Citation
Mapping the External Terrain of Marital Satisfaction: A Review and Meta-Analysis of Contextual Stressors

Material Information

Title:
Mapping the External Terrain of Marital Satisfaction: A Review and Meta-Analysis of Contextual Stressors
Creator:
SPRINGER, SHAUNA H. ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Marital satisfaction ( jstor )
Marriage ( jstor )
Medical conditions ( jstor )
Psychological stress ( jstor )
Social psychology ( jstor )
Spouses ( jstor )
Stress functions ( jstor )
Stress tests ( jstor )
Unemployment ( jstor )
Wives ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Shauna H. Springer. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
8/31/2016
Resource Identifier:
658230824 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

MAPPING THE EXTERNAL TERRAIN OF MARITAL SATISFACTION: A REVIEW AND META-ANALYSIS OF CONTEXTUAL STRESSORS By SHAUNA H. SPRINGER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Shauna H. Springer

PAGE 3

This document is dedicated to the graduate students of the University of Florida.

PAGE 4

ACKNOWLEDGMENTS First and foremost, I would like to thank my husband for his incredible support and patience during the completion of this project. It is a wonderful thing to be married to a very handsome computer guru! I would also like to acknowledge the support of both sides of my family, those in California and in Wisconsin. I am grateful to both of my dissertation cochairs, Dr. Ben Karney and Dr. Greg Neimeyer, for their responsiveness and guidance on this project. Finally, I’d like to thank my colleagues and friends in the Counseling Psychology Department at the University of Florida, whose friendship has helped me as much as anything else to persevere until the completion of my degree. iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................iv LIST OF TABLES ...........................................................................................................viii LIST OF FIGURES ...........................................................................................................ix ABSTRACT .........................................................................................................................x CHAPTER 1 INTRODUCTION TO THE RESEARCH PROBLEM...............................................1 2 THEORY REVIEW AND DESCRIPTION OF COMPONENTS IN CURRENT STRESS THEORIES....................................................................................................6 Components of Current Stress Theories.......................................................................8 Context..................................................................................................................8 Resources and Vulnerabilities.............................................................................14 Coping.................................................................................................................16 Outcomes.............................................................................................................20 Summary of the Components of Stress Theories.......................................................23 3 THEORIZED ASSOCIATIONS BETWEEN COMPONENTS OF STRESS THEORIES.................................................................................................................25 Direct Associations between Stress and Physical and Psychological Functioning....25 Resources and Vulnerabilities Moderate the Association between Stress and Marital Functioning...............................................................................................26 Coping Processes Mediate the Association between Stressors and Outcomes..........31 Outcomes Moderate the Impact of Future Stressors...................................................33 Summary of Associations in Stress Theories and Justification for Empirical Review...................................................................................................................35 4 REVIEW OF THE EMPIRICAL LITERATURE OF ASSOCIATIONS BETWEEN STRESS AND MARITAL SATISFACTION........................................37 General Overview.......................................................................................................37 Sample of Studies.......................................................................................................39 v

PAGE 6

Selection Criteria........................................................................................................40 Coding Domains of Stress Studied in Samples in the Review...................................41 Description of Characteristics of the Empirical Literature.........................................43 Domains of Stress Studied in Samples in the Review.........................................43 Coding Characteristics of Studies in the Review................................................46 Publication type and sample size characteristics of studies in the review...47 Demographic characteristics of samples addressed by studies in the review......................................................................................................49 Design characteristics of studies in the review............................................54 Summary and Critique of Characteristics in the Empirical Literature.......................60 5 METHOD AND PROPOSED HYPOTHESES..........................................................68 Rationale.....................................................................................................................68 Meta-Analytic Inclusion Criterion..............................................................................70 Research Questions.....................................................................................................71 Analysis of Association between Stressors and Marital Satisfaction..................71 Analysis of Theorized Moderation of Associations between Stress and Marital Satisfaction..........................................................................................74 Meta-Analytic Procedure............................................................................................83 Aggregation of Effect Sizes.................................................................................83 Test of Homogeneity...........................................................................................86 Moderating Analysis of Gender..........................................................................87 Moderating Analyses based on Sampling Characteristics...................................87 6 RESULTS OF THE META-ANALYSIS...................................................................91 Broad Overview of Domains to be Meta-Analyzed...................................................91 Descriptive Characteristics of Studies in the Meta-Analysis.....................................91 Comparisons between Studies in the Review and Studies in the Meta-Analysis.......93 Presentation of Meta-Analytic Results.......................................................................95 Aggregation of Effect Sizes.................................................................................95 Test of Gender Differences...............................................................................100 Moderating Analyses based on Sampling Characteristics........................................105 Income Level.....................................................................................................105 Racial Composition of Sample..........................................................................109 National Representation....................................................................................114 Summary of Meta-Analytic Findings.......................................................................116 Factors that Limit Interpretations of these Results............................................119 Strategy for Final Chapter.................................................................................122 7 CONCLUSIONS AND FUTURE DIRECTIONS...................................................123 Brief Review of the Rationale..................................................................................123 Implications for Sampling and Methods...................................................................124 Suggested Improvements in Measurement........................................................125 Suggested Improvements in Design..................................................................127 vi

PAGE 7

Implications for Theory and Research......................................................................132 Implications for Public Policy and Intervention.......................................................142 Concluding Comments.............................................................................................145 LIST OF REFERENCES.................................................................................................147 BIOGRAPHICAL SKETCH...........................................................................................183 vii

PAGE 8

LIST OF TABLES Table page 4-1 Categories of Stress Measured in Existing Studies..................................................46 6-1 Number of Individual Correlations and Median Sample Size in each Domain of Stress........................................................................................................................91 6-2 Summary of Descriptive Characteristics for Studies in the Meta-analysis (n=113).....................................................................................................................94 6-3 Comparison of Characteristics of Articles in the Review with those in the Meta-analysis.....................................................................................................................95 6-4 Aggregate Effect Sizes of Specific Domains of Stress on Marital Satisfaction.....100 6-5 Aggregate Effect Sizes within each Domain for Females and Males and Tests of Gender Differences................................................................................................105 6-6 Effect Size by Domain and Income Collapsing over Gender................................109 6-7 Effect Size by Domain and Racial Composition Collapsing Across Gender........114 6-8 Effect Size by Domain and National Representation Collapsing over Gender......116 6-9 Statistically Significant Results at the p<.05 level across Stress Domains and Moderators.............................................................................................................119 viii

PAGE 9

LIST OF FIGURES Figure page 4-1 Publication Year Breakdown of Articles in the Empirical Review.........................47 4-2 Sample Size Distribution for Studies in the Empirical Review...............................48 4-3 Gender Distribution of Participant in Samples in the Empirical Review................49 4-4 National Representation of Samples in the Empirical Review................................50 4-5 International Sample Recruiting...............................................................................51 4-6 Racial Composition of Samples in the Empirical Review.......................................52 4-7 Median Income of Studies in the Empirical Review................................................53 4-8 Study Design in Articles in the Empirical Review..................................................55 4-9 Number of Waves in Longitudinal Studies in the Review.......................................56 4-10 Average Length of Longitudinal Studies in the Review..........................................56 4-11 Sampling Unit in Studies in the Empirical Review..................................................57 4-12 Separation of Gender in Analyses............................................................................58 4-13 Measurement Methods in the Studies in the Empirical Review..............................59 7-1 Profile Model of Adjustment to Crisis (Hill, 1949)..............................................127 7-2 Vulnerability-Stress-Adaptation Model.................................................................134 ix

PAGE 10

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MAPPING THE EXTERNAL TERRAIN OF MARITAL SATISFACTION: A REVIEW AND META-ANALYSIS OF CONTEXTUAL STRESSORS By Shauna H. Springer May 2006 Chair: Gregory Neimeyer Cochair: Benjamin Karney Major Department: Psychology Attention to mean rates and summary statistics has the potential to obscure the fact that, like other threats to public health, marital dissolution is a problem that disproportionately affects some populations more than others. Disparities in divorce rates across neighborhoods and levels of socioeconomic status highlight the fact that marriages are shaped by the contexts in which they form and develop. In light of the theoretical and practical benefits of gaining a clearer understanding of how marital quality may be affected by variables external to the marriage, the existence of a substantial literature on this topic, and the lack of a comprehensive review, the goal of this dissertation was to identify, evaluate, and analyze research on how stressors external to a marriage may be associated with perceptions of satisfaction within the dyad. First, existing theories exploring stress effects on marriages are described and areas of ongoing disagreement or consensus are summarized. Following the theoretical review is an examination of the publication, sampling, and methodological characteristics of 225 empirical articles x

PAGE 11

exploring associations between stress and marriage, representing 164,359 married individuals. The next section, featuring a meta-analysis of 113 articles, addresses differences in aggregate effect sizes of different types of stressors and explores potential moderators of stress effects. Finally, the concluding section discusses several clinical and policy implications of the accumulated results and recommends several immediate priorities for future theory development and research. xi

PAGE 12

CHAPTER 1 INTRODUCTION TO THE RESEARCH PROBLEM The quality of a marriage exerts a profound influence on the broader quality of spouses’ lives. Those who are happily married live longer (House, Landis, & Umberson, 1988), are healthier (Kiecolt-Glaser & Newton, 2001), and recover from stressful events more quickly (Waltz, Badura, Pfaff, & Schott, 1988) than people who are unmarried or unhappily married. The children of happy marriages are better adjusted (Grych & Fincham, 1990) and less likely to get divorced themselves (Amato & Keith, 1991). Research on general well-being reveals that being happily married contributes more strongly to life satisfaction than any other variable that has been measured, including health, financial security, and professional success (Glenn & Weaver, 1981). Given that a fulfilling marriage promises the best that life can offer, it is perhaps not surprising that in the United States 97% of people get married at some point in their lives (Kreider & Fields, 2001). Despite the promise that a healthy marriage offers, however, this promise goes unfulfilled for many of those who marry. Census data provide the oft-cited statistic that the risk of divorce or permanent separation for newly married couples hovers above 50% in the United States. Approximately 80% of those who get divorced are likely to marry again (Kreider & Fields, 2001), but those marriages are even more likely to dissolve (Cherlin, 1992). Even marriages that endure may not be satisfying, and marital distress has severe and negative consequences for partners’ physical health (Kiecolt-Glaser & Newton, 2001), their emotional health (Beach, 2001), and their children’s outcomes 1

PAGE 13

2 (Booth & Amato, 2001). The social impact of marital quality and the prevalence of marital distress and dissolution have led some to conclude that “preserving and enhancing the quality and stability of couples’ relationships is a major public health issue” (Dyer & Halford, 1998, p. 107). 1 Accordingly, lawmakers and social leaders have embarked on a national debate about the best educational programs and policies to shore up the institution of marriage. Yet attention to mean rates and summary statistics has the potential to obscure the fact that, like other threats to public health, marital dissolution is a problem that disproportionately affects some populations more than others (Bramlett & Mosher, 2001). In particular, rates of marital dissolution are strongly associated with levels of income, such that the risk for individuals of low-income (defined as within 200% of the poverty line) is several times greater than the risk for individuals of middleor high-income. Similarly, divorce rates vary substantially across neighborhoods and states. According to national data collected in 1995, in communities where rates of unemployment, numbers of families receiving public assistance, and rates of poverty are relatively high, and where median family incomes are relatively low, the chance of a first marriage breaking up is nearly twice as great as in communities where socioeconomic indicators are higher (Bramlett & Mosher, 2001, Figure 26). Disparities in divorce rates across neighborhoods and levels of socioeconomic status highlight the fact that marriages do not succeed or fail in a vacuum. Rather, they 1 Throughout this article, the terms marital quality , marital satisfaction , marital adjustment , marital distress , fulfilling marriage , and healthy marriage are used interchangeably to refer to spouses’ own global evaluations of their marital relationship (e.g., Fincham & Bradbury, 1987). In contrast, the terms marital stability and marital dissolution are used to refer to whether or not spouses decide to end the marriage via divorce or permanent separation.

PAGE 14

3 are shaped by the contexts in which they form and develop. Indeed, associations between environmental contexts and marital stability have long been recognized within research on marriage. As early as the 1930s, the first work to identify predictors of divorce established links between economic strain and an increased risk for marital dissolution (Bakke, 1940; Angell, 1934; Komarovsky, 1940; Liker & Elder, 1983). In the 1970s, demographic research demonstrated that poverty was associated with higher rates of divorce (e.g., Cherlin, 1979; Zill, 1978, as cited in Zill & Nord, 1994). In recent years, research exploring associations between contextual factors and marital outcomes has expanded to include other types of external stressors such as rural economic decline (e.g. Johnson & Booth, 1990), economic hardship (e.g., Conger, Rueter, & Elder, 1999), negative life events (e.g., Cohan & Bradbury, 1997), and post-traumatic stress disorder in Vietnam Veterans (e.g., Riggs, Byrne, Weathers, & Litz, 1998). Recognizing this history, researchers have increasingly suggested that understanding a marriage requires insight related to the larger network of social and physical contexts within which it is embedded (e.g., Berscheid, 1998; Huston, 2000; Levinger, 1994; Reis, Collins, & Berscheid, 2000; Vemer, Coleman, Ganong, & Cooper, 1989). Yet despite the well-developed literature on associations between stressful contexts and marital dissolution, there persists a sense that variables external to spouses and their behavioral interactions have been overlooked within research on determinants of marital quality. This would be a significant oversight, because the determinants of marital dissolution may be distinct from the determinants of marital quality (e.g., Karney and Bradbury, 1995) and a disproportionate focus on internal factors that influence decisions to end a marriage may offer only limited insight into the ways that external

PAGE 15

4 context may affect the health of the marital relationship. With respect to this latter issue, Huston (2000) has observed that “the portraits of marriage that social scientists createusually include little information about the psychological, social, and environmental contexts in which interaction[s] are embedded” (p. 299). Berscheid (1999) goes so far as to suggest that “we have been prey to the fundamental attribution error in our overemphasis on the causal role of individual dispositions and our neglect of relationships’ environments” (p. 264). Given these concerns, some may be surprised to learn that over 200 empirical reports have addressed associations between contextual variables and marital quality. The lack of awareness of these studies may stem in part from the fact that they have been conducted across several decades and across a wide range of disciplines. Moreover, I am aware of no attempts to assemble and integrate this literature so that established associations and unanswered questions might be identified. As a consequence, the existing research exploring associations between contextual factors and marital quality remains diffuse, and is thus unlikely to accumulate sufficiently to provide direction for current research or to inform current discussions regarding policies and interventions that may benefit marriage. In light of the theoretical and practical benefits of gaining a clearer understanding of how marital quality may be affected by variables external to the marriage, the existence of a substantial literature on this topic, and the lack of a broad review, the goal of this dissertation is to identify, evaluate, and analyze research on how stressors and supports external to a marriage may be associated with perceptions of satisfaction within the dyad. Towards this end, the rest of this dissertation is organized into four broad

PAGE 16

5 components. The first section describes the theoretical perspectives that have informed research on marriages in context to highlight areas of ongoing disagreement or consensus. The second section evaluates the research itself, examining the methods that have been used to address associations between contextual factors and marital satisfaction empirically. The third section presents the results of a meta-analysis conducted to estimate the associations of different sources of external stress and support on marital satisfaction across several studies, as well as to identify potential moderators of these associations across studies. Finally, the concluding section explores the clinical and policy implications of the accumulated results and recommends several immediate priorities for future theory development and research

PAGE 17

CHAPTER 2 THEORY REVIEW AND DESCRIPTION OF COMPONENTS IN CURRENT STRESS THEORIES Existing theoretical frameworks for understanding marriage have focused disproportionately on variables internal to the relationship. For example, perhaps the most influential theory of marital development in the last several decades has been behavioral theory, an approach that highlights the role of marital communication and problem-solving behaviors in determining marital outcomes (e.g., Gottman & Krokoff, 1989). Research in the cognitive-behavioral tradition expanded the original concentration on behavior to include the way spouses interpret and understand each other’s behavior (e.g., Lopez, 1993; Murray & Holmes, 1997), but this approach still remains centered on processes taking place exclusively within the dyad. The interdependence approach to understanding marriage, drawn in part from the social exchange theory, adopts a somewhat broader view. Social exchange theory (Thibaut & Kelley, 1959) suggests that relationship satisfaction is a function of each partner’s perceptions of relationship attractions relative to potential alternatives and barriers to leaving the relationship. Even though barriers to leaving and potential alternatives are certainly related to external contexts such as the local “relationship market,” the main focus of social exchange theory is on how partners’ internal perceptions affect the course of the relationship. Finally, attachment theory (Bowlby, 1969; Hazan & Shaver, 1987), another widely influential tradition (for a review, see Fraley & Shaver, 2000), suggests that the development of trust or distrust in infancy influences later romantic relationship behavior. Across all of these 6

PAGE 18

7 theories, then, there is a common tendency to focus primarily on internal factors as determinants of relationship satisfaction. Relative to the prevalent focus on variables internal to the spouses and the dyad, much less attention has been devoted to understanding the potential impact of variables external to marital relationships. Recently, a number of prominent researchers (e.g., Berscheid, 1999; Huston, 2000; Kurdek, 1991; Levinger & Huston, 1990; Reis, Collins, & Berscheid, 2000; White, 1990; Wilson, Larson, McCulloch, & Stone, 1997) have argued that this disproportionate focus on internal variables is a serious limitation. In the early 1990s, White (1990) called for “research that establishes the structural conditions of high versus low divorce” (p. 910). Kurdek (1991) echoed White’s sentiment in asserting that models of marital stability and quality must include demographic as well as psychological variables that influence how marital interactions are appraised. A decade later, Huston (2000) identifies a widespread tendency in the marital literature to overlook the psychological, social, and environmental contexts in which interactions occur. As these researchers have pointed out, neglect of the role of the environment may render existing theories of marriage incomplete or even misrepresentative of the range and influence of factors that contribute to the course of marital relationships. Given repeated assertions that current theories have ignored the role of context, some may be surprised to learn that there are more than 30 extant models of context. However, there is considerable consensus on what elements theories of context must contain. The first of these theories, and the one to which most subsequent models owe an explicit or implicit debt, is Hill’s (1949) Crisis theory. As an army psychologist during World War II, Hill derived his Crisis theory through observations of families separated

PAGE 19

8 by the war. Hill’s original goal was to understand variability in the family stress response. That is, he aimed to understand why some families dealing with an accumulation of stressors fare better than other families who are facing fewer stressors. Across a wide variety of families, he noted that four elements appeared to hold the key to understanding variability in family outcomes. These elements give shape to Hill’s model, which holds that a stressor event (A) is followed by the family’s crisis meeting with its resources (B) which leads to the family’s interpretation of the seriousness of the stressful event (C), producing the amount of crisis and the response to the stressor (X). Thus, families with more resources, fewer vulnerabilities, and a view of the stressor as a manageable challenge are better equipped to overcome crises. All subsequent theories have refined or expanded on the same components first outlined by Hill: a contextual factor (i.e., a stressor), a set of resources and vulnerabilities, a variety of coping responses, and an outcome(s). In fact, because the components of post-ABCX theories of context can be mapped onto those outlined in the ABCX model, the ABCX model will be used as a conceptual scaffolding for the following review of the components of existing theories. Components of Current Stress Theories Context In Hill’s model (Hill, 1949), the contextual factor (“A”) is defined in terms of an external stressor event (e.g., World War II) that may induce a state of crisis. However, the broader definition of “context” ranges from specific stressors and other proximal variables to variables as distal as the historical milieu in which a relationship is embedded. Urie Bronfenbrenner (e.g., Bronfenbrenner, 1986), another World War II psychologist, has organized context into a series of levels ranging from those most

PAGE 20

9 immediate to very distal levels of context. Bronfenbrenner’s five levels of context, the micro-system, meso-system, exo-system, macro-system, and chrono-system, provide an apt structure for understanding how other theorists have thought about context at different levels. According to Bronfenbrenner, the micro-system forms the immediate context of our lives. For example, research that explores the specific adjustments made in response to unemployment (e.g., Vinokur, Price, and Caplan, 1996; Voydanoff, 1983) focuses on the level of the micro-system. The meso-system refers to an interaction between two settings, such as the work-family interface (e.g., Blair, 1993). The exo-system describes experiences in social settings in which the individual does not have an active role but which indirectly influence the more immediate family context. The effect of work-related stress on a child through punitive parenting (e.g., Conger, Conger, Elder, & Lorenz, 1992) is as an example of research focused on the exo-system level. At the next level, the macro-system refers to overarching patterns of culture such as a racially oppressive environment. For example, Peters and Massey (1983) focused their research on the level of the macro-system by exploring the pervasive, racially oppressive environment that defines the lives of many African Americans. Finally, the broadest level of context is the chrono-system, which includes each individual’s life span as it is bounded within a given historical period. An example of chrono-system focused research is Furstenberg’s (1994) provocative analysis of trends in divorce and remarriage across time. Furstenberg points out the possibility that as life spans have increased over time, the rise in the prevalence of divorce in the United States may be partly accounted for by the fact that “survival rates

PAGE 21

10 have placed a greater strain on the ability of couples to manage marital stress or maintain marital contentment” (p. 30). A second way to classify context is thematically. Within each of Bronfenbrenner’s levels, contextual elements may be physical, social, or symbolic in nature. Physical elements of context may include things like income loss, living in a dilapidated home, or living in an unsafe neighborhood (e.g., South, 2001). A social contextual element is exemplified in decreases in interaction due to retirement (e.g., Higginbottom, Barling, & Kelloway, 1993). Finally, symbolic elements of context are illustrated in a racially oppressive environment (e.g., Peters & Massey, 1983). That is, racism illustrates a symbolic element of context because, although it is not physically tangible, the pervasive experience of discrimination often goes along with minority status in a given society. Yet another key distinction can be made between positive and negative contextual elements. Positive contextual elements are those that increase resources or contribute positively toward well being. Having physical security and a supportive family are positive contextual influences, while living in a high crime neighborhood and having inadequate social support are negative contextual influences. In fact, as in the case of Hill’s ABCX model, theories of “context” do not describe context as a range of possible influences that may be both positive and negative. These models are actually models of stressful or negative contexts. Within the range of stressful contexts, further conceptual distinctions can be drawn. The most widely referenced distinction relates to the longevity of a stressor’s impact. That is, within the theories, many distinguish between chronic and acute stressors. Acute stressors are “time limited stressors” (Gump & Matthews, 1999) such as combat

PAGE 22

11 exposure, natural disasters, pre-doctoral comprehensive examinations, and physical relocation. According to this definition, most life events, which Wheaton (1997) defines as “discrete, observable eventswith a relatively clear onset and offset” (p. 52), would qualify as acute stressors. However, at present, there is no consensus regarding the definition of “time limited.” For instance, some researchers (e.g., McGonagle & Kessler, 1990) have considered stressors that have lasted up to one year to be acute. Broadly speaking, then, chronic stressors are stressors that last longer than acute stressors. McGonagle and Kessler (1990) have conceptualized chronic stressors as those in effect for at least one year and any ongoing difficulties that have taken place for longer than a year. In some cases, stressors are chronic despite coping ability, as in the case of terminal cancer, while in other cases, chronic stressors are the result of unresolved acute stressors, such as the unexpected hardship of long-term unemployment. A number of theorists have attempted to categorize chronic stressors. Wheaton’s (1997) fourfold classification includes ongoing role occupancy, role inoccupancy (e.g., not being in a desired job), stressors that result in unwanted social roles, and disadvantageous life circumstances. Gottlieb (1997) pointed out that Wheaton’s classification of chronic stressors has notable overlap with the classification developed by McLean and Link (1994), which includes life difficulties, role strains, the chronic strain of discrimination, and community strains (e.g., natural disasters). Many theorists have also been sensitive to the notion that individuals deal simultaneously with a number of stressors, both chronic and acute. McCubbin and Patterson (1982) refer to the “pile-up” of both chronic and acute stressors and Taylor and Aspinwall (1996) state that ongoing stressors of both types form the “backdrop” (p. 91)

PAGE 23

12 against which future stressors occur. Noting that definitions of chronicity vary widely, Gump and Matthews (1999) have chosen the “supraordinate term, background stressors,” (p. 469) to refer to the enduring events that serve as ground to the figure of acute stress events. Other underlying features of stressors have also been theorized to be important. For example, Hill (1964) has emphasized the severity and suddenness of events, as well as the amount of previous experience an individual has had with similar types of stress. In terms of predicting the extent of the impact of stress, it seems that each of these stressor qualities directly relates to the ability to “muster” various types of resources. Figley (1983) and McCubbin & Figley (1983) have generated a list of possible reasons why some stressors may have a larger impact than others. This list includes stressors for which an individual has having little or no time to prepare, slight previous experience, few sources of guidance, and stressors that have been experienced by few others, those which prompt feelings of helplessness, or those that involve an interminable time in crisis or a great sense of loss, disruption or destruction. Boss (1987) has also suggested that stressor events may be classified according to their desirability, source, intensity, or randomness. The desirability of a stressor relates to a previous distinction that has been made, the distinction between “bad” stress and “good stress,” sometimes called “eustress” (Edwards & Cooper, 1987). As stated previously, however, current theories of context do not typically focus on forms of eustress. The source of a stressor may be another important distinction. For example, a stressor’s impact might be more severe if the source of the stress is highly personal (e.g., divorce) in comparison to one that is impersonal (e.g., a sinkhole beneath parts of one’s

PAGE 24

13 house). This example also illustrates the higher likelihood that certain types of stressors, like having a sinkhole on one’s property, may heighten the sense that one is alone. The relative rarity of certain stressors and the stigma associated with some stressors, such as having AIDS, may also exacerbate the impact of a stressor. Lastly, domains of stress have also been used to distinguish stressors. Wheaton’s stress measurement survey features several life domains where stress may have an impact, such as money and financial matters, work, love and marriage, family and children, social life and recreation, crime and legal matters, residence, religion, and health (Wheaton, 1997, p. 71). The suggestion implied by distinguishing domains of stressors is that the impact of a stressor may vary for each individual as a function of the domain(s) that are affected by a stressor event. To summarize then, context has been conceptualized in a variety of ways. Bronfenbrenner organized context into levels, ranging from the most immediate, proximal influence of the micro-system, to the overarching historical context referred to as the chrono-system. Context may be conceptualized thematically in terms of physical, social, and symbolic properties. Contextual influences can also be divided on the basis of their general positivity or negativity, but current theories have focused rather exclusively on negative or stressful circumstances that may affect psychosocial outcomes. A number of theorists have distinguished between acute and chronic stressors, and other stressor qualities (e.g., severity and suddenness), as important. Finally, contextual stressors have also been categorized according to domains of impact. To conclude then, notions of context vary widely and defy simple conceptualization.

PAGE 25

14 Resources and Vulnerabilities Resources and vulnerabilities may be broadly defined as elements in peoples’ lives that facilitate or impede the process of managing stress. In Hill’s model, the “B” element refers to key resources and vulnerabilities that determine the family’s ability to deal with a stressor event. As in the ABCX model, all subsequent models suggest that individuals possess, at any given time, a collection of strengths and vulnerabilities. The primary distinction that has been made in discussions of resources and vulnerabilities is between those that are internal and those that are external to the relationship. The term “internal” is used to refer to resources and vulnerabilities possessed by individual partners in a relationship or those shared by relationship partners. To illustrate, problem solving ability and intelligence are internal resources and emotional instability and a history of violent behavior are internal vulnerabilities. Resources and vulnerabilities shared by partners in a relationship may also be considered internal. Relationship resources might include stable and high marital quality, spousal social support, and adequate shared income, while vulnerabilities might include marital distress and inadequate financial resources. By contrast, external resources and vulnerabilities are those located outside of the individual partners or the relationship. For example, available social support received from the community would be an external resource while living in a climate of constant racial oppression signifies an external vulnerability. As this example illustrates, there is potential for conceptual overlap between some types of resources and vulnerabilities and the context. The presence of racial discrimination not only qualifies as a vulnerability, but also seems inextricable from how some have thought about the notion of “context” (e.g., Peters & Massey, 1983). To deal with the overlap, most theories focus on stressor events

PAGE 26

15 as the contextual element and separate resources and vulnerabilities from specific stressor events on this basis. Distinctions have also been made between tangible, informational, and emotional resources and vulnerabilities (e.g., Taylor & Aspinwall, 1996). Tangible resources are those like income, goods, and other forms of material support. Informational resources, defined as “support involv[ing] the provision of specific information about a stressful event and resources for coping with it” (Taylor and Aspinwall, 1996, p.92) notably overlap with the “appraisal support,” defined by Taylor and Aspinwall (1996) as “helping an individual to understand the stressful event better and what resources and coping strategies might be mustered to deal with the event” (p. 92). Emotional resources are best illustrated in the concept of social support. There is a vast literature on the area of social support (see Sarason, Sarason, Brock, & Pierce, 1996, for a review). An alternative way of thinking about these distinctions is that tangible and informational resources are likely to be utilized in instrumental coping, coping that involves “actions oriented to removing or resolving the stressful demands,” while emotional resources are likely to be employed in the service of emotion-focused coping, coping that involves actions directed to the “management” of emotions produced by stressors (Gottlieb, 1997, p. 28). In sum, current theories consider the roles of different types of internal and external resources and vulnerabilities. Although there may be some conceptual overlap between resources and vulnerabilities and notions of context, most theorists have focused on stressor events as the main context while resources and vulnerabilities are described as qualities internal and external to the relationship that affect the ability to manage stressful events.

PAGE 27

16 Coping Coping has been defined as “the efforts of the person to manage (reduce, minimize, master, or tolerate) the internal or external demands of the person-environment transaction” (Folkman, Lazarus, & Gruen, 1986, p. 572). There has been a tremendous amount of research on coping. A full review of this extensive literature is beyond the scope of the present dissertation, but the reader is referred to Gottlieb (1997) and Wheaton (1997) for informative discussions of coping. For the present purposes, the focus will be on coping processes identified in theories of context. Hill’s “C” element was the family’s interpretation of the seriousness of the event. However, some (e.g., Burr and colleagues, 1994; Carver, Scheier, and Weintraub, 1989; Lazarus & Folkman, 1984) have argued that “C” may also be seen as an element of the coping process. Lazarus and Folkman (1984) refer to “cognitive control” (p. 171) as the ability to change one’s interpretation of a harmful event. Carver, Scheier, and Weintraub (1989) propose that positive reinterpretation is one of thirteen coping strategies. Burr and colleagues (1994) suggest that reframing an event is a form of cognitive coping. Although these distinctions may seem somewhat fuzzy, if “C” is taken in a broader sense to refer to coping processes, then “C” can be loosely mapped onto the third element that existing theories incorporate. The most common distinction in this literature is problem-focused versus emotion-focused coping. Problem-focused coping is “aimed at problem solving or doing something to alter the source of the stress” (Carver, Scheier, & Weintraub, 1989, p. 267). Lazarus and Folkman (1984) have also referred to this type of coping as “primary control, the attempt to change the environment” (p.170). Others have made reference to instrumental coping (e.g., Lazarus, 1977; Taylor, Jackson, & Chatters, 1997), a similar

PAGE 28

17 concept. Taking an ailing car to a mechanic is an example of problem-focused coping. As illustrated in this example, problem-focused coping is outwardly focused and action based and is theorized to occur in situations that are viewed as changeable or controllable (e.g., Carver, Scheier, & Weintraub, 1989). Conversely, emotion-focused coping is inwardly focused and reaction based. Carver, Scheier and Weintraub (1989) conceptualize emotion-focused as “aimed at reducing or managing the emotional distress that is associated with (or cued by) the situation” (p. 267). Emotion-focused coping fits with Lazarus and Folkman’s “secondary control,” defined as “the attempt to fit in with the environment” (Lazarus & Folkman, 1984, p. 171). Thus, emotion-focused coping is an attempt to accommodate to a stressor by reinterpreting the stressor’s meaning. For instance, emotion-focused coping might be the optimal response to a diagnosis of a terminal illness. As in this example, emotion-focused coping is theorized to be the dominant and optimal response when it appears that a situation cannot be changed (e.g., Carver, Scheier, & Weintraub, 1989). Particular types of coping can also be sub-divided on the basis of their nature as cognitive, behavioral, or emotional. The first of these categories, cognitive coping, involves cognitive reactions to stressor events, the forming of appraisals about the significance of a stressor in one’s life. Several stress theorists have highlighted the importance of cognitive appraisals. Lazarus (1977) suggests that the “psychological significance of injury” determined by cognitive appraisals may be even more important than “physiologically noxious effects in producing the adrenal cortical changes associated with stress” (p.147). Taylor & Aspinwall (1996) cite research to support the idea that perceptions of the controllability of a situation is a critical determinant of coping

PAGE 29

18 strategies. According to Williams (1995), the perceived stress threat as well as the stressor demand determines the overall level of stress experienced. In each of these cases, cognitive appraisals function as critical determinants of the magnitude of a stressor’s impact. Behavioral coping is most similar to the notions of problem-focused or instrumental coping in that it involves engaging in behaviors in order to reduce the negative impact of stressors. Pursuing counseling, making lifestyle changes after a diagnosis of diabetes, and creating a household financial budget are forms of behavioral coping. Finally, emotional coping, most similar to emotion-focused coping, is the third type of coping response. As previously discussed, attempts to change affective responses to a stressor represent emotion-focused coping. Some theorists have suggested that broad distinctions such as those discussed above are not sufficient to capture the most important distinctions in coping responses. As Carver, Scheier, & Weintraub (1989) point out, denial and positive reinterpretation are both forms of emotional coping, yet the implications for these actions on coping success varies widely. As such, a further distinction to be made is between adaptive and maladaptive coping. Two points have been expressed about the adaptability of coping responses. First, some responses are generally more adaptive than others in any situation. For instance, Klein, Forehand, Armistead, & Wierson (1994) distinguish between active coping strategies and avoidant coping strategies. Active coping strategies, which incorporate both problem-focused and emotion-focused coping, are defined as “overt behavioral attempts to deal directly with a problemand attempts to manage one’s appraisal of the stressfulness of an event” (p. 256). Alternatively, avoidant coping refers

PAGE 30

19 to “attempts to avoid actively confronting the problem or to indirectly reduce emotional tension by engaging in behaviors such as eating or smoking more” (p. 256). Evidence suggests that across situations, avoidant coping is associated with increased psychological distress (e.g., Bloom & Spiegel, 1984; Vitaliano, Katon, Maiuro, & Russo, 1989). Second, the match between coping strategies and the situation at hand may also determine the adaptability of coping responses. As mentioned previously, it is usually adaptive to employ problem-focused coping for controllable, changeable stressors. Likewise, it is usually adaptive to employ emotion-focused coping for uncontrollable stressors such as a terminal illness. Conversely, it may be maladaptive to employ problem-focused coping for uncontrollable stressors such as spinal cord injuries (e.g., Moore, Bombardier, Brown, & Patterson, 1994). Similarly, it may be maladaptive to focus primarily on changing one’s emotional response to a stressor when a problem can be resolved with action. Several theorists have also attempted to organize the multitude of specific coping responses into cohesive frameworks. As Burr and colleagues (1994) point out, lists of coping strategies (e.g., Caplan, 1964; McCubbin & Dahl, 1985) are common in the family science literature. Recently, after identifying a list of more than 400 ways of coping, Skinner, Edge, Altman, & Sherwood (2003) concluded that there is “little consensusabout how to conceptualize or measure the central concepts in the [coping] field” (p. 216). In response to the lack of “an integrated, coherent taxonomy”(Burr et. al., 1994, p. 131), Burr and colleagues developed a seven-layered conceptual framework that

PAGE 31

20 includes several potential coping strategies based in cognitions, emotions, relationships, communication, community, spirituality, and individual development. In summary, while Hill’s “C” referred to a family’s interpretation of the seriousness of a stressor, other theorists have included this notion in broader definitions of coping. Dimensions of coping include problem-focused versus emotion-focused coping, cognitive, behavioral, and emotional coping processes, and adaptive versus maladaptive coping. There is some consensus (e.g., Carver, Scheier, & Weintraub, 1989; Halford, Scott, & Smythe, 2000; Klein, Forehand, Armistead, & Wierson, 1994; Spangenberg & Theron, 1999) that some forms of coping (e.g., avoidant coping) are generally maladaptive while others (e.g., active coping) are generally adaptive. However, the match between the situation and a chosen response may, at times, determine the relative adaptability of coping behavior. Outcomes There are many possible outcomes to stress. There are individual outcomes such as depression and there are also relationship outcomes. The present review will focus on relationship outcomes. Hill focused on family outcomes in conceptualizing the “X” component of his model. In reference to the family, Hill uses the word “integrity” to describe cohesiveness in the family and a sense of economic and emotional interdependence. Although Hill’s “X” primarily refers to the impact of crises on family integrity, he clearly states that marital adjustment is the foundation of family adjustment (1949, p. 16). Like the larger family system, the marital relationship also has its own level of “integrity” as well as economic and emotional interdependence. In both the larger family context, and in the marital relationship, which is the focus of this review, the three

PAGE 32

21 relational outcomes affected by stress are relationship processes, feelings about the relationship system, and the status of the relationship as dissolved or intact. When stressful circumstances, arise, the first perceptible outcome is often changes in normal processes in the marriage. These changes can range from more abstract to highly specific processes. On a more abstract level, Hill talked about the ways in which family role patterns are disturbed by the stress of separation due to military duty. Specific processes that may changes as a result of stress might include the frequency and patterns of sexual relations (Hill, 1949) or the nature of interactions between a couple. For example, Liker and Elder (1983) found a marked increase in conflictual interactions and higher levels of expressed hostility in marriages affected by the Great Depression. Repetti (1994) obtained similar findings in a sample of air traffic controllers who were more likely to have conflictual interactions with their children following a stressful day at work. Some of the theories in the present review focus on these kinds of outcomes, referred to variously by terms such as marital tension, marital harmony, and marital adversity. These relationship process outcomes may be assessed using self-report questionnaires, but are frequently assessed with affect rating systems coded by trained raters. When the nature of relationship processes is negatively impacted by stress, feelings about the relationship often shift in response. Hill states that most crises sooner or later come to involve a de-morale-ization [sic] (Hill, 1949, p. 10) which he defines as a loss of family morale and unity. The same concept could be applied to a marital system; as marital interactions become less warm and more hostile, feelings about the marriage are likely to change. Current theories characterize these feelings as subjectively assessed

PAGE 33

22 outcomes like perceptions of relationship satisfaction, relationship adjustment, and relationship quality. Such outcome variables are critical in that they draw on the “extensive and special knowledge [a relationship partner] has about the relationship” (Berscheid, Snyder, & Omoto, 1989, p. 67). However, these subjectively assessed outcomes are also vulnerable to memory-based distortions or social desirability motivations (e.g., Reis and Gable, 2000). Furthermore, it is likely that participants may not be evaluating their relationships using a common standard. So, while these “insider classifications are undoubtedly useful, what they predict is not entirely clear” (Berscheid, Snyder, & Omoto, 1989, p. 68). Finally, relationship status variables are the third type of outcome. Relative to feelings about the relationship, these status outcomes are generally seen as more objective and may include measures of relationship stability such as marital status over time or thoughts of divorce (e.g., Booth, Johnson, White, & Edwards, 1986). The advantages of focusing on this type of outcome are less susceptibility to confounding influences, reliance on fewer inferences, and potentially greater face validity and generalizability (Carlson, 1995). However, the lack of consensus related to definitions of marital stability is also a notable disadvantage in focusing on marital status as an outcome. As Karney and Bradbury (1995) have pointed out, some have used stability to refer to partners’ evaluations of their risk for separating (e.g., Booth, Johnson, White, & Edwards, 1986), while others have defined stability as the state of legal union over time (e.g., Rockwell, Elder, & Ross, 1979). Another unique issue related to the measurement of marital stability is whether outcomes are viewed from an individual or dyadic perspective. In some cases, outcomes that are detrimental to the marriage may be

PAGE 34

23 personally beneficial to one or both of the individuals in the marriage (e.g., Eldridge, Lawrence, & Christensen, 1999). For instance, “[marital] harmony can also be achieved at one spouse’s expense, as when a spouse abandons goals that conflict with those of the partner” (Levinger & Huston, 1990, p. 25). In sum, outcomes are the final element common to contextual theories. Relationship outcomes highlighted by current contextual theories span a wide range, reflecting a larger lack of consensus on how to measure marital outcomes in the field of marital research (Karney & Bradbury, 1995). However, most relationship outcomes addressed by the theories appear to fall under three categories: relationship processes, feelings about the relationship, and marital status outcomes. Summary of the Components of Stress Theories The elements highlighted in current theories of stress are contextual factors, resources and vulnerabilities, coping processes, and outcomes. Although there is widespread disagreement about how to best define these elements, since the formation of Hill’s (1949) Crisis theory, no subsequent theory has articulated any new element not included in Hill’s theory. As such, it is possible to map the components of current theories onto the original elements proposed by Hill. The fact that mapping occurs with relative ease suggests a basic structure to the theories. In general, the theories consistently suggest that contextual factors are weighed against resources and vulnerabilities, coping involves an attempt to reduce the negative impact of stressor events, and differential outcomes reflect disparities in the effectiveness of various coping strategies. Future growth, then, may not entail coming up with new elements but in increasing refinement of the way we understand the interactions between elements. As Perry-Jenkins, Repetti, & Crouter (2000) have articulated, “we need to identify key moderating

PAGE 35

24 variables that shape the process by which stress is transmitted” (p. 989). In light of this, the next section will review principles in the theories that suggest associations between elements in the theories.

PAGE 36

CHAPTER 3 THEORIZED ASSOCIATIONS BETWEEN COMPONENTS OF STRESS THEORIES Direct Associations between Stress and Physical and Psychological Functioning Some theorists suggest that stress has a direct impact on individual functioning. Evidence suggests that minor stresses may compromise psychological and physical functioning on their own (e.g., Kohn, Lefreniere, & Gurevich, 1990; Lazarus, DeLongis, & Folkman 1985). Similarly, chronic strain may compromise psychological and physical functioning (Brown & Harris, 1978; Gannon & Pardie, 1989; McGonagle & Kessler, 1990; Pearlin & Schooler, 1978). As early as 1918, Janis (as cited in Janis, 1993) described how the fight-or-flight response generated by stressful situations leads to maladaptive hyper vigilance, defensive avoidance and constricted decision making abilities. More recently, rodent studies have directly linked chronic stressors to tumor growth and poor immunological functioning (Kiecolt-Glaser, Kennedy, & Malkoff, 1988). In human samples, Kiecolt-Glaser and colleagues have repeatedly discovered that marital discord produces a variety of negative autonomic, endocrinological, and immunological changes such as elevated cortisol levels and poor immune response to mitogens and latent viruses (Kennedy, Kiecolt-Glaser, & Glaser, 1988; Kiecolt-Glaser, 1999; Kiecolt-Glaser, Fisher, Ogrocki, Stout, Speicher, & Glaser, 1987). Although this research shows that stress may have a direct impact on psychological and physical functioning, the impact of stress is often moderated or mediated by intervening variables. The following three principles reflect proposed or potential 25

PAGE 37

26 associations to describe the moderation or mediation of the impact of stress on marital functioning. Resources and Vulnerabilities Moderate the Association between Stress and Marital Functioning Originally, Hill (1949) theorized that the resources of the family determine whether a stressor event will cause a state of crisis (p. 9). Several others (e.g., Booth & Amato, 1991; Broman, Riba, & Trahan, 1996; Burr, 1982; Cohen & Edwards, 1989; Elder, Conger, Foster, & Ardelt, 1992; Ensel & Lin, 1991; Lazarus and Folkman, 1984; Leiter & Durup, 1996; Liker & Elder, 1983; Mechanic, 1998; Pearlin, Menaghan, Lieberman, & Mullan, 1981; Taylor & Aspinwall, 1996; Wheaton, 1985; Williams, 1995) have also agreed that a variety of internal and external resources and vulnerabilities moderate the impact of stressors. Hobfoll aptly summarizes the theorized associations between resources and vulnerabilities and stressors: “if individuals have the requisite resources, they will be able to withstand or overcome stress, or even prevent it from occurring” (1998, p. 92). As expressed by Hobfoll, resources and vulnerabilities may either directly or indirectly influence stress outcomes. Direct buffering of stress would occur when resources prevent exposure to stressful situations. For instance, living in a safe neighborhood reduces exposure to threats of physical harm. Other positive influences like social support may also directly buffer against stress (Wheaton, 1985). In the same way, vulnerabilities may increase susceptibility to stress directly. To use the same example, living in a low-income neighborhood means more exposure to stressors (Glass & Singer, 1977; Graig, 1993). Similarly, working in conditions of low autonomy and flexibility increases exposure to stressors (Mauno & Kinnunen, 1999; Perry-Jenkins, Repetti, &

PAGE 38

27 Crouter, 2000; Schaefer & Moos, 1993). As shown by these examples, resources and vulnerabilities can directly exacerbate or ameliorate the impact of stressors. There are several mechanisms through which resources and vulnerabilities may indirectly affect the impact of stressors. First, resources may buffer against stress by influencing well-being related behaviors that prevent stress from occurring in the first place. In this way, resources function as stress deterrents (Wheaton, 1985) with a behavioral mediator(s). For example, the emergence of health-related stresses may be deterred for people whose social circles actively promote a healthy lifestyle (e.g., Taylor & Aspinwall, 1996). In terms of marital stress, individuals whose social circle is supportive of their relationship may benefit from indirect buffering. Conversely, vulnerabilities may induce higher levels of stress by influencing negative behaviors (e.g., Taylor, Repetti, & Seeman, 1997). To use the same type of example, higher stress levels may be promoted when individuals exist in social “networks that promote destructive behavior such as alcohol dependence and delinquency” (Belle, 1982, p. 92). Consistent with a marital health example, the presence of divisive social influences such as frequent invitations to Las Vegas strip clubs may create additional stress in a marriage. Vulnerabilities may also make it harder for partners to make positive contributions to their relationships. To illustrate, consider the finding that PTSD following combat exposure made it harder for husbands to contribute to their marriages through educational attainment, steady employment, and a stable income (Gimbel & Booth, 1994). Second, resources and vulnerabilities may affect stress outcomes by influencing the ability to cope with stressors a bi-directional manner. On the one hand, internal resources and vulnerabilities may affect external coping resources (e.g., Boss, 1983; Leiter &

PAGE 39

28 Durup 1996). For instance, internal vulnerabilities such as low emotional intelligence may result in fewer external coping resources (e.g., a less supportive social network). On the other hand, the internal resource of having an attractive personality may help some individuals muster more social support in stressful times. In this example, having one type of resource (an attractive personality) increases another kind of resource (available social support). However, in some cases, having one type of resource may be associated with fewer resources of another kind. An illustration of this point is the finding that wives’ employment has been linked to declines in physical health as well as contentment with marriage, work and life in general for husbands (Burke & Weir, 1976b). So, although wives’ employment supplements family income, a physical resource shared by the couple, there is a commensurate decrease in services previously performed by the wife on behalf of the family and this may have ripple effects in the opposite direction. Resources and vulnerabilities may also affect personal coping abilities, which then influence the amount of stress experienced by individuals. For example, possessing adequate resources may influence secondary appraisals of stressor events, making them seem more manageable (e.g., Lazarus & Folkman, 1984; Taylor & Aspinwall, 1996). In like manner, vulnerabilities may affect stress outcomes by decreasing the ability to cope with stressors. The vulnerability of personal instability, for instance, decreased participants’ ability to cope with the financial devastation of the Great Depression (Liker & Elder, 1983). Essentially, then, resources and vulnerabilities internal and external to the marriage may potentially exacerbate or ameliorate each other. A number of counter-intuitive possibilities may also describe how resources and vulnerabilities may increase or decrease the impact of stressors. In at least three cases,

PAGE 40

29 resources may have a reverse buffering effect. First, apparent resources may be transformed into neutral or even detrimental forces with the passage of time. This possibility is illustrated in the finding that in a crowded house, the buffering effect of social support was strong at first but disappeared after eight months (Lepore, Evans, & Schneider, 1991). As Taylor and Aspinwall (1996) point out, the duration of the stressor may moderate this transformation of resources. An underlying issue, then, is whether apparent resources actually serve the function of stress buffering. This issue is particularly relevant to discussions of the social networks of underprivileged groups. While support networks offer advantages such as having relatives nearby who can help out with personal and financial problems, networks also have “prices and disadvantages” (Peters & Massey, 1983, p. 209). Belle’s (1982) work is informative in regard to a host of potential problems associated with poverty-based support systems. According to Belle, social support networks can lead to overloads of responsibility and heightened distress, exposure to a high level of ‘contagion of stress,’ (p. 94), and the preclusion of upward mobility due to the redistribution of any newly acquired resources throughout the network. A second way in which resources may increase stress is through cognitive expectations. An individual may have an extensive social network, which can clearly be a resource. Yet, if an individual has expectations about how and when support will be offered, it is possible that support provision will not meet expectations. Discrepancies between desired social support and subsequent perceptions of the social support received have been noted with some frequency in marital interactions observed in laboratory

PAGE 41

30 settings (e.g., Dehle, Larsen, & Landers, 2001; Wethington & Kessler, 1986). In this case, what is usually a resource indirectly accentuates the impact of stress. Even support that meets expectations may be detrimental in some cases. Bolger and colleagues have conducted research demonstrating that the awareness of receiving social support carries emotional costs. In fact, according to multiple studies cited by Bolger and colleagues (e.g., Barrera, 1986; Bolger, Foster, Vinokur, & Ng, 1996; B. B. Brown, 1978; Dunkel-Schetter & Bennett, 1990; Eckenrode & Wethington, 1990; Lieberman, 1986; Wethington & Kessler, 1986), “support receipt is either associated with poor adjustment or, at best, leaves the recipient no better off than if support had not been received” (Bolger, Zuckerman, & Kessler, 2000, p. 953). Two provocative mechanisms may account for this counter-intuitive finding. The authors indicate that one possibility is that people’s well-intentioned efforts to provide social support may fail or make matters worse for the person under stress (Coyne, Wortman, & Lehman, 1988). A second possibility they consider is that the awareness of accepting social support entails a loss of self-esteem on the part of the recipient (e.g., Fisher, Nadler, & Whitcher-Alagna, 1982; Shapiro, 1978). Although it is yet unclear why receiving support may be ineffective in reducing distress, the phenomenon conveys the message that apparent resources may not always be what they seem. Finally, just as resources and vulnerabilities influence the impact of stressors, stressors exert a reciprocal influence on resources and vulnerabilities (e.g., Taylor & Aspinwall, 1996). The links forging this reciprocal association are the same as those previously discussed. Stress may directly weaken internal resources such as feelings of self-efficacy. Stress may also diminish various external resources, such as social support

PAGE 42

31 (Wheaton, 1985). For example, the expression of high levels of distress may cause significant others to withdraw support (e.g., Bolger, Foster, Vinokur, & Ng, 1996). Stress also affects coping abilities which may then decrease resources or exacerbate vulnerabilities. A more detailed description of this association will be featured in the discussion that follows. In summary, then, the ways in which resources and vulnerabilities moderate the impact of stress are extremely complex. Both resources and vulnerabilities may directly and indirectly influence the degree to which stressors lead to varying outcomes. Moreover, even the most intuitive notions such as the idea that resources will decrease stress may not always be true. In some cases, the positive benefits of resources may level off after time and in other cases, what are considered resources (e.g., a readily available social support network) may also be considered vulnerabilities in some contexts, such as when such support entails physical and emotional costs. Thus, extensive, focused future effort will be required to tease apart the moderating conditions under which resources and vulnerabilities will increase or decrease the impact of stressors. Coping Processes Mediate the Association between Stressors and Outcomes Hill (1949) was the first to assert that the definition a family makes of a stressful event determines whether it becomes a crisis or not (p. 9). Several others (e.g., Boss, 1987; Lavee, McCubbin, & Olson, 1987; Lazarus & Folkman, 1984; McCubbin & McCubbin, 1989; Scheier, Weintraub, & Carver, 1986; Taylor & Aspinwall, 1996; Wheaton, 1985; Williams, 1995) have agreed that coping processes mediate stress outcomes. As in the case of previous associations of elements, coping may mediate the impact of stress in both direct and indirect ways.

PAGE 43

32 The direct mediation hypothesis is that poor coping leads to poor outcomes and good coping leads to good outcomes. Hill’s early work provides a good example of direct mediation. Hill (1949) reported that women who coped with war-induced separation by seeking other men as substitutes for their husband adjusted least well (p. 144). More recent work also supports the direct mediation hypothesis. In several studies, Conger and colleagues (e.g., Conger, Elder, Lorenz, Conger, Simons, Whitbeck, Huck, & Melby, 1990; Conger, Ge, & Lorenz, 1994; Conger, Reuter, & Elder, 1999; Matthews, Conger, & Wickrama, 1996) have shown that “economic hardship has a significant influence on marital quality through its effect on irritable or conflictual behaviors within the marital dyad” (Conger, et al., 1990, p. 645). One hypothesis involving indirect mediation involves cognitive mechanisms integral to the coping process. That is, coping may influence the amount of crisis perceived or objectively present (e.g., Burr, 1982; Lavee, McCubbin, & Olson, 1987; Taylor & Aspinwall, 1996). Bandura (1997) has argued that coping success forms a feedback loop that determines self-efficacy beliefs in relation to a given stressor. These self-efficacy beliefs are thought to predict both the success and perseverance of future coping efforts. Taylor and Aspinwall (1996) have suggested that perceptions of controllability may be particularly important in determining the ways in which coping may produce differential stress outcomes. Finally, coping may also affect concurrent and future resources, which in turn may affect outcomes. Bodenmann has shown that different coping strategies may impact the course of a marriage (e.g., Bodenmann, 1997a; Bodenmann, 1997b). An intervention developed by Bodenmann and colleagues (e.g., Bodenmann, Charvoz, Cina, & Widmer,

PAGE 44

33 2001) suggests that accurate problem appraisal, problem-solving skills, and skillful communication of stress-related information are among the strategies that may buffer a marriage from detrimental impacts of stressors. The use of these adaptive strategies is likely to contribute to relational self-efficacy, which serves as a future “internal” resource. One benefit attained from the resource of relational self-efficacy has been articulated by Karney, McNulty, & Bradbury (2000); spouses with low relational efficacy are generally more likely to make dispositional attributions for their partners’ unexpected negative behaviors when compared to spouses high in relational self efficacy. To summarize then, coping mediates in both direct and indirect ways to produce outcomes. Coping may also affect current or future resources and vulnerabilities, which may in turn affect outcomes. In a general way, past and current coping attempts continually set the stage for future coping. In these ways, then, coping is integral in determining short and long-term outcomes of stressors. Outcomes Moderate the Impact of Future Stressors Hill’s ABCX model has been criticized as a relatively linear, static model (e.g., Burr et al., 1994; McCubbin & Figley, 1983; McCubbin & Patterson, 1983). At first glance, it is evident that Hill’s model does not graphically represent the dynamic feedback loops later added by McCubbin and Patterson (1983). However, in the text that accompanies the presentation of the ABCX model, it is clear that Hill believed in a dynamic interaction between outcomes and the impact of future stressors. Specifically, Hill states, “successful experience with crisis tests and strengthens a family, but defeat in crisis is punitive on family structure and morale” (Hill, 1949, p. 16). Following Hill, several others have echoed the idea that outcomes moderate the impact of future stressors (e.g., Burr et al., 1994; McCubbin & Figley, 1983; McCubbin & Patterson, 1983; Taylor

PAGE 45

34 & Aspinwall, 1996). As in the case of previous associations, outcomes may influence the impact of future stressors in both direct and indirect ways. First, outcomes may directly create or prevent future stressors. As McCubbin and Patterson (1983) have stated, the consequences of a family’s efforts to cope with past stressors directly determine future stress pile-up. For instance, ineffective coping prolongs the duration of a stressor’s impact. When acute stressors become chronic stressors, the baseline allostatic load for a couple is increased. “Allostasis” refers to an adaptive response geared towards maintaining stability, or homeostasis in the body. Allostatic load, then, is the price that the body has to pay for either doing its job less efficiently or being overwhelmed by too many challenges (e.g., McEwen, 1998). When future stressors occur, then, they occur against this new backdrop of a greater ongoing stress load. Second, outcomes may also indirectly determine the impact of future stressors by affecting levels of resources and vulnerabilities, which then exacerbate or ameliorate the impact of future stressors. To illustrate, poor marital outcomes may induce more stress from within the relationship or exacerbate environmentally emergent stressors. Conversely, a good marital relationship may buffer against future stressors. Outcomes also reinforce coping patterns that will mediate future stress responses. A maladaptive coping pattern like blaming one’s partner for a stressor may make one partner in the relationship feel better temporarily, but such a tendency will tend to harm a marriage in the long run (e.g., Sabourin, Lussier, & Wright, 1991). Conversely, having a mastery experience may make a couple feel stronger, more solidified, and more efficacious in the face of future stressors (e.g., Bandura, 1997).

PAGE 46

35 In sum, outcomes moderate the impact of future stressors in both direct and indirect ways, suggesting a dynamic relationship between outcomes and future events. Summary of Associations in Stress Theories and Justification for Empirical Review To broadly summarize, there is a pattern of complexity and no consistent direction of associations between elements in current stress theories. In fact, either directly or indirectly, resources and vulnerabilities, coping processes and outcomes may interact with each other to buffer or exacerbate stress, to strengthen or diminish external or internal resources and vulnerabilities, to enhance or impair coping ability, to produce given stress outcomes, and to determine a couple’s level of threat or resilience in the face of future stressors. The complexity of associations between elements brings to mind the complexity of the study of relationships in general. As Reis and colleagues have articulated, “relationships are embedded in a social environmental system, a physical environmental system, and larger societal and cultural systems [and] all of these systems are simultaneously evolving and influencing each other” (Reis, Collins, & Berscheid, 2000, p. 847). The preceding review of theory has revealed both clarity and consensus about the critical elements to consider in exploring associations between stressors and marital outcomes as well as confusion about how these elements affect each other to account for the ways in which context affects marriages. The first step in resolving these theoretical issues will require assembling the empirical literature and summarizing what has already been established, as well as identifying any issues and variables that may have been overlooked. Thus, the following section will focus on the empirical literature. First will be a general overview of the scope of current research. Next, I will describe the methods and selection criteria used to locate and assemble the set of empirical articles. After

PAGE 47

36 describing the search methods and article selection criteria, I will discuss the domains of stress that have been studied, and the sample and design characteristics of the review articles. Finally, I will outline the strengths and limitations in the current literature on stress and marital satisfaction.

PAGE 48

CHAPTER 4 REVIEW OF THE EMPIRICAL LITERATURE OF ASSOCIATIONS BETWEEN STRESS AND MARITAL SATISFACTION General Overview Pioneering work dating back to the early 1930s which looked at associations between stress and marital outcomes focused on economic strain (Angell, 1934; Bakke, 1940; Cherlin, 1979; Komarovsky, 1940; Zill, 1978 as cited in Zill & Nord, 1994). Consistently, these researchers found that higher levels of economic strain put some couples at greater risk for divorce. In the 1950’s, Terman’s classic longitudinal study of gifted individuals suggested that stressful life circumstances may be associated with lower levels of marital quality and stability. The Great Depression of the 1930s in America gave two researchers (Liker and Elder, 1983) a unique and unprecedented opportunity to intensively study the effects of a sweeping historical change on marriages and families. These researchers found many provocative stress-generated outcomes such as more conflict, more hostility, and less warmth in marital interactions, lower levels of satisfaction within marriages, and higher rates of divorce. Others have also obtained similar results using broad historical events such as the aftermath of the Vietnam War (Riggs, Byrne, Weathers, & Litz, 1998) and the Midwest farm crisis (Johnson & Booth, 1990). Currently, researchers in the United States and abroad continue this line of study in populations experiencing such diverse stressors as high-pressure working conditions (Repetti, 1994), economic deprivation (Conger, Rueter, & Elder, 1999), negative life 37

PAGE 49

38 events and normative transitions (e.g., Cohan & Bradbury, 1997), work-family conflict (Doumas, Margolin, & John, 2003), children’s health problems (Dahlquist, Czyzewski, & Jones, 1996), severe illness in one spouse (Giese-Davis, Hermanson, Koopman, Weibel, & Spiegel, 2000), sudden and unexpected unemployment (Larson, Wilson, & Beley, 1994), forced relocation during wartime (Hraba, Lorenz, & Pechacova, 2000), Post-traumatic stress disorder following active military duty (Riggs, Byrne, Weathers, & Litz, 1998) and interpersonal strain generated by relationships with in-laws (Proulx, Helms, & Payne, 2004) . The following review will give some structure to these results and produce some meaningful generalizations of current findings related to stress and marital satisfaction. Throughout the following review and meta-analysis, the focus will be on marital satisfaction as opposed to marital stability. Marital satisfaction is being defined according to Fincham and Bradbury’s description of “an individual’s global evaluation of his or her marriage” (1987, p. 800). A similar but distinct variable, marital stability, an indicator of marital status, has been “variously operationalized as the absence of divorce, the absence of separation or divorce, and the reports of intact couples that have not considered separation or divorce” (Karney & Bradbury, 1995, p. 16). Although satisfaction and stability have been meaningfully linked (e.g., Kurdek, 1993), the decision to divorce entails a variety of factors beyond global evaluations of the marriage. As Rusbult (1980) has pointed out, commitment to a relationship is not merely a function of perceived relationship quality but also involves the quality of the best available alternative [to the relationship] and the magnitude of the individual’s investment in the relationship. Or, as Udry (1981) explains, marital alternatives are a “separate dimension which predicts

PAGE 50

39 disruption independently of the prediction made by marital satisfaction” (p. 890). In like manner, then, the decision to remain married does not indicate a healthy relationship. Barriers to divorce such as financial dependency or a lack of viable alternative partners may dampen the will of those who would otherwise wish to divorce. Ultimately, as described previously, a basic premise of this dissertation is that living in the context of an unsatisfying relationship produces ill effects for spouses and their children (e.g., Bloom, Asher, & White, 1978; Grych & Fincham, 1990). Therefore, this dissertation will focus on marital satisfaction as opposed to marital stability. Sample of Studies Assembling the articles included in the current review and meta-analysis required the use of multiple strategies. The first stage of the process involved a Psychological Literature (PsycLIT; 1887-2003) search of over 3,000 articles identified by the keywords marital satisfaction, marital quality, or marital adjustment and stress, stressful, stressful event, strain, impact, illness, financial, job insecurity, job loss, unemployment, socio-economic status, neighborhood, environment, context, contextual, resources, income, military, and war. Three hundred thirty of these articles, focused on associations between stress and marital satisfaction, were photocopied and inspected closely by the author. The second phase of the search process involved a careful scan of approximately 220 pages of references of a sub-set of empirical or review studies within the first 330 articles, resulting in the identification and photocopying of an additional 78 articles for further inspection. The third phase of the article search occurred six months later. A second search of PsycLIT was performed using only the broad keywords “marital satisfaction,” “marital quality,” or “marital adjustment” for the studies published from 2002-2004. This search

PAGE 51

40 of the recent literature yielded 13 new articles for closer review. At the same time, both MEDLINE and Sociological Abstracts were searched using the broad keyword terms “marital,” “marriage” and “satisfaction.” The MEDLINE search yielded an additional 95 articles and the Sociological Abstracts search resulted in the closer inspection of an additional 22 articles. Finally, the fourth phase involved another broad search of PsycLIT, MEDLINE and Sociological Abstracts using the keywords “marital” “marriage” for any additional article published between 2004-2005. This search of the recent literature yielded 70 new articles for closer review, 40 of which were included in the empirical review with a subset of 23 of these 40 articles also used in the meta-analysis. In total, for all four phases of the search process, 608 articles were photocopied and closely inspected as potentially relevant to the current review and/or meta-analysis and 225 of these articles were retained for inclusion in either just the review or both the review and the meta-analysis. Selection Criteria The following criteria were used to select studies for inclusion in the review and/or meta-analysis. 1. Articles written in non-English languages were excluded. 2. The search was confined to articles in refereed journals and books. Thus, I excluded dissertations and unpublished manuscripts. 3. I only included studies that were empirical. Thus, descriptive studies and reviews were eliminated. 4. Only those studies that specified a self-reported assessment of marital quality were included. Typically, the names for these outcomes varied across studies (e.g., marital satisfaction, marital happiness, marital harmony) but all selected outcomes were assessed by partners in the marital relationship. Thus, studies that solely featured marital status outcomes such as divorce rates, marital disruption, or marital

PAGE 52

41 dissolution or other possible outcomes such as physical or psychological adjustment to stress were excluded. 5. I only included those empirical studies that statistically analyzed the association between an external stressor (e.g., job loss, illness in one’s spouse) and a marital outcome. External stressors were conceptualized as stressors that are external to each individual in the couple or family system, and the relationship interactions they share. For example, depression, parenting stress, and marital conflict, although stressful, were not considered to be external stressors. Thus, studies that did not operationalize and analyze an external stressor were not included. 6. Finally, I excluded studies that reported the same analyses using the same data set in more than one publication. Now that the exclusion criteria have been specified, the following section will describe the process for coding domains of stress studied in the existing literature. Coding Domains of Stress Studied in Samples in the Review Stressors studied in the review articles were first identified by the graduate investigator as falling within eleven possible domains, represented in Table 4.1. Next, three separate undergraduate raters recruited through the University of Florida honors program were instructed to assign the stressors into one of these eleven discrete a priori categories. Raters were also told that multiple stressors may be measured within the same study but that the same stress measure could not simultaneously be assigned to more than one category. For example, one study looked at how life threatening illness, financial strain, and race correlate with marital satisfaction. So, this study was coded as including three associations between stress and marital satisfaction. Twenty-five of the 225 total articles were used in the coding training sessions. First, the research assistants coded ten articles, after which a training session followed. In this training session, the extent of coding consensus was assessed and discrepancies were discussed. At this point, research assistants were specifically instructed to differentiate between work stress and work-family stress as follows: work stress stems from

PAGE 53

42 conditions or problems located solely in the work environment (e.g., high job pressure) while work-family stress refers to tension between work demands and home demands (e.g., work-family conflict) or strain due to the need to fulfill simultaneous roles at home and work (e.g., mothers who are working and breastfeeding at the same time). The research assistants also had some difficulty distinguishing between economic stress and low-SES stress. They were instructed to err in the direction of assigning these variables to the economic stress category, provided that the stressor was clearly financial in nature. This included variables with terms such as “financial strain,” “poverty,” and “perceived financial stress” as economic stressors. Otherwise, if a variable referred to SES-level, then this variable was coded as a “low-SES stressor.” I also reminded the coders that any variables that were general in nature and did not fit into a specific category (e.g., typically those with names such as “life events,” “stressful life events,” “daily stressors,” “daily hassles,” “perceived stress scale,” and “life transitions”) should be coded as General/Life stress. Based on this discussion, consensus was reached on these ten articles, and these consensus codes were retained for the study database. Following this training session, the research assistants coded ten more articles, and agreement was again checked and followed by a discussion of discrepancies. At the time of this training session, agreement was very good, but there persisted some difficulty in distinguishing between work and work-family stressors. Consensus codes again contributed to the codes retained for the study database. After this session, the research assistants coded a final five articles and then agreement was assessed again. There was complete agreement on the coding for these five articles.

PAGE 54

43 Following the training sessions, the coders were instructed to code the remaining articles. Based on the remaining codes, representing 200 articles, an intra-class correlation coefficient of approximately .95 was computed, indicating a high level of reliability between coders. In general, based on the discussions during the training sessions, it appeared that certain types of stress were more easily identified and coded relative to other types. Specifically, from the start of the training sessions, there was perfect consensus on the coding of health problems in one partner in a marriage, health problems in a child, war/military stress, minority status, unemployment stress and interpersonal stress. On the other hand, work stress was somewhat difficult to differentiate from work-family stress, and low-SES stress was similarly difficult to distinguish from economic stress. It is also important to note that because some studies explored associations between multiple stressors and marital satisfaction, the total sum of 356 associations between stress and marital satisfaction in Table 4-1 includes overlap in the case of studies that looked at multiple stressors. Description of Characteristics of the Empirical Literature Domains of Stress Studied in Samples in the Review Spousal health problems, the most common type of stress studied, varied considerably, including such conditions as breast cancer, testicular cancer, myocardial infarction, Alzheimer’s disease, migraine headaches, renal disease, kidney disease, multiple sclerosis, rheumatoid arthritis, chronic lower back pain, head injury, spinal cord injury and burn trauma. Work stress, conceptualized as problems restricted to the work domain, included lack of desired job autonomy, working long hours, undesirable amount of required work

PAGE 55

44 travel, job burnout, and the high-risk associated with positions such as being a police officer. Economic stress, often referred to as “financial strain,” included impoverishment and a low income-to-needs ratio. Several sub-categories were subsumed under the category of general stress: daily stress, life stress, stressful life events, transitions, and life hassles. Clearly, there is likely to be great variance in such a broad category. Work-family stress, investigated with slightly less frequency than general stress, differs from work stress in that work-family stress occurs at the interface of the family and work settings. Specific examples of work-family stress include inter-role conflict and role overload due to contradictory pressures coming from the work and family domains. As in the case of spousal health problems, the category of children’s health problems involved wide variability. Children’s health problems included such varied issues as the death of a child, having a profoundly mentally retarded child, or medical conditions such as cancer or spina bifida in a child. Stress associated with low socioeconomic status was explored with about the same frequency as children’s health problems. Low SES was either directly reported by study authors or was inferred based on an extremely low mean sample income. For example, one study pulled its participants from a local state unemployment office in the Pacific Northwest. The average income for males in the sample was $7925, considerably lower than the national average of $13,113 for males in 1978. As such, the men in this sample were viewed as being of low socioeconomic status.

PAGE 56

45 Minority status, a proxy for the potential experience of racial discrimination, was explored with the same frequency as low SES-based stress. Race-based stress included both the stress of being a minority in a majority culture, and the stress of an inter-ethnic marriage. Relative to the preceding stress domains, war or military stress was measured somewhat less frequently. This category included stressors such as the deployment of a spouse during the Gulf War, the after-effects of exposure to war trauma, and forced relocation of families due to acts of war. Associations between unemployment and marital satisfaction were assessed with slightly less frequency than war/military stress and race-based stress. Unemployment was due to a number of reasons, but was most commonly due to large-scale layoffs. Finally, interpersonal stress was measured least often. Interpersonal stress was conceptualized as relational stress originating from outside of the primary dyadic pair or the nuclear family being studied. The most common example of interpersonal stress was stress due to poor relationships with in-laws or extended family. To summarize then, eleven different domains of stress have been studied with varying frequency. Spousal health problems, work stress, and economic stress have been studied most frequently, and interpersonal stress, unemployment stress, and war/military stress have been studied least often. Some domains, such as spousal health problems and unemployment stress, are more conceptually homogeneous, while others, such as general stress and war/military stress are more conceptually heterogeneous. The eleven domains in Table 4-1 include stressors of many different types. For example, there are acute (e.g., sudden unemployment) and chronic stressors (e.g., low socio-economic status), relatively

PAGE 57

46 rare (threat of relocation for Israeli couples living in th e Golan Heights) and normative stressors (work-family stress), and severe (diagnosis of can cer in a child) or less severe stressors (daily hassles). So me domains of stress may have more stressors of certain types, and to the extent that this is the case, these differences will be brought to bear in understanding results of the meta-analysis to come. Table 4-1. Categories of Stress Measured in Existing Studies Stress Category No. of studies measuring this type of stress Spouse’s health problem 79 Work stress 65 Economic stress 58 General stress/Life stress 36 Work-family stress 32 Child’s health problem 19 Low-SES stress 16 Minority status 16 War/Military stress 15 Unemployment stress 13 Interpersonal stress 7 Coding Characteristics of Studies in the Review As opposed to the coding of domains of stress, a number of sampling and design characteristics were coded solely by the grad uate investigator. These dimensions were judged to be “low inference codings,” define d as “information that is present in the primary study and that the synt hesist can readily transcribe to coding sheets for future analysis” (Hall, Tickle-Degnen, Rosenthal, & Mosteller, 1994, p. 25). As in the case of other low inference coding procedures such as “characteristics of publication, type of research design, measuring instruments, and subject characteristics, and the kind of

PAGE 58

47 setting where the study took place” (Hall et al., 1994, p.25), calculating inter-rater reliability was not deemed relevant to this coding process. Publication type and sample size characteristics of studies in the review Publication YearPublication Year20052003200119991997199519931991198919871985198319811979197719751965Frequency3020100 Figure 4-1. Publication Year Breakdown of Articles in the Empirical Review. Although the search parameter was set as the period of 1887-2005, the earliest article located using the previously specified selection criteria was published in 1965. Although the topic of contextual effects on marriages has been explored previously, as mentioned in the previous literature review, a scan of the earlier articles in the possession of the author suggests that researchers prior to 1965 may have focused on marital status outcomes such as divorce and marital stability rather than the outcome of marital satisfaction that is currently under investigation. Thus, the 225 articles were published from 1965 to 2005. Overall, there appears to be a trend towards increased interest in the topic of how stress affects marriages. The cross-disciplinary nature of this body of work is apparent in that more than one hundred separate journals have featured articles exploring associations between stress and marital satisfaction. Broadly speaking, these

PAGE 59

48 journals represent fields such as marriage and family studies (e.g., Journal of Marriage and the Family, Family Relations, Journal of Family Psychology, The American Journal of Family Therapy), occupational behavior (e.g., Journal of Occupational Behavior, Journal of Organizational Behavior), social psychology (Journal of Social and Personal Relations, Journal of Personality and Social Psychology), and health and medicine (Journal of Psychosomatic Research, Archives of Physical Medicine, Pain, Journal of Health and Social Behavior). However, the categories most widely represented are the marriage and family studies literature, in which 36% of the articles were published and the medical journals in which 26% of the articles were published. Sample sizeSample size3617166813448105654533943442722472081881521269881604016Frequency543210 Figure 4-2. Sample Size Distribution for Studies in the Empirical Review All together, the 225 studies represent a total of 164,359 married individuals. However, it should be noted that some samples, such as nationally representative panel studies, may have been used by more than one researcher, so this is not to say that there are 164,359 discrete cases. The range of sample sizes is wide, spanning a sample of 16 spouses of aphasic stroke patients to a sample of 59,896 married individuals in which at

PAGE 60

49 least one spouse was deployed during the Gulf War. It bears noting that the next largest study is a sample of 6,952 married individuals in a national probability sample exploring associations between nonstandard work schedules and marital satisfaction. The mean sample size within the pool of studies is 730. However, the median sample size of 206 is a better estimate of the average sample size in the pool of studies because 59,896 is a clear outlier. So, there is considerable diversity in the publication sources and sample sizes within studies in the review. The cross-disciplinary nature of this work and the wide variability in sample sizes may possibly indicate that a broad range of samples have been recruited. The next section will assess the extent of sampling diversity by describing the demographic characteristics of samples in light of gender, national representation, international representation, average income, and racial composition. Demographic characteristics of samples addressed by studies in the review Gender of ParticipantsGender of ParticipantsMaleFemaleFrequency6050403020100 Figure 4-3. Gender Distribution of Participant in Samples in the Empirical Review.

PAGE 61

50 The kind of conclusions we can draw from studies of stress and marriage are limited by the samples that have been studied. While it would be nice to be able to make broad conclusions about associations between stress and marital satisfaction, any generalizations of findings must be anchored to the characteristics of the samples in this body of literature. Along these lines, the following section will illustrate and describe the demographic characteristics of the samples in the review. In terms of gender distribution, males and females appear to be rather equally represented in the body of studies, with 78,415 (48%) females and 85,944 (52%) males. The fairly equal distribution across genders is likely due to the predominant tendency to recruit married couples as opposed to married individuals. The proportionate representation of both genders will allow for comparisons of effect sizes for males versus females in the meta-analysis that follows. Proportion of Nationally Rep. SamplesProportion of Nationally Rep. SamplesNatl. Rep.Not Natl. Rep.Frequency3002001000 Figure 4-4. National Representation of Samples in the Empirical Review. The majority of the studies, 200 studies (89%), are not nationally representative. Only 25 studies (11%) of the studies in the review used nationally representative samples. Most of the samples that have been collected have either been convenience samples or

PAGE 62

51 highly targeted samples. An example of a convenience sample that typifies the studies in this pool is the use of participants in a readily accessible work organization or academic setting. An example of a highly targeted sample is a group of wives with breast cancer who were recruited from an oncology clinic. Proportion of International SamplesProportion of International SamplesInternational SampleU.S. SampleFrequency2001000 Figure 4-5. International Sample Recruiting Another interesting demographic variable is whether the samples in the review were recruited in the United States or were of International origins. Five studies did not report on the national origins of their sample. Of the 220 that did provide these data, a large majority of the samples were based in the United States, 164 studies (75%) out of the 220 total studies. International samples were recruited in 56 studies, representing one quarter of the total pool. Of the international samples, the most frequently represented areas of the world were Canada, Asia, and the Netherlands followed by South Africa, Israel, and England. Other regions, including Australia, Germany, India, the Czech

PAGE 63

52 Republic, and the Soviet Union accounted for less than 1% of the International samples. Thus, most of the samples included in the review were recruited from the United States. More than 60 percent Caucasian?More than 60 percent Caucasian?YesNoFrequency120100806040200 Figure 4-6. Racial Composition of Samples in the Empirical Review Another critical demographic variable is racial composition of samples. Ninety-three studies out of the 225 total studies (41%) did not report racial composition data. However, of the 132 studies that did present the racial composition of their sample, 105 (80%) of these studies recruited groups that were more than 60% Caucasian. Thus, the demographic profile of the studies that presented racial data appears to be predominantly Caucasian. This does not necessarily mean that these samples are unrepresentative of the areas in which they were recruited or unrepresentative within the United States as a whole for a particular time period in history. There may have been a high percentage of Caucasian couples in areas where the studies occurred. However, the high proportion of largely Caucasian samples in studies that do report racial breakdowns, suggests that stress

PAGE 64

53 researchers may not be tapping the very groups that are likely to be dealing with the chronic pervasive environmental stress that non-Caucasians are faced with on a daily basis, or the very groups likely to be experiencing the highest levels of financial stress (e.g., Peters & Massey, 1983). So, extrapolating on the admittedly scant data that have been reported, it is likely that the majority of the samples are populated by Caucasian participants, a situation which has some potentially sobering implications for the study of stress. MEDIINCOMEDIINCO85000.0076000.0070000.0055000.0051000.0050000.0045000.0040000.0035000.0034980.0033868.0033399.0030000.0028000.0027500.0025000.0017500.007925.00Frequency2.52.01.51.0.50.0 Figure 4-7. Median Income of Studies in the Empirical Review In total, forty percent of the 225 studies provided income information in the form of either median household income, average household income, individual income for one married partner, or a range of incomes. Eleven internationally based studies were eliminated from the average income calculations because the metric for comparison with

PAGE 65

54 U.S. Census data would not have been the same. The U.S. Census Bureau website 2 provides median and mean incomes by year between the years of 1947-2003 according to race, region of the country, and gender. Due to the great variability in income levels due to these factors, and the amount of missing data in terms of racial background of the samples and the recruitment of participants from regions all over the United States, I decided to compare all reported median incomes with median incomes for all races and regions of the United States. The median income of the samples in the review was $41,758 (s.d. = $19,428), slightly higher than the national median household income of $40,156 for the corresponding years. So, despite notable variability in sample sizes and the cross-disciplinary nature of this literature, extrapolating on the admittedly scant data that have been reported, it is likely that a majority of the samples are not only populated by Caucasian participants, but by participants that are likely to be in the middle class income range. 3 The next section will move from sampling to design, and will assess the types of design and measurement that characterize the studies in the review. Design characteristics of studies in the review The design of samples in the review may also inform the conclusions we can draw. For example, cross-sectional designs present snapshots of marriages while longitudinal designs allow us to explore trajectories of marital satisfaction over time, the separation of genders in analyses allows for a computation of gender differences, and various measurement methods (e.g., self-report, observation and coding by independent trained 2 (Source: http://www.census.gov.hhes/income/histinc/h05.html; Table H-5.) 3 The term middle class is used to describe samples whose median income approximates the national median income for the corresponding year.

PAGE 66

55 raters) provide different kinds of data. The following section will explore several key elements of design within the samples in the review. Longitudinal or Cross-sectional studyLongitudinal or Cross-sectional studyCross-sectionalLongitudinalFrequency2001000 Figure 4-8. Study Design in Articles in the Empirical Review. Studies in the review can be broadly distinguished as featuring either a longitudinal or cross-sectional design. The majority of studies, 156 (70%) out of the 225 studies in the current review, are cross-sectional designs. Thus, the current literature bears a closer resemblance to an album of snapshots in time rather than a series of movies exploring the interplay of stressors with changes in marital satisfaction over time. A longitudinal study is defined as any study with two or more data collection points, or waves. Of the 69 longitudinal studies in the sample, an average of 3.75 (s.d. = 3.93) waves of data were collected. It is noteworthy that the standard deviation in this calculation is larger than the average. Another way to look at these data, which can be seen in the bar graph above, is that more than 50% (n = 36) of the studies involved only two waves of data collection.

PAGE 67

56 Number of WavesNumber of Waves301511865432Frequency403020100 Figure 4-9. Number of Waves in Longitudinal Studies in the Review. Months of DurationMonths of Duration20414496724830201816126310Frequency121086420 Figure 4-10. Average Length of Longitudinal Studies in the Review. As mentioned, the definition of a longitudinal design is any study with at least two data collection points. Therefore, a study that lasts for less than one week might be

PAGE 68

57 classified as longitudinal. A period of just over three years (mean = 3.2 years; s.d. = 4.74) was the mean length of the longitudinal designs in the current review. As in the case of the mean sample size, the mean length of studies is clearly influenced by outlying data. The shortest longitudinal studies were 5 and 7 days while the longest longitudinal studies were fourteen and thirty years. Based on a trimmed sample in which 10% of the data was removed from the lowest and highest ends of the spectrum, the mean study length of the longitudinal studies was 2 years (s.d.=1.58). Another way to parse these data is to break it down into three broad time lengths. Twenty one studies (31% of the total) lasted less than 1 year, twenty-three studies (33% of the total) lasted between one and two years, twenty-five studies (36% of the total) lasted more than two years. Either way, extremely lengthy studies are relatively rare in this literature. Sampling UnitSampling UnitCoupleIndividualFrequency140120100806040200 Figure 4-11. Sampling Unit in Studies in the Empirical Review

PAGE 69

58 Another distinct design feature relates to researchers’ choice to collect data from either one individual in a married pair or from both husbands and wives. In the 225 studies under review, 129 (57%) analyzed couples while 96 (43%) analyzed married individuals. Thus, more than half of the studies obtained data from both husbands and wives in a couple. Genders analyzed separately?Genders analyzed separately?YesNoFrequency2001000 Figure 4-12. Separation of Gender in Analyses. Another related issue is whether gender of respondents was viewed as a potentially important variable. That is, were both genders lumped together in the analysis stage or were males and females analyzed separately, allowing for the exploration of gender differences? The decision to collect data from both husbands and wives as opposed to just one partner is not isomorphic with the unit of analysis. For example, the unit of analysis was collapsed across gender in a study of partners, both male and female, of heart attack patients. In another case, however, researchers focused on spouses of women with breast cancer, therefore guaranteeing the homogeneity of sample gender. Another group of

PAGE 70

59 researchers collected data from both husbands and wives who had lost a child, but collapsed across gender in the analysis phase. However, in general, researchers did analyze genders separately; in the 225 studies in the current review, 178 (79%) studies split the data by gender, while 47 (21%) studies collapsed across gender. In addition to the aforementioned fact that samples in this review recruited a fairly equal distribution of males and females, the fact that a vast majority of researchers have split their data by gender will allow for meta-analytic comparisons of gender differences in effect sizes for various domains of stressors. How did they measure the variables?How did they measure the variables?1312self-report onlyFrequency3002001000 Figure 4-13. Measurement Methods in the Studies in the Empirical Review Finally, the mode of measurement used in the studies was coded according to whether variables were measured using self-report only, self-report plus observation or interview, or self report plus archival or record search. 4 A vast majority of the studies in 4 As in the case of other “low inference coding procedures” (Hall, Tickle-Degnen, Rosenthal, & Mosteller, 1994, p. 25) these data was coded by the graduate principal investigator.

PAGE 71

60 the current review, 198 studies (88%) used self-report measures only. Twenty-three studies (10%) used observation or interview procedures (denoted as ” in Figure 4-13) in addition to self-report and four studies (2%) accessed archival data (denoted as ” in Figure 4-13) in addition to self report. Now that the domains of stress, sampling, and design characteristics of the empirical literature have been described, the next section will summarize and critique the features of this literature, outlining several strengths and deficits that could mitigate potential findings based on this body of work. Summary and Critique of Characteristics in the Empirical Literature Arguably, the biggest strength of this empirical literature is its size and scope. Despite the fact that many have pointed to a neglect of the role of context in marital research (e.g., Huston, 2000; Kurdek, 1991; White, 1990), there are at least 225 studies, representing a total of 164,359 married individuals, in this literature. Thus, this literature represents the amalgamation of a massive amount of research looking at how stress impacts marital satisfaction. Furthermore, the cross-disciplinary nature of this body of this work is apparent in that more than one hundred separate journals have featured articles exploring associations between stress and marital satisfaction. There are at least two broad implications of the cross-disciplinary nature of these studies. First, on the positive side, it appears that the relationship between stress and marriage has been researched by investigators representing diverse philosophical and academic backgrounds. This body of research appears to be expanding both in the number of studies conducted and in the diversity of traditions related to the research. However, on the more negative side is the possibility that no one field seems to have gained a critical mass of studies that would allow for the recognition of patterns in

PAGE 72

61 the findings. The fact that these articles are spread over so many disciplines may account for the lack of a prior review of this scope. Furthermore, each discipline is likely to use different measures of variables, which might contribute to the lack of consensus about the operationalization of variables described by Karney and Bradbury (1995). For example, researchers in the marital and family studies field typically use the Dyadic Adjustment Scale (Spanier, 1976) or the Quality Marriage Index (QMI; Norton, 1983) to measure marital satisfaction. However, articles which represent other disciplines may use very brief, empirically un-validated assessments of marital satisfaction such as “On a scale of 1-6, how happy are you with your marriage?” Thus, although it would appear that the relationship between stress and marital satisfaction is being researched with increasing frequency across a wider variety of fields, no one field seems to have gained enough critical mass to allow for the description of patterns and holistic perspectives on convergent and contradictory findings. Because there is great sample size variability and a couple very large outliers, the median sample size of 206 is probably the best estimate of the average sample size. From the perspective of other consumers of this literature, this average sample size of about 200 participants could be considered a positive feature of this literature. However, the power of individual studies is not critical in the context of this meta-analysis for two reasons. First, the meta-analytic procedure of aggregating effect sizes weighs study results according to the sample size of contributing studies. Therefore, a correlation between stress and marital satisfaction in a study with a sample of 20 married individuals will not contribute as much weight to the aggregate effect size when compared to a study with a sample of 400 married individuals. Second, the power of individual studies to find

PAGE 73

62 significant results is not relevant because combining individual studies to form aggregate effect sizes results in a huge increase in power. There are also strengths and limitations within the demographic profile of this literature. Fairly equal representation of both genders is a positive feature of this literature, allowing for comparisons of gender differences in associations between stress and marital satisfaction. However, aside from variability in gender representation, considerable restrictions of range dominate the demographic characteristics of this research. The vast majority of these studies are not nationally representative and are based on samples recruited in the United States. Most of the samples that have been collected have either been convenience samples or very small, highly targeted samples with specific stressors such as spousal illness. The average income of the samples is slightly higher than the national median income for the years corresponding to the studies. When income data is reported, the median sample income suggests that the majority of the studies looking at associations between stressors and marital satisfaction have probably focused on middle class samples. Moreover, in terms of racial composition, there appears to be a widespread tendency to omit racial composition data, as ninety-three studies out of the 225 total studies (41%) did not report these data. Of the 132 studies that did present the racial composition of their sample, 105 (80%) of these studies recruited groups that were more than 60% Caucasian. In other words, only 20% of the samples that reported racial background data recruited samples that were less than 60% Caucasian. These demographic characteristics paint a picture of a population that is likely to be comprised of middle-class, Caucasian participants. In effect, these studies tell us little

PAGE 74

63 about very poor or very wealthy people, people of very high or very low SES, or people whose racial background is not Caucasian. This lack of fiscal, socio-economic, and racial diversity in existing studies of stress and marriage introduces a particularly sobering implication in regard to the lower end of these spectra: we simply do not have much data on the very samples that are likely to experience the highest levels of stress as well those without the requisite physical resources to facilitate coping with stressors (e.g., Peters & Massey, 1983). Furthermore, any results we obtain based on the current literature, which appears to be comprised mainly of middle class, Caucasian participants, may not generalize to other socio-economic or non-Caucasian populations. Moreover, the widespread tendency for researchers to neglect to report income and racial composition of their samples as well as the obvious restriction of range when that data is available may limit the extent to which it will be possible to conduct moderating analyses in the meta-analysis to follow because some domains of stress may lack sufficient variability. Finally, from a statistical point of view, restrictions of range attenuate correlations and may obscure the true magnitude of findings. Thus, any results that are obtained should hold even more weight due to homogeneous sampling characteristics that limit the potential for finding significant results. Figures 4.8-4.13 reveal a number of methodological strengths and limitations. First, about two thirds of the studies in the review are based on a cross-sectional design. On one hand, it is encouraging to see that one third of the studies in the current review provide data about changes in marriages over time. On the other hand, the predominance of cross-sectional designs entails limitations articulated by several researchers (e.g., Bradbury & Karney, 1993; Markman, 1979; Weiss & Heyman, 1990). Specifically, cross-sectional

PAGE 75

64 designs cannot address the causes of marital dissatisfaction or the causal mechanisms that induce harmful stress effects on marriages. Some (e.g., Weiss & Heyman, 1990) have even stated that cross sectional data is largely irrelevant and potentially misleading in understanding marital conflict. In the case of the current meta-analysis, these limitations will not be a problem because I will be aggregating correlations, but the prevalence of cross-sectional designs does have implications for future research, which will be discussed later. In general, the greater the number of data collection points, the more it is possible to see clearer trends as well as a more complete picture of a phenomena. Yet, Figure 4-9 shows that more than half of the longitudinal studies in the review involved only two waves of data collection. The number of studies with six or more waves of data is minimal, representing only 6 studies, less than 1% of the 69 longitudinal studies. Basing results on only two waves of data is not uncommon; Karney and Bradbury (1995) reviewed 115 longitudinal studies of marriage and found that over 70% of the studies used procedures that analyzed only two waves of data. Some have pointed out serious limitations of relying on only two waves of data. For instance, a two-wave design does not support complex hypotheses about how relationships change and development over time (Karney; Bradbury, & Johnson, 1999). Moreover, as Rogosa, Brant, & Zimowski noted, “Two waves of data contain an extremely limited amount of information about the change of each individual” (1982, p. 729). This dissertation does not focus on change over time, and thus, the collection of one-wave of data will not be a direct limitation in the current analyses. However, the paucity of studies with more than two waves of data is troubling as an indicator of the general quality of the studies in this literature.

PAGE 76

65 The average length of 2-3 years for longitudinal studies in the review is another clear limitation. In general, data derived over a period of months or years offers a much richer source upon which to base conclusions. Measuring stability and change over time may be especially important in the study of close relationships given Furman’s suggestion that “continuity over time” is the key feature that distinguishes close relationships (1984, p. 29). An average study length of 2-3 years is not very long in the course of marriage, and studies of this length may fail to assess important marital milestones such as pregnancy and childbirth. In addition to missing critical life events and normative transitions, the use of brief research designs may also obscure possible patterns in how spouses interact as they are faced with stressors and major life transitions over time. Finally, studies with very short time spans may also fail to fully capture or account for the presence of personality factors and individual differences that may emerge most clearly in response to stressful events. Thus, the abundance of cross-sectional designs and short-term longitudinal studies likens this literature more to a collection of snapshots as opposed to a multi-layered script that unfolds over time. However, a noteworthy strength of these studies is that the majority include data on both husbands and wives together as opposed to just one member of a marital dyad. Data collected from both members in a married couple theoretically provides a more complete picture of a marriage, allowing researchers to look at the interplay of variables elicited from both partners. Collection of couple data also allows for intra-couple analyses of stable personality factors and individual reactions to stressors. Another strength of this set of studies in that researchers have generally analyzed genders separately, allowing for explorations of

PAGE 77

66 possible gender differences in aggregate correlation magnitude in the meta-analysis to follow. Finally, the majority of studies relied exclusively on self-report to collect data on both stressors and marital satisfaction. Several researchers have expressed serious criticisms of exclusive reliance on self-report data. For example, Raush, Barry, Hertel, & Swain (1974, p. 5) asserted that “studying what people say about themselves is no substitute for studying how they behave. Questionnaireshave yielded very little. We need to look at what people do with one another.” When researchers depend exclusively on self-reports, the validity of findings may also be questionable to the extent that external variables can affect self-reports of stress and marital quality, leading to inflated correlations between the two. Another issue related to the use of self-report is whether participants are evaluating their relationships using a common standard. Furthermore, subjectively assessed outcomes may be vulnerable to memory-based distortions or social desirability motivations. An example of the former possibility is found in a fascinating study conducted by Karney and Frye (2002). These authors found evidence that “partners may construct histories to justify their current orientation toward the relationship” (p. 222). Specifically, when asked to evaluate their relationships, newlyweds tended to report that satisfaction had declined in the distant past but was currently marked by recent improvements. Prospective reports provided by the same couples, however, tended to decline linearly over time! The degree to which researchers employ more than one method of data collection speaks to the foundations of researchers’ conclusions, since the use of self-report alone does not provide the richness or convergent validity of data obtained with multiple

PAGE 78

67 methodologies such as observation or archival techniques. Gottman has referred to the use of “self-report data obtained from a single reporter” as “the glop problem” (1998, p. 172-173). Although “it is so much easier to hand out a packet of questionnaires” than using multiple methodologies, Gottman admonishes that marital researchers ought to “employ multiple methods to operationalize constructs” (p. 173). Given these concerns about the exclusive use of self-report measures, it is that troubling that nearly all of this research is based on self-report measures alone. The previous critique points to some noteworthy strengths in the empirical literature such as its size and cross-disciplinary nature, the relatively large average sample size of 200 participants, fairly equal representation of both genders, and the separation of husband and wife data in most of the statistical analyses. Despite these strengths, there are several serious limitations that may temper any conclusions based on these data, such as the likelihood that results may not generalize beyond a constricted population of middle-class, Caucasian participants. Bearing these strengths and limitations in mind, I will now focus on a meta-analysis exploring associations between eleven domains of stress and marital satisfaction as well as three potential moderators of these aggregate effect sizes. In the next chapter, I describe the rationale for conducting the meta-analysis and the inclusion criteria used in selecting articles, propose five research questions, and outline the meta-analytic strategies that will be used to address these questions.

PAGE 79

CHAPTER 5 METHOD AND PROPOSED HYPOTHESES Rationale The guiding aim of this dissertation is to gain a better understanding of how marital quality may be affected by variables external to the marriage. Although I have identified 225 articles on this topic, representing a large body of literature, the previous section features the first extensive review of this topic that I am aware of to date. Descriptions of the sampling, demographic, and design characteristics of samples in the empirical review suggest that researchers have explored associations between eleven different types of stressors and marital satisfaction. However, the relative impact of various types of stressors remains unclear. While fairly equal proportions of females and males have been sampled, any possible gender differences in the impact of stress on marital satisfaction are unknown. Researchers have also collected data on some potentially important moderators of associations between stress and marital satisfaction, including sampling characteristics such as national representation, and demographic variables like income level and racial background of participants in various samples. However, the degree to which these types of moderators may account for differences in the magnitude of associations between stressors and marital satisfaction remains to be discovered. These gaps in the literature warrant the use of meta-analytic techniques, which offer several specific advantages with respect to the goal of synthesizing and gaining a clearer perspective on the nature of these associations. 68

PAGE 80

69 The primary strength of meta-analyses is that they represent an efficient way to summarize large literatures (Green & Hall, 1984). In this case, each correlation between a stressor and a marital satisfaction outcome represents one data point. Meta-analysis aggregates each individual correlation into an average effect size for a given domain of stress while taking sample size into account. From a statistical point of view, the high level of aggregation involved in meta-analysis reduces a large portion of the sampling error that accounts for much of the instability of individual study findings (Schmidt, 1992). From a practical point of view, this synthesis of the literature also moves discussion away from individual studies toward an overview of the whole body of research bearing on a topic, allowing for a holistic view of the ways stress affects marriages. Such a “birds-eye” view of the literature has at least two distinct advantages. First, meta-analytic findings possess “a certain robustness that make them highly attractive as a basis for policy and practice guidelines” (Lipsey & Wilson, 2001, pp. 167). Second, meta-analysis often provides an ideal way to quickly locate the gaps in the literature and identify which populations have been studied extensively and which have been largely neglected. As such, meta-analyses are often a powerful source of insight into intriguing new directions that warrant serious consideration in future research (Green & Hall, 1984; Lipsey & Wilson, 2001). Second, another major advantage of the meta-analytic approach is “its ability to ask not simply whether there is overall support for a hypothesis, but also whether the extent of support varies with attributes of the studies” (Green & Hall, 1984). That is, when studies differ in their findings, meta-analysis allows researchers to determine the source of those differences and may demonstrate that key variables account for the magnitude of

PAGE 81

70 differences in effect sizes (e.g., Lipsey & Wilson, 2001). Furthermore, this ability to look for interactions in the data means that it is possible to test hypotheses that were never tested in individual studies, allowing for the extension of extant research findings into new terrain. Finally, in coding the variables that may serve as important moderators, a high quality meta-analysis makes its methods transparent. Lipsey & Wilson (2001) have argued that the explicit specification of criteria and methods is a noteworthy strength of meta-analysis, when compared to empirical studies that do not inform the consumer about how certain populations or various methodologies were selected. In light of the distinct advantages of meta-analytic techniques, the following section will focus on a meta-analysis exploring five research questions. In keeping with the principle of transparency, the section below outlines the inclusion criterion for articles in the meta-analysis. Then, I will describe the five research questions and outline empirical and theoretical justifications for any hypothesized findings. Finally, following the presentation of research questions, I will clarify the analysis procedure. Meta-Analytic Inclusion Criterion To ensure that all statistics being combined were comparable, I adopted a very strict inclusion criterion, including only articles that directly reported correlations that could be aggregated. Sometimes these correlations were reported in the text of the articles but often they were pulled from correlation matrices provided in the articles. Studies that failed to report zero-order correlations in the text of the article or within a correlation matrix were not included. So, studies that did not report a correlation between external stressors and marital quality were included in the empirical review but not in the meta-analysis.

PAGE 82

71 Based on this criterion, a total of 113 articles from the original pool of 608 articles were retained for inclusion in the meta-analysis, representing 44,346 married individuals. Research Questions Analyses addressed the following five research questions. Analysis of Association between Stressors and Marital Satisfaction Research Question 1. What is the aggregate effect size for associations of stress and marital satisfaction within each of the eleven domains of stress and how does the magnitude of aggregate effect sizes compare across domains? Empirical Justification. The preceding literature review and common sense suggest that stress will have a negative impact on relationship functioning. As stressful situations arise, specific processes in relationships are likely to deteriorate. Sexual intimacy may suffer (e.g., Hill, 1949) and interactions between couples may become more tense and discordant (e.g., Liker & Elder, 1983; Vinokur, Price, & Caplan, 1996). Individual and collective decision-making may be impaired (e.g., Janis, 1918 as cited in Janis, 1993) and self-efficacy may founder (Bandura, 1997). After observing families affected by separation of war, Hill (1949) concluded that most crises sooner or later come to involve a de-moralization (p. 10), which he defines as a loss of morale and unity. In general, then, stress could be expected to impede, rather than enhance, healthy relationship functioning. However, not all stressors are likely to affect couple functioning to the same extent. Ostensibly, being separated from one’s spouse is stressful. Yet, the “loneliness” that results from the medical student’s “inaccessibility” (e.g., Perlow & Mullins, 1976, p. 726; Warde, Moonesinghe, Allen, & Gelberg, 1999) or from “unpredictable time demands” for members of the clergy (e.g., Morris & Blanton, 1994) is not likely to be as stressful as

PAGE 83

72 separation because one’s spouse is a prisoner of war (e.g., Cook, Riggs, Thompson, Coyne, & Sheikh, 2004). Moreover, some separations may not even be that stressful at all. For example, one study showed that marital satisfaction was enhanced when a husband traveled for those couples that hold traditional gender role attitudes (Roehling & Bultman, 2002), suggesting that these wives may have enjoyed a break from steady interaction with their husbands. Similarly, while work strains such as time pressures, lack of autonomy and having a poor relationship with one’s boss (e.g., Mauno & Kinnunen, 1999) are no doubt stressful, it is probably more stressful to suddenly lose one’s job due to the closing of a major business (e.g., Broman, Hamilton, & Hoffman, 1990; Ridley & Wilhelm, 1988). Along the same lines, although family transitions may be a source of stress (e.g., Menaghan, 1983), the stress of Bosnians refugees transitioning to a new life in the United States (Spasojevic, Heffer, & Snyder, 2000) or couples living under the threat of forced relocation due to military tensions (e.g., Shamai & Lev, 1999) is likely to be greater. Thus, while stress is generally harmful to marriages, there may be some stressors that have a more harmful impact than others. The potential complexity regarding the impact of various stressors on marriages prompts questions about the conditions under which stress is likely to be harmful as opposed to beneficial. I have previously described theories that outline several key dimensions that differentiate types of stressors. A chief distinction between stressors types is longevity, referring to whether a stressor is acute or chronic. Acute stressors have variously been described as “time-limited” (Gump & Matthews, 1999), “discrete and observable” (Wheaton, 1997, p. 52), and having a relatively clear onset and offset (Wheaton, 1997). Examples of acute stressors are natural

PAGE 84

73 disasters, pre-doctoral comprehensive examinations, and physical relocation. Conversely, chronic stressors have been defined as stressors lasting for at least one year and any ongoing difficulties that have taken place for longer than a year (McGonagle & Kessler, 1990). In some cases, chronic stressors may form the “backdrop” (p. 91) against which acute stressors occur (Taylor and Aspinwall, 1996). Other distinctions may also provide a way to predict and understand differential impacts of stressor events. Several theorists have made the intuitive statement that stressors which are more severe or intense, such as having a child with leukemia, are likely to have a greater impact relative to less severe or intense stressors, such as having a child with chicken pox (Boss, 1987; Hill, 1964). There is also theoretical consensus that stressors which are more sudden, unexpected, relatively rare, or random, are likely to have a greater impact (e.g., Boss, 1987; Hill, 1964, Figley, 1983; McCubbin & Figley, 1983). In terms of predicting the extent to which stress affects individual or couple functioning, it seems that each of these stressor qualities directly relates to the ability to “muster” various types of resources. Lastly, the impact of stressors may also be more pronounced when stressors are sudden, unexpected, rare, or random because couples are less likely to have a lot of previous experience dealing with them. In regard to the proposed analysis at hand, although stress is expected to have a negative impact on marital satisfaction in general, it is not clear which types of stress may have a much greater impact when compared to other types. Therefore, research question one will identify the aggregate effect size for associations of stress and marital satisfaction for each of the eleven domains of stress, which will facilitate comparisons between the magnitude of effect sizes across domains. Finally, in making comparisons

PAGE 85

74 between domains, the theoretical dimensions of stressor types may elucidate any differences in the magnitude of associations between various stressors and marital satisfaction. Analysis of Theorized Moderation of Associations between Stress and Marital Satisfaction Research Question 2. For any of the eleven domains of stressors, will there be a significant gender difference in the magnitude of the association between stressors and marital satisfaction? Empirical Justification. It is possible that in many cases, stress affects people the same regardless of their gender. For example, daily stress appears to decrease marital quality for husbands and wives (Harper, Schaalje, & Sandberg, 2000). In a study of spouses who have a child diagnosed with Rett syndrome, husbands and wives perceived their family situation quite similarly (Perry, Sarlo-McGarvey, & Factor, 1992). Likewise, the transition to retirement appears to be related to declines in marital quality for both husbands and wives, and married men and women who move into retirement while their spouses remain employed report the greatest marital conflict, regardless of gender (Moen, Kim, & Hofmeister, 2001). However, it is also possible that certain types of stress may impact husbands and wives differently. It may not be particularly surprising that women with breast cancer have reported significantly more distress than their husbands (Northouse, Jeffs, Cracchiolo-Caraway, Lampman, & Dorris, 1995). In this case, breast cancer can be seen as a direct source of distress for these wives, whose bodies are ravaged by its presence.

PAGE 86

75 Yet, empirical examples support some interesting gender differences for various types of stressors in which the stressor could theoretically impact both partners to the same degree. As mentioned above, Moen, Kim and Hofmeister (2001) reported that retirement is associated with lower marital quality for males and females. However, another large national sample of males transitioning into retirement suggests that gender role reversals may lower marital quality (Myers & Booth, 1996). In terms of the domains of stress featured in the current meta-analysis, this finding might suggest gender differences could emerge for unemployment, given the likely upheaval of role changes the loss of a job could cause. In the domain of work, Pradhan and Misra (1995) found that job stress was associated with the quality of marital relationship in males but not in females. Opposite findings were observed in a Finnish study of work stress; women with job stressors had more psychosomatic symptoms than men (Mauno & Kinnunen, 1999). Doumas, Margolin, & John (2003) asked husbands and wives to complete daily diaries assessing work experiences and marital interactions over a period of 42 days. Results showed different patterns of spillover for husbands and wives, leading the authors of suggest that “wives may be more reactive to their husbands’ experiences and behaviors than vice versa” (p.3). Of course, it seems to me equally plausible that the same finding can indicate that grouchy, irritated husbands may be more difficult to deal with than long-suffering wives! In another study, deployment of male soldiers during the Gulf War conflict had no effect on marital dissolution, while deployment of female soldiers led to a large and statistically significant increase in divorce rates, “suggesting that deployment of women placed a marked strain on marriages” (Angrist & Johnson, 2000, p. 41). Thus,

PAGE 87

76 tests of gender differences in the domain of work stress have been mixed and I do not expect to find significant gender differences in this domain. However, evidence suggests that balancing work and family roles may consistently entail special challenges for women. For instance, Bedeian, Burke, & Moffett (1988) found that the relationship between parental demands and life satisfaction was mediated by satisfaction with childcare arrangements for women, but not men, with young children at home. Coverman (1989) showed that perceived role conflict, but not role overload, decreased women’s psychological health. Barnes, Pase, and VanLeeuwen (1999) stated that employed women with children have “unique” stresses such as a disproportionate share of child care responsibilities and gender-specific role experiences and generally “exhibit higher rates of distress” (p. 203). Findings based on a sample of 2,877 workers in the 1997 wave of the National Study of the Changing Workforce revealed that more working mothers than fathers experienced negative family to work spillover. Furthermore, time spent caring for children and doing household chores did not predict negative spillover for mothers, while caring for a sick child was a significant predictor of spillover for fathers (Dilworth, 2004). Therefore, although I am not expecting that gender differences will emerge in the domain of work, gender differences may emerge in the work-family stress domain. Medical problems in the family may also disproportionately affect wives, who often shoulder the bigger burden of family care-giving. For parents of disabled children, mothers, but not fathers, experienced higher levels of stress compared to parents of non-disabled children (Kazak, 1987). In a study of couples in which one partner has multiple sclerosis, the female partners were more dissatisfied with the relationship relative to male

PAGE 88

77 partners (Dupont, 1996). For couples in which one partner is experiencing chronic pain, other results suggest a stronger relationship for female than for male spouses between the spouse’s perception of patient dysfunction and the spouse’s marital adjustment (Romano, Turner, & Clancy, 1989). The disproportionate impact of stress on wives may also be explained by a stronger emotional attachment to certain medical issues. Specifically, it has been shown that mothers are more likely to be distressed than fathers following a baby’s death and are more likely than husbands to report later marital problems (Vance, Boyle, Najman, & Thearle, 2002). Other research suggests that infertility may also have a stronger negative impact on wives than on husbands, regardless of which spouse is biologically infertile (e.g., Andrews, Abbey, & Halman, 1991). However, there are exceptions to every generalization. In a study of 68 cancer patients and their partners, the male patients felt over-benefited in their relationship, whereas the female patients on average felt equitably treated! (Kuijer, Buunk, Ybema, & Wobbes, 2002). In general, however, these findings seem to indicate the possibility of gender differences in the domains of spouse and child health problems. To summarize then, empirical findings indicate potential gender differences for the domains of unemployment, work-family stress, spouse’s and children’s health problems. Findings do not suggest that there will be gender differences for work stress. Empirical findings do not inform particular expectations about the other six domains of stress, which are general stress, war/military stress, interpersonal stress, economic strain, non-white racial background, and low socio-economic status. Therefore, exploratory analyses will be conducted for these six domains but no particular gender differences will be

PAGE 89

78 anticipated. To test research question 2, I will compute aggregate effect sizes for each gender and will run tests of gender differences within each domain. Research Question 3. Will income level moderate the magnitude of the association between domains of stress and marital satisfaction, such that samples with lower incomes will have more negative aggregate correlations than those with higher incomes? Empirical Justification. Hobfoll (1998) has stated, “if individuals have the requisite resources, they will be able to withstand or overcome stress, or even prevent it from occurring” (p. 92). In this case, resources refer to income level. Longitudinal data support an association between lower incomes and decisions to separate or divorce (e.g., Kurdek, 1991). As described in the previous review of the literature, there are several ways in which financial resources can directly or indirectly increase or decrease the impact of stressors on marriages. In a direct way, financial resources can provide opportunities for leisure and temporary reprieves from stressful circumstances. Employed women with children may have unique stresses such as a disproportionate share of child care responsibilities and gender-specific role expectations. However, increasing income level was positively associated with greater satisfaction in a sample of working women (Barnes, Pase, & VanLeeuwen, 1999). Conversely, a major stressor perceived by spouses of medical students was “limited finances for recreation and non-essentials” (Perlow & Mullins, 1976, p. 726). Indirectly, financial resources may moderate associations between stress and marital satisfaction. Two large-scale studies of unemployment, collectively representing 1663 participants, indicate that the stress of unemployment was mediated by financial hardship (Broman, Hamilton, & Hoffman, 1990; Ridley & Wilhelm, 1988). In another

PAGE 90

79 unemployed sample, financial strain was linked to depressive symptoms and destructive relationship behaviors which, in turn, were associated with decreased relationship satisfaction (Vinokur, Price, & Caplan, 1996). Similarly, Cutrona and colleagues found that neighborhood-level economic disadvantage predicted observed warmth, hostility and marital quality in a sample of African American couples (Cutrona, Russell, Abraham, Gardner, Melby, Bryant, and Conger, 2003). Conversely, empirical findings demonstrate evidence of stress-buffering for those who have greater financial resources. Higher levels of family income seemed to partially protect marriages in a sample of parents of severely mentally ill children (Cook, Hoffschmift, Cohler, & Pickett, 1992). Another study revealed a significant positive relationship between financial well-being and marital satisfaction in a sample of military couples (Thoresen & Goldsmith, 1987). Sampling may be a critical consideration when theorizing about the ways in which income level may exacerbate or ameliorate a stressor’s impact. For instance, based on a sample of black married couples, researchers have reported that economic resources are important predictors of marital quality among low-income groups but are less relevant to middle-income groups (Clark-Nicolas & Gray-Little, 1991). In light of this finding, it is noteworthy that the vast majority of the studies in this literature seem to be based on middle class, Caucasian participants. Because of this restriction of range, aggregate correlations are expected to be attenuated and any moderating analyses of income level will be subject to any limitations associated with research that appears to be focused on a relatively affluent set of participants. Generally speaking, though, based on theoretical support for the notion that financial resources moderate the association between stressors

PAGE 91

80 and a variety of psycho-social outcomes, samples with lower income levels are expected to have more negative aggregate correlations between stress and marital satisfaction. To test research question 3, I identified studies that recruited low-income samples. I then coded these studies as “low-income” and grouped the rest of the studies, which were based on samples with either mid or high-level incomes, into a category I coded “non-low-income.” The justification for making this broad division is that, theoretically speaking, the important distinction made in this analysis is whether a sample is populated with individuals who are likely to be experiencing the highest levels of financial strain relative to samples near the median income level or above. The aim in this analysis was to locate and separate the samples that are likely to have relatively more difficulty making ends meet from the other samples with relatively higher levels of income (typically around the national median based on an extrapolation of the income data that are reported). Based on the data that was reported, the average income in the low-income condition was $25,041 per year compared to the national average of $38,346 for the same years. In other words, the low-income average is 65% of the national average income. Moderating analyses, therefore, will be conducted on the basis of this categorical classification of income levels. Research Question 4. Will racial composition of the samples moderate the magnitude of the association between domains of stress and marital satisfaction, such that samples that are less than 60% Caucasian will have more negative aggregate correlations relative to samples that are more than 60% Caucasian? Empirical Justification. Empirical literature suggests that African Americans are less likely than Caucasians to be satisfied with their marriages (Broman, 1993; Dillaway & Broman, 2001; Renne,

PAGE 92

81 1970; Scanzoni, 1975). In considering this trend, it is important to note that non-white racial background is not being viewed as an inherent stressor, but rather as a proxy for multiple sources of chronic stress due to living as a minority in a dominant culture. Peters and Massey (1983) have described the pervasive, racially oppressive environment that defines the lives of many African Americans and Orbuch and colleagues (Orbuch, Veroff, Hassan, & Horrocks, 2002) refer to oppressive social conditions, lower status positions in society, and the unique challenges of parenthood and family responsibilities facing non-Caucasian couples. Financial satisfaction is one structural variable which has received some support as a critical explanatory variable (Broman, 1993). Scanzoni (1975) has documented considerable evidence of race-based disparities across various socio-economic indices, with African Americans having significantly less education, job status, and income compared to whites. Evidence of discrimination emerges in that black males have significantly lower job status relative to equally trained whites. Considering these findings, it should not surprisingly that African Americans report global feelings of deprivation relative to whites (Scanzoni, 1975). In the context of marriages, such pervasive disparities based on racial background are likely to result in feelings of de-moralization, which Hill (1949, p. 10) has defined as a loss of morale and unity. In this vein, research question 4 is designed to explore whether samples that are less than 60% Caucasian have more negative associations between stressors and marital satisfaction relative to those that are more than 60% Caucasian. To test research question 4, I identified studies whose samples were more than 60% Caucasian. I then coded these studies as “greater than 60% Caucasian” and grouped the rest of the studies, which were based on samples that were less than 60% Caucasian into

PAGE 93

82 a category I coded “less than 60% Caucasian.” Notwithstanding the crudeness of this distinction, and the fact that a sample falling into the latter category could theoretically be up to 59% Caucasian, most of the samples had to be coded as “greater than 60% Caucasian” due to the dearth of samples that included anything approaching a majority of participants with a non-white racial background. Despite this admitted limitation, I will conduct moderating analyses on the basis of this categorical classification of the racial composition of samples. Research Question 5. Will national representation of samples moderate the magnitude of the association between domains of stress and marital satisfaction, such that those samples that are nationally representative will have more negative aggregate correlations than those that are not nationally representative? Theoretical Justification. Restriction of range in the studies contributing to a meta-analysis results in attenuated correlations (Matt & Cook, 1994). Some of the studies in the meta-analysis are based on large, nationally representative samples while others are based on much smaller convenience samples or samples experiencing a relatively rare stressor such as chronic fatigue syndrome. Relative to geographically constricted samples, range restriction is less likely to emerge in nationally representative samples, due to their large size and the broad demographic span of participants. As such, it is possible that correlations derived from nationally representative samples may feature larger estimates of associations between stressors and marital satisfaction. To test research question 5, I identified studies whose samples were nationally representative. I then coded these studies as “nationally representative” and grouped the rest of the studies, which were based on geographically-constricted samples, into a

PAGE 94

83 category I coded “not nationally representative.” I will conduct exploratory moderating analyses based on whether samples are nationally or non-nationally representative in order to determine whether nationally representative samples have larger effect sizes than geographically restricted samples. Meta-Analytic Procedure Aggregation of Effect Sizes According to Lipsey and Wilson (2001), when the research findings to be meta-analyzed involve associations between two continuously measured variables, as in the case of a variety of stress impact scales and marital satisfaction measures, the product-moment correlation is the straightforwardly appropriate effect size statistic. (p. 63). Henceforth, the terms “effect size” or “effect size statistic” will be used to describe individual correlation coefficients derived from various studies. Aggregate effect sizes across studies will be identified as “aggregate” or “average effect size(s).” Despite variation in the operationalization of stress impact and marital satisfaction, because the correlation coefficient is already a standardized index, ranging from -1 to +1, it is “usable as a meta-analytic effect size statistic in its raw form even if the variables being correlated are differently operationalized” (Lipsey & Wilson, 2001, p. 63). Thus, relevant effect size statistics were identified for associations between marital satisfaction and each of the eleven domains of stress, including spouses’ health problems, work stress, economic stress, general/life stress, work-family stress, children’s health problems, low-SES status, minority status, war/military stress, unemployment stress, and interpersonal stress. In the context of meta-analyses, Lipsey and Wilson (2001) have suggested that effect size statistics contributing to aggregate effect sizes can usually be assumed to be

PAGE 95

84 statistically independent if, for a given distribution, no more than one correlation comes from any subject sample. While some have argued that correlations for sub-samples from the same study and those from different studies conducted by the same investigators may share dependencies (e.g., Landman & Dawes, 1982; Wolf, 1990), Lipsey and Wilson assert that “these dependencies are likely to be small in most applications.and the standard practice in meta-analysis has been to define independence at the sample or study level” (p.112). Thus, I created independent sets of effect size statistics by constructing separate databases for each domain of stress, allowing each study to contribute only one correlation for the association between the stressor and marital satisfaction to the aggregate effect size computed for that domain. Within a given study, if multiple correlations are reported for a construct, they should not be included in the analysis as though they are different data points because these data are not independent (Lipsey & Wilson, 2001). In some cases, more than one association was reported for the correlation between marital satisfaction and a stressor. For example, a given study may have reported a correlation between a health stressor and two separate, but related, self report measures of marital satisfaction. Unless the magnitude of the correlations is visibly disparate, meta-analytic guidelines (e.g., Lipsey & Wilson, 2001) suggest averaging over the correlations so that the sample on which they are based contributes only one effect size statistic to the distribution. Based on this suggestion, I averaged over correlations expressing the same association if the correlations were less than .10 different. In the handful of cases where the correlations of the same association were more disparate than this criteria allowed, I selected the correlation for the most commonly used measure that was reported. For instance, if a

PAGE 96

85 study reported a correlation between unemployment stress and the Dyadic Adjustment Scale and another correlation between the same stress variable and a less established marital satisfaction inventory, I selected the correlation associated with the DAS to be used as the appropriate effect size statistic for the meta-analysis. Furthermore, because different scales were used to assess marital quality variables and indices of stressors, the signs of all of the correlations were set so that negative correlations mean that higher stress levels are associated with lower levels of marital satisfaction. The Comprehensive Meta-analysis Program developed by Biostat, Incorporated (2004) was used to aggregate the effect size statistics for each domain of stress. This program uses the Hedges and Olkin method (1985) which differentially weights the sample sizes of studies as effect size statistics are aggregated. Because the standard error is used to determine the inverse variance weight needed for the aggregation analysis, and “correlation coefficients have a problematic standard error formulation” the program also transforms the correlations using Fishers’ Z formula: The aggregation of effect sizes for each stressor domain is then computed with the following equation, in which each effect size value is multiplied by its respective weight, then summed and divided by the sum of the weights. Finally, the Zr is then transformed back into an easily interpretable aggregate effect size, with the following values (Lipsey & Wilson, 2001, p. 147):

PAGE 97

86 Small Medium Large r < .10 r = .25 r > .40 This methodology represents an important development in statistical procedure because studies with larger samples are likely to produce more accurate results than those with smaller samples (Risdal & Singer, 2004). That is, the correlations obtained in various studies are allowed to contribute to an aggregate effect size in proportion to the number of participants in that study relative to the total number of participants across all studies in the analysis. The utility of this feature is illustrated through the following example. In the career counseling literature, Brown and Krane (2000) reported that the aggregate effect size of the efficacy of career interventions was .85. However, when the relative sample sizes of contributing studies was accounted for with the Hedges and Olkin method, this aggregate effect size was reduced to .48. In this way, the Hedges and Olkin method improves the accuracy of estimation of aggregate effect sizes. Test of Homogeneity After obtaining an aggregate effect size for each domain of stress, I conducted a test of homogeneity in order to assess whether the correlations from different studies that have been averaged into an aggregated effect size estimate the same population effect size (Hedges, 1982). In a homogeneous distribution, an individual effect size differs from the population mean only by sampling error whereas heterogeneity indicates the presence of variability that cannot be accounted for by sampling error alone. The homogeneity test is based on the Q statistic which is distributed as a chi-square with k-1 degrees of freedom where k is the number of effect sizes (Hedges & Olkin, 1985). The formula for this Q statistic (Cooper and Hedges, 1994) is:

PAGE 98

87 ___ Q = w (ES – ES) If Q exceeds the critical value with k-1 degrees of freedom, then the null hypothesis of homogeneity is rejected, suggesting a heterogeneous distribution and the need to explore key moderating variables. Conversely, a non-significant heterogeneity test indicates that further analysis of moderating variables is not warranted because the various correlations that have been aggregated estimate the same population effect size (Hedges, 1982a; Rosenthal & Rubin, 1982). As a result, stressor domains with homogeneous distributions will be dropped from future moderating analyses. Moderating Analysis of Gender Next, I conducted an exploratory test of gender differences. That is, I assessed whether differences in the magnitude of aggregate effect sizes for each gender within each domain were statistically significant. The test of significant gender differences is similar to a one-way ANOVA. Total variance can be portioned into within (Qw) and between (Qb) group variability. The between gender Q is the weighted sum of squares of the mean aggregate effect size for each gender around the grand mean. Likewise, the within group Q is the weighted sum of squares of the individual effect sizes within each gender around the female and male group means, pooled over both genders. As in the one-way ANOVA, the degrees of freedom for Qb is j where j is the number of groups. Qb is found by subtracting Qw from Q(total). The degrees of freedom for Qw is k-j where k is the number of effect sizes and j is the number of groups (2 in this case). Moderating Analyses based on Sampling Characteristics In the case of heterogeneous distributions, averaging over contrary results would lead to further confusion about the true magnitude of associations between stressors and

PAGE 99

88 marital satisfaction. Such cases warrant “coding of study and effect size descriptors” (Lipsey & Wilson, 2001). Rather than assuming that effect size heterogeneity (between study differences) is due to unobserved random sources, I believe that it has systematic sources and can be explained by differences in the sampling characteristics for the studies which contributed to the aggregate effect size. This assumption is the basis for further analyzing effect size variation in order to explore whether excess variability can be explained by showing that it is associated with moderator variables that systematically differentiate studies with larger or smaller effect size statistics. I coded each study for the following three additional sampling variables that might account for heterogeneity in the effect size aggregates: Income level (low or not) 1. Racial composition of sample: greater than sixty percent Caucasian? (yes or no) 2. Nationally representative sample? (yes or no) In coding the studies, I experienced the “common frustration in meta-analysis” (Lipsey & Wilson, 2001, p. 126) of uneven reporting practices among researchers when information that should have been reported was often missing or, if it was reported, was too vague to permit adequate coding for meta-analysis. For instance, recall that 93 out of 225 articles (41%) in the earlier review did not report racial composition data. Along the same lines, in an earlier version of this draft, one reviewer suggested I code SES level according to the Hollingshead Sociological Index, but there was insufficient data about the particular careers of participants in the sample to follow through on this suggestion. When these data were missing, cases had to be excluded from moderating analyses. Choosing between a fixed-effects and a random effects meta-analytic model was a difficult decision. On the one hand, a fixed-effects analysis is based on the assumption

PAGE 100

89 that the correlations being compared are the only correlations of interest. Given that the articles in the meta-analysis are a sub-set of the articles in the review, this seems to indicate a random effects model. On the other hand, Hedges (1994) says that “if differences between studies that lead to differences in effects are not regarded as randomthen fixed effects methods are appropriate for the analysis” (p. 286). Thus, the a priori specification of moderating variables based on the theory review, variables that are expected to explain differences in effect size magnitude, seems to suggest a fixed-effects model. Even if “the fixed effects assumptions are rejected,” Lipsey & Wilson (2001) explain that one option, which characterizes the strategy taken in this dissertation, is to “continue to assume a fixed effects model, but add the assumption that the variability beyond subject-level sampling error is systematic, that is, derived from identifiable differences within studies” (p. 118). This assumption then forms the basis for the moderating analyses based on sample characteristics such as income level and racial composition. The number of studies to be aggregated in the analyses also plays a role in the decision to use either a fixed or random effects model. According to Raudenbush, in the “extreme case of a synthesis of only two studies, the fixed effect approach seems the most sensible” (1994, p. 307). Given that some of the comparisons between groups in the moderating analyses contained only a few studies, and because the strategy of this dissertation was to test a set of a priori variables that may explain differences in effect size magnitude, a fixed-effects approach seemed, on balance, to make more sense than a random effects model. Thus, the decision was made to use Hedges’ analogue to the ANOVA (1982a), which groups effect sizes into mutually exclusive categories on the

PAGE 101

90 basis of an independent variable and tests the homogeneity among effect sizes within and between the categories. As described in the gender analyses mentioned above, this technique partitions the total homogeneity statistic, Q, into the portion explained by the categorical variable and the residual pooled within-groups portion. The homogeneity statistic Q is the weighted sum of squares of the individual effect sizes around the grand mean. The between groups Q (Qb) is the weighted sum of squares of the mean effect sizes for each group around the grand mean and the within groups Q (Qw) is the weighted sum of squares of the individual effect sizes within each group around their respective group mean, pooled over the groups. The degrees of freedom for Qb is j where j is the number of groups and the degrees of freedom for Qw is k-j where k is the number of effect size statistics and j is the number of groups. Qb is found by subtracting Qw from Qtotal. If Qw is not significant, the categorical variable represented in the within category variance is sufficient to account for the excess variability in the effect size distribution. However, if the between category variance is significant, then the mean effect sizes across groups differ by more than sampling error (Lipsey & Wilson, 2001). Now that the rationale, inclusion criteria, proposed research questions, and meta-analytic strategies have been clarified, the following chapter will present the results of the meta-analysis.

PAGE 102

CHAPTER 6 RESULTS OF THE META-ANALYSIS Broad Overview of Domains to be Meta-Analyzed In total, the meta-analysis involved findings from 113 studies, representing 44,346 married individuals. As shown in Table 6-1, each of the eleven domains of stress involved an aggregation of no fewer than three studies or 1030 married individuals. On the other end of the spectrum, spouses’ health problems were investigated 31 times and economic strain involved 30 studies with combined findings on 17,133 married individuals. Table 6-1. Number of Individual Correlations and Median Sample Size in each Domain of Stress Domain No. of studies No. of participants in this domain Median sample size Spouse’s health issue 31 6970 136 Economic strain 30 17133 373 Work stress 26 10021 184 General stress 24 12378 191 Work-family stress 15 5223 296 Low SES 9 3034 240 Unemployment of a spouse 7 5562 370 Child’s health issue 6 1025 108 War/Military stress 6 3211 245 Race 4 3073 759 Interpersonal stress 3 1030 120 Descriptive Characteristics of Studies in the Meta-Analysis Descriptive characteristics for the 113 studies included in the meta-analysis are illustrated in Table 6-2. As shown, the median and mean years of publication are 1995 91

PAGE 103

92 and 1994, respectively. A median of 222 participants were recruited in the studies in the meta-analysis. Based on this average sample size, one might be tempted to assume that there is adequate power to find a medium sized effect. However, given that each domain of stress was studied separately, statistical power will vary somewhat as it will be based on the total number of studies in each domain. This is important because domains such as the illness of a child tend to include very small, highly targeted samples, such as a sample of parents whose children have been diagnosed with Rett Syndrome. The nature of the stressor lends itself to smaller samples in comparison to domains like general life stress, in which we might expect to find much larger samples. In general, however, the aggregation of multiple samples in meta-analytic technique greatly increases power to find effects (Matt & Cook, 1994). For information about the total number of participants and the average sample size for the studies included in each domain-specific analysis, the reader is referred to Table 6-1. The median income in the samples in the meta-analysis was 1.11 times the national median income reported in U.S. Census Bureau Historical Income Table H-5 for the years associated with those studies that reported incomes. Thus, this figure indicates that the median income for samples in the meta-analysis was slightly higher than the national median income. However, it is important to note that many authors failed to report income levels for participants in their samples. Another potential moderator is whether samples in the meta-analysis are nationally representative or not. Not surprisingly, most of the samples were not nationally representative (n = 102; 91.1%) while only a small proportion of samples (n = 10; 8.9%) were recruited from a national base. As one might also expect, most of the samples in the

PAGE 104

93 meta-analysis were recruited in the United States (n = 85, 75.9%) while a much smaller proportion were recruited in other countries (n = 27; 24.1%). Low income populations were also poorly represented, with only 24 (21.4%) of samples recruited from less than mid-level SES groups, and 86 (76.8%) of samples recruited from middle or high SES groups. Of course, this restriction of range limits the generalizability of potential findings and carries the same limitations that have been mentioned previously in the review: we simply do not have data on the very samples that are likely to be experiencing the highest levels of stress. Furthermore, from a practical point of view, analyzing low income as a potential moderating variable may not be possible in any domains of stress that lack sufficient income level variability. Along the same lines, there is a noticeable restriction in range for racial composition of samples in the meta-analysis. Fully 33% (n = 37) of the samples lacked racial composition data. When these data were reported, 57% (n = 64) of the studies were based on samples that were more than 60% Caucasian. Less than 10% (n = 11) were composed of samples that were less than 60% Caucasian. Again, this lack of racial background variability and the rarity of minority samples raises theoretical and practical limitations, limiting both generalizability and the extent to which it will be possible to conduct moderating analyses for racial background. Comparisons between Studies in the Review and Studies in the Meta-Analysis Even though half of the articles in the meta-analysis overlap with articles in the review, the total number of participants represented in the meta-analysis, 44,346, is nowhere near half the total of 164,359 participants in the larger pool. Yet, as illustrated in Table 6-3, there are many commonalities in the publication, demographic, and

PAGE 105

94 Table 6-2. Summary of Descriptive Characteristics for Studies in the Meta-analysis (n=113) General Characteristics Median year of publication 1995 (SD = 7.96) Mean year of publication 1994 (SD = 7.96) Average sample size (median) 222 (SD = 537) Average income in sample 1.11* (SD = .47) Theoretical Moderators Moderator No. of studies Percent of total studies National representation Nationally representative samples 10 8.9% Non-nationally representative samples 102 91.1% Low income Low income sample 24 21.4% Not low income sample 86 76.8% Not reported 2 1.8% International recruitment International samples 27 24.1% U.S. sample 85 75.9% Racial composition More than 60% Caucasian sample 64 57.1% Less than 60% Caucasian sample 11 9.8% Not reported 37 33.0% * Reported median incomes were compared to the national median income for each respective year, according to U.S. Census Bureau Historical Income Table H-5. 5 Thus, this figure indicates that median incomes for samples in the meta-analysis were 1.11 times as much as the national median income. design characteristics between these two groups despite the fact that the number of participants in the meta-analysis is only 27% of the total number of participants. At least two conclusions may be drawn from Table 6-3. First, on the positive side, the numerous commonalities between the sampling and demographic characteristics of these two sets of studies suggests that the meta-analytic population bears a clear resemblance to the population in the larger pool of articles. Therefore, results obtained in 5 (Source: http://www.census.gov.hhes/income/histinc/h05.html; Table H-5.)

PAGE 106

95 Table 6-3. Comparison of Characteristics of Articles in the Review with those in the Meta-analysis. Characteristic Articles in the Review Articles in the Meta-analysis Median sample size 206 222 Percent of studies conducted in the United States 75% 75% Percent of samples that are nationally representative 89% 90% Median sample income (in comparison with the national median income) 1.12 1.11 Percent of studies that lacked racial composition data 41% 33% For studies reporting these data, percent that had more than 60% Caucasian participants 80% 85% the meta-analysis may be generalized to the character of this literature. However, on the negative side, the commonalities that emerge in studies in both the review and meta-analysis suggest a clear lack of racial and socio-economic heterogeneity. Notwithstanding these notable limitations, a number of interesting results emerged as follows. Presentation of Meta-Analytic Results Aggregation of Effect Sizes In setting up Research Question 1, I suggested that, in general, stress is likely to decrease marital satisfaction. However, it seems likely that various types of stress may affect marital satisfaction to different degrees. Based on this theoretical possibility, I designed research question 1 to identify the relative magnitude of effect sizes between domains.

PAGE 107

96 First, I computed aggregate effect sizes for the association between stress and marital satisfaction for each domain of stress. Results are depicted in Table 6-4, with aggregate effect sizes for each domain listed in descending order of magnitude. Table 6-4 also presents the total number of separate studies that contributed to the aggregate correlations. Each study was allowed to contribute only one correlation to the aggregate analysis. In other words, if, in a particular study, researchers analyzed males and females separately, then these correlations were averaged together and that average correlation was the contributing correlation for that study. As shown in Table 6-4, each domain of stress was associated with decreased marital satisfaction. This converges with other findings suggesting that stress is generally detrimental to marital satisfaction. I also conducted a test of heterogeneity for each stressor domain (see Table 6-4) in order to determine whether the variability in effect size estimates exceeds that expected from sampling error alone (Matt & Cook, 1994). When a heterogeneity test is significant, the implication is that there may be important moderating variables that explain differences in the magnitude of effect sizes for subclasses of studies within a given domain of stress. This was the case for all domains except three: child’s health problems, interpersonal stress, and non-white racial background. It is noteworthy that these three domains were based on very low numbers of studies relative to other domains of stress. Three studies were aggregated in the case of interpersonal stress, four studies in the case of non-white racial background, and six in the case of children’s health problems. Another way to look at this is that the average number of studies contributing to these aggregate effect sizes was 4.3 for these three domains as compared to an average of 18 studies collapsing across the other domains. As Matt and Cook (1994) explain, a small

PAGE 108

97 number of studies may create a lack of statistical power which may account for non-significant findings. Ultimately, based on the evidence of homogeneous distributions in these three domains, I dropped these three domains from the future moderating analyses. In the case of heterogeneous distributions, averaging over contrary results would lead to further confusion about the true magnitude of associations between stressors and marital satisfaction. As such, I planned to conduct moderating analyses of gender, income level, racial composition and national representation for all other stressor domains. In terms of the relative magnitude of aggregate effect sizes, as shown in Table 6-4, the largest effect sizes were obtained for the domains of children’s health problems, unemployment of one spouse, and work-family stress, with aggregate correlations of -.30, -.28, and -.26, respectively. In other words, high levels of these kinds of stressors are associated with relatively low levels of marital satisfaction. Practically speaking, these are medium effect sizes in reference to Lipsey and Wilson’s assertion that an aggregate correlate of .10 represents a small effect size, .25 indicates a medium effect size, and .40 represents a large effect size (Lipsey & Wilson, 2001, p. 147). The negative impact of children’s health problems converges with other research findings that have explored the impact of a child’s death on married couples (e.g., Schwab, 1992). Feeley & Gottlieb (1988) have proposed that withdrawal due to mourning the loss of a child may cause breakdowns in marital communication which may help explain the deterioration of marital satisfaction. At the other end of the spectrum, the smallest effect sizes were observed for non-white racial background and low SES/occupational status, with aggregate correlations of -.06 each. In other words, it would appear that being non-white

PAGE 109

98 or having low SES/occupational status is not strongly associated with decreased marital satisfaction. Some of the dimensions of stress described in the preceding review and the set-up of Research Question 1 may elucidate differences between the sizes of aggregate correlations. Generally speaking, the magnitude of these associations may be understood in light of the acute or chronic nature of the stressors. Specifically, the stressors that are associated with the lowest levels of marital satisfaction, children’s health problems and unemployment of a spouse, are likely to be acute stressors, while the stressors associated with the highest levels of marital satisfaction, having a non-white racial background or low socio-economic status, are likely to be chronic stressors. Acute stressors tend to be time-limited (Gump & Matthews, 1999) with a “relatively clear onset and offset” (Wheaton 1997, p. 52). On the other hand, chronic stressors have been defined as stressors that have been in effect for at least one year (McGonagle and Kessler, 1990). In thinking about the interaction between acute and chronic stressors, Taylor and Aspinwall (1996) have stated that chronic stressors form the “backdrop” (p. 91) against which acute stressors occur. The distinction between chronic and acute stressors is critical for several reasons. Acute stressors like unemployment are more proximal to marital functioning relative to more distal chronic stressors like pervasive experiences of racism. Further, relative to chronic stressors like being a minority in a dominant culture or having low socio-economic status, acute stressors often come with no warning. A diagnosis of cancer in one’s child or the sudden loss of a job are events which cannot be predicted in most cases. As Figley (1983) has noted, a lack of time to prepare for the impact of a stressor is

PAGE 110

99 likely to exacerbate the impact of that stressor. In support of this notion, the anticipation of a child’s death has been shown to temper subsequent grief reactions in some cases (Parkes & Weiss, 1983; Shanfield, Benjamin, & Swain, 1988). Further, as Hill (1964) has pointed out, the amount of previous experience one has had with a stressor seems to be related to the ability to muster various types of coping resources. Couples are not likely to have much experience dealing with relatively rare acute stressors like serious medical problems in one’s child or the sudden loss of a job. Supportive evidence emerges in Caspi, Bolger, and Eckenrode’s finding that previous experience with major life events decreases the impact of stressful life events (1987). Moreover, these stressors entail a real or possible loss of great value to a couple. The relative rarity of certain types of acute stressors may also heighten a sense that couples are alone and have few sources of guidance in coping with a stressor. Conversely, although being a minority or having low socio-economic status is certainly a chronic strain, there are likely to be others in one’s community that can normalize experiences of discrimination or offer emotional support. Research on African American communities has identified the existence of elaborate social networks that include extended family, friends, neighbors, and church members (e.g., Chatters, Taylor & Jackson, 1986). According to Taylor (1986), these networks represent a vital system of mutual aid in the form of financial support, emotional comfort, advice, and services. Thus, it is possible that differences in the magnitude of associations between various stressors and marital satisfaction may be understood in light of the suddenness, severity, and rarity of various stressors.

PAGE 111

100 Table 6-4. Aggregate Effect Sizes of Specific Domains of Stress on Marital Satisfaction Domain Effect Size No. of Studies p-value for test of significant difference from zero Q value p-value for test of heterogeneity* Child’s health problem -.30 6 <.001 7.68 .175 Unemployment of a spouse -.28 7 <.001 206.37 <.001 Work-family stress -.26 15 <.001 139.56 <.001 General stress/Life events -.19 24 <.001 98.96 <.001 War/military stress -.18 6 <.001 43.81 <.001 Interpersonal stress -.18 3 <.001 3.48 .175 Spouse’s health problem -.16 31 <.001 221.94 <.001 Economic strain -.13 30 <.001 150.86 <.001 Work stress -.13 26 <.001 93.33 <.001 Non-white race -.06 4 <.001 1.41 .704 Low SES/Occupational status -.06 9 .003 34.01 <.001 * The test of heterogeneity determines whether the variability in effect size estimates exceeds that expected from sampling error alone (Matt & Cook, 1994). If the heterogeneity test is significant, the implication is that there may be important moderating variables that explain differences in the magnitude of effect sizes for subclasses of studies within a given domain of stress. Test of Gender Differences Research Question 2 asked whether there will be a significant gender difference in the magnitude of aggregate correlations between stressors and marital satisfaction for any of the domains of stress. While it is possible that stress affects males and females equally, empirical examples support some interesting gender differences for various types of stressors, most notably in the domains of work-family stress, unemployment, and spouse’s and children’s health issues. Findings do not suggest that there will be gender differences for work stress and there is a dearth of evidence regarding gender differences in the other four domains of stress, which are general stress, war/military stress, economic strain, and low socio-economic status. Since I have dropped child’s health problems from future moderating analyses, I hypothesized that gender differences would emerge for work-family stress, unemployment, and spouse’s health issues. I did not

PAGE 112

101 expect to find gender differences for work stress and I ran exploratory analyses for general stress, war/military stress, economic strain, and low socio-economic status. To address the question of gender differences, I computed aggregate effect sizes for males and for females within each stress domain and then conducted statistical tests of gender differences. Table 6-5 illustrates the results for the gender analyses. Unlike Table 6-3, which reported the number of studies contributing to the aggregate analysis, Table 6-5 reports the number of effects involved in each condition. In this analysis, for studies that analyzed each gender separately, the correlations of female participants were not averaged with the correlations of male participants in order to explore any possible gender differences. Therefore, studies that analyzed each gender separately contributed two correlations, one for females, which was aggregated with the correlations of other groups of females in a given domain, and one for males, which was aggregated with the correlations of other groups of males in a given domain. As illustrated in Table 6-5, significant gender differences were found for unemployment, work-family stress, and spouse’s health problems. Thus, results conformed remarkably well with respect to hypothesized gender differences, with females having significantly more negative aggregate correlations for the domains of unemployment, work-family stress, and spouse’s health problems. It might be tempting to conclude that these gender differences can be explained by disproportionate sampling of males and females that have been directly affected by these stressors. For instance, for unemployment, one might hypothesize that males are more likely to be affected by the loss of their own jobs. However, researchers collected data on unemployment for female spouses in fairly equal proportions. That is, as illustrated in

PAGE 113

102 Table 6-1, the total number of participants in studies of unemployment stress was 5562 across seven different studies. Three of these studies focused on employment of the male spouse in a relationship and these three studies included 2431 participants (44% of the total). However, one very large study of 2134 participants looked at unemployment for female spouses, which represents 38% of 5562. Three additional studies looked at unemployment in either the male or female spouse, and these samples totaled 997 participants (18% of the total). So, we cannot conclude that gender differences can be explained on the basis of disproportionate stressor impact of the unemployment on the male partners in the relationships represented. Similarly, assumptions about a disproportionate representation of males and females in the domains of work-family stress and spouse’s health problems are not warranted. Half of the studies in the work-family domain (representing 2642 participants) report data on both working husbands and wives. In the other half of the studies in this domain, 33% (representing 1709 participants) were focused on males’ work-family stress and 17% (representing 872 participants), were focused on females’ work-family stress. For the domain of spouse’s health problems, there were a total of 6970 participants in the 31 studies exploring spouse health problems. The proportion of females and males in this pool of participants was nearly equal, with 3546 female participants (51% of the total) and 3425 male participants (49% of the total). As such, any conclusions about stressor impact should be contextualized by the equal representation of females and males with health problems in this domain. Having dismissed the possibility that disproportionate sampling explains these differences, there are at least two possibilities that may explain these findings. First, a

PAGE 114

103 common theme in these domains is the likelihood that unemployment, work-family stress, and spouse’s health problems will cause disruptions in spouses’ marital roles. Consider the following perspective of a low-income mother who asked her partner to leave when he lost his job: I told [him] he had to leave, even though I knew it wasn’t his fault that [he wasn’t working]. But I had nothing in the house to feed the kids, no money to pay the bills. Nothing. And he was just sitting there, not working. I couldn’t take it, so I made him leave. (Edin, 2000, p. 119) The additional economic strain generated by stressors like unemployment and a spouse’s health problem may also lead to increased role strain. Stressors that cause disruptions in marital roles may be especially challenging for couples whose marriages are characterized by more traditional gender roles. For example, it might seem intuitive that seeking external support in the midst of economic strain would be a wise and positive course of action. In fact, when wives in families with economic strain seek external support, their distress is often lessened (Robertson, Elder, Skinner, & Conger, 1991). However, if the husband is the primary breadwinner, seeking this external support may simultaneously underscore his failure to fulfill his role as a good provider, and, not surprisingly, wives’ seeking of external support has been associated with husbands’ negativity towards their wives (Robertson, Elder, Skinner, & Conger, 1991). In further support of the role strain explanation is the assertion that role theory may explain the disproportionately higher incidence of psychological distress and emotional pathology found among female physicians and trainees as compared to their male counterparts (Coombs & Hovanessian, 1988, p. 21). Second, it may be that these types of stressors may have a disproportionately negative impact on marital satisfaction for wives, who often shoulder the bigger burden

PAGE 115

104 of emotional and physical care-giving in the family. Discriminant validity for this possibility surfaces in the non-significant gender differences for the other five domains of stress, which are general stress, war/military stress, low socio-economic status, work stress, and economic strain. Relative to spousal health problems, work-family stress, and stress due to unemployment, these domains seem more removed from the home setting. Research findings indicate that wives often bear the brunt of “mental hygiene” responsibilities and physical care-taking in marriages. For parents of disabled children, mothers, but not fathers, experienced higher levels of stress compared to parents of non-disabled children (Kazak, 1987). In a study of couples in which one partner has multiple sclerosis, the female partners were more dissatisfied with the relationship relative to male partners (Dupont, 1996). Gottman and Krokoff (1989) hypothesized that wives have to manage a complex dialectic in their role as “the manager of marital disagreements” (p. 52). Along similar lines, Burke and Weir (1977) found that wives have a greater awareness of concrete helping behaviors which they can employ to help their husbands deal with stress and were more active in the performance of these behaviors. These authors suggest that different socialization experiences may better prepare women to serve in helping roles and set norms that women should “set the stagefor the alleviation and resolution of anxieties and tensions arising for either partner in the marital pair” (Burke & Weir, 1977, p. 920). To summarize, then, gender difference for the domains of unemployment, work-family stress, and spouse’s health problems may be explained by disruptions in marital roles and/or disproportionate burdens of care-giving responsibilities placed on females.

PAGE 116

105 Table 6-5. Aggregate Effect Sizes within each Domain for Females and Males and Tests of Gender Differences Domain Males (# Effects) Females (# Effects) Females have a more negative aggregate correlation? Between groups Q-value (df = 1) p-value for test of difference Unemployment of a spouse -.27 (6) -.39 (6) Yes 15.61 <.001 Work-family stress -.21 (8) -.32 (8) Yes 15.36 <.001 Spouse’s health problem -.15 (15) -.22 (16) Yes 5.03 .025 Economic strain -.09 (18) -.11 (19) N/A 1.42 .234 War/Military stress -.18 (4) -.24 (3) N/A 1.20 .274 General stress/Life events -.19 (17) -.20 (18) N/A .163 .686 Work stress -.13 (13) -.12 (12) N/A .087 .769 Low SES/Occ. status -.02 (4) -.02 (4) N/A .004 .951 Moderating Analyses based on Sampling Characteristics In addition to exploring gender as a potential moderating variable, I also conducted moderating analyses of three additional sampling characteristics. The following results were obtained for each of the moderating analyses. Income Level Theory and empirical research have suggested that financial resources moderate the impact of stress on marital satisfaction (e.g., Broman, Hamilton, & Hoffman, 1990; Cook, Hoffschmift, Cohler, & Pickett, 1992; Cutrona, Russell, Abraham, Gardner, Melby, Bryant, and Conger, 2003; Ridley & Wilhelm, 1988; Thoresen & Goldsmith, 1987; Vinokur, Price, & Caplan, 1996). There are at least two mechanisms to explain the theorized moderating impact of income level. One possibility is that financial resources can provide opportunities for leisure and temporary reprieves from stressful circumstances. For example, a major stressor perceived by spouses of medical students was “limited finances for recreation and non-essentials” (Perlow & Mullins, 1976, p.

PAGE 117

106 726). Along similar lines, Park and colleagues (2002) interviewed married individuals with low income levels and were told: If you have no money, it’s difficult tobe together, to do fun things, to be at peace, to come home to a havenbecause if you have no money, bills are not paid, you not gonna get rest when you get home [sic]. (Park, Turnbull, & Turnbull, 2002, p. 151) Another empirically-supported possibility is that financial resources buffer the impact of external stressors. For instance, higher levels of family income seemed to partially protect marriages in two samples of parents of children with severe mentally illnesses or developmental disabilities (Cook, Hoffschmift, Cohler, & Pickett, 1992; Willoughby & Glidden, 1995). On the other end of the spectrum, the impact of stressors is likely to be exacerbated for families without financial resources. Again, interview data succinctly describes this idea: [Economic deprivation leads to] “feelings of insecurity which floods over into other problems[causing] anger, bitternessand then it jumps on the other family members [sic].” (Beach Center, 1999, in Park, Turnbull, and Turnbull, 2002, p. 151) Therefore, research question 3 proposed that lower income samples will have more negative aggregate correlations between stress and marital satisfaction. Unlike the gender analyses, in the income analyses, each study was allowed to contribute only one correlation to the aggregate analysis. In other words, if, in a particular study, researchers analyzed males and females separately, then these correlations were averaged together and that average correlation was the contributing correlation for that study. As shown in Table 6-6, income emerges as a significant moderator of associations between stress and marital satisfaction in 5 out of the 8 domains for which analyses were conducted, but not always in the direction of expected effects. Low-income samples have more negative aggregate correlations in the domains of unemployment, war/military

PAGE 118

107 stress, and work stress, while higher income samples have more negative aggregate correlations in the domains of work-family stress and spouse’s health problems. In reference to Lipsey and Wilson’s assertion that an aggregate correlation effect size of .10 represents a small effect size, .25 indicates a medium effect size, and .40 represents a large effect size (Lipsey & Wilson, 2001, p. 147), the difference between low-income and non-low income groups in the case of war/military stress (.22) was medium sized. Differences were relatively smaller in the domains of work stress (.16), unemployment (.14), work-family stress (.14) and spouse’s health problem (.11). Finally, income does not moderate associations between general stress, low socio-economic status, and economic strain and marital satisfaction. It is noteworthy, however, that that there were only two studies that could be coded as “low-income” in several analyses. The total number of participants in the unemployment stress analysis was 2351, and most of these participants were gleaned from a large geographically diverse sample based in the United States (n=2101). The other study in the unemployment domain was recruited in Finland (n=250). The war/military stress analysis also included only two “low-income” groups, one group recruited in California (n=406) and one group recruited in either Houston or Chicago (n=80) for a total of 486 participants across the two studies. The two studies in the work-family stress analysis were based in Iowa (n=674) and New Mexico (n=59) and contained a combined number of 733 participants. Finally, the work stress analysis included only one “low-income” group, recruited in the Pacific Northwest (n=178). Ultimately, such mixed results defy simple interpretations. Those with lower levels of financial resources have more negative aggregate correlations when the stressor is

PAGE 119

108 unemployment, war/military stress, or work stress, but less negative aggregate correlations relative to their more affluent counterparts when the stressor is work-family strain or a spouse’s health problem. One possible way to explain such results is that income level may be operating as a proxy variable for relative status in society. If this is true, then samples populated by individuals with lower incomes are likely to be populated by individuals with lower status in society. In the work and military settings, status determines one’s place in the organizational hierarchy, which is then a key determinant of autonomy, flexibility, and various forms of compensation, both financial and social. Applying, and admittedly stretching, the same logic, it may also be the case that those with lower status are more likely to be laid off, and endure more stress while looking for a suitable replacement position. Conversely, in the case of work-family stress, and spouse’s health problems, relative status may not be as salient, as these stressors are fairly normative and equally distributed throughout the general population. A second possible way to explain these findings is that community-based social support may be more easily accessed or more likely to be mobilized for certain types of stressors. Specifically, in the case of work-family stress, childcare duties may be shared by extended family networks. In the same way, the extra burden of caring for a sick spouse may be shouldered in part by others in one’s kin or community network. Conversely, stressors like seeking a new job or dealing with the stressful job that one has may be seen as activities best left to individuals. However, any interpretations based on these data must be tempered because results are mixed and the magnitude of the effects which were found is in the small to medium range.

PAGE 120

109 Table 6-6. Effect Size by Domain and Income Collapsing over Gender Domain Low (# Effects) Not Low (# Effects) Between groups Q-value (df=1) Low-income samples have more negative aggregate correlations? p-value for test of difference Unemployment of a spouse -.34 (5) -.20 (2) 28.98 Yes <.001 War/Military stress -.37 (2) -.15 (4) 23.97 Yes <.001 Work-family stress -.14 (2) -.28 (13) 12.70 No <.001 Spouse’s health problem -.06 (6) -.17 (24) 9.83 No .002 Work stress -.29 (1) -.13 (25) 4.73 Yes .030 General stress -.16 (5) -.20 (18) 3.80 N/A .051 Low SES/Occ. status -.08 (2) -.05 (7) .759 N/A .384 Economic strain -.13 (11) -.14 (17) .417 N/A .519 Racial Composition of Sample Research has documented that African Americans have significantly less education, lower job status, and lower income levels compared to Caucasians (e.g., Scanzoni, 1975). Therefore, the pervasive effects of racism were expected to function as “background” stressors (Gump & Matthews, 1999), causing greater stress pile-up for those of non-white racial background. As McAdoo (1983) has asserted, “while degrees of racial oppression vary with each situation, the potential for being devalued and put down is always present, dangling constantly over Blacks, beyond their control.” (p. 180). The awareness of such pervasive disparities is likely to result in feelings of demoralization, which Hill (1949) has defined as a loss of morale and unity within the family. Convergent validity for this proposition may be observed in consistent empirical findings that African Americans are less satisfied with their marriages relative to Caucasians (Broman, 1993; Dillaway & Broman, 2001; Renne, 1970). Thus, research question 4 proposed that racial composition will moderate the magnitude of the effect size between domains of stress and marital satisfaction, such that samples that are less than 60% Caucasian will have more negative

PAGE 121

110 aggregate correlations between stress and marital satisfaction relative to those that are more than 60% Caucasian. A test of moderation was conducted for six domains: work stress, war/military stress, economic strain, general stress, spouse’s health problems, and work-family stress. Analyses could not be conducted for unemployment stress or low SES/Occupational status due to a lack of information needed to code these studies or a lack of representation in more than one level of the moderating variable. As in the case of the income analyses, each study was allowed to contribute only one correlation to the aggregate analysis. In other words, if, in a particular study, researchers analyzed males and females separately, then these correlations were averaged together and that average correlation was the contributing correlation for that study. As illustrated in Table 6-7, racial background emerges as a significant moderator of associations between work stress, war/military stress, and economic stress and marital satisfaction. For the work stress and war/military stress domains, the direction of associations was as hypothesized; samples that were less than 60% Caucasian had more negative aggregate correlations between stress and marital satisfaction. For the domain of economic stress, the results were in the opposite direction of what was hypothesized; samples that were less than 60% had less negative aggregate correlations. The differences in the magnitude of aggregate correlations were .15 for work stress, .26 for war/military stress, and .05 for economic stress. According to Lipsey and Wilson’s criteria (2001), the .15 difference (work stress) lies in the small to medium range, the .26 difference (war/military stress) represents a medium difference, and the .05 difference (economic stress) indicates a very small difference.

PAGE 122

111 Even though studies were coded using the theoretically permissive criteria of whether samples were “less than 60% Caucasian,” it is important to note that there were only a few studies that could be coded as “less than 60% Caucasian” in several domains. For example, in the work stress analysis, there were only two studies, representing a total of 1486 participants with most of these participants gleaned from a large geographically diverse sample based in the United States (n=1386). The other study in the work stress domain was a sample of working doctors in Lucknow or Kanpur, India. Thus, due to the meta-analytic weighting process, the significant moderation in the domain of work stress is based largely on the United States sample. Further, this finding is corroborated by McAdoo’s contention that the “pervasive daily reality” of racism for African Americans “occurs when employers do not consider Blacks to be capable of doing a job regardless of their individual abilities or training” (1983, p. 180). Further, there is a much higher likelihood that African-Americans have demoralizing, stressful jobs because they are “concentrated in low-level jobs with poor benefits” (McAdoo, 1983, p. 183). Likewise, in the domain of war/military stress, there were only two studies that could be coded as “less than 60% Caucasian.” The total number of participants in the “less than 60% Caucasian” condition was 292 for war/military stress and these samples were Israeli families living under the threat of forced re-location in the Golan Heights (n=212), and a group of Bosnian refugees who had recently re-located to two large American cities (n=80). Based on the nature of these samples, it is not clear that racial composition is the variable that drives the significant moderation effect. These two samples are either currently facing, or have recently endured, the severe stress of a sudden uprooting and forced re-location to another geographical region. Thus, the threat

PAGE 123

112 or actual event of re-location under duress may explain why these samples, which were also “less than 60% Caucasian” had more negative aggregate correlations between war/military stress and marital satisfaction relative to the “more than 60% Caucasian” comparison groups, which were mostly comprised of active military personnel. In the case of economic strain, samples that are less than 60% Caucasian samples had a less negative aggregate correlation between economic strain and marital satisfaction. It may be that samples with a greater proportion of non-Caucasian participants are more familiar with economic strain, and have habituated somewhat to this type of chronic stress. In support of this notion, African-Americans receive less education, and are faced with more job discrimination and higher unemployment rates, all resulting in lower incomes for Black families. (McAdoo, 1983, p. 180). However, this finding was only marginally significant and the difference between the aggregate effect sizes was very small, so firm conclusions based on this finding are not warranted until further research is conducted. It was also somewhat surprising that racial background did not moderate the association between stress and marital satisfaction in the domains of work-family stress, spouse’s health problems, and general stress. As mentioned previously, the pervasive effects of racism were expected to function as “background” stressors (Gump & Matthews, 1999), causing greater general stress pile-up for those of non-white racial background. Furthermore, I particularly anticipated significant moderation for the domain of spousal health issues due to racial disparities in access to healthcare. Specifically, black families are less likely than white families to have adequate health insurance and, not surprisingly, surveys of black families indicate that they tend to feel that their health

PAGE 124

113 care is poor (McAdoo, 1982). The overall mortality rate for African American has also been consistently higher than for Caucasians, and a survey of black parents found that many partially attributed this higher death rate to the psychological and social pressures of racism in their everyday environment (McAdoo, 1982). One possible explanation for these null findings is that restrictions of range in the samples in the meta-analysis did not allow enough power to obtain any significant findings for these domains. In fact, as Table 6-7 illustrates, there is only one sample that could be coded as “less than 60% Caucasian” in the domains of both work-family stress and spouse’s health problems. The sample in the work-family analysis was a group of working class (e.g., plumbers, carpenters, electricians) Mexican Americans (n=59) and the relatively non-Caucasian sample in the spouse’s health domain was recruited in Hong Kong (n=132). In fact, a lack of racial diversity in these samples is a repeating issue across several domains, making the significant findings that were obtained all the more surprising. However, in the case of general stress, there were six samples that were not more than 60% Caucasian, which should have allowed for adequate power to find potentially significant results. It may be the case that excessive heterogeneity in the category of general stress prevented meaningful patterns from emerging. Overall, then racial composition may be an important moderating variable in that samples who were less than 60% Caucasian had significant associations with notable differences in effect size magnitude for two domains of stress, work stress and war/military stress. These results mirror the results of the income level analyses, which also showed significant moderation in the same direction. However, due to low power in several domains, and the crudeness of the measure of racial composition, future research

PAGE 125

114 is needed to determine whether and how racial background moderates associations between stress and marital satisfaction. Table 6-7. Effect Size by Domain and Racial Composition Collapsing Across Gender Domain Less than 60% Caucasian (# Effects) More than 60% Caucasian (# Effects) Samples less than 60% Caucasian have a more negative aggregate correlation? Between groups Q-value (df=1) p-value for test of difference Work stress -.25 (2) -.10 (15) Yes 30.92 <.001 War/Military stress -.42 (2) -.16 (4) Yes 21.01 <.001 Economic strain -.10 (4) -.15 (20) No 5.47 .019 Work-family stress -.28 (1) -.19 (9) N/A .547 .459 Spouse’s health problem -.12 (1) -.17 (21) N/A .332 .564 General stress -.17 (6) -.18 (8) N/A .06 .811 Unemployment of a spouse n/a* n/a* n/a* n/a* n/a* Low SES/Occ. status n/a* n/a* n/a* n/a* n/a* * Could not conduct this analysis due to lack of information needed to code studies by this moderator or lack of representation in more than one level of the moderating variable National Representation Restriction of range has been a recurrent theme in this dissertation. Some have explained that restriction of range results in attenuated correlations (e.g., Matt & Cook, 1994). Research Question 5 explores restriction of range inherent in sampling procedures for the studies in the meta-analysis. Relative to geographically constricted samples, range restriction is less likely to be a problem in nationally representative samples, due to their large size and broad demographic span of participants. Therefore, it is possible that the correlations derived from nationally representative samples may yield larger estimates of associations between stressors and marital satisfaction. This analysis was designed to ascertain whether national representation of samples will moderate the magnitude of the effect size between domains of stress and marital satisfaction, such that those samples that are nationally representative will have more negative aggregate correlations than those that are not nationally representative. As in the case of previous analyses, in the national representation analyses, each study was allowed to contribute only one

PAGE 126

115 correlation to the aggregate analysis. In other words, if, in a particular study, researchers analyzed males and females separately, then these correlations were averaged together and that average correlation was the contributing correlation for that study. Moderating analyses based on national representation were conducted in four domains: work stress, work-family stress, economic strain, and general stress. Analyses could not be conducted for low SES/Occupational status, spouse’s health problems, unemployment stress, or war/military stress due to a lack of information needed to code these studies or a lack of representation in more than one level of the moderating variable. As illustrated in Table 6-8, nationally representative samples had a more negative aggregate correlation in the domain of work stress, but not in the domain of work-family stress. In practical terms, the -.11 and +.09 differences in correlation magnitude for work stress and work-family stress could be considered relatively small differences (Lipsey & Wilson, 2001). National representation did not moderate the magnitude of the association between stress and marital satisfaction in the case of economic stress or general stress. Although there were only 2-3 nationally representative samples in each condition for these analyses, restriction of range is not likely to be as much of a problem as in past analyses, because nationally representative samples are very large and the meta-analytic technique accords greater weight to studies with larger sample sizes. For the work stress analysis, the total combined sample size of the three nationally representative studies was 2747 with two of these samples recruited in the United States and the third recruited in the Netherlands. For the work-family stress analysis, the total combined sample size of the three nationally representative studies was 1175 with two of these samples recruited

PAGE 127

116 in the Netherlands and the third recruited in the United States. For the economic stress analysis, the total combined sample size of the three nationally representative studies was 2086 with one sample recruited in the Netherlands and one sample recruited in Czechoslovakia. Finally, for the general stress analysis, the total combined sample size of the three nationally representative studies was 3624, with both of these samples recruited in the United States. Overall, then mixed results were obtained when national representation was tested as a moderating variable. As such, future research is needed before any firm conclusions are warranted. Table 6-8. Effect Size by Domain and National Representation Collapsing over Gender Domain NatRep Sample (# Effects) Not NatRep (# Effects) Nationally representative samples have a more negative aggregate correlation? Between groupsQ-value (df=1) p-value for test of difference Work stress -.21 (3) -.10 (23) Yes 28.29 <.001 Work-family stress -.19 (3) -.28 (12) No 8.81 .003 Economic strain -.17 (2) -.13 (28) N/A 3.12 .077 General stress -.21 (2) -.18 (22) N/A 1.66 .198 Low SES/Occ. status n/a* n/a* n/a* n/a* n/a* Spouse’s health problem n/a* n/a* n/a* n/a* n/a* Unemployment of a spouse n/a* n/a* n/a* n/a* n/a* War/Military stress n/a* n/a* n/a* n/a* n/a* * Could not conduct this analysis due to lack of information needed to code studies by this moderator or lack of representation in more than one level of the moderating variable Summary of Meta-Analytic Findings To summarize, the stressors associated with the lowest levels of marital satisfaction, children’s health problems and unemployment of a spouse, are likely to be acute, while the stressors associated with comparatively higher levels of marital satisfaction, having a non-white racial background or low socio-economic status, are likely to be chronic stressors. Stressors that are proximal, unexpected, severe, and relatively rare may also have a more negative impact on marriages. Although being a

PAGE 128

117 minority or having low socio-economic status is certainly a chronic strain, there may be others in one’s community that can normalize experiences of discrimination or offer emotional support. Alternatively, habituation to chronic stress may mute the impact of stressors. Gender differences in the domains of unemployment, work-family stress, and spouse’s health problems cannot be written off as an artifact of disproportionate representation of males and females in these domains. It is more likely that gender differences in these domains may be explained by disruptions in marital roles and/or disproportionate burdens of care-giving responsibilities placed on females. Results of moderation analyses of income suggest that samples with lower levels of income have more negative aggregate correlations between stress and marital satisfaction for the domains of unemployment, work, and war/military stress but not in the domains of work-family stress and spouse’s health problems. Similarly, the analysis of racial composition indicates that samples that are less than 60% Caucasian have more negative aggregate correlations between stress and marital satisfaction in the domains of work stress and war/military stress but not in the domain of economic stress. When national representation was tested as a moderating variable, nationally representative samples had a more negative aggregate correlation in the domain of work stress but not in the domain of work-family stress. It is possible that discriminatory power hierarchies in some settings, such as work and military settings, may help explain why those samples with lower income levels and those samples that were relatively less Caucasian may have a more negative aggregate correlation between stress and marital satisfaction in some cases but not in others. However, the two samples in the racial composition analyses were also

PAGE 129

118 threatened by, or undergoing, a process of cultural displacement so this may have been driving these significant moderation results. Alternatively, extended family or community social support may be offered more frequently when stress is normative and visible, as when a member in a couple has health problems, or childcare help is needed when both care-takers are employed. However, another possible interpretation is that these mixed moderation results may be due to low power in several domains, and future research is needed to determine whether and how variables such as income level, racial composition, and national representation moderate associations between stress and marital satisfaction. Table 6-9 offers a parsimonious, holistic view of the meta-analytic findings. Looking up and down the columns, relative to other moderators that were tested in these analyses, the gender analyses are most consistent in terms of finding hypothesized results. Correlations between unemployment, work-family, and spouse’s health problem stress and marital satisfaction were more negative for females than for males. Income level and racial composition also have some explanatory power to account for consistent differences in effect sizes for war/military stress and work stress. Looking across the rows, it appears that work stress, war/military stress, and unemployment are the domains that are most often moderated in the expected direction. Viewing these results together, an intriguing possibility is that relative status may help explain some of the discrepant findings. In the home domain, the greater proportion of household and care-taker duties often fall to females, who typically earn less money and hold lower-status positions relative to their male partners. In other domains such as the work and military domains, racial composition and income level may simply be proxy variables for relative status in hierarchies, which is a key determinant of autonomy,

PAGE 130

119 flexibility, and various forms of compensation, both financial and social. Conversely, status may be less salient in domains like general stress or economic strain, where the more important question may be whether one has the requisite “resources” (financial and emotional) to cope with these strains. Table 6-9. Statistically Significant Results at the p<.05 level across Stress Domains and Moderators. DOMAIN OF STRESS Gender Income Level Racial Composition National Representation Unemployment of a spouse + + + + Work-family stress + General stress + War/Military stress + + + Spouse’s health problem + + + + Work stress + + + + + + Economic strain + Low SES/Occ. status * Positive (+) signs indicate that results obtained conformed to the hypothesized direction of results. Negative ( ) signs indicate that results were contrary to the expected direction of results. ** According to Lipsey and Wilson (2001), .10 represents a small effect size, .25 represents a medium effect size, and .40 represents a large effect size (2001). Therefore, one (+) or ( ) sign indicates an effect size difference that .10 or less, two (+ +) or ( ) signs indicate an effect size difference that is between .11 and .25 and three ( + + + ) or ( ) signs indicates an effect size difference that is greater than .25. Factors that Limit Interpretations of these Results There are several limitations, both general to meta-analyses and specific to this meta-analysis, that temper interpretations of these results. In general, meta-analysis is only as good as the studies that go into computations of aggregate correlations. This has often been referred to as the “garbage in, garbage out” phenomenon. In lieu of this, the empirical review of articles in the meta-analysis revealed some troubling deficits in the quality of many of these studies. For example, although numerous problems have been linked to the exclusive reliance on self-report measures, the vast majority of the studies in the meta-analysis have depended solely on self-report data. Of those studies that are longitudinal, researchers commonly collected only two waves of data, which does not

PAGE 131

120 allow for the exploration of meaningful patterns over time. Moreover, an average study length of 2-3 years is not very long in the course of marriage, and the use of brief research designs may obscure possible patterns in how spouses handle stressors and major life transitions over time. These last two methodological issues were not directly problematic in the current meta-analysis because results are based on cross-sectional correlations. However, these methodological issues may serve as study quality indices of a sort. Furthermore, meta-analysis cannot summarize research that is not part of the base of studies included in the meta-analysis (Lipsey & Wilson, 2001, pp. 163). Studies are often conducted in non-naturalistic settings, and in ways that may distort the results that are obtained. For instance, social desirability motivations may be one factor that biases self-reports of marital satisfaction in these studies. However, this problem is inherent to most social science research. Another problem particularly relevant to meta-analyses has been called the “file drawer problem.” That is, the publication process often self-selects for significant results. This was not deemed to be a significant problem in this meta-analysis for two reasons. First, in many cases, the correlations that were plucked from the 113 studies in the meta-analysis were simply by-products of other analyses. That is, correlation matrices were often published within studies that primarily focused on associations other than stress and marital satisfaction. A good example of this is the numerous studies in the medical field that looked at coping with various health problems and included marital satisfaction as one variable among many. Therefore, because the association between stress and marital satisfaction was not central, non-significant findings would not have relegated these studies to the “file drawer.” Furthermore, the

PAGE 132

121 restriction of range that I have repeatedly described was expected to attenuate any correlations, possibly countering the threat of overestimating effect sizes. There are also several limitations specific to this meta-analysis. First, some domains of stress may have been too heterogeneous to allow meaningful patterns to emerge. For example, several sub-categories were subsumed under the category of general stress, including daily stress, life stress, stressful life events, transitions, and life hassles. Similarly, war/military stress included stressors such as the deployment of a spouse during the Gulf War, the after-effects of exposure to war trauma, and forced relocation of families due to ongoing war. In order to gain power for the analyses, it was necessary to limit the number of domains of stress. Clearly, though, the price that was paid is evidence of substantial within-domain variance when different types of stress are categorized together. Second, even though the number of domains was limited to eleven, some domains, such as interpersonal stress which included just three studies, were visibly anemic. Although common lore suggests that in-law relationships cause married couples a lot of extra stress, apparently researchers have generally neglected to study of this type of stress! Third, another compromise involved coding studies categorically. It may have been preferable to run analyses of variables along spectra rather than according to very broad, somewhat crude classifications. However, the meta-analytic program I used to compute the analyses required that data be entered in categorical fashion. Fourth, the moderating variables that I tested are not completely independent and are really just proxy variables for elements of demography and design that seem

PAGE 133

122 theoretically important. Variables that are more distal relative to a phenomenon of interest produce a greater “signal to noise ratio.” An example of this is the racial composition moderator, as the categorical coding of “greater than 60% Caucasian” was somewhat arbitrary and a relatively crude measure of the additional strain of pervasive discrimination faced by members of minority groups living in a dominant culture. Finally, as I have stated repeatedly, the results I obtained appear to be based on a very restricted sample of predominantly middle class, white participants so findings are not largely based on the samples likely to be experiencing the highest levels of stress. Strategy for Final Chapter Yet, despite all of these limitations and compromises, this dissertation makes a critical contribution to our understanding of associations between stress and marital satisfaction. I have personally inspected over 600 articles to locate 113 articles that provided correlations that could be meta-analyzed. Thus, this dissertation represents the amalgamation of data on 44,346 married individuals, a massive amount of research looking at how stress impacts marital satisfaction. Further, the use of meta-analytic techniques has led to a dramatic increase in power relative to findings reported in individual studies (Schmidt, 1992). Each of the eleven domains of stress involved an aggregation of no fewer than three studies or 1030 married individuals. On the other end of the spectrum, spouse’s health problems have been investigated in 31 cases and findings for the impact of economic strain are based on data for 17,133 married individuals. In addition to yielding much greater statistical power, this meta-analysis has been valuable not only in terms of synthesizing current findings, but also in gaining a clear view of gaps in the literature and directions for future research, which will be discussed in the next chapter.

PAGE 134

CHAPTER 7 CONCLUSIONS AND FUTURE DIRECTIONS Brief Review of the Rationale This dissertation was sparked by the recognition that the frequently-cited 50% divorce rate statistic obscures a critical reality: marital dissolution is a problem that disproportionately affects some populations more than others (Bramlett & Mosher, 2001). According to census data, rates of marital dissolution are nearly twice as high in low-income populations relative to those with average or above-average income levels. (Bramlett & Mosher, 2001, Figure 26). Reflection on this finding gave rise to the guiding premise of this dissertation: marriages are shaped by the contexts in which they occur. The idea of considering context is not new. However, although several have called for increased emphasis on the larger contexts within which marriages are embedded (e.g., Berscheid, 1998; Huston, 2000; Kurdek, 1991; Levinger, 1994; Reis, Collins, & Berscheid, 2000), extant theories have primarily focused on variables that are internal to the relationship. Furthermore, while there are more than 200 articles addressing the associations of various stressors on marital quality, to my knowledge, no one has yet assembled and integrated this literature so that established associations and unanswered questions can be identified. As such, the goal of this dissertation has been to identify, evaluate, and analyze research on how stressors and supports external to a marriage may be associated with perceptions of satisfaction within the dyad. 123

PAGE 135

124 Towards this end, I began by describing extant theories of stress and marriage. Then, I reviewed an empirical literature of 225 articles, representing a total of 164,359 married individuals. After presenting strengths and limitations of the publication, sampling, and design characteristics of these studies, I conducted a meta-analysis of the aggregate effect sizes of eleven different domains of stressors as well as analyses of several potential moderating variables of these effect sizes. Now that the results of this meta-analysis have been elaborated, the final section of this dissertation will focus on implications for methods, theory, research, public policy, and interventions. Implications for Sampling and Methods The earlier review of this literature revealed diversity in the sources of publication and variety of sample sizes. Despite these strengths, there are also several limitations in this literature, such as a very constricted population of middle-class, Caucasian participants. In light of the deficits in sampling, measurement and design, there are several opportunities for future improvements. In general, future research would benefit from greater sampling diversity. Obtaining data on low income or low socioeconomic populations will inform us about the impact of stressors within the groups that are both most likely to experience a disproportionate number of stressors as well those without the requisite physical resources to facilitate coping with stressors. More international research efforts or the recruitment of nationally representative samples, which provide a more varied demographic profile, are likely to result in greater sampling diversity, which will facilitate a better understanding of how stress affects individuals in less privileged positions relative to those represented in the majority of existing samples.

PAGE 136

125 Suggested Improvements in Measurement There are at least five ways that measurement methods can be improved. First is a need for further refinement and greater consensus about how to operationalize variables (e.g., Moles, 1979). The pervasive experience of racial discrimination seems to be a major stressor for minority individuals (e.g., Peters & Massey, 1983). However, there is no consensus in the present literature about how to measure racial discrimination. In the current meta-analysis, I explored a variable that I hoped would tap this construct, the prevalence of Caucasian or non-Caucasian participants represented in various studies. In these analyses, samples that were less than 60% Caucasian had more negative associations between work stressors and war/military stress and marital satisfaction. However, greater refinement is needed in conceptualizing the stress of insidious and persistent disadvantage experienced by some populations relative to others. Third, conventions for measuring and reporting data will help future researchers to replicate and extend current findings. In this literature, there is a widespread tendency for researchers to neglect to report income and racial composition of their samples. A lack of conventions regarding samples and data reporting makes it difficult to compare across studies (e.g., Vemer, Coleman, Ganong, & Cooper, 1989). Furthermore, as meta-analytic approaches become more common, it becomes increasingly important for future researchers to include basic data (e.g., means, standard deviations, zero-order correlation matrices, sample sizes for all groups) needed to compute effect sizes (Stith, Rosen, Middleton, Busch, Lundeberg, & Carlton, 2000). Innovative ways to measure stress and marital quality will also facilitate future progress. Technological advancements have created opportunities to measure variables in more spontaneous and naturalistic ways. For example, Bolger and Eckenrode (1991) have

PAGE 137

126 used daily diary methods to measure fluctuations in stress levels surrounding an MCAT medical exam. Marco and Suls (1993) provided participants with signal watches and diaries. The signal watches were set to beep eight times a day for eight consecutive days, cuing participants to make entries in their stress diaries. These designs are likely to capture daily variability in stress levels and moods. Moreover, because participants are studied during the normal process of their lives, such designs may have greater ecological validity relative to studies that are conducted in lab settings. Finally, future researchers are urged to avoid relying solely on self-report measures (e.g., Gottman, 1998). I have previously described several limitations of exclusive reliance on self-report measures. When researchers avoid the mono-method bias, it is possible to assess convergence between sources of data and patterns of findings may emerge. A recent study conducted by McNulty and Karney (2002) exemplifies the richness and convergent validity of data obtained with self-report and observation. The study design featured a videotaped discussion in which newlyweds were asked to work towards resolving a problem in their relationship. Each spouse recorded their pre-interaction expectations as well as their post-interaction appraisals about the problem solving discussion. Pairing this self report and observational data revealed that spouses’ prior expectations directly affected post-interaction appraisals and completely mediated an indirect effect on marital satisfaction while behavior did not mediate the association between expectations and appraisals, but instead influenced appraisals independently. As illustrated in this example, then, avoiding the mono-method bias may facilitate more sophisticated analyses of the interplay between cognitive, emotional, and behavioral variables.

PAGE 138

127 Suggested Improvements in Design In addition to suggestions for improving measurement, I also offer five suggestions for improving future designs. First, as several others have noted (e.g., Belsky & Pensky, 1988; Bowerman, 1964; Glenn, 1990; Hicks & Platt, 1970; Karney & Bradbury, 1995; Levinger, 1980; Raush, Barry, Hertel, & Swain, 1974), longitudinal designs have a number of positive features. Longitudinal designs afford opportunities to establish causal connections and study changes in marital satisfaction over time. Studying couples over time may also allow researchers to identify different trajectories of marital satisfaction. For example, Karney and Bradbury (1995) theorize that “marital quality within couples may follow a number of different courses, including initial rapid declines followed by stabilization, gradual linear decline, wide variability over time, or progressive increases in time” (p. 27). In the study of how stress affects marriages, longitudinal data would be useful for several reasons. First, examining couples over time will provide a better perspective on any patterns of reactions to stress within couples. More than half a century ago, Hill (1949) postulated a typical “roller coaster profile of adjustment to crises” (p. 14). As shown below, the crisis is followed by a period of disorganization, then recovery, and reorganization Crisis Figure 7-1. Profile Model of Adjustment to Crisis (Hill, 1949) Angle of recovery Period of Disorganization Level of Reorganization

PAGE 139

128 Longitudinal research would be needed to validate this type of model or any other models that purport to explain the stress reaction over time. Further, longitudinal research would facilitate the testing of models that can explain how each partner in a marriage may have a different pattern of reaction to stress. For example, Burr and colleagues (1994) have described three possible couple-level patterns on two dimensions, couples’ level of functioning and the energy they expend to cope with a stressor (p. 88). The parallel pattern is where energy levels increase and decrease in tandem with functioning. When peaks in energy expenditure match valleys in family functioning, this is referred to as a reciprocal pattern. Finally, the third possibility is that energy expenditure is independent of functioning. Longitudinal research on stress also presents an opportunity to conduct research at different levels of Bronfenbrenner’s levels of context. For example, Olson, Lavee, & McCubbin (1988) have looked at types of families and family response to stress across the life cycle. This research is further enriched by “chrono-level” findings that suggest that marriage is “becoming less valued as a source of economic stability” (e.g., Teachman, Tedrow, & Crowder, 2000, p. 1234). Hypothetically, future researchers who collect longitudinal data at the contextual levels in this example would be able to gain perspective on specific themes and patterns within individual couple interactions, patterns in stress reactions over the lifetime of a given marriage, and possible interactions with changing values in society at large. In addition to looking at stress reactions over time, future researchers are urged to collect multiple waves of data. Although two wave designs are most commonly represented in existing longitudinal studies in this literature, they do not support complex hypotheses about how relationships change and development over time (Karney,

PAGE 140

129 Bradbury, & Johnson, 1999). The benefit of collecting multiple waves of data can be seen in a couple examples. Based on twenty assessments of married couples over a six month period, DeLongis, Folkman, & Lazarus (1988) found that the negative impact of stress on mood was limited to a single day, with the following day characterized by mood scores that were better than usual. Consistent evidence of these lag effects of stress are not likely to have surfaced in a basic two wave design. Multiple waves of data would also be required to test Bohannon’s suggestion that “discrepancies in the timing or intensity of grief may be a “major source of secondary stress for the couple” (Bohannon, 1990, In Oliver, 1999, p. 215). Schwab interviewed parents who had lost a child and found that fathers frequently expressed concern and frustration about their wives’ grief while wives expressed anger over husbands’ not sharing their grief (1992). Thus, it may that one partner’s grieving may be out of synch with the other partner’s grieving and that gender plays a role in grieving patterns. In this example, grieving cycles may bring waves of stress, and waves of assessment would be required to identify any patterns. Furthermore, there may be daily variation for many types of stress. For instance, Repetti (1994) has identified temporal associations between conflictive parental interactions and higher levels of work stress earlier in the same day. Considering these types of provocative questions, and the fact that stress levels may vary greatly even within a given day, multi-wave studies may be uniquely suited for the study of stress. Third, future researchers are urged to collect data over much longer periods of time. The average study length of 2-3 years for studies in this literature is not very long in the course of marriage, and studies of this length may fail to assess important marital milestones such as pregnancy and childbirth. Furthermore, for any given couple, there

PAGE 141

130 will be periods of greater stress and periods of lesser stress throughout the course of marriage. As Wheaton has stated, in order to understand the role of stress properly, “it is necessary to see that the periods of greater stress could have an impact only if there is a contrast to periods of lesser stress” (1999, p. 193). Collecting data over a significant amount of time may clarify patterns in how spouses interact as they are faced with stressors and major life transitions. Fourth, designing studies to assess a number of naturally occurring stressors will provide greater ecological validity and deepen our current understanding of how normative or relatively rare stressors affect marital satisfaction. A sizable portion of studies in this review were conducted under the auspices of universities. Other researchers recruited samples from large organizations such as local businesses and hospitals. In these types of settings, where populations are likely to be transient, researchers could study some fairly common normative stressors like moving to a new city, transitioning to a new job or entering a graduate school program. Enterprising researchers will find abundant opportunities to sample relatively rare stressors. For instance, Bolger and Eckenrode (1991) measured fluctuations in stress levels surrounding an MCAT medical exam in a sample they recruited at an MCAT registration session. When longitudinal research extends over a significant amount of time, it is likely that unexpected “nuggets of gold” will be “sifted from the pan.” For example, researchers connected to the Florida Study on Newlywed Marriage and Adult Development will be able to explore the impact of sweeping historical events such as September 11 th as well as more localized stressful events such as hurricane Frances of fall 2004. Thus, creativity in

PAGE 142

131 designing studies and accessing populations exposed to stressors will only facilitate progress in the study of external contexts. Finally, future research would greatly benefit from the development of research designs that can explore increasingly sophisticated levels of interaction. In the current dissertation, I had originally expected financial resources and non-minority status to buffer against the stress of spousal health problems and work-family stressors. I had imagined that the socio-economic gap in healthcare between middle and low-income individuals might exacerbate the stress of spouses’ health problems. I had also imagined that financial resources could buffer the impact of work-family stress insofar as couples with higher incomes would be better able to afford nannies, housekeepers, cooks, and daycare services. The counter-intuitive findings I observed might indicate a higher tolerance level for stress, more past experience with stressors, or greater availability of family and community social support for lower income couples. Clarifying the complexity of these phenomena will require increasingly sophisticated designs capable of exploring many multiple layers of potential interaction effects. I also discovered gender differences in the domains of unemployment, work-family stress, and spouse’s health problems that could not be written off as an artifact of disproportionate representation of males and females in these domains. It is more likely that gender differences in these domains may be explained by disruptions in marital roles and/or disproportionate burdens of care-giving responsibilities placed on females, but future studies will need to study intervening variables to gain clarity in this regard. In general, then, methods will be strengthened to the extent that researchers design studies that assess context at various levels of Bronfenbrenner’s contextual model. As I

PAGE 143

132 have stated, a basic premise of this work is that marriages do not exist in a vacuum. Various levels of extra-marital stressors, from more proximal contexts such as living in a high crime neighborhood, to the more distal context of the “starter marriage” culture that espouses a new vow to stay committed “as long as we both shall love” may impinge upon spouses’ ability to maintain high levels marital satisfaction. The assessment of these levels of context, using longitudinal designs with multiple waves of data collected over a significant period of time will allow researchers to move beyond simplistic analyses of main effects. As Karney and Bradbury have noted, “while the demands of methodically rigorous research are great, the payoff in understanding marriage and alleviating marital and family distress justifies the investment” (Karney & Bradbury, 1995, p. 28). Implications for Theory and Research The earlier review of stress theories revealed a pattern of complexity and no consistent direction of associations between elements in current stress theories. In fact, either directly or indirectly, resources and vulnerabilities, coping processes and outcomes may interact with each other to buffer or exacerbate stress, to strengthen or diminish external or internal resources and vulnerabilities, to enhance or impair coping ability, to produce a variety of psychosocial outcomes, and to determine a couple’s level of threat or resilience in the face of future stressors. Thus, as I have stated, future growth may not entail coming up with new elements but in increasing refinement of the way we understand the interactions between elements. As Perry-Jenkins, Repetti, & Crouter (2000) have articulated, “we need to identify key moderating variables that shape the process by which stress is transmitted” (p. 989). Some have responded to calls to develop theories than can investigate the roles of key moderating variables in the stress response. For example, Karney and Bradbury

PAGE 144

133 (1995) have proposed the Vulnerability-Stress-Adaptation model (see Figure 7-2). The VSA model incorporates variables representing enduring vulnerabilities (i.e., backgrounds and traits that spouses bring to the marriage such as a history of clinical depression), stressful events, and adaptive processes (e.g., making benign attributions for a spouse’s occasional irritability). In general, the model suggests that couples with effective adaptive processes who face relatively few stressful events and have fewer enduring vulnerabilities will experience a satisfying and stable marriage, while couples with ineffective adaptive processes beset by a greater number of stressful events who have more enduring vulnerabilities will experience declining marital quality, separation or divorce. The VSA model has at least three notable strengths. First, the framework links broad and specific levels of analysis, facilitating explorations of interactions between various levels of context such as the micro-system level of behavioral exchanges and the meso-system level of the work-family interface. The VSA model also postulates mechanisms through which stress and vulnerability lead to changes in marital satisfaction. Specifically, the model highlights a reciprocal relationship between the ongoing process of adaptation within marriage and judgments of marital quality, such that a spouse’s marital satisfaction should change as a function of his or her accumulated experiences with and reactions to behavioral exchanges in the marriage. Third, the VSA model can account for variations in marital outcomes, both between and within couples. Thus, the VSA model represents a holistic, yet parsimonious stress theory and has begun to receive some empirical validation (e.g., Cohan & Bradbury, 1997). As Berscheid (1999) has stated, “much more research on the effects of environment on relationships is

PAGE 145

134 needed.happily, the prospect for obtaining such research has been enhanced by Karney and Bradbury’s (1995) [VSA] model” (p. 264). Yet, despite the strengths of the VSA, the authors point to several refinements that future models could address. Although the authors suggest that “acute stressors may engage a couple’s adaptive capabilities directly, [while] chronic stressors may affect marriage only in the presence of enduring vulnerabilities” (p. 26), the model does not Figure 7-2. Vulnerability-Stress-Adaptation Model Enduring traits Adaptive processes Change in Marital Initial Satisfaction Marital Dissolution External Stressors distinguish between chronic and acute stress. Further, the model does not specify whether the impact of multiple stressors is additive, interactive, or some combination of the two. The experience of stress may vary depending not only on the nature of stressors (e.g., their intensity, source, unpredictability, rarity, etc.), but also according to the moderating effects of variables like income level and racial composition. In response to the need for elaboration of these elements of the VSA model, future models may attempt to specify when, and for whom, pile-up of chronic and acute stressors exacerbates the impact of stress relative to the possibility that some couples may habituate to chronic stressors. Finally, the model does not directly address between-spouse differences in the experience of stress, vulnerability, or marital quality. The current meta-analysis has shown some significant gender differences for certain types of stressors, such as unemployment and

PAGE 146

135 spousal health problems. Future specification may predict differential impacts of certain types of stressors based on role sharing arrangements within a marriage. Future theories may also be able to shed light on the nature of differential patterns of response to stress for each spouse, possibly stemming from enduring vulnerabilities (e.g., a history of major depression), the degree to which a stress directly affects one partner (e.g., prostate cancer), and the meaning of a stressor for each individual (e.g., an existential crisis prompted by the loss of a child for one partner but not for the other, or the unique triggering potential of certain stressors for spouses exposed to war combat). Research exploring associations between stress and marital satisfaction lags behind theories like the VSA model and other models that suggest interactions between mediators and moderators of stress. For example, several domains of stress in the current meta-analysis included very few studies. Due to low power, it was often impossible to accurately or even tentatively assess whether and how variables like racial composition may moderate associations between stress and marital satisfaction. Furthermore, frequent patterns of mixed results may indicate a need for increasingly sophisticated moderation analyses in order to tease apart the key differences within domains of stress or to identify and test potential moderating variables. Given such deficits, there are several opportunities for future improvements in research. Several suggestions for future researchers are described as follows. First, future researchers are encouraged to assess stress as an important marital variable. As many have argued (e.g., Berscheid, 1998; Huston, 2000; Levinger, 1994, Reis, Collins, & Berscheid, 2000; Vemer, Coleman, Ganong, & Cooper, 1989), understanding a marriage requires understanding the larger network of social and

PAGE 147

136 physical contexts within which it is embedded. Even though the majority of the samples in the meta-analysis appear to have been based on white, middle-class participants, results showed that stress was consistently associated with lower marital satisfaction. As described previously in the section on methodological implications, the development of conventions to measure stressors will benefit future explorations. One useful tool for assessing stressors may be Almeida, Wethington, & Kessler’s daily inventory of stressful events (2002, p. 50). This instrument features five dimensions: 1. A broad content classification (i.e., interpersonal tensions, work stressors, home stressors, network, and miscellaneous stressors) 2. The focus of involvement with a stressor (respondent, other, joint) 3. Threat dimensions (loss, danger, disappointment, frustration, opportunity) 4. Stressor severity (objective and subjective assessments) and 5. Primary appraisal domains (disruptions of daily routines, financial situations, way feel about self, way others feel about you, physical health or safety, health/well-being of someone you care about, plans for the future) Second, results of this meta-analysis converge with other evidence suggesting that some stressors are more stressful than others. For example, Pillow, Zautra and Sandler (1996) looked at three types of stress: death of a spouse, divorce stress, and having an asthmatic child. Results showed that the nature of the mediating relationships linking major life events with psychological distress through minor stressors was found to vary as a function of the stressor under consideration. Future research, then, would benefit from going beyond the main effect of stress by paying greater attention to qualities of stressors that explain differences the magnitude of associations between stress and marital satisfaction. In this literature, contradictory findings about the impact of given stressors abound. For example, some evidence suggests that the death of a child has a

PAGE 148

137 profoundly negative impact on marriage (e.g., Kaplan, Grobstein, & Smith, 1976; Simpson, 1979). However, others (e.g., Lang & Gottlieb, 1991; Najman, Vance, Boyle, Embleton, Foster, & Thearle, 1993) have suggested that death of a child has little negative impact and can even improve the relationship. In one longitudinal study (Bohannon, 1990-1991), a substantial minority of couples experienced marital difficulties after the death of a child. Inconsistent findings suggest that important moderators have not been identified. For instance, in this example, other research suggests that anticipation of a child’s death may temper subsequent grief reactions (e.g., Parkes & Weiss, 1983; Shanfield, Benjamin, & Swain, 1988). Ultimately, to gain a better understanding of the critical qualities of stressors, it may be wise to follow Wheaton’s suggestion to dis-aggregate stress concepts, extricate coping from its dependence on context, and consider a full range of stressors that vary in their typical time course and occur at different stages in the life course and at different levels of social reality (1999, p. 197). In addition to stressor qualities, it seems important to assess other key moderating and mediating variables. For instance, in the current meta-analysis, I found significant gender differences for unemployment, spousal health problems, and work-family stress. I have theorized that these gender differences may be explained by disruptions in marital roles and/or disproportionate burdens of care-giving responsibilities placed on females. However, further research on gender differences is needed to provide convergent or discriminant validity for these hypotheses. Contradictory results were also prevalent for moderation analyses based on income level and racial background. I have suggested that contradictory results might indicate a

PAGE 149

138 higher tolerance level for stress (Gump & Matthews, 1999), more past experience with stressors (e.g., McAdoo, 1983), or greater availability of family and community social support for some types of stress or in some types of samples (e.g., Taylor, 1986). Future research, then, could also explore the question of whether stress levels are higher for those with fewer financial resources who may have habituated to chronic stress, or those with more financial resources for whom stress may be relatively new. Some researchers are already beginning to map this terrain. A review of research focused on ongoing stressors revealed that background stressors were associated with increased acute stressor reactivity in slightly over half of the 19 studies (Gump & Matthews, 1999). Along similar lines, Marco, & Suls (1993) discovered that mood in response to a current problem was worse if the prior time has been problem free than if the prior time had been stressful. Interestingly, though, nearly half of the studies in the aforementioned review of ongoing stressors demonstrated “reduced acute stressor reactivity, suggesting habituation for certain people in certain situations” (Gump & Matthews, 1999, p. 469). Supportive of the habituation hypothesis, Caspi, Bolger, & Eckenrode (1987) found that previous exposure to major life events decreased the impact of stressful life events. So, because results support both possibilities, this question remains open for future research. Although I have more often focused speculations on less privileged groups, there is also the possibility that stress levels may be exacerbated or ameliorated in interesting ways for affluent populations. For instance, Lewis, Barnhart, Nace, Carson, & Howard (1993) sampled a large group of doctors and their spouses and found that about 85% are satisfied with their marriages. In this study, high marital satisfaction was associated with high work satisfaction. So, research on this privileged sample might indicate that factors

PAGE 150

139 like job autonomy and occupational prestige may mitigate the negative impact of high pressure and long hours on the job. Further, facing unemployment may more be more stressful for a relatively privileged individual insofar as the loss of status is felt more keenly. As McAdoo has explained, “unemployment for Black urban youth is such a common experience that it is often seen in the Black community as a ‘normative’ stressor” (McAdoo, 1983, p. 181). Thus, for privileged individuals, certain types of stress may be experienced as both non-normative and more stressful due to perceived social devaluation. Future research may also benefit from explorations of personality factors and stable individual differences that enhance or decrease the negative impact of stress. A number of researchers have also begun to map this terrain. For example, Scheier, Weintraub, & Carver (1986) found global pessimism to be associated with denial, distancing, and disengagement from the goal with which a stressor was interfering. Negative affect and neuroticism have also been linked to greater higher levels of stress. Marco and Suls (1993) discovered that participants that are high in negative affect (NA) were more reactive to concurrent stressors relative to low NA individuals. Gunthert, Cohen, and Armeli’s (1999) two week diary study of college students showed that neurotic people reported more interpersonal stressors, had more negative appraisals of stress, and used less adaptive coping strategies compared to less neurotic people. Convergent validity emerges in a study conducted by Bolger and Zuckerman (1995); participants high in neuroticism had greater exposure and more reactivity to interpersonal conflicts. Relationship beliefs may be another provocative individual difference variable that warrants future exploration. Knee (1998) has described two primary sets of beliefs about

PAGE 151

140 relationships. On the one hand, individuals with a belief in destiny tend to feel that each person has an ideal soul mate and that potential relationship partners are either meant for each other or they are not. On the other hand, individuals with a belief in growth tend to feel that successful relationships are “constructed by conquering obstacles and growing closer” and thus “evolve from the resolution of risks, challenges, and difficulties rather than their absence” (p. 361). It is easy to see how relationship beliefs could have a profound impact on how spouses view and cope with stressors. Individuals who believe in growth may engage in more active forms of coping as well as positive reinterpretation of a stressor event while individuals who believe in destiny may be more prone to make less charitable, and more negative dispositional inferences when their partners are irritable in the face of a stressful event. In fact, this link to attributions may be critical in that other research suggests that partners in satisfying relationships maintain positive illusions about their partners that reduce doubt and uncertainty about the future of the relationship (Murray & Holmes, 1993). As Murray and Holmes (1993) explain, partners’ sense of security “depends on their continued struggle to weave stories that depict potential faults in their partners in the best possible light” (p. 719). In these ways, then, relationship beliefs may moderate attributions that serve to protect relationships from dissatisfaction and dissolution. Future research may also benefit from investigations of the conditions under which various resources may be more or less helpful. For instance, one question that remains open is whether and when social support benefits couples under stress. Addressing this question will require going beyond main effects and simplistic assertions that social support is always good for relationships. On the one hand, individuals with strong family

PAGE 152

141 and friendship networks are less likely to succumb to the damaging impact of stress (For a review, see Cohen & Wills, 1985). According to Taylor (1986), elaborate social networks in African American communities provide a crucial system of mutual aid. However, Belle (1982) has argued that social support networks can expose individuals to high levels of “contagion of stress” and may “force many individuals to engage in survival networks which preclude upward mobility and often exact emotional penalties” (p. 94). Furthermore, Bolger, Zuckerman, & Kessler (2000) have demonstrated that the awareness of receiving support entails an emotional cost and that invisible support promotes the best adjustment to a major stressor. Empirical support for this possibility emerges in a study conducted by Robertson and colleagues (e.g., Robertson, Elder, Skinner, & Conger, 1991). Specifically, when there is economic pressure in the family, and when wives seek external support, wives’ distress is often lessened. At the same time, seeking this support can affirm the husband’s sense of being a failure and is typically associated with husbands’ negativity towards wives. Thus, important gains will be achieved when future research goes beyond unsophisticated claims that social support is always supportive. Finally, when greater clarity regarding important mediators and moderators has been achieved, future researchers will then be able to compare the relative explanatory power of various mechanisms that may reduce marital satisfaction. There are potentially several ways to explain how stress has a negative on relationship functioning. For instance, sexual intimacy may suffer, individual and collective decision-making may be impaired, and relationship self-efficacy may founder. Spouses may make less positive attributions for their partner’s bad behavior when they themselves are taxed by stress.

PAGE 153

142 Further research is needed to explore other specific processes in relationships that are likely to deteriorate in the face of stressful circumstances and to identify those which are most likely to negatively impact marital satisfaction. Implications for Public Policy and Intervention Meta-analytic findings also possess “a certain robustness that make them highly attractive as a basis for policy and practice guidelines” (Lipsey & Wilson, 2001, pp. 167). This review and meta-analysis has shed light on the inherent complexity of a stressor’s impact on marital satisfaction. Moreover, findings suggest that some samples may be indeed experience higher levels of stress which may be associated with greater decreases in marital satisfaction relative to other samples. Therefore, three suggestions for policymakers and four suggestions for practitioners are offered as follows. First, as Park, Turnbull, and Turnbull (2002) have noted, on a systems level, there should be greater recognition that no single professional or agency can meet the multi-faceted needs of poor couples and their families. Parker and colleagues have made specific suggestions such as partnerships between schools, families, and communities to allow for coordinated delivery of services. Second, it is important to actively create policies that alleviate the disproportionate presence of stress in some peoples’ lives and marriages. Such policies may represent “upriver prevention” of marital dissolution. Reducing divorce rates may require less marital education and more attempts to actively reduce poverty (Karney & Springer, 2004, p. 249). For instance, one specific suggestion is allowing home and community based Medicaid waiver funds to pay families to support their members who have disabilities (Park, Turnbull, & Turnbull, 2002).

PAGE 154

143 Finally, when policies have been created, it is critical that policymakers build in accountability mechanisms in order to assess whether changes are having their intended effect. At this point, there have not been enough studies of the cost-benefit ratio or effectiveness of current [welfare] policies and practices (Park, Turnbull, & Turnbull, 2002). Interviews of individuals on welfare illustrate the sizable gap between the intention of policies to be helpful and their actual effects. Cherlin (2004) describes the case of one participant: Evelyn Santero, 6 a full-time working mother, was required to fill out a monthly form with information such as her income and work hours for the previous four weeks and who was living in her household. After she failed to submit the first form on time, the welfare office sent her a notification that she was in danger of losing her benefits. With the help of her daughters, she filled out the form and left it at the welfare office. The cycle of tardiness, warning, and just-in-time submission of the form continued for a few months. Finally, after apparently missing the filing deadline completely, she received a notice that her TANF and food stamp benefits had been suspended for “refusal to return the completed monthly report.” (p. 269) In light of cases like this, outcome data, assessed with longitudinal designs, would provide the best test of whether public aid policies are actually helpful to struggling couples. Finally, I will conclude this dissertation by making some suggestions for interventions. First, this review and the subsequent meta-analytic findings highlight the importance of assessing both stressors and coping styles as a regular part of interventions. Pre and post-marital assessments of couple functioning would benefit from the addition of inventories that measure the presence of stressors and the habitual coping styles of each spouse. For example, the daily inventory of stressful events (Almeida, Wethington, 6 Pseudonym

PAGE 155

144 & Kessler, 2002) includes a broad content classification (e.g., interpersonal tensions, work stressors etc), the focus of stressor involvement (e.g., respondent, other, joint), the dimensions a threat (e.g., loss, danger, frustration etc.), stressor severity (both objective and subjective assessments), and primary appraisal domains (e.g., disruption of daily routines, financial situations physical health or safety etc.). The multidimensional coping inventory (MCI; Carver, Scheier, & Weintraub, 1989) is an assessment of coping strategies which elicits data on different ways people respond to stress. The MCI divides coping into five adaptive coping strategies, which are seeking emotional social support, positive reinterpretation, acceptance, denial, turning to religion, and three maladaptive coping strategies, which are focusing on and venting emotions, behavioral disengagement, and mental disengagement. Second, in addition to assessing for the presence of stressors and styles of coping, practitioners are urged to facilitate a greater awareness of the potential impact of stress on the part of married partners. Janis (1918 as cited in Janis, 1993) has suggested that stress may impair functioning in a number of ways by causing maladaptive hyper vigilance, defensive avoidance and constricted decision making abilities. If couples are made aware of the presence of stress and the possible ways that stress might affect them, they may be able to ward off destructive, misguided attributions. Armed with a greater awareness of stressors, couples may be able to engage in novel coping strategies. As a whimsical example, the last month has been a “paper plate only” month as I pressed towards the goal of submitting this dissertation and my husband simultaneously submitted and defended his master’s thesis. As in this example, agreeing on even simple coping

PAGE 156

145 strategies, such as one that lightens the usual amount of chores, may strengthen the marital bond as couples feel they are working as a team. Third, practitioners are urged to help couples develop skills for coping with stressors. Part of coping may involve teaching couples how to maintain positive communication and engage in effective problem solving. There are a number of programs that have been developed to teach these skills and they play a role in supporting healthy relationships. Longitudinal outcome research may help consumers differentiate between programs that are more or less effective. For example, Bodenmann and colleagues’ found long-term positive effects in a one-year follow up study of a couples coping enhancement training program (Bodenmann, Charvoz, Cina, & Widmer, 2001). Practitioners should also be aware that even the most effective coping training programs may not make a dent in the face of pervasive, chronic stress. Recall one individual’s statement: If you have no money, it’s difficult tobe together, to do fun things, to be at peace, to come home to a havenbecause if you have no money, bills are not paid, you not gonna get rest when you get home [sic]. (Park, Turnbull, & Turnbull, 2002, p. 151) In general, though, for couples that are not experiencing the most extreme or chronic forms of stress, applying empirically supported training programs to help couples effectively cope with stressors may be of some benefit. Concluding Comments Ultimately, while this dissertation has helped illuminate some of the external terrain of marital satisfaction, this research is only a small step in an effort to achieve a more holistic understanding of how stressors affect marital satisfaction. As a concluding thought, Gottman (1993) has been cited for the discovery that stable, happy relationships are characterized by a 5 to 1 ratio of positive to negative affect displays. Perhaps a similar

PAGE 157

146 ratio applies in the prediction of marital satisfaction based on the proportion of negative stressors to positive pleasures that various couples face?

PAGE 158

LIST OF REFERENCES Alexander, C. J., Hwang, K., & Sipski, M. L. (2002). Mothers with spinal cord injuries: impact on marital, family, and children's adjustment. Archives of Physical Medicine Rehabilitation, 83, 24-30. Almeida, D. M., Wethington, E., & Kessler, R. C. (2002). The Daily Inventory of Stressful Events: An interview-based approach for measuring daily stressors. Assessment, 9(1), 41-55. Amato, P. R., & Keith, B. (1991). Parental divorce and adult well-being: A meta-analysis. Journal of Marriage and the Family, 53, 43-58. Anderson, E. A., & Leslie, L. A. Coping with stress: Differences among working families. Stressful Life Events, 697-710. Andrews, F. M., Abbey, A., & Halman, L. J. (1991). Stress from infertility, marriage factors, and subjective well-being of wives and husbands. Journal of Health and Social Behavior, 32(3), 238-253. Angell, R. C. (1934). The influence of severe and apparently lasting decrease in income upon family life. Publications of the American Sociological Society, 28, 85-89. Angrist, J. D., & Johnson, J. H. (2000). Effects of work-related absences on families: evidence from the Gulf War. Industrial and Labor Relations Review, 54, 41-58. Aube, N., & Linden, W. (1991). Marital disturbance, quality of communication and socioeconomic status. Journal of Behavioural Science, 23(2), 125-132. Aubry, T., Teft, B., & Kingsbury, N. (1990). Behavioral and psychological consequences of unemployment in blue collar couples. Journal of Community Psychology, 18, 99-109. Bakke, E. W. (1940). Citizens without work: A study of the effects of unemployment upon the worker’s social relations and practice. New Haven, CT: Yale University Press. Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist, 44, 1175-1184. Bandura, A. (1997). Self-efficacy: The exercise of control. New York: W. H. Freeman and Company. 147

PAGE 159

148 Barbarin, O. A., & Chesler, M. A. (1984). Coping as interpersonal strategy: Families with childhood cancer. Family Systems Medicine, 2, 279-289. Barbarin, O. A., Hughes, D., & Chesler, M. A. (1985). Stress, coping, and marital functioning among parents of children with cancer. Journal of Marriage & Family, 473-480. Barling, J. (1984). Effects of husbands' work experiences on wives' marital satisfaction. Journal of Social Psychology, 124(2), 219-225. Barling, J. (1986). Interrole conflict and marital functioning amongst employed fathers. Journal of Occupational Behaviour, 7(1), 1-8. Barling, J., & MacEwen, K. E. (1992). Linking work experiences to facets of marital functioning. Journal of Organizational Behavior, 13(6), 573-583. Barnes, M. D., Pase, M., & VanLeeuwen, D. (1999). The relationship of economic factors and stress among employed, married women with children. American Journal of Health Promotion,13(4), 203-206. Barnett, K. A., Del Campo, R. L., & Del Campo, D. S. (2003). Work and family balance among dual earner working-class Mexican-Americans: Implications for therapists. Contemporary Family Therapy: An International Journal, Vol 25(4), 353-366. Barrera, M. (1986). Distinctions between social support concepts, measures, and models. American Journal of Community Psychology, 14, 413-445. Basolo-Kunzer, M., Diamond, S., Maliszewski, M., & Weyermann, L. (1991). Chronic headache patients' marital and family adjustment. Issues in Mental Health Nursing, 12(2), 133-148. Beach, S. R. (2001). Marital and family processes in depression: A scientific foundation for clinical practice. Washington, D.C.: American Psychological Association. Bedeian, A. G., Burke, B. G., & Moffett, R. G. (1988). Outcomes of work-family conflict among married male and female professionals. Journal of Management,14(3), 475-491. Belle, D. E. (1982). The impact of poverty on social networks and supports. Marriage & Family Review, 5, 89-103. Belsky, J., & Pensky, E. (1988). Marital change across the transition to parenthood. Marriage and Family Review, 12, 133-156. Benazon, N., Wright, J., & Sabourin, S. (1992). Stress, sexual satisfaction, and marital adjustment in infertile couples. Journal of Sex and Marital Therapy, 18(4), 273-284.

PAGE 160

149 Bentler, P. M., & Newcomb, M. D. (1978). Longitudinal study of marital success and failure. Journal of Consulting & Clinical Psychology, 46, 1053-1070. Berant, E., Mikulincer, M., & Florian, V. (2003). Marital satisfaction among mothers of infants with congenital heart disease: The contribution of illness severity, attachment style, and the coping process. Anxiety, Stress, and Coping: An International Journal, 16 (4), 397-415. Berkowitz, A. D., & Perkins, H. W. (1985). Correlates of psychosomatic stress symptoms among farm women: A research note on farm and family functioning. Journal of Human Stress, 11(2), 76-81. Bermas, B. L., Tucker, J. S., Winkelman, D. K., & Katz, J. N. (2000). Marital satisfaction in couples with rheumatoid arthritis. Arthritis Care and Research, 13(3), 149-155. Berry, R. E., & Williams, F. L. (1987). Assessing the relationship between quality of life and marital and income satisfaction: A path analytic approach. Journal of Marriage and the Family, 49(1), 107-116. Berscheid, E. (1998). A social psychological view of marital dysfunction and stability. In T. N. Bradbury (Ed.), The developmental course of marital dysfunction (pp. 441-459). New York: Cambridge University Press. Berscheid, E. (1999). The greening of relationship science. American Psychologist, 54, 260-266. Berscheid, E., Snyder, M., & Omoto, A. M. (1989). Issues in studying close relationships: Conceptualizing and measuring closeness. In C. Hendrick (Ed.), Close relationships (pp. 63-91). Thousand Oaks, CA: Sage Press. Blacher, J., Nihira, K., & Meyers, C. E. (1987). Characteristics of home environment of families with mentally retarded children: Comparison across levels of retardation. American Journal of Mental Deficiency, 91(4), 313-320. Blair, S. L. (1993). Employment, family, and perceptions of marital quality among husbands and wives. Journal of Family Issues, 14, 189-212. Block, A. R., & Boyer, S. L. The spouse's adjustment to chronic pain: Cognitive and emotional factors. Social Science & Medicine, 19(12), 1313-1317. Bloom, B. L., Asher, S. J., & White, S. W. (1978). Marital disruption as a stressor: A review and analysis. Psychological Bulletin, 85, 867-894. Bloom, J. R., & Spiegel, D. (1984). The relationship of two dimensions of social support to the psychological well-being and social functioning of women with advanced breast cancer. Social Science & Medicine, 19, 831-837.

PAGE 161

150 Bodenmann, G. (1997a). Can divorce be prevented by enhancing the coping skills of couples? Journal of Divorce & Remarriage, 27, 177-194. Bodenmann, G. (1997b). Dyadic coping: A systemic-transactional view of stress and coping among couples: Theory and empirical findings. European Review of Applied Psychology, 47, 137-141. Bodenmann, G., Charvoz, L., Cina, A., & Widmer, K. (2001). Prevention of marital distress by enhancing the coping skills of couples: 1-year follow-up-study. Swiss Journal of Psychology, 60, 3-10. Bohannon, J. R. (1990). Grief responses of spouses following the death of a child: A longitudinal study. Omega: Journal of Death & Dying, 22(2), 109-121. Bokemeier, J., & Maurer, R. (1987). Marital quality and conjugal labor involvement of rural couples. Family Relations, 36(4), 417-424. Bolger, N., DeLongis, A., Kessler, R. C., & Schilling, E. A. (1989). Effects of daily stress on negative mood. Journal of Personality and Social Psychology, 57, 808-818. Bolger, N., DeLongis, A., Kessler, R. C., & Wethington, E. (1990). The microstructure of daily role-related stress in married couples. In J. Eckenrode & S. Gore (Eds.), Stress between work and family (pp. 95-115). New York: Plenum Press. Bolger, N., & Eckenrode, J. (1991). Social relationships, personality, and anxiety during a major stressful event. Journal of Personality and Social Psychology, 61(3), 440-449. Bolger, N., Foster, M., Vinokur, A. D., & Ng, R. (1996). Close relationships and adjustment to life crisis: The case of breast cancer. Journal of Personality and Social Psychology, 70, 283-294. Bolger, N., & Zuckerman, A. (1995). A framework for studying personality in the stress process. Journal of Personality and Social Psychology, 69, 890-902. Bolger, N., Zuckerman, A., & Kessler, R. C. (2000). Invisible support and adjustment to stress. Journal of Personality and Social Psychology, 79, 953-961. Booth, A., & Amato, P. (1991). Divorce and psychological stress. Journal of Health & Social Behavior, 32, 396-407. Booth, A., & Amato, P. (2001). Parental predivorce relations and offspring postdivorce well-being. Journal of Marriage and the Family, 63,197-212. Booth, A., & Johnson, D. R. (1994). Declining health and marital quality. Journal of Marriage and the Family, 56(1), 218-223.

PAGE 162

151 Booth, A., Johnson, D. R., White, L. K., & Edwards, J. N. (1986). Divorce and marital instability over the life course. Journal of Family Issues, 7, 421-442. Boss, P. (1983). The marital relationship: Boundaries and ambiguities. In C. R. Figley & H. I. McCubbin (Eds.), Stress and the family: Coping with normative transitions, Vol.1 (pp. 26-40). New York: Brunner/Mazel. Boss, P. (1987). Family stress. In M.B. Sussman & S. K. Steinmetz (Eds.), Handbook of marriage and the family (pp. 695-723). New York: Plenum Press. Bowen, G. L. (1987). Wives' employment status and marital adjustment in military families. Psychological Reports, 61(2), 467-474. Bowerman, C. E. (1964). Prediction studies. In H. T. Christensen (Ed.), Handbook of marriage and the family (pp. 215-246). Chicago: Rand McNally. Bowlby, J. (1969). Attachment and loss: Volume I. New York: Basic Books. Brackney, B. E. (1979). The impact of home hemodialysis on the marital dyad. Journal of Marital and Family Therapy, 5(1), 55-60. Bradbury, T. N., & Karney, B. R. (1993). Longitudinal study of marital interaction and dysfunction: Review and analysis. Clinical Psychology Review, 13, 15-27. Bramlett, M. D., & Mosher, W. D. (2001). First marriage dissolution, divorce, and remarriage: United States (Advance data from vital and health statistics No. 323). Hyattsville, Maryland: National Center for Health Statistics. Brinkerhoff, D. B., & White, L. K. (1978). Marital satisfaction in an economically marginal population. Journal of Marriage and the Family, 40(2), 259-267. Broman, C. L. (1988a). Satisfaction among Blacks: The significance of marriage and parenthood. Journal of Marriage & the Family, 50(1), 45-51. Broman, C. L. (1988b). Household work and family life satisfaction in Blacks. Journal of Marriage and the Family, 50, 743-748. Broman, C. L. (1993). Race differences in marital well-being. Journal of Marriage and the Family, 55, 724-732. Broman, C. L., Hamilton, V. L., & Hoffman, W. S. (1990). Unemployment and its effects on families: Evidence from a plant closing study. American Journal of Community Psychology, 18(5), 643-659. Broman, C. L., Riba, M. L., & Trahan, M. R. (1996). Traumatic events and marital well-being. Journal of Marriage & the Family, 58, 908-916.

PAGE 163

152 Bronfenbrenner, U. (1986). Ecology of the family as a context for human development: Research perspectives. Developmental Psychology, 22, 723-742. Brown, B. B. (1978). Social and psychological correlates of help-seeking behavior among urban adults. American Journal of Community Psychology, 6, 425-439. Brown, G. W., & Harris, T. O. (1978). Social origins of depression: A study of psychiatric disorder in women. New York: Free Press. Brown, R. T., Kaslow, N. J., Hazzard, A. P., & Madan-Swain, A. (1992). Psychiatric and family functioning in children with leukemia and their parents. Journal of the American Academy of Child & Adolescent Psychiatry, 31(3), 495-502. Brown, S. D., & Krane, N. E. (2000). Four (or five) sessions and a cloud of dust: Old assumptions and new observations about career counseling. In S. D. Brown & R. W. Lent (Eds.), Handbook of counseling psychology 3 rd ed. (pp. 740-766). New York, NY: John Wiley and Sons. Burke, R. J. (1982). Occupational demands on administrators and spouses' satisfaction and well-being. Psychological Reports, 51(3, Pt 1), 823-836. Burke, R. J., & Weir, T. (1976a). Marital helping relationships: The moderators between stress and well-being. Journal of Marriage and the Family, 95(1), 121-130. Burke, R. J., & Weir, T. (1976b). Relationship of wives' employment status to husband, wife and pair satisfaction and performance. Journal of Marriage & the Family, 38, 279-287. Burke, R. J., & Weir, T. (1977). Husband-wife helping-relationships: The "mental hygiene" function in marriage. Psychological Reports, 40(3, Pt 1), 911-925. Burr, W. R. (1982). Families under stress. In H. I. McCubbin, A. E. Cauble, & J. M. Patterson (Eds.), Family stress, coping, and social support (pp. 5-25). Springfield, Ill: Thomas. Burr, W. R., Klein, S. R., Burr, R. G., Doxey, C., Harker, B., Holman, T. B., Martin, P. H., McClure, R. L., & Weiler, S. (1994). Reexamining family stress: New theory and research. Thousand Oaks, CA: Sage Press. Caplan, G. (1964). Principles of preventive psychiatry. New York: Basic Books. Cano, A., Gillis, M., & Heinz, W. (2004). Marital functioning, chronic pain, and psychological distress. Pain, 107(1-2), 99-106. Cano, A., & Vivian, D. (2003). Are life stressors associated with marital violence? Journal of Family Psychology, 17(3), 302-314.

PAGE 164

153 Carlson, C. I. (1995). Families as the focus of assessment: theoretical and practical issues. In J.J. Conoley & E.B. Werth (Eds), Family measurement (pp. 19-63). Lincoln, NB: Buros Institute of Mental Measurements. Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). Assessing coping strategies: A theoretically based approach. Journal of Personality & Social Psychology, 56, 267-283. Caspi, A., Bem, D. J., & Elder, G. H. (1989). Continuities and consequences of interactional styles across the life course. Journal of Personality, 57, 376-406. Caspi, A., Bolger, N., & Eckenrode, J. (1987). Linking person and context in the daily stress process. Journal of Personality & Social Psychology, 52, 184-195. Cassidy, M. L. (1985). Role conflict in the postparental period: The effects of employment status on the marital satisfaction of women. Research on Aging, 7(3), 433-454. Chadiha, L. A., Rafferty, J., & Pickard, J. (2003). The influence of caregiving stressors, social support, and caregiving appraisal on marital functioning among African American wife caregivers. Journal of Marital and Family Therapy, 29(4), 479-490. Chan, R. C. K., Lee, P. W. H., & Lieh-Mak, F. (2000). The pattern of coping in persons with spinal cord injuries. Disability & Rehabilitation: An International Multidisciplinary Journal, 22(11), 501-507. Chandler, B. J., & Brown, S. (1998). Sex and relationship dysfunction in neurological disability. Journal of Neurology, Neurosurgery, and Psychiatry, 65(6), 877-880. Chatters, L. M., Taylor, R. J., Jackson, J. S. (1986). Aged Blacks’ choices for an informal helper network. Journal of Gerontology, 41, 94-100. Cherlin, A. J. (1979). Work life and marital dissolution. In G. Levinger & O. C. Moles (Eds.), Divorce and separation: Context, causes, and consequences (pp. 151-166). New York: Basic Books. Cherlin, A. J. (1992). Marriage, divorce, and remarriage (2nd ed.). Cambridge, MA: Harvard University Press. Cherlin, A. J. (2004). The growing compliance burden for recipients of public assistance. In A. C. Crouter & A. Booth (Eds.), Work-family challenges for low-income parents and their children. (pp. 265-272). Mahwah, NJ: Lawrence Erlbaum Associates. Chesler, M. A., & Barbarin, O. A. (1987). Childhood cancer and the family: Meeting the challenge of stress and support. Philadelphia, PA: Brunner/Mazel, Inc.

PAGE 165

154 Christensen, A., & Nies, D. C. (1980). The Spouse Observation Checklist: Empirical analysis and critique. American Journal of Family Therapy, 8, 69-79. Clark-Nicolas, P., & Gray-Little, B. (1991). Effect of economic resources on marital quality in Black married couples. Journal of Marriage and the Family, 53(3), 645-655. Cohan, C. L., & Bradbury, T. N. (1997). Negative life events, marital interaction, and the longitudinal course of newlywed marriage. Journal of Personality & Social Psychology, 73, 114-128. Cohen, S., & Edwards, J. R. (1989). Personality characteristics as moderators of the relationship between stress and disorder. In R. J. W. Neufeld (Ed.) Advances in the investigation of psychological stress (pp. 235-283). New York: Wiley. Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin, 98, 310-357. Comprehensive Meta-analysis. (beta version). Biostat Incorporated (2004). Conger, R. D., Conger, K. J., Elder, G. H., & Lorenz, F. O. (1992). A family process model of economic hardship and adjustment of early adolescent boys. Child Development, 63, 526-541. Conger, R. D., Elder, G. H., Lorenz, F. O., Conger, K. J., Simons, R. L., Whitbeck, L. B., Huck, S., & Melby, J. N. (1990). Linking economic hardship to marital quality and instability. Journal of Marriage & the Family, 52, 643-656. Conger, R. D., Ge, X.-J., & Lorenz, F. O. (1994). Economic stress and marital relations. In R. D. Conger & G. H. Elder (Eds.), Families in troubled times: Adapting to change in rural America (pp. 187-203). New York: A. de Gruyter. Conger, R. D., Lorenz, F. O., Elder, G. H., & Simons, R. L. (1993). Husband and wife differences in response to undesirable life events. Journal of Health & Social Behavior, 34(1), 71-88. Conger, R. D., Reuter, M. A., & Elder, G. H. (1999). Couple resilience to economic pressure. Journal of Personality & Social Psychology, 76, 54-71. Cook, J. A., Hoffschmidt, S., Cohler, B. J., & Pickett, S. (1992). Marital satisfaction among parents of the severely mentally ill living in the community. American Journal of Orthpsychiatry,62(4), 552-563. Cook, J. M., Riggs, D. S., Thompson, R., Coyne, J. C., & Sheikh, J. I. (2004). Posttraumatic stress disorder and current relationship functioning among World War II ex-prisoners of war. Journal of Family Psychology, 18, 36-45.

PAGE 166

155 Coombs, R. H., & Hovanessian, H. C. (1988). Stress in the role constellation of female resident physicians. Journal of American Medical Women's Association, 43, 21-27. Cooper, H., & Hedges, L. V. (1994). Research synthesis as a scientific enterprise. In H. Cooper & L.V. Hedges (Eds.), The Handbook of research synthesis (pp. 3-14). New York: Russell Sage Foundation. Coverman, S. (1989). Role overload, role conflict, and stress: Addressing consequences of multiple role demands. Social Forces, 67(4), 965-982. Coyne, J. C., & Smith, D. A. (1991). Couples coping with a myocardial infarction: A contextual perspective on wives' distress. Journal of Personality and Social Psychology, 61(3), 404-412. Coyne, J. C., Wortman, C. B., & Lehman, D. R. (1988). The other side of support: Emotional over-involvement and miscarried helping. In B. H. Gottlieb (Ed.), Marshaling social support: Formats, processes, and effects (pp. 305-330). Newbury Park: Sage Publications. Croake, J. W., & Lyon, R. S. Factors related to tours of duty and marital adjustment. International Journal of Sociology of the Family, 259-261. Croog, S. H., & Fitzgerald, E. F. (1978). Subjective stress and serious illness of a spouse: Wives of heart patients. Journal of Health & Social Behavior, 19(2), 166-178. Crouter, A. C., Bumpas, M. F., Head, M. R., & McHale, S. M. (2001). Implications of overwork and overload for the quality of men's family relationships. Journal of Marriage and the Family, 63(2), 404-416. Cutrona, C. E., Russell, D. W., Todd, A. W., Gardner, K. A., Melby, J. M., Bryant, C., & Conger, R. D. (2003). Neighborhood context and financial strain as predictors of marital interaction and marital quality in African American couples. Personal Relationships, 10, 389-409. Dahlquist, L. M., Czyzewski, D. I., & Jones, C. L. (1996). Parents of children with cancer: A longitudinal study of emotional distress, coping style, and marital adjustment two and twenty months after diagnosis. Journal of Pediatric Psychology,21(4), 541-554. Daneker, B., Kimmel, P. L., Ranich, T., & Peterson, R. A. (2001). Depression and marital dissatisfaction in patients with end-stage renal disease and their spouses. American Journal of Kidney Disease, 38(4), 839-846. Dehle, C., Larsen, D., & Landers, J. E. (2001). Social support in marriage. American Journal of Family Therapy, 29, 307-324.

PAGE 167

156 DeLongis, A., Folkman, S., & Lazarus, R. S. (1988). The impact of daily stress on health and mood: Psychological and social resources as mediators. Journal of Personality and Social Psychology, 54, 486-496. Dillaway, H., & Broman, C. L. (2001). Race, class, and gender differences in marital satisfaction and divisions of household labor among dual-earner couples. Journal of Family Issues, 22, 309-327. Dilworth, J. E. (2004). Predictors of negative spillover from family to work. Journal of Family Issues, 25, 241-261. Doherty, W. J., Su, S., & Needle, R. (1989). Marital disruption and psychological well-being. Journal of Family Issues, 10, 72-85. Dorval, M., Maunsell, E., Taylor-Brown, J., & Kilpatrick, M. (1999). Marital stability after breast cancer. Journal of the National Cancer Institute, 91(1), 54-59. Doumas, D. M., Margolin, G., & John, R. S. (2003). The relationship between daily marital interaction, work, and health-promoting behaviors in dual-earner couples: an extension of the work-family spillover model. Journal of Family Issues, 24(1), 3-20. Duncan, S. F., & Goddard, H. W. (1993). Stressor and enhancers in the marital/family life of family professionals and their spouses. Family Relations, 42(4), 434-441. Dunkel-Schetter, C., & Bennett, T. L. (1990). Differentiating the cognitive and behavioral aspects of social support. In B. R. Sarason, I. G. Sarason, & G. R. Pierce (Eds.), Social support: An interactional view (pp. 267-296). New York: J. Wiley & Sons. Dupont, S. (1996). Sexual function and ways of coping in patients with multiple sclerosis and their partners. Sexual & Marital Therapy, 11(4), 359-372. Durlak, J. A. (1995). Understanding meta-analysis. In L.G. Grimm and P.R. Yarnold (Eds.), Reading and understanding multivariate statistics (pp.319-352). Washington DC: American Psychological Association Press. Dyer, C., & Halford, W. K. (1998). Prevention of relationship problems: Retrospect and prospect. Behaviour Change, 15(2), 107-125. Eckenrode, J., & Wethington, E. (1990). The process and outcome of mobilizing social support. In S. W. Duck & R. C. Silver (Eds.), Personal Relationships (pp. 83-103). Thousand Oaks, CA: Sage. Edin, K. (2000). What do low-income single mothers say about marriage? Social Problems, 47, 112-133.

PAGE 168

157 Edwards, J. R., & Cooper, C. L. (1987). The impacts of positive psychological states on physical health: A review and theoretical framework. Social Science & Medicine, 27, 1447-1459. Elder, G. H., Conger, R. D., Foster, E. M., & Ardelt, M. (1992). Families under economic pressure. Journal of Family Issues, 13, 5-37. Eldridge, K. A., Lawrence, E., & Christensen, A. (1999). Focus chapter: Research methods with couples. In P. C. Kendall, J. N. Butcher, & G. N. Holmbeck (Eds). Handbook of research methods in clinical psychology (2nd ed.) (pp. 681-699). New York: J. Wiley & Sons. Ensel, W. M., & Lin, N. (1991). The life stress paradigm and psychological distress. Journal of Health and Social Behavior, 32, 321-341. Fang, C. Y., Manne, S. L., & Pape, S. J. (2001). Functional impairment, marital quality, and patient psychological distress as predictors of psychological distress among cancer patients' spouses. Health Psychology, 20(6), 452-457. Feeley, N., & Gottlieb, L. N. (1988-1989). Parents’ coping and communication following their infants’ death. Omega: Journal of Death and Dying, 19, 51-67. Figley, C. R. (1983). Catastrophes: An overview of family reactions. In C. R. Figley & H. I. McCubbin (Eds.), Stress and the family (pp. 3-20). New York: Brunner/Mazel. Fincham, F. D., & Bradbury, T. N. (1987). The assessment of marital quality: A reevaluation. Journal of Marriage & the Family, 49(4), 797-809. Fincham, F. D., & Bradbury, T. N. (1993). Marital satisfaction, depression, and attributions: A longitudinal analysis. Journal of Personality & Social Psychology, 64, 442-452. Fisher, J. D., Nadler, A., & Whitcher-Alagna, S. (1982). Recipient reactions to aid. Psychological Bulletin, 91, 27-54. Flor, H., Turk, D. C., & Rudy, T. E. (1989). Relationship of pain impact and significant other reinforcement of pain behaviors: The mediating role of gender, marital status and marital satisfaction. Pain, 38(1), 45-50. Flor, H., Turk, D. C., & Scholz, O. B. Impact of chronic pain on the spouse: Marital, emotional and physical consequences. Journal of Psychosomatic Research, 31(1), 63-71. Floyd, F. J., & Markman, H. J. (1983). Observational biases in spouse observation: Toward a cognitive/behavioral model of marriage. Journal of Consulting and Clinical Psychology, 51, 450-457.

PAGE 169

158 Folkman, S. (1984). Personal control and stress and coping processes: A theoretical analysis. Journal of Personality & Social Psychology, 46, 839-852. Folkman, S., Lazarus, R. S., Dunkel-Schetter, C., DeLongis, A., & Gruen, R. J. (1986). Dynamics of a stressful encounter: Cognitive appraisal, coping, and encounter outcomes. Journal of Personality and Social Psychology, 50, 992-1003. Folkman, S., Lazarus, R. S., & Gruen, R. J. (1986). Appraisal, coping, health status, and psychological symptoms. Journal of Personality and Social Psychology, 50, 571-579. Fraley, R. C., & Shaver, P. R. (2000). Adult romantic attachment: Theoretical developments, emerging controversies, and unanswered questions. Review of General Psychology, 4, 132-154. Frankel, H., Snowden, L. S., & Nelson, L. S. (1992). Wives' adjustment to military deployment: an empirical evaluation of a family stress model. International Journal of Sociology of the Family, 22, 93-117. Friedman, D., Holmbeck, G. N., & Jandasek, B. (2004). Parent functioning in families of preadolescents with spina bifida: Longitudinal implications for child assessment. Journal of Family Psychology, 18(4), 609-619. Friedrich, W. N., Cohen, D. S., & Wilturner, L. S. (1987). Family relations and marital quality when a mentally handicapped child is present. Psychological Reports, 61(3), 911-919. Furman, W. (1984). Some observations on the study of personal relationships. In J. C. Masters & K. Yarkin-Levin (Eds.), Boundary areas in social and developmental psychology (pp. 15-42). Orlando: Academic Press. Furstenberg, F. F. (1994). History and current status of divorce in the United States. Future of Children, 4, 29-43. Gabbard, G. O., Menninger, R. W., & Coyne, L. (1987). Sources of conflict in the medical marriage. American Journal of Psychiatry, 144(5), 567-572. Gannon, L., & Pardie, L. (1989). The importance of chronicity and controllability of stress in the context of relationships. Journal of Behavioral Medicine, 12, 357-372. Garts, K., & Garland, S. (1983). Marital satisfaction of the post-rehabilitation burn patient. Occupational Health Nursing, 35-37. Giese-Davis, J., Hermanson, K., Koopman, C., Weibel, D., & Spiegel, D. (2000). Quality of couples' relationship and adjustment to metastatic breast cancer. Journal of Family Psychology, 14(2), 251-266.

PAGE 170

159 Gill, D. S., Christensen, A., & Fincham, F. D. (1999). Predicting marital satisfaction from behavior: Do all roads really lead to Rome? Personal Relationships, 6, 369-387. Gimbel, C., & Booth, A. (1994). Why does military combat experience adversely affect marital relations? Journal of Marriage & the Family, 56, 691-703. Glass, D. C., & Singer, J. E. (1977). Environmental stress and the adaptive process. In A. Monat & R. Lazarus (Eds.), Stress and coping: An anthology (pp. 132-144). New York: Columbia University Press. Glenn, N. D. (1990). Quantitative research on marital quality in the 1980's: A critical review. Journal of Marriage and the Family, 52, 818-831. Glenn, N. D., & Weaver, C. N. (1981). The contribution of marital happiness to global happiness. Journal of Marriage and the Family, 43, 61-68. Glick, I. D., Weiss, R. S., Parkes, C. M. (1974). The first year of bereavement. Oxford, England: John Wiley and Sons. Glickman, L., Tanaka, J. S., & Chan, E. (1991). Life events, chronic strain, and psychological distress: Longitudinal causal models. Journal of Community Psychology, 19(4), 283-305. Godwin, D. D., Draughn, P. S., Little, L. F., & Marlowe, J. (1991). Wives' off-farm employment, farm family economic status, and family relationships. Journal of Marriage & the Family, 53(2), 389-402. Goode, W. J. (1993). World changes in divorce patterns. New Haven: Yale University Press. Goodwin, S. (1997). The marital relationship and health in women with chronic fatigue and immune dysfunction syndrome: Views of wives and husbands. Nursing Research, 46(3), 138-146. Gosling, J., & Oddy, M. (1999). Rearranged marriages: Marital relationships after head injury. Brain Injury, 13(10), 785-796. Gottlieb, B. H. (1997). Coping with chronic stress. New York: Plenum Press. Gottman, J. M. (1993). A theory of marital dissolution and stability. Journal of Family Psychology, 7, 57-75. Gottman, J. M. (1998). Psychology and the study of the marital processes. Annual Review of Psychology,49, 169-197. Gottman, J. M., Coan, J., Carrere, S., & Swanson, C. (1998). Predicting marital happiness and stability from newlywed interactions. Journal of Marriage & the Family, 60, 5-22.

PAGE 171

160 Gottman, J. M., & Krokoff, L. J. (1989). Marital interaction and satisfaction: A longitudinal view. Journal of Consulting & Clinical Psychology, 57, 47-52. Graig, E. (1993). Stress as a consequence of the urban physical environment. In L. Goldberger & S. Breznitz (Eds.) Handbook of stress: Theoretical and clinical aspects (2nd ed) (pp. 316-332). New York: Free Press. Green, B. F., & Hall, J. A. (1984). Quantitative methods of literature reviews. Annual Review of Psychology, 35, 37-53. Gritz, E. R., Wellisch, D. K., Siau, J., & Wang, H. (1990). Long-term effects of testicular cancer on marital relationships. Psychosomatics: Journal of Consultation Liaison Psychiatry, 31, 301-312. Grych, J. H., & Fincham, F. D. (1990). Marital conflict and children's adjustment: A cognitive-contextual framework. Psychological Bulletin, 108(2), 267-290. Grzywacz, J. G., Almeida, D. M., & McDonald, D. A. (2002). Work-family spillover and daily reports of work and family stress in the adult labor force. Family Relations: Interdisciplinary Journal of Applied Family Studies, 51, 28-36. Gump, B. B., & Matthews, K. A. (1999). Do background stressors influence reactivity to and recovery from acute stressors? Journal of Applied Social Psychology, 29, 469-494. Gunthert, K. C., Cohen, L. H., & Armeli, S. (1999). The role of neuroticism in daily stress and coping. Journal of Personality and Social Psychology, 77, 1087-1100. Hafstrom, J. L., & Schram, V. R. (1984). Chronic illness in couples: Selected characteristics, including wife's satisfaction with and perception of marital relationships. Family Relations, 33(1), 195-203. Hagedoorn, M., Kuijer, R. G., Buunk, B. P., DeJong, G. M., Wobbes, T., & Sanderman, R. (2000). Marital satisfaction in patients with cancer: Does support from intimate partners benefit those who need it most? Health Psychology, 19(3), 274-282. Halford, W. K., Gravestock, F. M., Lowe, R., & Scheldt, S. (1992). Toward a behavioral ecology of stressful marital interactions. Behavioral Assessment, 14(2), 199-217. Halford, W. K., Scott, J. L., & Smythe, J. (2000). Couples and coping with cancer: Helping each other through the night. In K. B. Schmaling & T. B. Sher (Eds.), The psychology of couples and illness: Theory, research, & practice (pp. 135-170). Washington D.C.: American Psychological Association. Hall, J. A., Tickle-Degnen, L., Rosenthal, R., & Mosteller, F. (1994). Hypotheses and problems in research synthesis. In H. Cooper & L.V. Hedges (Eds.), The Handbook of research synthesis (pp. 17-28). New York: Russell Sage Foundation.

PAGE 172

161 Halverson, C. F. (1995). Measurement beyond the individual. In J.J. Conoley & E.B. Werth (Eds.), Family measurement (pp. 3-18). Lincoln, NE: Buros Institute of Mental Measurements. Hammen, C., & Brennan, P. A. (2002). Interpersonal dysfunction in depressed women: Impairments independent of depressive symptoms. Journal of Affective Disorders, 72, 145-156. Hannum, J. W., Giese-Davis, J., Harding, K., & Hatfield, A. K. Effects of individual and marital variables on coping with cancer. Journal of Psychosocial Oncology, 9(2), 1-20. Harper, J. M., Schaalje, B. G., & Sandberg, J. G. (2000). Daily hassles, intimacy, and marital quality in later life marriages. American Journal of Family Therapy,28(1), 1-18. Harrison, T., Stuifbergen, A., & Eishi, A. (2004). Marriage, impairment, and acceptance in persons with multiple sclerosis. Western Journal of Nursing Research, 26(3), 266-285. Hazan, C., & Shaver, P. R. (1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52, 511-524. Hedges, L. V. (1982a). Estimation of effect size from a series of independent experiments. Psychological Bulletin, 92, 490-499. Hedges, L. V. (1982b). Fitting categorical models to effect sizes from a series of experiments. Journal of Educational Statistics, 7, 119-137. Hedges, L. V. (1994). Statistical considerations. In H. Cooper & L. V. Hedges (Eds.), Handbook of research synthesis. (pp. 29-38). New York, NY: Russell Sage Foundation. Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. Orlando, FL: Academic Press. Hicks, M. W., & Platt, M. (1970). Marital happiness and stability: A review of the research in the sixties. Journal of Marriage and the Family, 32, 553-574. Higginbottom, S. F., Barling, J., & Kelloway, E. K. (1993). Linking retirement experiences and marital satisfaction: A mediational model. Psychology & Aging, 8, 508-516. Hill, R. (1949). Families under stress. New York: Harper & Row. Hill, R. (1964). Methodological Issues in Family Development Research. Family Process, 3, 186-206.

PAGE 173

162 Hobfoll, S. E. (1998). Stress, culture, and community: The psychology and philosophy of stress. New York: Plenum Press. Hodgson, J. H., Shields, C. G., & Rousseau, S. L. (2003). Disengaging communication in later-life couples coping with breast cancer. Families, Systems, and Health, 21(2), 145-163. Horwitz, A. V., McLaughlin, J., & White, H. R. (1997). How the negative and positive aspects of partner relationships affect the mental health of young married people. Journal of Health and Social Behavior, 39, 124-136. House, J. S., Landis, K. R., & Umberson, D. (1988). Social relationships and health. Science, 241, 540-545. Howe, G. E., Levy, M. L., & Caplan, R. D. (2004). Job loss and depressive symptoms in couples: Common stressors, stress transmission, or relationship disruption? Journal of Family Psychology, 18(4), 639-650. Hraba, J., Lorenz, F. O., & Pechacova, Z. (2000). Family stress during the Czech transformation. Journal of Marriage and the Family, 62, 520-531. Hughes, D. L., & Galinsky, E. (1994). Work experiences and marital interactions: Elaborating the complexity of work. Journal of Organizational Behavior, 15(5), 423-438. Hurvitz, N. Marital roles strain as a sociological variable. Family Life Coordinator, 39-42. Huston, T. L. (2000). The social ecology of marriage and other intimate unions. Journal of Marriage and the Family, 62, 298-320. Hyun, E. M., Bauer, J. W., & Hogan, M. J. (1993). Resource adequacy perception and marital satisfaction of rural wives and husbands: A nonrecursive model. Journal of Family and Economic Issues, 14(3), 215-236. Isaac, R., & Shah, A. (2004). Sex roles and marital adjustment in Indian couples. International Journal of Social Psychiatry, 50(2), 129-141. Jackson, S. E., & Maslach, C. (1982). After-effects of job-related stress: Families as victims. Journal of Occupational Behaviour, 3(1), 63-77. Janis, I. L. (1993). Decision-making under stress. In L. Goldberger & S. Breznitz (Eds.), Handbook of stress: Theoretical and clinical aspects (2 nd ed.) (pp. 56-74). New York, NY: Free Press. Johnson, C. L. (1985). The impact of illness on late-life marriages. Journal of Marriage and the Family, 165-172.

PAGE 174

163 Johnson, D. R., & Booth, A. (1990). Rural economic decline and marital quality: A panel study of farm marriages. Family Relations: Interdisciplinary Journal of Applied Family Studies, 39(2), 159-165. Jones, L. (1989). The relationship between unemployment and divorce. Journal of Divorce, 12, 99-112. Jorgensen, S. R. (1979). Socioeconomic rewards and perceived marital quality: A re-examination. Journal of Marriage and the Family, 41(4), 825-835. Julien, D., & Markman, H. J. (1991). Social support and social networks as determinants of individual and marital outcomes. Journal of Social & Personal Relationships, 8(4), 549-568. Kaplan, D., Grobstein, R., & Smith, A. (1976). Predicting the impact of severe illness in families. Health and Social Work, 1, 1-82. Karney, B. R., & Bradbury, T. N. (1995). The longitudinal course of marital quality and stability: A review of theory, methods, and research. Psychological Bulletin, 118, 3-34. Karney, B. R., & Bradbury, T. N. (1997). Neuroticism, marital interaction, and the trajectory of marital satisfaction. Journal of Personality & Social Psychology, 72(5), 1075-1092. Karney, B. R., Bradbury, T. N., & Johnson, M. D. (1999). Deconstructing marital stability: Distinguishing between marital dissolution and the trajectory of marital satisfaction. In J. M. Adams & W. H. Jones (Eds.), Handbook of interpersonal commitment and relationship stability (pp. 481-499). New York: Plenum Publishing. Karney, B. R., & Frye, N. E. (2002). "But we've been getting better lately": Comparing prospective and retrospective views of relationship development. Journal of Personality & Social Psychology, 82(2), 222-238. Karney, B. R., McNulty, J. K., & Bradbury, T. N. (2000). Cognition and the development of close relationships. In G. J. Fletcher & M. S. Clark (Eds.), Blackwell handbook in social psychology, Vol.2: Interpersonal processes (pp. 32-59). Oxford: Blackwell Publishers Limited. Karney, B. R., & Springer, S. H. (2004). Should promoting marriage be the next stage of welfare reform? In A. C. Crouter & A. Booth (Eds.), Work-family challenges for low-income parents and their children. (pp. 243-250). Mahwah, NJ: Lawrence Erlbaum Associates.

PAGE 175

164 Karney, B. R., Story, L. B., & Bradbury, T. N. (2005). Marriages in context: Interactions between chronic and acute stress among newlyweds. In T. A. Revenson, K. A. Kayser, & G. Bodenmann (Eds.). Couples coping with stress: Emerging perspectives on dyadic coping (pp. 13-32). Washington, DC: American Psychological Association. Katz, J., Monnier, J., Libet, J., Shaw, D., & Beach, S. R. H. (2000). Individual and crossover effects of stress on adjustment in medical student marriages. Journal of Marital and Family Therapy, 26(3), 341-351. Kazak, A. E. (1987). Families with disabled children: Stress and social networks in three samples. Journal of Abnormal Child Psychology, 15(1), 137-146. Kazak, A. E. (1989). Family functioning in families with older institutionalized retarded offspring. Journal of Autism & Developmental Disorders, 19(4), 501-509. Kazak, A. E., & Clark, M. W. (1986). Stress in families of children with myelomeningocele. Developmental Medicine and Child Neurology, 28(2), 220-228. Kazak, A. E., Reber, M., & Snitzer, L. (1988). Childhood chronic disease and family functioning: a study of phenylketonuria. Pediatrics, 81(2), 224-230. Keller, M., Henrich, G., Sellschopp, A., & Beutel, M. Between distress and support: Spouses of cancer patients. Cancer and the family., 187-223. Kelloway, E. K., & Barling, J. (1994). Stress, control, well-being, and marital functioning: A causal correlational analysis. Job stress in a changing workforce: Investigating gender, diversity, and family issues, (Ed. Hurrell), 241-251. Kennedy, S., Kiecolt-Glaser, J. K., & Glaser, R. (1988). Immunological consequences of acute and chronic stressors: Mediating role of interpersonal relationships. British Journal of Medical Psychology, 61, 77-85. Kiecolt-Glaser, J. K. (1999). Stress, personal relationships, and immune function: Health implications. Brain, Behavior and Immunity, 13, 61-72. Kiecolt-Glaser, J. K., Fisher, L. D., Ogrocki, P., & Stout, J. C. (1987). Marital quality, marital disruption, and immune function. Psychosomatic Medicine, 49, 13-34. Kiecolt-Glaser, J., & Glaser, R. (1988). Major life changes, chronic stress, and immunity. Psychological, neuropsychiatric, and substance abuse aspects of AIDS, 217-224. Kiecolt-Glaser, J. K., Kennedy, S., & Malkoff, S. (1988). Marital discord and immunity in males. Psychosomatic Medicine, 50, 213-229. Kiecolt-Glaser, J., & Newton, T. L. (2001). Marriage and health: his and hers. Psychological Bulletin, 127, 472-503.

PAGE 176

165 Kinnunen, U., & Feldt, T. (2004). Economic stress and marital adjustment among couples: Analyses at the dyadic level. European Journal of Social Psychology, 34(5), 519-532. Kinnunen, U., Geurts, S., & Mauno, S. (2004). Work-to-family conflict and its relationship with satisfaction and well-being: A one-year longitudinal study on gender differences. Work & Stress, 18(1), 1-22. Kinnunen, U., & Pulkkinen, L. (1998). Linking economic stress to marital quality among Finnish marital couples. Journal of Family Issues, 19(6), 705-724. Kinnunen, U., Vermulst, A., & Gerris, J. (2003). Work-family conflict and its relation to well-being: The role of personality as a moderating factor. Personality and Individual Differences, 35(7), 1669-1683. Kinsella, G., & Duffy, F. D. (1979). Psychosocial readustment in the spouses of aphasic patients. Scandinavian Journal of Rehabiltative Medicine, 11, 129-132. Klein, K., Forehand, R., Armistead, L., & Wierson, M. (1994). The contributions of social support and coping methods to stress resiliency in couples facing hemophilia and HIV. Advances in Behaviour Research and Therapy, 16, 253-275. Knee, C. R. (1998). Implicit theories of relationships: Assessment and prediction of romantic relationship initiation, coping and longevity. Journal of Personality and Social Psychology, 74, 360-370. Kohn, P. M., Lafreniere, K., & Gurevich, M. (1990). The Inventory of College Student's Recent Life Experiences: A decontaminated hassles scale for a special population. Journal of Behavioral Medicine, 13, 619-630. Komarovsky, M. (1940). The unemployed man and his family: The effect of unemployment upon the status of the man in fifty-nine families. Ft. Worth, TX: Dryden Press. Konstam, V., Surman, O., Hizzazi, K. H., Fierstein, J., Konstam, M., Turbett, A., et al. (1998). Marital adjustment in heart transplantation patients and their spouses: A longitudinal perspective. The American Journal of Family Therapy, 26(2), 147-158. Kreider, R. M., & Fields, J. M. (2001). Number, timing, and duration of marriages and divorces: Fall 1996. In Current population reports (pp. 70-80). Washington, DC: U.S. Census Bureau. Kuijer, R. G., Buunk, B. P., Ybema, J. F., & Wobbes, T. (2002). The relation between perceived inequity, marital satisfaction and emotions among couples facing cancer. British Journal of Social Psychology, 41(1), 39-56. Kunzer, M. B. (1987). Marital adjustment of headache: Sufferers and their spouses. Journal of Psychosocial Nursing and Health, 25(5), 12-17.

PAGE 177

166 Kurdek, L. A. (1991). Marital stability and changes in marital quality in newlywed couples: A test of the contextual model. Journal of Social and Personal Relationships, 8(1), 27-48. Kurdek, L. A. (1993). Predicting marital dissolution: A five-year prospective longitudinal study of newlywed couples. Journal of Personality and Social Psychology, 64, 221-242. Kwon, H., Rueter, M. A., & Lee, M. (2003). Marital relationships following the Korean economic crisis: Applying the family stress model. Journal of Marriage and the Family, 65(2), 316-325. Landman, J. T., & Dawes, R. M. (1982). Psychotherapy outcome: Smith and Glass’ conclusions stand up under scrutiny. American Psychologist, 37, 504-516. Lang, A., & Gottlieb, L. (1991). Marital intimacy in bereaved and nonbereaved couples: A comparative study. In D. Papadatou & C. Papadatous (Eds.), Children and death (pp. 267-275). New York: Hemisphere Publishing Corporation. Larsen, A. S., & Olson, D. H. (1989). Predicting marital success using PREPARE: A replication study. Journal of Marital and Family Therapy, 15, 311-322. Larson, J. H. (1984). The effect of husband's unemployment on marital and family relations in blue-collar families. Journal of Applied Family and Child Studies, 33(4), 503-511. Larson, J. H., Wilson, S. M., & Beley, R. (1994). The impact of job insecurity on marital and family relationships. Family Relations: Interdisciplinary Journal of Applied Family Studies, 43, 128-143. Larson, R. W., & Almeida, D. M. (1999). Emotional transmission in the daily lives of families: A new paradigm for studying family process. Journal of Marriage & the Family, 61, 5-20. Laufer, R. S., & Gallops, M. S. (1985). Life-course effects of Vietnam combat and abusive violence: marital patterns. Journal of Marriage & Family, 839-853. Lavee, Y., McCubbin, H. I., & Olson, D. H. (1987). The effect of stressful life events and transitions on family functioning and well-being. Journal of Marriage & the Family, 49, 857-873. Lavee, Y., & Mey-Dan, M. (2003). Patterns of change in marital relationships among parents of children with cancer. Health & Social Work, 28(4), 255-263. Lazarus, R. S. (1977). Cognitive and coping processes in emotion. In A. Monat & R. Lazarus (Eds.), Stress and coping: An anthology (pp. 145-158). New York: Columbia University Press.

PAGE 178

167 Lazarus, R., S., Delongis, A., & Folkman, S. (1985). Stress and adaptational outcomes: The problem of confounded measures. American Psychologist, 40, 770-779. Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal and coping. New York: Springer Publishing Company. Lee, C., & Iverson-Gilbert, J. (2003). Demand, support, and perception in family-related stress among Protestant clergy. Family Relations: Interdisciplinary Journal of Applied Family Studies, 52(3), 249-257. Lehman, D. R., Lang, E. L., Wortman, C. B., & Sorenson, S. B. (1989). Long-term effects of sudden bereavement: Marital and parent-child relationships and children's reactions. Journal of Family Psychology, 2(3), 344-367. Leiblum, S. R., & Hamer, R. (1998). Life after infertility treatment: a long-term investigation of marital and sexual function. Human Reproduction, 13(12), 3569-3574. Leiter, M. P., & Durup, M. J. (1996). Work, home, and in-between: A longitudinal study of spillover. Journal of Applied Behavioral Science, 32, 29-47. Lepore, S. J., Evans, G. W., & Palsane, M. N. (1991). Social hassles and psychological health in the context of chronic crowding. Journal of Health & Social Behavior, 32, 357-367. Lepore, S. J., Evans, G. W., & Schneider, M. L. (1991). Dynamic role of social support in the link between chronic stress and psychological distress. Journal of Personality and Social Psychology, 61, 899-909. Leslie, L. A. & Letiecq, B. L. (2004). Marital quality of African American and white partners in interracial couples. Personal Relationships, 11(4), 559-574. Lester, D., & Guerin, T. W. (1982). Further explorations of police officers' satisfaction with their marriages. Psychological Reports, 50(2), 608. Lester, D., & Karsevar, M. (1980). Perceived satisfaction of police officers with their marriages. Perceptual & Motor Skills, 51(2), 634. Levinger, G. (1980). Toward the analysis of close relationships. Journal of Experimental Social Psychology, 16, 510-544. Levinger, G. (1994). Attachment theory as a paradigm for studying close relationships. Psychological Inquiry, 5, 45-47. Levinger, G., & Huston, T. L. (1990). The social psychology of marriage. In F. D. Fincham & T. N. Bradbury (Eds.), The psychology of marriage: Basic issues and applications (pp. 19-58). New York: Guilford.

PAGE 179

168 Lewin, K. (1951). Field theory in social science. New York: Harper. Lewis, J. M., Barnhart, F. D., Nace, E. P., & Carson, D. I. (1993). Marital satisfaction in the lives of physicians. Bulletin of the Menninger Clinic, 57(4), 458-465. Lewis, F. M., Woods, N. F., Hough, E. E., & Bensley, L. S. The family's functioning with chronic illness in the mother: The spouse's perspective. Social Science & Medicine, 29(11), 1261-1269. Lieberman, M. A. (1986). Social supports: The consequences of psychologizing: A commentary. Journal of Consulting & Clinical Psychology, 54, 461-465. Liem, R., & Liem, J. H. (1988). Psychological effects of unemployment on workers and their families. Journal of Social Issues, 44(4), 87-105. Liker, J. K., & Elder, G. H., Jr. (1983). Economic hardship and marital relations in the 1930s. American Sociological Review, 48, 343-359. Lipsey, M. W., & Wilson, D. B. (2001). Practical meta-analysis. Thousand Oaks, CA: Sage Press. Lopez, F. G. (1993). Cognitive processes in close relationships: Recent findings and implications for counseling. Journal of Counseling & Development, 71, 310-315. Lou, V. W., Chow, J. C., & Chan, C. L. (2004). Impact of expectation fulfillment on post-migration marital happiness among mainland Chinese wives with Hong Kong husbands. Journal of Social Work Research & Evaluation, 5(1), 99-112. MacLean, A. P., & Peters, R. D. (1995). Graduate student couples: Dyadic satisfaction in relation to type of partnership and demographic characteristics. Canadian Journal of Behavioral Science, 27(1), 120-124. Manne, S. L., Alfieri, T., Taylor, K. L., & Dougherty, J. (1999). Spousal negative responses to cancer patients: The role of social restriction, spouse mood, and relationship satisfaction. Journal of Consulting & Clinical Psychology, 67(3), 352-361. Manne, S., Sherman, M., & Ross, S. (2004). Couples’ support-related communication, psychological distress, and relationship satisfaction among women with early stage breast cancer. Journal of Consulting and Clinical Psychology, 72(4), 660-670. Marco, C. A., & Suls, J. (1993). Daily stress and the trajectory of mood: Spillover, response assimilation contrast, and chronic negative affectivity. Journal of Personality and Social Psychology, 64, 1053-1063. Markman, H. J. (1979). Application of a behavioral model of marriage in predicting relationship satisfaction of couples planning marriage. Journal of Consulting and Clinical Psychology, 47, 743-749.

PAGE 180

169 Martin, J. A., & Ickovics, J. R. (1987). The effects of stress on the psychological well-being of army wives: Initial findings from a longitudinal study. Journal of Human Stress, 13(3), 108-115. Maruta, T., Osborne, D., Swanson, D. W., & Halling, J. M. (1981). Chronic pain patients and spouses. Mayo Clinical Proc, 56, 307-310. Matt, G. E., & Cook, T. D. (1994). Threats to the validity of research synthesis. In H. Cooper & L. V. Hedges (Eds). (pp. 503-520). Handbook of research synthesis. New York, NY: Russell Sage Foundation. Matthews, L. S., Conger, R. D., & Wickrama, K. A. S. (1996). Work-family conflict and marital quality: Mediating processes. Social Psychology Quarterly, 59, 62-79. Mauno, S., & Kinnunen, U. (1999). The effects of job stressors on marital satisfaction in Finnish dual-earner couples. Journal of Organizational Behavior, 20, 879-895. McAdoo, H. P. (1982). Stress absorbing in Black families. Family Relations: Journal of Applied Family and Child Studies, 31, 479-488. McAdoo, H. P. (1983). Societal stress: the black family. In S. Figley & H. McCubbin (Eds.), Stress and the family, Vol. 1, 178-187. McCabe, M. P. (2004). Exacerbation of symptoms among people with multiple sclerosis: Impact on sexuality and relationships over time. Archives of Sexual Behavior, 33(6), 593-601. McCubbin, H. I. (1967). Resiliency in African-American Families. Thousand Oaks, CA: Sage.McCubbin, H. I., & Dahl, B. B. (1985). Marriage and family: Individuals and life cycles. New York: Wiley. McCubbin, H. I., & Dahl, B. (1985). Marriage and family: Individuals and life cycles. New York: John Wiley. McCubbin, H. I., & Figley, C. R. (1983). Bridging normative and catastrophic family stress. In H. I. McCubbin & C. R. Figley (Eds.), Stress and the family: Coping with normative transitions, Vol. 1 (pp. 218-228). New York: Brunner/Mazel. McCubbin, M. A., & McCubbin, H. I. (1989). Theoretical orientations to family stress and coping. Treating stress in families (pp. 3-43). New York: Brunner/Mazel. McCubbin, H. I., & Patterson, J. M. (1982). Family adaptation to crises. In H. I. McCubbin, A. E. Cauble & J. M. Patterson (Eds.), Family stress, coping, and social support (pp. 26-47). Springfield, IL: Thomas. McCubbin, H. I., & Patterson, J. M. (1983a). Family transitions: Adaptation to stress. In H. I. McCubbin & C. R. Figley (Eds.), Stress and the family: Coping with normative transitions, Vol.1 (pp. 5-25). New York: Brunner/Mazel.

PAGE 181

170 McCubbin, H. I., & Patterson, J. M. (1983b). The Family Stress Process: The Double ABCX Model of adjustment and adaptation. Marriage & Family Review, 6, 7-37. McEwen, B.S. (1998). Protective and damaging effects of stress mediators. New England Journal of Medicine, 338, 171-179. McGonagle, K. A., & Kessler, R. C. (1990). Chronic stress, acute stress, and depressive symptoms. American Journal of Community Psychology, 18, 681-706. McLean, D. E., & Link, B. G. (1994). Unraveling complexity: Strategies to refine concepts, measures, and research designs in the study of life events and mental health. In W. R. Avison & I. H. Gotlib (Eds). Stress and mental health: Contemporary issues and prospects for the future (pp. 15-42). New York: Plenum Press. McNulty, J. K., & Karney, B. R. (2002). Expectancy confirmation in appraisals of marital interactions. Personality & Social Psychology Bulletin, 28(6), 764-775. Mechanic, D. (1998). Stress-moderating and amplifying factors. In P. B. Dohrenwend, B. P. (Ed.) Adversity, stress, and psychopathology (pp. 371-389). Cambridge: Oxford University Press. Meeks, S., Arnkoff, D. B., Glass, C. R., & Notarius, C. I. (1986). Wives' employment status, hassles, communication, and relational efficacy: Intraversus extra-relationship factors and marital adjustment. Family Relations: Journal of Applied Family & Child Studies, 35(2), 249-255. Menaghan, E. (1983). Marital stress and family transitions: A panel analysis. Journal of Marriage and the Family, 45, 371-386. Miller, B., Townsend, A. L. & Ishler, K. J. (2004). Change in marital dissatisfaction, health and depression in older married couples. Journal of Mental Health and Aging, 10(1), 65-77. Moen, P., Kim, J. E., & Hofmeister, H. (2001). Couples' work/retirement transitions, gender, and marital quality. Social Psychology Quarterly, 64(1), 55-71. Mohamed, S. N., Weisz, G. M., & Waring, E. M. (1978). The relationship of chronic pain to depression, marital adjustment, and family dynamics. Pain, 5(3), 285-292. Moles, O. C. (1979). Public welfare payments and marital dissolution: A study of recent reviews. In G. Levinger & O. C. Moles (Eds.), Divorce and separation: Context, causes, and consequences (pp. 167-180). New York: Basic Books, Inc. Monga, M., Alexandrescu, B., Katz, S. E., Stein, M., & Ganiats, T. (2004). Impact of infertility on quality of life, marital adjustment, and sexual function. Urology, 63(1), 126-130.

PAGE 182

171 Moore, A. D., Bombardier, C. H., Brown, P. B. & Patterson, D. R. (1994). Coping and emotional attributions following spinal cord injury. International Journal of Rehabilitation and Research, 17, 39-48. Morris, L. M., & Blanton, P. W. (1994). The influence of work-related stressors on clergy husbands and their wives. Family Relations: Interdisciplinary Journal of Applied Family Studies, 43(2), 189-195. Mortimer, J. T. (1980). Occupation-family linkages as perceived by men in the early stages of professional and managerial careers. Research in the Interweave of Social Roles: Women and Men, 1, 99-117. Murray, S. L., & Holmes, J. G. (1993). Seeing virtues in faults: Negativity and the transformation of interpersonal narratives in close relationships. Journal of Personality and Social Psychology, 65, 707-722. Murray, S. L., & Holmes, J. G. (1997). A leap of faith? Positive illusions in romantic relationships. Personality and Social Psychology Bulletin, 23, 586-604. Myers, S. M., & Booth, A. (1996). Men's retirement and marital quality. Journal of Family Issues, 17(3), 336-357. Najman, J. M., Vance, J. C., Boyle, F., Embleton, G., Foster, & Thearle. (1993). The impact of a child death on marital adjustment. Social Science & Medicine, 37(8), 1005-1010. Neff, L. A., & Karney, B. R. (2004). How does context affect intimate relationships? Linking external stress and cognitive processes within marriage. Personality and Social Psychology Bulletin, 30(2), 134-148. Negy, C., & Snyder, D. K. (2000). Relationship satisfaction of Mexican American and non-Hispanic White American interethnic couples: Issues of acculturation and clinical intervention. Journal of Marital & Family Therapy, 26(3), 293-304. Northouse, L. L., Jeffs, M., Cracchiolo-Caraway, A., Lampman, L., & Dorris (1995). Emotional distress reported by women and husbands prior to breast biopsy. Nursing Research, 44, 196-201. Northouse, L. L., Mood, D., Templin, T., Mellon, S., & George, T. (2000). Couples' patterns of adjustment to colon cancer. Social Science & Medicine, 50(2), 271-284. Norton, R. (1983). Measuring marital quality: A critical look at the dependent variable. Journal of Marriage and the Family, 45, 141-151. Oliver, L. E. (1999). Effects of a child's death on the marital relationship: A review. Omega: Journal of Death & Dying, 39(3), 197-227.

PAGE 183

172 Olson, D. H., Lavee, Y., & McCubbin, H.I. (1988). Types of families and family response to stress across the family life cycle. In D. M. Klein & J. Aldous (Eds.), Social stress and family development (pp. 16-43). New York: Guilford Press. Orbuch, T. L., House, J. S., Mero, R. P., & Webster, P. S. (1996). Marital quality over the life course. Social Psychology Quarterly, 59(2), 162-171. Orbuch, T. L., Veroff, J., Hassan, H., & Horrocks, J. (2002). Who will divorce: A 14-year longitudinal study of Black couples and White couples. Journal of Social and Personal Relationships, 19(2), 179-202. Park, J., Turnbull, A. P., & Turnbull, H. R. (2002). Impacts of poverty on quality of life in families of children with disabilities. Exceptional children, 68(2), 151-170. Parkes, C., & Weiss, R. (1983). Recovery from bereavement. New York: Basic Books. Paulus, P. B., Nagar, D., Larey, T. S., & Camacho, L. M. (1996). Environmental, lifestyle, and psychological factors in the health and well-being of military families. Journal of Applied Social Psychology, 26(23), 2053-2075. Pearlin, L. I., Menaghan, E. G., Lieberman, M. A., & Mullan, J. T. (1981). The stress process. Journal of Health & Social Behavior, 22, 337-356. Pearlin, L. I., & Schooler, C. (1978). The structure of coping. Journal of Health & Social Behavior, 19, 2-21. Perlow, A. D., & Mullins, S. C. (1976). Marital satisfaction as perceived by the medical student's spouse. Journal of Medical Education, 51(9), 726-734. Perrucci, C. C., & Targ, D. B. (1988). Effects of a plant closing on marriage and family life. Families and economic distress: Coping strategies and social policy., 55-71. Perry, A., Sarlo-McGarvey, N., & Factor, D. C. (1992). Stress and family functioning in parents of girls with Rett syndrome. Journal of Autism and Developmental Disorders, 22(2), 235-248. Perry-Jenkins, M., Repetti, R. L., & Crouter, A. C. (2000). Work and family in the 1990s. Journal of Marriage & the Family, 62, 981-998. Peters, M. F., & Massey, G. (1983). Mundane extreme environmental stress. In Family stress theories: The case of black families in white America (pp. 193-218). Haworth Press. Peterson, B. D., Newton, C. R., & Rosen, K. H. (2003). Examining congruence between partners’ perceived infertility-related stress and its relationship to marital adjustment and depression in infertile couples. Family Process, 42(1), 59-70.

PAGE 184

173 Pillow, D. R., Zautra, A. J., & Sandler, I. (1996). Major life events and minor stressors: Identifying mediational links in the stress process. Journal of Personality and Social Psychology, 70, 381-394. Pittman, J. F. (1994). Work/family fit as a mediator of work factors on marital tension: Evidence from the interface of greedy institutions. Human Relations, 47(2), 183-209. Pittman, J. F., & Lloyd, S. A. (1988). Quality of family life, social support, and stress. Journal of Marriage and the Family, 50(1), 53-67. Pond, S. B., & Green, S. B. (1983). The relationship between job and marriage satisfaction within and between spouses. Journal of Occupational Behavior, 4(2), 145-155. Powers, A. S., Myers, J. E., & Tingle, L. R. (2004). Wellness, perceived stress, mattering, and marital satisfaction among medical residents and their spouses: Implications for education and counseling. Family Journal: Counseling and Therapy for Couples & Families, 12(1), 26-36. Pradhan, M., & Misra, N. (1995). Spouse support and quality of marital relationship as correlates of stress. Journal of the Indian Academy of Applied Psychology, 21(1), 43-50. Presser, H. B. (2000). Nonstandard work schedules and marital instability. Journal of Marriage and the Family, 62, 93-110. Previti, D., & Amato, P. R. (2003). Why stay married? Rewards, barriers, and marital stability. Journal of Marriage & Family, 65(3), 561-573. Proulx, C. M., Helms, H. M., & Payne, C. C. (2004). Wives’ domain-specific “marriage work” with friends and spouses: Links to marital quality. Family Relations: Interdisciplinary Journal of Applied Family Studies, 53, 393-404. Quek, K. F., Loh, C. S., & Low, W. Y. (2003). The male marital satisfaction following treatment for lower urinary tract symptoms. Journal of Sex and Marital Therapy, 29(2), 173-176. Quittner, A. L., Espelage, D. L., Opipari, L. C., Carter, B., Eid, N., & Eigen, H. (1998). Role strain in couples with and without a child with a chronic illness: Associations with marital satisfaction, intimacy, and daily mood. Health Psychology, 17(2), 112-124. Raudenbush, S. W. (1994) Random effects models. In H. Cooper & L. V. Hedges (Eds.), Handbook of research synthesis. (pp. 301-321). New York, NY: Russell Sage Foundation.

PAGE 185

174 Raush, H. L., Barry, W. A., Hertel, R. K., & Swain, M. A. (1974). Communication, conflict, and marriage. San Francisco: Jossey-Bass. Reis, H. T., Collins, W. A., & Berscheid, E. (2000). The relationship context of human behavior and development. Psychological Bulletin, 126, 844-872. Reis, H. T., & Gable, S. L. (2000). Event-sampling and other methods for studying everyday experience. In H. T. Reis & C. M. Judd (Eds.), Handbook of research methods in social and personality psychology (pp. 190-222). New York: Cambridge University Press. Renne, K. S. (1970). Correlates of dissatisfaction in marriage. Journal of Marriage & Family, 54-67. Renne, K. S. (1974). Measurement of social health in a general population survey. Social Science Research, 3, 25-44. Renne, K. S. (1976). Childlessness, health, and marital satisfaction. Social Biology, 23, 183-197. Repetti, R. L. (1989). Effects of daily workload on subsequent behavior during marital interaction: The roles of social withdrawal and spouse support. Journal of Personality and Social Psychology, 57, 651-659. Repetti, R. L. (1994) Short-term and long-term processes linking job stressors to father-child interaction. Social Development, 3, 1-15. Revenson, T. A., & Majerovitz, S. D. (1991). The effects of chronic illness on the spouse: Social resources as stress buffers. Arthritis Care and Research, 4(2), 63-72. Ricer, R. E. (1983). Marital satisfaction among military and civilian family practice residents. Journal of Family Practice, 17(2), 303-307. Ridley, C. A., & Wilhelm, M. S. (1988). Adaptation to unemployment: Effects of a mine closure on husbands and wives. Lifestyles: Family and Economic Issues, 9(2), 145-160. Riggs, D. S., Byrne, C. A., Weathers, F. W., & Litz, B. T. (1998). The quality of the intimate relationships of male Vietnam veterans: Problems associated with posttraumatic stress disorder. Journal of Traumatic Stress, 11(1), 87-101. Rimmerman, A., & Stanger, V. (2001). Parental stress, marital satisfaction and responsiveness to children: a comparison between mothers of children with and without inborn impairment. International Journal of Rehabilitation Research, 24, 317-320.

PAGE 186

175 Risdal, D., & Singer, G. H. (2004). Marital adjustment in parents of children with disabilities: A historical review and meta-analysis. Research and Practice for Persons with Severe Disabilities, 29, 95-103. Roberto, K. A., Gold, D. T., & Yorgason, J. B. (2004). The influence of osteoporosis on the marital relationships of older couples. Journal of Applied Gerontology, 23(4), 443-456. Robertson, E. B., Elder, G. H., Skinner, M. L., & Conger, R. D. (1991). The costs and benefits of social support in families. Journal of Marriage & the Family, 53(2), 403-416. Rockwell, R. C., Elder, G. H., & Ross, D. J. (1979). Psychological patterns in marital timing and divorce. Social Psychology Quarterly, 42, 399-404. Roehling, P. V., & Bultman, M. (2002). Does absence make the heart grow fonder? Work-related travel and marital satisfaction. Sex Roles, 46, 279-293. Rogers, S. J. (1996). Mothers' work hours and marital quality: Variations by family structure and family size. Journal of Marriage & the Family, 58(3), 606-617. Rogers, S. J., & May, D. C. (2003). Spillover between marital quality and job satisfaction: Long-term patterns and gender differences. Journal of Marriage & Family, 65(2), 482-495. Rogosa, D., Brant, D., & Zimowski, M. (1982). A growth curve approach to the measurement of change. Psychological Bulletin, 92, 726-748. Rohrbaugh, M. J., Shoham, V., & Coyne, J. C. (2004). Beyond the “self” in self-efficacy: Spouse confidence predicts patient survival following heart failure. Journal of Family Psychology, 18(1), 184-193. Romano, J. M., Turner J. A., & Clancy, S. L. (1989). Sex differences in the relationship of pain patient dysfunction to spouse adjustment. Pain, 39, 289-295. Rosenthal, R., & Rubin, D. B. (1982). A simple, general purpose display of magnitude of experimental effects. Journal of Educational Psychology, 74, 166-169. Rusbult, C. E. (1980). Commitment and satisfaction in romantic associations: A test of the investment model. Journal of Experimental Social Psychology, 16(2), 172-186. Saarijarvi, S., Rytokoski, U., & Karppi, S.-L. (1990). Marital satisfaction and distress in chronic low-back pain patients and their spouses. The Clinical Journal of Pain, 6, 148-152. Sabatelli, R. M., Meth, R. L., & Gavazzi, S. M. (1988). Factors mediating the adjustment to involuntary childlessness. Family Relations, 37, 338-343.

PAGE 187

176 Sabourin, S., Lussier, Y., & Wright, J. (1991). The effects of measurement strategy on attributions for marital problems and behaviors. Journal of Applied Social Psychology, 21, 734-746. Sarason, I. G., Sarason, B. R., Brock, D. M., & Pierce, G. R. (1996). Social support: Current status, current issues. In Stress and emotion: Anxiety, anger, and curiosity, Vol. 16 (pp. 3-27). New York: Hemisphere Publishing Company. Scanzoni, J. (1975). Sex roles, economic factors, and marital solidarity in Black and White marriages. Journal of Marriage & the Family, 37(1), 130-144. Schaefer, J. A., & Moos, R. H. (1993a). Relationship, task and system stressors in the health care workplace. Journal of Community & Applied Social Psychology, 3, 285-298. Schaefer, J. A., & Moos, R. H. (1993b). Work stressors in health care: Context and outcomes. Journal of Community & Applied Social Psychology, 3, 235-242. Scheier, M. F., Weintraub, J. K., & Carver, C. S. (1986). Coping with stress: Divergent strategies of optimists and pessimists. Journal of Personality & Social Psychology, 51, 1257-1264. Schmidt, F. L. (1992). What do data really mean? Research findings, meta-analysis, and cumulative knowledge in psychology. American Psychologist, 47, 1173-1181. Schulz, M. S., Cowan, P. A., & Cowan, C. (2004). Coming home upset: Gender, marital satisfaction, and the daily spillover of workday experience into couple interactions. Journal of Family Psychology, 18(1), 250-263. Schumm, W. R., Bell, D. B., & Gade, P. A. (2000). Effects of a military overseas peacekeeping deployment on marital quality, satisfaction, and stability. Psychological Reports, 87(3, Pt 1), 815-821. Schumm, W. R., Bell, D. B., Knott, B., & Rice, R. E. (1996). The perceived effect of stressors on marital satisfaction among civilian wives of enlisted soldiers deployed to Somalia for Operation Restore Hope. Military Medicine, 161(10), 601-606. Schumm, W. R., & Bugaighis, M. A. (1985). Marital quality and marital stability: Resolving a controversy. Journal of Divorce, 9(1), 73-77. Schwab, R. (1992). Effects of a child's death on the marital relationship: A preliminary study. Death Studies, 16(2), 141-154. Shamai, M., & Lev, R. (1999). Marital quality among couples living under the threat of forced relocation: The case of families in the Golan Heights. Journal of Marital and Family Therapy, 25(2), 237-252.

PAGE 188

177 Shanfield, S. B., Benjamin, J. A., & Swain, B. J. (1988). The family under stress: The death of adult children. In O. S. Margolis and A. H. Kutscher (Eds.), Grief and loss of an adult child. (pp. 3-7). New York, NY: Praeger Publishers. Shapiro, E. G. (1978). Embarrassment and help-seeking. In B. M. DePaulo, A. Nadler, & J. D. Fisher (Eds.), New directions in helping, Vol. 2 (pp. 143-164). New York: Academic Press. Simpson, M. (1979). The facts of death. Englewood Cliffs: Prentice Hall. Skinner, E. A., Edge, K., Altman, J., & Sherwood, H. (2003). Searching for the structure of coping: a review and critique of category systems for classifying ways of coping. Psychological Bulletin, 129, 216-269. Small, S. A., & Riley, D. (1990). Toward a multidimensional assessment of work spillover into family life. Journal of Marriage & the Family, 52(1), 51-61. Solomon, Z., Waysman, M., Belkin, R., Levy, G., Mikulincer, M., & Enoch, D. (1992). Marital relations and combat stress reaction: The wives' perspective. Journal of Marriage and the Family, 54, 316-326. Sotile, W. M., & Sotile, M. O. (2004). Physicians’ wives evaluate their marriages, their husbands, and life in medicine: Results of the AMA-Alliance Medical Marriage Survey. Bulletin of the Menninger Clinic, 68(1), 39-59. South, S. J. (2001). The geographic context of divorce: Do neighborhoods matter? Journal of Marriage and the Family, 63, 755-766. Spangenberg, J. J., & Theron, J. C. (1999). Stress and coping strategies in spouses of depressed patients. Journal of Psychology, 133, 253-262. Spanier, G. B. (1976). Measuring dyadic adjustment: New scales for assessing the quality of marriage and similar dyads. Journal of Marriage and the Family, 38, 15-28. Spasojevic, J., Heffer, R. W., & Snyder, D. K. (2000). Effects of posttraumatic stress and acculturation on marital functioning in Bosnian refugee couples. Journal of Traumatic Stress, 13(2), 205-217. Sprecher, S., Metts, S., Burleson, B., Hatfield, E., & Thompson, A. (1995). Domains of expressive interaction in intimate relationships: Associations with satisfaction and commitment. Family Relations: Journal of Applied Family & Child Studies, 44, 203-210. Steffy, B. D., & Ashbaugh, D. (1986). Dual career planning, marital satisfaction, and job stress among women in dual-career marriages. Journal of Business and Psychology, (2), 114-123.

PAGE 189

178 Stern, M. J., & Pascale, L. Psychosocial adaptation post-myocardial infarction: The spouse's dilemma. Journal of Psychosomatic Research, 23(1), 83-87. Stith, S. M., Rosen, K. H., Middleton, K. A., Busch, A. L., Lundeberg, K., & Carlton, R. P. (2000). The intergenerational transmission of spouse abuse: A meta-analysis. Journal of Marriage and the Family, 62, 640-654. Strawbridge, W. J., Shema, S. J., & Roberts, R. E. (2004). Impact of spouses’ sleep problems on partners. Journal of Sleep and Sleep Disorders Research, 27(3), 527-531. Suchet, M., & Barling, J. (1986). Employed mothers: Interrole conflict, spouse support and marital functioning. Journal of Occupational Behaviour, 7(3), 167-178. Szinovacz, M. (1996). Couples' employment/retirement patterns and perceptions of marital quality. Research on Aging, 18(2), 243-268. Taylor, R. J. (1986). Receipt of support from family among Black Americans: Demographic and familial differences. Journal of Marriage and the Family, 48, 67-77. Taylor, S. E., & Aspinwall, L. G. (1996). Mediating and moderating processes in psychological stress. In H. B. Kaplan (Ed.) Psychosocial stress: Trends in theory and research (pp. 88-110). New York: Academic Press. Taylor, R. J., Jackson, J. S., & Chatters, L. M. (1997). Family life in black America. Thousand Oaks, CA: Sage. Taylor, S. E., Repetti, R. L., & Seeman, T. (1997). Health psychology: What is an unhealthy environment and how does it get under the skin? Annual Review of Psychology, 48, 411-447. Teachman, J. D., Tedrow, L. M., & Crowder, K. D. (2000). The changing demography of the family. Journal of Marriage and the Family, 62, 1234-1246. Terman, L. M. (1938). Psychological Factors in Marital Happiness. New York: McGraw-Hill.Terman, L. M. (1950). Prediction data: Predicting marriage failure from test scores. Marriage and Family Living, 12, 51-54. Thibaut, J. W., & Kelley, H. H. (1959). The Social Psychology of Groups. New York: Wiley. Thompson, A., & Bolger, N. (1999). Emotional transmission in couples under stress. Journal of Marriage and the Family, 61, 38-48. Thoresen, R. J., & Goldsmith, E. B. (1987). The relationship between army families' financial well-being and depression, general well-being, and marital satisfaction. Journal of Social Psychology, 127(5), 545-547.

PAGE 190

179 Udry, J. R. (1981). Marital alternatives and marital disruption. Journal of Marriage and the Family, 43, 889-897. Umberson, D. (1995). Marriage as support or strain? Marital quality following the death of a parent. Journal of Marriage and the Family, 57, 709-723. Urey, J. R., & Henggeler, S. W. (1987). Marital adjustment following spinal cord injury. Archives of Physical Medicine Rehabilitation, 68, 69-74. Vance, J. C., Boyle, F. M., & Najman, J. M., & Thearle. (2002). Couple distress after sudden infant or perinatal death: A 30-month follow up. Journal of Pediatrics and Child Health, 38, 368-382. Van Der Poel, A., & Greeff, A. P. (2003). The influence of coronary bypass graft surgery on the marital relationship and family functioning of the patient. Journal of Sex and Marital Therapy, 29(1), 61-77. Vangelisti, A. L., Reis, H. T., & Fitzpatrick, M. A. (2002). Stability and change in relationships. New York: Cambridge University Press. Vemer, E., Coleman, M., Ganong, L. H., & Cooper, H. (1989). Marital satisfaction in remarriage: A meta-analysis. Journal of Marriage and the Family, 51, 713-725. Vermulst, A. A., & Dubas, J. S. (1999). Job stress and family functioning; the mediating role of parental depression and the explanatory role of emotional stability. Gedrag & Gezondheid, 27, 96-102. Vinokur, A. D., Price, R. H., & Caplan, R. D. (1996). Hard times and hurtful partners: How financial strain affects depression and relationship satisfaction of unemployed persons and their spouses. Journal of Personality & Social Psychology, 71, 166-179. Vitaliano, P. P., Katon, W., Maiuro, R. D., & Russo, J. (1989). Coping in chest pain patients with and without psychiatric disorders. Journal of Consulting & Clinical Psychology, 57, 338-343. Voydanoff, P. (1983). Unemployment: Family strategies for adaptation. In C. R. Figley & H. I. McCubbin (Eds.), Stress and the family Vol.1 (pp. 91-102). New York: Brunner/Mazel. Voydanoff, P., & Donnelly, B. W. (1988). Economic distress, family coping, and quality of family life. Families and Economic Distress: Coping Strategies and Social Policy, 97-115. Wai-Ming, V. M. (2002). Psychological predictors of marital adjustment in breast cancer patients. Psychology, Health & Medicine, 7(1), 37-51.

PAGE 191

180 Waltz, M., Badura, B., Pfaff, H., & Schott, T. (1988). Marriage and the psychological consequences of a heart attack: A longitudinal study of adaptation to chronic illness after 3 years. Social Science and Medicine, 27, 149-158. Warde, C. M., Moonesinghe, K., Allen, W., & Gelberg, L. (1999). Marital and parental satisfaction of married physicians with children. Journal of General Internal Medicine, 14, 157-165. Watanabe, H. K., Jensen, P. S., Rosen, L. N., & Newby, J. (1995). Soldier functioning under chronic stress: Effects of family member illness. Military Medicine, 160(9), 457-461. Weihs, K., Enright, T., Howe, G., & Simmens, S. J. Marital satisfaction and emotional adjustment after breast cancer. Journal of Psychosocial Oncology, 17(1), 33-49. Weinberger, M., Hiner, S. L., & Tierney, W. M. (1987). In support of hassles as a measure of stress in predicting health outcomes. Journal of Behavioral Medicine, 10, 19-31. Weiss, R. L., & Heyman, R. E. (1990). Observation of marital interaction. In F. D. Fincham & T. N. Bradbury (Eds.), The Psychology of Marriage (pp. 87-117). New York: Guilford. Wertlieb, D., In, Gotlib, I. H., & Wheaton, B. (1997). Children whose parents divorce: Life trajectories and turning points. In I. H. Gotlib & B. Wheaton (Eds.), Stress and adversity over the life course: Trajectories and turning points (pp. 179-196). New York: Cambridge University Press. Westman, M., Vinokur, A. D., & Hamilton, V. L. (2004). Crossover of marital dissatisfaction during military downsizing among Russian army officers and their spouses. Journal of Applied Psychology, 89(5), 769-779. Wethington, E., & Kessler, R. C. (1986). Perceived support, received support, and adjustment to stressful life events. Journal of Health & Social Behavior, 27(1), 78-89. Wheaton, B. (1985). Models for the stress-buffering functions of coping resources. Journal of Health & Social Behavior, 26, 352-364. Wheaton, B. (1994). Sampling the stress universe. In W. R. Avison & I. H. Gotlib (Eds.), Stress and mental health: Contemporary issues and prospects for the future (pp. 77-114). New York: Plenum Press. Wheaton, B. (1997). The nature of chronic stress. In B. H. Gottlieb (Ed.), Coping with chronic stress (pp. 43-71). New York: Plenum Press.

PAGE 192

181 Wheaton, B. (1999). The nature of stressors. In A. V. Horwitz & T.L. Scheid (Eds.), A handbook for the study of mental health: Social contexts, theories, and systems (pp. 176-197). New York: Cambridge University Press. Whiffen, V. E., & Gotlib, I. H. (1989). Stress and coping in maritally distressed and nondistressed couples. Journal of Social & Personal Relationships, 6(3), 327-344. White, L. K. (1990). Determinants of divorce: A review of research in the 80's. Journal of Marriage and the Family, 52, 904-912. White, L. K., & Keith, B. (1990). The effect of shift work on the quality and stability of marital relations. Journal of Marriage and the Family, 52, 453-462. Wickrama, K. A. S., Lorenz, F. O., Conger, R. D., Matthews, L., & Elder, G. H. (1997). Linking occupational conditions to physical health through marital, social, and intrapersonal processes. Journal of Health & Social Behavior, 38(4), 363-375. Widrich, L., & Ortlepp, K. (1994). The mediating role of job satisfaction in the work stress and marital interaction relationship. South African Journal of Psychology, 24(3), 122-130. Wilhelm, M. S., & Ridley, C. A. (1988). Unemployment induced adaptations: Relationships among economic responses and individual and marital well-being. Lifestyles, 9(1), 5-20. Williams, L. (1995). The impact of stress on marital quality: A stress-vulnerability theory. Contemporary Family Therapy: An International Journal, 17(2), 217-227. Williams, S. E., & Freer, C. A. (1986). Aphasia: Its effect on marital relationships. Archives of Physical Medicine Rehabilitation, Psychological Reports, 67, 250-252. Willoughby, J. C., & Glidden, L. M. (1995). Fathers helping out: Shared child care and marital satisfaction of parents of children with disabilities. American Journal on Mental Retardation, 99, 399-406. Wilson, S. M., Larson, J. H., McCulloch, B. J., & Stone, K. L. (1997). Dyadic adjustment: An ecosystemic examination. American Journal of Family Therapy, 25, 291-306. Woelfel, J. C. (1979). The relationship between the army life and family life of soldiers. Public Data Use, 7(1), 20-29. Wolf, F. M. (1990). Methodological observations on bias. In K. W. Wachter & M. L. Straf. Future of meta-analysis (pp. 139-151). New York, NY: Russell Sage Foundation. Wright, L. K. (1991). The impact of Alzheimer's disease on the marital relationship. The Gerontologist, 31(2), 224-237.

PAGE 193

182 Yim, S. Y., Lee, I. Y., Lee, J. H., Kim, T. H., Bae, S. C., & Yoo, D. H. (2003). Quality of marital life in Korean patients with spondyloarthropathy. Clinical Rheumatology, 22(3), 208-212. Yim, S. Y., Lee, I. Y., Song, M. S., Rah, E. W., & Moon, H. W. (1998). Quality of marital life in Korean spinal cord injured patients. Spinal Cord, 36(12), 826-831. Yip, S.K., Chan, A., Pang, S., Leung, P., Tang, C., Shek, D., et al. (2003). The impact of urodynamic stress incontinence and detrusor overactivity on marital relationship and sexual function. American Journal of Obstetrics and Gynecology, 188(5), 1244-1248. Yogev, S. (1986). Relationships between stress and marital satisfaction among dual-earner couples. Women and Therapy, 5(2-3), 313-330. Zill, N., (1978) as cited in Zill, N. & Nord, C. W. (1994). Running in place: How American families are faring in a changing economy and an individualistic society. Washington, DC: Child Trends, Inc. Zvonkovic, A. M., Guss, T., & Ladd, L. (1988). Making the most of job loss: Individual and marital features of underemployment. Family Relations: Journal of Applied Family & Child Studies, 37(1), 56-61.

PAGE 194

BIOGRAPHICAL SKETCH Shauna Howarth Springer completed her bachelor’s degree in English and American literature at Harvard University in Cambridge, Massachusetts. She married Utaka Springer (Harvard class of 1999) and began her Ph.D. studies in counseling psychology at Iowa State University, in Ames, Iowa, under the mentorship of Dr. Lisa Larson. After the completion of her master’s degree, she and Utaka moved to Gainesville, Florida, where she completed her Ph.D. in counseling psychology, under the mentorship of Drs. Benjamin Karney and Gregory Neimeyer. Her area of interest has been the study of how external stress affects marital satisfaction, and the completion of her dissertation provided both academic knowledge and personal experience on this topic! 183