The influences of negative external stressors on cognitive processes within marital relationships

MISSING IMAGE

Material Information

Title:
The influences of negative external stressors on cognitive processes within marital relationships
Physical Description:
Book
Creator:
Neff, Lisa A
Publication Date:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 28824032
oclc - 50396820
System ID:
AA00020483:00001

Full Text










INFLUENCES OF NEGATIVE EXTERNAL STRESSORS ON COGNITIVE
PROCESSES WITHIN MARITAL RELATIONSHIPS














ACKNOWLEDGMENTS

I would like to thank the members of my committee, Benjamin Karney (chair),

Barry Schlenker, Dolores Albarracin, Greg Neimeyer and James Algina, for the time an

insights they have provided to this project.














TABLE OF CONTENTS
page

ACKNOWLEDGMENTS ....................................................................... ii

ABSTRACT............................................... V

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

LITERATURE REVIEW........................................................................ 5

Stress and Relationship Quality.............................................................. 5
Stress and the Individual: Spillover Effects........................................... 6
Stress and the Partner: Crossover Effects............................................ 8
Critique of the Literature on Stress and Relationship Well-Being................. 9
The Mechanisms Underlying Stress and Declines in Relationship Satisfaction..... 11
Stress and Cognitive Content......................................................... 13
Stress and Cognitive Structure...................................................... 15
Incorporating the Dyad: Additive and Interactive Effects of Intimates' Stress...... 18
Is Stress Always Bad? Stressful Life Events and Longitudinal Outcomes...........21
Overview of the Current Study............................................................. 24
Review of Hypotheses ...................................................................... 26

M ETH O D .................................................. ... ............ ....................... 29

-Dart,,mntntu ........---- 29















D ISCU SSION ............................................... ............................. ....... 74

Study Rationale and Summary of Results................................................. 74
Strengths and Limitations of the Study.................................................... 82
Marriage as a Safe Haven: The Successful Adaptation to Stress....................... 86
Two Routes to Change in Satisfaction: Expanding the Model.......................... 88
Directions for Future Research.............................................................. 89
Conclusions .................... ............................................................... 90

APPENDIX

A SEMANTIC DIFFERENTIAL MEASURE
OF MARITAL SATISFACTION.................................................... 92

B INVENTORY OF SPECIFIC MARITAL PROBLEMS.......................... 93

C INVENTORY OF SPECIFIC RELATIONSHIP STANDARDS............... 94

D RELATIONSHIP ATTRIBUTIONS MEASURE.................................. 98

E SURVEY OF LIFE EVENTS....................................................... 100

REFEREN CES............................................ ..................................... 106

BIOGRAPHICAL SKETCH.................................................................. 112














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

INFLUENCES OF NEGATIVE EXTERNAL STRESSORS ON COGNITIVE
PROCESSES WITHIN MARITAL RELATIONSHIPS

By

LISA A. NEFF

May 2002

Chairperson: Benjamin R. Karney, Assistant Professor
Major Department: Psychology

Traditionally, most research on relationship maintenance has focused on the effect

of intrapersonal factors on relationship outcomes. What this perspective overlooks,

however, is that part of maintaining a relationship involves navigating the negative

stressors external to the relationship that may nevertheless strain the relationship.

Consequently, the effects of intrapersonal factors on relationship outcomes may not be

able to be full understood without reference to the stressful circumstances surrounding















specific relationship perceptions in a less relationship-enhancing manner. Thus, external

stress seems to lead to declines in satisfaction not only by providing spouses with more

negative perceptions of the marriage, but also by affecting spouses' ability to

subsequently cope with this increase in negative perceptions. Support for the prediction

that each spouse's stress would interact to affect marital satisfaction was not found.

When controlling for spouses' own stress, the stress experienced by the partner did not

have an additional influence on spouses' satisfaction. Finally, some evidence suggested

that spouses may become resilient against the effects of stress. Successfully coping with

low stress was associated with less vulnerability to future stress. However, under high

stress, a positive coping strategy did not protect spouses from the adverse affects of later

stressors. Overall, the current findings suggest that a clear understanding of relationship

maintenance and deterioration is limited without taking into account the broader

circumstances surrounding the marriage to which couples must adapt.

















vi














INTRODUCTION

Marriages tend to begin happily, with both spouses expressing highly positive

evaluations of each other and the relationship. Despite this early optimism, however,

marriages today are more likely to end in separation or divorce than to continue

(Bumpass, 1990). Thus, for many people, the course of a marriage is characterized by a

shift in relationship beliefs, such that initially positive relationship beliefs deteriorate and

transform into negative beliefs. How does this shift occur? In other words, how is it that

some couples are able to maintain their initial feelings of satisfaction despite the

challenges of a long-term relationship, whereas other couples are not?

Theories of close relationships indicate that the experience of change or stability

in relationship satisfaction is, at its heart, a cognitive phenomenon. Though the

development of a relationship is affected by a broad range of variables, from the enduring

characteristics of each partner to the behaviors that partners exchange (Karney &

Bradbury, 1995), these variables nevertheless tend to exert their influence on future

outcomes through their effects on how individuals think about the relationship (Bradbury

& Fincham, 1991; Karney, McNulty, & Frye, 2001). In other words, changes in

intimates' global attitudes toward their relationship ultimately should stem from changes

in the specific relationship beliefs that give rise to the overall relationship evaluation.

Consequently, research on relationship maintenance and deterioration has focused a great

deal of attention on cognitive processes within relationships. This research has

demonstrated reliable links among particular cognitions (i.e., expectations, perceptions









of a partner, relationship memories) and relationship outcomes (Karney & Coombs,

2000; Knee, 1998; Murray, Holmes, & Griffin, 1996). Moreover, this literature has

drawn attention to the way intimates integrate different levels of cognition to construct a

global evaluation of the relationship. The organization of intimates' specific relationship

beliefs has been shown to affect relationship quality independent of the content of those

beliefs (Murray & Holmes, 1999; Neff & Karney, unpublished manuscript; Showers &

Kevlyn, 1999). From this perspective, then, progress in understanding change and

stability in relationship quality necessitates an accurate assessment of the content of

intimates' relationship beliefs as well as the development of models of how intimates

integrate those beliefs within an overall impression of the relationship.

This detailed attention to cognitive processes within relationships, however,

draws attention away from the external context in which the relationship is embedded.

Nevertheless, many aspects of the broader environmental context surrounding the

marriage are likely to influence relationship quality and stability. In particular, the

effects of intrapersonal factors on marriage may not be able to be fully understood

without reference to the stressful circumstances and events to which couples must adapt

(Karney & Bradbury, 1995). Research on stressful events, traditionally defined as events

that challenge an individual's adaptive capacity (Cohen, Kessler, & Gordon, 1997),

suggests that even the happiest couples will likely experience stressful events external to the

relationship that may strain the marriage, despite the initial absence of difficulties within the

relationship (Robinson & Jacobson, 1987). Theoretically, these stressful events may include

any event that requires adaptation by the individual, regardless of whether the event is

positive or negative. Empirical evidence, however, suggests that the adaptation to negative
J .I... A- -- -- *-- -.-. I --n- -IIT --L. --. ----.. -- -








been found to have clear, consistent patterns with psychological distress, tnldmgs with

respect to positive events tend to be weak and contradictory (Turner & Wheaton, 1997). In

other words, adapting to winning the lottery seems to be less of a challenge than adapting to

the loss of a job. For this reason, research examining the influence of stress on marital quality

has tended to focus on the impact of negative life events.

In fact, negative life events consistently have been associated with lowered marital

adjustment (Lavee, McCubbin, & Olson, 1987; Whiffen & Gottlib, 1989). Marital instability

tends to be higher among couples experiencing negative external stresses, such as financial

difficulties (Bahr, 1979). Moreover, couples reporting more negative life events reap fewer

long-term benefits from marital therapy than those not faced with such challenging

circumstances (Jacobson, Schmaling, & Holtzworth-Munroe, 1987). Consequently, some of

the antecedents of marital deterioration may stem from contextual, rather than intrapersonal

influences. Yet, little is known about the mechanisms through which the external

circumstances surrounding a marriage influence relationship outcomes over time

(Bradbury, Cohan, & Karney, 1998; Karney & Bradbury, 1995). In other words, how do














perceptions within the relationship. Furthermore, stress should prevent intimates from

successfully coping with this increase in negative relationship content by hindering

intimates' ability to re-organize their perceptions in a manner that would preserve their

overall relationship satisfaction. Next, we take these ideas out of the context of the

individual and place them within the relationship dyad by examining how the stress of

each individual in the relationship may combine to affect intimates' cognitive processes.

From here, we explore the possibility that stress may not always result in declines in

satisfaction. Though most stress research has emphasized the harmful effects of stress,

under some circumstances stress actually may serve to enhance rather than deteriorate

relationship functioning. Finally, we describe a study designed to evaluate these ideas by

examining the longitudinal effects of negative, external stressful events on intimates'

relationship cognitions.














LITERATURE REVIEW

Stress and Relationship Quality

The general literature on stress frequently has examined the association between

an individual's stress level and his or her personal well-being. Thus, this literature has

sought to answer the question: How does exposure to stress influence an individual's

thoughts and behaviors? Within the context of a romantic relationship, however,

intimates' thoughts and behaviors tend to affect the thoughts and behaviors of their

relationship partners. As a result, the external stressors experienced by one individual are

likely to create circumstances that will influence each individual within the relationship

(O'Brien & DeLongis, 1997). Considering the perspective of both the individual and the

relationship partner thus broadens the original question. Within a relationship, the

question becomes: How does stress influence the relationship functioning of the

individual and how does the stress experienced by one individual influence the

relationship functioning of the spouse? As a result of this broader question, theory and

research geared toward explaining the association between stress and relationship well-

being has grown to include the perspective of both individuals in a relationship through

the study of two distinct, yet related phenomena. The first phenomenon, known as stress

spillover, refers to a situation in which the stress generated in one setting affects the

thoughts and behaviors of an individual within a different setting (Bolger, DeLongis,

Kessler, & Wethington, 1989; Repetti & Wood, 1997). Thus, stress spillover suggests

that individuals' external stressful experiences may affect their own judgments of








their relationships. The second phenomenon, known as stress crossover or emotional

transmission, refers to a situation in which the stress being experienced by one individual

leads to heightened distress in the partner (Larson & Almeida, 1999). Stress crossover,

then, suggests that individuals' external stressful experiences may affect their partners'

judgments of the relationship. The next section addresses each of these processes in

greater detail.

Stress and the Individual: Spillover Effects

Two lines of research suggest that stressful life events external to the relationship

may spill over to affect an individual's functioning within the relationship. First,

increases in stress consistently have been associated with changes in relationship

behaviors. For instance, a common response to job stress appears to be social withdrawal

(Repetti & Wood, 1997). A study of male air traffic controllers' daily work stress

revealed that both the air traffic controllers and their wives described the husbands'

behavior as more withdrawn after work shifts that husbands described as busier and more

difficult than after work shifts that were relatively stress free (Repetti, 1989).

Specifically, under higher levels of work stress, individuals tend to reduce their

involvement at home by engaging in fewer household tasks and fewer leisure activities

(Bolger et al., 1989; Crouter, Perry-Jenkins, Huston, & Crawford, 1989). This

withdrawal coping response, however, tends only to occur when spouses support this

behavior (Repetti, 1989). When withdrawal is not facilitated, the experience of stress in

the workplace may give rise to negative interactions at home. Husbands in blue-collar

occupations display more negative affect in their marital interactions than do husbands in

white-collar occupations (Krokoff, Gottman, & Roy, 1988). This association was

nonsignificant when controlling for job distress, suggesting that stressful working








conditions rather than status intimuencea inese oenaviors. in auiuoun, igUliumiis III uL

home are more likely to be reported on days in which individuals report more distressing

encounters with coworkers and supervisors (Bolger et al., 1989; Repetti & Wood, 1997).

Second, increases in stress have been directly associated with diminished

relationship evaluations. Individuals' work stress has been linked to less accepting views

of family members (Crouter & Bumpus, 2001). Moreover, the accumulation of external

stressors negatively affects not only concurrent but also future judgments of relationship

satisfaction (Bodenmann, 1997). However, recent evidence suggests a more complex

association between stressful life events and relationship satisfaction. In a series of

studies, Tesser and Beach (1998) examined the between-subjects association between

negative life events and relationship satisfaction. Results revealed that as the number of

negative life events experienced increased from low to moderate, stress spillover

occurred, such that individuals also reported increasingly negative perceptions of their

relationship. Interestingly, however, under moderate levels of stress, negative life events

did not seem to reflect on individuals' relationships. Rather, under moderate stress,

individuals seemed to prevent their stressful experiences from spilling over into their

relationship judgments, resulting in a jump in relationship satisfaction. As negative life

events continued to accumulate beyond this point, though, increases in stress again were

associated with lowered relationship satisfaction. To explain these findings, Tesser and

Beach (1998) postulate that under conditions of moderate stress, individuals may become

aware of the possibility that their stress is contaminating their relationship judgments. As

a result of this awareness, individuals will attempt to discount their feelings of stress

when making their relationship evaluations, thereby minimizing stress spillover. This









levels are high, individuals' stress may tax their cognitive resources, thereby

overwhelming their ability to separate their stress from their satisfaction (e.g., Martin,

Seta, & Crelia, 1990). As a result, increased stress spillover occurs. In other words,

when individuals are aware of their stress and possess the cognitive capacity to do so,

they may be able to nullify the influence of their stress on their relationship evaluations.

Stress and the Partner: Crossover Effects

In addition to affecting one's own thoughts and behaviors, the stressful life events

of one individual may lead to changes in the emotions and behaviors of significant others

(Larson & Almeida, 1999). For instance, in one study, independent observers rated

children's affect as increasingly dysphoric on days in which their mothers reported

experiencing higher levels of work stress (Repetti & Wood, 1997). Similarly, mothers'

anxiety predicts the subsequent anxiety of their adolescent children even when

controlling for adolescents' initial levels of anxiety. This transmission of anxiety from

mother to child was particularly likely when mothers were under higher levels of stress

(Larson & Gillman, 1999).

As seen in stress spillover research, however, in some situations individuals may

succeed in limiting the negative influences of stress. Downey and colleagues (Downey,

Purdie, & Schaffer-Neitz, 1999) found that mothers suffering from chronic pain report

experiencing higher distress and more anger than a sample of control mothers without

chronic pain. However, the within-day correspondence between mother's anger and

child's anger was significantly lower in families coping with chronic pain, relative to the

control families. In other words, the anger experienced by mothers with chronic pain was

less likely to affect the anger of their children than was the anger experienced by the








control mothers, even though these control mothers reported lower overall levels of

anger.

Similar findings were found in a study of romantic relationships in which one

partner was preparing for the bar exam. Thompson and Bolger (1999) measured the

examinees' mood as well as their partners' feelings about the relationship for the 35 days

preceding and immediately after the exam. Prior to the exam, the examinee's depressed

mood led to a subsequent reduction in positive feelings about the relationship in the

partner. However, as the day of the exam drew closer and examinees' distress reached its

highest level, this association no longer remained significant. In fact, on the day

immediately before the exam, the association between examinees' mood and partners'

relationship evaluations was essentially zero. Thus, partners seemingly made allowances

for the examinee's distress by tolerating negative emotions they did not previously

tolerate. Together with the previous study, these findings imply that when an individual

clearly may attribute the source of a partner's distress to the stressful situation, the

individual may not react as strongly to the partner's distress as he or she otherwise would,

thereby preventing stress crossover effects (Thompson & Bolger, 1999).

Critique of the Literature on Stress and Relationship Well-Being

Together, research on stress spillover and stress crossover argues that stressful life

circumstances external to the relationship frequently affect couples' thoughts and

behaviors within the relationship. Nevertheless, our understanding of the complex

interplay between external events and internal relationship processes is hindered by three

major limitations of this research. First, the current literature is limited by its focus on

describing rather than explaining the spillover/crossover phenomena. Most of the stress

literature simply delineates when spillover or crossover occurs without regard for how it








occurs. An emerging theme from the stress literature is that when the source of a

negative emotion is attributable to a justified cause and individuals possess the cognitive

capacity to do so, individuals may think and act in ways that prevent stress from

influencing their relationship evaluations. How, then, do individuals successfully protect

their relationship evaluations from the influence of external stress? Moreover, what

happens when stress overwhelms this ability to protect relationship satisfaction? In other

words, how does relationship satisfaction ultimately break down in the face of stress?

The answers to these questions have yet to be examined directly. Consequently, there is

a gap between theory and research preventing a more thorough understanding of the

mechanisms underlying stress spillover and crossover processes.

A second limitation concerns a failure to examine how the external circumstances

of the relationship partner may influence stress crossover processes. By definition,

relationships involve dyadic processes. Thus, the factors each individual brings to a

relationship should combine to influence relationship outcomes. For instance, as

previously mentioned, spouses may not be affected by their partners' stress if they are

able to attribute their partners' behavior to the stressor (Thompson & Bolger, 1999).

However, if spouses are experiencing high levels of stress themselves, they may lack the

cognitive resources to discount their partners' distress in this manner (Tesser & Beach,

1998). Thus, the manner in which spouses' stress levels interact to affect each

individual's relationship well-being warrants further attention.

Though also an important strength of the current research, a final limitation of the

stress spillover and crossover literature is the almost exclusive reliance on daily diary

methods. Daily diaries have advanced the understanding of spillover/crossover processes

by allowing for the within-subjects examination of the association between stress and








well-being. This methodology controls for numerous extraneous variables, such as

personality or general response tendencies, by examining changes in an individual's

relationship functioning according to whether the person is experiencing more or less

stress than usual. However, as daily diary data is difficult to obtain, current research has

examined only the short-term longitudinal consequences (e.g., several days to one month)

of stress for relationship well-being. Given that systematic changes in satisfaction occur

over years rather than days, a broader perspective is necessary to investigate how external

stress may be linked to the deterioration of relationship well-being over the course of a

long-term marriage.

Overall, the existing stress research leaves three important unanswered questions

concerning the interplay between external stress and relationship well-being. First,

through what mechanism does stress affect relationship satisfaction? Second, how do

spouses' stress experiences interact to produce relationship outcomes? Finally, how is

stress linked to satisfaction over time? The remainder of the introduction examines each

of these three questions in greater detail.

The Mechanisms Underlying Stress and Declines in Relationship Satisfaction

The first limitation of the stress spillover/crossover literature is a lack of

understanding concerning how stressful life events produce changes in intimates'

satisfaction. Traditionally, research on stress and relationship well-being has been

conducted without regard for the existing theories on relationship development. Linking

stress research to the broader literature on relationship maintenance and deterioration

suggests a mechanism through which stressors external to the relationship may influence

judgments of satisfaction within the relationship. As mentioned, this literature argues that







12

intimates' cognitive processes within the relationship. In particular, relationship

satisfaction appears to be shaped both by what intimates believe about their relationships

as well as by how those relationship perceptions are integrated within an overall

representation of the relationship (Karney, McNulty, & Frye, 2001). Clearly, possessing

a large number of positive relationship beliefs is associated with higher relationship

satisfaction (Murray, Holmes, & Griffin, 1996). Nevertheless, all couples tend to

acknowledge some specific problems or disappointments in their relationships (McNulty

& Karney, 2001). Maintaining satisfaction over the course of a continuing relationship,

then, requires that intimates resolve their positive global evaluation of the relationship

with the negative specific beliefs and experiences that inevitably arise. The difference

between satisfaction that endures and satisfaction that declines may lie in the different

ways that this process of reconciliation can take place (Murray & Holmes, 1999; Neff &

Karney, unpublished manuscript; Showers & Kevlyn, 1999). When a specific relationship

perception is positive, linking that perception to the global evaluation of the relationship will

likely promote satisfaction. However, when a specific perception is negative, linking that

perception to the global evaluation will likely result in a deterioration of relationship

satisfaction. In other words, any cognitive organization that serves to separate specific

negative perceptions from the broader positive view of the relationship enhances relationship

outcomes.

This perspective on relationship development implies there may be two general

routes to declines in satisfaction. The first route involves a change in intimates' cognitive

content. As the number of negative relationship perceptions increases, structurally re-

organizing those beliefs may no longer protect judgments of satisfaction from the

implications of the beliefs, leading to lower satisfaction. The second route involves a








change in intimates' cognitive structure. Intimates may experience a deterioration of the

ability to re-organize their specific relationship perceptions in a manner that would serve

to enhance their relationship satisfaction.

The current study argues that stress influences each of these paths of relationship

decline. Stressful life circumstances should affect the content of intimates' relationship

beliefs by increasing the number of negative perceptions intimates hold about their

relationship. Perhaps more importantly, however, stressful life circumstances should

affect the structure of intimates' beliefs by hindering intimates' ability to organize their

beliefs in a relationship-enhancing fashion. In other words, the experience of external

stressors should provide individuals with more negativity to deal with in the relationship

as well as affect individuals' ability to subsequently cope with this increase in negative

relationship perceptions. The following sections describe hypotheses derived from this

general framework.

Stress and Cognitive Content

The first premise of this study was that increases in stressful life events would

negatively affect the content of intimates' specific relationship perceptions. Frequently,

satisfied intimates are rather enhancing in their overall impressions of their relationships,

even describing their relationships as ideal (Ruvulo & Veroff, 1997). This positive bias,

however, does not seem to extend to intimates' perceptions of specific aspects of the

relationship (Neff & Karney, in press). Unlike global impressions, which allow intimates

to choose from a wide range of specific examples to justify a positive self-view, specific

aspects of the relationship tend to be defined more concretely, and thus restrict intimates'

flexibility to justify a desired belief (Dunning, Meyerowitz, & Holzberg, 1989).








individual's experiences in a way that gloDai impressions are not. ror msianuce, ui

communications within the relationship become increasingly abrupt and critical,

intimates will likely begin to hold the negative specific belief that their communication

skills are suffering. Evidence from a daily diary study of satisfied newlywed couples

supports this idea (McNulty & Kamrney, 2001). This study revealed that spouses' specific

perceptions of the marriage tended to be less positive and more likely to fluctuate from

day to day than spouses' global perceptions, suggesting that changes in daily experiences

may be associated with changes in specific relationship beliefs. Thus, an increase in

negative experiences within the relationship should lead to an accumulation of negative

relationship perceptions, thereby resulting in the eventual deterioration of relationship

satisfaction.

The first goal of the study, then, was to examine the role of cognitive content in

the stress spillover process. Our first hypothesis was that stress should contribute to the

deterioration of satisfaction through its effects on spouses' specific relationship

perceptions. Replicating previous work on stress spillover, we first predicted that

increases in spouses' external stressful life circumstances should be associated with

corresponding decreases in their marital satisfaction over time (Hypothesis la).

Second, we predicted that specific relationship perceptions would mediate this

expected stress spillover effect (Hypothesis lb). Namely, given that stress should

provide spouses with increased negative experiences within the relationship, increases in

spouses' external stressful life circumstances should be accompanied by corresponding

increases in the negativity of spouses' specific relationship perceptions. This association

was expected to account for the association between stress and global satisfaction.







15

Stress and Cognitive Structure

The second premise of the study was that stressful life events would influence the

manner in which intimates organize and integrate their specific relationship perceptions.

Growing evidence suggests that intimates who evaluate their relationships positively or

negatively at the global level may not differ in the content of their specific relationship

perceptions, but rather in the way those perceptions are integrated. This research

demonstrates that the ability to organize relationship cognitions in a manner that limits

the influence of specific negative beliefs on the global relationship evaluation, allows

satisfaction to remain high, despite the presence of these negative perceptions (Murray &

Holmes, 1999; Neff& Kamrney, unpublished manuscript; Showers & Kevlyn, 1999).

Intimates may reduce the impact of their negative specific perceptions on

relationship satisfaction using a variety of organizational techniques. For instance,

intimates may attribute great importance to their positive specific perceptions, while

dismissing the importance of their negative specific perceptions. This process of

differentially weighing positive and negative beliefs, or differential importance, ensures

that positive beliefs will contribute more to overall satisfaction than negative beliefs

(Neff& Karney, unpublished manuscript; Pelham & Swann, 1989). Similarly, intimates

may integrate specific beliefs with a global evaluation through the use of causal

attributions. Attributing a partner's relationship transgressions to temporary or external

causes in effect weakens the link between this specific negative perception and global

relationship satisfaction (Holzworth-Munroe & Jacobson, 1989). In other words, any

organization that serves to weaken the link between the global relationship evaluation and

specific negative perceptions contributes to the maintenance of relationship quality.








In fact, individuals' cognitive organization may represent a strategic response to

negative experiences. Research on the self-concept shows that though individuals report

an increase in negative self-views when in a negative mood, they will attempt to

counteract this negativity by altering the structure of their self-views (Showers,

Abramson, & Hogan, 1998). Similarly, a longitudinal study of differential importance in

marriage demonstrated that intimates' specific relationship perceptions were changing

significantly over the course of the marriage. Among satisfied spouses, however, those

fluctuations in specific relationship perceptions were accompanied by corresponding

changes in the importance of those perceptions. Specifically, spouses who maintained

flexible cognitive structures, such that positive perceptions were always viewed as more

important than negative perceptions despite any changes in the content of those

perceptions, exhibited more stable levels of satisfaction over the first 2 1/2 years of

marriage (Neff & Karney, unpublished manuscript). Thus, whereas changes in the

content of intimates' beliefs may reflect intimates' objective experience within the

relationship, changes in cognitive structure may signify intimates' attempt to cope with

those changing experiences.

This coping response to negative perceptions within the relationship, however,

may be adversely affected by increases in the stressful life circumstances outside the

relationship. Previous research has shown that cognitive organization may buffer

individual's well-being from the negative effects of stress (Linville, 1987; Showers &

Kling, 1996). Nevertheless, the experience of a number of stressful events also may tax

intimates' energy and cognitive resources, thereby leaving intimates with fewer resources

to handle successfully their negative relationship beliefs (e.g., McCubbin & Patterson,

1983). As a result, intimates' information processing within the relationship may be







17
simplified when they are distracted by external stress (Hammond, 2000). Evidence

suggests that if individuals are devoting energy to other tasks while making evaluations,

they are less likely to partial out extraneous influences from those judgments (Martin et

al., 1990). For instance, individuals primed with irrelevant negative information will

judge a target other more negatively if they are distracted during the rating process than if

they are not distracted (Martin et al., 1990). Distracted individuals also tend not to correct

for situational influences when judging the behavior of others, instead relying on

dispositional attributions of behavior (Gilbert, Pelham, & Krull, 1988). Thus, to the

extent that maintaining a relationship-enhancing cognitive organization requires cognitive

effort, individuals may find it difficult to separate their specific negative perceptions from

their global relationship satisfaction while also attempting to manage high levels of

external stress.

The second goal of the study, then, was to examine the role of cognitive structure

in the stress spillover process. Our second hypothesis was that stress should contribute to

the deterioration of satisfaction through its effects on the organization of spouses'

specific relationship perceptions. As previously discussed, increases in stress should be

accompanied by increases in negative relationship perceptions. However, under low to

moderate levels of external stress, intimates may retain the resources necessary to cope

successfully with these negative relationship perceptions. In other words, low to

moderate stress should be associated with a relationship-enhancing cognitive structure,

such that positive perceptions are more closely linked to the global evaluation than are

negative perceptions. In fact, prior research demonstrates that cognitive organization

effects are strongest when the content of beliefs is somewhat negative (Showers et al.,

1998). Thus, the relationship-enhancing nature of intimates' cognitive organization








actually may increase as stress increases from low to moderate and intimates have more

negativity to cope with in the relationship. Under these conditions, then, stress spillover

should be low. On the contrary, as the number of external life stressors continues to grow

larger, intimates may find themselves overwhelmed by their stress. Conditions of high

stress may deplete intimates' cognitive resources and thus interfere with intimates' ability

to organize their relationship perceptions in a relationship-enhancing manner.

Consequently, as stress increases from moderate to high, the relationship-enhancing

nature of spouses' cognitive organization may decrease, resulting in greater stress

spillover.

Consequently, we predicted that fluctuations in spouses' stress would be

associated with fluctuations in their cognitive organization over time, independent of

changes in cognitive content (Hypothesis 2a). Specifically, external stress was expected

to be curvilinearly related to spouses' cognitive organization. Furthermore, we predicted

that spouses' organization of their specific relationship perceptions also would mediate

the association between stress and overall satisfaction (Hypothesis 2b).

Incorporating the Dyad: Additive and Interactive Effects of Intimates' Stress

A second limitation of the stress spillover/crossover literature is the lack of

attention given to the possible additive and interactive effects of each spouse's stress on

relationship evaluations. Previous research has argued that spouses' external stress not

only may affect their own relationship judgments, but also may affect the relationship

judgments of their partners. However, this research has examined the stress crossover

effect without regard to the stressful circumstances surrounding the relationship partner.

Nevertheless, the understanding of stress crossover effects is likely to be complicated by







19

in partners' satisfaction. In other words, we predict that intimates' stress will affect the

way their partners think about the relationship. Yet, the influence of intimates' stress on

their partners' relationship cognitions should depend on the amount of stress that partners

are experiencing themselves.

For instance, an increase in a spouse's stress should be associated with an increase

in the number of negative relationship perceptions held by the partner. However, the

partner's own stress level should moderate this effect. Namely, if John is under high

stress and thus likely to behave poorly in the relationship, Jane should experience an

increase in her negative relationship perceptions. If Jane is also under high levels of

stress, though, she may be more likely to reciprocate this negativity. Thus, both John's

stress and Jane's stress should work to contribute to Jane's negative relationship

perceptions.

Similarly, partners' own stress level should influence how they cope with the

increase in negative relationship perceptions brought about by their spouses' stress.

When intimates' stress is high, yet their partners' stress is low, partners should have the

cognitive resources at their disposal to make allowances for intimates' negativity. For

instance, if Jane is experiencing low stress, she should be able to cope successfully with

the increase in negative perceptions caused by John's stress by separating those

perceptions from her satisfaction. As a result, when intimates' stress is high and partners'

stress is low, stress crossover should be low. However, if partners are also faced with

high levels of stress, they may find themselves unable to accomplish the cognitive

reorganization needed to preserve their satisfaction. In other words, if Jane is also under

high stress, she may lack the resources necessary to maintain a relationship-enhancing








cognitive organization. Thus, when intimates' stress is high and their partners' stress is

high, stress crossover should also be high.

Importantly, however, intimates' stress levels should not directly affect their

partners' cognitive organization. Namely, intimates' stress should have little influence

on the cognitive resources partners have to re-organize their specific relationship

perceptions. Thus, unlike the influence of stress on cognitive content, intimates' stress

should not interact with their partners' stress to produce changes in partners' cognitive

organization. Rather, partners' own stress should influence their own ability to cope with

negative perceptions in the relationship, which in turn should moderate the effect of their

spouses' stress on their relationship satisfaction.

The third goal of the study, then, was to examine the role of the partner's own

stress in the stress crossover process. Specifically, our third hypothesis was that partners'

own stressful experiences would moderate the association between spouses' stress level

and partners' relationship satisfaction. Replicating previous work on stress crossover, we

first predicted that increases in spouses' external stressful life circumstances would be

associated with corresponding decreases in their partners' marital satisfaction over time

(Hypothesis 3a). In addition, we predicted that partners' own stress should moderate this

stress crossover effect, such that when partners' own stress is low, stress crossover should

be low. However, when partners are also experiencing high levels of external stress,

stress crossover should be high (Hypothesis 3b).

Parallel to Hypothesis Ib, we also predicted that partners' specific relationship

perceptions would mediate the expected stress crossover process (Hypothesis 3c). In

other words, spouses' stress should contribute to the deterioration of their partners'

satisfaction through its effects on partners' specific relationship perceptions. Moreover,








we predicted that the partner's own stress should moderate the association between

partners' specific relationship perceptions and their spouses' stressful life circumstances

(Hypothesis 3d). In other words, the interaction of each individual's stress should

produce even greater increases in the number of negative perceptions held by the

relationship partner.

Turning to the role of cognitive organization in the stress crossover process, we

hypothesized that changes in spouses' stressful life circumstances would not be directly

associated with changes in their partners' cognitive structures (Hypothesis 3e). Namely,

spouses' stress should have no direct effects on their partners' ability to cope with

negative specific perceptions within the relationship. Rather, as hypothesized earlier,

partners' own stress level should influence their cognitive organization (see Hypothesis

2b). Consequently, partners' own stress was expected to influence the stress crossover

process through its effects on partners' cognitive organization. Specifically, partners'

cognitive organization should moderate the association between spouses' stressful life

circumstances and partners' relationship satisfaction (Hypothesis 3f). In other words,

stress crossover should be highest when partners' cognitive organization is more

negative.

Is Stress Always Bad? Stressful life Events and Longitudinal Outcomes

A third limitation of the stress spillover/stress crossover literature is a lack of

longitudinal research investigating the effects of stress spillover and crossover over the

course of a long-term relationship. Daily diary studies indicate that the short-term effects

of stress on relationship functioning tend to be negative. These findings, then, imply that

stress will be detrimental to relationship quality over time. However, is this always the

case? This perspective on stress and relationship well-being fails to account for why








some relationships may emerge from stressful experiences relatively unscathed, while

other relationships crumble in the face of such hardships. Longitudinal research is

necessary to determine the ultimate effects of stressful life events on future relationship

outcomes.

In fact, some theories of stress have begun to shift away from an emphasis on the

harmful effects of stress toward a focus on the potential of stress to actually enhance

individuals' well-being. Stressful life events can provide opportunities for growth by

promoting new coping skills or mobilizing previously untapped personal and social

resources (Holahan & Moos, 1990; McCubbin & Patterson, 1983). As a result,

individuals who are exposed to stressful experiences and cope with them effectively may

develop a resilience to future stress (Holahan & Moos, 1990). In other words, successful

adaptation to a stressful event should stimulate positive changes and contribute to

improved functioning after the stressor by strengthening the individual's coping

resources, making the successful adaptation to future stressors more likely (McCubbin &

Patterson, 1983).

Empirical evidence for the positive effects of stressful experiences is growing.

Studies of daily stress have revealed that individuals report a more positive mood on the

day following a stressful event than on other stress-free days (Bolger, DeLongis, Kessler,

& Schilling, 1989; DeLongis, Folkman & Lazarus, 1988). This effect is particularly

strong when individuals receive high levels of social support for the event (Caspi, Bolger,

& Eckenrode, 1987). Moreover, individuals who behave adaptively under conditions of

high stress have shown an increase in resources, such as improved family support and

reduced family conflict, one year later (Holahan & Moos, 1990). Together, this research

suggests that successfully coping with a stressor should lead to more success in







23

surmounting stressful experiences in the future. Nevertheless, this assumption of the

stress resilience literature has not been studied directly.

The fourth and final goal of the study, then, was to examine whether intimates'

responses to stressful life events encountered early in the relationship affect their future

relationship functioning. Our fourth hypothesis was that intimates who successfully cope

with stress by preventing their stressful life experiences from spilling over into their

relationship satisfaction may emerge from the experience as less susceptible to the

adverse effects of later stressors. Specifically, we first predicted that spouses who

maintain a positive organization of relationship perceptions in the face of external stress

would be less vulnerable to future declines in relationship satisfaction (Hypothesis 4a).

In other words, spouses who experience stress early in the relationship and respond

effectively to that stress should maintain more stable levels of satisfaction over time.

Moreover, we predicted that spouses who maintain a positive organization of relationship

perceptions in the face of external stress early in the relationship would be less vulnerable

to future stress spillover effects (Hypothesis 4b). In other words, spouses who experience

stress early in the relationship and respond effectively to that stress should demonstrate

lower levels of stress spillover over time.

We also predicted that successful coping early in the relationship should serve to

bolster intimates' ability to successfully cope with stress in the future (Hypothesis 4c).

Intimates who counteract the influence of stress by separating their negative relationship

perceptions from their overall satisfaction should emerge from the stressful experience

with a positive cognitive organization. As a result, the experience of stress will serve to

augment intimates' future coping resources, thereby enhancing their probability of

surmounting future stressors. By the same token, intimates who fail to successfully cope







24
with their negative specific relationship perceptions in the face of stress will emerge with

a weakened cognitive organization. In this case, the experience of stress will serve to

deteriorate intimates' coping resources, rendering them unlikely to adapt successfully to

other stressors. Thus, successful coping should be associated with better coping in the

future.

Overview of Current Study

The current study attempted to address the limitations of the existing literature on

external stress and relationship evaluations by examining the influence of stressful life

events on cognitive processing within the relationship over the course of a continuing

marriage. First-married, newlywed couples participating in a broader study of marital

development provided information concerning their stressful experiences, their specific

relationship perceptions, and their overall relationship satisfaction every six months over

the first 3 /z years of their marriage. To ensure that the results of the study were not

unique to a single method of assessing a construct, several variables in the study were

assessed using multiple measures.

The use of a fairly homogenous sample of newlywed couples provided several

advantages. First, this sample allows us to distinguish between the initial onset of marital

dissatisfaction and the continuing course of marital dissatisfaction (Bradbury, 1998). In

samples that vary in their marital duration, a decline in satisfaction may represent the

beginning of marital difficulties or the further deterioration of the marriage. The study of

newlyweds, on the other hand, enables research to clarify the origins of marital

instability. Second, newlywed couples are an appropriate sample in which to examine

issues of change and stability. Compared to more established marriages, newlyweds

experience more dramatic changes in relationship quality and are at elevated risk of







25
marital disruption (Bradbury, 1998; Cherlin, 1992). Similarly, couples in the early years

of marriage may be more likely to be exposed to a variety of stressful life events, as a

number of external stressors tend to accompany the transition to marriage (e.g.,

relocation, starting a new job).

Evaluating the role of external stressful life circumstances on cognitive processing

within the relationship requires attention to two important methodological issues. The

first issue involves the use of between-subjects versus within-subjects designs. Previous

research has utilized both types of designs when examining the association between stress

and relationship quality. However, it is important to recognize that these designs address

distinct questions. Between-subjects designs examine how stressed versus non-stressed

couples differ in their relationship functioning. Within-subjects designs, on the other

hand, provide a more precise examination of stress effects by investigating how changes

in stress influence an individual's relationship functioning over time, while controlling

for the individual's average relationship functioning. The current study, then, addressed

all questions at the within-subjects level.

The second methodological issue involves the distinction between chronic and

acute stress. By definition, chronic stress refers to a stable stressor experienced over an

extended duration of time, such as living in a dangerous neighborhood or having a low

income. Conversely, acute stress refers to stressful events that occur at one point in time

and have a clear onset and offset, such as a temporarily heavy workload or a personal

injury. Unlike chronic stress, which tends to remain constant over long periods of time,

acute stress is likely to vary substantially over time, and thus seems particularly suited for

examining the types of questions associated with within-subjects designs. Given that the

current study intended to examine whether variations in spouses' stress are associated









with variations in spouses' cognitive processing, the current study relied on measures ot

acute stress rather than of chronic stress.

Review of Hypotheses

Hypothesis 1

Our first hypothesis was that stress should contribute to the deterioration of

marital satisfaction through its effects on spouses' specific relationship perceptions.

Hypothesis la. Over time, fluctuations in spouses' stressful life circumstances

should be accompanied by corresponding changes in their marital satisfaction, controlling

for spouses' average level of marital satisfaction. Thus, this hypothesis tested for stress

spillover.

Hypothesis lb. Increases in spouses' stressful life circumstances should also be

associated with increases in the negativity of spouses' specific relationship perceptions

over time. This association was expected to mediate the stress spillover effect.

Hypothesis 2

Our second hypothesis was that stress should contribute to the deterioration of

marital satisfaction through its effects on the organization of spouses' specific

relationship perceptions.

Hypothesis 2a. Over time, fluctuations in spouses' stressful life circumstances

should be accompanied by changes in their cognitive organization, controlling for

spouses' average level of cognitive organization and for changes in spouses' cognitive

content. Specifically, we predicted that the association between spouses' stress and their

cognitive organization would be curvilinear.

Hypothesis 2b. Spouses' cognitive organization should mediate the stress

spillover effect.







27

Hypothesis 3

Our third hypothesis was that partners' own stressful experiences would moderate

the association between spouses' stress level and partners' relationship satisfaction.

Hypothesis 3a. Parallel to Hypothesis I a, over time fluctuations in spouses'

external stressful life circumstances should be accompanied by corresponding changes in

their partners' marital satisfaction. Thus, this hypothesis tested for stress crossover.

Hypothesis 3b. Partners' own stress should moderate the association between

partners' overall satisfaction and their spouses' stressful circumstances. Namely, stress

crossover should be greatest when partners also are experiencing high levels of stress.

Hypothesis 3c. Parallel to Hypothesis lb, increases in spouses' stressful life

circumstances should also be associated with increases in the negativity of their partners'

specific relationship perceptions over time. This association was expected mediate the

stress crossover effect.

Hypothesis 3d. Partners' own stress should moderate the association between

partners' specific relationship perceptions and their spouses' stressful life circumstances.

Namely, spouses' stress should lead to the greatest increase in the negativity of their

partners' specific perceptions when partners are experiencing high levels of stress.














Hypothesis 4

Our fourth hypothesis was that intimates who successfully cope with stressors

early in the relationship by preventing their stressful life experiences from spilling over

into their relationship satisfaction should exhibit resilience to future stressors.

Hypothesis 4a. Spouses who are able to maintain a positive cognitive

organization in the face of stressors encountered early in the relationship should exhibit

more stable levels of marital satisfaction over time.

Hypothesis 4b. Spouses who are able to maintain a positive cognitive

organization in the face of stressors encountered early in the relationship should exhibit

lower levels of stress spillover over time.

Hypothesis 4c. Spouses who are able to maintain a positive cognitive

organization when faced with stress early in the relationship should exhibit a more

positive cognitive organization in the face of stressors encountered later in the

relationship.













METHOD

Participants

Newlywed couples were recruited for this study using two methods. First,

advertisements were placed in community newspapers and bridal shops, offering up to

$300 to couples willing to participate in a study of the early years of marriage. Second,

letters were sent to couples who had applied for marriage licenses in Alachua County,

Florida. Couples responding to either method of solicitation were screened in a telephone

interview to determine whether they met the following criteria: (a) this was the first

marriage for each partner, (b) the couple had been married less than 6 months, (c) neither

partner had children, (d) each partner was at least 18 years of age and wives were less

than 35 years of age (to allow that all couples were capable of conceiving children over

the course of the study), (e) each partner spoke English and had completed at least 10

years of education (to ensure comprehension of the questionnaires), and (f) the couple

had no immediate plans to move away from the area. The final sample consisted of 82

couples. Analyses revealed no significant differences in age or years of education

between couples recruited through the different types of solicitations.

On average, husbands were 25.1 ($JSD = 3.3) years old, and had received 16.3 (SD

= 2.4) years of education. Forty percent were employed full time and 54% were full time

students. Wives averaged 23.7 (SD = 2.8) years old and had received 16.3 (SD = 1.2)








83% of husbands and 89% of wives were white. The average combined income of

couples was less than $20,000 per year.

Procedure

Couples meeting eligibility requirements were scheduled to attend a 3-hour

laboratory session. Before the session, they were mailed a packet of questionnaires to

complete at home and bring with them to their appointment. This packet contained self-

report measures of stress and of relationship perceptions as well as a letter instructing

couples to complete all questionnaires independently of one another.

Every six months following the initial assessment, couples were contacted by

phone and mailed additional packets of questionnaires along with postage-paid return

envelopes and a letter of instruction reminding couples to complete all forms

independently of one another. Couples were paid $25 to continue participating at each

follow-up. This study will examine seven waves of data, covering approximately the firsi

3 years of marriage. At Time 7, the final wave of data collection described here, 66

couples were still married, eight couples had divorced, and eight couples had withdrawn

from the study. Of the 66 couples who were still married and participating in the study,

54 couples (82.0%) returned completed packets at Time 7. This slight attrition over time,

however, should not affect the results presented here as all analyses relied on growth

curve modeling. One advantage of growth curve modeling is that this type of analysis

includes both participants providing full data as well as those participants who did not

provide a full seven waves of data. All subsequent analyses, then, are based on data from

all 82 couples.








Materials

Global Marital Satisfaction

Most frequently administered measures of relationship satisfaction (e.g., the

Marital Adjustment Test; Locke & Wallace, 1959) include items that assess intimates'

global relationship evaluations as well as items assessing perceptions of specific aspects

of the relationship (e.g., evaluation of communication skills). To ensure that global and

specific ratings were not confounded in the present study, marital satisfaction was

measured using a 15-item version of the Semantic Differential (SMD; Osgood, Suci, &

Tannenbaum, 1957; see Appendix A) that assessed global evaluations of the relationship

exclusively. At each time point, spouses were asked to indicate their current feelings

about their marriage on 7-point scales between two opposing adjectives (e.g., "satisfied-

dissatisfied," "unpleasant-pleasant," "rewarding-disappointing,"). Scores on the measure

can range from 15 to 105, with higher scores indicating higher satisfaction. The internal

consistency of the measure was high across all seven waves of data collection, ranging

from .91 to .98 for husbands and from .93 to .98 for wives.

Specific Relationship Perceptions

Spouses' specific perceptions of the relationship were assessed at each time point

using the Marital Problems Inventory (MPI; Geiss & O'Leary, 1981; see Appendix B).

The measure lists nineteen potential problem areas in a marriage (e.g., trust,

communication, household management) and asks participants to rate each item on a

scale from I ("not a problem") to 11 ("major problem"). Of the nineteen areas of

difficulty included on the original measure, we selected only those problems that are

internal to the relationship to be included in the final composite score. Thus, the

following items were not included in the composite score as these items may represent








external stressors on the relationship: in-laws, parents, relatives; recreation and leisure

time; friends, money management; drugs and alcohol; career decisions; and amount of

time spent together. The remaining twelve items were summed to form an index of the

negativity of spouses' specific relationship perceptions. Composite scores can range from

12 to 132, with higher scores representing more negative perceptions of the relationship.

Internal consistency of the measure was high across all seven waves of data collection,

ranging from .85 to .92 for both husbands and wives.

Cognitive Organization

Differential importance. One way spouses may organize their specific

perceptions is to attribute differential importance to their positive and negative specific

perceptions. Attributing greater importance to positive relationship perceptions than to

negative relationship perceptions serves to limit the contribution of negative perceptions

to the global evaluation, allowing satisfaction to remain high despite the presence of

specific negative aspects of the relationship (Neff & Karney, unpublished manuscript).

To assess spouses' use of a differential importance strategy to organize their specific

relationship perceptions, spouses completed the Inventory of Specific Relationship

Standards at each time point (Baucom, Epstein, Rankin, & Burnett, 1996; See Appendix

C). The measure presents spouses with sixteen specific relationship standards, such as

"My partner and I should spend a lot of time and energy expressing physical affection for

each other," and "My partner and I should have the same ideas about how the housework

should be done."

For each item, spouses were asked two questions. First, spouses were asked to

indicate whether the standard was currently being met in their relationship. Thus, this

question assessed spouses' current perceptions of how the marriage was meeting or







33
failing to meet each specific standard. For the first three waves of data collection (Tl-

T3), this question was measured on a dichotomous scale (1 = yes, 0 = no) as originally

designed by Baucom and colleagues (1996). This dichotomous scale was changed to a

five-point continuous scale for the last four waves of data collection (T4-T7). However,

given that we were interested in comparing spouses' responses across the waves of data,

the continuous scale was changed back into a dichotomous scale for the purposes of

computing the differential importance index described below.' Second, spouses were

asked to indicate on the same page how upset they would be if the standard were not met

(1 = not at all; 3 = very much). Thus, this question assessed the importance spouses

attributed to each standard. Specifically, a response indicating that the individual would

be very upset if the standard were not met suggests that the standard is highly important

to the person. To calculate a differential importance index, the within-subjects

association between specific relationship perceptions and the importance of those

perceptions was then computed according to the following equation:

Specific Perception = 3j + O3ij (Importance of Perception) + error

Thus, a higher 01 would represent a higher, more positive differential importance index,

indicating that spouses' view their positive perceptions as more important than their

negative perceptions. This measure of cognitive organization was computed for each

spouse at every time point.

Attributions. Attributions represent a process in which spouses determine whether

their partners' specific behavioral failings are taken as an indication of broader faults in



To change the 5-point continuous scale to a dichotomous scale, the frequency
distribution of each item was examined for husbands and wives at each of the first three
time points. In general, approximately 80% of individuals responded "yes" to each item
atinl 0/nlO vlf-xrflin -,f 4r l "^ "6I-- Tr n,-k*-,- +. +1-* C- - .. .i-K i n4 +'








the relationship, or whether these behaviors should be separated from overall judgments

of the relationship. Relying on temporary, situational attributions to describe a partner's

transgression should serve to weaken the link between the negative specific behavior and

spouses' global relationship evaluation (McNulty & Karney, 2001). The manner in which

spouses use attributions to link specific behaviors to their global satisfaction was assessed

at each time point using the Relationship Attributions Measure (RAM; Fincham &

Bradbury, 1992; see Appendix D). This 24-item measure presents spouses with four

negative stimulus events that are likely to occur in all marriages (e.g., "Your spouse

criticizes something you say" and "Your spouse begins to spend less time with you").

For each event, spouses are asked to rate their agreement, on a 7-point scale ranging from

"Agree strongly" to "Disagree strongly," with statements that reflect six attribution

dimensions. The causal attribution sub-scale consists of 12 judgments (3 dimensions X 4

stimulus events) and the responsibility attributions sub-scale consists of 12 judgments.

For causal attributions, the three dimensions relate to the perceived locus, globality, and

stability of the cause of the negative partner behavior. For responsibility attributions, the

three dimensions capture the extent to which spouses consider their partners' behaviors as

intentional, selfishly motivated, and blameworthy. For each sub-scale, a composite score

was computed by summing the 12 judgments, resulting in two scores for each spouse

with possible ranges of 12 to 84. Higher scores indicate attributions that view the partner

in a more negative light. Internal consistency of each sub-scale was relatively high

across the seven waves of data collection. Coefficient alphas for causality attributions

ranged from .85 to .92 for husbands and from .73 to .86 for wives. Coefficient alphas for

when converting the continuous scale to a dichotomous scale, we recorded a response of 1
or 2 to be a 0 (i.e., "no") and a 3, 4, or 5, to be a 1 (i.e., yes). After this recoding, the
frequency distribution of each item tended to be 85% "yes" and 15% "no."








responsibility attributions ranged from .89 to .95 for husbands and from .88 to .91 for

wives. Causality and responsibility attributions were significantly correlated across the

time points for husbands (ranging from .58 to .78) and for wives (ranging from .52 to

.66).

Stressful Life Circumstances

To assess external acute stress at each time point, couples completed a subset of

the Stressful Life Events checklist (SLE; Bradbury, unpublished manuscript; see

Appendix E) designed to assess life events in the previous 6 months. Ninety events were

selected from other standardized life events checklists, with an emphasis on objective

events likely to occur in a young, married population. Events were grouped to represent

nine life domains: marriage, work, school, family and friends, finances, health, personal

events, living conditions, and legal. For each event, spouses were first asked to indicate

whether the event occurred. If the event occurred, spouses then indicated the impact the

event had on their lives on a 7-point scale ranging from extremely negative (-3) to

extremely positive (+3). Each stressful event then had to meet two criteria to be included

in the final composite score. First, the event could not represent a likely consequence of

marital satisfaction or marital distress. Fourteen items were not included in the final

score for this reason.2 In this way, the measure should tap only those stressors external to

(i.e., less likely to be caused by) the marriage. For instance, whereas being fired from a

job or being hospitalized may affect the marriage, these events are less likely to have


2 The following items were not included on the Survey of Stressful Life Events as they
were likely to represent consequences of marital satisfaction or marital distress: change
in quality of relationship with spouse; change in number of arguments with spouse; had
an affair; spouse had an affair; reconciliation with spouse after separation; pregnancy;
had a baby; abortion; sexual difficulties; dropped out of school for personal reasons;








UViV vauUcu uy 11 laia AAJ, UAUuwAI > 'AI. -- -* I..- ---

and Wheaton (1997), the event had to represent a negative life stressor. Turner and

Wheaton (1997) found in their review of the stress literature that clear patterns of results

emerge in studies of negative events whereas the effects of positive events are weak and

inconsistent. On the basis of this review, the authors have recommended that stressors

generally considered as positive be excluded from stress measures. To identify negative

stressful events, we examined the average impact rating of each event at each wave of

data collection. To be included in the final composite score, the event had to be rated as

having a negative impact on average by both husbands and wives each time the item was

endorsed. Thus, a total of 51 stressful life events were used to calculate the final stress

score (see bolded items in Appendix F). Analysis of the events revealed that at each time

point, most (75%) of these 51 items were endorsed by at least one individual, suggesting

that the items are a reasonable sampling of life events likely to be experienced by couples

in the early stages of marriage. The final stress score was computed by adding together

the number of negative events the spouse reported had occurred. Stress scores could















provides reliable estimates of within-subject parameters even when sample sizes are

relatively small. Second, HLM provides maximally efficient estimates of these

parameters by weighting individual estimates according to empirical Bayes theory. When

the within-subject parameter for an individual can be estimated precisely, the final

estimate relies heavily on the individual data. When the parameter cannot be estimated

precisely (e.g., because of missing data), the final estimate relies more heavily on the

mean of the sample. Because the most precise estimates therefore contribute more to the

final estimated variance of the sample, variances estimated in this way tend to be more

conservative than those obtained through traditional OLS methods.

In general, data from each spouse was used to estimate the association between

spouses' stressful life experiences and the content and organization of their relationship

cognitions.













RESULTS

Descriptive Statistics and Correlations

Table I presents descriptive statistics for measures of global satisfaction and of

specific relationship perceptions. As would be expected from a sample of newlywed

couples, on average both husbands and wives reported high levels of global marital

satisfaction and low levels of negative specific relationship perceptions. Spouses'

perceptions of specific problems in the relationship were significantly negatively

associated with their global marital satisfaction, with correlations ranging from -.60 to

-.83 for husbands and from -.77 to -.86 for wives. A repeated-measures ANOVA with a

linear contrast test was conducted to test for linear change in global satisfaction scores

over time. Results revealed that both husbands' and wives' satisfaction tended to

decrease significantly over the seven waves of data, F (1,47) = 5.5, p = .02, T2 =. 11 for

husbands and F (1,49) = 13.9, p =.001, T2 = .22 for wives. Tests for linear change in

spouses' perceptions of specific marital problems were not significant for husbands, F

(1,48) = .22, p = .64, T12 = .01, but were significant for wives, F (1,48) = 4.5, p = .04, l2 =

.09. Thus, on average, husbands' negative specific perceptions of the relationship

seemed to remain fairly stable across the seven waves of data. Conversely, wives tended

to perceive more specific problems in the relationship over time.

Table 2 presents descriptive statistics for all measures of cognitive organization.

On average, husbands and wives appeared to make relatively positive attributions for






39

their partners' negative behaviors, seeing external causes for negative events and freeing

their partners from blame. Tests for linear change in causal attributions were not
+

significant for husbands, but were significant for wives, F (1,38) = .09, p = .76, 112 = .003,

and F (1,40) = 4.0, p = .05, 12 = .09, respectively. On average, then, husbands' causal

attributions of their wives' transgressions remained fairly stable over time. However,

wives' tendency to view their husbands as the cause of negative behaviors increased over

time. Tests for linear change in spouses' responsibility attributions were not significant

for husbands or for wives, F(1,38) = 3.S,p = .06, T12 = .09, and F(1,40) = 2.8, p= .10, T12

=.07, respectively. Thus, on average, husbands' and wives' tendencies to perceive their

partners as responsible for negative behaviors remained fairly stable over time.

Turning to spouses' use of differential importance to organize their specific

perceptions, the average differential importance index was positive and significant for

both spouses across time, suggesting that, on average, husbands and wives tended to

attribute more importance to their positive relationship perceptions than to their negative

relationship perceptions. Tests for linear change in spouses' differential importance index

were significant for wives but not for husbands, F(1,31) = 77.6,p <.001, 12 = .72, and F

(1,31) = 2.l,p = .16, 12 = .06, respectively. Thus, on average, wives seemed to exhibit a

weaker tendency to attribute more importance to their positive relationship perceptions

than to their negative relationship perceptions over time. Husbands' differential

importance index, however, seemed to remain fairly stable over time. Differential

importance was inconsistently significantly associated with both causal and responsibility

attributions. Husbands' tendency to attribute more importance to positive relationship

perceptions than to negative relationship perceptions was significantly associated with a

walker tendencrv tn make internal and Naming attributions for a nartner's negative








behavior at three of the seven time points, with correlations ranging from .08 to -.39 for

causal attributions and from .03 to -.45 for responsibility attributions. Wives' use of

differential importance was significantly associated with a weaker tendency to make

internal and blaming attributions for their partners' negative behavior at four of the seven

time points, with correlations ranging from -.09 to -.46 for causal attributions and from

.09 to -.45 for responsibility attributions.

Table 3 presents descriptive statistics for the measure of acute stress. On average,

husbands and wives reported experiencing low numbers of acute negative stressors.

However, the range of the number of stressors reported at each time point was fairly

large, suggesting there was at least some variability in the number of stressors that

spouses were experiencing. Tests for linear change in acute stress scores were significant

for husbands and for wives, F (1, 36) = 11.5, p =.002, TI = .25, and F (1,39) = 5.5, p =

.02, T2 = .13, respectively. Thus, the number of acute stressors spouses reported tended to

decrease over the seven waves of data.

With respect to the independent and hypothesized mediating variables, wives'

stress was significantly associated with the negativity of their specific relationship

perceptions, with correlations ranging from .23 to .57. However, the associations

between husbands' stress and their specific relationship perceptions rarely reached

significance, with correlations ranging from .01 to .40. Stress was only inconsistently

significantly associated with less relationship-enhancing cognitive organizations. In most

cases, the association between stress and the tendency to make internal causal attributions

for a partner's negative behavior did not reach significance, with correlations ranging

from .08 to .30 for husbands and from .13 to .28 for wives. Likewise, the association

between stress and the tendency to perceive the partner as responsible for negative






41

behaviors rarely reached significance, with correlations ranging from .05 to .25 for

husbands and from .02 to .28 for wives. Nevertheless, stress did tend to be significantly

associated with a weaker tendency to attribute more importance to positive relationship

perceptions than to negative relationship perceptions (i.e., a lower differential importance

index), with correlations ranging from -.08 to -.25 for husbands and from -.41 to .22 for

wives.

Spouses' cognitive content tended to be associated with their cognitive

organization. Spouses' negative relationship perceptions were significantly associated

with the tendency to make internal attributions for a partner's negative behavior, with

correlations ranging from .34 to .50 for husbands and from .33 to .57 for wives.

Similarly, spouses' negative relationship perceptions were significantly associated with a

tendency perceive the partner as responsible for negative behaviors, with correlations

ranging from .33 to .45 for husbands and from .22 to .49 for wives. Finally, spouses'

negative relationship perceptions were significantly associated with a weaker tendency to

attribute more importance to positive relationship perceptions than to negative

relationship perceptions, with correlations ranging from -. 19 to -.60 for husbands and

from -. 11 to -.50 for wives. Thus, all analyses examining the role of cognitive

organization in stress spillover and crossover process controlled for spouses' cognitive

content.

Cross-spouse correlations for the independent and mediating variables reveal that

husbands' and wives' stress scores were significantly associated, with correlations

ranging from .16 to .44. With regard to cognitive content, husbands' and wives' negative

perceptions of the relationship were significantly associated, with correlations ranging

from .37 to .73, suggesting that spouses tended to perceive their relationships in similar






42

ways. Turning to measures of cognitive organization, husbands' and wives' tendencies to

attribute their partners' negative behavior to internal causes were significantly associated

at three of the seven time points, with correlations ranging from -.03 to .32. In addition,

husbands' and wives' tendencies to view their partners as responsible for their negative

behaviors tended to be significantly associated, with correlations ranging from -.09 to .43.

Finally, husbands' and wives' tendencies to attribute more importance to their positive

perceptions than to their negative perceptions were significantly associated, with

correlations ranging from .52 to .81. Thus, spouses seemed to organize their specific

relationship beliefs in similar ways.

In most cases, the association between husbands' stress and wives' negative

relationship perceptions did not reach significance, with correlations ranging from .07 to

.29. However, wives' stress was significantly associated with husbands' negative

relationship perceptions, with correlations ranging from .05 to .48. Overall, spouses'

stress was not significantly associated with their partners' cognitive organization.

Husbands' and wives' stress were not significantly associated with their partners'

tendency to make internal causal attributions for negative behaviors, with correlations

ranging from .07 to .31 for husbands' stress and from -.18 to .22 for wives' stress.

Husbands' and wives' stress also were not significantly associated with their partners'

tendency to make blaming attributions for negative behaviors, with correlations ranging

from -.10 to .29 for husbands' stress and from -.05 to .27 for wives' stress. Finally,

husbands' and wives' stress were not significantly associated with their partners'

tendency to attribute more importance to positive relationship perceptions than to








Overall, then, preliminary analyses indicate that all measures performed generally

as expected. In general, spouses' stress was positively and significantly related both to

spouses' own negative relationship perceptions. Contrary to Hypothesis 2, spouses'

stress was not consistently significantly associated with spouses' own cognitive

organization. However, these bivariate correlations do not threaten subsequent analyses

as they only test for a linear relationship between stress and cognitive organization rather

than the predicted curvilinear relationship. Moreover, these correlations do not address

the within-subjects association between changes in stress and changes in cognitive

organization. In line with Hypothesis 3, spouses' stress was associated with their

partners' negative specific perceptions of the relationship, but was not significantly

related to their partners' cognitive organization. To more thoroughly examine the

hypotheses of the current study, the following sections present the results of analyses

conducted to investigate the within-subjects association between changes in stress and

changes in cognitive content or cognitive organization over time.

Do Spouses' Specific Relationship Perceptions Mediate the Stress Spillover Process?

Previous research on stress and relationship quality consistently has demonstrated

that the experience of stress may affect individuals' judgments of their relationships.

However, this literature has failed to address the processes through which stress may

negatively influence relationship judgments. The first goal of these analyses, then, was to

examine the role of cognitive content in the stress spillover process. Hypothesis 1

suggests that stress should contribute to the deterioration of marital satisfaction through

its effects on spouses' specific relationship perceptions.








Hypothesis la: The Stress Spillover Hypothesis

Hypothesis la sought to replicate and extend previous work on stress spillover.

Specifically, we predicted that increases in spouses' external negative life events would

be associated with decreases in their marital satisfaction over the first 3 /2 years of

marriage. Preliminary analyses revealed that spouses' marital satisfaction seemed to

significantly decrease over time. Thus, to examine the stress spillover hypothesis, we

first examined the appropriate baseline model of satisfaction over time for husbands and

for wives by comparing the following two equations:

Satisfaction= P3j + r ij [Equation 1]

Satisfaction= 0oj + 3ij (time) + r [Equation 2]

Again, results revealed that, on average, husbands' and wives' satisfaction significantly

declined over the first years of marriage, 3Ij = -.80, SE = .23, t (81) = -3.4,p = .001,

effect size r = .35 for husbands and ij = -1.3, SE= .27, 1 (81) = -4.8, p < .001, effect size

r = .47 for wives. Moreover, including time in the model significantly improved the fit of

the model for both husbands and wives, x2 (2) = 72.8, p < .001 and X2 (2) = 82.2, p <

.001, respectively. Consequently, to address the stress spillover hypothesis, we examined

the within-person association between spouses' stress and their marital satisfaction

according to the following model:

Satisfaction = O + Pij (time) + Nj (stress) + r [Equation 3]

where time and stress were group-mean centered. In this equation, O3q represents an

estimate of the average positivity of a spouse's global marital satisfaction. Oij represents

the slope of a spouse's satisfaction over the first years of marriage. f6j, then, captures the

within-person association between changes in stress and changes in marital satisfaction






45

over the first years of marriage for a given spouse, controlling both for a spouse's

tendency to view the relationship as more or less satisfying and for the tendency for

satisfaction to decrease linearly over time. In other words, a negative 32j would indicate

that increases in a spouse's negative external stressors are associated with decreases in

marital satisfaction above and beyond the tendency for satisfaction to decrease simply as

a function of time. Finally, rij is the residual variance in satisfaction for a spouse, assumed

to be independent and normally distributed across spouses. This equation was estimated

for each spouse and the significance of the average 2 term across spouses was

investigated.

Results revealed that, for wives, increases in exposure to negative external

stressors were significantly associated with decreases in marital satisfaction, 32j = -.42,

SE = .18, t (81) = -2.3,p = .02, effect size r = .23. Thus, extending previous work, which

has demonstrated stress spillover processes over the course of several days, these results

provided evidence of stress spillover over the first 3 '/2 years of marriage. However,

stress spillover was not found for husbands. For husbands, changes in stress were not

significantly associated with changes in marital satisfaction, 32j = -.002, SE= .16, t (81)

= -.02, p = .99, effect size r = .002.

Hypothesis lb: Specific Perceptions and the Stress Spillover Process

Given that wives' stress was found to be significantly associated with their marital

satisfaction, Hypothesis lb predicted that wives' specific relationship perceptions should

mediate this stress spillover effect. Specifically, increases in stress were expected to be

associated with corresponding increases in the negativity of specific relationship

perceptions. To examine this hypothesis, the procedures for testing mediation outlined








Baron and Kenny (1986) argue that three additional equations must be estimated in order

to test for mediation effects. First, the proposed mediator variable must be significantly

associated with the outcome variable. In other words, specific relationship perceptions

should be significantly associated with global marital satisfaction. This effect was

modeled according to the following equation:

Satisfaction= 3oj + 3ij (time) + 32j (specific perceptions) + r ij [Equation 4]

where time and specific perceptions were group-mean centered. In this equation, 3oj

represents an estimate of the average positivity of a wife's global marital satisfaction. 32j

represents the slope of a wife's satisfaction over the first years of marriage. 132j represents

the within-person association between perceptions of specific problems in the

relationship and marital satisfaction for a given wife. A negative 32j would indicate that

in increases in the negativity of specific relationship perceptions are associated with

decreases in overall marital satisfaction, controlling both for a wife's tendency to view

the relationship as more or less satisfying and for the tendency of satisfaction to decrease

linearly over time. Finally, rij is the residual variance in satisfaction for a wife, assumed

to be independent and normally distributed across wives. This equation was estimated

for each wife and the significance of the average 132 term across wives was investigated.

This association was in fact significant, 32j = -.42, SE = .05, t (81) = -8.3,p < .001, effect

size r = .68, suggesting that changes in specific perceptions are associated with

corresponding changes in global satisfaction over time.

Second, the independent variable must be significantly associated with the

proposed mediating variable. In other words, stress should be significantly associated

with specific relationship perceptions. To model this association, we first examined the








appropriate baseline model of specific relationship perceptions over time for wives by

comparing the following two equations:

Specific Perceptions = 3oj + r ij [Equation 5]

Specific Perceptions = 3o + 13ij (time) + r ij [Equation 6]

Results revealed a significant tendency for wives' specific perceptions of the relationship

to become more negative over the first years of marriage, pij = .71, SE = .24, t (81) = 2.9,

p = .004, effect size r = .30. Moreover, including time in the model significantly

improved the fit of the model, x2 (2) = 25.5, p < .001. Consequently, the within-person

association between stress and specific relationship perceptions was modeled according

to the following equation:

Specific Perceptions = O3j + P3ij (time) + 32j (stress) + r ij [Equation 7]

where time and stress were group-mean centered. In this equation, 03j represents an

estimate of the average negativity of a wife's specific relationship perceptions. 3ij

represents the slope of a wife's specific perceptions over the first years of marriage. 2j,

then, captures the within-person association between stress and specific relationship

perceptions over the first years of marriage for a given wife, controlling both for a wife's

tendency to view the relationship more or less negatively and for the tendency for

negative relationship perceptions to increase linearly over time. In other words, a

positive 32j would indicate that increases in a wife's negative external stressors are

associated with corresponding increases in a wife's perceptions of specific problems in

the relationship, above and beyond the tendency for perceptions of specific problems to

;nnnic- i;rmnlv nc a fiintm;nn rnf tim p m;nllv r- ic thp rpaitrlal varnnriP in enp-ifir








This equation was estimated for each wife and the significance of the average 032 term

across wives was investigated. Results showed that, over the first years of marriage,

increases in wives' exposure to negative external stressors were associated with increases

in the negativity of wives' specific relationship perceptions, 02j = .67, SE = .17, t (81) =

3.9, p < .001, effect size r = .40.

Finally, Baron and Kenny (1986) argue that both the independent variable and the

predicted mediator variable should be regressed simultaneously onto the outcome

variable. In other words, the association between stress and satisfaction and specific

relationship perceptions and satisfaction should be estimated simultaneously. If the

association between stress and satisfaction is lower in this equation than in Equation 3,

this would provide evidence for mediation effects. Thus, to test for mediation, the

following equation was modeled:

Satisfaction= Poj+ ij (time) + P,2j (specific perceptions) + 03j (stress) +r i [Equation 8]

where time, specific perceptions and stress were group-mean centered. In this equation,

Ooj represents an estimate of the average positivity of a wife's global marital satisfaction.

pij represents the slope of a wife's satisfaction over the first years of marriage. p2j

represents the within-person association between perceptions of specific problems in the

relationship and marital satisfaction. 33j represents the within-person association

between stress and marital satisfaction. Finally, rij is the residual variance in satisfaction

for a wife, assumed to be independent and normally distributed across wives. This

equation was estimated for each wife and the significance of the average p2 and 33 terms

across wives was investigated.








Results demonstrated that increases in the negativity of wives' specific

relationship perceptions remained significantly associated with decreases in their marital

satisfaction, 02j = -.43, SE = .05, t (81) = -8.2, p < .001, effect sizer = .67. However,

changes in wives' stress were no longer associated with changes in wives' marital

satisfaction, 33j = -.03, SE = .16, t (81) = .18, p = .86, effect size r = .02. Consequently,

wives' specific relationship perceptions seemed to fully mediate the stress spillover

process.1

Overall, then, evidence for stress spillover was found for wives, but not for

husbands. For wives, changes in the number of negative external stressors being

experienced were significantly associated with changes in their global marital satisfaction

over the first 3 years of marriage, such that increases in stress were associated with

decreases in satisfaction. Moreover, wives' specific relationship perceptions seemed to

mediate this stress spillover process. Namely, changes in the number of external

stressors being experienced were significantly associated with changes in wives' specific

relationship perceptions, such that increases in stress were associated with increases in

the negativity of wives' specific perceptions. This association between stress and specific

perceptions seemed to account for the relationship between stress and overall satisfaction.

Does Spouses' Cognitive Organization Mediate the Stress Spillover Process?

The second goal of these analyses was to examine the role of cognitive organization in

the stress spillover process. Hypothesis 2 suggests that stress also should contribute to

the deterioration of marital satisfaction through its effects on spouses' cognitive


1 Given the limitations of cross-sectional data for testing mediation, we further tested the
hypothesis that specific perceptions mediate the stress spillover process by examining the







50
organization. Again, given that evidence for stress spillover was found for wives but not

for husbands, the following analyses relied on data from wives only. Moreover, as this

study measured three types of cognitive organization (causal attributions, responsibility

attributions and differential importance), each of the following analyses was conducted

three times, using each measure of cognitive organization.

Hypothesis 2a: Stress and Cognitive Organization

Hypothesis 2a predicted that changes in spouses' external negative life events

would be associated with corresponding changes in their cognitive organization over the

first 3 1, years of marriage, independent of changes in their cognitive content

Specifically, we predicted that spouses' cognitive organization would be cuvilinearly

related to their stress, such that as stress increases from low to moderate, spouses'

cognitive organization should become more positive. As stress increases from moderate

to high, however, spouses' cognitive organization should become more negative. To

examine the association between stress and cognitive organization for wives, we first

examined the appropriate baseline model for modeling cognitive organization over time

by comparing the following two equations for each measure of cognitive organization:



Cognitive Organization = Poj + rg [Equation 9]

Cognitive Organization= Pj + 0ij (time) + rij [Equation 10]

Results indicated that wives' tendency to make internal attributions for their husbands'

behavioral transgressions tended to increase significantly over time, Oij = .39, SE =. 16, t

(81) = 2.4, p = .02, effect size r = .26. Moreover, including time in the model

significantly improved the fit of the model, x2 (2) = 19.9, p < .001. Results also revealed








p = .06, effect size r = .21, however, including time in the model nevertheless improved

the fit of the model, X2 (2) =21.3, p < .001. Finally, wives' tendency to view positive

relationship perceptions as more important than negative relationship perceptions

significantly declined overtime, 3ij = -.09, SE= .01, t (81) = -9.9, p< .001, effect size r

=.74. Again, including time in the model significantly improved the fit of the model, x2

(2) = 91.7, p < .001.

To test for a curvilinear relationship between stress and cognitive organization,

the within-person association between wives' stress and their cognitive organization was

then modeled according to the following equation:

Cognitive Organization= 03o, + Pj (time) + 32j (stress) + 03j stresss2+ rij [Equation 11]

where time and stress were group-mean centered. In this equation, Oj represents an

estimate of the average positivity of a wife's cognitive organization. 3ij represents the

slope of a wife's cognitive organization over the first years of marriage. 02j, then,

captures the association between cognitive organization and stress over the first years of

marriage for a given wife, controlling both for a wife's tendency to organize specific

perceptions more or less positively and for the tendency of cognitive organization to

change linearly over time. In other words, in the case of causal or responsibility

attributions, a positive 32j would indicate that increases in a wife's negative external

stressors are associated with decreases in the positivity of a wife's cognitive organization.

In the case of differential importance, a negative 02j would indicate that increases in a

wife's negative external stressors are associated with decreases in the positivity of a

wife's cognitive organization. 133j captures the curvilinear association between stress and

cognitive organization- again controlling for a wife's tendency to organize specific








perceptions more or less positively and for the tendency of cognitive organization to

change linearly over time. In the case of causal and responsibility attributions, positive

03j would indicate the predicted U-shape curve. In the case of differential importance, a

negative 133j would indicate the predicted U-shape curve. Finally, rij is the residual

variance in cognitive organization for a wife, assumed to be independent and normally

distributed across wives. This equation was estimated for each wife and the significance

of the average 33 term across wives was investigated.

Results revealed no significant curvilinear associations between stress and

cognitive organization, 03j = -.05, SE= .03, t (81) = -. 1.8, p = .08, effect size r= .20 for

causality attributions; 3j = .001, SE= .04, t (81) = .02, p = .99, effect size r = .002 for

responsibility attributions; and 03j = .-.0001, SE = .002, t (81) = -.09, p =.93, effect size r

= .01 for differential importance.

To test for a linear relationship between stress and cognitive organization, then,

the within-person association between stress and cognitive organization was modeled

according to the following equation:

Cognitive Organization= 1oj + p j (time) + 02j (stress) + r [Equation 12]

where time and stress were group-mean centered. As in the previous equation, 2j,,

captures the association between cognitive organization and stress over the first years of

marriage for a given wife, controlling both for a wife's tendency to organize specific

perceptions more or less positively and for the tendency of cognitive organization to

change linearly over time. Again, in the case of causal or responsibility attributions, a

positive 02j would indicate that increases in a wife's negative external stressors are

associated with decreases in the positivitv of a wife's cognitive organization. In the case






53

of differential importance, a negative 12j would indicate that increases in a wife's

negative external stressors are associated with decreases in the positivity of a wife's

cognitive organization. This equation was estimated for each wife and the significance of

the average 32 term across wives was investigated.

Results indicated no significant linear association between wives' causality

attributions and their stress, 2j = .17, SE= .13, t (81) = 1.3,p = .18, effect size r = 14.

However, increases in wives' stress were significantly associated with a stronger

tendency to blame the partner for negative behaviors, I0j = .46, SE= .17, t (81) = 2.7, p =

.007, effect size r = .29. Wives' stress was not significantly associated with their

tendency to attribute more importance to positive relationship perceptions than to

negative relationship perceptions, 12j = .01, SE= .01, t (81) = .52,p =.60, effect size r =

.06.

Preliminary analyses revealed that spouses' cognitive organization tended to be

significantly associated with their cognitive content. Thus, to determine whether the

association between wives' stress and their responsibility attributions was independent of

wives' cognitive content, the following equation was estimated:

Responsibility = Poj + pij (time) + 02j (stress) + ,3j (specific perceptions)+ r ij

[Equation 13]

where time, stress, and specific perceptions were group-mean centered. Thus, in this

equation, 0zj represents the association between responsibility attributions and stress over

the first years of marriage for a given wife, controlling for the association between a

wife's specific relationship perceptions and her responsibility attributions. Results








associated with their responsibility attributions for their husbands' negative behaviors, 03j

= .13, SE = .05, t (81) = 2.9,p = .004, effect size r = .31. Thus, as wives' specific

perceptions of the relationship became more negative, they also exhibited a stronger

tendency to perceive their husbands as responsible for negative behaviors. Controlling

for this association, however, wives' stress remained significantly associated with their

responsibility attributions, such that increases in wives' stress were associated with a

stronger tendency for wives to perceive their husbands as responsible for negative

behaviors, 32j = .33, SE =. 17, t (81) = 2.0,p = .05, effect size r = .22.

Overall, then, no evidence was found for the predicted curvilinear relationship

between stress and cognitive organization. However, some evidence revealed a linear

association between stress and cognitive organization, such that increases in wives' stress

were associated with decreases in the positivity of wives' cognitive organization.

Namely, changes in wives' stress were significantly associated with changes in wives'

tendency to perceive their husbands as responsible for negative behaviors, even when

controlling for the general negativity of wives' cognitive content.

Hypothesis 2b: Cognitive Organization and the Stress Spillover Process

Hypothesis 2b suggested that spouses' cognitive organization should mediate the

stress spillover process. To examine this hypothesis, the procedures for testing mediation

outlined by Baron and Kenny (1986) were followed. In addition to Equation 3, which

modeled the stress spillover phenomenon, and Equation 13, which modeled the

association between stress and responsibility attributions controlling for cognitive

content, two additional equations must be estimated in order to test for mediation effects.

First, the proposed mediator variable, responsibility attributions, must be significantly






55

associated with the outcome variable, global marital satisfaction. This effect was

modeled according to the following equation:

Satisfaction = Poj + Pij (time) + 02i (specific perceptions)

+ 33j (responsibility attributions)+ rij [Equation 14]

where specific perceptions and responsibility attributions were group-mean centered. In

this equation, 3oj represents an estimate of the average positivity of a wife's global

marital satisfaction. 31j represents the slope of a wife's satisfaction over the first years of

marriage. 02j represents the within-person association between specific relationship

perceptions and global marital satisfaction. 133j represents the within-person association

between responsibility attributions and marital satisfaction. A negative 03j would

indicate that increases in the tendency to perceive a partner as more responsible for

negative behaviors are associated with decreases in overall marital satisfaction,

controlling for a wife's tendency to view the relationship as more or less satisfying, for

the tendency for satisfaction to decrease linearly over time, and for wives' cognitive

content. In other words, 03j examines the association between cognitive organization and

marital satisfaction, controlling for the association between cognitive content and marital

satisfaction. Finally, rij, is the residual variance in satisfaction for a wife, assumed to be

independent and normally distributed across wives. This equation was estimated for each

wife and the significance of the average 03 term across wives was investigated. The

association was in fact significant, 03j = -.09, SE = .03, t (81) = -2.7,p = .008, effect size

r = .29, suggesting that changes in wives' responsibility attributions were associated with

changes in wives' marital satisfaction, controlling for wives' specific perceptions of

problems in the relationship.








Second, the association between stress and satisfaction and cognitive organization

and satisfaction were estimated simultaneously. If the association between stress and

satisfaction is lower in this equation than in Equation 3, this result would provide

evidence for mediation effects. Thus, to test for mediation, the following equation was

modeled:

Satisfaction = oj + 13 ij (time) + 132j (responsibility

attributions)+ 03j (stress) +rij [Equation 15]

where time, responsibility attributions, and stress were group-mean centered. In this

equation, f3oj represents an estimate of the average positivity of a wife's global marital

satisfaction. 13ij represents the slope of a wife's satisfaction over the first years of

marriage. 132j represents the within-person association between responsibility attributions

and marital satisfaction. 03j represents the within-person association between stress and

marital satisfaction. Finally, rij is the residual variance in satisfaction for a wife, assumed

to be independent and normally distributed across wives. Given that previous results

indicated that wives' perceptions of specific problems in the relationship fully mediated

the stress spillover process, this variable was not included in the present equation. This

equation was estimated for each wife and the significance of the average 133 terms across

wives was investigated.

Results demonstrated that increases in the tendency to perceive a partner as

responsible for negative behaviors remained significantly associated with decreases in

marital satisfaction, 32j = -.20, SE = .05, t (81) = -3.8,p < .001, effect size r = .39.

However, the association between wives' stress and their marital satisfaction was no

longer significant, 3j = -.22, SE= .14, t (81) = -1.5,p = .13, effect size r = .16.








Consequently, wives' responsibility attributions seemed to fully mediate the stress

spillover process. Namely, increases in stress were associated with less positive

cognitive organization (i.e., a stronger tendency to perceive a partner as responsible for

negative behaviors), even when controlling for the negativity of spouses' specific

relationship perceptions. This association seemed to account for the stress spillover

effect.2

Do Spouses' External Stressors Interact to Affect Marital Satisfaction?

Previous research has found that spouses' stress may affect not only spouses' own

marital satisfaction, but also the marital satisfaction of their partners. However, this

research has examined stress crossover without regard for the amount of stress partners

may be experiencing themselves. The third goal of these analyses, then, was to examine

the role of the partner's own stress in the stress crossover process. Hypothesis 3 suggests

that partners' own stressful experiences may moderate the association between spouses'

stress and partners' marital satisfaction.

Hypothesis 3a: The Stress Crossover Hypothesis

Hypothesis 3a first sought to replicate and extend previous work on stress crossover.

Specifically, we predicted that increases in spouses' external negative life events would

be associated with decreases in their partners' marital satisfaction over the first 3 years

of marriage. Results from Hypothesis l a revealed that including time in the baseline

model of satisfaction significantly improved the fit of the model for both husband and

wives. Thus, the within-person association between spouses' stress and their partners'


2 Given the limitations of cross-sectional data for testing mediation, we further tested the
hypothesis that responsibility attributions mediate the stress spillover process by








marital satisfaction was examined using the following equation:

Partner's Satisfaction = oj + 3ij (time) + 3z2j (spouse's stress) + r3 [Equation 17]

where time and spouse's stress were group-mean centered. In this equation, 3oj

represents an estimate of the average positivity of a partner's global marital satisfaction.

Oij represents the slope of a partner's satisfaction over the first years of marriage. Izj,

then, captures the within-person association between partners' marital satisfaction and

their spouses' stress over the first years of marriage for a given partner, controlling both

for a partner's tendency to view the relationship as more or less satisfying and for the

tendency of satisfaction to decrease linearly over time. In other words, a negative 132j

would indicate that increases in a spouse's negative external stressors are associated with

decreases in a partner's marital satisfaction above and beyond the tendency for

satisfaction to decrease simply as a function of time. Finally, r, is the residual variance in

satisfaction for a partner, assumed to be independent and normally distributed across

partners. This equation was estimated for each partner and the significance of the

average 132 term across partners was investigated.

Evidence of stress crossover was not found for husbands, 32j = -.07, SE= .13, t

(81) = -.55,p = .58, effect size r = .06. Thus, changes in wives' stress were not

associated with changes in their husbands' marital satisfaction. However, increases in

husbands' stress were associated with corresponding decreases in wives' marital

satisfaction, 132j = -.30, SE =.13, t (81) = -2.2,p .03, effect size r = .24. Thus, these

results provide evidence of stress crossover over the first 3 V2 years of marriage.

Given that wives' stress tended to be moderately correlated with their husbands'








wives' stress spillover effect according to the following model:

Wives' Satisfaction = P3oj + 13ij (time) + 02j (wives' stress) +03j (husbands' stress) + rij

[Equation 18]

where time, husbands' stress and wives' stress were group-mean centered. In this

equation, O3j represents an estimate of the average positivity of a wife's global marital

satisfaction. Pij represents the slope of a wife's satisfaction over the first years of

marriage. 02j captures the within-person association between a wife's satisfaction and her

own stress. In other words, 32j is an estimate of stress spillover. 03j, then, captures the

within-person association between a wife's satisfaction and a husband's stress. In other

words, 33j estimates stress crossover controlling for stress spillover. Finally, rij is the

residual variance in satisfaction for a wife, assumed to be independent and normally

distributed across wives. This equation was estimated for each wife and the significance

of the average 02 and 03 terms across wives was investigated.

Results indicated that the association between wives' stress and their own

satisfaction remained significant, 32j = -.38, SE=.18, t (81) = -2.1, p =.04, effect size r

= .23. Controlling for this association, however, the association between husband's stress

and wives' satisfaction was no longer significant, 03j = -.16, SE = .14, t (81) = -1.1, p =

.26, effect size r =. 12. Thus, these results suggest that the previous evidence for a stress

crossover effect may be due simply to the correlation between husbands' and wives'

external stressors. As evidence of an independent stress crossover effect was not found,

the results of further analyses regarding the mediators and moderators of this effect (i.e.,

Hypotheses 3c-3f) are not reported.







60
Hypothesis 3b: Partners' Own Stress and the Stress Crossover Process

Though evidence of an independent crossover effect was not found, the possibility

remains that husbands' and wives' stress may interact to affect wives' satisfaction.

Hypothesis 3b predicted that the greatest declines in satisfaction should occur when both

spouses are experiencing high levels of external stress. This hypothesis was modeled

according to the following equation:

Wives' Satisfaction = poj+ Pi3j (time) + 12j (husband's stress) +33j (wives'

stress) +-4j (husbands' stress X wives' stress)+ rij [Equation 19]

where time, husbands' stress and wives' stress were group-mean centered. In this

equation, 3oj represents an estimate of the average positivity of a wife's global marital

satisfaction. Pij represents the slope of a wife's satisfaction over the first years of

marriage. 2j, then, captures the within-person association between a wife's satisfaction

and a husband's stress. In, other words, 02j is an estimate of stress crossover. 33j captures

the within-person association between a wife's satisfaction and her own stress. In other

words, 03j is an estimate of stress spillover. 134j captures the within-person association

between the interaction of spouses' stress and the wife's satisfaction, controlling both for

the wife's tendency to view the relationship more or less positively and for the tendency

for satisfaction to decrease linearly over time. Finally, rij is the residual variance in

satisfaction for a wife, assumed to be independent and normally distributed across wives.

This equation was estimated for each wife and the significance of the average 34 term

across wives was investigated. Results indicated that the association between the

interaction of stress and wives' satisfaction was not significant, 34j = -.14, SE = .10, t (81)








= -1.4,p= .16, ettect size r= I inlUS, evidence ror me iniCeUacrive cui Ut 1 L uca u

satisfaction was not found.

Is Stress Always Detrimental to Relationship Outcomes?

Results thus far suggest that the experience of external stressors may be harmful

for one's own relationship judgments. However, recent theories have argued that under

some circumstances, stress may actually enhance well-being. The final goal of these

analyses was to examine whether successful coping with stress early in the relationship

may lead spouses to be resilient to future stress. Hypothesis 4 suggests that spouses who

successfully cope with stress may emerge from the experience as less susceptible to the

adverse effects of later stressors.

Hypotheses 4a and 4b: Coping. Satisfaction, and Stress Spillover

Hypothesis 4a suggested that spouses who maintain a positive organization of

relationship perceptions in the face of external stress should be less vulnerable to declines

in relationship satisfaction over time. Similarly, Hypothesis 4b predicted that spouses

who maintain a positive organization of relationship perceptions in the face of external

stress early in the relationship would be less vulnerable to stress spillover effects over

time. These hypotheses were addressed simultaneously by re-examining the following

previously estimated equation:

Satisfaction = Pj + 3 ij (time)+ 02j (stress) + r ij [Equation 3'

where time and stress were group-mean centered. Again, in this equation, Oij represents

the slope of a spouse's satisfaction over the first years of marriage. 02j captures the

within-person association between spouses' satisfaction and their stress. In other words,

32j is an estimate of stress spillover over the first years of marriage.








To determine the association between coping with early external stress and

marital satisfaction over time, the following equation was then estimated at the between-

subjects level of the HLM analysis:

Pij = Yo+ yn (TI Stress)+ Y12 (TI Cognitive Organization)

+ Y13 (TI Stress X T 1 Cognitive Organization)+ gij [Equation 20]

where stress and cognitive organization were grand-mean centered. In this equation, y7o

represents an estimate of the average slope of spouses' marital satisfaction over time. yI

represents the association between spouses' Time 1 stress and the slope of their

satisfaction. Thus, a negative y, 1 would indicate that spouses with the highest level of

stress at Time 1 also experienced the greatest declines in satisfaction over time. 122

represents the association between spouses' Time 1 cognitive organization and the slope

of their satisfaction. For causality and responsibility attributions, a negative 712 would

indicate that spouses with the poorest cognitive organization at Time 1 also experienced

the greatest declines in satisfaction over time. For differential importance, apositiie y12

would indicate that spouses with the poorest cognitive organization at Time I also

experienced the greatest declines in satisfaction over time. 713 captures the association

between the interaction of stress and cognitive organization at Time 1 and the slope of

satisfaction over time. In other words, 713 represents the association between coping with

early stress and satisfaction over time. Finally, pi j is the residual variability in the slope

of satisfaction that remains to be explained after controlling for stress, cognitive

organization and the interaction of these two variables. This equation was estimated using

each measure of cognitive organization and the significance of the average y13 term was

investigated.








Results revealed only one significant main effect for husbands. Namely,

husbands' Time 1 differential importance index was associated with the stability of their

satisfaction over time, such that a more positive organization was associated with less

decline in satisfaction over the early years of marriage, Y12= .83, SE = .42, t (79) = 2.0,p

=.05, effect size r = .22. No significant main effects were found for wives. Turning to

the interaction terms, results indicated that the interaction between stress and causality

attributions was not significantly associated with the slope of satisfaction over time for

husbands or for wives, y13 = .002, SE = .005, t (78) = .44, p = .66, effect size r = .002 and

Y13 = .009, SE= .008, t (78) = 1.09, p = .28, effect size r = .12, respectively. However,

the interaction between stress and responsibility attributions was significantly associated

with the slope of satisfaction over time for husbands, Y13 = -.01, SE = .006, (78) = -1.9, p

= .05, effect size r = .21. This interaction was plotted using stress scores and

responsibility attribution scores that were one standard below the mean and one standard

deviation above the mean. As seen in Figure 1, husbands who maintained a more

positive cognitive organization in the face of stress at Time 1 (i.e., tended not to blame

their partners for negative behaviors) also experienced less decline in their satisfaction

over time. In other words, for husbands, adaptively coping with stress by maintaining a

positive cognitive organization was associated with more stable satisfaction over time

than was unsuccessfully coping with that stress. Moreover, the size of this effect appears

to be larger under conditions of low stress than under conditions of high stress. The type

of attributions husbands made under low stress seemed to have a large influence on future

satisfaction. Namely, the ability to make positive attributions under low stress seemed to

contribute to the maintenance of satisfaction over time. However, under high stress, the






04

smaller protective effect on future satisfaction. In other words, attributions appeared to

play less of a role when husbands were experiencing high stress. This interaction was not

significant for wives, 713 = .006, SE = .004, t (78) = 1.7, p = .08, effect size r = .19.

Finally, the interaction between stress and differential importance was not significantly

associated with the slope of satisfaction for husbands or for wives, 713 = -.22, SE = .25, t

(78) = -.90,p = .37, effect size r = 10 and 713= -.26, SE= .33, t (78) = -.80,p = .42,

effect size r = .09, respectively.

To determine the association between coping with early external stress and stress

spillover over time, the following equation also was estimated at the between-subjects

level:

2j = y2 + y21 (TI Stress)+ 722 (TI Cognitive Organization)

+ 723 (TI Stress X TI Cognitive Organization)+ g.2j [Equation 21]

where stress and cognitive organization were grand-mean centered. In this equation, 720

represents an estimate of the average stress spillover effect. 721 represents the association

between spouses' Time 1 stress and their stress spillover. Thus, a negative 721 would

indicate that spouses with the highest level of stress at Time 1 also experienced the

greatest stress spillover over time. 722 represents the association between spouses' Time 1

cognitive organization and their stress spillover. For causality and responsibility

attributions, a negative 722 would indicate that spouses with the poorest cognitive

organization at Time 1 also experienced the greatest stress spillover overtime. For

differential importance, a positive 722 would indicate that spouses with the poorest

cognitive organization at Time 1 also experienced the greatest stress spillover over time.







65

Time 1 and stress spillover. Finally, ut2j is the residual variability in stress spillover that

remains to be explained after controlling for stress, cognitive organization and the

interaction of these two variables. This equation was estimated using each measure of

cognitive organization and the significance of the average y23 term was investigated.

Results revealed one significant main effect for husbands. Contrary to

expectations, husbands with the most negative responsibility attributions at Time 1

tended to experience the least stress spillover over time, '22 = .02, SE = .01, 1 (79) = 2.4, p

= .02, effect size r = .26. One marginally significant main effect was found for wives.

There was a trend for wives with the most negative causality attributions for their

partners' negative behavior at Time 1 also to experience the most stress spillover over

time, Y22 = -.03, SE = .02, t (79) = -1.9,p = .06, effect size r = .21. Turning to the

interaction terms, results indicated that the interaction between stress and causality

attributions was not significantly associated with stress spillover over time for husbands,

Y23= -.002, SE= .001, t (78) = -.71,p = .48, effect size r = .08. However, this interaction

was significant for wives, Y23 = .005, SE = .02, t (78) = 2.3, p = .02, effect size r = .25.

This interaction was plotted using stress scores and causality attribution scores that were

one standard deviation below the mean and one standard deviation above the mean.

Consistent with the previous results for husbands, attributions tended to play a stronger

role under conditions of low stress than under conditions of high stress. As seen in

Figure 2, wives who maintained a more positive cognitive organization in the face of low

stress at Time I (i.e., tended not to see their partners as the cause of their negative

behaviors) also experienced less stress spillover over time. Thus, coping well with low

stress at Time 1 seemed to have a protective effect on future spillover. However, wives'







66
Time 1 cognitive organization under conditions of high stress did not have this protective

effect. Under high stress, wives, initial coping seemed to have little effect on future

stress spillover. Importantly, though there appears to be a significant main effect of Time

1 stress on future spillover, this main effect was not significant. The interaction between

responsibility attributions and stress was not significantly associated with stress spillover

for husbands or for wives, y23 = .001, SE = .002, t (78) = .34, p = .73, effect size r = .08

and Y23 = .003, SE = .002, t (78) = 1.3, p =. 18, effect size r =. 14, respectively. Finally,

the interaction between stress and differential importance was not significantly associated

with stress spillover for husbands or for wives, y23 = .03, SE = .16, t (78) = .18,p =.85,

effect size r = .02 andy23=. 12,SE=.13, t (78) = .97, p = .33, effect size r=.11,

respectively. Overall, then, only some support was found for the hypotheses that initial

coping is associated with future satisfaction and stress spillover.

Hypotheses 4c: Initial Coping and Future Coping

Hypothesis 4c predicted that successful coping with stress early in the relationship

should serve to bolster intimates' ability to successfully cope with stress in the future. In

other words, spouses who maintain a positive organization of relationship perceptions in

the face of external stress should be even more likely to maintain a positive organization

when faced with stress later in the relationship. To address this hypothesis, the following

equation was estimated at the between-subjects level using hierarchical regression

analyses:

T7 Cognitive Organization = p3oj + P3ij (T7 Stress)+ 32j (TI Stress)+03j (TI

Organization)+ 34j (TI Stress X TI Organization)+error [Equation 22]

This equation was estimated using each measure of cognitive organization.







67
With regard to causality attributions, results revealed no significant main effects

of Time 7 stress on Time 7 causality attributions for husbands or for wives, 31j = -.53, SE

=.64, t (81) = -.84,p = .41, effect size r = .11 and P3j = .004, SE= .74, t (54) = -.06,p =

.96, effect size r = .01, respectively. A significant main effect of Time 1 stress on Time 7

cognitive organization was found for husbands, 132j = 1.2, SE = .48, t (54) = 2.4, p =.02,

effect size r= .31, though not for wives, 32j = .77, SE =.49, t (54)= 1.6,p = 12, effect

size r =. 17. Thus, husbands experiencing higher stress at Time 1 also exhibited more

negative causality attributions at Time 7. Moreover, a significant main effect of Time 1

causality attributions on Time 7 causality attributions for both husbands and wives, 133j =

.48, SE = .17, t(54) = 2.9,p = .006, effect size r = .21 and 13j = .49, SE = .18, t (54) =

2.8, p = .007, effect size r = .21, such that spouses with more negative causality

attributions at Time 1 also reported more negative causality attributions at Time 7.

Contrary to our hypothesis, the interaction between Time 1 causality attributions and

Time 1 stress was not significantly associated with Time 7 causality attributions for

husbands or for wives, 04j = -.007, SE = .05, t (52) = -.84,p = .40, effect size r = .12 and

R =.. _nAAA V= n-- ft(V =-1 A nr = 1 Aff'i aei;pa r= 10 rienPtiwulv








SE= .06, t(52) = .83, p = .41, effect size r = .11 and 04j =-.004,SE= .05, t(52) = -.89,p

= .38, effect size r = .12, respectively.

Finally, with regard to differential importance, no main effects on Time 7

differential importance were found for husbands. For wives, however, there was a

significant main effect of Time 7 stress on Time 7 differential importance, such that

wives reporting more stress at Time 7 also demonstrated a less positive cognitive

organization at Time 7, pj = -.06, SE = .03, t (52) = -2.0, p = .05, effect size r =.27. In

addition, there was a marginal main effect of Time 1 stress on wives' Time 7 differential

importance index, such that wives reporting more stress at Time 1 demonstrated a

tendency to maintain a less positive cognitive organization at Time 7, :zj = -.03, SE =

.02, t (52) = -1.9,p = .06, effect size r = .25. Nevertheless, the interaction between Time

1 stress and Time 1 differential importance was not significantly associated with Time 7

differential importance for husbands or for wives, 04j = -.20, SE= .14, t (50) = -1.4, p =

.16, effect sizer= .14, and 14j =-.13, SE= .08, t (50) = -1.6,p= .12, effect sizer= .22,

fA ,.,,w44..l.r +I.. -- *l, -, *, *.wi.,,,-.. ..- , C-k-,.. ..,,A ^t*A tP- ,fi .-4 ha aAae%+r^ ar-, "








Table 1

Mean of Global Marital Ouality and Specific Marital Problem Scores Across Seven
Waves of Measurement for Husbands and Wives


Spouse Time I Time 2 Time 3 Time 4 Time 5 Time 6 Time 7

Global Marital Satisfaction
Husbands
M 96.3 92.0 92.5 92.1 93.5 92.1 91.1
SD 8.8 14.1 14.8 14.7 13.9 15.5 16.9
N 81 76 74 67 64 59 60
Wives
M 97.7 94.8 93.3 92.1 93.8 90.0 89.1
SD 10.7 12.9 16.0 14.7 15.6 19.4 19.6
N 82 77 73 68 66 61 62

Inventory of Specific Marital Problems
Husbands
M 31.3 30.1 29.7 29.8 29.9 30.4 31.3
SD 16.2 17.1 15.8 17.2 17.3 17.7 20.0
N 82 75 74 67 64 59 60
Wives
M 29.2 27.8 30.5 29.1 27.4 30.8 31.5
SD 17.3 15.8 18.7 17.1 14.5 18.1 18.7
N 82 76 73 66 66 61 62
Note: For the Inventory of Specific Marital Problems, higher scores indicate a more
negative view of the relationship.









Table 2

Mean of Attribution Scores and the Differential Importance Index Across Seven Waves
of Measurement for Husbands and Wives


Spouse Time I Time 2 Time 3 Time 4 Time 5 Time 6 Time7

Causal Attributions
Husbands
M 42.9 46.8 45.9 45.2 46.1 44.5 43.5
SD 10.9 10.4 10.5 11.4 10.3 13.2 13.8
N 82 75 64 58 58 53 54
Wives
M 44.9 45.4 46.3 46.2 45.9 47.3 47.9
SD 9.8 10.2 10.2 10.0 11.3 12.4 11.9
N 82 76 64 63 62 55 55















T

































M- U- AMI/M


r//ff,/ff/


0.5-


-1-


-24-'


I Negative
Organization

* Positive Organization


[,'////////., -/


I AIII* ;NMI t'X% n lI U11.M















Spillover



-0.5,




-1


[] Negative Organizatioi

[] Positive Organization







More
Stress
Spillover -2.5 -
Low Stress High Stiess




Figure 2. The Interaction of Stress and Causality Attributions on Stress Spillover for
Wives.














DISCUSSION

Study Rationale and Summary of Results

Part of maintaining a close relationship over time involves navigating the negative

life events external to the relationship, such as work stress or financial stress, that may

nevertheless strain the relationship. Understanding change and stability in relationship

satisfaction may therefore require an understanding of the broader context in which the

relationship is embedded. In fact, ample research has linked the external stressful

circumstances surrounding a marriage to relationship outcomes. This research has

demonstrated that individuals' stress may have a detrimental effect on their relationship

evaluations, a phenomenon known as stress spillover (Tesser & Beach, 1998). Moreover,

individuals' stress may have negative consequences for their partners' relationship

evaluations, a phenomenon known as stress crossover (Thompson & Bolger, 1999). This

dissertation attempted to further our understanding of the association between external

stress and relationship quality by addressing three important limitations of the existing

stress spillover/crossover literature.

First, the existing literature on stress and relationship quality has failed to

investigate the potential mechanisms through which negative external stressors may

affect individuals' relationship satisfaction. We predicted that external stress should

influence spouses' satisfaction by affecting spouses' specific relationship beliefs, as well

as the organization of those beliefs. In other words, the experience of external stressors

should provide individuals with more negativity to deal with in the relationship and








attect individuals' aDility to suosequenuy cope wim mis increase in negative reiauniusmp

perceptions. The first goal of the study, then, was to examine the role of spouses'

cognitive content in the stress spillover process. In pursuit of this goal, Hypothesis 1

stated that external stress should affect spouses' relationship satisfaction through its

effects on spouses' specific relationship perceptions. This prediction was supported for

wives, though not for husbands. Replicating and expanding previous research on stress

spillover, which has demonstrated the effects of stress spillover over the course of several

days, the current study found evidence for stress spillover over the first 3 / years of

marriage. Within-subjects analyses revealed that increases in wives' stress were

associated with decreases in their marital satisfaction over time. Moreover, wives'

specific relationship perceptions fully mediated this association. Wives' stress was

associated with the negativity of wives' specific relationship perceptions, such that as

wives' stress increased, they also tended to perceive more specific problems in the

relationship. This association seemed to account for the stress spillover effect. Thus, the

current findings suggest that one way in which external stress may lead to declines in

satisfaction is by increasing the negativity of spouses' cognitive content.

The second goal of the study was to examine the role of spouses' cognitive

organization in the stress spillover process. Hypothesis 2 stated that external stress also

should affect spouses' relationship satisfaction through its effects on spouses' cognitive

organization, controlling for spouses' cognitive content. This hypothesis was partially

supported for wives. Specifically, Hypothesis 2 predicted a curvilinear relationship

between stress and cognitive organization, such that as stress increased from low to

moderate, stress was not expected to interfere with spouses' ability to organize their

specific perceptions. Thus, as stress increased from low to moderate, spouses' cognitive








organization was expected to become more relationship-enhancing. However, as stress

increased from moderate to high, we predicted that coping with stress would tax

intimates' cognitive resources, and therefore interfere with spouses' ability to cope with

negative specific perceptions of the relationship. Thus, as stress increased from moderate

to high, spouses' cognitive organization was expected to become less relationship-

enhancing. Evidence for a curvilinear relationship between stress and cognitive

organization was not found on any measure of cognitive organization. However, further

analyses provided some evidence for a linear relationship between stress and cognitive

organization, such that increases in stress were associated with a less relationship-

enhancing cognitive organization. For wives, increases in external stress were associated

with an increased tendency to make blaming attributions for a partner's behavioral

transgressions over the first years of marriage, even when controlling for the general

negativity of wives' specific relationship perceptions. In other words, wives' stress was

associated with the nature of their responsibility attributions, independent of wives'

cognitive content. Moreover, wives' responsibility attributions fully mediated the stress

spillover process. Thus, in line with the second prediction of Hypothesis 2, some

evidence indicated that stress may lead to declines in satisfaction by limiting spouses'

ability to separate their negative specific relationship perceptions from their global

relationship satisfaction.

Overall, then, results revealed modest support for the proposed model suggesting

that external stress may affect marital satisfaction through two general routes. First, stress

was predicted to affect spouses' specific perceptions of the relationship. In fact, strong

support was found for the role of cognitive content in the stress spillover process.

Second, stress was predicted to affect the structure of spouses' specific relationship







7'
perceptions. Evidence for the role of cognitive organization in the stress spillover

process, though, was somewhat weaker. The current data provided no evidence for the

predicted curvilinear relationship between stress and cognitive organization. One reason

for the failure to find this predicted association may involve a lack of statistical power.

Though the current study measured seven waves of data, only 62 spouses provided data

for at least four of the seven time points. Consequently, the current study had low power

for detecting a curvilinear relationship. Moreover, though there was respectable

between-subjects variability in stress scores, there may have been a restricted range in

within-subjects stress scores over time. Reliability analyses across the seven waves of

stress data revealed a high Cronbach's alpha (ao = .81), suggesting that spouses may not

have been experiencing much variability in their acute stress between assessments. A

sample of spouses experiencing more variability in their stress over time may have

revealed the expected curvilinear relationship.

A second reason for the failure to find the predicted curvilinear association may

be theoretical. We predicted that under low to moderate levels of external stress, not only

would spouses retain the cognitive resources necessary to reorganize their specific

perceptions, but also that spouses' cognitive organization actually may become more

positive as stress increased from low to moderate. This prediction was based on research

demonstrating that cognitive organization effects are strongest when the content of

beliefs is somewhat negative (Showers et al., 1998). In other words, in order for spouses

to structure their beliefs in a relationship-enhancing manner, they first must hold negative

specific perceptions of the relationship. We suggested that under low stress, there would

be little negativity in the relationship, resulting in weak cognitive organization effects.

This Dersoective incorrectly assumes that eyteRnmal qtrPc rPnrPcPntc the nnhlw enrtrp nrf






/5

negative specific relationship perceptions. However, many other factors, such as spouses'

enduring vulnerabilities (e.g., neuroticism), may contribute to the experience of

negativity in the relationship (Karney & Bradbury, 1995). Even spouses under low

stress, then, may have negative relationship perceptions that must be integrated within a

generally positive framework of relationship beliefs. Consequently, if negativity is

already present in the relationship, spouses' cognitive organization would be unlikely to

improve as their stress increased. Rather, increases in stress may simply interfere with

spouses' ability to organize their perceptions in a relationship-enhancing manner. The

linear relationship between stress and cognitive organization found in the current study

supports this alternative conceptualization.

In fact, though a curvilinear relationship was not found, the finding that

responsibility attributions mediate the stress spillover effect is encouraging. Previous

research has theorized that attributions may play an important role in the stress

spillover/crossover process (Tesser & Beach, 1998; Thompson & Bolger, 1999). The

current study provides the first empirical evidence suggesting that when individuals may

attribute their negative relationship perceptions to an external source, such as the stressful

situation, they may limit the association between stress and their overall relationship

evaluation. However, high stress seemingly interferes with individuals' ability to

maintain positive attributions for negative perceptions, and thus is associated with greater

stress spillover. Consequently, the current results suggest that further research on this

issue is warranted.

For instance, future research may want to examine the association between stress

and other cognitive organizational strategies that may be more central to the stress

spillover process. Some research has argued that individuals vary in the complexity with







79
which they organize their specific relationship beliefs (Murray & Holmes, 1999; Showers

& Kevlyn, 1999). In a compartmentalized organization, negative specific beliefs are

lumped together and separated from positive specific beliefs. Thus, the activation of any

one negative belief may lead to the activation of a flood of negative beliefs, possibly

lowering judgments of satisfaction. Conversely, in a complex, integrative organization,

negative specific beliefs are linked to positive specific beliefs. Accordingly, the

activation of a negative belief will bring to mind other positive beliefs, and thus minimize

the influence of the negative belief on judgments of satisfaction. This type of complex

organization has been found to buffer individuals from the negative effects of stress

(Showers & Kling, 1996). Future research may want to examine whether stress also may

limit the complexity with which intimates organize their specific relationship beliefs, thus

increasing the likelihood of stress spillover.

A second limitation of the existing literature on stress and relationship quality is

the failure to investigate whether the stress of each spouse may interact to influence

individuals' relationship satisfaction. In other words, research on stress crossover has

examined how spouses' stress may affect their partners' satisfaction without regard for

the amount of stress partners may be experiencing themselves. Hypothesis 3 predicted

that partners' own stress would moderate the stress crossover effect. This prediction was

not supported as the current study failed to find evidence of a stress crossover effect.

Namely, changes in wives' stress over the first years of marriage were not associated

with changes in their husbands' satisfaction. Similarly, though initial evidence indicated

a significant association between husbands' stress and wives' marital satisfaction, this

association did not remain significant when controlling for the association between

wives' own stress and their satisfaction. In other words, the stress crossover effect found








for wives was not independent of wives' stress spillover effect, indicating that the stress

crossover effect simply may have resulted from the correlation between husbands' and

wives' stress.

There are two possible reasons why the current study failed to replicate the stress

crossover effect. First, the finding that husbands' stress did not have an independent

association with wives' satisfaction is not completely surprising given that evidence for

stress spillover was not found for husbands. Theories of stress spillover and crossover

suggest that individuals' stress should affect their own thoughts and behaviors. This

change in individuals' own thoughts and behaviors should result in changes in the

thoughts and behaviors of their partners. In other words, to affect the relationship

partner, stress must first affect the individual. Consequently, given that husbands' stress

was not found to affect their own relationship evaluations, it seems unlikely that

husbands' stress would affect the relationship evaluations of their wives. Second,

previous research on stress crossover has not examined the effects of stress crossover

independent of stress spillover. Thus, the possibility remains that previous findings of

stress crossover really are simply further evidence of stress spillover effects. Further

research is necessary to determine whether stress crossover effects truly are independent

of stress spillover effects.

Finally, the existing literature on stress and relationship quality has failed to

examine the long-term effects of stress on relationship satisfaction. Namely, this

literature has failed to distinguish why for some couples, stress may be detrimental for

the relationship, while for other couples stress may actually lead to enhanced relationship

functioning. Hypothesis 4 suggested that spouses who successfully adapt to stress early in

the relationship should exhibit resilience to future stressful experiences. Specificallv.y, we







81

first predicted that spouses who maintained a positive organization of relationship

perceptions in the face of external stress would be less vulnerable to declines in

relationship satisfaction over time. Some evidence confirmed this prediction. Namely,

husbands who adaptively coped with stress by maintaining less blaming attributions of

their wives' behavioral transgressions (i.e., a more positive cognitive organization) also

tended to maintain more stable satisfaction over time than did husbands who

unsuccessfully coped with that stress. In addition, the size of this effect appeared to be

larger under conditions of low stress than under conditions of high stress, suggesting that

positive attributions may not be as effective in protecting relationship satisfaction when

husbands are experiencing high levels of stress.

Second, we predicted that spouses who maintained a positive cognitive

organization in the face of external stress would be less vulnerable to future stress

spillover. Again, some evidence confirmed this prediction. Wives who maintained a

more positive cognitive organization in the face of low stress early in the relationship

(i.e., tended not to see their partners as the cause of their negative behaviors) also

experienced less stress spillover over time. Thus, coping well with low stress seemed to

have a protective effect on future stress spillover. However, consistent with the previous

finding, positive attributions seemed to have little effect on future stress spillover under

conditions of high stress. Finally, we predicted that adaptive coping early in the

relationship would serve to bolster spouses' ability to cope with stress encountered later

in the relationship. No support was found for this prediction, as spouses' early coping

was not found to be associated with their future coping.

Overall, support for our hypotheses of stress resilience was modest at best.

Results were not found consistently across spouses or across measures of cognitive






82
organization. Nevertheless, when significant results were found, the pattern of results was

notably similar. In line with our theory, coping well with low stress (i.e., maintaining

positive attributions) did seem to lead to future stress resilience. Moreover, consistent

with the idea that high stress may overwhelm spouses' ability to prevent stress from

affecting their relationship evaluations, maintaining positive attributions in the face of

high stress did little to protect spouses from the adverse affects of stress. In other words,

when faced with high external stress, spouses' cognitive organization simply didn't seem

to matter as much.

Given this consistent pattern of results, further research on stress resilience seems

warranted. In particular, future research may want to examine alternative methods of

coping with stress. The current study operationalized successful coping with stress as

maintaining a positive cognitive organization. However, given that only modest support

was found for an association between stress and cognitive organization, using cognitive

organization as a proxy for coping may not have been the best method for examining

issues of stress resilience. Rather, the manner in which spouses behave under conditions

of stress may have an influence on future susceptibility to stress. Alternatively, future

research may want to examine whether individuals who successfully prevent stress

spillover early in the relationship are better able to prevent stress spillover over the course

of the relationship. For instance, a longitudinal diary study examining the association

between stress and satisfaction early in the relationship then again several years later may

illuminate the advantages of successful coping.

Strengths and Limitations of the Study

Our confidence in the results of this study is enhanced by a number of strengths in

its methodology and design. Foremost among these was the use of within-subjects






83

analyses to examine the associations between stress and relationship cognitions. Within-

subjects analyses allowed for the estimation of the association between changes in stress

and changes in relationship cognitions, controlling for spouses' stable tendencies to view

their stress and their relationship in a particular manner. Namely, within-subjects

analyses allowed us to control for potentially confounding between-subjects variables

such as marital satisfaction, stress level, and general negative affectivity, enabling us to

limit the possibility that the association between stress and relationship cognitions was

the result of these variables. Second, when examining the role of cognitive content and

cognitive organization in the stress spillover process, the HLM approach allowed for the

estimation of the association between stress and cognitive organization, controlling for

the influence of cognitive content, ensuring that these parameters were not confounded.

Third, in contrast to prior research that has relied almost exclusively on short-term diary

data, the current study used longitudinal data that allowed us to address whether stress

and relationship cognitions are associated over the course of the first 3 i/ years of

marriage. Fourth, also in contrast to much prior research that has addressed samples

varying widely in marital duration, the analyses reported here examine data from a

relatively homogeneous sample of couples, reducing the likelihood that the effects

observed here result from uncontrolled differences in marital duration. Moreover, the use

of a fairly homogeneous sample provided a more conservative test of our hypotheses.

Despite these strengths, several factors nevertheless limit interpretations of the

current findings. First, all of the data examined here were correlational. The current

paper suggests that spouses' stress should lead to changes in their relationship

evaluations. However, these data cannot rule out the alternative perspective that the

nature of spouses' marriages may lead to changes in the amount of external stress they






84

experience. Nevertheless, this interpretation seems less likely for two reasons. First, all

of the events listed on our measure of acute stress were chosen to represent stresses that

are not likely to be a consequence of marital satisfaction. For instance, whereas being

hospitalized or the death of a family member may affect spouses' satisfaction, the reverse

is less likely to be true. Second, the majority of the events on the measure represent

concrete, objective events. Thus, troubles in the marriage are unlikely to lead spouses to

simply perceive more external stress in their lives. In other words, having a bad marriage

is unlikely to lead spouses to perceive that they were fired from their job or that their

application to school was rejected if these events did not actually occur.

In line with this reasoning, a second limitation of the current study is the use of

self-report measures of stress. The use of self-report measures opens up the possibility

that third variables may be affecting spouses' views of the relationship as well as

spouses' perceptions of stress. For instance, spouses high in neuroticism may exhibit a

stable tendency to view their relationship and their stress more negatively, leading to a

spurious association between these two variables. Again, the current data cannot rule out

the possibility that a third variable, such as neuroticism, may account for the associations

between stress and relationship cognitions. However, for reasons mentioned above, this

interpretation seems less likely. As mentioned, the use of within-subjects analyses

allowed us to partial out spouses' stable tendencies to view their stress and their

relationship in a particular manner. Moreover, when measuring stress, the current study

did not rely on spouses' subjective ratings of the negativity of the stressful event. Rather,

the focus was on whether the spouse reported that the event had occurred. Again, as

mentioned, given that the stress measure tended to tap concrete, objective events, it seems








stressors if those events did not actually occur. Nevertheless, future research may want to

examine issues of stress spillover and crossover using objective measures of stress, such

as interviewer ratings of stress. The use of objective stress measures may help clarify the

directional link between stress and relationship processes.

Though also an important strength of the current research, a third limitation to the

current study involves the use of a relatively homogeneous sample of satisfied couples.

The current study was limited in the range of satisfaction and stress scores being

reported. Thus, generalizations to other samples should be made with caution. For

instance, research examining less satisfied couples may find different results that the ones

reported here. The newlywed spouses in the current sample likely were motivated to

maintain the positivity of their overall satisfaction. For couples that do not have the same

motivation to perceive the relationship positively, stress may have an even stronger effect

on relationship cognitions. However, the fact that stress was significantly associated with

spouses' relationship cognitions even in this sample of uniformly happy couples serves to

enhance our confidence in these findings.

Finally, although our sample size compared favorably to other longitudinal

studies of marriage, a larger sample size with additional waves of data would have

provided greater power to detect additional effects not detected in the current study. For

example, in a study with a larger sample size, the interactive effects of spouses' stress on

relationship satisfaction may have been significant. In addition, a larger sample may

have revealed further evidence of stress resilience. Nevertheless, the fact that several of

our predictions were supported, despite the conservative nature of our tests, suggests the

current findings are robust.








Marriage as a Safe Haven: The Successful Adaptation to Stress

The current paper argues that stressful circumstances external to the marriage may

have detrimental effects on cognitive processes within the marriage. Namely, spouses

who are experiencing a lot of stress in their lives also tend to view their relationships in a

negative light. Successfully adapting to stress, then, involves preventing that stress from

negatively affecting judgments of the relationship, or preventing stress spillover.

However, other theorists take a different perspective, arguing that stress may not always

lead to negative relationship perceptions. Rather, a good marriage may provide a source

of comfort when external circumstances are difficult (Brunstein, Dangelmayer, &

Schultheiss, 1996; Coyne & DeLongis, 1986). When faced with a number of stressors,

spouses may contrast their marriage against their stressful external circumstances, leading

them to view their relationships more positively than ever (cf. Bless & Schwartz, 1998).

For instance, a man experiencing stress at work may begin to more fully appreciate the

warmth and stability of his marriage. In other words, his marriage may come to represent

a "safe haven" from the turmoil and stress he encounters at the office. Thus, in contrast to

the stress spillover perspective, which predicts a negative association between stress and

marital satisfaction, the safe haven perspective argues that as external stress increases,

satisfaction with the marriage may also increase.

Though these two perspectives appear contradictory, the model estimated in the

current study could in fact address each of these stress effects. Results indicated that, on

average, spouses' stress seemed to spillover into their marriage, as increases in stress

were associated with decreases in marital satisfaction. In the current study, significant

variability in the strength of this association was not found. However, in a larger study,

results may indicate that though stress has negative effects on average, for some, stress






87

may have a positive association with marital satisfaction. If this is the case, future

research may want to examine sources of between-subjects variance for the within-

subjects association between stress and marital satisfaction. In other words, what

distinguishes those experiencing stress spillover from those for whom the marriage serves

as a safe haven from stressful experiences?

One promising answer to this question may be the overall quality of spouses'

lives. Spouses who enjoy an overall positive quality of life tend to be less affected by

external acute stressors than spouses who are faced with a number of chronic life

stressors, such as living in poverty or coping with a long-term illness (Caspi et al., 1987).

Perhaps spouses with low chronic stress not only experience less stress spillover, but also

are able to rely on the marriage as a safe haven from the experience of acute stress. A

second answer to this question may be spouses' trait cognitive complexity. Individuals

who define themselves using a greater number of independent self-aspects tend to be less

affected by negativity in any one domain than individuals with less complex views of the

self (Linville, 1987). In other words, when spouses hold a greater number of

differentiated self-aspects, stress in one domain, such as work, tends to be less likely to

spill over to affect thoughts and feelings of a different domain, such as the marriage.

Again, future research may want to examine whether this type of self-complexity not

only limits stress spillover effects, but also allows individuals to compensate for the stress

felt in one domain by focusing on other positive self-domains.

Two Routes to Change in Satisfaction: Expanding the Model

The hypotheses of the current paper were based on the idea that there may be two

general routes to declines in satisfaction. The first route involves a change in intimates'

cognitive content, while the second route involves a change in intimates' cognitive








structure. Throughout this paper, these two routes were discussed as if they were parallel

processes. However, future research may want to question this assumption. For instance,

there may be a temporal order of the effects of stress on cognitive content and cognitive

organization. If evidence of a curvilinear relationship between stress and cognitive

organization is found, this could suggest that stress may first affect cognitive content,

then affect cognitive structure. Namely, as stress increased from low to moderate, stress

would affect spouses' specific perceptions, yet not interfere with the organization of

those perceptions. However, as stress continued to increase from moderate to high, stress

would then also begin to affect spouses' organizational abilities. However, if further

research corroborates that the relationship between stress and organization is linear, this

may indicate that the effects of stress on content and on organization are simultaneous.

Similarly, expanding the model even further, there may be a temporal order to the

types of cognitive organizational strategies spouses use for coping with negative

relationship perceptions (e.g., Robin & Beer, 2001). For instance, when faced with a

negative specific perception of a relationship partner, spouses may first attempt to cope

with that perception though the use of relationship-enhancing attributions. Viewing the

partner as not responsible for a behavioral transgression allows the spouses to separate

that negative perception from an overall relationship evaluation, thereby maintaining a

positive view of the relationship. This strategy, however, may only be effective when

coping with isolated, negative events. Attributions are unlikely to be successful in coping

with particularly salient or pervasive negative relationship perceptions that are

encountered repeatedly over time. In this case, spouses may need to move to a different

coping strategy, such as the use of differential importance. Namely, once it is no longer








spouses may choose to limit the negativity of the perception by dismissing the

importance of the negative act. Thus, our understanding of marital stability may benefit

from further investigation of both when and how various cognitive strategies may

successfully protect marital satisfaction from the implications of specific negative beliefs

Additional Directions for Future Research

The current study examined two intervening variables in the stress spillover

process: spouses' cognitive content and spouses' cognitive organization. These particular

variables were chosen because most factors that shape the development of a relationship

exert their influence on future relationship outcomes through their effects on how

individuals think about the relationship (Karney, McNulty, & Frye, 2001). However, a

variety of other possible intervening variables were not explored in this study. For

instance, prior research has established links between stress and negative mood (Bolger,

DeLongis, Kessler, & Schilling, 1989) and between stress and negative behaviors

(Bolger, DeLongis, Kessler, & Wethington, 1989; Repetti & Wood, 1997). Future

research is needed to piece together the large literature on stress and to untangle the

causal chains linking each of the potential intervening variables of the stress spillover

process. For example, one possibility is that stress may first lead to negative mood,

which results in changes in behaviors. These changes in behaviors may then lead to

changes in spouses' specific relationship cognitions, which in turn affect spouses' global

relationship evaluations. Most likely, however, a number of reciprocal relationships may

exist between these variables. Overall, then additional research is necessary to construct

a more complete picture of the intervening variables underlying the stress spillover

process.







90

Our understanding of the stress spillover process may also benefit from

alternative measurements of spouses' exposure to stressors. The current study examined

a composite of spouses' stress across a variety of life domains, such as work, health, and

family. Future research may want to consider whether stress spillover may be moderated

















phenomena.














APPENDIX A
SEMANTIC DIFFERENTIAL MEASURE OF MARITAL SATISFACTION

For each of the following items, fill in the circle (0) that best describes HOW YOU
FEEL ABOUT YOUR MARRIAGE. Base your responses on your first impressions and
immediate feelings about the item.


INTERESTING 0000000 BORING
BAD 0000000 GOOD
UNPLEASANT 00 0 00000 PLEASANT


SATISFIED C
LONELY C
STURDY C

REWARDING (
DISCOURAGING (
ENJOYABLE (
TENSE (

STABLE (
HAPPY (
STRESSFUL (


)0 0 0 0 DISSATISFIED
)0 0 0 0 FRIENDLY
)0 000 FRAGILE

)0 0 0 0 DISAPPOINTING
)0 0 0 0 HOPEFUL
)0 0 0 0 MISERABLE
)0 0 0 0 RELAXED

)0 0 0 0 UNSTABLE
)0 0 0 0 SAD
)0 0 0 0 PEACEFUL














APPENDIX B
INVENTORY OF SPECIFIC MARITAL PROBLEMS

























r AXTri ADTC