richter, 1988). Milliken, Dutton, and Beyer (1990) argue that the amount of organizational attention devoted to work-family issues, and the interpretation of current demographic trends (e.g., more working women.with dependents) depends on the characteristics of the organizational context (e.g., managerial values), and on how work-family issues are framed. Kossek; Bass, and DeMarr (1994) found.three components of an overall management orientation or dominant logic of employer-sponsored child care. In other words, the reasons organizations adopt dependent care programe. The first component is "management control" which reflects the notion that providing dependent care assistance helps improve productivity, efficiency, and morale which results in increased management control over workers. The second component is termed."environmental" which basically'amounts to organizations adopting programs to mimic what others who are considered to be "leading edge" are doing. The third component, termed "coercive", reflects the belief that organizations should only adopt dependent care programs when the government or society demands it. Kossek and her colleagues (1994) further found that the extent of adoption.of employer-sponsored child care was positively related to the strength of human resource manager's own orientation toward 89 employer-sponsored child care, and his/her interpretations of favorable executive attitudes toward employer—sponsored child care. How work—family issues are framed by an organization is also an important factor to consider. Hall (1989) asserts that when top management's basic values and experiences are very much those of the traditional one-earner family with an enabling spouse (i.e., wife) at home, and when those values remain unexamined, true progress on the advancement of women and work-family issues will be blocked by top management. For example, when policies such as parental leave and part-time work are perceived as appropriate only for women, their use is accompanied by decreased opportunities for some types of advancement (Voydanoff, 1988a). Thus, the way an organization implements an option is likely to influence its acceptance (Kossek, et al, 1993). Although more employers are offering benefits to assist employees with dependent care responsibilities, when it comes to work-family balance corporate cultures are still largely inflexible (Solomon, 1994). Galinsky (1991) maintains that most company policies have been built on the assumption that there is a spouse at home to manage family responsibilities, and the common notion seems to be that competent workers can handle work-family problems and employees who can’t manage them shouldn't work. Miller, et al (1991) studied the differences between the perceptions of top managers and middle or supervisory managers 90 and found that the top managers, who were typically male and not personally involved in caregiving, believed that job performance was not significantly affected by dependent care responsibilities, and that firms have little responsibility in helping to alleviate the dependent care burdens of employees. Conversely, the middle and supervisory managers, who were predominately female and were personally involved in daily caregiving, believed that job performance was significantly influenced.by dependent care responsibilities, and that firms do have a responsibility to help alleviate employees' dependent care burdens (Miller, et al, 1991). If a person's basic beliefs differ significantly from that of top management, they may be more likely to keep those beliefs to themselves to appear to others to fit int Wharton.and.Erickson (1993) have proposed that the more dissimilar the display norms emphasized at work and home, the higher the level of work-family conflict. Organizations should therefore be able to help reduce the level of work—family conflict experienced by making their culture more family-friendly. The notion of an organization's culture being more family-friendly goes well beyond offering .a few dependent care assistance programs. Greenhaus (1988) argued that organizations must also consider more fundamental changes in their structure, reward systems, and culture to be truly responsive to contemporary work-family issues. While current thought typically favors a greater integration of the 91 work and home domains, Hall and Richter (1988) argue that there is a greater need for separation of work and families, and that organizations need to legitimize the boundaries between work and home. Focusing on daily transitions between work and home, they maintain that an organization can make the transitions easier by things such as not scheduling breakfast meetings, and that boundaries can be legitimized by placing constraints on. work-related. phone calls (e.g., no calls between 5 & 7 pm or after 10 pm) and not expecting weekend travel (Hall & Richter, 1988). While most of the current literature seems to favor either integration or segmentation of the work and family roles, Kossek and DeMarr (1996) argue that the appropriate approach depends on the individual, and propose that work- family conflict will be higher when an individual’s orientation does not fit with that of the organization. Basically, an employee’s orientation toward work and family matters is just another aspect of diversity with some employees preferring to integrate, while00000000 x000 06>060060 .06 4460. 60. 4460. 66. 66.6 00000000600 00600000>00 0000 .66 60. 00. 4400. 66. 66.6 60006066 6060 06006 .66 4460. 4460 4460. 66. 66.6 60006066 6060 00000 .66 60. 4460. 4400. 00. 66.6 66600600 0600000000060 .66 4460. 4460. 60. 66. 60.0 666006 0000 .66 4466. 4400. 4460. 66. 66.0 666006 6060 006006060 .06 60. 60. 4460. 06. 66.6 6000000 000060 0000 .06 00. 00. 460. 66. 60.6 0000006 0060>06006 .00 60. 460. 00. 06. 66.6 660006060 066060 .60 4400. 4460. 4400. 60. 66.6 0000000 60006006 .00 00. 60. 60. 66.06 60.0 0600 6060 06000 .60 460. 00. 60. 06.6 60. 60660 6060 06006 00 060002 .60 4460. 460 4460. 00.6 00.6 6060 06006 00 00606 6000 .60 60. 00 60. 06. 60. 06000 0000 66>00 .60 00. 00. 00. 66. 66. 606000 00 060002 .60 00am. «0mo. i060. om.¢m hm.mm umOU OHMU UHHSU .HH 60. 4400 60. 06. 00.6 6060 00000 \3 000006060066 .00 «6mm. 66mH aimo. Hm. mm. Gmuvaflnu MO HOQESZ .m 4466 4460 4400. 06.06606 06.60066 600000 000066006 .6 -- -- 60. 66. 66. 600000 6 .06000000 02 .0600062 .0 -- -- 4400. 66. 00. 600000 0 .06000000 02 .0600062 .6 -- -- 00. 06. 60. 600000 6 .06000000 .0600062 .6 -- -- 4400. 00. 60. 600000 0 .06000000 .0600062 .6 -- -4 460. 66. mm. 006000 mamswm .m -- 4460. 06. 60. 006060 600006 .6 -- 66. 66.0 060060 .0 6 6 0 .o.6 0662 6006006> .66006006> 006 no 60000606000000000 006 .6000060>0o 00600606 46:60: .0 00069 122 0.66660006060 :0 03056 606 600606 000 6600000060060 600040 A060060-60 00. 60 .000006 - 44 60. 60 .000006 - 4 *mo.: 0*mo.: mo. ¢mo.: «65o. 6M0. «060. 06mo.: *«vo. .5N Ho. 0050. No. 60mo. Ho.: 00. No.: 6*mo. Ho.: .wN 60mm. 4M0. 0¥mo.: No. Ho. Ho. No.: mo. mo.: .mN 06mo.: «00m. No.: 00mm. 0050.: No.: 0tmo.: 04NN. 00.: .¢N «6mo. *«mo. 06OH.: atmo. 0050.: Ho.: 60mo.: iamo. «mo.: .MN 66mo. 660a. 6mo.: «0HH. itmo.: No.: 66mo.: 66HH. oo.: .NN 60mm. 06mm. 00mH.: 00mv. «0mo. 0*mo.: tmo.: «6mm. «6wo. .HN «6mo.: mo.: 00. Ho.: «mo. 6000.: «mo. No.: 6M0. .om Ho.: Ho.: «50. 00.: 60mo. 00.: Ho. oo.: mo. .mH Ho. 660H. No.: aamo. «tmfl. NO. NO.: 60mo. No.: .mH Ho.: 0000. mo. 00mo. 06mm. Ho. «6mo. 60¢o. 00m0. .5H 00mv. 00.: Ho.: 0060.: «0mo. No. Ho. Ho.: No.: .mH 06Hw. iamo.: 6050.: 0050.: «mo. Ho. No. 6000.: Ho.: .mH 66am. 6060.: 00. «0mo.: 6060.: Ho.: Ho. 6460.: Ho.: .¢H 06mm. 6660.: #0. «660.: No.: 6mo.: 40mo. 4mo.: 00. .MH :: 66mo.: mo.: 60mo.: «6mo. Ho. «mo. 6mo.: Ho.: .NH :: «*HN. 60m¢. 66mm. «650.: 60mo.: «05m. Ho.: .HH Aom.v mo.: 66¢H. Ho.: Ho. «tNH. mo.: .OH :: *«MH. «*mH.: No.: 06mm. 00mm. .m :: 660m. «6mo. *«ma. 00OH. .w II. II II II 0N. II II II om II II om I-I nfi .m .N .H NH HH OH m m 5 m m m : : 600000000600 GO006000000 : : .o03000000 : 0 00068 123 60mH.: 660m. 060a. 60HN.: 06am. 0050.: 006m. 6600.: 6060. 6600.: 665m. 6606.: A50.v 60HH.: Am5.v Hm 0N A.6060nuc0u6m Ga 63026 AflmHHMu:NV H0. ma 066d. 660d.: 6660. N0. 66HN.: 600d. 6650.: 00. H0.: mo. H0.: 66MH. 0060.: 600. 0660.: 000. 060N.: 60mm. 00HH.: No. 0000.: 060m. 660H.: 6660. 6650.: 6600. H0. 00mH. 6666. 60H6.: 00mH. 0m0. A00.v 6000.: 0600. H0. A05.v 600.: 600. :: m0. 0H 0H 5H 0H 0H6 60H606 Hem 60fluflafin6fla0u 6£QH¢V .uflcmflm - 44 mo. 66 .uflsmflm - 4 : 0660.: 6600.: N0. .5N H0. 00. H0.: .0N 065m. 665a. 660H. .mN : 6000.: N0.: 6060.: .6N 6050. 6600. H0. .mm 6650. 6000. H0. .NN 660m. 060a. 660a. .HN : 6600.: 600.: H0.: .0N : H0.: 00.: H0. .mH H0. 00.: N0.: .0H 00.: 00.: 00.: .5H 0606. 6606. 666m. .0H :: 600m. 6606. .mH :: 6606. .6H :: .0H .NH .HH .0H .m .0 .5 .0 .m .6 .m .N .H 0H 6H 0H : : 606060600000 coflu6amuuou : : .o0scfiucou : H 0HQ6B 124 A.6060nuc0u6m 6H 63056 066 60H606 600 60flufiafln6fia0u 6£0H40 Aomafimu-mv Ho. 66 .Mficmfim - 44 mo. 64 .uflnmflm - 4 Am0.0 5N 06HH.: 6660.: 0650.: 66Hm.: 0600.: .5N A00.v 66am. 66mm. 00mm. 6660. .0N A50.v 6606. 666a. 66HH. .mN A6m.v 000a. 66mH. .6N A60.v 600m. .MN AHO.V .NN .HN .ON .0H .0H .FH .0H .mH .6H .MH .NH .HH .OH HNMVLDUDI‘GJO‘ 0N mm 60 mm mm 60:0flofiuu0ou coflu6a0uuou : : .00scfiucoo : H 0HQ6B .00006000 068006> 0003 606>0606 6000000800 060000000 . m.m0 m.om b.0m 0006006> 00 0000000 N5m6m.0 006N0.m boomm.m . 0006>0000m ommmm. 600660 0060 00000 mvbmm. 600660 0060 000000000\0000m 6mnom. 600660 0009 m660006 00> 006060 000300000 000 00 >06 0>60 .600008 00000 0660 000 00 ommmm. «0068006 00600 00 .60080600 .600060600 66 0006 68000000 000600 00008 >0 00000000 0000 600mb. m0000 000> 00 600000 006000080 000 0000000 00 000600 0003 00> 0600 0000 06mow. ~00 00 060 00> 600000 000 006 0003 0000 000 00000 00> 0600 00000 0mo0m. m00660006 006 600>000 0000 mwmmm. ~8000 080000>0 000 00000 00> 0600 0000 06 00 000000 0003 600000000000 0000 ...00> 0>60 00000 300 .600008 00000 0660 000 000000 125 mmomm. 0060>00Q06\0006068 000> mm060. 60600mx0 >000008 000000: MO600. 0008800 >0060 000> mmmmb. 000 000> 000600 00006 6000003 mummb. 0600 0003 00000. 600000006 0003 m660006 00> 006060 000300000 000 00 >06 0>60 .600008 00000 0660 000 00 m 000060 N 000000 0 000060 68000 .68000 00600 660000 0060 000000000 006 .66000600 0600000000>60 .660000 0003 00 .606>0600 000060 00 6000600 .N 00069 126 productivity. Convergence of the Lisrel routine was achieved for each of the three models for each of the four dependent variables. A summary of the goodness of fit measures is provided in Table 3. The chi-square test assesses the extent to which the structure of the observed covariances corresponds to those predicted by the model. The null hypothesis is that the covariance structure predicted by the model is not significantly different from the observed covariance structure. Thus, nonsignificant chi-square values indicate the model provides a good fit with the observed data. The chi- square statistic is, however, seriously inflated by larger sample sizes and departures from normality (Baldwin, 1989). Mulaik, James, VanAlstine, Bennett, Lind, and Stilwell (1989) recommend using thems seem. mxmme no upz Ho. 606308 menusuu< .460. :466. hemmm mass Ho. 1 . 4 < .me.- wuemoflmws-oo Ho.- mmmsummllllllu .4me. assess Mo snz .me.- asoa. mnmu "mu“ may uomumsu ucmpcmmwo \, .4eo. 6060 Mo umou .4QH. %Vlllll ¢¢mH.- mumo\3 .umm mo.-. .4wm. smhufihsu no snz ..m~. "mofiumflnmuomumnu sumo UHflSU smo. mEoocfl paonwmso: s:ma.- wo.- Ca N .nu o: .HME Ho. mo.- as H .zu 0c .umE oo. seva. CH N .20 .HmE two. «0.- CH H .n0 .umE «0.- ismH.- cemuma HmCm mo.- ssoa. ucmuma HmCm :mo. "cofiumuzmfimcou Ufiocmmso: ttvo.n HmUCmU tsHH.n .mHnmflum> quUCmeU wzu mm mufluocmm mumu paflno cue: Hobo: coflumfiomz-am«uumm .m shaman 132 omHHmu-ozu .Ho. v d u .. .mo. v m u . iiHm.u .OHDMHZU mmz NO. |_ IflmdflMNMMdded eulilnllnlllllll ..0H.- «mummmmw .00. mmmwwmwmmm mmmuum x003 all...HH. mmHuHmch Hmmumu Ho.- stem. Ho.- Houucou oHDUmnom oo. It symH. munwcmm Al| mmmuumHo show. wxmmh m0 HQZ ssHm. sumo hmon ..0H. HsonoHos0>ms sumzoe mosuHuu< ..00. "uammmmqlammwmmwwmmd _ .00.- mosmsHmmu-ou 00.- \,\7\, mmmsumkllllllu ..00. msmnHw mo 052 .00.- ..0H. sumo . ucoocmamo \/ \fl ..00. 0060 no 0000 .00. ..0H.- 0000\3 .000 ..00.- n ..0N. smHnHHno no hnz 00. .mo. meoocH oHonmmso: No.- 00.- 0H N .00 on .005 m0. mo.- CH H .no 0: .Hme No. ..VH. :H m .20 .005 me. No.- :H H .20 .Hme mo.- ..MH.- cemumm HmCm mo.- .«OH. ucwumm HmCm vo. ”coHumusmHucoo uHonmmso: ..vo.- prcmw ..mo.- ”masquwmuuummmmuuwaamm .mHanum> ucmocmamo mnu mm munmcom oumu uprm :qu Hobo: :oHumeoz-HmHuHmm .m musmHm 133 00H000-030 .H0. v a u .. .mo. v a n . #th.- "OHDMHZU mug limo.“— Admdfimmmmmmde GWIIIIIIIIIIII iamH.- "UHmmmDm tkMH.- HOWH>Hummw mmmuum x003 «III ..HH. 0000Hms0a 000000 ..00. ..0m. H0.- Hohusoo 0Hss0sum H0.- "0 H0 H00 0 s to x0 3 IV ismH. mchHmue >uHHHbe0Hm AIIIIIIIIIIIIII- mmmuumHo H0. 0nmo mo umou 00. 0:0ssosh>sm x003 ..0H. HsuHmoHosu>0m ..0N. 0x000 no 092 00. 000300 00200004 ..00. ..00. 0:000 0500 «0.- I . _ .00.- 00:00H000-oo H0.- mmmuum .000. wumUHm mo unz H0. sioa. mHmU " H MR ucmpcomma \/ 0.00. mumo mo umou N0. AVIIIII- ..0H.- 0000\3 .000 ..m0. . ..0m. coanHso no 002 00. u H H00 0 s 0 Has .00. oeoocH pHonmmsom H0.- 00.- CH N .co 0: .ume 00. 00.- :H H .50 0: .005 m0. ..0H. :H m .:0 .005 00. m0.- cH H .20 .005 m0. ..MH.- COmuwQ Hmcm m0. ..0H. ucmumm HmCm 00. "coHumpsmHucoo pHoswmaom iivo.n HmUEUU ##HH.I .mHannm> ucmpcoqmp 02b mm mchHmph >uHHHnHXmHm utmEcouH>cm x003 cuHs H000: coHumewz-HmHuumm .v onsmHm 134 00H000-030 .H0. v a u .. .mo. v a u . Ix >uH>HuUDUO0m MIIIIIIIIIIIIII- mmm0uwHo x002 Uw>fimUH®m ”0500030 HmcHUDuH pd \z\7\/ iuHm.- "mHDuHDU mm: same“; HmcoHumNHcm 0o lfillllllllllllllll «smH.- "whommzm No. mmmwmmmmmw mmm0uw 0003 All. ..HH. 0000H0c0a 000000 ..00.- ..vm. H0.- Houucou mHspmnom N0.- "moHu Huouomumno x003 tin. 00. 0000 00 umoo 00.- ..0~.- H000moHoso>00 ..0m. 00000 00 022 00.- ..00. ..00. 0:000 0500 00.- 1 fi .00.- 000000000-00 ..00. mw00ummlllllll ..00. mumUHm Mo unz H0. .00H. 0000 “m H mHum om0mnu 0000 ucmpcmmmo \/ t0m0. mumo MO umOU yumo.- \WIIIIII- .40H.- 0000\3 .umm N0. - samm. CMHUHHSU mo 092 mo. ” UH .mo. mEoucH UHocmmso: .000. 00.. cH m .20 o: .005 H0. m0.- cfl H .so 0: .005 m0. 0.0H. CH N .no .005 00.- m0.- :H H .30 .005 m0. ..MH.- com0mm Hmcm 00.- ..0H. ucm0ma HmCm H0.- ucoHum0smHucoo pHonmmso: «ivo.- HUUC00 stmo. .memflum> ucmpcwamp 0:0 mm >uH>Huu:UO0m x003 Um>HmU0wm 20H: Hmpoz coHumemz-H0H00mm .m m0DmHm 135 attitudes toward family-friendly benefits and perceptions of negative work productivity; Limited support.was found for this hypothesis in that single parents and those in dual earner families experienced significantly more, and single persons significantly less, dependent care stress than those without children or those with a nonworking spouse. A direct effect was found on employer—sponsored child care benefits in that single parents and those in dual earner households were also significantly more likely to value such benefits. Household configuration, however, did not have significant direct effects on attitude toward elder care benefits, work environment flexibility training, or perceived work productivity. Income It was predicted that lower household incomes would be related to more favorable attitudes toward family-friendly benefits and higher levels of dependent care stress. Surprisingly, there was a small, but significant, positive relationship between household income and dependent care stress. In other words, those with higher incomes are likely to experience~more, not less, dependent care stress than those with lower incomes. With respect to the direct effects on the dependent variables, as expected income displayed a significant negative effect on attitude toward child care benefits such that those with lower incomes placed more value on child care benefits. Income did not have a significant 136 effect on attitude toward elder care benefits or work environment flexibility training. It did, however, have a significant positive effect on perceived work productivity in that those with higher incomes see themselves as more productive. child cars characteristics: Ntmbsr 9f ghildren Ungsr Ags 14 The partial mediation model predicted that the number of children under age 14 would have both direct and indirect effects on.-attitudes toward. family-friendly' benefits and perceptions of work productivity. The path coefficients do indeed show that the number of children has a significant positive effect on both the level of dependent care stress experienced and the value placed on child care benefits. The number of children does not, however have a direct effect on one’s attitude toward elder care benefits, work environment flexibility training, or perceived productivity. Satis attign with tars The degree to which an employee is satisfied with child care arrangements was predicted to be negatively related to dependent care stress, attitudes toward benefits, and positively related to perceived work productivity. Satisfaction with child care arrangements was significantly negatively related to dependent care stress and the degree to which one values both child and.elder care benefits. There was a significant, albeit small, direct positive effect on 137 attitude toward work environment flexibility training which was opposite of the predicted relationship. Finally, with perceived work productivity as the dependent variable, there were no significant direct effects. ggst Qf child tars The partial-mediation model predicted that those who incurred higher costs for child care would exhibit higher dependent care stress and more favorable attitudes toward benefits, and lower perceived productivity at work. The hypothesis is supported by the path coefficients except when attitude toward work environment flexibility training was the dependent variable when no significant direct effect was found. E1 r re har ri i 5: Number gf Eldsrs The number of elders for whom one is providing care was predicted to be significantly positively related to dependent care stress and attitudes toward benefits, and negatively related to perceived productivity. As with the number of children, there was a significant although smaller, positive relationship with dependent care stress. Surprisingly, there was a significant, but negative, direct effect on attitude toward child and elder care benefits. This indicates that those with fewer elder dependents actually place a higher value on such benefits than those who are caring for more elders. With attitude toward work environment flexibility 138 training and perceived productivity as the dependent variables the number of elders did not display significant direct effects. gg-Rssigsnts With Elgar Dspsndsnts It was hypothesized that those who share a residence with an elder dependent would.experience higher levels of dependent care stress, hold more favorable attitudes toward family- friendly benefits, and perceive themselves to be less productive than those who do not share a household. This hypothesis did not receive support. While there was a significant, but small relationship with dependent care stress it was negative instead of positive. Also contrary to the hypothesis was a small positive direct effect on perceived work productivity. There were no significant direct effects when attitude toward child care or elder care benefits, or work environment flexibility training was the dependent variable. Tims Spsnt Qn Ears Similarly, the amount of time a caregiver spends providing care for an elder dependent did not prove to be a strong predictor. As expected, the more time one spends providing care the more dependent care stress experienced. While this relationship was significant, the path coefficient of .06 was quite small. While the amount of time spent providing care was predicted to be positively related to one' s attitude toward family-friendly benefits and negatively 139 related to perceived productivity, it did not exhibit significant direct effects in any of the analyses. Nu r f T sk Compared to co-residence with an elder dependent and the amount of time spent providing care, the number of tasks with which a caregiver is providing assistance had a much larger (.20) effect on the amount of dependent care stress experienced. Also consistent with the hypothesis was a positive direct effect on the value one places on elder care benefits. However, therdirect effects on-attitude toward child care benefits, work environment flexibility training, and perceived work productivity were not significant. Cgst 9f Elder gare Unlike the cost of child care, the cost of elder care was not significantly related to dependent care stress. Further, there ‘were no significant direct effects on. any' of the dependent variables. The fact that all path coefficients for this variable were either zero or very near zero suggests that the cost of elder care has no predictive value in this type of model. Work gharattsristigs: h l n r l The partial-mediation model predicted that the amount of control one has over his/her work schedule would be negatively related to work stress and positively related to perceived productivityu However, none: of the relationships, either 140 direct or indirect, were significant for this variable. As with the cost of elder care, all path coefficients for this variable were either zero or very near zero. The results could be due to the way the variable was operationalized. Given that this variable was based on the response to a single question, it may be premature to dismiss the amount of control one has over his/her work schedule as an unimportant variable. garssr Banaltiss As hypothesized, employee perceptions of career penalties was significantly positively related to work stress. While it did.not have a direct effect on one's attitude toward child or elder care benefits, there was a significant positive direct effect on one's attitude toward work environment flexibility training. Also consistent with the hypothesis, was a significant negative direct effect when perceived work productivity was used as the dependent variable. rvi r u ort The degree to which an employee believes his/her supervisor’ is supportive ‘was predicted. to lbe :negatively related to the level of work stress experienced, and positively related to perceptions of work productivity. Consistent with the hypothesis, the path coefficients indicate that the less supportive one’s supervisor'thewmore work stress will be experienced. A significant negative direct effect was also found with attitude toward work environment flexibility training as the dependent variable. No direct effects were 141 found, however, for attitude toward child or elder care benefits, or perceptions of work productivity. Qrganizatignal Work-Family Qtlturs Similarly, organizational work—family culture displayed a significant negative effect on work stress, which indicates that the less supportive the culture, the more stress will be experienced. A significant negative direct effect was found when attitude toward work environment flexibility training was theedependent variable. This suggests that the less supportive the culture, the more employees see value in such training. A small, but significant, positive direct effect was found for perceived work productivity, suggesting that employees see themselves as more productive when.they'believe the culture is supportive. Organizational work-family culture did not exhibit significant direct effects on employee attitudes toward either child or elder care benefits. Dapsndent Care Strsss, Work Strsss, and Psyghglggigal Di r s. As noted. earlier, employee jperceptions of work. and dependent care stress, and psychological distress partially mediate the relationship between domain specific characteristics and all of the outcome variables. As hypothesized, the path coefficients indicate that there is a significant positive reciprocal relationship between.work and dependent care stress. Dependent care stress has a larger effect on work stress (.16) than work stress has on dependent 142 care stress (.06). Also as expected, there is a positive relationship between both work and dependent care stress and psychological distress, although the relationship is stronger for work stress (.54) than dependent care stress (.10). Attitdds Toward Bsnefits Employee attitudes toward child and elder care benefits and work environment flexibility training are influenced by work and nonwork characteristics both directly, and indirectly through stress as predicted by the partial-mediation model. With attitude toward child care benefits as the dependent variable, there were significant positive effects from psychological distress (.10), single parents (.05), those in a dual earner family (.06), number of children (.23), and cost of child care (.18). There were significant negative effects for household income (-.13), and.gender (-.11) such.that women hold more favorable attitudes toward employer-sponsored child care benefits than men. When attitude toward elder care benefits was used as the dependent variable, there were significant positive effects from psychological distress (.10), cost of child care (.03)- the number of tasks for which one is providing assistance to the elder(s) (.21), and the level of supervisor support (.03). Significant negative effects were found for satisfaction.with child care (-.07), number of elders (-.05), and gender (-.08) which again indicates that women are more likely to value elder care benefits. 143 Lastly, with attitude toward work environment flexibility training as the dependent variable, psychological distress (.14), satisfaction with child care (.05), and career penalties (.05) all displayed significant positive effects. Conversely, significant negative effects were found for both supervisor support (-.13) and gender (-.11). Partsivsd Wdrk Prdduttivity As predicted by the partial-mediation model, employee perceptions of work productivity are predicted by psychological distress as well as work and non-work characteristics. With a path coefficient of -.27, psychological distress appears to have the strongest influence, although. several other' work. and. nonwork characteristics display significant direct effects. Specifically, women, and those with higher household incomes, lower child care costs, and. who live with their elder dependents perceive themselves as more productive. Conversely, those who believe there are strong career penalties or a less supportive organizational work-family culture perceive themselves as being less productive. CHAPTER SIX DISCUSSION This dissertation investigated the relationship between employees' work and non-work characteristics, and attitudes toward employer-sponsored family-friendly benefits and perceptions of work productivity in an effort to provide a link between the study of work and family stress and dependent care. There were two basic objectives. The first was to determine if the relationship was direct, or fully- or partially-mediated by the amount of stress experienced, and to assess the impact of individual variables. The second was to provide insight for practitioners in the development of programs to help employees balance their work and personal lives. With those objectives in mind, this chapter will discuss the results of the analyses focusing on the study’s support for spillover theory and the importance of measuring both the direct and indirect effects of individual variables. Implications for managers will then be presented, as well as limitations of the study and suggestions for future research. r f il v Th Path analysis using Lisrel VII was used to test no- mediation, full-mediation, and partial-mediation models. Recall that the no-mediation model is based on the segmentation approach which holds that the work and home domains are separate or segmented, and conditions in one 144 145 domain do not affect outcomes in the other. In contrast, both the full- and partial-mediation models are based on the spillover approach which holds that conditions in the work (home) domain spillover to affect outcomes in the home (work) domains Each model was tested separately with.attitudes toward child care benefits, elder care benefits, work environment flexibility training, and.perceptions of work.productivity as the dependent variable. The strongest support was found for the partial-mediation model for each.of the four dependent variables. These findings discount the argument that home and work.domains are separate, and supports Lambert's (1990) contention that given today's changing gender roles, segmentation is unlikely to occur naturally. Indeed the results of this study showed that the no-mediation model provided the worst fit and explained the least amount of variance for each of the dependent variables. Thus, in order to effectively analyze employee needs future research must include stress as a mediator of the relationship between nonwork and work characteristics and employee attitudes toward benefits. The results of this study also support Greenhaus and Beutell's (1985) contention that there is a positive reciprocal relationship between family stress and work stress, and is consistent with other research findings (cf. Prone, et a1, 1992; Gutek, et al, 1991; Williams & Alliger, 1994) . Given that stress spills over from the family domain to the work 146 domain, it seems that employers do have a vested interest in helping' employees cope with. their' dependent care needs. Similarly, since stress from the work domain spills over to the family domain, employees need to try to manage that spillover if they wish to protect their families from negative consequences such as lower family satisfaction (Rice, et al, 1992), marital satisfaction (Aryee, 1992), and quality of family life (Higgins, et al, 1992), and higher levels of drinking (Prone, et al, 1993) and depression (Frone, et al, 1991). Both. dependent care stress and. work stress also displayed significant positive effects on psychological distress, although work stress had a much larger impact. Psychological distress in turn influences employee attitudes toward family-friendly benefits and perceived work productivity such that employees who are experiencing higher levels of psychological distress hold more favorable attitudes toward family-friendly benefits and.perceive themselves to be less productive. It appears that as employees become more distressed the more likely they are to look to their employer for support. Given that distressed employees also see themselves as less productive they may believe it is in the employer's best interest to help employees cope with stressful conditions. It may also be that employees see their employers as having the resources available to provide services. These findings highlight the importance of including stress variables in research on employee needs for family-friendly 147 benefits, in addition to the obvious practical implications for employers. The importance of the 'mediating role of stress is emphasized by the large difference in the amount of variance explained between the partial-mediation model and the no- mediation model. The amount of variance explained by the partial-mediation model was 60.5%, 56.0%, 54.5%, and 53.8%, for attitudes toward.child care benefits, elder care benefits, work environment flexibility training, and perceived work productivity, respectively. In contrast, the amount of variance explained by the no-mediation model was only 20.4% for attitudes toward child care benefits, 9.2% for attitudes toward elder care benefits, 7.5% for attitudes toward work environment flexibility training, and 6.6% for perceived.work productivity. Clearly, demographic characteristics do a better job of predicting attitudes toward child care benefits than they do as predictors of the other dependent variables. This may explain why some research has found significant differences in employee preferences for employer-sponsored child care assistance based on demographic characteristics (cf. Kossek, 1990). While the full-mediation.model did not hold up as well as the partial-mediation model, it too explained substantially more of the variance (52.6% for each of the dependent variables) than the no—mediation model. The fact that the partial-mediation model is robust 148 across the four dependent variables used in this study is significant in several respects. First, it highlights the stability of the predictive value of the model with the coefficient.of determination ranging only from.60.5% for child care benefits to 53.8% for perceived work productivity. It also suggests that the model could.be used with.a wide variety of dependent variables, which means that there is not a need to develop a unique model for every individual benefit program or other outcome variable of interest. While some of the individual variables were not significant in one or more of the analyses, or had very small effects indicating that in the future some minor revisions could be made to the model, overall the model held up well. With that in mind I turn the focus to the importance of measuring the indirect as well as direct effects of individual variables. Ths Impdrtante df Direct and Indirstt Effstts The independent variables used in this study encompass a variety of work and nonwork characteristics of employees. While most of these variables have been used to some extent in other studies, no one study as yet has used such a wide variety. A summary of the direct and indirect effects of the independent variables is shown in Table 4. The results show that work and non-work characteristics exhibit both indirect effects through stress, as well as direct effects on each of the dependent variables. The implication is that assessing employees' needs for employer-sponsored family-friendly 149 Table 4. Ddrect Versus Indirect Effects of all Independent Variables In n n V ri 1 Family Characteristics: Gender Household configuration Household income Child Care Characteristics: Number of children Satisfaction w/ care Cost of child care Elder Care Characteristics: Number of elders Co-residence Time spent on care Number of tasks Cost of elder care Work Characteristics: Schedule control Career penalties Supervisor Support Organizational W/F Culture Child Care Ben. D/I D/I D/I D/I I D/I D/I I I I none none Elder Care B n. D/I I I D/I D/I D/I D/I none none D/I D/I I HHHH none none D/I D/I D/I ------- Dependent Variable ------- Work Perceived Flex. Tr in. Pr u tivit Work D/I I D/I none D/I D/I 150 benefit programs is more complex than simply using demographic characteristics to predict needs. Although simple models based on employees’ demographic characteristics are more convenient to use than complex, integrated models which also require more sophisticated 'methods of analysis, the information they provide is not as robust. All of the family characteristics had both direct and indirect effects on at least one of the dependent variables with gender displaying both types of effects for each of the dependent variables. Gender is a commonly used variable in both the work-family conflict and.dependent care literatures. Consistent with most other research, this study found that women hold ‘more favorable attitudes toward all of the employer-sponsored benefit programs and experience more dependent care stress than men. The direct effects were, however; much.stronger, suggesting that in today’s environment men and women are experiencing 'more similar levels of dependent care stress. Surprisingly, the study found that women actually perceive themselves as being more productive than men do. When considering both work and nonwork responsibilities women are often responsible for more in total than men. If a woman sees herself as being productive over a multitude of nonwork responsibilities, that sense of productivity may spillover to the work environment. This is consistent with Barnett's (1994) finding that positive experiences in nonwork roles can buffer the effects of job 151 experiences. Employees’ household configuration for the most part had only indirect effects highlighting the need to include stress as a moderator. Indeed the only direct effects that were found were on employee attitudes toward child care benefits, such that single parents and parents in a dual income household were more likely to favor such benefits. Surprisingly, there were no significant negative direct effects for other household configurations discounting the popular notion that those who do not benefit directly from work-family programs (e.g. employees who do not have children) resent them. Also surprising was that parents in dual-earner households experienced slightly more dependent care stress than single parents. In addition to a small, yet positive indirect effect, household income displayed direct effects on attitudes toward child care benefits and perceived work productivity. Those with lower household incomes were more likely to favor child care benefits suggesting they have a special need for child care assistance, while those with higher incomes perceived themselves as more productive. It may be that those in higher income brackets are more likely to be salaried and to work additional hours for which they are not compensated, which may lead them to believe that they are more productive. Conversely, it may be that those with higher incomes have a need to justify their income, and thus convince themselves 152 that they are more productive which is consistent with equity theory (Adams, 1965). All of the child care characteristics also displayed both types of effects for at least one of the dependent variables. Having responsibility for more children under age 14, being less satisfied with child care arrangements, and incurring larger costs for child care all contributed to higher levels of dependent care stress. The fact that the number of children had the strongest influence is consistent with the notion that larger families place more demands on a person's time (Cartwright, 1978; Keith & Schafer, 1980), and result in more strain (Katz & Piotrkowski, 1983). Of the elder care characteristics, only the cost of elder care did not have either direct or indirect effects on any of the dependent variables. This is surprising since one study (Scharlach.& Boyd, 1989) found that 45% of employed caregivers were providing from. $20 to $3,000 per ‘month in. direct financial assistance, and.over half of all caregivers reported some degree of financial strain. One explanation for the nonsignificant results is that this study included all employees and not just those with eldercare responsibilities. Thus, the effects would.be diluted..Among those providing care for an elder dependent, the weekly out of pocket costs ranged from $1.00 to $999.00, and 35.5% of caregivers reported that it was either difficult or very difficult to pay those expenses. 153 In contrast, the number of elders for whom one is providing care, co-residence with an elder dependent, the amount of time spent on care, and the number of tasks with which one is assisting all had indirect effects through stress. By far the strongest of these effects was for the number of tasks, which indicates that assisting with a wide variety of tasks is far more stressful than spending more time on fewer tasks or even providing care for more elders. This may be a result of simply having to keep track of and juggle more things. Surprisingly, co-residence with an elder dependent resulted in slightly less stress. This is contrary to previous research which has shown co-residence to be a strong predictor of strain (Brody, 1985; George & Gwyther, 1986) . It may be that when a caregiver lives with their elder dependent(s) s/he has greater access to the elder and therefore is more aware of the elder’s condition, reducing the amount of anxiety over how well the elder is doing. By far the strongest of the direct effects was the positive effect of the number of tasks with which one is assisting on attitude toward elder care benefits. This is not surprising given the large influence of this variable on dependent care stress. The number of elders for whom one is providing care displayed a direct effect on attitudes toward child and elder care benefits, while co-residence with an elder dependent had a direct effect on perceived work productivity. Surprisingly, providing care for more elders 154 resulted in slightly less favorable attitudes toward child and elder care benefits, and co-residence resulted in slightly higher'perceived.productivityu It may be that those caring for more elders may have been providing care for a longer period of time and be more familiar with outside services. Since the most common forms of employer-sponsored eldercare assistance are information and referral services, employees who care for more elders may see themselves as already having that knowledge and thus not needing their employer’s assistance. That caregivers who live with their elder dependents perceive themselves to be slightly more productive may be related to the fact that with co-residence there is only one household for which to care. In other words, caregivers may think that their caregiving responsibilities would require even more time if they had a second household to maintain. Of the remaining independent variables, only the amount of control one has over his/her work schedule did not display either direct or indirect effects. Given that previous research has found that lack of flexibility in, and control over, one's work schedule induces work-family conflict (Greenhaus & Beutell, 1985; Greenhaus, et al, 1989) and increases stress (Friedman, 1990; Williams & Alliger, 1994), it seems that the lack; of significant effects for’ this variable are likely due to another cause. The probable cause is that this variable was measured with a single survey item which can lead to unreliable results. Thus, a psychometrically 155 sound scale should be developed and tested before any conclusions can be drawn. In contrast, career penalties, supervisor support, and organizational work-family culture each provided strong indirect effects through stress. The finding that the more one believes s/he will be penalized for using flexible work arrangements increases work stress, is consistent with Greenhaus and Beutell's (1985) proposition that work-family conflict will be strongest when there are negative sanctions for noncompliance with role expectations. That less supportive supervisors and cultures result in higher levels of stress is consistent with much of the research on social support (cf. Zarit, et al, Burke, 1988). While career' jpenalties, supervisor‘ support, and organizational work-family culture did not have direct effects on attitude toward child care benefits, each had direct effects on.the other dependent variables. Career penalties had direct effects on attitude toward work environment flexibility training and perceived work productivity such that respondents who believed that they would be penalized for using flexible work arrangements favored flexibility training for supervisors and managers and saw themselves as less productive. Supervisor support had a positive direct effect on the value placed on work environment flexibility training suggesting that when a supervisor is not viewed as being supportive, their employees would like to have them.trained to 156 become more flexible. While a small positive direct effect on attitude toward elder care benefits such that a more supportive supervisor was related to a more favorable attitude toward elder care assistance, the effect is too small to draw any firm conclusions. Somewhat surprising was the lack of a direct effect on perceived work productivity indicating that a supportive supervisor does not necessarily lead to higher productivity. This is similar to a finding of the Goff, et al (1990) study, that supportive supervision was weakly related to higher absenteeism. As with supervisor support, a less supportive environment was directly associated with more favorable attitudes toward work environment flexibility training. However, unlike supervisor support, a supportive work-family culture did have a small direct positive effect on perceived work.productivity suggesting that employers may realize productivity gains when they actively promote a work environment that is more supportive of employees’ family responsibilities. In sum, each category of independent variables displayed both direct and indirect effects that were significant for each of the dependent variables. Clearly, the common.practice of relying strictly on direct effects to predict employee attitudes toward family—friendly benefits overlooks the important indirect influence of the independent variables through stress. Similarly, if research were to rely strictly on indirect effects, as is the case with the full-mediation 157 model, valuable information is also lost. Thus, to get a complete picture research must allow for both direct and indirect effects. L'm' ion f h While this study makes a number of important contributions, there are weaknesses in virtually all research designs and methods. One criticism frequently cited by editors (Campbell, 1982) and reviewers (Spector, 1994) alike is that of using a self-report questionnaire to measure all the variables in a cross-sectional study as is the case in this study. It might even be argued that monomethod bias would predispose this study to finding the strongest support for the partial-mediation model in that method bias inflates correlations between all variables and the partial-mediation model relies on more of the variables being significantly correlated. Thus, some might be inclined to completely dismiss the results of this study as simply an artifact of the method used, however, more thoughtful consideration is needed for several reasons. First, is that the endogenous variables are perceptual (e.g., stress, perceived work productivity) or attitudinal (e.g., attitudes toward benefits). Most would agree that the use of self-reports is appropriate when the constructs involved are attitudinal or perceptual (Schmitt, 1994). While one might argue that work productivity is not a perceptual construct, recall that this variable was operationalized with 158 scale items that asked respondents t1) assess their productivity as a result of their personal responsibilities. While a more objective global measure of work productivity might be obtained from.organizational performance records, it would not be possible to determine whether any given level of productivity was a result of an employee’s personal responsibilities or some other cause(s). Several of the independent variables (e.g., satisfaction with child care, supervisor support, organizational work- family culture) are also perceptual in nature. Other independent variables included in this study (e.g., family, child care, and elder care characteristics), although also based on self-reports, were demographic or factual. A meta- analytic study of percept-percept inflation in published research found that self-report ‘methods are unlikely to increase the correlation between two variables when at least one of the variables is based on demographic data (Crampton & Wagner, 1994). The same study also considered the extent to whidh correlations among various nondemographic constructs were inflated. Specifically as it related to the present study, no significant evidence of inflation was found in correlations between stress or anxiety, performance, and organizational culture (Crampton & Wagner, 1994). It should also be noted that the survey was completely anonymous which reduces the potential for social desirability bias since individual responses could not be linked to any of the 159 participants. Finally, with respect to minimizing the potential effects of method bias, the questions concerning the respondents' child and elder care responsibilities were positioned at the end of the questionnaire to avoid a priming effect on the dependent variables. A weakness in this study can be found in the schedule control and perceived work productivity variables. Specifically, that schedule control was measured with a single item measure and that the productivity measure had relatively weak reliability (alpha = .62). Future research in this area should ensure that psychometrically sound multi-item measures are used. Instead of using the number of minor or elder dependents, future studies might employ a single measure of family responsibilities such as that recently developed by Rothausen (1995). Further, some of the variables in the model did.not display significant effects. Whilewmany models used in this type of research do include variables that are not statistically significant, the results suggest that in the future the model could be refined. A final weakness in this study may be that dependent care stress, work stress, and psychological distress were used instead of measures of family conflict, work conflict, and work-family conflict. Use of the more traditional measures of work—family conflict would better assess the directionality of the relationships in that such measures ask address more specifically the impact of home (work) responsibilities on 160 work (home). A drawback in those measures, however, is that they typically focus on the negative effects of spillover (e.g. the extent to which family responsibilities interfere with work or work responsibilities interfere with family) and ignore potential positive spillover effects. Implications fdr Employers There is an increasing focus today on the practical significance of research and the results of this study do have a number of practical implications for employers. First, the study shows that the amount of work and dependent care stress experienced by employees does have an effect on how employees view employer-sponsored benefits. While some employers have been reluctant to get involved in employees’ personal lives, the boundaries between the work and home domains are becoming increasingly blurred. By accepting this and providing benefits that are truly valued and needed by employees, employers will obtain. greater “value for' their .benefit dollars. Prudent allocation of benefit dollars is especially important in today's cost conscious business environment. While some have tried to achieve this by using demographic characteristics as predictors, this study shows that it is not that simple and that a more complex model such as the one used in this study is needed. Similarly, the amount of work and dependent care stress experienced affects how employees view their productivity at work such that higher levels of stress were associated with 161 lower productivity. Thus, as stress decreases employee productivity is likely to increase. While this study used employee perceptions of productivity as opposed to more objective measures, given the strength of the relationship, other“productivity“measures‘would.likely'show'similar'results. As with insuring that benefit dollars are spent wisely, increasing employee productivity is also a major concern for employers today. In testing the model with several dependent variables this study demonstrated that the model was robust across the four dependent variables suggesting that the model could also be used for other dependent variables. There is value for practitioners and researchers alike in having a generic model that can be used for a number of dependent variables instead of developing new models for every type of benefit program or outcome variable of interest. The use of a single model makes it easier for employers to survey employees about their benefit needs without the need to construct separate surveys. It also allows for a direct comparison of the need for various benefit programs. The results also show that certain groups of employees may be more prone to stress than others. In particular, employees with many children or who are assisting an elder dependent with.a number of tasks, as well as those who believe their supervisor or the organizational culture is not supportive. This is important since the stress experienced by 162 employees is likely to increase as more employees assume elder care responsibilities and organizations continue to expect more from employees to remain competitive. While employers typically do not and should not attempt to control an employee’s family situation, they can and some would argue should, influence the level of support employees receive in the workplace. By training and encouraging supervisors to be more flexible, and promoting a more supportive work environment organizations can help minimize the stress levels of their employees. This study also suggests that employers need to reconsider the allocation of their investment in benefit programs among various employee groups. In most organizations, higher income employees typically have more attractive benefit packages. Yet this study showed in the case of child care benefits, employees with lower incomes place a higher value on such benefits. The question for employers is how much value is there in providing additional benefits for higher income employees if those benefits are not as highly valued? It seems that employers would realize a greater return on their benefit dollars by offering such benefits to those employees who truly need them. I pli . ion_ for Th-o .n- ---_tions for ' r- '-,-.r h One of the key themes in the social sciences is the importance of individual differences in understanding various phenomenon. Ironically, the dependent care literature has thus 163 far implicitly dismissed individual differences by focusing almost exclusively on demographic characteristics to predict needs for benefits and ignoring what has been learned by work- family conflict and stress researchers. The support this study found for the partialemediation model indicates that individual differences in the way people experience work and dependent care stress do influence their attitudes toward employer-sponsored benefits. While some employees may find having children to be very stressful and have a greater need for employer assistance, others may find parenthood to be a source of rewards (Cooke & Rousseau, 1984) and be able to effectively shoulder their dependent care responsibilities without employer-sponsored dependent care assistance programs . Clearly, dependent care research can be enhanced by integrating concepts from the work-family conflict and stress literatures. At the same time work-family conflict and stress researchers have not yet extended their work to assess the effects of providing care for elder dependents or to investigate the impact of work-family conflict and stress on the need for employer-sponsored family-friendly benefits. As this study has shown the influence of work and family responsibilities extends beyond psychological distress and ‘more research is needed to see just how far these effects go. Similarly, researchers should investigate the impact of various benefit programs on stress and work-family conflict. 164 If employers adopt family-friendly benefit programs that are truly needed by employees it may even be that the use of such benefits benefit programs will help to ameliorate the stress experienced by employees. Although the work-family conflict and dependent care literatures have evolved separately, this study highlights the importance of integrating them. Clearly, when the two literatures begin to speak more clearly to one another each will be enhanced. Although, stress and employee attitudes toward family-friendly benefits appear to be an important link between the work-family conflict and dependent care literatures, there may also be others. Thus, while this study makes an important first step in the integration of the two literatures, much remains to be done. Researchers who are interested in extending this line of work should consider using subjects from other industries. While this study involved a very large heterogeneous population, it was drawn from employees of a single large corporation in the financial services industry. Although the participants were from a variety of job classifications and geographic locations, there may be some questions of generalizabilty to employed persons in general. Based on the comparison of the respondents to all employed persons discussed in the methods chapter, these results are more likely representative of peOple who are above the national averages in terms of socioeconomic status. 165 There are also several ways in which the model might be broadened in the future. Using a measure of total nonwork stress instead of dependent care stress would allow for the inclusion of other independent variables such as time commitments related to community, religious, or social organizations, or even physical fitness. It may also be beneficial to include a measure of peer support that would reflect the degree to which a person has co-workers who are willing to help one another as a need arises. By using common ‘measures of both child.and.elder care responsibilities, future studies would also be able to provide a direct comparison of the impact of elder versus care responsibilities on employed caregivers. Operationalizing child and elder care expenses as a proportion of household income would also provide more insight into the relative impact of these costs on individuals than using an actual dollar amount as was done in this study. Future research could also test the model with a wider variety of dependent variables to examine the extent to which the basic model is generic. Some of the possibilities include: absence, tardiness, willingness to travel or relocate, turnover intentions, and actual productivity, as well as measures of family functioning and physiological well-being. Finally, as with any emerging area of study, longitudinal research is needed to assess the stability of the observed relationships over time. In conclusion, while there is still much to be learned 166 about the work-family nexus in general, and in particular the mediating role of stress, this study makes several contributions. First, it provides a link between the more theoretical work-family conflict and stress literatures and the more practitioner—oriented dependent care literatures. Second, this study tested and assessed the impact of a wide variety of variables believed to influence stress and attitudes toward.employer-sponsored.family-friendly'benefits. Finally, this study provides insight for employers WhO‘WiSh to develop programs to help employees balance their work and personal responsibilities while obtaining the largest return on their investment in employee benefits. References Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.). Advances i Eapsrimsntal gdtial gsythdldgy; 267-299. New York: Academic Press Aldous, S. (1969). Occupational characteristics and male role performance in the family. dournal df Marriags aud ths Family, 31, 701-712. Allen, V. L., Wilder, D. A., & Atkinson, M. (1983). Multiple group membership and social identity. In T. R. Sarbin & Kr E. Scheibe (Eds.). Studiss in Sdcial Idantity; 92-115. New York: Praeger. Anastas, J. W., Gibeau, J. L., & Larson, P. J. (1990). Working families and eldercare: A national perspective in an aging America. Sdtial Wdrk, 35 (5), 405-411. Aryee, S. (1992). Antecedents and outcomes of work-family conflict among married professional women: Evidence from Singapore. Human Rslatidns, 45 (8), 813-837. Ashforth, B. E., & Mael, F. (1989). Social identity theory and the organization. f M n m n R v' w, 14, 20-39. Azarnoff, R. S., & Scharlach, A. E. (1988). Can employees carry the eldercare burden? Psrsdnnsl ddurnal, 67 (9), 60-66. Bacharach, S. B., Bamberger, P., & Conley, S. (1991). Work- home conflict among nurses and engineers: Mediating the impact of role stress on burnout and satisfaction at work. dournal df Qrganizatidnal Bshavidr, 12, 39-53. Baldwin, B. (1989). A primer in the use and interpretation of structural equation models. M r m n n Ev i n in gdunssling and stsldpmsnt, 22, 100-112. Barnes, C. L., Given, B. A», & Given, C. W. (1992). Caregivers of elderly relatives: Spouses and adult children. Hsalth ems-2915.11093- 17 (4). 282-289. Barnett, R. C. (1994). Home-to-work spillover revisited: A study of full-time employed women in dual-earner couples. ddurnal df Martiags and ths Family, 56, 647-656. Barnett, R. C., s. Baruch, G. K. (1987). Determinants of father’s participation in family work. ddutual df Emmi-10112.49. 29-40. 167 168 Belsky, J., Perry-Jenkins, M., & Crouter,.A. (1985). The work- family interface and marital change across the transition to parenthood. ddurnal of Family Issues, 6, 205-220. Brett, J. M., & Yogev, S. (1988). Restructuring work for family: How dMal-earner couples with children manage. ddurnal of Sotial Bshavior and Psrsonality, 3 (4), 159- 174 . Brody, E. M. (1981). Women in the middle and family help to older people. Ths Gardntdlogist, 21, 471-480. Brody, E. M. (1985). Parent care as normative family stress. The Gardntdldgist, 25 (1), 19-29. Brody, E. M., Kleban, M. H., Johnsen, P. T., Hoffman, C., & Schoonover, C. B. (1987). Work status and parent care: A comparison of four groups of women. W, 27 (2) I 201-208 0 Broman, C. L. (1991). Gender, work-family roles, and psychological well-being of blacks. ddutual__df Marriage and the Family, 53, 509-520. Brooke, Jr., P. P. & Price, J. L. (1989). The determinants of employees absenteeism: An empirical test of a causal model. ddurnal df thupatidnal Psychdldgy, 62, 1-19. Bunting, S. M. (1989). Stress on caregivers of the elderly. Advangss in Nutsiug Sgisnds, 11 (2), 63-73. Burden, D., & Googins, J. (1986). Boston University balancing job and homelife study. Boston: Boston University School of Social Work. Bureau of National Affairs, (1989). Bulletin td Mauagsusnt, 40, (7), 1-4. Burke, R. J. (1988). Some antecedents and consequences of work-family conflict. Journal of Sotial Bshavidt and Parsdnality, 3 (4), 287-302. Burke, R. J., Weir, T., & Duwors, R. E. (1979). Type A behavior of administrators and wives’ reports of marital satisfaction and well-being. Jo rnl A 1' 252509199X, 64, 57-65. 1Burke, R. J., Weir, T., & Duwors, R. E. (1980a). Perceived type A behavior of husbands and wives' reports of marital satisfaction and well-being. ddurnal df thupatidna; WI 11 139-150: 169 Burke, R. J., Weir, T., & Duwors, R. E. (1980b) . Work demands on administrators and spouse well-being. Human Rslatidns, 33, 253-278. Burud, S., Aschbacher, P., & McCroskey, J. (1984). Empldysr— suppdttsd thild tars - investing in human rssdurtss. Dover, Mass.: Auburn House Publishing Co. Campbell, J. P. (1982). Editorial: Some remarks from the outgoing editor. ddurnal df Applied Psythdldgy, 67, 691- 700. Cartwright, L. K. (1978) . Career satisfaction and role harmony in a sample of young women physicians. ddurual pf Vddatidnal Bahavidr, 12, 184-196. Cooke, R., 8: Rousseau, D. (1984). Stress and strain from family roles and work-role expectations. W Appligd Psyghgldgy, 69, 252-260. Crampton, S. M., & Wagner, J. A., III (1994) . Percept-percept inflation in microorganizational research: An investigation of prevalence and effect. ddurnal df Appligd Psyghdlggy, 79, 67-76. Creedon, Michael A. (1988). The Corporate Response to the Working Caregiver, Aging, 358, 16—19+. Crouter, A. C. (1984). Spillover from family to work: The neglected side of the work-family interface. Human Relations. 37. 425-442. Curry, J. P., Wakefield, D. 8., Prince, J. P., Mueller, C. W., & McCloskey, J. C. (1985) . Determinants of turnover among nursing department personnel. Rsssardh in Nursing and Hgalth, 8, 397-411 Deaux, K. (1985) . Sex and gender. Annual Ravisw df Psyghdldgy, 36, 49-81. Denton, K., Love, L. T., & Slate, R. (1990). Eldercare in the '903: Employee responsibility, employer challenge. Familiss in Sdtisty, June 1990, 349-359. Duxbury, L. E. & Higgins, C. A. (1991) . Gender differences in work-family conflict. Jdurnal df Applisd Psytudlogy, 76 (1): 60-74. Edwards, J. R., & Rothbard, N. P. (1995) . An integrative model of stress, coping, and well-being within and between work and family domains. Paper presented at the 55th annual meeting of the Academy of Management, Vancouver, British 170 Columbia, August 1995. Falkenberg, L. , & Monachello, M. (1990) . Dual—career and dual- income families: Do they haverdifferent needs? dournal of Business Ethics, 9, 339-351. Finley, N. J. (1989). Theories of family labor as applied to gender differences in caregiving for elderly parents. ddurnal df Marriags and Family, 51, 79-86. Friedman, D. E. (1986). Eldercare: The employee benefit of the 19908? Adrdss the Board, 23 (6), 45-51. Friedman, D. E. (1990). Work and family: The new strategic plan. Human Rssourts Planning, 13 (2), 79-89. Frone, M. R., Russell, M., & Cooper, M. L. (1991). Relationship of work and family stressors to psychological distress: The independent moderating influence of social support, mastery, active coping, and self-focused attention. Journal of Sgtial Bahavidr and Psrsonality, 6 (7), 227-250. Frone, M. R., Russell, M., & Cooper, M. L. (1992). Antecedents and outcomes of work-family conflict: Testing a model of the work-family interface. ddurnal of Applisd Psyghdlggy, 77 (1), 65-78. Frone, M. R., Russell, M., & Cooper, M. L. (1993). Relationship of work-family conflict, gender, and alcohol expectancies to alcohol use/abuse. ddutnal df Qrganizatidnal Bshavidr, 14, 545-558. Frone, M. R., Russell, M., & Cooper, M. L. (1994). Relationship between job and family satisfaction: Causal or noncausal variation? Journal of Managsmsnt, 20 (3), 565-579. Galinsky, E. (1988). The impact of supervisors’ attitudes and company culture on work/ family adjustment. Paper presented at the Annual Convention of the American Psychological Association, Atlanta, Ga. Galinsky, E. (1989). Child care and corporate productivity. Unpublished paper for the child care action campaign. Trenton: NJ: Resources for Child Care Management. Galinsky, E. (1991). The globa1.perspective and.future trends: What employers are doing. Book chapter in Emldyst- Spgnsdtsd thld Cars ; A Stratagit Wdrkplats Issue (E. E. Kossek, Ed.) . Proceedings of the First Annual Governor's Conference on Employer-Sponsored.Child 171 Care for the State of Michigan. Fort Washington, PA: LRP Press, 35-47. George, L. K., and Gwyther, L. P. (1986). Caregiver well-being: A multidimensional examination of family caregivers of demented adults. The Gerontologist, 26 (3) . 253-259. Glass, J., &.Camarigg, V. (1992). Gender, parenthood, and.job- family compatibility. American Journal df §ocigldgy, 98 (1), 131-151. Goff, S. J., Mount, M.K., & Jamison, R. L. (1990). Employer supported. child. care, 'work-family' conflict, and absenteeism: A field study. Psrsdnnel Psythdldgy, 43, 793-809. Granrose, C. S., Parasuraman, S., & Greenhaus, J. H. (1992). A. proposed. model of support. provided. by two-earner couples. Human Relatidns, 45 (12), 1367-1393. Greenhaus, J. H. (1988). The intersection of work and family roles: Individual, Interpersonal, and organizational issues. ddurnal df §dgial Bshavidr and Pstsdnality, 3 (4), 23-44. Greenhaus, J. H. (1989). The intersection of work and family roles: Individual, Interpersonal, and organizational issues. In E. B. Goldsmith (Ed.), Wdrk and family; Thsdty, rsssargh, and applidations; 23-44. Newbury Park, CA: Sage. Greenhaus, J. H., &.Beutell, N; J. (1985). Sources of conflict between work and family roles. Academy of Managsmsnt Rsvisw, 10, 76-88. Greenhaus, J. H., & Parasuraman, S. (1986). A work-nonwork interactive perspective of stress and its consequences. ddurnal pf Qrgauizatidnal Bshavior‘Managsment, 8, 37-60. Greenhaus, J . H. , & Parasuraman, S. , (1994) . Work-family conflict, social support and well-being. In M. J. Davidson & R . J . Burke (Eds . ) , Woman in uauagsmsnt . London: Paul Chapman Publishing Ltd. Greenhaus, J. H., Parasuraman, S., Granrose, C., Rabinowitz, S., & Buetell, N. (1989). Sources of*work-family conflict among two-career couples. Qdurnal df thatidnal Bauavidr, 34, 133-153. Gross, N., Mason, W. S., & McEachern, A. W. (1958). Expldratidns iu tdle analysis: Studiss df ths sshdol 172 supsrintendency rdls. New York: Wiley. Grover, S. L. (1991). Predicting the perceived fairness of parental leave policies. dournal of Applied Psyghdldgy, 76 (2), 247-255. Gutek, B. A., Searle, S., & Klepa, L. (1991). Rational versus gender role explanations for work-family conflict. ddurnal df Applisd Psychology, 76 (4), 560-568. Hall, D. T. (1972). A.model of coping with role conflict: The role behavior of college educated women. Administrative stisncs Quartsrly, 17, 471-486. Hall, D. T. (1989). Moving beyond the "mommy track": An organization.change-approach. Parsonnel, 66 (12), 23-29. Hall, F. S., & Hall, D. T. (1978). Dual careers - How do couples and companies cope *with the ‘problems? Qrganizatidnal Dyuamigs, 6 (4), 57-77. Hall, D. T., & Richter, J. (1988). Balancing work life and home life: What organizations can do to help. W of Management Exegutive, 2 (3), 213-223. Higgins, C. A., & Duxbury, L. E. (1992) . Work-family conflict: A comparison of dual-career and traditional-career men. ddurnal df Organizational Bahavior, 13, 389-411. Higgins, C. A., Duxbury, L. E., & Irving, R. H. (1992). Work- family conflict in the dual-career family. Qrgauizatidnal Bahavidr and Human Dsgisidu Prdgss , 51, 51-75. Hock, E., Christman, K., & Hock, M. (1980). Factors associated with return to work in mothers and infants. stsldpmsntal Psydholggy, 16 , 535-536 Hofferth, S. L. (1988). Child care in the United States. In Amaridan familiss in tdmdrrdw’s economy, Hearing before the Select Committee on Children, Youth and Families, July 1, 1987. Washington, D.C.: U.S. Government Printing Office. Homans, G. C. (1976). Fundamental processes of social exchange. In E. P. Hollander & R. G. Hunt (Eds.) . gamut psrspettives in sdtial psythology (4th ed.) : 161-173 . New York: Oxford University Press. Horowitz, A. (1985). Sons and daughters as caregivers to parents: Differences in role performance and consequences. Ths Gardntdldgist, 25, 612-617. 173 Jick, T. D. and Mitz, L. F. (1985). Sex differences in work stress. Atadsmy df Management Review, 10, (3), 408-420. Jones, A. P., & Butler, M. C. (1980). A role transition approach to the stresses of organizationally-induced family role disruption. Journal df Marriags and ths Family, 42, 367-376. Joreskog, K. G., & Sorbom, D. (1989) . LISREL 7: A guids t9 the prdgram and applitations (2nd ed.). Chicago, IL: SPSS, Inc. Katz, M. H. & Piotrkowski, C. S. (1983). Correlates of family role strain among employed black women. Family Relatidns, 32, 331-339. Keats, B. W., & Hitt, M. A. (1988). A causal model of linkages among environmental dimensions, macro organizational characteristics, and performance. Acadsmy of Managsmsnt figurnal, 31 (3), 570-598. Keith, P. M., & Schafer, R. B. (1980). Role strain and depression in two job families. Family Rslatidns, 29, 483-488. Keith, P. M., & Schafer, R. B. (1984). Role behavior and psychological well-being: A comparison of men in one-job and two-job families. American ddurual df Qrthepsxshiatrz- 54 (1). 137—154. Kelly, R. F. & Voydanoff, P. (1985). Work/family role strain among employed parents. Family Relations, 34, 367-374. Kirchmeyer, C. (1992). Nonwork participation and work attitudes: A test of scarcity vs. expansion models of personal resources. Humau Rslatidns, 45 (8), 775-795. Kopelman, R., Greenhaus, J., &.Connolly, T. (1983). Aumodel of work, family and interrole conflict: A construct validation study. Qrganizatidnal Behavidr and Human Psrfdrmants, 32 (2), 198-215. Kossek, E. E. (1990). Diversity in child care assistance needs: Employee problems, preferences, and work-related outcomes. Eatsdnns; Psyghdldgy, 43, 769-791. Kossek, E. E., Dass, P., & DeMarr, B. J., (1994). The dominant logic of employer-sponsored.work and family initiatives: Human resource managers’ institutional role. Human Belatio_s. 47. (9). 1121-1148. Kossek, E. E., & DeMarr, B. J., (1996). Work and family 174 matters: Overlooked influences on individual orientations toward work and family interaction. Unpublished Working Paper. Kossek, E. E., DeMarr, B. J., Backman, K., 8. Kollar, M. (1993). Assessing employees’ emerging elder care needs and reactions to dependant care benefits. Publit Psrsonnsl Managsment, 22 (4), 617-638. Kossek, E. E. & Nichol, V. (1992). The effects of on-site child care on employee attitudes and performance. Parsdnnsl Psythdldgy, 45, 485-509. Lambert, S. J., (1990). Processes linking Work and family: A critical review and research agenda. Human Rslatipus, 43 (3), 239-257. Lambert, S. J., (1991). The combined effects of job and family characteristics on the job satisfaction, job involvement, and intrinsic motivation of men and women workers. ddurnal pf Organizational Bahavior, 12, 341-363. Lobel, S. A. (1991). Allocation of investment in work and family roles: Alternative theories and implications for research. Academy of Managsmsnt Revisw, 16 (3), 507-521. Lobel, S. A., & St. Clair, L., (1992). Effects of family responsibilities, gender and career identity salience on performance outcomes. Academy of Managsmsnt ddurnal, 35 (5), 1057-1069. Locksley, A. (1980). On the effects of wives’ employment on marital adjustment and companionship. dournal df Mattiage and the Family, 42, 337-346. Long, J. S. (1989). Covariance Structurs Mddsls; An Introdution to LISREL. Beverly Hills, CA: Sage. Magid, Rs Y. (1983). Child, dare initiatives for"wdrking parsnts: Why smployers get involved. New York: American Management Association. Magnus, M. (1988). Eldercare: Corporate awareness but little action. Patsdnusl Jdurnal, 67 (6), 19-23. Mattis, M. C. (1990). New forms of flexible work arrangements for managers and professionals: Myths and realities. Human Rssdurds Plauning, 13 (2), 133-146. Miller, J. J., Stead, B. A., & Pereira, A. (1991). Dependent care and the workplace: An analysis of management and employee perceptions. ddutual pf Businsss Ethids, 10, 175 863-869. Milliken, F. J., Dutton, J. E., & Beyer, J. M. (1990). Understanding organizational adaptation to change: The case of work-family issues. Human Resource Planning, 13 (2), 91-106. Mottaz, C. (1986). Gender differences in work satisfaction, work-related rewards and values, and the determinants of work satisfaction. Human Rslations, 39, 359-378. Mulaik, S. A., James, L. R., Van Alstine, J., Bennett, N., Lind, S., & Stilwell, C. D. (1989). Evaluation of goodness-of-fit indices for structural equation models. Egyghglggigal Bullgtin, 105 (3), 430-445. Noelker, L. S., & Poulshock, S. W. (1982). The effects on families of caring for impaired elderly in residence. Final Report submitted to AOA. Cleveland, OH: Benjamin Rose Institute. O’Driscoll, M. P., Ilgen, D. R., & Hildreth, K. (1992). Jdurnal df Applisd Psychdldgy, 77 (3), 272-279. Parasuraman, S., Greenhaus, J. H., & Granrose, C. S. (1992). Role stressors, social support, andnwell-being'among'two- career couples. Journal of Qrganizational Bshavidr, 13, 339-356. Parasuraman, S., Greenhaus, J. H., Rabinowitz, S., Bedeian, A. G., & Mossholder, K. W. (1989). Work and family variables as mediators of the relationship between wives’ employment and husbands’ well-being. Acadsmy df Managemsnt Jdurnal, 32 (1), 185-201. Piotrkowski, C. (1979). Wdrk and ths Family Systsms New York: The Free Press. Piotrkowski, C. S., & Katz, M. H. (1982). Indirect socialization of children: The effects of mothers’ jobs on academic behaviors. Child.steldpment, 53, 1520-1529. Piotrkowski, C. S., Rapaport, R. N., & Rapaport, R. (1987). Families and work. In M. B. Sussman & S. K” Steinmetz (Eds.), Haudbddk df martiags and the family (pp. 251- 283). New York: Plenum Press. Pleck, J. H. (1977). The work-family role system. $29113. Prdplsms, 24, 417-427. Pleck, J. H. (1985). Watking wivss[working husbands. Beverly Hills, CA: Sage. 176 Pleck, J. H., Staines, G. L., & Lang, L. (1980). Conflicts between work and family life. Monthly Labor Review, 103 (3), 29-32. Powell, G. N., & Mainiero, L. A. (1992). Cross-currents in the river of time: Conceptualizing the complexities of women’s careers. Journal of Management, 18 (2), 215-237. Pratt, C., Schmall, V., & Wright, S. (1987). Ethical concerns of family caregivers to dementia patients. Ths Gardntdlogist, 27 (5), 632-638. Quinn, R. P., & Staines, G. L. (1978). The 1277 guality pf smpldymsnt survsy, Ann Arbor, Michigan: ISR, University of Michigan, 1978. Rice, R. W., Frone, M. R., & McFarlin, D. B. (1992). Work- nonwork conflict and the perceived quality of life. Journal df Organizational Behavior, 13, 155-168. Rodgers, F. S., & Rodgers, C. (1989). Business and the facts of family life. Harvard Business Review, 67 (6), 121-129. Rosin, H. M. (1990). The effects of dual career participation on men: Some determinants of variation in career and personal satisfaction. Human Rslatidns, 43 (2), 169-182. Rosin" H., &.Korabik, K. (1990). Marital and family correlates of women managers’ attrition from organizations. Jdutnal df Vdcational Behavior, 37 (1), 104-120. Rothausen, T. J. (1995). Measuring the impact of family on work attitudes and behaviors: Review and development and construct validation of family responsibility level (FRL). Paper presented at the 55th annual meeting of the Academy of Management, Vancouver, British Columbia, August 1995. Scharlach, .A. E. (1987). Role Strain. in. Mother-Daughter Relationships in Later Life. The Gerontoldgist, 27 (5), 627-631. Scharlach, A. E., & Boyd, 8. L. (1989). Caregiving and Employment: Results of an employee survey. Ths Gardntdldgist, 29 (3), 382-387. Schmitt, N. (1994). Method bias: The importance of theory and measurement. Jdurnal of Organizational Behavidr, 15, 393- 398. Schneer, J. A., & Reitman, F. (1993). Effects of alternate family structures on managerial career paths. Adadsmy_dt 177 Management Jdurnal, 36 (4), 830-843. Schwartzberg, N. S., & Dytell, R. S. (1988). Family stress and psychological well-being among employed and nonemployed mothers. Journal df Social Behavior and Personality, 3 (4): 175-190. Shaw, S. B. (1987). Parental aging: Clinical issues in adult psychotherapy. Social Casework: The Journal df Contsmporaty Social Work, 68, 406-412. Sizemore, FL (P., & Jones, A. B. (1990). Eldercare and the workplace: Short-term training preferences of employees. Educatidnal Carontdlogy, 16, 97-104. Solomon, C. M. (1994). Work/family’s failing grade: Why today’s initiatives aren’t enough. Personnsl Jdurual, 73 (5), 72-87. Spector, P. E. (1994). Using self-report questionnaires in OB research: A comment on.the use of a controversial method. Jdurnal df Qrganizational Bshavior, 15, 385-392. Staines, G. L. (1980). Spillover versus compensation; A review of the literature on the relationship between work and nonwork. Human Relations, 33, 111-129. Staines, G. L., Pleck, J., Shepard, L., & O’Connor, P. (1978). Wives’ employment status and marital adjustment: Yet another look. Psychology df Women Quarterly, 3, 90-120. Stamp, D. (1989) . Agenda for the 19908. Businsss Quartsrly, 54 (1), 111-115. Stone, R., Cafferata, G. L., & Sangl, J. (1987). Caregivers of the Frail Elderly: A National Profile, The Gsrontdldgist, 27 (5), 616-626. Strickland, W. J. (1992). A typology of career wife roles. Human Relatidns, 45, 797-811. Sullivan, S. E., & Gilmore, J. B. (1991). Employers begin to accept eldercare as a business issue. Personnel, 68 (7), 3-4. Troll, L. E. (1971). The family of later life: A decade review. Jdutual df Marriags and the Family, 33, 263-290. U.S. Bureau of the Census (1994). Statistical Absttatt 9f the Unitsd Statss; 1224 (114th sdition). Washington, DC. U.S. Department of Labor Women’s Bureau (1986). Fatts dn U.S. 178 Working Women, Fact Sheet No. 86-4. Voydanoff, P. (1988a). Work and family: A review and expanded conceptualization. Journal of Social Behavidr and 252501151122- 3 (4), 1-22. Voydanoff, P. (1988b). Work. role Characteristics, family structure demands, and work/family conflict. Journal df Marriage and the Family, 50 (3), 749-761. Voydanoff, P. & Kelly, R. F. (1984). Determinants of work- related family problems among employed parents. Jdurnal df Marriags and the Family, 46, 881-892. Wharton, A. S., & Erickson, R. J. (1993). Managing emotions on the job and at home: Understanding the consequences of multiple emotional roles. Academy of Managsmsnt Rsvisw, 18 (3), 457-486. Wiersma, U. J} (1990). Gender differences in job attribute preferences: Work-home role conflict and job level as mediating variables. Journal df Occupatidnal Psyshdldgy, 63, 231-243. Williams, K. J., & Alliger, G. M. (1994). Role stressors, mood spillover, and perceptions of work-family conflict in employed.parents. Atademy df Management Journal, 37 (4), 837-868. Williams, K. J., Suls, J., Alliger, G. M., Learner, S. M., & Wan, C. K. (1991). Multiple role juggling and daily mood states in working mothers: An experience sampling study. Jdurnal pf Applied Psychology, 76 (5), 664-674. Winfield, F. E. (1987). Workplace Solutions for Women Under Eldercare Pressure. Persdnnsl, 64 (7), 31-39. Wolf, D. A, & Soldo, B. J. (1994). Married women’s allocation of time to employment and care of elderly parents. Journal pf Human Rssourtes, 29 (4), 1259-1275. Zarit, S. H., Reever, K. E. , 8: Bach-Peterson, J. (1980) . Relatives of the impaired elderly: Correlates of feelings of burden. Ths Gardntologist, 20, 260-266. Zedeck, S. & Mosier, K. L. (1990). Work in the family and employing organization. Amsrican Psychologist, 45 (2), 240-251. Appendix A APPENDIX A satisfagtidn with Child Care Scalea How satisfied are you with the following aspects of the main child care arrangement you currently have for your youngest child? Convenience of location Safety Attention General quality of care Educational content My relationship with caregiver(s) (hU'IthJNH a Responses ranged from "very dissatisfied" (1) to "very satisfied" (4). Nu r f El er re Ta k Scal b What kinds of care and assistance have you.provided.during the last 6 months? Taking someone to doctors or other services Arranging medical appointments or other kinds of appointments 3 Providing meals 4 Visiting in person or over the telephone 5 Helping around the house 6 Personal care (e.g., bathing or dressing) 7 Filling out legal or insurance forms 8 Helping the person look for a new home 9. Giving medications 10. Shopping 11 12 13 NH Providing emotional support Providing financial assistance Other (Please describe.) b Respondents were asked to circle all that applied. 179 180 Carssr Penaltiss ScaleC To what extent do you agree or disagree with the following statements about flexible or alternative work arrangements (e.g., flextime, part-time, job sharing) at The Company, whether or not these are currently available in your department? 1. Employees who use flexible work arrangements will be perceived as less committed to The Company. 2. Co-workers resent those who have flexible work arrangements. 3. Using a flexible work arrangement will hurt your career. C Responses ranged from "strongly disagree" (1) to "strongly agree" (4). Supervisdr guppdrt Stalsd To what extent do you agree or disagree with each of the following statements about your immediate manager or supervisor' and your work—personal life responsibilities? (Please answer the questions with your immediate supervisor in mind.) 1. My manager/supervisor treats everyone fairly in responding to employees’ personal needs. 2. My manager/supervisor is helpful when I have a personal emergency. 3. My manager/supervisor is helpful when I have to take care of a routine personal matter. 4. My manager/supervisor appears to know a lot about Bank policies that help employees manage their personal responsibilities. 5. My manager/supervisor generally lets me make and receive important personal telephone calls at work. 6. My manager/supervisor actively involves my work group in figuring out how to balance the needs of the business with people’s personal responsibilities (e.g., scheduling issues). d Responses ranged from "strongly disagree" (1) to "strongly agree" (4). 181 Qrganizatidnal Wdrk-Family Culture Scalse To what extent do you agree or disagree with each of the following statements? 1. 2. 3. "The Company" is concerned about the well-being of employees and their families. "The Company" has a workaholic culture. At "The Company", employees have to choose between advancing in their jobs or devoting attention to their personal lives. Employees are judged more on the quantity (e.g., how many hours they work) than on the quality of their work. There are managers at "The Company" who set good examples of how to balance work and personal life. There are some employees (e.g., men, single employees, childless employees)who are expected to do more work because it is thought they do not have outside commitments. e Responses ranged from "strongly disagree" (1) to "strongly agree" (4). Dependant Cars Strsss Scalef In the last 3 months, have any of the following issues caused you stress? 1. 2. 3. Child care issues Teen issues Elder/dependent care issues f Responses ranged from "no stress" (1) to "a lot of stress" (3). 182 Wdrk Stress Stalsg In the last 3 months, have any of the following issues caused you stress? OlU'ltbbJNl-J Your manager/supervisor Meeting monthly expenses Worries about losing your job Work schedules Work load Your daily commute 9 Responses ranged from "no stress" (1) to "a lot of stress" (3). Psythdldgigal Distrsss Staleh During the past three months, how often have you: 1. been bothered by minor health problems such as headaches, insomnia, or upset stomach? felt that you were unable to control the important things in your life? felt nervous and stressed? found that you could not cope with all the things you had to do? felt difficulties were piling up so high that you could not overcome them? h Responses ranged from "never" (1) to "very often" (4). 183 Child Care Benefitsi 1. 2. 3. 10. 11. 12. 13. 14. 15. Offer (or improve existing) child care referral services to help parents find care Sponsor a child care center at or near the worksite Increase the supply and quality of community child care programs to make them more accessible to Company employees Create a child care center for mildly ill children (on site or in the community) Create an in-home service where trained.nurses care for mildly ill children in your home Provide a back-up child care center at the worksite for last-minute emergencies Help expand or establish the supply of back-up or emergency child care services in the community Increase the availability of early-morning and late evening child care in the community Help expand or establish summer child care/camp options Arrange for care of children during school holidays and vacations Help expand or establish after-school care Arrange for transportation for children to get from school to after-school activities Provide financial assistance through vouchers or discounts at local programs to reduce child care costs Offer seminars for parents with experts talking about specific topics Encourage support groups for' parents 'where employees serve as a resource to one another 1 Responses ranged from."The Company should not develop this option" (1) to "this would be helpful to me personally" (4). 184 Eldsr Care Bensfits Scalej Elder/Dependent Care 1. Offer (or improve existing) elder care resource and referral (e.g., information to help employees identify services) Offer case management for long-distance caregiving (e.g., someone to arrange care for elderly or other dependents who live far away from employees) Provide financial assistance through vouchers (or discounts at local programs to reduce elder or other dependent care costs) Offer seminars for elder or other dependent care providers Provide information for people who expect to provide care in the future to elderly or adult dependents. Offer support groups for employees with elder or dependent care responsibilities j Responses ranged from "The Company should not develop this option" Work Flsxibility Training Stale 1 . 2. 3. (1) to "this would be helpful to me personally" (4). k Train managers/supervisors on how to deal with work-personal issues facing employees Build managing flexibly into managers’ and supervisors’ performance evaluation Provide managers and supervisors with written guidelines on managing flexibility k Responses ranged from "The Company should not develop this option" (1) to "this would be helpful to me personally" (4). 185 Psrceivsd Wdrk Productivity Scalel In the past 3 months, have you done any of the following because of your personal responsibilities? 1. Produced lower quality work 2. Made errors or had on—the-job accidents 3. Been distracted so that productivity declined 1' Responses were "yes" or "no". “llllllllllllllr