Citation
"Should I expect the best or brace for the worst?" : the effects of newlyweds' expectations on their future relationship satisfaction

Material Information

Title:
"Should I expect the best or brace for the worst?" : the effects of newlyweds' expectations on their future relationship satisfaction
Portion of title:
Effects of newlyweds' expectations on their future relationship satisfaction
Creator:
McNulty, James K., 1975-
Publication Date:
Language:
English
Physical Description:
vi, 101 leaves : ; 29 cm.

Subjects

Subjects / Keywords:
Covariance ( jstor )
Diaries ( jstor )
Husbands ( jstor )
Marital satisfaction ( jstor )
Marriage ( jstor )
Personality psychology ( jstor )
Social psychology ( jstor )
Spouses ( jstor )
Trajectories ( jstor )
Wives ( jstor )
Dissertations, Academic -- Psychology -- UF ( lcsh )
Psychology thesis, Ph. D ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2001.
Bibliography:
Includes bibliographical references (leaves 93-100).
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by James K. McNulty.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
027737450 ( ALEPH )
48470608 ( OCLC )

Downloads

This item has the following downloads:


Full Text










"SHOULD I EXPECT THE BEST OR BRACE FOR THE WORST.'?
THE EFFECTS OF NEWLYWEDS' EXPECTATIONS ON THEIR FUTURE
RELATIONSHIP SATISFACTION
















By

JAMES K. MCNULTY


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE
UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2001














ACKNOWLEDGEMENTS

First off, I would like to thank the members of my committee, Benjamin Karney (chair), Jams Shepperd, Barry Schlenker, Mark Fondacaro, and James Algina, for the time and insights they have provided to this project. I would also like to thank my parents, James and Kathy McNulty. Without their continued support and encouragement, I would not be where I am today. Finally, I would like to thank my future wife Melissa for the overwhelming love and support she has provided me throughout my graduate career. I look forward to the many years of happiness we have ahead of us.














TABLE OF CONTENTS
page
ACKNOWLEDGMENTS ..................................................................... ii

A B STR A C T ..................................................................................... v

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

REVIEW OF LITERATURE ............................................................... 4

Expectancy Confirmation .......................................................... 4
Positive Expectations and Disappointment ......................................... 8
Reconciliation and the Distinction between
Global and Specific Expectations ......................................... 11
Overview of the Current Study ..................................................... 16
Hypotheses ........................................................................... 17

METHOD ..................................................................................... 23

Participants ........................................................................... 23
Procedure ................................................................................ 24
Materials .............................................................................. 25
Data Analysis ......................................................................... 30

RESULTS ..................................................................................... 32

Descriptive Statistics and Correlations .............................................. 32
Were Specific and Global Expectations Associated with
Attributions and Negative Behavior? .................................... 35
Were Specific and Global Expectations Associated with
Marital Outcomes Over Time? ........................................... 37
Did Positive Expectations Lead to Positive Changes in Marital
Satisfaction When They Were Likely to Get Confirmed? ............... 47

DISCUSSION ................................................................................ 69

Study Rationale and Summary of Results .......................................... 69
When Are Positive Expectations Likely to Get Confirmed? .................. 74
Additional Directions for Future Research ...................................... 76
Strengths and Limitations .......................................................... 79
Conclusion ........................................................................... 81








APPENDIX A


APPENDIX B APPENDIX C APPENDIX D APPENDIX E APPENDIX F APPENDIX G APPENDIX H APPENDIX I APPENDIX J


SEMANTIC DIFFERENTIAL MEASURE OF MARITAL SATISFACTION ................................... 82

QUALITY MARRIAGE INDEX ............................... 83

INVENTORY OF SPECIFIC MARITAL PROBLEMS........ 84 EXPECTATIONS FOR FURTURE RELATIONSHIP SATISFACTION ..................................................... 85

EXPECTED TRAJECTORY OF MARITAL SATISFACTION ................................................. 86

EXPECTATIONS FOR STABILITY .............................. 87

EXPECTATIONS FOR STEADINESS ........................ 88

EXPECTATIONS FOR PARTNER ............................... 89

EXPECTATIONS FOR SPECIFIC RELATIONSHIP PROBLEM S ....................................................... 90

RELATIONSHIP ATTRIBUTIONS MEASURE ............... 91


REFERENCES ................................................................................ 93

BIOGRAPHI CAL SKETCH ................................................................... 101














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

"SHOULD I EXPECT THE BEST OR BRACE FOR THE WORST.'
THE EFFECTS OF NEWLYWEDS' EXPECTATIONS ON THEIR FUTURE RELATIONSHIP SATISFACTION

By

JAMES K. MCNULTY

August, 2001

Chairperson: Benjamin R. Kamey
Major Department: Psychology

What are the implications of positive expectations about a relationship? The answer to this question may depend on whether the expectations get confirmed or disconfirmed. Positive expectations should lead to positive outcomes to the extent that they are confirmed through processes of expectancy confirmation, but may lead to negative outcomes if they are disconfirmed. One factor that may be associated with whether expectations get confirmed or disconfirmed is the specificity of the outcome. Expectations about global outcomes, because they are more open to interpretation and more likely to lead to confirmatory behaviors, should be more likely to get confirmed, leading to positive feelings about the relationship over time. In contrast, expectations about specific outcomes, because they are less open to interpretation and less likely to lead to confirmatory behaviors, should be more likely to get disconfirmed, leading to less positive feelings about the relationship over time. This dissertation addressed this possibility by examining the implications of positive expectations about relationship








outcomes that vary in specificity. Newly married couples reported their expectations for global and specific relationship outcomes and subsequently reported their satisfaction with the relationship every six months for two years. Contrary to the prediction that global and specific expectations would have different effects on the trajectory of marital satisfaction, results indicated that, on average, positive expectations about both global and specific outcomes led to declines in marital satisfaction. However, consistent with the idea that the association between expectations and change over time would depend on the likelihood that the expectations would get confirmed, objective and subjective measures of relationship quality moderated the association between expectations and change in satisfaction over time. When positive expectations were confirmed by a tendency to make positive attributions and engage in few negative behaviors, positive expectations led to stable marital satisfaction. In contrast, when positive expectations were disconfirmed by a tendency to make negative attributions and engage in frequent negative behaviors, positive expectations led to declines in marital satisfaction. The current findings suggest that a clear understanding of the role of expectations in relationships will be limited without taking into account the broader factors likely to be associated with the confirmation and disconfirmation of positive expectations.














INTRODUCTION

Things become better when you expect the best instead of the worst.

- Norman Vincent Peale, The power of positive thinking (p. 110)



Blessed is the man that expects nothing, for he shall never be disappointed.

- Alexander Pope



Contemporary wisdom suggests that the expectations partners have about their close relationships should play a role in determining whether their relationships will succeed or fail. Consistent with this idea, researchers have demonstrated that expectations influence a wide variety of interpersonal factors, including cognitions, behaviors, and emotions (Olson, Roese, & Zanna, 1998). Accordingly, what newlyweds expect from their relationships should affect how they ultimately feel about them.

Although both theory and research suggest that expectations should influence relationship outcomes, there is no consensus about the direction of the influence. Consistent with the idea that people who think about their relationship in positive terms should obtain positive outcomes, research on expectancy confirmation shows that people interpret events and behave in a manner consistent with their prior expectations (Darley & Fazio, 1980; Miller & Turnbull, 1986; Snyder, 1984). Thus, compared to those with more pessimistic expectations, people who expect positive things from their relationships should behave more constructively, interpret specific events more charitably, and








therefore be happier over time. In contrast, consistent with the idea that positive expectations may be detrimental to relationships, research on counterfactual thinking demonstrates that people with positive expectations can be disappointed when their outcomes do not meet their expectations (e.g., Mellers, Schwartz, Ho, & Ritov, 1997). Accordingly, partners with positive expectations about their relationships may react more negatively to specific relationship problems, and consequently grow less happy over time.

Understanding the factors that contribute to the functioning and stability of close relationships is important. Although all relationships begin very satisfying, most become less satisfying over time. Moreover, marital distress is the number one reason why people seek counseling (Veroff, Kulka, & Douvan, 1981). Yet, it remains unclear exactly what causes relationships to become unhappy and what may cause them to be happy again. Given the close connection between expectations and interpersonal outcomes, expectations frequently have been the target of interventions to promote healthy relationships (e.g., Baucom, Epstein, Rankin, & Burnett, 1996; Eidelson & Epstein, 1982; Markman, Stanley, & Blumberg, 1994; Olson, Fournier, & Druckman, 1986). Unfortunately, however, due to the lack understanding about the role of expectations in relationships, recommendations are often inconsistent with one another. For example, one approach to preventing marital distress argues that "couples are at increased risk when expectations are unreasonable" (Stanley, Blumberg, & Markman, 1999, p. 285). In contrast, other work on marital therapy techniques suggests that "holding extreme standards does not seem to doom couples to dissatisfaction. Instead... holding extreme








standards that ask a great deal of the marriage is positively related to marital adjustment" (Baucom et al., 1996, p.83).

The goal of this dissertation is to determine the longitudinal implications of positive expectations about a relationship on relationship outcomes. Specifically, this research addresses the question: is it adaptive to have positive expectations at the beginning of a relationship, or will such optimism lead to disappointment? To this end, the remainder of this introduction is organized into four sections. The first section reviews research on the process of expectancy confirmation, whereby expectations lead to consistent outcomes and consistent evaluations, suggesting that positive expectations will lead to positive outcomes. The second section reviews the literature on counterfactual reasoning whereby positive expectations can lead to disappointment by providing a contrast to actual outcomes. The third section will attempt to reconcile these contradictory findings by distinguishing between expectations about global outcomes and expectations about more specific outcomes. The final section will describe a study designed to evaluate whether the distinction between global and specific expectations matters by examining the longitudinal implications of positive expectations about relationship outcomes that vary in specificity.














LITRERATURE REVIEW


Expectancy Confirmation

One of the oldest and most well-established principles in social psychology is the idea that people's outcomes and evaluations are consistent with their prior expectations. In a classic study of this effect, Harold Kelley (1950) asked students to evaluate a guest lecturer whom they had been led to expect would be either "warm" or "cold." After participating in a group discussion, those students who expected a warm lecturer evaluated him more positively than students who expected a cold lecturer. In other words, students evaluated the lecturer in a manner consistent with their prior expectations. Rosenthal and Jacobson (1968) extended this finding by demonstrating that expectancy confirmations affect important social situations. These authors told elementary school teachers that based on results from an IQ test some of their students would "bloom" intellectually over the course of the academic year. In actuality, "bloomers" were randomly selected and thus no more likely to show increases in intelligence than the other students in the class. When tested with the IQ test again at the end of the school year, however, the students labeled "bloomers" demonstrated a significantly greater increase in IQ than students not labeled "bloomers." Since the early work of Kelley (1950) and Rosenthal and Jacobson (1968), researchers have shown that expectancy confirmations influence numerous other important social phenomena including stereotyped perceptions (e.g., Duncan, 1976; Hasorf& Cantril, 1954; Rosenhan, 1973), discriminatory behaviors








(e.g., Word, Zanna, & Cooper, 1974), health factors (e.g., Ditto & Hilton, 1990; Taylor & Brown, 1988), close relationship processes (e.g.Downey, Freitas, Michaelis, & Khouri, 1998; McNulty & Karney, under review; Murray, Holmes, & Griffin, 1996), and academic achievements (e.g., Bandura, 1982; Henderson & Dweck, 1990).

Research has revealed two mechanisms through which expectations can be confirmed. One mechanism, behavioral confirmation, refers to the idea that prior expectations for an event give rise to expectancy consistent behavior during the event (Merton, 1948). For example, in the Kelley (1950) study, one reason students expecting a warm lecturer may have evaluated him more positively is that students' prior expectations led them to behave more positively towards the lecturer, encouraging him to behave more positively in return (also see Word et al., 1974). Follow-up research on the confirmation of teachers' expectations about their students' performance provides evidence of behavioral confirmation. Harris and Rosenthal (1985) conducted a metaanalysis of 135 studies of teacher expectancy effects and found that 16 different behaviors (e.g., praise, input, ignoring) significantly mediated the effects of teachers' expectations on students' performance.

A second mechanism, perceptual confirmation, refers to the idea that prior expectations lead people to interpret the details of an event in ways that give rise to expectancy consistent evaluations of the event (Darley & Fazio, 1980; Snyder, 1984). For example, an alternative explanation for the results of the Kelley study is that students expecting a warm lecturer interpreted his behavior more charitably, even if they behaved no differently toward the lecturer and the lecturer behaved no differently toward them. Research on the effects of social stereotypes has confirmed that preexisting beliefs can








exert a powerful effect on people's perceptions and evaluations of others through perceptual confirmation (for a review, see Higgins & Bargh, 1987). In one study, white participants interpreted an "ambiguous" shove as more hostile when performed by an African American man than when performed by a white man (Duncan, 1976). The author argued that the white participants interpreted the ambiguous behavior according to the stereotype that African Americans are impulsive and prone to crimes and violence.

Both mechanisms are responsible for expectancy confirmations in interpersonal relationships. Consistent with the idea of perceptual confirmation, Weiss (1980) argued that, through the process of "sentiment override," partners in generally satisfying relationships should tend to evaluate specific aspects of the relationship positively, whereas partners in more distressed relationships should interpret specific aspects more negatively. These ideas have received empirical support. For instance, research on attributions demonstrates that partners who hold positive beliefs about their relationships tend to interpret specific relationship events in a positive light, whereas partners who hold less positive beliefs about their relationships tend to interpret specific events more negatively (e.g., Bradbury & Fincham, 1990; Fincharn, 1985; Holtzworth-Munroe & Jacobson, 1985; Jacobson, McDonald, Follette, & Berley, 1985; Karney & Bradbury, 2000).

Interpersonal expectancies may also get confirmed through behavioral

confirmation. Snyder, Tanke, and Berscheid (1977) recorded male college students having a telephone conversation with women they were led to believe were either more or less physically attractive. Objective raters found that men who believed they were interacting with attractive women elicited more sociable and friendly behavior from their








partners than men who believed they were interacting with less attractive women (for replication, see Anderson & Bern, 1981). In other words, participants' beliefs about their telephone partners affected the way they behaved toward their partners, which subsequently affected how their partners behaved in return. Research on more established relationships has examined similar processes in problem-solving discussions (Downey et al., 1998). This research examined the effects of partners' sensitivity to rejection, conceived as an individual difference variable, on their problem-solving behaviors and their partners' mood immediately following a problem-solving discussion. Results indicated that women who scored higher in rejection sensitivity engaged in behaviors that predicted greater negative feelings in their partners following the interaction.

Consistent with the idea that expectations may affect relationship outcomes by affecting the way couples approach problem-solving in their relationships, Knee (1998) found that partners' implicit theories about relationships affected the way they responded to relationship stressors He examined dating partners who endorsed different theories about whether relationships were either destined (i.e., determined by fate or the immutable characteristics of the partners) or grown (i.e., developed through effort and communication). Although neither type of belief was associated with satisfaction, results revealed that those who believed relationships require time to grow were less likely to disengage from the relationship in response to stressors and more likely to report adaptive coping strategies. In other words, those who expected their relationship to require effort reacted more positively to relationship problems.

In sum, partners' expectations about their relationships appear to be confirmed

through perceptual and behavioral confirmation. Accordingly, positive expectations about








a relationship should lead to positive relationship outcomes for two reasons. First, through perceptual confirmation, positive expectations should lead partners to interpret the events that occur in their relationships in positive ways. Second, through behavioral confirmation, positive expectations should lead partners to behave in ways that elicit positive outcomes. For the same reasons, less positive expectations about a relationship should lead to less positive outcomes; partners with less positive outcomes should interpret events more negatively and fail to engage in behaviors that are adaptive, or engage in behaviors that are less adaptive.


Positive Expectations and Disappointment

In contrast to the idea that positive expectations for a relationship may be

adaptive, theory and research in other domains suggest that people with high expectations may be disappointed when outcomes fall short of their expectations. As early as 1890, William James argued that people's self-evaluations can be expressed as a ratio of their successes to their pretensions (James, 1890). If people's achievements fall short of their aspirations, their evaluations are low, whereas if their achievements exceed their aspirations, their evaluations are high.

Research suggests that, like James, people believe their expectations affect the way they evaluate their outcomes. In one study, participants reported that they believed people would feel worse following a negative outcome (e.g., receiving a "C" on a test) that was unexpected versus expected, and feel better following a positive outcome (e.g., receiving an "A" on a test) that was unexpected versus expected (Shepperd & McNulty, in press, study 1). Additional research shows that people may avoid being disappointed by lowering their expectations in the face of feedback, ensuring that their expectations do








not exceed their outcomes (Mellers et al., 1997, study 2; Shepperd, Findley-Klein, Kwavnick, Walker, & Perez, 2000; Shepperd, Ouellette, & Fernandez, 1996; Taylor & Shepperd, 1998). In a study by Shepperd et al. (1996, study 2), participants predicted their scores on an upcoming exam on four occasions: two weeks before the exam, just after the exam, 45 minutes before feedback, and seconds before feedback. Results revealed that the average predictions of the students were optimistic before the exam, but became more realistic and eventually pessimistic just before feedback. Subsequent research suggests that this tendency stems from a desire to avoid receiving unexpected negative feedback (Shepperd & McNulty, 2000).

Clearly people believe that their expectations affect the way they evaluate their outcomes, but do they? The answer appears to be yes. In an experimental test of this hypothesis, Shepperd and McNulty (in press, study 3) led participants to believe they were either likely or unlikely to test positive for a threatening medical condition. All participants then tested themselves and received feedback that they had tested positively or negatively for the disease. Of the participants who learned they had the disease, those who expected not to have it felt the worst. Of the participants who learned they did not have the disease, those who expected to test negative felt best. In other words, following both positive and negative outcomes, people who expected the worst felt the best. In similar research, among participants taking part in a series of gambles, unexpected wins were more satisfying than expected wins and unexpected losses were more disappointing than expected losses (Mellers et al., 1997, study 1). Research also suggests that even people who are objectively better off can feel worse when their outcomes do not meet their expectations. In a study of medal winners at the Olympic games, bronze medallists








showed a tendency to appear happier than silver medallists. Bronze medallists focused on the alternative of winning no medal, whereas silver medallists focused on the alternative of winning a gold medal (Medvec, Madey, & Gilovich, 1995). In sum, as James suggested over a century ago, people tend to be disappointed by outcomes that do not meet their expectations.

Do partners who hold high expectations about their relationships similarly risk

being disappointed? Some research suggests that they might. Eidelson and Epstein (1982) developed an inventory to measure five beliefs they thought were dysfunctional to relationships: the belief that disagreement is destructive, the belief that mindreading is expected, the belief that partners can not change, the belief in sexual perfectionism, and the belief that the sexes are different. Such beliefs are likely to lead spouses to develop specific expectations about their own relationships (Olson et al., 1998). Consistent with the notion that positive relationship expectations may lead to disappointment, partners who report higher agreement with these beliefs are more likely to report being unhappy with their relationships (see also, Epstein & Eidelson, 1981; Jones & Stanton, 1988).

In sum, research on counterfactual thinking suggests that when outcomes exceed expectations, people are pleased and satisfied, but when outcomes do not meet expectations, people are let down and disappointed. Accordingly, positive expectations about a relationship may lead to future relationship distress by leading partners to react more negatively to specific relationship problems. Instead, more moderate expectations may be adaptive because they may allow partners to dismiss specific problems, maintaining a relatively positive level of satisfaction.








Reconciliation and the Distinction between Global and Specific Expectations

When are positive expectations about a relationship likely to lead to positive

outcomes and when are they likely to lead to negative outcomes? The research reviewed thus far suggests that to the extent that they get confirmed, positive expectations about a relationship should be beneficial. To the extent that they do not get confirmed, however, not only will they not be beneficial, but they may lead to disappointment. Thus, determining when positive expectations are beneficial to a relationship requires determining when they will get confirmed and when they will not get confirmed. Researchers in other domains have addressed this issue, revealing two factors that affect whether expectancy confirmations occur. First, perceivers' motivation can influence the extent to which expectancy confirmations will occur. Specifically, research has shown that the goal of forming an accurate impression of a target can reduce the likelihood that people's expectations will influence their evaluations of a particular target or outcome (Fiske & Neuberg, 1990; Miller & Tumbull, 1986). For example, participants in one study were asked to interview two job applicants, and given prior negative information about one of them. Results indicated that when participants were told to be accurate, they were less likely to behave in ways that confirmed their initial expectations (Neuberg, 1989). Second, perceivers' a also can affect the extent to which they will rely on their expectations. Specifically, certain factors make perceivers more or less able to interpret events or behave in a manner consistent with their prior expectations. For example, research has shown that people who lack sufficient cognitive resources tend to make judgments that rely more heavily on their existing beliefs and expectations than








those who have ample cognitive resources (Chaiken, 1987; Gilbert & Hixon, 1991; Pratto & Bargh, 1991).

Similarly, Miller and Turbull (1986) argued that certain outcomes may allow perceivers to be more or less able to confirm their initial expectations. Although these authors did not expand on this possibility, one quality that may affect the extent to which expectancy confirmations occur is the specificity of the outcome; some expectations involve outcomes that are very global whereas others involve outcomes that are more specific. For the purpose of this dissertation, global outcomes will be defined as outcomes that can be assessed in terms of a large number of characteristics. For example, happiness, a global outcome, can be assessed by many specific behaviors (e.g., smiling, laughing, behaving charitably, being supportive). In contrast, specific outcomes will be defined as outcomes that can be assessed in terms of a more limited range of characteristics. For example, punctuality, a more specific outcome, can be defined by fewer characteristics (e.g., arriving on time; for related discussion, see Hampson, Goldberg, & John, 1987; Hampson, John, & Goldberg, 1986; Neff & Karney, in press).

For two reasons, the specificity of an outcome should be associated with whether positive expectations about relationships get confirmed or disconfirmed. First, outcomes that are more specific are less open to interpretation. For example, Dunning and his colleagues (Dunning, 1995; Dunning & McElwee, 1995; Dunning, Meyerowitz, & Holzberg, 1989) have shown that individuals describe themselves more positively on global traits than on specific ones. Explaining this finding, the researchers note that global traits can be defined in idiosyncratic ways, allowing the perceiver to choose from a wide range of specific examples to justify a positive self-view. Specific traits, in contrast,








are defined more concretely, and so restrict the flexibility of the perceiver to justify a desired belief. Partners' evaluations of their relationships also seem to be constrained by the specificity of the belief (Karney, McNulty, & Bradbury, 2001; McNulty & Karney, 2001; Neff& Karney, in press). In one study (McNulty & Kamey, 2001), newlywed partners reported their global satisfaction and their satisfaction with more specific aspects of their relationships (e.g., communication, level of support) every day for seven consecutive days. Results revealed that global evaluations were more positive and less variable than evaluations of specific aspects, suggesting that partners' global beliefs about the relationship may be more flexible and open to interpretation than their beliefs about specific aspects of the relationship. In the same way, specific outcomes, because they can be defined by a limited range of characteristics, may limit the extent to which perceptual confirmation is likely to occur.

Second, outcomes that are more specific are less likely to be associated with

behavior. To the extent that the outcome is global, it is likely that there is a wide variety of behaviors that could potentially lead to it. In contrast, to the extent that the outcome is very specific, it is likely that there is only a limited range of behaviors that could potentially lead to it. For example, expecting a relationship to be positive generally is more global than expecting to resolve relationship conflicts well. Accordingly, there are many more behaviors under one's control that can lead to a happy relationship (e.g., support, having fun, romance, sharing, problem-solving) than behaviors that can lead to resolving conflicts (problem-solving behavior). In this way, the specificity of the belief may limit the extent to which behavioral confirmation can occur.








In light of these issues, expectations about more global aspects of a relationship, because they are more open to interpretation and capable of being confirmed by a wider variety of behaviors, may be more likely to lead to consistency effects. For example, the positive belief that a relationship will succeed should result in supportive behaviors that will likely create more positive relationship satisfaction. In contrast, positive expectations about more specific outcomes, because they are less open to interpretation and capable of being confirmed by only a limited variety of behaviors, may be more likely to be disconfirmed and lead to disappointment. For example, the positive expectations for punctuality should be disconfirmed to the extent that one is late for important engagements.

The previously-reviewed research on expectancy confirmation and counterfactual thinking is consistent with these ideas. For instance, the research on counterfactual thinking, showing that positive expectations can be disconfirmed and lead to disappointment, involved outcomes that were more specific and concrete (e.g., test results, wins and losses, Olympic medals). In contrast, the research on expectancy confirmation, showing that positive expectations can be confirmed and lead to positive outcomes, involved expectations about more global and ambiguous outcomes. For instance, in the Kelley (1950) study, participants led to expect a "warm" lecturer evaluated him more positively. This expectancy was easy to confirm because there are a wide variety of characteristics and behaviors that can be used to justify the belief that someone is "warm" (e.g., supportiveness, humor, eye contact, touching). Similarly, the research on relationship attributions demonstrates that people who are generally satisfied with their relationships tend to interpret marital events more positively. Again,








relationship satisfaction is a global belief that can be confirmed by a wide variety of marital events and partner behaviors. Likewise, the research on behavioral confirmation also involved expectations about outcomes that were very global (e.g., "blooming" intellectually), and therefore associated with multiple behaviors under perceivers' control. For instance, the meta-analysis by Harris and Rosenthal (1985) identified a total of 21 different behaviors that mediated the effects of teachers' expectation on students' performance.

In addition to this indirect support, two studies provide direct support for the idea that specific expectations, more than global expectations, are likely to get disconfirmed. The first study examined how partners' prior expectations for a specific interaction influence their appraisals of the interaction (McNulty & Karney, under review). In this study, newlywed participants reported their expectations about how a specific upcoming conversation would go, participated in a 10 -minute videotaped discussion of a relationship problem, and finally reported their evaluations of the discussion. Results revealed no evidence for behavioral confirmation. Partners who expected the interaction to go well were no less likely to behave negatively than people who expected the conversation to go negatively. Consistent with the idea that specific outcomes may be under less behavioral control than more global outcomes, the authors' post-hoc explanation for this finding was that the ability to solve problems in close relationships is best conceived not as a controlled response, but rather as a skill that may develop independently of partners' beliefs and expectations. The second study providing direct support for the idea that expectations for specific outcomes are less likely to be confirmed was a study examining the effects of expectations on achievement. In this study, Armor








and Taylor (1997) found that people with high expectations were disappointed by their outcomes, but only when expectations were specific.

In sum, whether positive expectations are beneficial to relationships may depend on whether they involve global or specific outcomes. Positive expectations about global outcomes, because they are more open to interpretation and capable of being confirmed by a wide variety of behaviors, may be more likely to lead to consistency effects. In contrast, positive expectations about more specific outcomes, because they are less open to interpretation and capable of being confirmed by only a limited variety of behaviors, may be more likely to lead to disappointment. Based on this idea, the primary goal of this study is to test the hypothesis that positive global expectations about a relationship will lead relationships to be more satisfying over time, whereas positive specific expectations about a relationship will lead relationships to be less satisfying over time.


Overview of the Current Study

Given the uncertainty about how positive expectations about a relationship affect relationship functioning, and given the prospect that the specificity of the expected outcomes may help clarify this issue, the current study examined how expectations about outcomes at various levels of specificity affect relationship outcomes over time. Married couples participating in a broader study of relationship satisfaction were asked to report their expectations for various relationship outcomes at the beginning of a longitudinal study. Subsequently, these couples responded to measures of relationship satisfaction every six months for two years. Two conceptually similar dependent measures were used to operationalize relationship outcomes. The first was change in satisfaction over time. Analyses examined whether each type of expectation was associated with within-subject








change in marital satisfaction. The second dependent variable was reactivity to relationship problems. Analyses assessed the extent to which variability in spouses' satisfaction with the relationship covaried with variability in the number and severity of relationship problems, and the extent to which expectations moderated this association. All of the couples participating in this research were newlyweds, assessed within the first six months of both partners' first marriage. This is an appropriate sample because, as Jussim and Eccles (1995) point out, over the course of an ongoing relationship, expectations are likely to become accurate reflections of past experience. In a sample of newlyweds, expectations and prior experience were less likely to be confounded.

At the initial assessment, partners reported their expectations for global and specific relationship events. First, they reported their expectations for two global outcomes: (a) whether they expected their global satisfaction to get better, stay the same, or get worse over time and (b) how confident they were that their relationships would remain intact or stable. Second, they reported their expectations for three specific relationship outcomes: (a) expectations about specific changes in relationship satisfaction, (b) expectations about specific partner qualities and (c) expectations about specific relationship problems.


Hypotheses

Hypothesis 1. Consistent with the idea that that expectations about global

outcomes are likely to be confirmed through the processes of perceptual and behavioral confirmation, I predicted that spouses who expected global outcomes to be positive would (a) behave less negatively and (b) interpret events more charitably in their relationships. Thus, positive global expectations should be associated with (a) a lower








proportion of negative behavior during their interactions and (b) more positive scores on the measure of relationship attributions, whereas less positive global expectations would be associated with (a) a higher proportion of negative behavior during their interactions and (b) lower scores on the measure of relationship attributions. In other words, scores on each measure of global expectations will be negatively associated with observations of behavior and positively associated with scores on the measure of relationship attributions.

Hypothesis 2. Consistent with the idea that that expectations about specific outcomes are less likely to be confirmed through the processes of perceptual and behavioral confirmation, I predicted that spouses who expected specific outcomes to be positive would be no more likely to (a) behave negatively in their relationships over time or (b) interpret events more charitably in their relationships than spouses who expect less positive specific outcomes. In other words, scores on each measure of specific expectations should not be associated with observations of behavior or scores on the measure of relationship attributions.

Hypothesis 3. Based on the idea that expectations about specific outcome are less susceptible to the confirming effects of expectations, spouses' positive expectations about specific relationship outcomes were likely to be disconfirmed over time leading to disappointment. Thus, I predicted that spouses with positive expectations about more specific relationship outcomes would experience negative changes in satisfaction over time and a stronger negative covariance between global satisfaction and specific relationship problems. In contrast, because they are less likely to be disappointed, spouses with less positive specific expectations should experience more stable satisfaction and a weaker negative covariance between satisfaction and specific problems








over time. Predictions in terms of each measure of specific expectations are presented in the following paragraphs.

Participants were asked how they expected their feelings toward their relationship to change over time: whether their feelings would experience ups and downs or remain steady from day to day. I predicted that partners who expected their feelings to be very steady would be more likely to (a) experience declines in satisfaction over time and (b) experience a stronger covariance between reported problems and satisfaction over time. In contrast, partners who expected some ups and downs should be more likely to (a) be happier over time and (b) experience a weaker covariance between reported problems and satisfaction over time. This prediction will be supported to the extent that expectations for steadiness are negatively associated with change in satisfaction and negatively associated with the covariance between global satisfaction and specific problems.

Participants were asked to report their expectations about specific qualities of their partners (e.g., my partner will never disappoint me). I predicted that participants who held more positive expectations for their partners would be more likely to (a) experience declines in satisfaction over time and (b) experience a stronger covariance between reported problems and satisfaction over time. In contrast, partners who expected their problems to remain the same or get worse should be more likely to (a) be happier over time and (b) experience a weaker covariance between reported problems and satisfaction over time. This prediction will be supported to the extent that expectations for partner are negatively associated with change in satisfaction and negatively associated with the covariance between global satisfaction and specific problems.








Participants were asked to report how they expected a number of potential

relationship problems to develop over time: whether they would get worse or get better. I predicted that partners who expected their problems to get better would be more likely to

(a) experience declines in satisfaction over time and (b) experience a stronger covariance between reported problems and satisfaction over time. In contrast, partners who expected their problems to remain the same or get worse should be more likely to (a) be happier over time (b) and experience a weaker covariance between reported problems and satisfaction over time. This prediction will be supported to the extent that expectations for specific problems are negatively associated with change in satisfaction and negatively associated with the covariance between global satisfaction and specific problems.

Hypothesis 4. Based on the idea that expectations about global outcome are more susceptible to the confirming effects of expectations, spouses' positive expectations about global relationship outcomes were likely to be confirmed over time. Thus, I predicted that spouses with positive expectations about more global relationship outcomes would experience more stable satisfaction over time and a weaker negative covariance between global satisfaction and specific relationship problems. Because less positive global expectations are also likely to be confirmed, spouses with less positive global expectations should experience declines in satisfaction and a stronger negative covariance between satisfaction and specific problems over time. Predictions in terms of each measure of global expectations are described in the following paragraphs.

Participants were asked how they expected their feelings toward their relationship to change over time: whether their feelings would become better or become worse. I predicted that partners who expected their feelings to grow positively over time would be








more likely to experience (a) positive changes in marital satisfaction over time and (b) a weaker negative covariance between reported problems and satisfaction over time. In contrast, partners who expected their relationships to stay the same or get worse should be more likely to (a) be less happy over time and (b) experience a stronger covariance between reported problems and satisfaction over time. This prediction will be supported to the extent that expectations for growth are positively associated with change in satisfaction and positively associated with the covariance between global satisfaction and specific problems.

Participants were asked to predict the likelihood that they would still be married in four years. I predicted that partners who were more confident that they would remain married would be more likely to experience (a) positive changes in marital satisfaction over time and (b) a weaker negative covariance between reported problems and satisfaction over time. In contrast, partners who were less confident that they should stay remain married should be more likely to (a) be less happy over time and (b) experience a stronger covariance between reported problems and satisfaction over time. This prediction will be supported to the extent that expectations for stability are positively associated with change in satisfaction and positively associated with the covariance between global satisfaction and specific problems.

Hypothesis 5. Consistent with the idea that positive global expectations should lead to positive relationship outcomes through perceptual and behavioral confirmation, I predicted that the effects of positive global expectations on relationship outcomes would be mediated by positive interpretations (perceptual confirmation) and fewer negative behaviors (behavioral confirmation). Specifically, observations of behavior and scores on





22


the measure of relationship attributions should mediate the effects of global expectations on changes in marital satisfaction and the covariance between reported problems and satisfaction such that spouses with global positive expectations should behave less negatively and make more positive attributions.













METHOD


Participants

Newlywed couples were recruited for this study using two methods. The first

method was to place advertisements in community newspapers and bridal shops, offering up to $300 to couples willing to participate in a study of newlyweds. The second method was to review the applications of couples who had applied for marriage licenses in Alachua County, Florida. In Alachua County, marriage licenses are available to the public and contain data on spouses' ages, whether or not this is their first marriage, and the date of the wedding. Couples who were eligible for the study based on these criteria were sent letters offering them up to $300 to participate in a study of newlyweds. 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 3 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. Nearly 300 couples responded to these solicitations; the first 82 eligible couples who arrived for their scheduled interview comprised the current sample. Analyses revealed no significant differences in age or








years of education between couples recruited through the different types of solicitations (D'Angelo & Karney, 1999).

At the time of initial data collection, husbands were 25.1 (SD = 3.3) years old,

and had completed 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 completed 16.3 (SD = 1.2) years of education. Thirty-nine percent were employed full time, and 50% were full time students. Slightly over 70% of the sample was Christian (over 45% were Protestant) and 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 selfreport measures of marital satisfaction and a letter instructing couples to complete all questionnaires independently of one another. As part of the laboratory session, each spouse independently identified an area of difficulty in the marriage, and agreed to spend 10 minutes discussing the issues related to each area of difficulty. Spouses then participated in two 10-minute, videotaped discussions in which they were left alone to "work towards some resolution or agreement" for each area of difficulty. The order of the two interactions was determined through a coin flip. Couples were paid $50 for participating in this part of the study.

At approximately six-month intervals subsequent to 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 dissertation examined five waves of data, covering approximately the first two and a half years of marriage.


Materials

Marital satisfaction. The most common measures used to assess marital

satisfaction ask spouses to report their global sentiments towards the marriage as well as their level of agreement about specific problem areas (e.g., the Marital Adjustment Test; Locke & Williamson, 1958). As several authors have pointed out (e.g., Fincham & Bradbury, 1987), the use of such omnibus measures can lead to inflated associations with other variables that also address relationship processes. To ensure that these two ideas were not confounded in the current study, two global measures of satisfaction were used here. The first measure was a version of the Semantic Differential (SMD; Osgood, Suci, & Tannenbaum, 1957), that asks spouses to rate their perceptions of their relationship on 7-point scales between fifteen pairs of opposing adjectives (e.g., "Bad-Good," "Dissatisfied-Satisfied," "Unpleasant-Pleasant"). The SMD yields scores from 15 to 105. In the current sample, internal consistency of this measure was high (coefficient alpha = .91 for husbands and .93 for wives). (A sample of this measure can be found in Appendix A.) The second global measure was the Quality Marriage Index (QMI; Norton, 1983), a six item scale asking spouses to report the extent to which they agree or disagree with general statements about their marriage (e.g., "We have a good marriage" and "My relationship with my partner makes me happy"). The QMI yields scores from 6 to 45.








Internal consistency of this measure also was high (coefficient alpha = .94 for husbands and .94 for wives). (A sample of this measure can be found in Appendix B.)

Inventory of Specific Marital Problems. Participants' perceptions of specific

relationship problems were measured using the Marital Problems Inventory (MPI; Geiss & O'Leary, 1981). This measure lists nineteen potential problem areas in a marriage (e.g., children, communication, household management) and asks participants to rate each on a scale from I ("not a problem") to 11 ("major problem"). These items were summed to form an index of the severity of relationship problems. Internal consistency of this measure was high (coefficient alpha = .88 for husbands and .89 for wives). (A sample of this measure can be found in Appendix C.)

Expectations for Growth in Relationship Satisfaction. Two measures were

developed to assess spouses' expectations about global relationship outcomes. The first was a three-item measure designed to assess how partners expected their relationship satisfaction to develop globally over time. The first item from this measure asked partners to respond to the following question: "over the next six months, do you expect that your overall feelings about your marriage will become:" 1 = "much worse," 2 = "a little worse," 3 = "stay the same," 4 = "a little better," 5 = "much better." The second item asked spouses to respond to the same question with respect to the "next 4 years." (A sample of these items can be found in Appendix D.) For the third item, spouses selected from among nine pictures reflecting different patterns of growth the one pattern that best represented their expected trajectory of marital satisfaction over the upcoming six months. (A sample of these pictures is contained in Appendix E.) For the current measure, pictures were coded for the degree of expected growth in satisfaction: 1 = flat








trajectory, 2 = positive trajectory. These three items were summed to form an index of expectations for growth. Internal consistency of these three items was adequate (coefficient alpha = .63 for husbands and .72 for wives).

Expectations for Stability. The second measure of global expectations was one item designed to assess spouses' expectations about the likelihood that the relationship would remain intact or stable over the upcoming four years. Specifically, the item read "How likely is it that the two of you will still be married four years from now?": 1= "not at all likely," 2 = "not too likely," 3 = "pretty likely," 4 = "very likely," 5 = "completely certain." (A sample of this measure can be found in Appendix F.)

Expectations for Steadiness of Relationship Satisfaction. Three measures were

developed to assess spouses' expectations about specific relationship outcomes. The first measure was a three-item measure designed to assess very specifically how partners expected their relationship satisfaction to develop from day to day. In contrast to expectations for growth, expectations for steadiness assessed how much spouses expected their satisfaction to change from day to day. The first item asked spouses to respond to the following question: "over the next six months, which of the following describes how your feelings towards your marriage are likely to change?": 1 = "major highs/major lows," 2 = "some ups and downs," 3 = "a few ups and downs," 4 = "pretty steady," 5 "very steady." The second item asked spouses to respond to the same question with respect to the "next four years." (A sample of these items can be found in Appendix G.) Additionally, for the third item, the pictures contained in Appendix E were coded for the degree of expected steadiness: 1 = unsteady, 2 = very steady. These three items were








summed to form an index of expectations for steadiness. Internal consistency of these three items was relatively high (coefficient alpha = .77 for husbands and .78 for wives).

Expectations for Partner. The second measure of specific expectations was

designed to assess participants' specific expectations for how their partners would behave in the future. This measure asked spouses to respond to nine items designed to describe fairly unrealistic, specific qualities that people might expect their partners to possess (e.g., my partner will never disappoint me, my partner will always be attractive to me) on a scale from 1 (strongly disagree) to 7 (strongly agree). These items were summed to form an index of expectations for partner. Internal consistency of this measure was adequate (coefficient alpha = .71 for husbands and .80 for wives). (A sample of this measure can be found in Appendix H.)

Expectations for Specific Problems. The second measure of specific expectations was designed to assess participants' expectations for how the problems on the MPI would develop over time. The measure lists the same nineteen problems listed on the MPI and asks participants to "indicate how you think each problem is going to develop for you and your spouse over the next six months:" I = "get worse," 2 = "stay the same," 3 = "get better." These items were summed to form a single index of expectations for problems. Internal consistency of this measure was relatively high (coefficient alpha =.83 for husbands and .74 for wives). (A sample of this measure can be found in Appendix I.)

Attributions. The manner in which partners interpret the events in their marriages was assessed using a measure of marital attributions: the Relationship Attributions Measure (RAM; Fincham & Bradbury, 1992). This 24-item measure presents spouses with four negative stimulus events that are likely to occur in all marriages (e.g., "Your








spouses 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 partner's behaviors intentional, selfishly motivated, and blameworthy. For each subscale, 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 positive light. Internal consistency of each sub-scale was relatively high (for causality attributions, coefficient alpha = .85 for husbands and .73 for wives; for responsibility attributions, coefficient alpha = .89 for husbands and .90 for wives). (A sample of this measure can be found in Appendix J.)

Marital interaction behavior. The videotaped discussions were coded using a

modified version of the Verbal Tactics Coding Scheme (VTCS; Sillars, Coletti, Parry, & Rogers, 1982). This version of the VTCS assigns each spouse's speaking turn one of four possible codes. A speaker received an Avoidant code for speaking turns that were offtopic or moved the discussion away from the problem at hand. A speaker received one of two negative codes for speaking turns that either directly faulted, rejected, or criticized the partner (Direct Negative), or indirectly criticized the partner through presumptive attributions, avoiding responsibility, or hostile questions (Indirect Negative). A speaker








received a Constructive code for speaking turns that were on-topic and not negative. In light of prior research showing that negative behaviors are more reliably linked to relationship outcomes than positive behaviors (Bradbury & Karney, 1993), only the negative codes will be examined in the current study.

The reliability of this system was assessed by randomly choosing 30% of the interactions to be coded by a second rater. Degree of agreement between raters was estimated with an intra-class correlation coefficient (ICC) comparing the amounts of each code observed by each rater across the interactions. ICCs indicated adequate inter-rater reliability for the codes analyzed in the current study (for direct negative, ICC = .66 for husbands and .83 for wives; for indirect negative, ICC = .83 for husbands and .65 for wives). Because they were correlated for each spouse in each interaction (rs ranged from .34 to .54), the two negative codes were summed to form a single index of negative behavior for each spouse during each interaction. Because these indices were correlated for each interaction (rs ranged from .40 to .50), scores were summed across each interaction to form a single index of negative behavior for each spouse.


Data Analysis

Hierarchical Linear Modeling (HLM; Bryk & Raudenbush, 1992), implemented with the HLM/2L computer program (Bryk, Raudenbush, & Congdon, 1994) was used to analyze these data for several reasons. First, in contrast to other approaches to analyzing multilevel models (e.g., structural equation modeling), HLM 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 (1973) 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 can not 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.

Parameters describing husbands' and wives' data will be estimated

simultaneously in a couple-level model, according to procedures described by Raudenbush, Brennan, & Barnett (1995). Generally, two types of analyses will be conducted, one for each of the two dependent measures of relationship outcomes. First, to test the hypotheses that expectations affected change in satisfaction over time, growth curve analyses were conducted to model individual change over two years. In the first stage, within-subjects trajectories were computed for each individual's marital satisfaction over time. In the next stage, expectation scores were used to account for between-subjects differences in these trajectories. Second, multilevel modeling was conducted to test the hypothesis that expectations were associated with the way spouses reacted to their relationship problems over two and half years. In the first stage of this analysis, repeated-measures data from each spouse were used to estimate the extent to which variability in reported problems was associated with variability in satisfaction. In the second stage, expectations were used to account for between-subjects differences in this covariance.













RESULTS


Descriptive Statistics and Correlations

Descriptive statistics for all independent variables are reported in Table 1. As would be expected from a sample of newlyweds, on average husbands and wives held very positive expectations about their relationships. Furthermore, both spouses tended to hold the most positive expectations about global outcomes. For instance, among husbands, 85% and 27% reported the highest possible score for expectations for stability and expectations for growth respectively, whereas only 11% reported the highest possible score for expectations for steadiness and none reported the highest possible score for expectations for partner or expectations for problems. Among wives, 90% and 37% reported the highest possible score for expectations for stability and expectations for growth respectively, whereas only 17% reported the highest possible score for expectations for steadiness and none reported the highest possible score for expectations for partner or expectations for problems. Furthermore, comparing the only two global and specific expectations that were assessed on the same scale, expectations for growth and expectations for steadiness, reveals that both husbands and wives expected more growth from their relationship than they expected steadiness, t = 4.1, p < .01 for husbands, and t = 4.6, p <.01 for wives. Despite the positivity of these expectations, standard deviations reveal substantial variability, indicating that further analyses were appropriate. Of note, husbands and wives did not differ in the positivity of their expectations about the








relationship, except that husbands had marginally higher expectations for their wives than wives had for their husbands, t = 1.6, p < 10.

Descriptive statistics for the hypothesized mediating variables are also reported in Table 1. Husbands and wives appeared to make relatively positive attributions, seeing external causes for negative events and freeing their partners from blame. Similarly, husbands and wives did not behave very negatively during their interactions, with mean proportions of negative behavior M = 10.2% for husbands and M = 13.4 for wives. Nevertheless, standard deviations reveal substantial variability, indicating that firther analyses were appropriate. Paired-sample t-tests revealed no differences between husbands and wives in their tendency to make positive versus negative attributions or to behave negatively.

Correlations among the independent and mediating variables are reported in Table

2. Both measures of global expectations were somewhat correlated for husbands, r =

0.20, but not for wives. Also, two specific expectations, expectations for variability and expectations for partner, were correlated for husbands and wives, r = 0.36 for husbands and r = 0.33 for wives. Importantly, however, these measures did not overlap completely suggesting that they did not simply tap the same construct. Some measures of specific expectations were also correlated with some measures of global expectations. For instance, husbands' and wives' expectations for growth and expectations for problems were somewhat correlated, r = 0.27 for husbands and r = 0.29 for wives. Also, husbands' and wives' expectations for stability were correlated with their expectations for steadiness, r = 0.22 for husbands and r = 0.39 for wives. However, although these








measures of global and specific expectations were associated with one another, again they did not overlap completely.

With respect to the hypothesized mediating variables, consistent with previous research, attributions for responsibility and attributions for causality were strongly associated with one another for both husbands and wives, r = 0.62 for husbands and r = 0.56 for wives (Bradbury & Fincham, 1990). Also, for wives, both types of attributions were negatively correlated with negative behavior, r = -0.20 for causality attributions and r = -0.40 for responsibility attributions. In other words, wives who tended to make positive attributions also tended to behave less negatively. Of note, inconsistent with the prediction of Hypothesis 1 which stated that expectations about global outcomes would be significantly associated with attributions and behavior, attributions and behavior did not appear to be associated with global expectations. Also, inconsistent with the prediction of Hypothesis 2 that stated that specific expectations would not be associated with either attributions or behavior, both types of attributions were associated with specific expectations. However, these bivariate correlations did not control for marital satisfaction. Because previous research reveals that satisfied couples tend to make positive attributions (see Bradbury & Fincham, 1990, for a review), and because satisfied couples may also tend to expect positive specific outcomes from their relationships, it remains possible that these correlations were the result of shared variance with marital satisfaction.

Cross-spouse correlations for the independent and mediating variables are

reported in Table 3. The correlations on the diagonal reveal that husbands' and wives' scores were at least marginally correlated on all measures except attributions for








responsibility. In other words, husbands and wives expected similar outcomes from their relationships, both globally and specifically, behaved similarly in their relationships, and made similar attributions about whether or not their partners were the cause of negative specific relationship events.


Were Specific and Global Expectations Associated with Attributions and
Negative Behavior?

Hypothesis I stated that spouses who expected positive global outcomes from their relationship would tend to make positive attributions, attributing negative relationship events to external circumstances and freeing their partners from blame, and tend to behave less negatively. As previously mentioned, bivariate correlations estimating these relationships are presented in the bottom section of Table 2. However, because attributions, behavior, and expectations all may have been related to marital satisfaction, partial correlations were computed between each measure of global expectations and attributions and between each measure of global expectations and behavior, controlling for marital satisfaction. These correlations are reported in the top half of Table 4. As is evident from these partial correlations, Hypothesis 1 was not supported. There is no evidence in these data that spouses who expected positive global outcomes from their relationships tended to make positive attributions or to behave less negatively. Of note, because a significant association between an independent variable and a hypothesized mediating variable is an essential criterion for establishing mediation, Hypothesis 5, that suggested attributions and behavior would mediate the effects of global expectations on marital satisfaction, was not pursued.








Hypothesis 2 stated that spouses' expectations for specific relationship outcomes would not be associated with their tendency to make positive attributions or their tendency to behave negatively. The positive bivariate correlations reported in the bottom half of Table 2 between specific expectations and attributions suggest that spouses who expected positive specific outcomes tended to make positive attributions. To rule out the possibility that these correlations may have been artifacts resulting because attributions, behavior, and expectations all are related to marital satisfaction, partial correlations were computed between each measure of specific expectations and attributions and each measure of specific expectations and behavior, controlling for marital satisfaction. These partial correlations are reported in the bottom half of Table 4. As can be seen in this section of Table 4, most correlations between specific expectations and attributions and specific expectations and behavior were non-significant when controlling for marital satisfaction. Furthermore, the correlations that were significant do not reveal any consistent patterns. For instance, the correlation between husbands' expectations for problems and responsibility attributions was positive, suggesting that more positive expectations were associated with more positive attributions. But the correlation between wives' expectations for partner and causality attributions was negative, suggesting that more positive expectations were associated with more negative attributions. Given these inconsistencies, spouses' expectations for specific relationship outcomes do not appear to have been systematically associated with the way they explained the events in their relationships or the way they behaved in their relationships.








Were Specific and Global Expectations Associated with Marital Outcomes
Over Time?

Because all subsequent hypotheses involve either changes in satisfaction over

time or the covariance between specific problems and marital satisfaction over time, the mean scores on each measure of marital satisfaction and the Inventory of Specific Marital Problems are presented in Table 5 for each wave of data collection. As can be seen in Table 5, scores on both measures of marital satisfaction became less positive over time for both husbands and wives suggesting a negative linear decline in satisfaction over time. Likewise, scores on the problem inventory became less positive over time, indicating a slight decrease in marital problems over time. One explanation for this pattern may be that although problems did not get more numerous or more severe over time, they became more important over time and thus became more strongly associated with marital satisfaction. Of note, attrition in the current study was relatively low. Whereas the average rate of attrition in prior research on marriage is 31% (Karney & Bradbury, 1995), 78% of husbands and 80% of wives responded at Time 5. Furthermore, because HLM can estimate trajectories based as few as three waves of data, final analyses were based on 79 out of 82 couples, or 96% of the sample.

To compute a within-subject trajectory of satisfaction for each individual, the longitudinal data were analyzed using HLM. Change for each individual was estimated using the following model:

Yij = f30j + 3 1j (Time) + rij [Equation 1] where Yij is the marital satisfaction of individual j at Time i; r30j is the marital satisfaction of individual j at Time 1 (i.e., the initial satisfaction of individual j); j is the rate of linear change in marital satisfaction of individual j; and rij is the residual variance in








repeated measurements for spouse j, assumed to be independent and normally distributed across spouses. Husbands' and wives' parameters were estimated simultaneously using a multivariate technique suggested by Raudenbush, Brennan, and Barnett (1995). Fitting this model to the data provided estimated generalized least squares of the average of the intercept and trajectory of marital satisfaction for husbands and wives, as well as restricted maximum likelihood estimates of the variances of these parameters. These estimates are reported in Table 6 along with t-statistics that test whether the means of each parameter are different from zero and chi-square statistics that test whether the variances of each parameter are different from zero.

With respect to the intercept, the estimates presented in Table 6 suggest that on average husbands and wives reported relatively high marital satisfaction as newlyweds. Comparison of Table 6 with Table 5 reveals that the mean intercept for each measure is similar to the mean satisfaction reported at Time 1. This is to be expected because the intercept represents the starting point of the trajectory. However, as revealed by the chisquare tests of variance, the mean initial satisfaction varied significantly across husbands and across wives on both measures of marital satisfaction. In other words, some spouses were initially more happy than other spouses.

With respect to the trajectory, the estimates presented in Table 6 suggest that on average husbands and wives tended to experience significant declines in satisfaction over the first two and a half years of marriage, consistent with prior studies of marital satisfaction (e.g., Karney & Bradbury, 1997). The t-statistics presented in Table 6 reveal significant negative trajectories for husbands and wives on each of the two measures of marital satisfaction. Furthermore, according to the significance of the chi-square tests








reported in Table 6, the trajectories varied across husbands and across wives for both measures. In other words, although on average husbands' and wives' satisfaction declined over two and a half years, satisfaction declined more for some spouses than it did for others. This is important because Hypothesis 3 and Hypothesis 4 suggest that specific and global expectations will account for this between-subject variability in change in satisfaction over time.

As a measure of reactivity to problems, HLM analyses were also used to compute within-subject estimates of the extent to which variability in marital satisfaction covaried with variability in specific problems over time, controlling for the trajectory of satisfaction over time. To estimate this covariance, scores on the Marital Problem Inventory were centered around the mean for each spouse and entered into the following model:

Yij = [oj + 3ij(Time) + [2j(Problems) + rij [Equation 2] where Yij is the global marital satisfaction score for individual j at time i; [30j estimates the satisfaction of individual j at Time 1; 3 Ij is the rate of linear change in marital satisfaction of individual j; [32j captures the covariance between variability in reported specific problems and variability in reported global marital satisfaction for spouse j, controlling for the trajectory of satisfaction; and rj is the residual variance in repeated measurements for spouse j, assumed to be independent and normally distributed across spouses. Husbands' and wives' parameters were estimated simultaneously using a multivariate technique suggested by Raudenbush, Brennan, and Barnett (1995).

The estimated generalized least squares of the average of the covariances between problems and global marital satisfaction for husbands and wives are reported in Table 7








along with restricted maximum likelihood estimates of the variances of these parameters. Also reported are t-statistics that test whether the means are different from zero and chi-square statistics that test whether the variances are different from zero. According to the t-statistics in Table 7, variability in specific problems was significantly negatively associated with variability in global satisfaction, controlling for the trajectory of satisfaction over time. When problems were more numerous and severe, marital satisfaction was more negative. Additionally, as evidenced by the chi-square statistics in Table 7, the extent to which problems covaried with satisfaction varied across husbands and across wives. In other words, although on average variability in husbands' and wives' satisfaction was associated with variability in their specific problems, the magnitude of this association was stronger for some spouses than for others. This is important because Hypothesis 3 and Hypothesis 4 suggest that specific and global expectations will account for between-subjects variability in the extent to which satisfaction and problems covary over time.

Hypothesis 3. The first prediction of Hypothesis 3 stated that positive

expectations about specific relationship outcomes would be associated with declines in satisfaction over time. To test this hypothesis, each measure of specific expectations was entered into a second-stage model to account for between-subject variability in the parameters of change estimated by equation 1.

The effects of specific expectations on the parameters of change in satisfaction are reported in Table 8. The top half of Table 8 reports the associations between specific expectations and initial satisfaction. In general, the t-statistics in this section reveal that spouses who were initially more satisfied with the relationship tended to expect specific








relationship outcomes to be more positive. Specifically, husbands' and wives' expectations for steadiness were positively associated with their initial satisfaction with the relationship, and husbands' expectations about the qualities of their wives were positively associated with their initial satisfaction with the relationship. Of note, the association between initial satisfaction and expectations for partner was significantly stronger for husbands than for wives on both measures of marital satisfaction, Z = 7.1, p < .01 for the SMD, Z = 4.4, p < .05 for the QMI.

The associations between specific expectations and change in marital satisfaction are reported in the bottom half of Table 8. Consistent with the predictions of Hypothesis 3, this section of Table 8 reveals several significant or marginally significant negative associations between specific expectations and change in satisfaction over time. Specifically, expectations for steadiness were significantly associated with steeper negative trajectories for wives on the QMI, and marginally associated with steeper negative trajectories for husbands on the QMI. The more that spouses expected their relationship satisfaction to be very steady over time, the less satisfied they became with their relationships over time. In contrast, the less spouses expected their relationship satisfaction to be very steady over time, the more satisfied they became with their relationships over time. Additionally, positive expectations about specific partner qualities were associated with steeper negative trajectories for husbands on both measures of marital satisfaction. The more that husbands expected their wives to posses positive qualities, the less satisfied they became with their relationships over time. In contrast, the less that husbands expected their wives to possess positive qualities, the more satisfied they became with their relationships over time. Finally, wives'








expectations for problems were also associated with negative changes in satisfaction over time on the SMD. The more that wives expected their problems to get better over time, the less satisfied they became with their relationships over time. In contrast, the less that wives expected their problems to get better over time, the more satisfied they became with their relationships over time. Although some measures of expectations were significantly associated with change in satisfaction for one spouse but not the other, analyses revealed no significant gender differences in the extent to which specific expectations were associated with change in satisfaction over time.

The second prediction of Hypothesis 3 stated that spouses with positive

expectations about specific relationship outcomes would be more reactive to relationship problems and thus experience a stronger covariance between their problems and their marital satisfaction over time. To test this prediction, each measure of specific expectations was entered separately into a second-stage model to account for betweensubjects variability in the covariance estimated by equation 2. Support for this prediction would be provided by significant negative associations between specific expectations and the covariance between problems and satisfaction, indicating that positive specific expectations were associated with a stronger association between relationship problems and satisfaction.

The effects of specific expectations on the covariance between problems and

satisfaction are reported in Table 9. The t-statistics in Table 9 do not reveal any support for Hypothesis 3. Although the negative direction of the effects is consistent with predictions, the estimates did not reach significance.








In sum, the prediction that positive expectations about specific relationship outcomes would lead to declines in marital satisfaction over time was supported. Husbands' and wives who tended to have very positive expectations about specific outcomes tended to become less satisfied over time. However, the prediction that positive expectations about specific outcomes would be associated with a stronger negative covariance between marital problems and marital satisfaction was not supported. Spouses who expected specific outcomes to be positive were not significantly more reactive to relationship problems than spouses who expected specific outcomes to be less positive.

Hypothesis 4. The first prediction of Hypothesis 4 stated that positive

expectations about more global relationship outcomes would be associated with more positive changes in satisfaction over time. To test this prediction, each measure of global expectations was entered into a second-stage model to account for between-subjects variability in the parameters of change estimated by equation 1.

The effects of global expectations on the parameters of change in satisfaction are reported in Table 10. The top half of Table 10 contains the associations between global expectations and initial satisfaction. In general, the t-statistics in this section reveal that spouses who were more satisfied with their relationships initially, tended to expect global relationship outcomes to be more positive. Specifically, husbands who were more satisfied with the relationship were more likely to expect their feelings about the relationship to grow over time and more confident that their relationship would remain intact or stable over time. Wives who were more satisfied with their relationships according to the SMD were marginally more confident that their relationship would remain stable over time. Husbands and wives did not differ in the extent to which








expectations for stability were associated with initial satisfaction, but the association between expectations for growth and initial satisfaction was significantly stronger for husbands than for wives on both measures of satisfaction, X = 5.0, p < .05 for the SMD,

= 4.4, p < .05 for the QMI.

The associations between global expectations and change in marital satisfaction are reported in the bottom half of Table 10. Contrary to the predictions of Hypothesis 4, this section of Table 10 reveals that husbands' global expectations are negatively associated with changes in satisfaction over time. The more that husbands expected their feelings about the relationship to grow over time and the more that husbands expressed confidence that their relationship would remain stable over time, the less satisfied they became with their relationship over time. In contrast, the less that husbands expected their feelings about the relationship to grow over time and the less that husbands expressed confidence that their relationship would remain stable over time, the more satisfied they became with their relationship over time. Although the association between global expectations and change in marital satisfaction was not significant for wives, there were no significant gender differences in the association between global expectations and change in satisfaction over time.

The second prediction of hypothesis 4 stated that spouses with positive

expectations about global outcomes would be less reactive to specific problems and thus experience a weaker covariance between problems and marital satisfaction over time. To test this prediction, each measure of global expectations was entered into a second-stage analysis to account for between-subjects differences in the covariance between problems and satisfaction estimated by equation 2. Support for the second prediction of hypothesis








4 would be provided by significant positive associations between global expectations and the covariance between problems and satisfaction, indicating that positive global expectations resulted in a weaker negative covariance between problems and satisfaction.

The effects of global expectations on the covariance between problems and

satisfaction are reported in Table 11. The t-statistics in Table 11 provide no support for Hypothesis 4 that spouses with positive global expectations would be less reactive to problems and experience a weaker covariance between problems and satisfaction. Instead, contrary to predictions, positive scores on several measures of global expectations were associated with a stronger negative covariance between problems and satisfaction. Specifically, husbands who expected their feelings about the relationship to grow more positive over time, and wives who were more confident that the relationship would remain intact over time, were at least marginally more reactive to specific problems over time.

In sum, the first prediction of Hypothesis 4 that positive expectations about global outcomes would lead to more stable marital satisfaction over time was not supported. Instead, husbands who tended to have very positive expectations about global relationship outcomes, such as the likelihood that their relationship would remain intact and the extent to which their feelings about the relationship would grow over time, tended to become less satisfied with the relationship over time. Importantly, although this association was not significant for wives, the size of the effect did not differ significantly across husbands and wives. The second prediction of Hypothesis 4 that positive expectations would be associated with a weaker negative covariance also was not supported. In fact, several associations between global expectations and the covariance








between problems and satisfaction were significant, suggesting that spouses with the most positive global expectations were more reactive to relationship problems than spouses with less positive expectations about global outcomes.

Did Global and Specific Expectations Interact to Affect Changes in Satisfaction?

Results indicate that positive expectations about both global and specific outcomes tended to lead to negative changes in satisfaction over time, suggesting that the effects of positive expectations on changes in marital satisfaction do not depend on the specificity of the expected outcome. However, spouses are likely to hold expectations about both global and specific outcomes simultaneously. Whereas some spouses may hold positive or less positive expectations about both types of outcomes, other spouses may hold positive expectations about one type of outcome and less positive expectations about the other type of outcome. Given these different combinations of expectations, it remains possible that global and specific expectations interact to lead to changes in marital satisfaction. For instance, because specific expectations are predicted to lead to negative changes in marital satisfaction, positive expectations about global outcomes may lead to stable marital satisfaction for spouses who have less positive expectations about specific outcomes. To test this possibility, each measure of global expectations was entered into a second stage model along with each measure of specific expectations and an interaction term to account for the changes in satisfaction estimated in equation 1.

Effect sizes of these interactions are reported in Table 12. As can be seen, there were several significant positive interactions between expectations for stability and the measures of specific expectations. To deconstruct these interactions, several values of each variable were substituted into the regression equation and the results were plotted








(see Aiken & West, 1991). As suggested by Cohen and Cohen (1983), three values were chosen for each of the two variables: the mean, I standard deviation above the mean, and I standard deviation below the mean. Figure 1 presents a depiction of two of these deconstructions. Figure A represents the interaction between wives' expectations for stability and wives' expectation for steadiness. Figure B represents the interaction between husbands' expectations for stability and husbands' expectations for their partner. As can be seen from Figure 1, these two interactions do not reveal consistent patterns. According to Figure A, wives who expected positive global outcomes and positive specific outcomes tended to have the most positive slopes, whereas according to Figure B, spouses who expected less positive global outcomes and less positive specific outcomes tended to have the most positive slopes. Although these figures represent only two of the interactions, the remaining interactions reveal similar inconsistencies. For this reason, it is difficult to draw strong conclusions about the interactive effects between global and specific expectations.


Did Positive Expectations Lead to Positive Changes in Marital Satisfaction When They
Were Likely to Get Confirmed?

A premise of this dissertation has been that positive expectations should be

adaptive to the extent that they are confirmed, but maladaptive to the extent that they are disconfirmed. Based on this idea, it was predicted that expectations about more global relationship outcomes, because they should be relatively likely to get confirmed, would lead to more stable marital satisfaction over time, whereas expectations about specific relationship outcomes, because they should be relatively likely to get disconfirmed, would lead to declines in marital satisfaction over time. This prediction was not








supported. Instead, both positive global and positive specific expectations were associated with negative relationship outcomes over time, suggesting that both types of expectations were likely to be disconfirmed over time. However, it remains possible that when they were confirmed positive expectations led to positive changes in satisfaction. Although on average spouses may have experienced specific outcomes that disconfirmed their expectations, some couples may have experienced positive relationship events that confirmed their expectations. Thus, the association between expectations and change in marital satisfaction may be moderated by the quality of the relationship, or the likelihood that the expectations will be confirmed. When the quality is low, positive expectations are likely to get disconfirmed, leading to decreases in satisfaction. In contrast, when the quality is high, positive expectations are likely to get confirmed, leading to more stable trajectories of marital satisfaction over time.

To test this possibility, two processes associated with relationship quality that should lead to the disconfirmation of positive expectations were examined as potential moderators of the relationship between each measure of expectations and changes in marital satisfaction: negative behavior and relationship attributions. Negative behaviors during interactions should lead to disconfirmation of a positive expectation because they are negative relationship events. To the extent that spouses or their partners behave more negatively, spouses' positive expectations are likely to be disconfirmed. To the extent that spouses and their partners behave less negatively, spouses' positive expectations are likely to be confirmed. Thus, because the proportions of each spouse's negative behaviors were highly correlated (r = .57), indices of husbands and wives negative behavior were combined to form a single index of relationship quality: the proportion of negative








behavior in the relationship. Because they represent the perception of negative relationship events, more negative attributions should also lead to disconfirmation of positive expectations. To the extent that spouses perceive their partners as the cause of a negative event or as blameworthy for a negative event, their positive expectations are likely to be disconfirmed. To the extent that spouses free their partners from blame for events, their positive expectations are likely to be confirmed. To test these possibilities, interactions between these expectations and negative behavior and expectations and attributions were entered into a second-stage model to account for between-subject variability in the changes in satisfaction estimated in equation 1.

Consistent with the idea that positive expectations should lead to negative

satisfaction only when they are likely to get disconfirmed, negative interactions between expectations and negative behavior would suggest that negative changes in satisfaction were most likely when positive expectations were high and negative behavior was high. The effect sizes of the interactions between expectations and behavior are presented in the first column of Table 13. Consistent with predictions, results reveal a significant pattern of negative interactions between expectations and negative behavior for wives. Figure 2 depicts a representation of one of these interactions, the interaction between wives' negative behavior and wives' expectations for stability. As can be seen, in Figure 2, when behavior was most negative, very positive expectations were associated with the steepest declines in marital satisfaction whereas less positive expectations were associated with the most stable satisfaction. However, when behavior was the least negative, the least positive expectations were associated with the steepest declines whereas the most positive expectations were associated with the most stable satisfaction.








In other words, when rates of negative behavior were high, positive expectations led to declines in marital satisfaction, but when rates of negative behavior were low, positive expectations led to stable satisfaction. In general husbands and wives did not differ significantly in the extent to which their expectations and negative behavior interacted to affect changes in their marital satisfaction over time. The exception was that the interactive effect for wives' expectations for partner and negative behavior on changes in scores on the SMD was significantly greater than the interactive effect for husbands, t =

6.6, p < .05. Of note, one interaction between wives expectations for growth and negative behavior was significantly positive, suggesting the opposite pattern of results. However, because all other significant and marginally significant interactions were negative, it is difficult to draw conclusions about this interaction because it may be spurious.

Analyses also tested for interactions between expectations and attributions on

changes in satisfaction. Consistent with the idea that positive expectations should lead to declines when they are likely to get disconfirmed, positive interactions between expectations and attributions would suggest that changes in satisfaction are most negative when expectations are positive and attributions scores negative. To test this possibility, each sub-scale of the RAM was entered separately into a second-stage analysis along with each measure of expectations and an interaction term to predict the changes in satisfaction estimated in equation 1. The effect sizes of the interactions between each measure of attributions and each measure of global expectations are reported in the second and third columns of Table 13. Results reveal a significant pattern of positive interactions for both types of attributions. Figure 3 provides a representation of the interaction between wives' expectations for their partner and their attributions for








responsibility, and Figure 4 provides a representation of the interaction between wives' expectations for steadiness and their attributions for causality. As can be seen in these figures, when spouses made negative attributions, positive expectations were associated with the steepest declines in satisfaction whereas less positive expectations were associated with the most stable satisfaction. However, when spouses made positive attributions, less positive expectations were associated with the steepest declines in satisfaction whereas the most positive expectations were associated with the most stable marital satisfaction. In other words, positive expectations led to negative changes in marital satisfaction when spouses made attributions for their partner's behavior that were likely to disconfirm their expectations, but led to positive changes in marital satisfaction when they made attributions that were likely to confirm their expectations. Although more interactions between expectations and attributions were significant for wives than for husbands, in general husbands and wives did not differ significantly in the extent to which their expectations and their attributions interacted to affect changes in their marital satisfaction over time. The only exception was that that the interactive effect for husbands' expectations for problems and husbands' responsibility attributions on changes in satisfaction according to the SMD was significantly different from the interactive effect for wives, t = 3.8, p < .05








Table 1. Descriptive Statistics for Independent Variables at Time 1 Husbands Wives
Possible Range M SD M SD Global Expectations

Expectations for growth 2 - 12 10.1 1.8 10.3 1.6

Expectations for stability 1 - 5 4.9 0.36 4.9 0.30 Specific Expectations

Expectations for steadiness 2 - 12 9.0 2.1 9.1 2.1 Expectations for partner 9 - 63 45.5 6.5 43.9 8.6 Expectations for problems 11 - 57 42.7 4.7 41.9 3.8 Mediators

Attributions for causality 12 - 84 53.3 11.2 51.4 9.9 Attributions for responsibility 12 - 84 63.6 12.1 60.8 14.6

Negative behavior 0% - 100% 10.2% 10.2% 13.4% 11.8%







Table 2. Correlations among Independent Variables at Time I


1 2 3 4 5 6 7 8 Global Expectations
Expectations for growth 0.20+ 0.20+ 0.19+ 0.27* 0.03 -0.05 0.03 Expectations for stability -0.03 --- 0.22* 0.18 0.14 0.15 0.21+ 0.05

Specific Expectations
Expectations for steadiness 0.12 0.39" --- 0.36** -0.13 0.24* 0.24* -0.03
Expectations for partner -0.04 0.23* 0.33** --- 0.03 0.27* -0.29** -0.02
Expectations for problems 0.42** -0.08 -0.08 -0.21+ --- -0.20+ 0.27* -0.17

Mediators
Causality attributions -0.04 0.09 0.26* 0.34** 0.15 --- 0.62"* 0.03

Responsibility attributions -0.02 0.17 0.31 ** 0.22* -0.04 0.56** --- -0.04

Negative behavior -0.01 -0.25* -0.26* -0.11 -0.05 -0.20+ -0.40** --Note. Husbands' correlations are above the diagonal and wives' correlations are below the diagonal. +p <.10, "p < .05, ** p < .01, two tailed








Table 3. Cross-Spouse Correlations among Independent Variables at Time 1 WIVES
1 2 3 4 5 6 7 8 HUSBANDS

Global Expectations

Expectations for growth 0.22* 0.11 0.29"� 0.19 0.08 0.26* 0.11 0.01 Expectations for stability 0.14 0.21+ 0.22* -0.08 0.00 -0.07 0.04 0.05 Specific Expectations

Expectations for steadiness 0.03 0.25* 0.53"* 0.33** -0.23* 0.15 0.05 0.01 Expectations for partner -0.11 0.32* 0.19 0.35"" -0.08 0.11 -0.04 0.01

Expectations for problems 0.12 0.01 -0.05 -0.14 0.19+ -0.02 -0.05 -0.07

Mediators

Causality attributions -0.07 0.13 0.24* 0.19+ -0.11 0.25* 0.08 -0.06

Responsibility attributions -0.07 0.12 0.12 0.12 -0.14 -0.03 -0.09 -0.02

Proportion of negative behavior 0.05 0.01 -0.14 -0.09 -0.08 -0.10 -0.31 0.57**
Note. p <.10, "p <.05, ' <.01, two tailed








Table 4. Associations between Expectations and Mediating Variables, Controlling for Marital Satisfaction Negative Responsibility Causality Behavior Attributions Attributions

Husband Wife Husband Wife Husband Wife Global Expectations

Expectations for trajectory 0.07 -0.05 -0.12 0.02 -0.04 0.00 Expectations for stability 0.10 -0.12 0.13 0.03 0.06 -0.07 Specific Expectations

Expectations for steadiness 0.03 -0.08 0.14 0.16 0.12 0.09 Expectations for partner 0.08 0.03 -0.14 -0.10 -0.09 -0.24* Expectations for problems -0.18 -0.11 0.26* "0.01 0.20+ 0.12 Note. +p <.10, 1p <.05, p < .01, two tailed








Table 5. Mean Marital Quality Scores Across Five Waves of Measurement for Husbands and Wives

Spouse Time 1 Time 2 Time 3 Time 4 Time 5

Semantic Differential
Husbands
M 96.3 92.0 92.5 92.1 93.5 SD 8.8 14.1 14.8 14.7 13.9 N 81 76 74 67 64 Wives
M 97.7 94.8 93.3 92.1 93.8 SD 10.7 12.9 16.0 14.7 15.6 N 82 77 73 68 66

Quality Marriage Index
Husbands
M 42.1 40.3 40.3 40.1 40.8 SD 4.0 6.7 6.6 7.0 6.5 N 81 76 74 67 64 Wives
M 42.1 40.8 39.9 39.0 40.2 SD 5.3 6.0 7.5 8.6 7.3 N 82 77 73 68 66

Inventory of Specific Marital Problems
Husbands
M 52.0 49.1 48.1 48.7 48.6 SD 24.0 24.1 22.6 24.0 25.8 N 82 75 74 67 64 Wives
M 48.2 44.2 48.6 46.4 44.0 SD 24.4 20.4 26.8 24.0 21.5 N 82 76 73 68 66 Note. Means using only participants reporting at time 5 are similar and demonstrate similar trends.








Table 6. Change in Marital Satisfaction


Chi-Square test
Measure M SD ta of variance

Intercept
SMD
Husbands 95.7 1.1 149.4"* Wives 97.9 1.3 196.0"** QMI
Husbands 41.9 0.52 126.0"* Wives 42.3 0.60 166.6"**

Slope
SMD
Husbands -0.78 0.32 -2.4* 131.3** Wives -1.2 0.36 -3.3 161.5*** QMI
Husbands -0.38 0.15 -2.5* 114.8** Wives -0.62 0.19 -3.3"* 181.8***
Note. For t test, df = 81; for chi-square tests, df = 75. SMD = Semantic Differential (Osgood et al., 1957); QMI = Quality Marital Index (Norton & Glick, 1979). a The t test of the intercepts addresses the hypothesis that the intercepts differ significantly form zero. Because the lowest possible score on each of these measures is greater than zero, these tests are not meaningful and hence are not reported. +1p < .10, *p <.05, **p <.01 *** p <.001, two tailed








Table 7. Covariance between Specific Marital Problems and Marital Satisfaction Controlling for the Trajectory of Marital Satisfaction

Chi-Square test
Measure M SD t of variance

SMD
Husbands -0.27 0.03 -8.5 147.5** Wives -0.30 0.03 -9.7* 126.6** QMI
Husbands -0.10 0.02 -6.7* 123.7*"
Wives -0.12 0.01 -9.4" 192.3*"
Note. For t test, df = 81; for chi-square tests, df = 70. SMD = Semantic Differential (Osgood et al., 1957); QMI = Quality Marital Index (Norton & Glick, 1979).
p .10 p <.05 p <.01 p < .001, two tailed








Table 8. Effects of Specific Expectations on the Trajectory of Marital Satisfaction


Husband Wife

t effect size t effect size

Satisfaction intercepts

Expectations for steadiness
SMD 3.6 0.37 2.9 0.31 QMI 3.0* 0.32 2.7* 0.29 Expectations for partner
SMD 4.8 0.47 1.6 0.18 QMI 4.3 0.43 1.6 0.18 Expectations for problems
SMD 0.16 0.02 -0.33 -0.04 QMI 0.40 0.04 -1.0 -0.11

Satisfaction slopes

Expectations for steadiness
SMD -1.6 -0.18 -1.6 -0.18 QMI -1.8+ -0.20 -2.3* -0.25 Expectations for partner
SMD -2.4 -0.26 0.30 0.03 QMI -2.5* -0.27 0.02 0.00 Expectations for problems
SMD -0.16 -0.02 -2.1 -0.23 QMI -0.28 -0.03 -0.82 -0.09 Note. + p <.10 * p <.05 p < .01 p <.001, two tailed








Table 9. Effects of Specific Expectations on the Covariance between Specific Problems and Marital Satisfaction


Husband Wife

t effect size t effect size

Expectations for steadiness
SMD -0.71 -0.08 -1.5 -0.17 QMI -0.30 -0.03 -1.5 -0.17 Expectations for partner
SMD 0.54 0.06 -0.76 -0.08 QMI 0.65 0.07 -0.75 -0.08 Expectations for problems
SMD 1.7 0.19 -0.69 -0.08 QMI 1.6 0.18 0.36 0.04 Note. +p< .10 * p< .05 p <.01 p <.001, two tailed








Table 10. Effects of Global Expectations on the Trajectory of Marital Satisfaction.


Husband Wife

t effect size t effect size

Satisfaction intercepts
Expectations for growth
SMD 2.9 0.31 -0.76 -0.08 QMIl 2.9** 0.31 -0.41 -0.05 Expectations for stability
SMD 2.5* 0.27 1.7+ 0.19
QMI 2.5* 0.27 1.6 0.18

Satisfaction slopes

Expectations for growth
SMD -2.5* -0.27 0.23 0.03 QMI -2.9** -0.31 0.12 0.01 Expectations for stability
SMD -1.8+ -0.20 -0.23 -0.03 QMI -2.9** -0.31 -0.52 -0.06


p < .001, two tailed


Note. ' p < .10 ' p < .05 "" p < .01








Table 11. Effects of Global Expectations on the Covariance between Specific Problems and Marital Satisfaction.


Husband Wife

t effect size t effect size

Expectations for growth
SMD -1.8+ -0.20 -0.88 -0.10 QMI -1.3 -0.14 -0.29 -0.03 Exectations for stability
SMD -0.73 -0.08 -2.7" -0.29 QMI -1.5 -0.17 -1.6 -0.18 Note. +p < .10 * p < .05 p < .01 p < .001, two tailed








Table 12. Interactions between Specific Expectations and Global Expectations


Specific Expectations

Steadiness Partner Problems Global Expectations Husband Wife Husband Wife Husband Wife

Trajectory of Marital Satisfaction
Growth
SMD 0.12 -0.05 -0.04 -0.01 0.08 -0.09 QMI 0.08 -0.10 0.13 -0.05 0.06 -0.06 Stability
SMD 0.04 0.24* 0.16 0.04 0.33** 0.23* QMI 0.13 0.28* 0.22* 0.02 0.18 0.15 Note. 'p <.10 " p <.05 "p <.01 "p< .001, two tailed








Table 13. Moderating Effects of Negative Behavior and Attributions on the Association between Expectations and the Trajectory of Satisfaction

Negative Responsibility Causality Behavior Attributions Attributions

Husband Wife Husband Wife Husband Wife

Expectations for growth
SMD -0.09 0.22* 0.15 0.01 0.06 -0.03 QMI 0.10 0.08 0.15 0.10 0.13 0.06 Expectations for stability
SMD -0.05 -0.23* 0.02 0.31 0.01 0.23* QMI -0.09 -0.24* 0.11 0.20+ 0.06 0.12 Expectations for steadiness
SMD 0.16 -0.17 0.02 0.23* 0.20+ 0.34** QMI 0.05 -0.20+ 0.07 0.18 0.19+ 0.23* Expectations for partner
SMD 0.16 -0.26* 0.05 0.30* 0.11 0.13 QMI 0.09 -0.19+ 0.07 0.22* 0.15 0.12 Expectations for problems
SMD -0.18 0.02 0.26* -0.08 0.08 0.05 QMI -0.11 0.02 0.13 -0.10 0.04 -0.08 Note. +p < .10 * p < .05 "" p < .01l' p < .001, two tailed









Figure A Figure B


-nepkm 8- - e C * q~


C
-Q2



0

-Q6

0


P~dm


* 4 - eddw
ep as

-nepbtem


0

CO

Cn-1.5
a) -2


c-25
Ca
-C
0


Figure 1. Interactive Effects of Global and Specific Expectations on the Trajectory of Marital Satisfaction


nve Podm


pdm


P~dm


43Edfic epxgm


Figure A


Figure B


qm6c-ePticrs











- 4 - less positive
expectations
U "more positive
expectations


high low
proportion p rion
negative beWPv11orr


0
-0.2
-0.4
-0.6
-0.8
-1.2


Figure 2. Interactive Effects of Wives' Expectations for Stability and Negative Behavior on the Trajectory of Their Marital Satisfaction











� 4 - less positive
expectations
- more positive
expectations


positive


responsibility attributions


negative


Figure 3. Interactive Effects between Wives' Expectations for Partner and Responsibility Attributions on the Trajectory of Their Marital Satisfaction





68


.2 4-6.4
2
0- 0 " less positive M expectations
2 -2
-4 - more positive (D expectations

U
negative positive causality attributions


Figure 4. Interactive Effects of Wives' Expectations for Steadiness and Causality Attributions on the Trajectory of Their Marital Satisfaction













DISCUSSION


Study Rationale and Summary of Results

Given that expectations influence a wide variety of interpersonal phenomena

(Olson et al., 1998), the expectations that spouses have for their relationships should have implications for the development of those relationships over time. Consistent with this idea, research on expectancy confirmation suggests that positive expectations should lead to positive outcomes because they are likely to get confirmed through the processes of behavioral and perceptual confirmation (see Darley & Fazio, 1980; Miller & Turnbull, 1986; Snyder, 1984, for reviews). In contrast, however, research on counterfactual reasoning suggests that positive expectations are likely to lead to disappointment when they are disconfirmed (Kahneman & Miller, 1986). In light of these findings, the current study contended that the benefits of positive expectations may depend on the likelihood that they are confirmed or disconfirmed. Positive expectations should lead to positive feelings about the relationship when they are likely to get confirmed and lead to less positive feelings about the relationship when they are likely to get disconfirmed. This dissertation examined whether the specificity of the expected outcome may be associated with the likelihood that expectations are confirmed or disconfirmed, and therefore influence future relationship outcomes. Global outcomes, because they are more open to interpretation and capable of being confirmed by a wide variety of behaviors, may be








more likely to be confirmed. Specific outcomes, in contrast, because they are less open to interpretation and capable of being confirmed by only a limited variety of behaviors, may be more likely to be disconfirmed.

The first goal of this study was to test the idea that expectations about outcomes at different levels of specificity should be differentially likely to lead to perceptual and behavioral confirmation. In pursuit of this goal, Hypothesis I stated that measures of global expectations would be positively associated with marital attributions (perceptual confirmation) and negatively associated with observations of negative behavior (behavioral confirmation). In other words, spouses who expected global outcomes to be positive should have perceived and interpreted events in positive ways and behaved less negatively during their interactions with their partners. This prediction was not supported. Measures of global expectations were not significantly associated with attributions or with observations of negative behavior. In other words, spouses with positive expectations about global outcomes were no more likely to make positive attributions and no less likely to behave negatively than spouses with less positive global expectations.

Hypothesis 2 stated that expectations about more specific outcomes, because they are less likely to get confirmed, should not lead to perceptual confirmation and behavioral confirmation. Thus, it was predicted that measures of specific expectations would be unrelated to measures of marital attributions and observations of negative behavior. This prediction was supported. Measures of specific expectations were not significantly associated with measures of attributions and observations of behavior.

Thus, in contrast to the general idea that specific and global expectations should be differentially likely to get confirmed by attributions and behavior, results suggest that








both global and specific expectations were likely to get disconfirmed. There are several possible explanations for this finding. First, measures of attributions and behaviors may have been relatively weak operationalizations of perceptual and behavioral confirmation. For instance, the RAM assessed spouses' attributions about specific events, but did not assess their attributions about global events. Perhaps a measure that assessed spouses' attributions about more global outcomes would show that such attributions are associated with expectations about global outcomes. Second, measures of global and specific expectations may have been relatively weak operationalizations. Although the measures of global expectations were conceptually distinct, several correlations between global and specific measures were just as large as correlations among global measures and among specific measures, suggesting that the measures may have been too similar. Future research using measures that tap outcomes even more global than the outcomes assessed in the current study may reveal support for the idea that global expectations are more likely to get confirmed through perceptual and behavioral confirmation. One better way to measure global expectations may be to have spouses predict their future responses to measures of global marital satisfaction.

The second goal of this dissertation was to examine the effects of positive

expectations on changes in marital satisfaction and spouses' reactivity to relationship problems. Hypothesis 3 stated that positive expectations about specific outcomes, because they are more likely to get disconfirmed, would lead to negative changes in satisfaction and a stronger reactivity to relationship problems. The first part of this hypothesis was supported. Although measures of specific expectations were positively associated with initial marital satisfaction, positive specific expectations led to declines in








marital satisfaction over the first two and a half years of marriage. In other words, spouses who held the most positive expectations about specific outcomes were the most happy initially, but were also the most likely to experience declines in marital satisfaction over time. With respect to the second part of Hypothesis 3, positive specific expectations did not lead spouses to be more reactive to relationship problems. Measures of specific expectations were not associated with the covariance between satisfaction and relationship problems, or the extent to which their global satisfaction changed in response to specific problems.

Hypothesis 4 stated that positive expectations about global outcomes, because they are more likely to get confirmed, would lead to positive changes in marital satisfaction and less reactivity to relationship problems. This hypothesis was not supported. Instead, for husbands, although measures of global expectations were positively associated with initial satisfaction, positive expectations about global outcomes led to declines in satisfaction over the first two and a half years of marriage, similar to the effect for specific expectations. Husbands who held the most positive expectations about global outcomes were the most happy initially, but were the most likely to experience declines in marital satisfaction over time. Although the associations between measures of global expectations and decreases in satisfaction were only significant for husbands, there were no significant gender differences in these associations. With respect to the second part of Hypothesis 4, positive global expectations did not lead spouses to be less reactive to relationship problems. Instead, a few measures of global expectations were significantly associated with the covariance between satisfaction and relationship








problems, suggesting husbands and wives with positive expectations about some global outcomes were more reactive to relationship problems.

Thus, regardless of their level of specificity, positive expectations led to decreases in marital satisfaction, suggesting that both positive global and positive specific expectations were likely to be disconfirmed. However, further analyses revealed that when positive expectations were likely to be confirmed, they led to positive changes in satisfaction. Results revealed a consistent pattern of significant interactions between expectations and two measures of relationship quality likely to be associated with the likelihood that positive expectations were or disconfirmed: negative behavior and attributions. When behavior was least negative or when attributions were most positive, likely confirming positive expectations, positive expectations led to more stable satisfaction over time and less positive expectations led to declines in satisfaction over time, suggesting that positive expectations may be beneficial. In contrast, when behavior or attributions were most negative, likely disconfirming positive expectations, less positive expectations led to more stable satisfaction over time whereas more positive expectations led to declines in marital satisfaction. Importantly, although less positive expectations were associated with more stable satisfaction for spouses likely to encounter negative experiences, because expectations were also positively associated with satisfaction initially, spouses with less positive expectations were less satisfied initially. In other words, when less positive expectations maintained satisfaction, they did so in less satisfying relationships.

In sum, the findings of the current research suggest that the effects of positive expectations on relationship outcomes may depend on the likelihood that such








expectations will get confirmed. When positive expectations are readily confirmed, they lead partners to experience stable relationship satisfaction over time. In contrast, when positive expectations are likely to get disconfirmed, they lead spouses to experience declines in marital satisfaction over time.


When Are Positive Expectations Likely to Get Confirmed?

One factor that may affect the likelihood that positive expectations will get

confirmed is the quality of the relationship. The current study showed that two aspects of relationship quality, negative attributions and negative behavior, moderated the association between expectations and marital happiness. However, in addition to these factors, there are likely to be other factors associated with marital quality that lead positive expectations to get disconfirmed. Because researchers and therapists must have a complete understanding of these factors in order to know the implications of positive expectations in relationships, these findings highlight the importance of research to develop a better understanding of these factors.

In addition to qualities of the relationship, qualities of the expected outcomes may affect the extent to which positive expectations get confirmed. The current study argued that the specificity of the expected outcome was one factor that should affect the likelihood that an expectation is confirmed. This idea was not supported. Instead, global and specific expectations had similar effects on relationship satisfaction and spouses' reactivity to relationship problems. As discussed earlier, one reason that measures of global and specific expectations may not have demonstrated different effects on relationship outcomes may be that measures of global and specific expectations did not validly capture the distinction between them. Future research may benefit by examining








the effects of global and specific expectations using measures that more properly capture this difference.

In addition to the specificity of the outcome, the controllability of the expected outcome may affect the extent to which expectations get confirmed. Research on behavioral confirmation shows that spouses behave in a manner consistent with their prior expectations thereby creating outcomes consistent with their expectations (Darley & Fazio, 1980; Snyder, 1984). Thus, expectations about outcomes that are controllable may be more likely to get confirmed than expectations about outcomes that are not controllable. Previous research offers support for this idea. As discussed previously, research on expectancy confirmation in marital interactions has revealed that spouses' expectations for an upcoming problem-solving discussion were not related to the way they behaved during the discussion (McNulty & Karney, under review). The authors speculated that problem-solving behavior is a skill that develops independently of expectations, and thus is not under some spouses' control. Future research may benefit by examining the how the controllability of the expected outcomes influences the extent to which expectancy confirmation processes are likely to operate.

Some authors have argued that positive beliefs are beneficial only to the extent that they are realistic (see Baumeister, 1989, for related discussion). Consistent with this idea, the qualities of the relationship and qualities of the expected outcome that make some expectations more likely to get confirmed than other outcomes may make some expectations more realistic than others. When qualities of the relationship are likely to produce positive outcomes, or when positive outcomes are under spouses' control, positive expectations are realistic and beneficial to the relationship. In contrast, when








qualities of the relationship are unlikely to render spouses vulnerable to negative experiences, or when positive outcomes are not under spouses' control, positive expectations are somewhat unrealistic and may be detrimental to the relationship. In this light, realistic positive expectations are likely to lead to positive outcomes, whereas unrealistic positive expectations are likely to lead to disappointment.


Additional Directions for Future Research

The current study showed that spouses' expectations about their relationships affected their own satisfaction over the first two and a half years of marriage. Spouses' expectations may also influence their partners' satisfaction over time. Consistent with this idea, some research suggests that spouses' expectations lead expectancy-consistent outcomes for their partners (Downey et al., 1998; Drigotas, Rusbult, Wieselquist, & Whitton, 1999). For example, Downey et al. (1998) showed that spouses who expected their partners to behave negatively, responded in ways that predicted greater negative feelings in their partners. It is also possible, however, that holding unrealistic positive expectations about the partner or the relationship may lead to negative outcomes for partners by putting strain on the partner to meet the expectations. Consistent with this idea, spouses often report that their partners' unrealistic expectations are a problem in the relationship (Geiss & O'Leary, 1981). Future research may help reconcile these inconsistent findings by revealing the conditions under which positive expectations are likely to lead to positive versus negative outcomes for partners. As was the case in the current study, it is possible that the effects of positive expectations may depend on the spouses' beliefs about likelihood that the expectation can be confirmed. When spouses feel they that they can meet their partners' high expectations, such expectations may








make them feel good. However, when spouses feel that they can not meet their partners' high expectations, such expectations may be stressful (see Neff & Karney, in press, for related discussion).

Future research may also benefit by examining the sources of expectations. Given the link between expectations and changes in marital satisfaction, the sources of expectations are likely to be distal causes of change in marital satisfaction. Because people learn what to expect through their experiences (Olson et al., 1998), one source of spouses' expectations about their relationship may be their experiences with the relationship. Consistent with this idea, research by Fincham, Harold, and Gano-Phillips (2000) suggests that spouses' explanations of the events in their relationships influence their future expectations. In particular, spouses who made more positive attributions were more likely to believe that they had the ability to resolve conflicts with their partners and were happier with their relationships. Thus, longitudinal research may reveal that specific relationship experiences are important distal sources of change in marital satisfaction.

In addition to their experiences with their current relationships, another source of spouses' expectations about the relationship may be spouses' experiences with other relationships. Research and theory on attachment style (Ainsworth, Blehar, Waters, & Wall, 1978; Bowlby, 1969; Hazan & Shaver, 1987) suggests that the relationship that spouses had with their primary caregiver affects the expectations they develop about their current relationships (Baldwin, Keelan, Fehr, Enns, & Koh-Rangarajoo, 1996). In this study, partners who had positive relationships with their primary caregivers and therefore developed positive attachments styles held positive expectations about their relationships, whereas spouses who had more negative relationships with their primary caregivers and








therefore developed more negative attachment styles held more negative expectations about their relationships. Thus, given that attachment style has been linked to marital satisfaction and stability (see Collins & Allard, 2001, for a review), expectations may mediate the effects of attachment style on changes in marital satisfaction. Longitudinal research is needed to evaluate this possibility.

Another source of spouses' expectations about their relationships may be their parents' relationships with one another. Spouses may develop their own beliefs about relationships based on their experiences with their parents' relationships. If their parents' relationship was positive, they may develop positive expectations about relationships. On the other hand, if their parents' relationship was negative, they may develop less positive expectations about relationships. Research has provided some preliminary evidence consistent with this idea. In one study, students' optimism about their own future marriages was positively associated with their perceptions of the quality of their parents' relationship (Carnelley & Janoff-Bulman, 1992). Future research could expand on this finding by examining whether partners perceptions of their parents' relationships affects the quality other own relationships, and whether expectations mediate this association.

In addition to broader beliefs, personality characteristic may also lead people to expect various relationship outcomes. If so, given that personality has been shown to be associated with relationship satisfaction [e.g., \(Huston & Houts, 1998; Huston & Vangelisti, 1991; Karney & Bradbury, 1997; Karney, Bradbury, Fincham, & Sullivan, 1994), expectations may mediate the effects of various personality characteristics on future relationship satisfaction. Moreover, it is also possible that the effects of personality on satisfaction may completely account for the association between expectations and








satisfaction observed in the current study. Spouses' personality characteristics may lead them to hold positive or less positive expectations and to experience steep or flat trajectories of satisfaction, making any observed associations between expectations and satisfaction spurious. For instance, research suggests that spouses high in neuroticism are likely experience negative relationship outcomes (Karney & Bradbury, 1997). Because neurotic spouses may also expect less positive outcomes, any observed association between expectations and satisfaction may be due to shared variance between expectations and neuroticism. This possibility is important to consider because it has implications for therapy. Interventions aimed at changes expectations will be futile if expectations are not uniquely associated with satisfaction. Thus, future research may benefit by examining the unique effects of expectations and personality on satisfaction.


Strengths and Limitations

Several strengths of this research enhance my confidence in the findings. First, the general pattern of results replicated across conceptually distinct measures of expectations and conceptually distinct measures of disconfirmation, suggesting that results were not unique to one construct or one method of measurement. Second, whereas the average rate of attrition in prior research on marriage is 31% (Karney & Bradbury, 1995), during the two years of this study attrition was relatively low (21%), reducing the likelihood the results can be attributed to attrition bias. Third, all spouses entering the study were newlyweds, reducing the risk that important differences were confounded with marital duration (see, Booth, Johnson, White, & Edwards, 1985; South & Spitze, 1986). Fourth, whereas the majority of longitudinal research on marriage has collected and analyzed








only two waves of data, the current analyses examined five waves of data and analyzed all waves simultaneously using GCA.

Despite these strengths, several factors nevertheless limit interpretations of the

current findings. First, whereas the homogeneity of this sample enhances our confidence in the pattern of associations, generalizations to other samples should be made with caution. For example, although the expectations spouses have in the beginning of their relationships influence the later development of their relationships, the expectations of more established couples may be more accurate and thus less likely to be disconfirmed (see Jussim & Eccles, 1995). Second, the homogeneity of the sample was also limited the range of expectations assessed in the current study. Because all couples assessed in the current study were newlyweds, they were unlikely to hold negative expectations about their relationships. Research examining more negative expectations may yield effects different from those obtained in the current study. For instance, although the current study revealed that less positive expectations were more adaptive than very positive expectations for couples likely to experience negative outcomes, it is unlikely that negative expectations would also be adaptive. Instead, negative expectations may be more likely to lead couples to abandon the efforts to resolve relationship problems leading to the eventual dissolution of the relationship (see Knee, 1998). Third, although sample size compared favorably to other longitudinal studies of marriage, a larger sample size 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, associations between global expectations and perceptual and behavioral confirmation may have been significant.








Conclusion

Several writers have argued that spouses should lower their expectations to avoid being disappointed by outcomes that do not meet them (e.g., Talbot, 1997). Results of the current study argue against this recommendation. Spouses who held positive expectations that were likely to get confirmed by positive relationship experiences were more satisfied over time than spouses who held less positive expectations. Furthermore, although spouses who held less positive expectations that were unlikely to get confirmed did maintain a steadier trajectory of satisfaction over time, because expectations were positively associated with initial satisfaction, these spouses were less happy initially. Thus, less positive expectations only maintained satisfaction in less satisfying relationships. Because results suggest that the benefits of positive expectations depend on other relationship factors, the current findings highlight the importance of taking into account the broader factors associated with the likelihood that expectations are confirmed and disconfirmed in close relationships.














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 BAD UNPLEASANT FULL

WEAK SATISFIED LONELY STURDY

REWARDING DISCOURAGING ENJOYABLE TENSE

STABLE HAPPY
STRESSFUL


0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0


0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0


BORING GOOD PLEASANT EMPTY

STRONG
DISSATISFIED FRIENDLY FRAGILE

DISAPPOINTING HOPEFUL MISERABLE RELAXED

UNSTABLE SAD
PEACEFUL














APPENDIX B
QUALITY MARRIAGE INDEX

Please indicate how well the following statements describe you and your marriage.

Very Strong Very Strong DISAGREEMENT AGREEMENT
1 2 3 4 5 6 7


We have a good marriage. My relationship with my partner
is very stable. Our marriage is strong. My relationship with my partner
makes me happy.

I really feel like part of a team
with my partner.


0 0 0 0 0 0 0


0 0 0 0 0 0 0 0 0 0 0 0 0 0


All things considered, how happy are you in

Very UNHAPPY
1 2 3 4 5 0 0 0 0 0


your marriage? Happy
6 7

0 0 4


Perfectly HAPPY
9 10 0 0













APPENDIX C
INVENTORY OF SPECIFIC MARITAL PROBLEMS

All couples experience some difficulties or differences of opinion in their marriage, even if they are only very minor ones. Listed below are a number of issues that might be difficulties in your marriage. For each issue fill in a bubble to indicate how much it is a source of difficulty or disagreement for you and your spouse.


Not a Problem
1 2 3 4


0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0


Major Problem
5 6 7 R 9 10 11


Children
Religion
In-laws, parents, relatives Recreation and leisure time

Communication Household management Showing Affection Making decisions

Friends
Unrealistic expectations Money management


Jealousy Solving problems Trust
Independence


Drugs and alcohol Career decisions Amount of time spent together













APPENDIX D
EXPECTATIONS FOR FUTURE RELATIONSHIP SATISFACTION Please fill in the circles that best describe your opinions about your relationship:

1. Over the NEXT six months, do you expect that your overall feelings about your
marriage will become:

O 0 0 0 0

Much worse A little Stay A little Much better worse the same better


2. Over the NEXT FOUR YEARS, do you expect that your overall feelings about
your marriage will become:

0 0 0 0 0

Much worse A little Stay A little Much better worse the same better














APPENDIX E
EXPECTED TRAJECTORY OF MARITAL SATISFACTION

Please fill in the circle beneath the picture that best describes how you think your relationship will go during the next SIX MONTHS.


0


0


0


0


0











APPENDIX F EXPECTATIONS FOR STABILITY
I. How likely is it that the two of you will still be married four years from now?
O Very unlikely O Not too likely O Pretty likely O Very likely O Completely certain














APPENDIX G
EXPECTATIONS FOR STEADINESS

Over the NEXT six months, which of the following describes how your feelings towards your marriage are likely to change:


Major highs/ major lows


Some ups and downs


A few ups and downs


Pretty steady


Very steady


2. Over the NEXT FOUR YEARS, which of the following describes how your
feelings towards your marriage are likely to change:


Major highs/ major lows


Some ups and downs


A few ups and downs


Pretty steady


Very steady













APPENDIX H
EXPECTATIONS FOR PARTNER

In response to the following statements, indicate how strongly you agree by filling in the appropriate bubble.
Strongly Strongly Disagree Agree
1 2 3 4 5 6 7
My partner will always take time for me when I need him/her. 0 0 0 0 0 0 0

My partner will rarely make mistakes. 0 0 0 0 0 0 0


My partner will agree with me about the important things.

My partner will always get along well with my parents.

My partner will sometimes lose his/her temper. My partner will never disappoint me. My partner will always take care of me. My partner will always be attractive to me. My partner will always make me happy.


0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0













APPENDIX I
EXPECTATIONS FOR SPECIFIC RELATIONSHIP PROBLEMS

For each of the marital problems listed below, please indicate how you think that problem is going to develop for you and your spouse over the NEXT SIX MONTHS.


Get Stay Get
worse the same better

Children 0 0 0 Religion 0 0 0 In-laws, parents, relatives 0 0 0 Recreation and leisure time 0 0 0


Communication 0 0 0 Household management 0 0 0 Showing Affection 0 0 0 Making decisions 0 0 0

Friends 0 0 0 Unrealistic expectations 0 0 0 Money management 0 0 0 Sex 0 0 0

Jealousy 0 0 0 Solving problems 0 0 0 Trust 0 0 0 Independence 0 0 0

Drugs and alcohol 0 0 0 Career decisions 0 0 0 Amount of time spent together 0 0 0














APPENDIX J
RELATIONSHIP ATTRIBUTIONS MEASURE


This questionnaire describes several things that your spouse might do. Imagine your spouse performing each behavior and then bubble in the number that indicates how much you agree or disagree with each statement, using the following scale:
1 2 3 4 5 6 7


disagree strongly


disagree


disagree somewhat


neutral


agree somewhat


YOUR SPOUSE CRITICIZES SOMETHING YOU SAY:

My spouse's behavior was due to something about him/her
(e.g., the type of person he/she is, his/her mood) ...........
My spouse's behavior was due to something about me
(e.g., the type of person I am, the mood I was in) ..........
The reason my spouse criticized me is not likely to change ......... The reason my spouse criticized me is something that
affects other areas of our marriage ......................
My spouse criticized me on purpose rather
than unintentionally ..................................
My spouse's behavior was motivated by selfish
rather than unselfish concerns ..........................
My spouse deserves to be blamed for criticizing me ..............


YOUR SPOUSE BEGINS TO SPEND LESS TIME WITH YOU:

My spouse's behavior was due to something about him/her
(e.g., the type of person he/she is, his/her mood) ...........
My spouse's behavior was due to something about me
(e.g., the type of person I am, the mood I was in) ..........
The reason my spouse spent less time with me is not
likely to change ....................................
The reason my spouse spent less time with me is something that
affects other areas of our marriage ......................
My spouse spent less time with me on purpose rather
than unintentionally ..................................
My spouse's behavior was motivated by selfish
rather than unselfish concerns ..........................
My spouse deserves to be blamed for spending less time with me ....


agree agree strongly


1234567

0000000 0000000 0000000 0000000 0000000 0000000 0000000


1234567

0000000 0000000 0000000 0000000 0000000 0000000 0000000








1
disagree strongly


2
disagree


3
disagree somewhat


4
neutral


YOUR SPOUSE DOES NOT PAY ATTENTION TO WHAT YOU ARE SAYING:


My spouse's behavior was due to something about him/her
(e.g., the type of person he/she is, his/her mood) ...........
My spouse's behavior was due to something about me
(e.g., the type of person I am, the mood I was in) ..........
The reason my spouse spent less time with me is not
likely to change ....................................
The reason my spouse spent less time with me is something that
affects other areas of our marriage ......................
My spouse spent less time with me on purpose rather
than unintentionally ..................................
My spouse's behavior was motivated by selfish
rather than unselfish concerns ......................
My spouse deserves to be blamed for spending less time with me ....


1234567

0000000 0000000 0000000 0000000 0000000 0000000 0000000


YOUR SPOUSE IS COOL AND DISTANT:


My spouse's behavior was due to something about him/her
(e.g., the type of person he/she is, his/her mood) ...........
My spouse's behavior was due to something about me
(e.g., the type of person I am, the mood I was in) ..........
The reason my spouse spent less time with me is not
likely to change ....................................
The reason my spouse spent less time with me is something that
affects other areas of our marriage ......................
My spouse spent less time with me on purpose rather
than unintentionally ..................................
My spouse's behavior was motivated by selfish
rather than unselfish concerns ..........................
My spouse deserves to be blamed for spending less time with me ....


1234567


0000000 0000000

0000000 0000000 0000000 0000000 0000000


5
agree somewhat


6
agree


7
agree strongly













REFERENCES

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage.

Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the Strange Situation. Hillsdale, NJ: Lawrence Erlbaum.

Anderson, S. M., & Bem, S. L. (1981). Sex typing and androgeny in dyadic
interaction: Individual differences in responsiveness to physical attractiveness. Journal of Personality and Social Psychology, 41, 74-86.

Armor, D. A., & Taylor, S. E. (1997). Mindset, prediction, and performance: The causes and consequences of optimism, Unpublished manuscript, University of California, Los Angeles.

Baldwin, M. W., Keelan, J. P. R., Fehr, B., Enns, V., & Koh-Rangarajoo, E.
(1996). Social-cognitive conceptualization of attachment working models: Availability and accessibility effects. Journal of Personality and Social Psychology, 71, 94-109.

Bandura, A. (1982). The psychology of chance encounters and life paths. American Psychologist, 37, 747-755.

Baucom, D. H., Epstein, N., Rankin, L. A., & Burnett, C. K. (1996). Assessing relationship standards: The Inventory of Specific Relationship Standards. Journal of Family Psychology, 10, 72-88.

Baumeister, R. F. (1989). The optimal margin of illusion. Journal of Social and Clinical Psychology. 8, 176-189.

Booth, A., Johnson, D. R., White, L. K., & Edwards, J. N. (1985). Predicting divorce and permanent separation. Journal of Family Issues, 6, 331-346.

Bowlby, J. (1969). Attachment and loss: Volume I. Attachment. New York: Basic Books.

Box, G. E. P., & Tiao, G. C. (1973). Bayesian inference in statistical analysis. Reading, MA: Addison-Wesley.








Bradbury, T. N., & Fincham, F. D. (1990). Attributions in marriage: Review and critique. Psychological Bulletin, 107, 3-33.

Bradbury, T. N., & Karney, B. R. (1993). Longitudinal study of marital
interaction and dysfunction: Review and analysis. Clinical Psychology Review, 13, 1527.

Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models: Applications and data analysis methods. Newbury Park: Sage.

Bryk, A. S., Raudenbush, S. W., & Congdon, R. T. (1994). HLM: Hierarchical linear modelling with the HLM/2L and HLM13L programs. Chicago: Scientific Software International.

Camelley, K. B., & Janoff-Bulman, R. (1992). Optimism about love relationships: General vs specific lessons from one's personal experiences. Journal of Social and Personal Relationships, 9, 5-20.

Chaiken, S. (1987). The heuristic model of persuasion. In M. P. Zanna & J. M. Olson & C. P. Herman (Eds.), Social Influence: The Ontario Symposium (Vol. 5, pp. 339). Hillsdale, NJ: Erlbaum.

Cohen, J., & Cohen, P. (1983). Applied multiple regression: Correlational analysis for the behavioral sciences (2nd Edition). Hillsdale, NJ: Erlbaum.

Collins, N. L., & Allard, L. M. (2001). Cognitive representations of attachment: the content and functioning of working models. In G. J. 0. Fletcher & M. S. Clark (Eds.), Blackwell Handbook of Social Psychology, Vol 2: Interpersonal Processes (pp. 60-85). Oxford: Blackwell Publishers Limited.

D'Angelo, C., & Kamey, B. R. (1999). A comparison between sampling methods in marital research. Paper presented at the Southeastern Psychological Association, Savannah, GA.

Darley, J. M., & Fazio, R. H. (1980). Expectancy confirmation processes arising in the social interaction sequence. American Psychologist, 35, 867-881.

Ditto, P. H., & Hilton, J. L. (1990). Expectancy processes in the health care interaction sequence. Journal of Social Issues, 46(2), 97-124.

Downey, G., Freitas, A. L., Michaelis, B., & Khouri, H. (1998). The self-fulfilling prophecy in close relationships: Rejection sensitivity and rejection by romantic partners. Journal of Personality and Social Psychology, 75, 545-560.




Full Text

PAGE 1

"SHOULD I EXPECT THE BEST OR BRACE FOR THE WORST?' THE EFFECTS OF NEWLY WEDS' EXPECTATIONS ON THEIR FUTURE RELATIONSHIP SATISFACTION By JAMES K. MCNULTY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2001

PAGE 2

ACKNOWLEDGEMENTS First off, I would like to thank the members of my committee, Benjamin Karney (chair). Jams Shepperd, Barry Schlenker, Mark Fondacaro, and James Algina, for the time and insights they have provided to this project. I would also like to thank my parents, James and Kathy McNulty. Without their continued support and encouragement, I would not be where I am today. Finally, I would like to thank my flitvire wife Melissa for the overwhelming love and support she has provided me throughout my graduate career. I look forward to the many years of happiness we have ahead of us. ii

PAGE 3

TABLE OF CONTENTS page ACKNOWLEDGMENTS " ABSTRACT v INTRODUCTION 1 REVIEW OF LITERATURE 4 Expectancy Confirmation 4 Positive Expectations and Disappointment 8 Reconciliation and the Distinction between Global and Specific Expectations 11 Overview of the Current Study 16 Hypotheses 17 METHOD 23 Participants 23 Procedure 24 Materials 25 Data Analysis 30 RESULTS 32 Descriptive Statistics and Correlations 32 Were Specific and Global Expectations Associated with Attributions and Negative Behavior? 35 Were Specific and Global Expectations Associated with Marital Outcomes Over Time? 37 Did Positive Expectations Lead to Positive Changes in Marital Satisfaction When They Were Likely to Get Confirmed? 47 DISCUSSION 69 Study Rationale and Summary of Results 69 When Are Positive Expectations Likely to Get Confirmed? 74 Additional Directions for Future Research 76 Strengths and Limitations 79 Conclusion 81 iii

PAGE 4

APPENDIX A SEMANTIC DIFFERENTIAL MEASURE OF MARITAL SATISFACTION 82 APPENDIX B QUALITY MARRIAGE INDEX 83 APPENDIX C INVENTORY OF SPECIFIC MARITAL PROBLEMS 84 APPENDIX D EXPECTATIONS FOR FURTURE RELATIONSfflP SATISFACTION 85 APPENDIX E EXPECTED TRAJECTORY OF MARITAL SATISFACTION 86 APPENDIX F EXPECTATIONS FOR STABILITY 87 APPENDIX G EXPECTATIONS FOR STEADINESS 88 APPENDIX H EXPECTATIONS FOR PARTNER 89 APPENDIX I EXPECTATIONS FOR SPECIFIC RELATIONSHIP PROBLEMS 90 APPENDIX J RELATIONSHIP ATTRIBUTIONS MEASURE 91 REFERENCES 93 BIOGRAPHICAL SKETCH 101 iv

PAGE 5

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 "SHOULD I EXPECT THE BEST OR BRACE FOR THE WORST?' THE EFFECTS OF NEWLYWEDS' EXPECTATIONS ON THEIR FUTURE RELATIONSHIP SATISFACTION By JAMES K. MCNULTY August, 2001 Chairperson: Benjamin R. Karney Major Department: Psychology What are the implications of positive expectations about a relationship? The answer to this question may depend on whether the expectations get confirmed or disconfirmed. Positive expectations should lead to positive outcomes to the extent that they are confirmed through processes of expectancy confirmation, but may lead to negative outcomes if they are disconfirmed. One factor that may be associated with whether expectations get confirmed or disconfirmed is the specificity of the outcome. Expectations about global outcomes, because they are more open to interpretation and more likely to lead to confirmatory behaviors, should be more likely to get confirmed, leading to positive feelings about the relationship over time. In contrast, expectations about specific outcomes, because they are less open to interpretation and less likely to lead to confirmatory behaviors, should be more likely to get disconfirmed, leading to less positive feelmgs about the relationship over time. This dissertation addressed this possibility by examining the implications of positive expectations about relationship V

PAGE 6

outcomes that vary in specificity. Newly married couples reported their expectations for global and specific relationship outcomes and subsequently reported their satisfaction with the relationship every six months for two years. Contrary to the prediction that global and specific expectations would have different effects on the trajectory of marital satisfaction, results indicated that, on average, positive expectations about both global and specific outcomes led to declines in marital satisfaction. However, consistent with the idea that the association between expectations and change over time would depend on the likelihood that the expectations would get confirmed, objective and subjective measures of relationship quality moderated the association between expectations and change in satisfaction over time. When positive expectations were confirmed by a tendency to make positive attributions and engage in few negative behaviors, positive expectations led to stable marital satisfaction. In contrast, when positive expectations were disconfirmed by a tendency to make negative attributions and engage in frequent negative behaviors, positive expectations led to declines in marital satisfaction. The current findings suggest that a clear understanding of the role of expectations in relationships will be limited without taking into account the broader factors likely to be associated with the confirmation and disconfirmation of positive expectations. vi

PAGE 7

INTRODUCTION < Things become better when you expect the best instead of the worst. Norman Vincent Peale, The power of positive thinking (p. 1 10) Blessed is the man that expects nothing, for he shall never be disappointed. Alexander Pope Contemporary wisdom suggests that the expectations partners have about their close relationships should play a role in determining whether their relationships will succeed or fail. Consistent with this idea, researchers have demonstrated that expectations influence a wide variety of interpersonal factors, including cognitions, behaviors, and emotions (Olson, Roese, & Zanna, 1998). Accordingly, what newlyweds expect from their relationships should affect how they ultimately feel about them. Ahhough both theory and research suggest that expectations should influence relationship outcomes, there is no consensus about the direction of the influence. Consistent with the idea that people who think about their relationship in positive terms should obtam positive outcomes, research on expectancy confirmation shows that people interpret events and behave in a manner consistent with their prior expectations (Darley & Fazio, 1980; Miller & Tumbull, 1986; Snyder, 1984). Thus, compared to those with more pessimistic expectations, people who expect positive things from their relationships should behave more constructively, interpret specific events more charitably, and 1

PAGE 8

2 therefore be happier over time. In contrast, consistent with the idea that positive expectations may be detrimental to relationships, research on counterfactual thinking demonstrates that people with positive expectations can be disappointed when their outcomes do not meet their expectations (e.g., Mellers, Schwartz, Ho, & Ritov, 1997). Accordingly, partners with positive expectations about their relationships may react more negatively to specific relationship problems, and consequently grow less happy over time. Understanding the factors that contribute to the functioning and stability of close relationships is important. Although all relationships begin very satisfying, most become less satisfying over time. Moreover, marital distress is the number one reason why people seek coxonseling (Veroff, Kulka, & Douvan, 1981). Yet, it remains unclear exactly what causes relationships to become unhappy and what may cause them to be happy again. Given the close connection between expectations and interpersonal outcomes, expectations frequently have been the target of interventions to promote healthy relationships (e.g., Baucom, Epstein, Rankin, &. Burnett, 1996; Eidelson & Epstein, 1982; Markman, Stanley, &. Blumberg, 1994; Olson, Foumier, & Druckman, 1986). Unfortunately, however, due to the lack understanding about the role of expectations in relationships, recommendations are often inconsistent with one another. For example, one approach to preventing marital distress argues that "couples are at increased risk when expectations are unreasonable" (Stanley, Blumberg, & Markman, 1999, p. 285). In contrast, other work on marital therapy techniques suggests that "holding extreme standards does not seem to doom couples to dissatisfaction. Instead . . . holding extreme

PAGE 9

3 standards that ask a great deal of the marriage is positively related to marital adjustment" (Baucom et al, 1996, p.83). The goal of this dissertation is to determine the longitudinal implications of positive expectations about a relationship on relationship outcomes. Specifically, this research addresses the question: is it adaptive to have positive expectations at the beginning of a relationship, or v^ill such optimism lead to disappointment? To this end, the remainder of this introduction is organized into four sections. The first section reviews research on the process of expectancy confirmation, whereby expectations lead to consistent outcomes and consistent evaluations, suggesting that positive expectations will lead to positive outcomes. The second section reviews the literature on counterfactual reasoning whereby positive expectations can lead to disappointment by providing a contrast to actual outcomes. The third section will attempt to reconcile these contradictory fmdings by distinguishing between expectations about global outcomes and expectations about more specific outcomes. The final section will describe a study designed to evaluate whether the distinction between global and specific expectations matters by examining the longitudinal implications of positive expectations about relationship outcomes that vary in specificity.

PAGE 10

LITRERATURE REVIEW Expectancy Confirmation One of the oldest and most well-established principles in social psychology is the idea that people's outcomes and evaluations are consistent with their prior expectations. In a classic study of this effect, Harold Kelley (1950) asked students to evaluate a guest lecturer whom they had been led to expect would be either "warm" or "cold." After participating in a group discussion, those students who expected a warm lecturer evaluated him more positively than students who expected a cold lecturer. In other words, students evaluated the lecturer in a manner consistent with their prior expectations. Rosenthal and Jacobson (1968) extended this finding by demonstrating that expectancy confirmations affect important social situations. These authors told elementary school teachers that based on results from an IQ test some of their students would "bloom" intellectually over the course of the academic year. In actuality, "bloomers" were randomly selected and thus no more likely to show increases in intelligence than the other students in the class. When tested with the IQ test again at the end of the school year, however, the students labeled "bloomers" demonstrated a significantly greater increase in IQ than students not labeled "bloomers." Since the early work of Kelley (1950) and Rosenthal and Jacobson (1968), researchers have shown that expectancy confirmations influence numerous other important social phenomena including stereotyped perceptions (e.g., Duncan, 1976; Hasorf & Cantril, 1954; Rosenhan, 1973), discriminatory behaviors 4

PAGE 11

5 (e.g., Word, Zanna, & Cooper, 1974), health factors (e.g.. Ditto «& Hilton, 1990; Taylor & Brown, 1988), close relationship processes (e.g.Downey, Freitas, Michaelis, & Khouri, 1998; McNulty & Kamey, under review; Murray, Holmes, & Griffin, 1996), and academic achievements (e.g., Bandura, 1982; Henderson & Dweck, 1990). Research has revealed two mechanisms through which expectations can be confirmed. One mechanism, behavioral confirmation , refers to the idea that prior expectations for an event give rise to expectancy consistent behavior during the event (Merton, 1948). For example, in the Kelley (1950) study, one reason students expecting a warm lecturer may have evaluated him more positively is that students' prior expectations led them to behave more positively towards the lecturer, encouraging him to behave more positively in return (also see Word et al., 1974). Follow-up research on the confirmation of teachers' expectations about their students' performance provides evidence of behavioral confirmation. Harris and Rosenthal (1985) conducted a metaanalysis of 135 studies of teacher expectancy effects and found that 16 different behaviors (e.g., praise, input, ignoring) significantly mediated the effects of teachers' expectations on students' performance. A second mechanism, perceptual confirmation , refers to the idea that prior expectations lead people to interpret the details of an event in ways that give rise to expectancy consistent evaluations of the event (Darley & Fazio, 1980; Snyder, 1984). For example, an alternative explanation for the results of the Kelley study is that students expecting a warm lecturer interpreted his behavior more charitably, even if they behaved no differently toward the lecturer and the lecturer behaved no differently toward them. Research on the effects of social stereotypes has confirmed that preexisting beliefs can

PAGE 12

6 exert a powerful effect on people's perceptions and evaluations of others through perceptual confirmation (for a review, see Higgins & Bargh, 1987). In one study, white participants interpreted an "ambiguous" shove as more hostile when performed by an African American man than when performed by a white man (Duncan, 1976). The author argued that the white participants interpreted the ambiguous behavior according to the stereotype that African Americans are impulsive and prone to crimes and violence. Both mechanisms are responsible for expectancy confirmations in interpersonal relationships. Consistent with the idea of perceptual confirmation, Weiss (1980) argued that, through the process of "sentunent override," partners in generally satisfying relationships should tend to evaluate specific aspects of the relationship positively, whereas partners in more distressed relationships should interpret specific aspects more negatively. These ideas have received empirical support. For instance, research on attributions demonstrates that partners who hold positive beliefs about their relationships tend to interpret specific relationship events in a positive light, whereas partners who hold less positive beliefs about their relationships tend to interpret specific events more negatively (e.g., Bradbury & Fincham, 1990; Fincham, 1985; Holtzworth-Munroe & Jacobson, 1985; Jacobson, McDonald, FoUette, & Berley, 1985; Kamey & Bradbury, 2000). Interpersonal expectancies may also get confirmed through behavioral confirmation. Snyder, Tanke, and Berscheid (1977) recorded male college students having a telephone conversation with women they were led to believe were either more or less physically attractive. Objective raters found that men who believed they were interacting with atfractive women elicited more sociable and friendly behavior from their

PAGE 13

partners than men who beUeved they were interacting with less attractive women (for replication, see Anderson & Bem, 1981). In other words, participants' beliefs about their telephone partners affected the way they behaved toward their partners, which subsequently affected how their partners behaved in return. Research on more established relationships has examined similar processes in problem-solving discussions (Dovraey et al., 1998). This research examined the effects of partners' sensitivity to rejection, conceived as an individual difference variable, on their problem-solving behaviors and their partners' mood immediately following a problem-solvmg discussion. Results indicated that women who scored higher in rejection sensitivity engaged in behaviors that predicted greater negative feelings in their partners following the interaction. Consistent with the idea that expectations may affect relationship outcomes by affecting the way couples approach problem-solving in their relationships. Knee (1998) found that partners' implicit theories about relationships affected the way they responded to relationship stressors He examined dating partners who endorsed different theories about whether relationships were either destined (i.e., determined by fate or the immutable characteristics of the partners) or grown (i.e., developed through effort and communication). Although neither type of belief was associated with satisfaction, results revealed that those who believed relationships require time to grow were less likely to disengage from the relationship in response to stressors and more likely to report adaptive coping strategies. In other words, those who expected their relationship to require effort reacted more positively to relationship problems. In sum, partners' expectations about their relationships appear to be confirmed through perceptual and behavioral confirmation. Accordingly, positive expectations about

PAGE 14

8 a relationship should lead to positive relationship outcomes for two reasons. First, through perceptual confirmation, positive expectations should lead partners to interpret the events that occur in their relationships in positive ways. Second, through behavioral confirmation, positive expectations should lead partners to behave in ways that ehcit positive outcomes. For the same reasons, less positive expectations about a relationship should lead to less positive outcomes; partners with less positive outcomes should interpret events more negatively and fail to engage in behaviors that are adaptive, or engage in behaviors that are less adaptive. Positive Expectations and Disappointment In contrast to the idea that positive expectations for a relationship may be adaptive, theory and research in other domains suggest that people with high expectations may be disappointed when outcomes fall short of their expectations. As early as 1 890, William James argued that people's self-evaluations can be expressed as a ratio of their successes to their pretensions (James, 1 890). If people's achievements fall short of their aspirations, their evaluations are low, whereas if their achievements exceed their aspirations, their evaluations are high. Research suggests that, like James, people believe their expectations affect the way they evaluate their outcomes. In one study, participants reported that they believed people would feel worse following a negative outcome (e.g., receiving a "C" on a test) that was unexpected versus expected, and feel better following a positive outcome (e.g., receiving an "A" on a test) that was unexpected versus expected (Shepperd & McNulty, in press, study 1). Additional research shows that people may avoid being disappointed by lowering their expectations in the face of feedback, ensuring that their expectations do

PAGE 15

9 not exceed their outcomes (Mellers et al, 1997, study 2; Shepperd, Findley-Klein, Kwavnick, Walker, & Perez, 2000; Shepperd, Ouellette, & Fernandez, 1996; Taylor &, Shepperd, 1998). In a study by Shepperd et al. (1996, study 2), participants predicted their scores on an upcoming exam on four occasions: two weeks before the exam, just after the exam, 45 minutes before feedback, and seconds before feedback. Results revealed that the average predictions of the students were optimistic before the exam, but became more realistic and eventually pessmiistic just before feedback. Subsequent research suggests that this tendency stems from a desire to avoid receiving unexpected negative feedback (Shepperd & McNulty, 2000). Clearly people believe that their expectations affect the way they evaluate their outcomes, but do they? The answer appears to be yes. In an experimental test of this hypothesis, Shepperd and McNulty (m press, study 3) led participants to believe they were either likely or unlikely to test positive for a threatening medical condition. All participants then tested themselves and received feedback that they had tested positively or negatively for the disease. Of the participants who learned they had the disease, those who expected not to have it felt the worst. Of the participants who learned they did not have the disease, those who expected to test negative felt best. In other words, following both positive and negative outcomes, people who expected the worst felt the best. In similar research, among participants taking part in a series of gambles, unexpected wins were more satisfymg than expected wins and unexpected losses were more disappointing than expected losses (Mellers et al., 1997, study 1). Research also suggests that even people who are objectively better off can feel worse when their outcomes do not meet their expectations. In a study of medal wiimers at the Olympic games, bronze medallists

PAGE 16

showed a tendency to appear happier than silver medaUists. Bronze medallists focused on the alternative of winning no medal, whereas silver medallists focused on the alternative of winning a gold medal (Medvec, Madey, & Gilovich, 1995). In sum, as James suggested over a century ago, people tend to be disappointed by outcomes that do not meet their expectations. Do partners who hold high expectations about their relationships similarly risk being disappointed? Some research suggests that they might. Eidelson and Epstein (1982) developed an inventory to measure five beliefs they thought were dysfunctional to relationships: the belief that disagreement is destructive, the belief that mindreading is expected, the belief that partners can not change, the belief in sexual perfectionism, and the belief that the sexes are different. Such beliefs are likely to lead spouses to develop specific expectations about their own relationships (Olson et al., 1998). Consistent with the notion that positive relationship expectations may lead to disappointment, partners who report higher agreement with these beliefs are more likely to report being unhappy with their relationships (see also, Epstein & Eidelson, 1981 ; Jones & Stanton, 1988). In sum, research on counterfactual thinking suggests that when outcomes exceed expectations, people are pleased and satisfied, but when outcomes do not meet expectations, people are let down and disappointed. Accordingly, positive expectations about a relationship may lead to future relationship distress by leading partners to react more negatively to specific relationship problems. Instead, more moderate expectations may be adaptive because they may allow partners to dismiss specific problems, maintaining a relatively positive level of satisfaction.

PAGE 17

11 Reconciliation and the Distinction between Global and Specific Expectations When are positive expectations about a relationship likely to lead to positive outcomes and when are they likely to lead to negative outcomes? The research reviewed thus far suggests that to the extent that they get confirmed, positive expectations about a relationship should be beneficial. To the extent that they do not get confirmed, however, not only will they not be beneficial, but they may lead to disappointment. Thus, determining when positive expectations are beneficial to a relationship requires determining when they will get confirmed and when they will not get confirmed. Researchers in other domains have addressed this issue, revealing two factors that affect whether expectancy confirmations occur. First, perceivers' motivation can mfluence the extent to which expectancy confirmations will occur. Specifically, research has shown that the goal of forming an accurate impression of a target can reduce the likelihood that people's expectations will influence their evaluations of a particular target or outcome (Fiske & Neuberg, 1990; Miller & TumbuU, 1986). For example, participants in one study were asked to interview two job applicants, and given prior negative information about one of them. Results indicated that when participants were told to be accurate, they were less likely to behave in ways that confirmed their initial expectations (Neuberg, 1989). Second, perceivers' ability also can affect the extent to which they will rely on their expectations. Specifically, certain factors make perceivers more or less able to interpret events or behave in a manner consistent with their prior expectations. For example, research has shown that people who lack sufficient cognitive resources tend to make judgments that rely more heavily on their existing behefs and expectations than

PAGE 18

12 those who have ample cognitive resources (Chaiken, 1987; Gilbert & Hixon, 1991; Pratto & Bargh, 1991). Similarly, Miller and Turbull (1986) argued that certain outcomes may allow perceivers to be more or less able to confirm their initial expectations. Although these authors did not expand on this possibility, one quality that may affect the extent to which expectancy confirmations occur is the specificity of the outcome; some expectations involve outcomes that are very global whereas others involve outcomes that are more specific. For the purpose of this dissertation, global outcomes will be defined as outcomes that can be assessed in terms of a large number of characteristics. For example, happiness, a global outcome, can be assessed by many specific behaviors (e.g., smiling, laughing, behaving charitably, being supportive). In contrast, specific outcomes will be defmed as outcomes that can be assessed in terms of a more limited range of characteristics. For example, punctuality, a more specific outcome, can be defmed by fewer characteristics (e.g., arriving on time; for related discussion, see Hampson, Goldberg, & John, 1987; Hampson, John, & Goldberg, 1986; Nefif & Kamey, in press). For two reasons, the specificity of an outcome should be associated with whether positive expectations about relationships get confirmed or disconfirmed. First, outcomes that are more specific are less open to interpretation . For example. Dunning and his colleagues (Dunning, 1995; Dunning & McElwee, 1995; Dunning, Meyerowitz, & Holzberg, 1989) have shown that individuals describe themselves more positively on global traits than on specific ones. Explaining this fmding, the researchers note that global traits can be defmed in idiosyncratic ways, allowing the perceiver to choose from a wide range of specific examples to justify a positive selfview. Specific traits, in contrast,

PAGE 19

13 are defined more concretely, and so restrict the flexibility of the perceiver to justify a desired belief. Partners' evaluations of their relationships also seem to be constrained by the specificity of the belief (Karney, McNulty, & Bradbury, 2001; McNulty &Xamey, 2001; Neff & Karney, in press). In one study (McNulty & Kamey, 2001), newlywed partners reported their global satisfaction and their satisfaction with more specific aspects of their relationships (e.g., communication, level of support) every day for seven consecutive days. Results revealed that global evaluations were more positive and less variable than evaluations of specific aspects, suggesting that partners' global beliefs about the relationship may be more flexible and open to interpretation than their beliefs about specific aspects of the relationship. In the same way, specific outcomes, because they can be defmed by a limited range of characteristics, may limit the extent to which perceptual confirmation is likely to occur. Second, outcomes that are more specific are less likely to be associated with behavior . To the extent that the outcome is global, it is likely that there is a wide variety of behaviors that could potentially lead to it. In contrast, to the extent that the outcome is very specific, it is likely that there is only a limited range of behaviors that could potentially lead to it. For example, expecting a relationship to be positive generally is more global than expecting to resolve relationship conflicts well. Accordingly, there are many more behaviors under one's control that can lead to a happy relationship (e.g., support, having fun, romance, sharing, problem-solving) than behaviors that can lead to resolving conflicts (problem-solving behavior). In this way, the specificity of the belief may limit the extent to which behavioral confirmation can occur.

PAGE 20

14 In light of these issues, expectations about more global aspects of a relationship, because they are more open to interpretation and capable of being confirmed by a wider variety of behaviors, may be more likely to lead to consistency effects. For example, the positive belief that a relationship will succeed should result in supportive behaviors that will likely create more positive relationship satisfaction. In contrast, positive expectations about more specific outcomes, because they are less open to interpretation and capable of being confumed by only a limited variety of behaviors, may be more likely to be disconfirmed and lead to disappointment. For example, the positive expectations for punctuality should be disconfirmed to the extent that one is late for important engagements. The previously-reviewed research on expectancy confirmation and counterfactual thinking is consistent with these ideas. For instance, the research on coimterfactual thinking, showing that positive expectations can be disconfirmed and lead to disappomtment, involved outcomes that were more specific and concrete (e.g., test results, wins and losses, Olympic medals). In contrast, the research on expectancy confirmation, showing that positive expectations can be confirmed and lead to positive outcomes, involved expectations about more global and ambiguous outcomes. For instance, m the Kelley (1950) study, participants led to expect a "warm" lecturer evaluated him more positively. This expectancy was easy to confirm because there are a wide variety of characteristics and behaviors that can be used to justify the belief that someone is "warm" (e.g., supportiveness, humor, eye contact, touching). Similarly, the research on relationship attributions demonstrates that people who are generally satisfied with their relationships tend to interpret marital events more positively. Again,

PAGE 21

15 relationship satisfaction is a global belief that can be confirmed by a wide variety of marital events and partner behaviors. Likewise, the research on behavioral confirmation also involved expectations about outcomes that were very global (e.g., "blooming" intellectually), and therefore associated with multiple behaviors under perceivers' control. For instance, the meta-analysis by Harris and Rosenthal (1985) identified a total of 21 different behaviors that mediated the effects of teachers' expectation on students' performance. In addition to this indirect support, two studies provide direct support for the idea that specific expectations, more than global expectations, are likely to get disconfirmed. The first study examined how partners' prior expectations for a specific interaction influence their appraisals of the interaction (McNulty & Karney, under review). In this study, newlywed participants reported their expectations about how a specific upcoming conversation would go, participated in a 10-minute videotaped discussion of a relationship problem, and fmally reported their evaluations of the discussion. Results revealed no evidence for behavioral confirmation. Partners who expected the interaction to go well were no less likely to behave negatively than people who expected the conversation to go negatively. Consistent with the idea that specific outcomes may be under less behavioral control than more global outcomes, the authors' post-hoc explanation for this finding was that the ability to solve problems in close relationships is best conceived not as a controlled response, but rather as a skill that may develop independently of partners' beliefs and expectations. The second study providing direct support for the idea that expectations for specific outcomes are less likely to be confirmed was a study examining the effects of expectations on achievement. In this study. Armor

PAGE 22

16 and Taylor (1997) found that people with high expectations were disappointed by their outcomes, but only when expectations were specific. In sum, whether positive expectations are beneficial to relationships may depend on whether they involve global or specific outcomes. Positive expectations about global outcomes, because they are more open to interpretation and capable of being confirmed by a wide variety of behaviors, may be more likely to lead to consistency effects. In contrast, positive expectations about more specific outcomes, because they are less open to interpretation and capable of being confirmed by only a limited variety of behaviors, may be more likely to lead to disappointment. Based on this idea, the primary goal of this study is to test the hypothesis that positive global expectations about a relationship will lead relationships to be more satisfying over time, whereas positive specific expectations about a relationship will lead relationships to be less satisfying over time. Overview of the Current Study Given the uncertainty about how positive expectations about a relationship affect relationship functioning, and given the prospect that the specificity of the expected outcomes may help clarify this issue, the current study examined how expectations about outcomes at various levels of specificity affect relationship outcomes over time. Married couples participating in a broader study of relationship satisfaction were asked to report their expectations for various relationship outcomes at the beginning of a longitudinal study. Subsequently, these couples responded to measures of relationship satisfaction every six months for two years. Two conceptually similar dependent measures were used to operationalize relationship outcomes. The first was change in satisfaction over time. Analyses examined whether each type of expectation was associated with within-subject

PAGE 23

17 change in marital satisfaction. The second dependent variable was reactivity to relationship problems. Analyses assessed the extent to which variability in spouses' satisfaction with the relationship covaried with variability in the number and severity of relationship problems, and the extent to which expectations moderated this association. All of the couples participating in this research were newlyweds, assessed within the first six months of both partners' first marriage. This is an appropriate sample because, as Jussim and Eccles (1995) point out, over the course of an ongoing relationship, expectations are likely to become accurate reflections of past experience. In a sample of newlyweds, expectations and prior experience were less likely to be confounded. At the initial assessment, partners reported their expectations for global and specific relationship events. First, they reported their expectations for two global outcomes: (a) whether they expected their global satisfaction to get better, stay the same, or get worse over time and (b) how confident they were that their relationships would remain intact or stable. Second, they reported their expectations for three specific relationship outcomes: (a) expectations about specific changes in relationship satisfaction, (b) expectations about specific partner qualities and (c) expectations about specific relationship problems. Hypotheses Hypothesis 1 . Consistent with the idea that that expectations about global outcomes are likely to be confirmed through the processes of perceptual and behavioral confirmation, I predicted that spouses who expected global outcomes to be positive would (a) behave less negatively and (b) interpret events more charitably in their relationships. Thus, positive global expectations should be associated with (a) a lower

PAGE 24

18 proportion of negative behavior during their interactions and (b) more positive scores on the measure of relationship attributions, w^hereas less positive global expectations would be associated with (a) a higher proportion of negative behavior during their interactions and (b) lower scores on the measure of relationship attributions. In other words, scores on each measure of global expectations will be negatively associated with observations of behavior and positively associated with scores on the measure of relationship attributions. Hypothesis 2 . Consistent with the idea that that expectations about specific outcomes are less likely to be confu-med through the processes of perceptual and behavioral confirmation, I predicted that spouses who expected specific outcomes to be positive would be no more likely to (a) behave negatively in their relationships over time or (b) interpret events more charitably in their relationships than spouses who expect less positive specific outcomes. In other words, scores on each measure of specific expectations should not be associated with observations of behavior or scores on the measure of relationship attributions. Hypothesis 3 . Based on the idea that expectations about specific outcome are less susceptible to the confirming effects of expectations, spouses' positive expectations about specific relationship outcomes were likely to be disconfirmed over time leading to disappointment. Thus, I predicted that spouses with positive expectations about more specific relationship outcomes would experience negative changes in satisfaction over time and a stronger negative covariance between global satisfaction and specific relationship problems. In contrast, because they are less likely to be disappointed, spouses with less positive specific expectations should experience more stable satisfaction and a weaker negative covariance between satisfaction and specific problems

PAGE 25

19 over time. Predictions in terms of each measure of specific expectations are presented in the following paragraphs. Participants were asked how they expected their feelings toward their relationship to change over time: whether their feelings would experience ups and downs or remain steady from day to day. I predicted that partners who expected their feelings to be very steady would be more likely to (a) experience declines in satisfaction over time and (b) experience a stronger covariance between reported problems and satisfaction over time. In contrast, partners who expected some ups and downs should be more likely to (a) be happier over time and (b) experience a weaker covariance between reported problems and satisfaction over time. This prediction will be supported to the extent that expectations for steadiness are negatively associated wdth change in satisfaction and negatively associated with the covariance between global satisfaction and specific problems. Participants were asked to report their expectations about specific qualities of their partners (e.g., my partner will never disappoint me). I predicted that participants who held more positive expectations for their partners would be more likely to (a) experience declines in satisfaction over time and (b) experience a stronger covariance between reported problems and satisfaction over time. In contrast, partners who expected their problems to remain the same or get worse should be more likely to (a) be happier over time and (b) experience a weaker covariance between reported problems and satisfaction over time. This prediction will be supported to the extent that expectations for partner are negatively associated with change in satisfaction and negatively associated with the covariance between global satisfaction and specific problems.

PAGE 26

20 Participants were asked to report how they expected a number of potential relationship problems to develop over time: whether they would get worse or get better. I predicted that partners who expected their problems to get better would be more likely to (a) experience declines in satisfaction over time and (b) experience a stronger covariance between reported problems and satisfaction over time. In contrast, partners who expected their problems to remain the same or get worse should be more likely to (a) be happier over time (b) and experience a weaker covariance between reported problems and satisfaction over time. This prediction will be supported to the extent that expectations for specific problems are negatively associated with change in satisfaction and negatively associated with the covariance between global satisfaction and specific problems. Hypothesis 4 . Based on the idea that expectations about global outcome are more susceptible to the confirming effects of expectations, spouses' positive expectations about global relationship outcomes were likely to be confirmed over time. Thus, I predicted that spouses with positive expectations about more global relationship outcomes would experience more stable satisfaction over time and a weaker negative covariance between global satisfaction and specific relationship problems. Because less positive global expectations are also likely to be confirmed, spouses with less positive global expectations should experience declines in satisfaction and a stronger negative covariance between satisfaction and specific problems over time. Predictions in terms of each measure of global expectations are described in the following paragraphs. Participants were asked how they expected their feelings toward their relationship to change over time: whether their feelmgs would become better or become worse. I predicted that partners who expected their feelings to grow positively over time would be

PAGE 27

21 more likely to experience (a) positive changes in marital satisfaction over time and (b) a weaker negative covariance between reported problems and satisfaction over time. In contrast, partners who expected their relationships to stay the same or get worse should be more likely to (a) be less happy over time and (b) experience a stronger covariance between reported problems and satisfaction over time. This prediction will be supported to the extent that expectations for growth are positively associated with change in satisfaction and positively associated with the covariance between global satisfaction and specific problems. Participants were asked to predict the likelihood that they would still be married in four years. I predicted that partners who were more confident that they would remain married would be more likely to experience (a) positive changes in marital satisfaction over time and (b) a weaker negative covariance between reported problems and satisfaction over time. In contrast, partners who were less confident that they should stay remain married should be more likely to (a) be less happy over time and (b) experience a stronger covariance between reported problems and satisfaction over time. This prediction will be supported to the extent that expectations for stability are positively associated with change in satisfaction and positively associated with the covariance between global satisfaction and specific problems. Hypothesis 5 . Consistent with the idea that positive global expectations should lead to positive relationship outcomes through perceptual and behavioral confirmation, I predicted that the effects of positive global expectations on relationship outcomes would be mediated by positive interpretations (perceptual confirmation) and fewer negative behaviors (behavioral confirmation). Specifically, observations of behavior and scores on

PAGE 28

22 the measure of relationship attributions should mediate the effects of global expectations on changes in marital satisfaction and the covariance between reported problems and satisfaction such that spouses with global positive expectations should behave less negatively and make more positive attributions.

PAGE 29

METHOD Participants Newlywed couples were recruited for this study using two methods. The first method was to place advertisements ui community newspapers and bridal shops, offering up to $300 to couples willing to participate in a study of newlyweds. The second method was to review the applications of couples who had applied for marriage licenses in Alachua County, Florida. In Alachua County, marriage licenses are available to the public and contain data on spouses' ages, whether or not this is their first marriage, and the date of the wedding. Couples who were eligible for the study based on these criteria were sent letters offering them up to $300 to participate in a study of newlyweds. 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 3 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 1 0 years of education (to ensure comprehension of the questionnaires), and (f) the couple had no immediate plans to move away fi-om the area. Nearly 300 couples responded to these solicitations; the first 82 eligible couples who arrived for their scheduled interview comprised the current sample. Analyses revealed no significant differences in age or 23

PAGE 30

24 years of education between couples recruited through the different types of solicitations (D'Angelo «& Karney, 1999). At the tune of initial data collection, husbands were 25.1 (SD = 3.3) years old, and had completed 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 completed 16.3 (SD = 1.2) years of education. Thirty-nine percent were employed fiiU time, and 50% were full time students. Slightly over 70% of the sample was Christian (over 45% were Protestant) and 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 selfreport measures of marital satisfaction and a letter instructing couples to complete all questionnaires independently of one another. As part of the laboratory session, each spouse independently identified an area of difficulty in the marriage, and agreed to spend 10 minutes discussing the issues related to each area of difficulty. Spouses then participated in two 10-minute, videotaped discussions in which they were left alone to "work towards some resolution or agreement" for each area of difficulty. The order of the two interactions was determined through a coin flip. Couples were paid $50 for participating in this part of the study. At approximately six-month intervals subsequent to the initial assessment, couples were contacted by phone and mailed additional packets of questionnaires along

PAGE 31

25 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 dissertation examined five waves of data, covering approximately the first two and a half years of marriage. Materials Marital satisfaction . The most common measures used to assess marital satisfaction ask spouses to report their global sentiments towards the marriage as well as their level of agreement about specific problem areas (e.g., the Marital Adjustment Test; Locke & Williamson, 1958). As several authors have pointed out (e.g., Fincham & Bradbury, 1987), the use of such omnibus measures can lead to inflated associations with other variables that also address relationship processes. To ensure that these two ideas were not confounded in the current study, two global measures of satisfaction were used here. The first measure was a version of the Semantic Differential (SMD; Osgood, Suci, & Tannenbaum, 1957), that asks spouses to rate their perceptions of their relationship on 7-point scales between fifteen pairs of opposing adjectives (e.g., "Bad-Good," "Dissatisfied-Satisfied," "Unpleasant-Pleasant"). The SMD yields scores from 15 to 105. In the current sample, internal consistency of this measure was high (coefficient alpha = .91 for husbands and .93 for wives). (A sample of this measure can be found in Appendix A.) The second global measure was the Quality Marriage Index (QMI; Norton, 1983), a sk item scale asking spouses to report the extent to which they agree or disagree with general statements about their marriage (e.g., "We have a good marriage" and "My relationship with my partner makes me happy"). The QMI yields scores from 6 to 45.

PAGE 32

26 Internal consistency of this measure also was high (coefficient alpha = .94 for husbands and .94 for wives). (A sample of this measure can be found in Appendix B.) Inventory of Specific Marital Problems. Participants' perceptions of specific relationship problems were measured using the Marital Problems Inventory (MPI; Geiss & O'Leary, 1981). This measure lists nineteen potential problem areas in a marriage (e.g., children, communication, household management) and asks participants to rate each on a scale from 1 ("not a problem") to 1 1 ("major problem"). These items were summed to form an index of the severity of relationship problems. Internal consistency of this measure was high (coefficient alpha = .88 for husbands and .89 for wives). (A sample of this measure can be found in Appendix C.) Expectations for Grov^h in Relationship Satisfaction . Two measures were developed to assess spouses' expectations about global relationship outcomes. The first was a three-item measure designed to assess how partners expected their relationship satisfaction to develop globally over time. The first item from this measure asked partners to respond to the following question: "over the next sbc months, do you expect that your overall feelings about your marriage will become:" 1 = "much worse," 2 = "a little worse," 3 = "stay the same," 4 = "a little better," 5 = "much better." The second item asked spouses to respond to the same question with respect to the "next 4 years." (A sample of these items can be found in Appendix D.) For the third item, spouses selected from among nine pictures reflecting different patterns of growth the one pattern that best represented their expected trajectory of marital satisfaction over the upcoming sbc months. (A sample of these pictures is contained in Appendbc E.) For the current measure, pictures were coded for the degree of expected growth in satisfaction: 1 = flat

PAGE 33

27 trajectory, 2 = positive trajectory. These three items were summed to form an index of expectations for growth. Internal consistency of these three items was adequate (coefficient alpha = .63 for husbands and .72 for wives). Expectations for Stability . The second measure of global expectations was one item designed to assess spouses' expectations about the likelihood that the relationship would remain intact or stable over the upcoming four years. Specifically, the item read "How likely is it that the two of you will still be married four years from now?': 1= "not at all likely," 2 = "not too likely," 3 = "pretty likely," 4 = 'Very likely," 5 = "completely certain." (A sample of this measure can be found in Appendix F.) Expectations for Steadiness of Relationship Satisfaction . Three measures were developed to assess spouses' expectations about specific relationship outcomes. The first measure was a three-item measure designed to assess very specifically how partners expected their relationship satisfaction to develop from day to day. In contrast to expectations for growth, expectations for steadiness assessed how much spouses expected their satisfaction to change from day to day. The first item asked spouses to respond to the following question: "over the next six months, which of the following describes how your feelings towards your marriage are likely to change?": 1 = "major highs/major lows," 2 = "some ups and downs," 3 = "a few ups and downs," 4 = "pretty steady," 5 = "very steady." The second item asked spouses to respond to the same question with respect to the "next four years." (A sample of these items can be found in Appendix G.) Additionally, for the third item, the pictures contained in Appendix E were coded for the degree of expected steadiness: 1 = unsteady, 2 = very steady. These three items were

PAGE 34

28 summed to form an index of expectations for steadiness. Internal consistency of these three items was relatively high (coefficient alpha = .77 for husbands and .78 for wives). Expectations for Partner . The second measure of specific expectations was designed to assess participants' specific expectations for how their partners would behave in the future. This measure asked spouses to respond to nine items designed to describe fairly unrealistic, specific qualities that people might expect their partners to possess (e.g., my partner will never disappoint me, my partner will always be attractive to me) on a scale from 1 (strongly disagree) to 7 (strongly agree). These items were summed to form an index of expectations for partner. Internal consistency of this measure was adequate (coefficient alpha = .71 for husbands and .80 for wives). (A sample of this measure can be found in Appendix H.) Expectations for Specific Problems . The second measure of specific expectations was designed to assess participants' expectations for how the problems on the MPI would develop over time. The measure lists the same nineteen problems listed on the MPI and asks participants to "mdicate how you think each problem is going to develop for you and your spouse over the next sbc months:" 1 = "get worse," 2 = "stay the same," 3 = "get better." These items were summed to form a single index of expectations for problems. Internal consistency of this measure was relatively high (coefficient alpha = .83 for husbands and .74 for wives). (A sample of this measure can be found in Appendix I.) Attributions . The manner in which partners interpret the events in their marriages was assessed using a measure of marital attributions: the Relationship Attributions Measure (RAM; Fincham & Bradbury, 1992). This 24-item measure presents spouses with four negative stimulus events that are likely to occur in all marriages (e.g., "Your

PAGE 35

spouses 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 partner's behaviors intentional, selfishly motivated, and blameworthy. For each subscale, a composite score was computed by summing the 1 2 judgments, resulting in two scores for each spouse with possible ranges of 1 2 to 84. Higher scores indicate attributions that view the partner in a more positive light. Internal consistency of each sub-scale was relatively high (for causality attributions, coefficient alpha = .85 for husbands and .73 for wives; for responsibility attributions, coefficient alpha = .89 for husbands and .90 for wives). (A sample of this measure can be fovmd in Appendix J.) Marital interaction behavior . The videotaped discussions were coded using a modified version of the Verbal Tactics Coding Scheme (VTCS; Sillars, Coletti, Parry, & Rogers, 1982). This version of the VTCS assigns each spouse's speaking turn one of four possible codes. A speaker received an Avoidant code for speaking turns that were offtopic or moved the discussion away from the problem at hand. A speaker received one of two negative codes for speaking turns that either directly faulted, rejected, or criticized the partner (Direct Negative), or indirectly criticized the partner through presumptive attributions, avoiding responsibility, or hostile questions (Indirect Negative). A speaker

PAGE 36

30 received a Constructive code for speaking turns that were on-topic and not negative. In light of prior research showing that negative behaviors are more rehably linked to relationship outcomes than positive behaviors (Bradbury & Kamey, 1993), only the negative codes will be examined in the current study. The reliability of this system was assessed by randomly choosing 30% of the interactions to be coded by a second rater. Degree of agreement between raters was estimated with an intra-class correlation coefficient (ICC) comparing the amounts of each code observed by each rater across the interactions. ICCs indicated adequate inter-rater reliability for the codes analyzed in the current study (for direct negative, ICC = .66 for husbands and .83 for wives; for indirect negative, ICC = .83 for husbands and .65 for wives). Because they were correlated for each spouse in each interaction (rs ranged from .34 to .54), the two negative codes were summed to form a single index of negative behavior for each spouse during each interaction. Because these indices were correlated for each interaction (rs ranged from .40 to .50), scores were summed across each interaction to form a single index of negative behavior for each spouse. Data Analvsis Hierarchical Linear Modelmg (HLM; Bryk & Raudenbush, 1992), implemented with the HLM/2L computer program (Bryk, Raudenbush, & Congdon, 1994) was used to analyze these data for several reasons. First, in contrast to other approaches to analyzing multilevel models (e.g., structural equation modeling), HLM provides reliable estimates of within-subject parameters even when sample sizes are relatively small. Second, HLM provides maximally efficient estimates of these parameters by weightmg individual estimates according to empirical Bayes (1973) theory. When the within-subject parameter

PAGE 37

31 for an individual can be estimated precisely, the final estimate relies heavily on the individual data. When the parameter can not 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 fmal estimated variance of the sample, variances estimated in this v^ay tend to be more conservative than those obtained through traditional OLS methods. Parameters describing husbands' and wives' data will be estimated simultaneously in a couple-level model, according to procedures described by Raudenbush, Brennan, & Bamett (1995). Generally, two types of analyses will be conducted, one for each of the two dependent measures of relationship outcomes. First, to test the hypotheses that expectations affected change in satisfaction over time, growth curve analyses were conducted to model individual change over two years. In the first stage, withinsubjects trajectories were computed for each individual's marital satisfaction over time. In the next stage, expectation scores were used to account for between-subjects differences in these trajectories. Second, multilevel modeling was conducted to test the hypothesis that expectations were associated with the way spouses reacted to their relationship problems over two and half years. In the first stage of this analysis, repeated-measiores data fi-om each spouse were used to estimate the extent to which variability in reported problems was associated with variability in satisfaction. In the second stage, expectations were used to account for between-subjects differences in this covariance.

PAGE 38

RESULTS Descriptive Statistics and Correlations Descriptive statistics for all independent variables are reported in Table 1 . As would be expected from a sample of newlyweds, on average husbands and wives held very positive expectations about their relationships. Furthermore, both spouses tended to hold the most positive expectations about global outcomes. For instance, among husbands, 85% and 27% reported the highest possible score for expectations for stability and expectations for growth respectively, whereas only 1 1 % reported the highest possible score for expectations for steadiness and none reported the highest possible score for expectations for partner or expectations for problems. Among wives, 90% and 37% reported the highest possible score for expectations for stability and expectations for growth respectively, whereas only 1 7% reported the highest possible score for expectations for steadiness and none reported the highest possible score for expectations for partner or expectations for problems. Furthermore, comparing the only two global and specific expectations that were assessed on the same scale, expectations for growth and expectations for steadiness, reveals that both husbands and wives expected more growth from their relationship than they expected steadiness, t = 4.1, p < .01 for husbands, and t = 4.6, p < .01 for wives. Despite the positivity of these expectations, standard deviations reveal substantial variability, indicating that ftirther analyses were appropriate. Of note, husbands and wives did not differ in the positivity of their expectations about the 32

PAGE 39

33 relationship, except that husbands had marginally higher expectations for their wives than wives had for their husbands, t = 1 .6, p < . 1 0. Descriptive statistics for the hypothesized mediating variables are also reported in Table 1 . Husbands and wives appeared to make relatively positive attributions, seeing external causes for negative events and freeing their partners from blame. Similarly, husbands and wives did not behave very negatively during their interactions, with mean proportions of negative behavior M = 10.2% for husbands and M =^ 13.4 for wives. Nevertheless, standard deviations reveal substantial variability, indicating that fiirther analyses were appropriate. Paired-sample t-tests revealed no differences between husbands and wives in their tendency to make positive versus negative attributions or to behave negatively. Correlations among the independent and mediating variables are reported in Table 2. Both measures of global expectations were somewhat correlated for husbands, r = 0.20, but not for wives. Also, two specific expectations, expectations for variability and expectations for partner, were correlated for husbands and wives, r = 0.36 for husbands and r = 0.33 for wives. Importantly, however, these measures did not overlap completely suggestmg that they did not simply tap the same construct. Some measures of specific expectations were also correlated with some measures of global expectations. For instance, husbands' and wives' expectations for growth and expectations for problems were somewhat correlated, r = 0.27 for husbands and r = 0.29 for wives. Also, husbands' and wives' expectations for stability were correlated with their expectations for steadiness, r = 0.22 for husbands and r = 0.39 for wives. However, although these

PAGE 40

34 measures of global and specific expectations were associated with one another, again they did not overlap completely. With respect to the hypothesized mediating variables, consistent with previous research, attributions for responsibility and attributions for causality were strongly associated with one another for both husbands and wives, r = 0.62 for husbands and r = 0.56 for wives (Bradbury & Fincham, 1990). Also, for wives, both types of attributions were negatively correlated with negative behavior, r = -0.20 for causality attributions and r = -0.40 for responsibility attributions. In other words, wives who tended to make positive attributions also tended to behave less negatively. Of note, inconsistent with the prediction of Hypothesis 1 which stated that expectations about global outcomes would be significantly associated with attributions and behavior, attributions and behavior did not appear to be associated vsdth global expectations. Also, inconsistent with the prediction of Hypothesis 2 that stated that specific expectations would not be associated with either attributions or behavior, both types of attributions were associated with specific expectations. However, these bivariate correlations did not control for marital satisfaction. Because previous research reveals that satisfied couples tend to make positive attributions (see Bradbury & Fincham, 1990, for a review), and because satisfied couples may also tend to expect positive specific outcomes fi-om their relationships, it remains possible that these correlations were the result of shared variance with marital satisfaction. Cross-spouse correlations for the independent and mediating variables are reported in Table 3. The correlations on the diagonal reveal that husbands' and wives' scores were at least marginally correlated on all measures except attributions for

PAGE 41

35 responsibility. In other words, husbands and wives expected similar outcomes from their relationships, both globally and specifically, behaved similarly in their relationships, and made similar attributions about whether or not their partners were the cause of negative specific relationship events. Were Specific and Global Expectations Associated with Attributions and Negative Behavior ? Hypothesis 1 stated that spouses who expected positive global outcomes from their relationship would tend to make positive attributions, attributing negative relationship events to external circumstances and freeing their partners from blame, and tend to behave less negatively. As previously mentioned, bivariate correlations estimating these relationships are presented in the bottom section of Table 2. However, because attributions, behavior, and expectations all may have been related to marital satisfaction, partial correlations were computed between each measure of global expectations and attributions and between each measure of global expectations and behavior, controlling for marital satisfaction. These correlations are reported in the top half of Table 4. As is evident from these partial correlations. Hypothesis 1 was not supported. There is no evidence in these data that spouses who expected positive global outcomes from their relationships tended to make positive attributions or to behave less negatively. Of note, because a significant association between an independent variable and a hypothesized mediating variable is an essential criterion for establishing mediation. Hypothesis 5, that suggested attributions and behavior would mediate the effects of global expectations on marital satisfaction, was not pursued.

PAGE 42

36 Hypothesis 2 stated that spouses' expectations for specific relationship outcomes would not be associated with their tendency to make positive attributions or their tendency to behave negatively. The positive bivariate correlations reported in the bottom half of Table 2 between specific expectations and attributions suggest that spouses who expected positive specific outcomes tended to make positive attributions. To rule out the possibility that these correlations may have been artifacts resulting because attributions, behavior, and expectations all are related to marital satisfaction, partial correlations were computed between each measure of specific expectations and attributions and each measure of specific expectations and behavior, controlling for marital satisfaction. These partial correlations are reported in the bottom half of Table 4. As can be seen in this section of Table 4, most correlations between specific expectations and attributions and specific expectations and behavior were non-significant when controlling for marital satisfaction. Furthermore, the correlations that were significant do not reveal any consistent patterns. For instance, the correlation between husbands' expectations for problems and responsibility attributions was positive, suggesting that more positive expectations were associated with more positive attributions. But the correlation between wives' expectations for partner and causality attributions was negative, suggesting that more positive expectations were associated with more negative attributions. Given these inconsistencies, spouses' expectations for specific relationship outcomes do not appear to have been systematically associated with the way they explained the events in their relationships or the way they behaved in their relationships.

PAGE 43

37 Were Specific and Global Expectations Associated with Marital Outcomes Over Time ? Because all subsequent hypotheses involve either changes in satisfaction over time or the covariance between specific problems and marital satisfaction over time, the mean scores on each measure of marital satisfaction and the Inventory of Specific Marital Problems are presented in Table 5 for each wave of data collection. As can be seen in Table 5, scores on both measures of marital satisfaction became less positive over time for both husbands and wives suggesting a negative linear declme m satisfaction over time. Likewise, scores on the problem inventory became less positive over time, indicating a slight decrease in marital problems over time. One explanation for this pattern may be that although problems did not get more numerous or more severe over time, they became more important over time and thus became more strongly associated with marital satisfaction. Of note, attrition in the current study was relatively low. Whereas the average rate of attrition in prior research on marriage is 31% (Kamey & Bradbury, 1995), 78% of husbands and 80% of wives responded at Time 5. Furthermore, because HLM can estimate trajectories based as few as three waves of data, fmal analyses were based on 79 out of 82 couples, or 96% of the sample. To compute a within-subject trajectory of satisfaction for each individual, the longitudinal data were analyzed using HLM. Change for each individual was estimated using the following model: Yij = 3oj + f3ij (Time) + ry [Equation 1] where Yy is the marital satisfaction of individual j at Time i; poj is the marital satisfaction of individual j at Time 1 (i.e., the initial satisfaction of individual j); Pij is the rate of linear change in marital satisfaction of individual j; and r^ is the residual variance in

PAGE 44

38 repeated measurements for spouse j, assumed to be independent and normally distributed across spouses. Husbands' and wives' parameters were estimated simultaneously using a multivariate technique suggested by Raudenbush, Brennan, and Bamett (1995). Fitting this model to the data provided estimated generalized least squares of the average of the intercept and trajectory of marital satisfaction for husbands and wives, as well as restricted maximum likelihood estimates of the variances of these parameters. These estimates are reported in Table 6 along with t-statistics that test whether the means of each parameter are different from zero and chi-square statistics that test whether the variances of each parameter are different from zero. With respect to the intercept, the estimates presented in Table 6 suggest that on average husbands and wives reported relatively high marital satisfaction as newlyweds. Comparison of Table 6 with Table 5 reveals that the mean intercept for each measure is similar to the mean satisfaction reported at Time 1 . This is to be ejqjected because the intercept represents the starting point of the trajectory. However, as revealed by the chisquare tests of variance, the mean initial satisfaction varied significantly across husbands and across wives on both measures of marital satisfaction. In other words, some spouses were initially more happy than other spouses. With respect to the trajectory, the estimates presented in Table 6 suggest that on average husbands and wives tended to experience significant declines in satisfaction over the first two and a half years of marriage, consistent with prior studies of marital satisfaction (e.g., Kamey & Bradbury, 1997). The t-statistics presented in Table 6 reveal significant negative trajectories for husbands and wives on each of the two measures of marital satisfaction. Furthermore, according to the significance of the chi-square tests

PAGE 45

39 reported in Table 6, the trajectories varied across husbands and across wives for both measures. In other words, although on average husbands' and wives' satisfaction declined over two and a half years, satisfaction declined more for some spouses than it did for others. This is important because Hypothesis 3 and Hypothesis 4 suggest that specific and global expectations will account for this between-subject variability in change in satisfaction over time. As a measure of reactivity to problems, HLM analyses were also used to compute withinsubject estimates of the extent to which variability in marital satisfaction co varied with variability in specific problems over time, controlling for the trajectory of satisfaction over time. To estimate this co variance, scores on the Marital Problem Inventory were centered around the mean for each spouse and entered into the following model: Yij (3oj + P ij(Time) + (32j(Problems) + ry [Equation 2] where Yy is the global marital satisfaction score for individual j at time i; (3oj estimates the satisfaction of individual j at Time 1; Pij is the rate of linear change in marital satisfaction of individual j; 32j captures the covariance between variability in reported specific problems and variability in reported global marital satisfaction for spouse j, controlling for the trajectory of satisfaction; and ry is the residual variance in repeated measurements for spouse j, assumed to be independent and normally distributed across spouses. Husbands' and wives' parameters were estimated simultaneously using a multivariate technique suggested by Raudenbush, Brennan, and Bamett (1995). The estimated generalized least squares of the average of the covariances between problems and global marital satisfaction for husbands and wives are reported in Table 7

PAGE 46

40 along with restricted maximum likelihood estimates of the variances of these parameters. Also reported are t-statistics that test whether the means are different from zero and chi-square statistics that test whether the variances are different from zero. According to the t-statistics in Table 7, variability in specific problems was significantly negatively associated with variability in global satisfaction, controlling for the trajectory of satisfaction over time. When problems were more numerous and severe, marital satisfaction was more negative. Additionally, as evidenced by the chi-square statistics in Table 7, the extent to which problems covaried with satisfaction varied across husbands and across wives. In other words, although on average variability in husbands' and wives' satisfaction was associated with variability in their specific problems, the magnitude of this association was stronger for some spouses than for others. This is important because Hypothesis 3 and Hypothesis 4 suggest that specific and global expectations will account for between-subjects variability in the extent to which satisfaction and problems covary over time. Hypothesis 3 . The first prediction of Hypothesis 3 stated that positive expectations about specific relationship outcomes would be associated with declines m satisfaction over time. To test this hypothesis, each measure of specific expectations was entered into a second-stage model to account for between-subject variability in the parameters of change estimated by equation 1 . The effects of specific expectations on the parameters of change in satisfaction are reported in Table 8. The top half of Table 8 reports the associations between specific expectations and initial satisfaction. In general, the t-statistics in this section reveal that spouses who were initially more satisfied with the relationship tended to expect specific

PAGE 47

41 relationship outcomes to be more positive. Specifically, husbands' and wives' expectations for steadiness were positively associated with their initial satisfaction with the relationship, and husbands' expectations about the qualities of their wives were positively associated with their initial satisfaction with the relationship. Of note, the association between initial satisfaction and expectations for partner was significantly stronger for husbands than for wives on both measures of marital satisfaction, = 7.1, p < .01 for the SMD, = 4.4, p < .05 for the QMI. The associations between specific expectations and change in marital satisfaction are reported in the bottom half of Table 8. Consistent with the predictions of Hypothesis 3, this section of Table 8 reveals several significant or marginally significant negative associations between specific expectations and change in satisfaction over time. Specifically, expectations for steadiness were significantly associated with steeper negative trajectories for wives on the QMI, and marginally associated with steeper negative trajectories for husbands on the QMI. The more that spouses expected their relationship satisfaction to be very steady over time, the less satisfied they became with their relationships over time. In contrast, the less spouses expected their relationship satisfaction to be very steady over time, the more satisfied they became with their relationships over time. Additionally, positive expectations about specific partner qualities were associated with steeper negative trajectories for husbands on both measures of marital satisfaction. The more that husbands expected their wives to posses positive qualities, the less satisfied they became with their relationships over time. In contrast, the less that husbands expected their wives to possess positive qualities, the more satisfied they became with their relationships over time. Finally, wives'

PAGE 48

42 expectations for problems were also associated with negative changes in satisfaction over time on the SMD. The more that wives expected their problems to get better over time, the less satisfied they became with their relationships over time. In contrast, the less that wives expected their problems to get better over time, the more satisfied they became with their relationships over time. Although some measures of expectations were significantly associated with change in satisfaction for one spouse but not the other, analyses revealed no significant gender differences in the extent to which specific expectations were associated with change in satisfaction over time. The second prediction of Hypothesis 3 stated that spouses with positive expectations about specific relationship outcomes would be more reactive to relationship problems and thus experience a stronger covariance between their problems and their marital satisfaction over time. To test this prediction, each measure of specific expectations was entered separately into a second-stage model to account for betweensubjects variability in the covariance estimated by equation 2. Support for this prediction would be provided by significant negative associations between specific expectations and the covariance between problems and satisfaction, indicating that positive specific expectations were associated with a stronger association between relationship problems and satisfaction. The effects of specific expectations on the covariance between problems and satisfaction are reported in Table 9. The t-statistics in Table 9 do not reveal any support for Hypothesis 3. Although the negative direction of the effects is consistent with predictions, the estimates did not reach significance.

PAGE 49

43 In sum, the prediction that positive expectations about specific relationship outcomes would lead to declines in marital satisfaction over time was supported. Husbands' and wives who tended to have very positive expectations about specific outcomes tended to become less satisfied over time. However, the prediction that positive expectations about specific outcomes would be associated with a stronger negative covariance between marital problems and marital satisfaction was not supported. Spouses who expected specific outcomes to be positive were not significantly more reactive to relationship problems than spouses who expected specific outcomes to be less positive. Hypothesis 4 . The first prediction of Hypothesis 4 stated that positive expectations about more global relationship outcomes would be associated with more positive changes m satisfaction over time. To test this prediction, each measure of global expectations was entered into a second-stage model to account for between-subjects variability in the parameters of change estimated by equation 1 . The effects of global expectations on the parameters of change m satisfaction are reported in Table 10. The top half of Table 10 contains the associations between global expectations and initial satisfaction. In general, the t-statistics in this section reveal that spouses who were more satisfied with their relationships initially, tended to expect global relationship outcomes to be more positive. Specifically, husbands who were more satisfied with the relationship were more likely to expect their feelings about the relationship to grow over time and more confident that their relationship would remain intact or stable over time. Wives who were more satisfied with their relationships according to the SMD were marginally more confident that their relationship would remain stable over time. Husbands and wives did not differ in the extent to which

PAGE 50

44 expectations for stability were associated with initial satisfaction, but the association between expectations for growth and initial satisfaction was significantly stronger for husbands than for wives on both measures of satisfaction, = 5.0, 2 < .05 for the SMD, x! = 4.4,E<.05fortheQMI. The associations between global expectations and change in marital satisfaction are reported in the bottom half of Table 10. Contrary to the predictions of Hypothesis 4, this section of Table 10 reveals that husbands' global expectations are negatively associated with changes in satisfaction over time. The more that husbands expected their feelings about the relationship to grow over time and the more that husbands expressed confidence that their relationship would remain stable over time, the less satisfied they became with their relationship over time. In contrast, the less that husbands expected their feelings about the relationship to grow over time and the less that husbands expressed confidence that their relationship would remain stable over time, the more satisfied they became with their relationship over time. Although the association between global expectations and change in marital satisfaction was not significant for wives, there were no significant gender differences in the association between global expectations and change in satisfaction over time. The second prediction of hypothesis 4 stated that spouses with positive expectations about global outcomes would be less reactive to specific problems and thus experience a weaker covariance between problems and marital satisfaction over time. To test this prediction, each measure of global expectations was entered into a second-stage analysis to account for between-subjects differences in the covariance between problems and satisfaction estimated by equation 2. Support for the second prediction of hypothesis

PAGE 51

45 4 would be provided by significant positive associations between global expectations and the covariance between problems and satisfaction, indicating that positive global expectations resulted in a weaker negative covariance between problems and satisfaction. The effects of global expectations on the covariance between problems and satisfaction are reported in Table 1 1 . The t-statistics in Table 1 1 provide no support for Hypothesis 4 that spouses with positive global expectations would be less reactive to problems and experience a weaker covariance between problems and satisfaction. Instead, contrary to predictions, positive scores on several measures of global expectations were associated with a stronger negative covariance between problems and satisfaction. Specifically, husbands who expected their feelings about the relationship to grow more positive over time, and wives who were more confident that the relationship would remain intact over time, were at least marginally more reactive to specific problems over time. In sum, the first prediction of Hypothesis 4 that positive expectations about global outcomes would lead to more stable marital satisfaction over time was not supported. Instead, husbands who tended to have very positive expectations about global relationship outcomes, such as the likelihood that their relationship would remain intact and the extent to which their feelings about the relationship would grow over time, tended to become less satisfied with the relationship over time. Importantly, although this association was not significant for wives, the size of the effect did not differ significantly across husbands and wives. The second prediction of Hypothesis 4 that positive expectations would be associated v^th a weaker negative covariance also was not supported. In fact, several associations between global expectations and the covariance

PAGE 52

46 between problems and satisfaction were significant, suggesting that spouses with the most positive global expectations were more reactive to relationship problems than spouses with less positive expectations about global outcomes. Did Global and Specific Expectations Interact to Affect Changes in Satisfaction ? Results indicate that positive expectations about both global and specific outcomes tended to lead to negative changes in satisfaction over time, suggesting that the effects of positive expectations on changes in marital satisfaction do not depend on the specificity of the expected outcome. However, spouses are likely to hold expectations about both global and specific outcomes simultaneously. Whereas some spouses may hold positive or less positive expectations about both types of outcomes, other spouses may hold positive expectations about one type of outcome and less positive expectations about the other type of outcome. Given these different combinations of expectations, it remains possible that global and specific expectations interact to lead to changes m marital satisfaction. For instance, because specific expectations are predicted to lead to negative changes in marital satisfaction, positive expectations about global outcomes may lead to stable marital satisfaction for spouses who have less positive expectations about specific outcomes. To test this possibility, each measure of global expectations was entered into a second stage model along with each measure of specific expectations and an interaction term to account for the changes in satisfaction estimated in equation 1 . Effect si2£s of these interactions are reported in Table 12. As can be seen, there were several significant positive interactions between expectations for stability and the measures of specific expectations. To deconstruct these interactions, several values of each variable were substituted into the regression equation and the results were plotted

PAGE 53

47 (see Aiken & West, 1991). As suggested by Cohen and Cohen (1983), three values were chosen for each of the two variables: the mean, 1 standard deviation above the mean, and 1 standard deviation below the mean. Figure 1 presents a depiction of two of these deconstructions. Figure A represents the interaction between wives' expectations for stability and wives' expectation for steadiness. Figure B represents the interaction between husbands' expectations for stability and husbands' expectations for their partner. As can be seen from Figure 1 , these two interactions do not reveal consistent patterns. According to Figure A, wives who expected positive global outcomes and positive specific outcomes tended to have the most positive slopes, whereas according to Figure B, spouses who expected less positive global outcomes and less positive specific outcomes tended to have the most positive slopes. Although these figures represent only two of the interactions, the remaining interactions reveal similar inconsistencies. For this reason, it is difficuh to draw strong conclusions about the interactive effects between global and specific expectations. Did Positive Expectations Lead to Positive Changes in Marital Satisfaction When They Were Likely to Get Confirmed ? A premise of this dissertation has been that positive expectations should be adaptive to the extent that they are confirmed, but maladaptive to the extent that they are disconfirmed. Based on this idea, it was predicted that expectations about more global relationship outcomes, because they should be relatively likely to get confirmed, would lead to more stable marital satisfaction over time, whereas expectations about specific relationship outcomes, because they should be relatively likely to get disconfirmed, would lead to declines in marital satisfaction over time. This prediction was not

PAGE 54

48 supported. Instead, both positive global and positive specific expectations were associated with negative relationship outcomes over time, suggesting that both types of expectations were likely to be disconfirmed over time. However, it remains possible that when they were confirmed positive expectations led to positive changes in satisfaction. Although on average spouses may have experienced specific outcomes that disconfirmed their expectations, some couples may have experienced positive relationship events that confirmed their expectations. Thus, the association between expectations and change in marital satisfaction may be moderated by the quality of the relationship, or the likelihood that the expectations will be confirmed. When the quality is low, positive expectations are likely to get disconfirmed, leading to decreases in satisfaction. In contrast, when the quality is high, positive expectations are likely to get confirmed, leading to more stable trajectories of marital satisfaction over time. To test this possibility, two processes associated with relationship quality that should lead to the disconfirmation of positive expectations were examined as potential moderators of the relationship between each measure of expectations and changes in marital satisfaction: negative behavior and relationship attributions. Negative behaviors during interactions should lead to disconfirmation of a positive expectation because they are negative relationship events. To the extent that spouses or their partners behave more negatively, spouses' positive expectations are likely to be disconfirmed. To the extent that spouses and their partners behave less negatively, spouses' positive expectations are likely to be confirmed. Thus, because the proportions of each spouse's negative behaviors were highly correlated (r = .57), indices of husbands and wives negative behavior were combined to form a single index of relationship quality: the proportion of negative

PAGE 55

behavior in the relationship. Because they represent the perception of negative relationship events, more negative attributions should also lead to disconfirmation of positive expectations. To the extent that spouses perceive their partners as the cause of a negative event or as blameworthy for a negative event, their positive expectations are likely to be disconfirmed. To the extent that spouses free their partners from blame for events, their positive expectations are likely to be confirmed. To test these possibilities, interactions between these expectations and negative behavior and expectations and attributions were entered into a second-stage model to account for between-subject variability in the changes in satisfaction estimated in equation 1 . Consistent with the idea that positive expectations should lead to negative satisfaction only when they are likely to get disconfirmed, negative interactions between expectations and negative behavior would suggest that negative changes in satisfaction were most likely when positive expectations were high and negative behavior was high. The effect sizes of the interactions between expectations and behavior are presented in the first column of Table 13. Consistent with predictions, results reveal a significant pattern of negative interactions between expectations and negative behavior for wives. Figure 2 depicts a representation of one of these interactions, the interaction between wives' negative behavior and wives' expectations for stability. As can be seen, in Figure 2, when behavior was most negative, very positive expectations were associated with the steepest declines in marital satisfaction whereas less positive expectations were associated with the most stable satisfaction. However, when behavior was the least negative, the least positive expectations were associated with the steepest declines whereas the most positive expectations were associated with the most stable satisfaction.

PAGE 56

50 In other words, when rates of negative behavior were high, positive expectations led to declines in marital satisfaction, but when rates of negative behavior were low, positive expectations led to stable satisfaction. In general husbands and wives did not differ significantly in the extent to which their expectations and negative behavior interacted to affect changes in their marital satisfaction over tune. The exception was that the interactive effect for wives' expectations for partner and negative behavior on changes in scores on the SMD was significantly greater than the interactive effect for husbands, t = 6.6, p < .05. Of note, one interaction between wives expectations for growth and negative behavior was significantly positive, suggesting the opposite pattern of results. However, because all other significant and margmally significant interactions were negative, it is difficult to draw conclusions about this interaction because it may be spurious. Analyses also tested for interactions between expectations and attributions on changes in satisfaction. Consistent with the idea that positive expectations should lead to declmes when they are likely to get disconfirmed, positive interactions between expectations and attributions would suggest that changes in satisfaction are most negative when expectations are positive and attributions scores negative. To test this possibility, each sub-scale of the RAM was entered separately into a second-stage analysis along with each measure of expectations and an interaction term to predict the changes in satisfaction estimated in equation 1. The effect sizes of the interactions between each measure of attributions and each measure of global expectations are reported in the second and third columns of Table 13. Resuhs reveal a significant pattern of positive interactions for both types of attributions. Figure 3 provides a representation of the interaction between wives' expectations for their partner and their attributions for

PAGE 57

51 responsibility, and Figure 4 provides a representation of the interaction between wives' expectations for steadiness and their attributions for causality. As can be seen in these figures, when spouses made negative attributions, positive expectations were associated with the steepest declines in satisfaction whereas less positive expectations were associated with the most stable satisfaction. However, when spouses made positive attributions, less positive expectations were associated with the steepest declines in satisfaction whereas the most positive expectations were associated with the most stable marital satisfaction. In other words, positive expectations led to negative changes in marital satisfaction when spouses made attributions for their partner's behavior that were likely to disconfirm their expectations, but led to positive changes in marital satisfaction when they made attributions that were likely to confirm their expectations. Although more interactions between expectations and attributions were significant for wives than for husbands, in general husbands and wives did not differ significantly in the extent to which their expectations and their attributions interacted to affect changes in their marital satisfaction over time. The only exception was that that the interactive effect for husbands' expectations for problems and husbands' responsibility attributions on changes in satisfaction according to the SMD was significantly different from the interactive effect for wives, t = 3.8, p < .05

PAGE 58

52 Table 1 . Descriptive Statistics for Independent Variables at Time 1 Husbands Wives Possible Range M SD M SD Global Expectations Expectations for growth 2 12 10.1 1.8 10.3 1.6 Expectations for stability 1 5 4.9 0.36 4.9 0.30 Soecific Exnectations Expectations for steadiness 2 12 9.0 2 1 9 1 2 1 Expectations for partner 9 63 45.5 6.5 43 9 8 6 Expectations for problems Mediators 11-57 42.7 4.7 41.9 3.8 Attributions for causality 12 84 53.3 11.2 51.4 9.9 Attributions for responsibility 12 84 63.6 12.1 60.8 14.6 Negative behavior 0% 100% 10.2% 10.2% 13.4% 11.8%

PAGE 59

i 53 00 6 H 1/3 as > u T3 C U a, T3 c I— I 00 c i CS C/3 c o 1 o U H c o o n o d d 1 • « 1 + * * « CN d o d O o 1 1 * * m fo 00 « • # d d d d d 1 * « OS m « 00 o OS o « lO d d d * d d d o « o CM o o d d d d d d CO C o CO ion o u ctat Exp pe o X s m 'o CO x> > 00 CD V q V V CO c o CO CO to C o 1 O o "(O CO JO CO 3 ni o T3 CO

PAGE 60

54 C/3 > 00 O o o o o >o O O o d d d d 1 o d o O o XI O o oo o o « d d d d o 1 d o 1 d * o (N o « rj O o d d d d o 1 o d d 1 IT) 00 o o • oo o + 00 o d d d d o d d d 00 o • • CO « « + On On O d d d o d d d d m • • ON fN * « "J o • d d o d d d d d 1 + « « o o d o d d d d d d .22* CO o o o o o d d d d d d 1 d CQ C/3 c O u X) u > u c <*-> o c o a S Oh O V q V V

PAGE 61

Vi o 1/1 o c :3 CO «4-l c 3 (U PQ -a c CO o o o ON o * d d 1 d d d o o On O + o d d d d d O d O d in o o C o o 5 o o w c2 CO C o o
PAGE 62

56 Table 5 . Mean Marital Quality Scores Across Five Waves of Measurement for Husbands and Wives Spouse Time 1 Time 2 Time 3 Time 4 Time 5 Semantic Differential Husbands M 96.3 92.0 92.5 92.1 93.5 SD 8.8 14.1 14.8 14.7 13.9 N 81 76 74 67 64 Wives M 97.7 94.8 93.3 92.1 93.8 SD 10.7 12.9 16.0 14.7 15.6 N 82 77 73 68 66 Quality Marriage Index Husbands M 42.1 40.3 40.3 40.1 40.8 SD 4.0 6.7 6.6 7.0 6.5 N 81 76 74 67 64 Wives M 42.1 40.8 39.9 39.0 40.2 SD 5.3 6.0 7.5 8.6 7.3 N 82 77 73 68 66 Inventory of Specific Marital Problems Husbands M 52.0 49.1 48.1 48.7 48.6 SD 24.0 24.1 22.6 24.0 25.8 N 82 75 74 67 64 Wives M 48.2 44.2 48.6 46.4 44.0 SD 24.4 20.4 26.8 24.0 21.5 N 82 76 73 68 66 Note. Means using only participants reporting at time 5 are similar and demonstrate similar trends.

PAGE 63

57 Table 6 . Change in Marital Satisfaction Chi-Square test Measure M SD t! of variance Intercept SMD T T 1 1 Husbands 95.7 1.1 149.4 Wives 97.9 1.3 196.0"* QMI Husbands 41.9 0.52 126.0" Wives 42.3 0.60 166.6*" O 1 Slope SMD Husbands -0.78 0.32 -2.4* 131.3*" Wives -1.2 0.36 -3.3" 161.5*" QMI Husbands -0.38 0.15 -2.5* 114.8" Wives -0.62 0.19 -3.3" 181.8*" Note. For t test, df 81; for chi-square tests, df = 75. SMD = Semantic Differential (Osgood et al, 1957); QMI = Quality Marital Index (Norton & Glick, 1979). 'The t test of the intercepts addresses the hypothesis that the intercepts differ significantly form zero. Because the lowest possible score on each of these measures is greater than zero, these tests are not meaningful and hence are not reported. < .10, *E < .05, "p < .01 *" p < .001, two tailed

PAGE 64

58 Table 7 . Covariance between Specific Marital Problems and Marital Satisfaction Controlling for the Trajectory of Marital Satisfaction ChiSquare test Measure M SD t of variance SMD Husbands -0.27 0.03 -8.5*** 147.5*** Wives -0.30 0.03 -9.7*** 126.6** QMI Husbands -0.10 0.02 -6.7*** 123.7*** Wives -012 0^01 -9.4*** 192.3*** Note . For t test, df = 81; for chi-square tests, df = 70. SMD = Semantic Differential (Osgood et al, 1957); QMI Quality Marital Index (Norton & Click, 1979). ^ p < .10 * p < .05 ** p < .01 *** p < .001, two tailed

PAGE 65

59 Table 8 . Effects of Specific Expectations on the Trajectory of Marital Satisfaction Husband Wife t effect size t effect size Satisfaction intercepts Expectations for steadiness SMD 3.6" 0.37 2.9" 0.31 QMI 3.0** 0.32 2.7** 0.29 Expectations for partner SMD 4.8*** 0.47 1.6 0.18 QMI 4.3*** 0.43 1.6 0.18 Expectations for problems SMD 0.16 0.02 -0.33 -0.04 QMI 0.40 0.04 -1.0 -0.11 Satisfaction slopes Expectations for steadiness SMD -1.6 -0.18 -1.6 -0.18 QMI -1.8^ -0.20 -2.3* -0.25 Expectations for partner SMD -2.4* -0.26 0.30 0.03 QMI -2.5* -0.27 0.02 0.00 Expectations for problems SMD -0.16 -0.02 -2.1* -0.23 QMI -0.28 -0.03 -0.82 -0.09 Note. ^ p < .10 ' p < .05 " p < .01 "' p < .001, two tailed

PAGE 66

60 Table 9 . Effects of Specific Expectations on the Covariance between Specific Problems and Marital Satisfaction Husband Wife t effect size t effect size Expectations for steadiness SMD -0.71 -0.08 -1.5 -0.17 QMI -0.30 -0.03 -1.5 -0.17 Expectations for partner SMD 0.54 0.06 -0.76 -0.08 QMI 0.65 0.07 -0.75 -0.08 Expectations for problems SMD 1.7 0.19 -0.69 -0.08 QMI 1.6 0.18 0.36 0.04 Note. ^ p < .10 ' p < .05 " p < .01 p < .001, two tailed

PAGE 67

61 Table 10 . Effects of Global Expectations on the Trajectory of Marital Satisfaction. Husband Wire t effect size t enect size Satisfaction intercepts Expectations for growth SMD 2.9" 0.31 -0.76 -0.08 2.9" 0.31 -0.41 -0.05 JZ/ApcCldllOna lOr SldUllUV 2.5* OiVLL/ 0.27 1.7 A 1 n 0.19 2.5* 0.27 1 1.0 n 1 o O.lo Satisfaction slopes Expectations for growth SMD -2.5* -0.27 0.23 0.03 QM -2.9" -0.31 0.12 0.01 Expectations for stability SMD -1.8" -0.20 -0.23 -0.03 QMI -2.9** -0.31 -0.52 -0.06 Note. ^p<. 10 ' p<.05 p < .01 p < .001, two tailed

PAGE 68

62 Table 1 1 . Effects of Global Expectations on the Covariance between Specific Problems and Marital Satisfaction. Husband Wife t effect size t effect size Expectations for growth SMD -1.8^ -0.20 -0.88 -0.10 QNG -1.3 -0.14 -0.29 -0.03 Expectations for stability SMD -0.73 -0.08 -2.r -0.29 QMI -1.5 -0.17 -1.6 -0.18 Note. ^p<. 10 ' p<.05 "p<.01 '"p<. 001, two tailed

PAGE 69

63 V3 c o u V gm X) o O CO CO o o u W o o a C/D C c p., •T3 C CO CO 3 X CO CO C CO a> t/3 C CO CO ON >0 o o o o 00 VO o o O CD CO o o <^ 00 o o o o o u >H — in o o o o o ^ o o o o o o d d >o o o ^ o o fN 00 — o <=) d Tj00 d d o ^ o o o a:s CO C/3 !/3 C T3 Ji i o o V a. V o V V D. h u o

PAGE 70

64 o o d d o ^ d d d d o o d d d d + + O On d d d d d d 00 o o d d OO o o d d ^ o o — d d • + —I o d d do d d CO 00 (N d d o o d d O fN d d o o d d OO o o — ' d d d d • 00 o « rn « + o « + ON O cs o iled d d d d d d d d d d -*-» 1 ON o o o On O o ON o 00 -*-» O O d d d 1 d 1 d d d d d d 1 V C/3 c 00 C o u CO partner nrohlerr for for ) ions ions o X W 1 a u 1 o X w o V o V V *-> o

PAGE 71

9i CN O CN CO CO a qj 9 c? qj UOIpe^SI^BS U! 9DUBL|0 CD i£> TID ai

PAGE 72

66 less positive expectations more positive expectations high low proportion . proportion hegative behavior Figure 2 . Interactive Effects of Wives' Expectations for Stability and Negative Behavior on the Trajectory of Their Marital Satisfaction

PAGE 73

67 negative positive responsibility attributions less positive expectations — nnore positive expectations Figure 3 . Interactive Effects between Wives' Expectations for Partner and Responsibility Attributions on the Trajectory of Their Marital Satisfaction

PAGE 74

68 negative positive causality attributions less positive expectations •more positive expectations Figure 4 . Interactive Effects of Wives' Expectations for Steadiness and Causality Attributions on the Trajectory of Their Marital Satisfaction

PAGE 75

DISCUSSION Study Rationale and Summary of Results Given that expectations influence a wide variety of interpersonal phenomena (Olson et al., 1998), the expectations that spouses have for their relationships should have implications for the development of those relationships over time. Consistent with this idea, research on expectancy confirmation suggests that positive expectations should lead to positive outcomes because they are likely to get confirmed through the processes of behavioral and perceptual confirmation (see Darley & Fazio, 1980; Miller & TumbuU, 1986; Snyder, 1984, for reviews). In contrast, however, research on counterfactual reasoning suggests that positive expectations are likely to lead to disappointment when they are disconfirmed (Kahneman & Miller, 1986). In light of these fmdings, the current study contended that the benefits of positive expectations may depend on the likelihood that they are confirmed or disconfirmed. Positive expectations should lead to positive feelings about the relationship when they are likely to get confirmed and lead to less positive feelings about the relationship when they are likely to get disconfirmed. This dissertation examined whether the specificity of the expected outcome may be associated with the likelihood that expectations are confirmed or disconfirmed, and therefore influence future relationship outcomes. Global outcomes, because they are more open to interpretation and capable of being confirmed by a wide variety of behaviors, may be 69

PAGE 76

70 more likely to be confirmed. Specific outcomes, in contrast, because they are less open to interpretation and capable of being confirmed by only a limited variety of behaviors, may be more likely to be disconfirmed. The first goal of this study was to test the idea that expectations about outcomes at different levels of specificity should be differentially likely to lead to perceptual and behavioral confirmation. In pursuit of this goal. Hypothesis 1 stated that measures of global expectations would be positively associated with marital attributions (perceptual confirmation) and negatively associated with observations of negative behavior (behavioral confirmation). In other words, spouses who expected global outcomes to be positive should have perceived and interpreted events in positive ways and behaved less negatively during their interactions with their partners. This prediction was not supported. Measures of global expectations were not significantly associated with attributions or with observations of negative behavior. In other words, spouses with positive expectations about global outcomes were no more likely to make positive attributions and no less likely to behave negatively than spouses with less positive global expectations. Hypothesis 2 stated that expectations about more specific outcomes, because they are less likely to get confirmed, should not lead to perceptual confirmation and behavioral confirmation. Thus, it was predicted that measures of specific expectations would be unrelated to measures of marital attributions and observations of negative behavior. This prediction was supported. Measures of specific expectations were not significantly associated with measures of attributions and observations of behavior. Thus, in contrast to the general idea that specific and global expectations should be differentially likely to get confirmed by attributions and behavior, results suggest that

PAGE 77

71 both global and specific expectations were likely to get disconfirmed. There are several possible explanations for this finding. First, measures of attributions and behaviors may have been relatively weak operationalizations of perceptual and behavioral confirmation. For instance, the RAM assessed spouses' attributions about specific events, but did not assess their attributions about global events. Perhaps a measure that assessed spouses' attributions about more global outcomes would show that such attributions are associated with expectations about global outcomes. Second, measures of global and specific expectations may have been relatively weak operationalizations. Although the measures of global expectations were conceptually distinct, several correlations between global and specific measures were just as large as correlations among global measures and among specific measures, suggesting that the measures may have been too similar. Future research using measures that tap outcomes even more global than the outcomes assessed in the current study may reveal support for the idea that global expectations are more likely to get confirmed through perceptual and behavioral confirmation. One better way to measure global expectations may be to have spouses predict their future responses to measures of global marital satisfaction. The second goal of this dissertation was to examine the effects of positive expectations on changes in marital satisfaction and spouses' reactivity to relationship problems. Hypothesis 3 stated that positive expectations about specific outcomes, because they are more Ukely to get disconfirmed, would lead to negative changes in satisfaction and a stronger reactivity to relationship problems. The first part of this hypothesis was supported. Although measures of specific expectations were positively associated with initial marital satisfaction, positive specific expectations led to dechnes in

PAGE 78

72 marital satisfaction over the first two and a half years of marriage. In other words, spouses who held the most positive expectations about specific outcomes were the most happy initially, but were also the most likely to experience declines in marital satisfaction over time. With respect to the second part of Hypothesis 3, positive specific expectations did not lead spouses to be more reactive to relationship problems. Measures of specific expectations were not associated with the covariance between satisfaction and relationship problems, or the extent to which their global satisfaction changed in response to specific problems. Hypothesis 4 stated that positive expectations about global outcomes, because they are more likely to get confirmed, would lead to positive changes in marital satisfaction and less reactivity to relationship problems. This hypothesis was not supported. Instead, for husbands, although measures of global expectations were positively associated with initial satisfaction, positive expectations about global outcomes led to declines in satisfaction over the fu-st two and a half years of marriage, similar to the effect for specific expectations. Husbands who held the most positive expectations about global outcomes were the most happy initially, but were the most likely to experience declines m marital satisfaction over time. Although the associations between measures of global expectations and decreases in satisfaction were only significant for husbands, there were no significant gender differences in these associations. With respect to the second part of Hypothesis 4, positive global expectations did not lead spouses to be less reactive to relationship problems. Instead, a few measures of global expectations were significantly associated with the covariance between satisfaction and relationship

PAGE 79

73 problems, suggesting husbands and wives with positive expectations about some global outcomes were more reactive to relationship problems. Thus, regardless of their level of specificity, positive expectations led to decreases in marital satisfaction, suggesting that both positive global and positive specific expectations were likely to be disconfirmed. However, further analyses revealed that when positive expectations were likely to be confirmed , they led to positive changes in satisfaction. Results revealed a consistent pattern of significant interactions between expectations and two measures of relationship quality likely to be associated with the likelihood that positive expectations were or disconfirmed: negative behavior and attributions. When behavior was least negative or when attributions were most positive, likely confirming positive expectations, positive expectations led to more stable satisfaction over time and less positive expectations led to declines in satisfaction over time, suggesting that positive expectations may be beneficial. In contrast, when behavior or attributions were most negative, likely disconfirming positive expectations, less positive expectations led to more stable satisfaction over time whereas more positive expectations led to declines in marital satisfaction. Importantly, although less positive expectations were associated with more stable satisfaction for spouses likely to encounter negative experiences, because expectations were also positively associated with satisfaction initially, spouses with less positive expectations were less satisfied initially. In other words, when less positive expectations maintained satisfaction, they did so in less satisfying relationships. In sum, the fmdings of the current research suggest that the effects of positive expectations on relationship outcomes may depend on the likelihood that such

PAGE 80

74 expectations will get confirmed. When positive expectations are readily confirmed, they lead partners to experience stable relationship satisfaction over time. In contrast, when positive expectations are likely to get disconfirmed, they lead spouses to experience declines in marital satisfaction over time. When Are Positive Expectations Likelv to Get Confirmed? One factor that may affect the likelihood that positive expectations will get confirmed is the quality of the relationship. The current study showed that two aspects of relationship quality, negative attributions and negative behavior, moderated the association between expectations and marital happiness. However, in addition to these factors, there are likely to be other factors associated with marital quality that lead positive expectations to get disconfirmed. Because researchers and therapists must have a complete understanding of these factors in order to know the implications of positive expectations in relationships, these findings highlight the importance of research to develop a better understanding of these factors. In addition to qualities of the relationship, qualities of the expected outcomes may affect the extent to which positive expectations get confirmed. The current study argued that the specificity of the expected outcome was one factor that should affect the likelihood that an expectation is confirmed. This idea was not supported. Instead, global and specific expectations had similar effects on relationship satisfaction and spouses' reactivity to relationship problems. As discussed earlier, one reason that measures of global and specific expectations may not have demonstrated different effects on relationship outcomes may be that measures of global and specific expectations did not validly capture the distinction between them. Future research may benefit by examining

PAGE 81

75 the effects of global and specific expectations using meastires that more properly capture this difference. In addition to the specificity of the outcome, the controllability of the expected outcome may affect the extent to which expectations get confirmed. Research on behavioral confirmation shows that spouses behave in a manner consistent with their prior expectations thereby creating outcomes consistent with their expectations (Darley & Fazio, 1980; Snyder, 1984). Thus, expectations about outcomes that are controllable may be more likely to get confirmed than expectations about outcomes that are not controllable. Previous research offers support for this idea. As discussed previously, research on expectancy confirmation in marital interactions has revealed that spouses' expectations for an upcoming problem-solving discussion were not related to the way they behaved during the discussion (McNulty & Kamey, under review). The authors speculated that problem-solving behavior is a skill that develops independently of expectations, and thus is not under some spouses' control. Future research may benefit by examining the how the controllability of the expected outcomes influences the extent to which expectancy confirmation processes are likely to operate. Some authors have argued that positive beliefs are beneficial only to the extent that they are realistic (see Baumeister, 1989, for related discussion). Consistent with this idea, the qualities of the relationship and qualities of the expected outcome that make some expectations more likely to get confirmed than other outcomes may make some expectations more realistic than others. When qualities of the relationship are likely to produce positive outcomes, or when positive outcomes are under spouses' control, positive expectations are realistic and beneficial to the relationship. In contrast, when

PAGE 82

76 qualities of the relationship are unlikely to render spouses vubierable to negative experiences, or when positive outcomes are not under spouses' control, positive expectations are somewhat unreaUstic and may be detrimental to the relationship. In this light, realistic positive expectations are likely to lead to positive outcomes, whereas unrealistic positive expectations are likely to lead to disappointment. Additional Directions for Future Research The current study showed that spouses' expectations about their relationships affected their own satisfaction over the first two and a half years of marriage. Spouses' expectations may also influence their partners' satisfaction over time. Consistent with this idea, some research suggests that spouses' expectations lead expectancy-consistent outcomes for their partners (Downey et al, 1998; Drigotas, Rusbult, Wieselquist, & Whitton, 1999). For example, Downey et al. (1998) showed that spouses who expected their partners to behave negatively, responded in ways that predicted greater negative feelings in their partners. It is also possible, however, that holding unrealistic positive expectations about the partner or the relationship may lead to negative outcomes for partners by putting strain on the partner to meet the expectations. Consistent with this idea, spouses often report that their partners' unrealistic expectations are a problem in the relationship (Geiss & O'Leary, 1981). Future research may help reconcile these inconsistent fmdmgs by revealing the conditions under which positive expectations are likely to lead to positive versus negative outcomes for partners. As was the case in the current study, it is possible that the effects of positive expectations may depend on the spouses' beliefs about Ukelihood that the expectation can be confirmed. When spouses feel they that they can meet their partners' high expectations, such expectations may

PAGE 83

77 make them feel good. However, when spouses feel that they can not meet their partners' high expectations, such expectations may be stressful (see Neff & Kamey, m press, for related discussion). Future research may also benefit by examming the sources of expectations. Given the link between expectations and changes in marital satisfaction, the sovirces of expectations are likely to be distal causes of change in marital satisfaction. Because people learn what to expect through their experiences (Olson et al., 1998), one source of spouses' expectations about their relationship may be their experiences with the relationship. Consistent with this idea, research by Fincham, Harold, and Gano-Phillips (2000) suggests that spouses' explanations of the events in their relationships mfluence their future expectations. In particular, spouses who made more positive attributions were more likely to believe that they had the ability to resolve conflicts with their partners and were happier with their relationships. Thus, longitudinal research may reveal that specific relationship experiences are important distal sources of change in marital satisfaction. In addition to their experiences with their current relationships, another source of spouses' expectations about the relationship may be spouses' experiences with other relationships. Research and theory on attachment style (Amsworth, Blehar, Waters, & Wall, 1978; Bowlby, 1969; Hazan & Shaver, 1987) suggests that the relationship that spouses had with their primary caregiver affects the expectations they develop about their current relationships (Baldwin, Keelan, Fehr, Enns, & Koh-Rangarajoo, 1996). In this study, partners who had positive relationships with their primary caregivers and therefore developed positive attachments styles held positive expectations about their relationships, whereas spouses who had more negative relationships with their primary caregivers and

PAGE 84

78 therefore developed more negative attachment styles held more negative expectations about their relationships. Thus, given that attachment style has been linked to marital satisfaction and stability (see Collins «& Allard, 2001, for a review), expectations may mediate the effects of attachment style on changes in marital satisfaction. Longitudinal research is needed to evaluate this possibility. Another source of spouses' expectations about their relationships may be their parents' relationships with one another. Spouses may develop their own beliefs about relationships based on their experiences with their parents' relationships. If their parents' relationship was positive, they may develop positive expectations about relationships. On the other hand, if their parents' relationship was negative, they may develop less positive expectations about relationships. Research has provided some preliminary evidence consistent with this idea. In one study, students' optimism about their own future marriages was positively associated with their perceptions of the quality of their parents' relationship (Camelley & Janoflf-Bulman, 1992). Future research could expand on this finding by examining whether partners perceptions of their parents' relationships affects the quality other own relationships, and whether expectations mediate this association. In addition to broader beliefs, personality characteristic may also lead people to expect various relationship outcomes. If so, given that personality has been shown to be associated with relationship satisfaction [e.g., \(Huston & Houts, 1998; Huston & Vangelisti, 1991; Karney & Bradbury, 1997; Karney, Bradbury, Fincham, & Sullivan, 1994), expectations may mediate the effects of various personality characteristics on future relationship satisfaction. Moreover, it is also possible that the effects of personality on satisfaction may completely account for the association between expectations and

PAGE 85

79 satisfaction observed in the current study. Spouses' personality characteristics may lead them to hold positive or less positive expectations and to experience steep or flat trajectories of satisfaction, making any observed associations between expectations and satisfaction spurious. For instance, research suggests that spouses high in neuroticism are likely experience negative relationship outcomes (Kamey & Bradbury, 1997). Because neurotic spouses may also expect less positive outcomes, any observed association between expectations and satisfaction may be due to shared variance between expectations and neuroticism. This possibility is important to consider because it has implications for therapy. Interventions aimed at changes expectations will be futile if expectations are not uniquely associated with satisfaction. Thus, future research may benefit by examining the unique effects of expectations and personality on satisfaction. Strengths and Limitations Several strengths of this research enhance my confidence in the fmdings. First, the general pattern of results replicated across conceptually distinct measures of expectations and conceptually distinct measures of disconfu-mation, suggesting that results were not unique to one construct or one method of measurement. Second, whereas the average rate of attrition in prior research on marriage is 31% (Kamey & Bradbury, 1995), during the two years of this study attrition was relatively low (21%), reducing the likelihood the results can be attributed to attrition bias. Third, all spouses entering the study were newlyweds, reducing the risk that important differences were confounded with marital duration (see, Booth, Johnson, White, & Edwards, 1985; South & Spitze, 1986). Fourth, whereas the majority of longitudinal research on marriage has collected and analyzed

PAGE 86

80 only two waves of data, the current analyses examined five waves of data and analyzed all waves simultaneously using GCA. Despite these strengths, several factors nevertheless limit interpretations of the c\irrent findings. First, whereas the homogeneity of this sample enhances our confidence in the pattern of associations, generalizations to other samples should be made with caution. For example, although the expectations spouses have in the beginning of their relationships influence the later development of their relationships, the expectations of more established couples may be more accurate and thus less likely to be disconfirmed (see Jussim & Eccles, 1995). Second, the homogeneity of the sample was also limited the range of expectations assessed in the current study. Because all couples assessed in the current study were newlyweds, they were unlikely to hold negative expectations about their relationships. Research examining more negative expectations may yield effects different from those obtained in the current study. For instance, ahhough the current study revealed that less positive expectations were more adaptive than very positive expectations for couples likely to experience negative outcomes, it is unlikely that negative expectations would also be adaptive. Instead, negative expectations may be more likely to lead couples to abandon the efforts to resolve relationship problems leading to the eventual dissolution of the relationship (see Knee, 1998). Third, although sample size compared favorably to other longitudinal studies of marriage, a larger sample size 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, associations between global expectations and perceptual and behavioral confirmation may have been significant.

PAGE 87

81 Conclusion Several writers have argued that spouses should lower their expectations to avoid bemg disappointed by outcomes that do not meet them (e.g., Talbot, 1997). Results of the current study argue against this recommendation. Spouses who held positive expectations that were likely to get confirmed by positive relationship experiences were more satisfied over time than spouses who held less positive expectations. Furthermore, although spouses who held less positive expectations that were unlikely to get confirmed did maintain a steadier trajectory of satisfaction over time, because expectations were positively associated with initial satisfaction, these spouses were less happy initially. Thus, less positive expectations only maintained satisfaction in less satisfying relationships. Because results suggest that the benefits of positive expectations depend on other relationship factors, the current findings highlight the importance of taking into accoimt the broader factors associated with the likelihood that expectations are confirmed and disconfirmed in close relationships.

PAGE 88

APPENDIX A SEMANTIC DIFFERENTIAL MEASURE OF MARITAL SATISFACTION For each of the following items, fill in the circle (O) that best describes HOW YOU FEEL ABOUT YOUR MARRIAGE. Base your responses on your first impressions and immediate feelings about the item. INTERESTING O o 0 0 o 0 0 BORING BAD 0 0 o o o 0 o GOOD UNPLEASANT o 0 0 o o 0 o PLEASANT FULL 0 o o o o o 0 EMPTY WEAK 0 o o o o o o STRONG SATISFIED 0 o o 0 o 0 o DISSATISFIED LONELY o o o o o o o FRIENDLY STURDY o o o 0 o o o FRAGILE REWARDING 0 o o o o o o DISAPPOINTING DISCOURAGING 0 o o 0 0 0 0 HOPEFUL ENJOYABLE o o o 0 0 o o MISERABLE TENSE o o o 0 o o 0 RELAXED STABLE 0 o 0 o o 0 o UNSTABLE HAPPY 0 o o 0 o 0 o SAD STRESSFUL 0 o 0 o o 0 o PEACEFUL 82

PAGE 89

APPENDIX B QUALITY MARRIAGE INDEX Please indicate how well the following statements describe you and your marriage. Very Strong DISAGREEMENT 1 2 3 Very Strong AGREEMENT 6 7 We have a good marriage. My relationship with my partner is very stable. Our marriage is strong. My relationship with my partner makes me happy. I really feel like part of a team with my partner. o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o All things considered, how happy are you in your marriage? Very UNHAPPY 1 2 3 Happy 5 6 7 Perfectly HAPPY 8 9 10 oooooooooo 83

PAGE 90

APPENDIX C INVENTORY OF SPECIFIC MARITAL PROBLEMS All couples experience some difficulties or differences of opinion in their marriage, even if they are only very minor ones. Listed below are a number of issues that might be difficulties in your marriage. For each issue fill in a bubble to indicate how much it is a source of difficulty or disagreement for you and your spouse. Not a Major Problem Problem 1 234 56 789 10 11 Children 0 o o o o o o 0 o o o Rehgion 0 o o o 0 o o o o o 0 In-laws, parents, relatives 0 o o o o o o o o o 0 Recreation and leisure time o 0 o 0 o o o 0 o o 0 Communication o o o o o o o o o o 0 Household management o o o o o o o o o o 0 Showing Affection 0 0 o o o o o 0 o o 0 Making decisions o 0 0 o o o o o o o o Friends 0 0 o o o o o o o 0 0 Unrealistic expectations o 0 0 0 o 0 o o 0 o 0 Money management o o 0 0 o o o o 0 0 o Sex 0 o 0 0 o o 0 o o o 0 Jealousy o o o o o o o 0 o o o Solving problems o 0 0 o o o o 0 o 0 o Trust o o 0 o 0 o o o o o 0 Independence o o o o o 0 o o 0 o o Drugs and alcohol o o o o o o o o o o o Career decisions o o o o o o o o o 0 o Amount of time spent together o o o o o o o o o o o 84

PAGE 91

APPENDIX D EXPECTATIONS FOR FUTURE RELATIONSfflP SATISFACTION Please fill in the circles that best describe your opinions about your relationship: 1 . Over the NEXT six months , do you expect that your overall feelings about your marriage will become: o o o o o Much worse AUttle Stay A little Much better worse the same better 2. Over the NEXT FOUR YEARS , do you expect that your overall feelings about your marriage will become: o o o o o Much worse A little Stay A little Much better worse the same better 85

PAGE 92

APPENDIX E EXPECTED TRAJECTORY OF MARITAL SATISFACTION Please fill in the circle beneath the picture that best describes how you think your relationship will go during the next SIX MONTHS. GOO 86

PAGE 93

APPENDIX F EXPECTATIONS FOR STABILITY 1 . How likely is it that the two of you will still be married four years from now? O Very unlikely O Not too likely O Pretty likely O Very likely O Completely certain 87

PAGE 94

APPENDIX G EXPECTATIONS FOR STEADINESS Over the NEXT six months , which of the following describes how your feelings towards your marriage are likely to change: o o o o o Major highs/ Some ups A few Pretty Very steady major lows and downs ups and steady downs Over the NEXT FOUR YEARS , which of the following describes how your feelings towards your marriage are likely to change: o o o o o Major highs/ Some ups A few Pretty Very steady major lows and downs ups and steady downs 88

PAGE 95

APPENDIX H EXPECTATIONS FOR PARTNER In response to the following statements, indicate how strongly you agree by filling in the appropriate bubble. Strongly Strongly Disagree Agree 1 2 3 4 5 6 7 My partner will always take time for me when I need him/her. O o o o o 0 o My partner will rarely make mistakes. u U u u u \J My partner will agree with me about the important things. O o O 0 0 0 o My partner will always get along well with my parents. 0 o o o o 0 o My partner will sometimes lose his/her temper. 0 o o 0 0 o o My partner will never disappoint me. o o o o o o o My partner will always take care of me. o o o o o o o My partner will always be attractive to me. o o o 0 o o o My partner will always make me happy. o o o o o 0 o 89

PAGE 96

APPENDIX I EXPECTATIONS FOR SPECIFIC RELATIONSHIP PROBLEMS For each of the marital problems Usted below, please indicate how you think that problem is going to develop for you and your spouse over the NEXT SIX MONTHS . Get Stay Get worse the same better Cniluren O O O Religion o o o In-laws, parents, relatives o o o Recreation and leisure time o o 0 Communication o o 0 Household management \3 u u Showing Affection o o 0 Making decisions o o 0 Friends o 0 o Unrealistic expectations o o o Money management o 0 o Sex 0 o 0 Jealousy o o o Solving problems o o 0 Trust o o 0 Independence 0 o o Drugs and alcohol o o 0 Career decisions o o o Amount of time spent together o o 0 90

PAGE 97

APPENDIX J RELATIONSfflP ATTRIBUTIONS MEASURE This questionnaire describes several things that your spouse might do. Imagine your spouse performing each behavior and then bubble in the number that indicates how much you agree or disagree with each statement, using the following scale: 1 2 3 4 5 6 7 disagree disagree disagree neutral agree agree agree strongly somewhat somewhat strongly YOUR SPOUSE CRITICIZES SOMETHING YOU SAY: 1 2 3 4 5 6 7 My spouse's behavior was due to something about him/her (e.g., the type of person he/she is, his/her mood) GOGOOGG My spouse's behavior was due to something about me (e.g., the type of person I am, the mood I was in) GGGOGGG The reason my spouse criticized me is not likely to change GGOGOGG The reason my spouse criticized me is something that affects other areas of our marriage GGGGOGO My spouse criticized me on purpose rather than unintentionally GOGGGGG My spouse's behavior was motivated by selfish rather than unselfish concerns GGGGGOG My spouse deserves to be blamed for criticizing me GGGGGGG YOUR SPOUSE BEGINS TO SPEND LESS TIME WITH YOU: 1 2 3 4 5 6 7 My spouse's behavior was due to something about him/her (e.g., the type of person he/she is, his/her mood) GGGGGGG My spouse's behavior was due to something about me (e.g., the type of person I am, the mood I was in) GGGOGGG The reason my spouse spent less time with me is not likely to change GGGGGGG The reason my spouse spent less time with me is something that affects other areas of our marriage GGGGGGG My spouse spent less time with me on purpose rather than unintentionally GGGGGGG My spouse's behavior was motivated by selfish rather than unselfish concerns GGGGGGG My spouse deserves to be blamed for spending less time with me GGGGGGG 91

PAGE 98

92 1 2 3 4 5 6 7 disagree disagree disagree neutral agree agree agree strongly somewhat somewhat strongly YOUR SPOUSE DOES NOT PAY ATTENTION TO WHAT YOU ARE SAYING: 1 2 3 4 5 6 7 My spouse's behavior was due to something about him/her (e.g., the type of person he/she is, his/her mood) OOOOOOO My spouse's behavior was due to something about me (e.g., the type of person I am, the mood I was in) OOOOOOO The reason my spouse spent less time with me is not likely to change OOOOOOO The reason my spouse spent less time with me is something that affects other areas of our marriage OOOOOOO My spouse spent less time with me on purpose rather than unintentionally OOOOOOO My spouse's behavior was motivated by selfish rather than unselfish concerns OOOOOOO My spouse deserves to be blamed for spending less time with me. ... OOOOOOO YOUR SPOUSE IS COOL AND DISTANT: 1 2 3 4 5 6 7 My spouse's behavior was due to something about him/her (e.g., the type of person he/she is, his/her mood) OOOOOOO My spouse's behavior was due to something about me (e.g., the type of person I am, the mood I was in) OOOOOOO The reason my spouse spent less time with me is not likely to change OOOOOOO The reason my spouse spent less time with me is something that affects other areas of our marriage OOOOOOO My spouse spent less time with me on purpose rather than unintentionally OOOOOOO My spouse's behavior was motivated by selfish rather than unselfish concerns OOOOOOO My spouse deserves to be blamed for spending less time with me. ... OOOOOOO

PAGE 99

93 REFERENCES Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions . Newbury Park, CA: Sage. Ainsworth, M. D. S., Blehar, M. C, Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the Strange Situation . Hillsdale, NJ: Lawrence Erlbaum. Anderson, S. M., & Bern, S. L. (1981). Sex typing and androgeny in dyadic interaction: Individual differences in responsiveness to physical attractiveness. Journal of Personality and Social Psychology, 41 . 74-86. Armor, D. A., & Taylor, S. E. (1997). Mindset, prediction, and performance: The causes and consequences of optimism. Unpublished manuscript. University of California, Los Angeles . Baldwin, M. W., Keelan, J. P. R., Fehr, B., Enns, V., & Koh-Rangarajoo, E. (1996). Social-cognitive conceptualization of attachment working models: Availability and accessibility effects. Journal of Personality and Social Psychology. 71 , 94-109. Bandura, A. (1982). The psychology of chance encounters and life paths. American Psychologist. 37 . 747-755. Baucom, D. H., Epstein, N., Rankin, L. A., & Burnett, C. K. (1996). Assessing relationship standards: The Inventory of Specific Relationship Standards. Journal of Family Psychology. 10 . 72-88. Baumeister, R. F. (1989). The optimal margin of illusion. Journal of Social and Clinical Psychology. 8 . 176-189. Booth, A., Johnson, D. R., White, L. K., & Edwards, J. N. (1985). Predictmg divorce and permanent separation. Journal of Family Issues. 6 . 331-346. Bowlby, J. (1969). Attachment and loss: Volume I. Attachment . New York: Basic Books. Box, G. E. P., & Tiao, G. C. (1973). Bayesian inference in statistical analysis . Reading, MA: AddisonWesley.

PAGE 100

94 Bradbury, T. N., & Fincham, F. D. (1990). Attributions in marriage: Review and critique. Psychological Bulletin, 107 . 3-33. Bradbury, T. N., & Kamey, B. R. (1993). Longitudinal study of marital interaction and dysfunction: Review and analysis. Clinical Psychology Review, 13 , 1527. Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models: Applications and data analysis methods . Newbury Park: Sage. Bryk, A. S., Raudenbush, S. W., & Congdon, R. T. (1994). HLM: Hierarchical linear modelling with the HLM/2L and HLM/3L programs . Chicago: Scientific Software International. Camelley, K. B., & Janoff-Bulman, R. (1992). Optimism about love relationships: General vs specific lessons from one's personal experiences. Journal of Social and Personal Relationships, 9 . 5-20. Chaiken, S. (1987). The heuristic model of persuasion. In M. P. Zanna & J. M. Olson &. C. P. Herman (Eds.), Social Influence: The Ontario Symposium (Vol. 5, pp. 339). Hillsdale, NJ: Erlbaum. Cohen, J., & Cohen, P. (1983). A pplied multiple regression: Correlational analysis for the behavioral sciences (2nd Edition) . Hillsdale, NJ: Erlbaum. Collins, N. L., & Allard, L. M. (2001). Cognitive representations of attachment: the content and fimctioning of working models. In G. J. O. Fletcher & M. S. Clark (Eds.), Blackwell Handbook of Social Psycho logv. Vol 2: Interpersonal Processes (pp. 60-85). Oxford: Blackwell Publishers Limited. D'Angelo, C, & Kamey, B. R. (1999). A comparison between sampling methods in marital research. Paper presented at the Southeastern Psychological Association, Savannah, GA. Darley, J. M., & Fazio, R. H. (1980). Expectancy confirmation processes arising in the social interaction sequence. American Psychologist, 35 . 867-881. Ditto, P. H., & Hilton, J. L. (1990). Expectancy processes in the health care interaction sequence. Journal of Social Issues. 46 (2). 97-124. Downey, G., Freitas, A. L., Michaelis, B., & Khouri, H. (1998). The self-fulfilling prophecy in close relationships: Rejection sensitivity and rejection by romantic partners. Journal of Personality and Social Psychology, 75 . 545-560.

PAGE 101

95 Drigotas, S. M., Rusbult, C. E., Wieselquist, J., & Whitton, S. W. (1999). Close partner as sculptor of the ideal self: Behavioral affirmation and the Michelangelo phenomenon. Journal of Personality and Social Psychology, 77 , 293-323. Duncan, S. L. (1976). Differential social perception and attribution of intergroup violence: Testing the lower limits of stereotyping blacks. Journal of Personality and Social Psychology, 34 , 590-598. Dunning, D. (1995). Trait importance and modifiability as factors influencing self-assessment and self-enhancement motives. Personality & Social Psychology Bulletin. 21 . 1297-1306. Dunning, D., & McElwee, R. O. (1995). Idiosyncratic trait definitions: Implications for self-description and social judgment. Journal of Personality and Social Psychology. 68 . 936-946. Dunning, D., Meyerow^itz, J. A., & Holzberg, A. D. (1989). Ambiguity and selfevaluation: The role of idiosyncratic trait definitions in self-serving assessments of ability. Journal of Personality and Social Psychology, 57 , 1082-1090. Eidelson, R. J., & Epstein, N. (1982). Cognition and relationship maladjustment: Development of a measure of dysfiinctional relationship beliefs. Journal of Consulting and Clinical Psychology. 50 , 715-720. Epstein, N., & Eidelson, R. J. (1981). Unrealistic beliefs of clinical couples: Their relationship to expectations, goals, and satisfaction. American Journal of Family Therapy. 9 . 13-22. Fincham, F. D. (1985). Attribution processes in distressed and non-distressed couples: 2. Responsibility for marital problems. Journal of Abnormal Psychology, 94 , 183-190. Fincham, F. D., & Bradbury, T. N. (1987). The assessment of marital quality: A reevaluation. Journal of Marriage and the Family, 49 , 797-809. Fincham, F. D., & Bradbury, T. N. (1992). Assessing attributions in marriage: The Relationship Attribution Measure. Journal of Personality and Social Psychology, 62 , 457-468. Fincham, F. D., Harold, G. T., & Gano-Phillips, S. (2000). The longitudinal association between attributions and marital satisfaction: direction of effects and the role of efficacy expectations. Journal of Family Psychology, 14 , 267-285.

PAGE 102

96 Fiske, S. T., & Neuberg, S. L. (1990). A continuum of impression formation, from category-based to individuating processes: Influences of formation and motivation on attention and interpretation. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 23, pp. 1-74). New York, NY: Academic Press. Geiss, S. K., & O'Leary, K. D. (1981). Therapist ratings of frequency and severity of marital problems. Journal of Marital and Family Therapy, 7 , 515-520. Gilbert, D. T., & Hkon, J. G. (1991). The trouble of thinking:Activation and application of stereotypic beliefs. Journal of Personality and Social Psychology. 60 . 509517. Hampson, S. E., Goldberg, L. R., & John, O. P. (1987). Category-breadth and social-desirability values for 573 personality terms. European Journal of Personality. 1 . 241-258. Hampson, S. E., John, O. P., & Goldberg, L. R. (1986). Category breadth and hierarchical structure in personality: Studies of asymmetries in judgments of trait implications. Journal of Personality and Social Psychology. 51 . 37-54. Harris, M. J., & Rosenthal, R. (1985). Mediation of interpersonal expectancy effects: 31 meta-analyses. Psychological Bulletin. 97 (3). 363-386. Hasorf, A., & Cantril, H. (1954). They saw a game: A case study. Journal of Applied Social Psychology. 49 . 129-134. Hazan, C, & Shaver, P. R, (1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology. 52 , 51 1-524. Henderson, V. L., & Dweck, C. S. (1990). Motivation and achievement. In S. S. Feldman & G. R. Elliot (Eds.), At the threshold: The developing adolescent (pp. 308329). Cambridge: MA: Harvard University Press. Higgins, E. T., &, Bargh, J. A. (1987). Social cognition and social perception. Annual Review of Psychology. 38 . 369-425. Holtzworth-Munroe, A., & Jacobson, N. S. (1985). Causal attributions of married couples: When do they search for causes? What do they conclude when they do? Journal of Personality and Social Psychology. 48 . 1398-1412. Huston, T. L., & Houts, R. M. (1998). The psychological infrastructure of courtship and marriage: The role fo personality and compatibility in romantic relationships. In T. N. Bradbury (Ed.), The developmental course of marital dysfunction. (pp. 114-151). New York, NY: Cambridge University Press.

PAGE 103

97 Huston, T. L., & Vangelisti, A., L. (1991). Socioemotional behavior and satisfaction in marital relationships: A longitudinal study. Journal of Personality and Social Psychology. 61 . 721-733. Jacobson, N. S., McDonald, D. W., FoUette, W. C, & Berley, R. A. (1985). Attribution processes in distressed and nondistressed married couples. Cognitive Therapy and Research. 9 . 33-50. James, W. (1890). The principles of psychology (Vol. 1). New York: Holt. Jones, M. E., & Stanton, A. L. (1988). Dysfimctional beHefs, belief similarity, and marital distress: a comparison of models. Journal of Social and Clinincal Psychology, 2(1), 1-14. Jussim, L., & Eccles, J. (1995). Naturally occurring interpersonal expectancies. In N. Eisenberg (Ed.), Social development: Review of personality and social psychology (Vol. 15, pp. 74-108). Thousand Oaks, CA: Sage. Kahneman, D., & Miller, D. T. (1986). Norm theory: Comparmg reality to its alternative. Psychological Review, 98 , 136-153. Karney, B. R., & Bradbury, T. N. (1995). The longitudinal course of marital quality and stability: A review of theory, method, and research. Psychological Bulletin, m 3-34. Karney, B. R., & Bradbury, T. N. (1997). Neuroticism, marital interaction, and the trajectory of marital satisfaction. Journal of Personality and Social Psychology, 72 , 1075-1092. Karney, B. R., & Bradbury, T. N. (2000). Attributions in marriage: State or trait? A growth curve analysis. Journal of Personality and Social Psychology. 78 . 295-309. Karney, B. R., Bradbury, T. N., Fincham, F. D., & Sullivan, K. T. (1994). The role of negative affectivity in the association between attributions and marital satisfaction. Journal of Personality and Social Psychology. 66 . 413-424. Karney, B. R., McNulty, J. K., & Bradbury, T. N. (2001). Cognition and the development of close relationships. In G. J. O. Fletcher & M. S. Clark (Eds.), Blackwell Handbook in Social Psychology, Vol.2: Interpersonal Processes (pp. 32-59). Oxford: Blackwell Publishers Lunited. Kelley, H. H. (1950). The warm-cold variable in first impressions of persons. Journal of Personality. 18 . 431-439.

PAGE 104

98 Knee, C. R. (1998). Implicit theories of relationships: Assessment and prediction of romantic relationship initiation, coping and longevity. Journal of Personality and Social Psychology, 74 . 360-370. Locke, H., & Williamson, R. (1958). Marital adjustment: A factor analysis study. American Sociological Review, 23 , 562-569. Markman, H. J., Stanley, S. M., & Blumberg, S. L. (1994). Fighting for vour marriage: Positive steps for preventing divorce and preserving a lasting love . San Francisco: Jossey-Bass. McNulty, J. K., & Kamey, B. R. (2001). Attributions in marriage: Integrating specific and global evaluations of a relationship. Personality and Social Psychology Bulletin. 27 . 943-955. McNulty, J. K., & Kamey, B. R. (under review). Expectancy confirmation in appraisals of marital interactions. Medvec, V. H., Madey, S. F., & Gilovich. (1995). When less or more: Counterfactual thinking and satisfaction among Olympic medalists. Journal of Personality and Social Psvchologv. 69 . 603-610. Mellers, B. A., Schwartz, A., Ho, K., & Ritov, I. (1997). Decision affect theory: Emotional reactions to the outcomes of risky options. Psychological Science. 8 . 423-429. Merton, R. K. (1948). The self-fulfilling prophecy. Antioch Review. 8 . 193-210. Miller, D. T., &, Tumbull, W. (1986). Expectancies and interpersonal processes. Annual Review of Psvchologv. 37 . 233-256. Murray, S. L., Hohnes, J. G., & Griffin, D. W. (1996). The self-fulfilling nature of positive illusions in romantic relationships: Love is not blind, but prescient. Journal of Personalitv and Social Psvchologv. 71 . 1155-1180. Neff, L. A., & Kamey, B. R. (in press). Self-evaluation motives in close relationships: A model of global enhancement and specific verification. In P. Noller & J. A. Fenney (Eds.), The intricacies of marital interaction . New York: Plenum. Neuberg, S. L. (1989). The goal of forming accurate impressions during social interactions: Attenuating the impact of negative expectancies. Joumal of Personality and Social Psvcholog v. 56. 374-386. Norton, A. J., & Glick, P. C. (1979). Marital instability in America: Past, present, and future. In G. Levinger & O. C. Moles (Eds.), Divorce and separation: Context, causes, and consequences (pp. 6-19). New York: Basic Books, Inc.

PAGE 105

Norton, R. (1983). Measuring marital quality: A critical look at the dependent variable. Journal of Marriage and the Family, 45 , 141-151. Olson, D. H., Foumier, D. G., & Druckman, J. M. (1986). Counselors manual for PREPARE-ENRICH (rev, ed.l Minneapolis, MN: PREP ARE-ENRICH, Inc. Olson, J. M., Roese, N. J., & Zarma, M. P. (1998). Expectancies. In E. T. Higgins & A. W. Kruglanski (Eds.), Social Psychology: Handbook of Basic Principles (pp. 211238). New York, NY: Guilford Press. Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957). The measurement of meaning . Urbana: University of Illinois Press. Peale, N. V. (1952). The povyer of positive thinking . New York, NY: PrenticeHall, Inc. Pratto, F., & Bargh, J. A. (1991). Stereotyping based on apparently individuating information: Trait and global components of sex stereotypes under attention overload. Journal of Experimental Social Psychology, 27 . 26-47. Raudenbush, S. W., Brennan, R. T., & Bamett, R. C. (1995). A multivariate hierarchical model for studying psychological change within married couples. Journal of Family Psychology, 9 . 161-174. Rosenhan, D. L. (1973). On being sane in insane places. Science, 179 , 250-258. Rosenthal, R., &. Jacobson, L. (1968). Pygmalion in the classroom: Teacher expectation and pupils' intellectual development . New York: Holt, Rinehart, and Winston. Shepperd, J. A., Findley-Klein, C, Kwavnick, K. D., Walker, D., & Perez, S. (2000). Bracing for loss. Journal of Personality and Social Psychology, 78 (4), 620-634. Shepperd, J. A., & McNulty, J. K. (2000). Why do people brace?. Unpublished Data, University of Florida . Shepperd, J. A., & McNulty, J. K. (in press). The affective consequences of expected and unexpected outcomes. Psychological Science . Shepperd, J. A., Ouellette, J. A., & Fernandez, J. K. (1996). Abandoning unrealistic optimism: Performance estimates and the temporal proximity of self-relevant feedback. Journal of Personality and Social Psychology, 70 . 844-855. Sillars, A. L., Coletti, S. F., Parry, D., & Rogers, M. A. (1982). Coding verbal conflict tactics: Nonverbal and perceptual correlates of the "avoidance-distributiveintegrative" distinction. Human Communication Research. 9 . 83-95.

PAGE 106

100 Snyder, M. (1984). When belief creates reality. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 18, pp. 248-305). Orlando, FL: Academic. Snyder, M., Tanke, E. D., & Berscheid, E. (1977). Social perception and interpersonal behavior: On the self-fulfilling nature of social stereotypes. Journal of Personality and Social Psychology, 35 , 656-666. South, S. J., & Spitze, G. (1986). Determinants of divorce over the marital Hfe course. American Sociological Review, 51 , 583-590. Stanley, S. M., Blumberg, S. L., & Markman, H. J. (Eds.). (1999). Helping couples fight for their marriages: the PREP approach . Philadelphia, PA: Brunner/Mazel. Talbot, M. (1997). Love, American style. The New Republic. 216. 30-38. Taylor, K. M., & Shepperd, J. A. (1998). Bracing for the worst: Severity, testing, and feedback as moderators of the optimistic bias. Personality and Social Psychology Bulletin. 24 .915-926. Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social psychological perspective on mental health. Psychological Bulletm. 103 , 193-210. Veroflf, J., Kulka, R. A., & Douvan, E. (1981). Mental heahh in America: Patterns of helpseeking fi-om 1957 to 1976 . New York: Basic. Weiss, R. L. (1980). Strategic behavioral marital therapy: Toward a model for assessment and intervention. In J. P. Vincent (Ed.), Advances in family intervention, assessment, and theory (Vol. 1, pp. 229-271). Greenwich, CT: JAI Press. Word, C. O., Zanna, M. P., & Cooper, J. (1974). The nonverbal mediation of selffulfilling prophecies in interracial interaction. Journal of Experimental Social Psychology. 10 . 109-120.

PAGE 107

BIOGRAPHICAL SKETCH Jim McNulty entered the University of Florida as a freshman in the fall of 1993. While attending the university, he majored in psychology and developed a strong interest in the field of social psychology. After graduating Phi Beta Kappa in the spring of 1997, he began his graduate work in the doctoral program in social psychology at the University of Florida and developed a program of research examining the development of marital satisfaction in newlywed couples under the supervision of Dr. Benjamin Karney. Jim McNulty received the degree of Master of Science in May of 1999. His master's research focussed on the distinction between global and specific evaluations about the relationship. Jim was granted the degree of Doctor of Philosophy from the University of Florida on August 1 1 , 2001 . His dissertation focussed on the longitudinal effects of positive expectations in relationships. 101

PAGE 108

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy, ff? vj "^J Benjamin Kamey, Chairman Assistant Professor of Psycho lo^ I certify that I have read this study and that in my opinioij,it conforms to acceptable standards of scholarly presentation and is fially adequate, msie^Q and qjja^yf^s_a. dissertation for the degree of Doctor of Philosophy. /Barry Schlenker^ Professor of Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fiilly adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. les Shepperd ociate Professor of Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosoph y. /^^^ Mark Fondacaro Assistant Professor of Criminology and Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Master of Science. ICS Algina fessor of Educatitmal Psychology This dissertation was submitted to the Graduate Faculty of the Department of Psychology in the College of Liberal Arts and Sciences and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August 2001 Dean, Graduate School