Quantitative Article assignment.
chady5
Article1.docxAuthors:
Schoebi, Dominik, ORCID 0000-0003-3991-2712 . Department of Psychology, University of Fribourg, Fribourg, Switzerland, [email protected] Karney, Benjamin R.. Department of Psychology, University of California, Los Angeles, CA, US Bradbury, Thomas N.. Department of Psychology, University of California, Los Angeles, CA, US
Address:
Schoebi, Dominik, Department of Psychology, University of Fribourg, Rue de Faucigny 2, 1700, Fribourg, Switzerland, [email protected]
Source:
Journal of Personality and Social Psychology, Vol 102(4), Apr, 2012. pp. 729-742.
NLM Title Abbreviation:
J Pers Soc Psychol
Publisher:
US : American Psychological Association
ISSN:
0022-3514 (Print) 1939-1315 (Electronic)
Language:
English
Keywords:
commitment, maintenance, marital interaction, relationships, satisfaction
Abstract:
Although commitment is theoretically distinct from relationship satisfaction, empirical associations between the concepts are high. After drawing from classic definitions of commitment to distinguish between commitment as the desire for a relationship to persist versus the behavioral inclination to maintain the relationship, we predicted that the former component would function much like satisfaction, whereas the latter component would operate independently of satisfaction to stabilize couple relationships. Using satisfaction and commitment data collected over the first 4 years of marriage (N = 172 couples), we demonstrate that only behavioral inclinations to maintain the marriage are related to observed marital interaction behaviors, to reported steps taken toward dissolution, and to 11-year divorce rates, independent of satisfaction. Consistent with dyadic 'weak-link'' conceptions of commitment, likelihood of divorce was found to increase as a function of the lower of the 2 partners' inclination to maintain the relationship. Commitment may stabilize declining intimate partnerships, particularly when it is conceptualized as the inclination to maintain the relationship. (PsycINFO Database Record (c) 2016 APA, all rights reserved)
Document Type:
Journal Article
Subjects:
*Commitment; *Interpersonal Interaction; *Marriage; *Relationship Satisfaction
Medical Subject Headings (MeSH):
Adult; Divorce; Family Conflict; Female; Follow-Up Studies; Humans; Interpersonal Relations; Male; Marriage; Models, Psychological; Personal Satisfaction; Probability; Problem Solving; Surveys and Questionnaires; Time Factors
PsycINFO Classification:
Marriage & Family (2950)
Population:
Human Male Female
Location:
US
Age Group:
Adulthood (18 yrs & older)
Tests & Measures:
Quality of Marriage Inventory Marital Status Inventory DOI: 10.1037/t17909-000
Grant Sponsorship:
Sponsor: National Institute of Mental Health Grant Number: Grant MH48674 Recipients: Bradbury, Thomas N. Sponsor: Swiss National Science Foundation, Switzerland Grant Number: Grants PA001-10899 and PZ00P1_121616 Recipients: Schoebi, Dominik
Methodology:
Empirical Study; Quantitative Study
Format Covered:
Electronic
Publication Type:
Journal; Peer Reviewed Journal
Publication History:
First Posted: Nov 21, 2011; Accepted: Oct 14, 2011; Revised: Oct 11, 2011; First Submitted: Feb 1, 2011
Release Date:
20111121
Correction Date:
20140414
Copyright:
American Psychological Association. 2011
Digital Object Identifier:
http://dx.doi.org.library.capella.edu/10.1037/a0026290
PMID:
22103577
PsycARTICLES Identifier:
psp-102-4-729
Accession Number:
2011-26987-001
Number of Citations in Source:
90
Stability and Change in the First 10 Years of Marriage: Does Commitment Confer Benefits Beyond the Effects of Satisfaction?
Contents
1. The Concept of Commitment: Strengths and Limitations
2. Possible Advantages to Distinguishing Between Forms of Commitment
3. Does Commitment Stabilize Marriages? Evidence From Longitudinal Studies
5. Method
6. Participants
7. Procedure
8. Measures
10. Results
12. How Are Fluctuations in Commitment Related to Changes in Relationship Satisfaction?
13. Are the Dimensions of Commitment Differentially Associated With Problem-Solving Behavior?
14. Does Commitment Predict Relationship Stability?
15. Discussion
16. Footnotes
17. References
Full Text
By: Dominik Schoebi Department of Psychology, University of Fribourg, Fribourg, Switzerland; Benjamin R. Karney Department of Psychology, University of California, Los Angeles Thomas N. Bradbury Department of Psychology, University of California, Los Angeles
Acknowledgement: This work was supported by National Institute of Mental Health Grant MH48674 to Thomas N. Bradbury. Dominik Schoebi's work was supported by Swiss National Science Foundation Grants PA001-10899 and PZ00P1_121616.
Arguably the simplest explanation for why a marriage dissolves is that one or both spouses become increasingly dissatisfied, diminishing the quality of couple interaction and prompting a separation or divorce in turn. Meta-analytic findings confirm the link between relationship distress and dissolution, but the magnitude of the association is modest (r ~ .3–.4; Karney & Bradbury, 1995), in part because many unhappy couples remain married. The concept of commitment is often invoked to explain the persistence of these unhappy marriages, under the assumption that relatively committed partners are motivated to continue their relationship for reasons other than their immediate emotional appraisals of the partnership (e.g., Kelley, 1983). A rich literature sheds light on the stabilizing role of commitment in dating relationships (e.g., Arriaga & Agnew, 2001), but the possibility that commitment operates differently in longer term partnerships has led some to call specifically for research on commitment in marriage and on the marital maintenance behaviors that commitment might motivate (e.g., Finkel, Rusbult, Kumashiro, & Hannon, 2002). This article responds to this call, using observational and 11-year longitudinal data from newlywed spouses to address two questions that are largely unaddressed in the marital commitment literature: First, does commitment stabilize marriage? Second, does commitment motivate interpersonal behaviors? Because lowered relationship satisfaction provides a more parsimonious explanation for why spouses neglect relationship maintenance and contemplate divorce, we tested whether any effects of commitment on relationship outcomes and processes remain after controlling for spouses' relationship satisfaction judgments.
The Concept of Commitment: Strengths and Limitations
In an early analysis of marital stability, Levinger (1965, 1976) proposed that the power of commitment derives from its ability to explain why two comparably satisfied or unsatisfied relationships might evolve in different ways: With otherwise equal forces of attraction in material, symbolic, and affectional domains, partners with more barriers to dissolution will be more committed and therefore will be more motivated to communicate better and to invest more energy in getting the relationship back on track when satisfaction dips, thereby experiencing a reduced risk of relationship dissolution (Berscheid, 1998; Lewis & Spanier, 1979). To the extent that the intense, passionate emotions enjoyed by partners early in their relationship give way to companionate love and perhaps disenchantment as a marriage unfolds (Aron, Fisher, & Strong, 2006), understanding how forces other than attractions and satisfaction operate to stabilize intimate relationships emerges as a crucial task.
The broadly influential investment model was developed by Rusbult and colleagues (Rusbult, 1980, 1983) specifically to address the “unjustified persistence” problem, or why people remain involved in relationships that are not particularly satisfying. Building upon Levinger's ideas and the principles of interdependence theory (Thibaut & Kelley, 1959), the investment model proposes that an individual depends upon a relationship—that is, an individual comes to need or rely upon a relationship as a source of desired experiences and outcomes—to the extent that he or she derives satisfaction from the relationship, has relatively few alternatives to the relationship, and has invested many resources in the relationship. With growing dependence comes a sense of commitment to the partner and to the relationship, according to the investment model, where commitment is understood to be “the sense of allegiance that is established to the source of one's dependence” (Rusbult, Coolsen, Kirchner, & Clarke, 2006, p. 618). Note, however, that the investment model characterizes commitment as more than mere dependence upon a relationship, encompassing as well the special bond or attachment that one forms to the partner and the partnership. Dependence and commitment do go hand in hand in the investment model, but whereas dependence is theorized to be a structural property of dyads, commitment is viewed as a psychological state that captures more directly the experience of being in the relationship and being joined to another person within it (Rusbult et al., 2006). In turn, commitment is hypothesized to have two related but crucial downstream effects, first directly influencing how an individual behaves toward his or her partner and second influencing decisions to persist in the relationship. Conceptualized in this way, commitment is assumed to play a pivotal role in relationships, not only reflecting the dependence that arises between two people but also mediating links between the three bases that combine to determine dependence, on the one hand, and key actions likely to promote or curtail relationship development on the other (e.g., Rusbult, Wieselquist, Foster, & Witcher, 1999). The unjustified persistence problem is therefore resolved by the investment model, in that commitment is rooted in experiences and judgments that extend beyond mere satisfaction; fluctuations in satisfaction therefore can occur without necessarily threatening relationship maintenance behaviors or relationship stability.
This theoretical representation of commitment, particularly the notion that commitment can be understood as the sense of allegiance one feels toward a unique and valued intimate bond, affords the concept a reasonable degree of precision as well as substantial explanatory power in relationship science. At the same time, the investment model provides more specificity with regard to hypothesized antecedents of dependence and commitment than to the means by which commitment affects interpersonal processes and decisions to persist. Although commitment level is assumed to influence a range of relationship maintenance mechanisms (e.g., accommodative behavior, willingness to sacrifice, derogation of alternative partners), the specific nature of these links is acknowledged to be underdeveloped, such that “future research should seek to further explicate the precise processes by which interdependent partners achieve long-term, well functioning relationships” (Rusbult et al., 1999, p. 446). This is an especially important challenge in that partners' perceptions of pro-relationship maintenance behaviors are assumed to provide the foundation for trust and deepening dependence, a process that has been described as mutual cyclical growth (Wieselquist, Rusbult, Foster, & Agnew, 1999).
A central premise of the current research is that while the breadth of commitment as a concept within the investment model is an obvious asset, its heterogeneity as a concept must be addressed before it can enable fruitful hypotheses linking commitment as a psychological state to particular forms of interpersonal transactions. To the extent that commitment is understood as allegiance to one's partner and relationship, we believe it is important to note that allegiance (and related concepts such as loyalty, faithfulness, and adherence) does not necessarily carry with it any behavioral efforts toward relationship maintenance. (Indeed, many people legitimately claim true allegiance to a democratic nation, to public broadcasting, to a religious faith, or even to an intimate partner without necessarily engaging in actions such as voting, donating money, praying, or apologizing for misdeeds.) We therefore propose that the concept of commitment can be profitably refined to include one element reflecting a desire for the relationship to persist and a related but distinct element reflecting the inclination to maintain the relationship. Early conceptualizations of the investment model alluded to this sort of distinction—Rusbult (1983, p. 102), for example, observed that “the definition of commitment includes two categories of definition advanced by other authors: behavioral intent and psychological attachment”—but these two elements have been subsumed and conflated in most conceptualizations of commitment. Formalizing a distinction between commitment as an inclination to engage in maintenance behaviors (which we refer to here as IM) and commitment as a longer term orientation (the desire for persistence, or DP) may be particularly important for understanding stability in marriage, as it combines a future-oriented goal or desire with the means by which that goal might be achieved. Three more specific implications of such a distinction are outlined next.
Possible Advantages to Distinguishing Between Forms of Commitment
One possible advantage to distinguishing IM and DP elements of commitment is that doing so may help clarify recent findings that directly contradict investment model predictions, notably, a lack of association between commitment and marital maintenance behaviors. For example, Tran and Simpson (2009) used Rusbult's five-item Commitment Scale to examine commitment in relation to 74 married couples' observed constructive and destructive behaviors during a problem-solving interaction. This research design had the advantage of combining self-report and observational measurements, thus eliminating concerns of shared method variance. Although relatively committed partners would be expected on theoretical grounds to engage in more effective communication, Tran and Simpson found that commitment no longer predicted observed behaviors after controlling for relationship satisfaction. This might be interpreted as a strong challenge to the hypothesis that commitment contributes to relationship maintenance efforts, but distinguishing between IM and DP elements of commitment suggests an alternate view. Specifically, conceptualizing one dimension of commitment as an inclination to engage in behavioral efforts to maintain one's relationship might result in commitment predicting relationship maintenance behaviors even after controlling for relationship satisfaction (cf. Finkel et al., 2002). Conceptualizing and assessing commitment in terms of one's desire for the relationship to continue, in contrast, might overlap considerably with partners' relationship satisfaction and therefore yield redundant effects.
This observation draws attention to a second possible benefit to distinguishing between IM and DP elements of commitment, which pertains to vexing measurement concerns. As noted earlier, the conceptual punch gained from the concept of commitment comes from it being distinct from relationship satisfaction. The investment model allows for covariation between satisfaction and commitment—for example, one would expect individuals to be more committed to relationships that are more fulfilling, all else being equal—yet there must be more to commitment than satisfaction if commitment is to operate by stabilizing a relationship that is not as satisfying as it once was. Unfortunately, although commitment is readily distinguished from satisfaction at a theoretical level (and even at an intuitive level: e.g., there are happy spouses who are uncommitted to their relationship and committed spouses who are unhappy with their relationship), measures of the two constructs correlate very highly. A meta-analysis of more than 50 studies and nearly 12,000 subjects found a correlation of .68 between satisfaction and commitment, with some correlations as high as .90 (Le & Agnew, 2003; also see Le, Dove, Agnew, Korn, & Mutso, 2010). In the face of such correlations, predicting relationship stability from measures of satisfaction and commitment provides limited information about their relative and unique contributions in the underlying model; until this overlap is resolved, the more parsimonious interpretation is that commitment and satisfaction measures tap a single underlying concept.
Compounding this concern is the heterogeneous item content used in measures of commitment. With the widely used five-item Commitment Scale (Rusbult, 1983), participants are asked questions about their expectations for the future of the relationship (“How likely is it you will end your relationship in the near future?”) and about their preferred duration of the relationship (“For what length of time would you like your relationship to last?”). People might want to end their relationship not because they are uncommitted to it but rather because they find it unfulfilling, because the partner has treated them badly, or because they have learned that prior efforts at relationship maintenance have failed. As these alternative interpretations reflect global sentiments toward the relationship, inclusion of these items might account for at least part of the strong observed overlap between commitment and satisfaction. A later seven-item measure of commitment with revised item content again yielded very high correlations with satisfaction (Rusbult, Martz, & Agnew, 1998), suggesting that this problem extends beyond one particular instrument. Here, we adopt the view that testing investment model predictions regarding the relative contributions of satisfaction and commitment to relationship stability is likely to benefit from distinguishing between IM and DP elements of commitment.
There is a third possible advantage to distinguishing between IM and DP elements of commitment, in that doing so might not only add precision to the investment model but also rectify limitations in another leading theoretical perspective on relationship maintenance and development. Two key predictions in the investment model—that commitment instigates relationship maintenance behaviors and that satisfaction plays no role beyond commitment in determining whether a relationship continues or ends—are particularly important predictions because they are at odds with predictions derived from social-learning conceptions of marriage. Studies conducted in the social-learning tradition focus almost exclusively on marital interaction as an independent variable rather than as a dependent variable, thus neglecting crucial questions about what factors aside from relationship satisfaction motivate ordinary relationship behaviors. Two seminal works in this tradition, those by Gottman (1979) and by Jacobson and Margolin (1979), never mention commitment, leaving this model largely unable to explain why spouses in unfulfilling marriages maintain their relationship and why those marriages might persist. The investment model does invoke commitment explicitly on these two points but offers the somewhat surprising prediction that relationship satisfaction does not predict relationship stability beyond any effects of commitment. In contrast, the social-learning conception makes no reference to commitment but is explicit in proposing that when marital interactions become less rewarding and more punitive and as coercive cycles engulf couples' positive feelings for one another, spouses become dissatisfied with the relationship and are likely to end the relationship as a consequence (Jacobson & Margolin, 1979). Thus, satisfaction should not predict relationship stability from the perspective of the investment model (after controlling for commitment), whereas satisfaction (and associated interpersonal processes) should predict relationship stability reliably from the perspective of social-learning-based analyses of marriage. Of course, to the extent that commitment and satisfaction are redundant, it is not possible to test the relative value of these perspectives; however, by distinguishing IM from DP, one might clarify whether the IM component of commitment predicts relationship processes and relationship stability independent of relationship satisfaction.
A unique contribution of the IM component to relationship functioning and development beyond satisfaction would provide valuable insight into the stabilizing effects of commitment. A stabilizing effect of the IM component could deepen an understanding of the behavioral processes that prevent erosion of a relationship in times of declining satisfaction levels. Building on such a view, an important focus of analysis is therefore the necessity of both partners to hold a minimum level of IM as a resource that can be mobilized to prevent negative behaviors and habits from further eroding the relationship. Either partner dropping low in IM, however, incurs considerable risk for the relationship. According to such a perspective, the critical momentum would be located in the lower of the two partners' IM or, in other words, the weaker link in the relationship (see also Attridge, Berscheid, & Simpson, 1995). Adopting such a dyadic perspective on how commitment acts on relationships becomes particularly important in the investigation of dyadic outcomes such as relationship stability.
Does Commitment Stabilize Marriages? Evidence From Longitudinal Studies
To date, empirical clarification of any unique contribution that commitment makes to stabilizing marriages has been hindered not only by the measurement issues just noted but also by the small number and relatively short duration of longitudinal studies addressing this problem. Using residualized-change analyses, one study of 65 newlywed married couples failed to find any association between commitment and relationship well-being 2 years later, perhaps because of the high degree of stability in relationship functioning (Drigotas, Rusbult, & Verette, 1999). Further analysis of this same sample failed to show that commitment predicted change in self-reported relationship maintenance behavior, “presumably because of insufficient change over time in the criteria” (Wieselquist et al., 1999, p. 956; also see Van Lange et al., 1997, Study 6). Commitment did contribute to the prediction of relationship duration in the 15-year Boston Couples Study (Bui, Peplau, & Hill, 1996), though in this case commitment was measured with items tapping the estimated likelihood of marrying the partner and items from Rubin's (1970) Love Scale, making it difficult to know whether commitment or relationship satisfaction was the driving force.
One longitudinal study has managed to circumvent the problem of high stability in relationship outcomes while also assessing specific dimensions of commitment (Stanley, Whitton, Sadberry, Clements, & Markman, 2006; but see Arriaga & Agnew, 2001, for a study on dating couples). Conducted with 38 couples in their 3rd year of marriage, this study demonstrated that two indices of commitment—one reflecting the desire for the relationship to continue and the other reflecting a willingness to maintain the relationship by making sacrifices—predicted residualized changes in husbands' satisfaction from the 3rd to the 5th year of marriage; for wives, only the “willingness to sacrifice” items predicted changes in satisfaction (Stanley et al., 2006). Husbands' willingness to sacrifice also mediated the association between their global desire for the relationship to continue and their changes in satisfaction. Although this sample is small and the couples participated in an educational workshop as newlyweds, it (a) provides valuable preliminary support for the idea that commitment can stabilize relationships, particularly when couples are studied through a period when their relationships are changing, and (b) underscores the value of distinguishing between commitment as a desire for the relationship to continue versus commitment as an inclination to engage in behaviors that will maintain the relationship. This study also draws attention to the importance of sacrifice as a form of relationship maintenance that might be characteristic of relatively committed individuals. Scholars working within the investment model tradition have suggested that the willingness to sacrifice is particularly influential as a relationship maintenance strategy because maintenance efforts perceived as incurring some cost for a partner are more likely to be attributed by the other partner to pro-relationship motives, thereby engendering trust and deepening dependence (e.g., Van Lange et al., 1997).
Thus, when conducted over sufficient spans of time, longitudinal studies permit analysis of changes in relationship satisfaction and how those changes might be tied to indices of commitment. Longitudinal studies are also valuable because they can provide information about relationship dissolution, which is presumed to be a key consequence of low commitment according to the investment model (e.g., Rusbult, 1983). Shifting from satisfaction (an individual-level variable) to dissolution (a characteristic of the dyad) is likely to also require a shift in the level at which commitment is conceptualized (Thibaut & Kelley, 1959). Specifically, as we noted earlier, commitment may predict dissolution not simply when it is low in an absolute sense but when it is low relative to that of the partner. This dyadic-level approach to commitment, first formulated by Waller and Hill (1951) as the principle of least interest, argues that the fate of a relationship should depend more strongly on the partner for whom there is less at stake—also known as the weak-link partner—if the relationship were to end. Data from a 6-month study of dating relationships support this view, even after controlling for main effects of commitment (Attridge et al., 1995; also see Oriña et al., 2011), and in the present study, we aimed to provide a further test of this dyadic model using married couples examined over an 11-year period.
The fact that couples remain in unhappy relationships presents relationship scientists with an important puzzle. Testing models of how commitment operates and establishing whether commitment operates independently of relationship satisfaction have proven difficult, however, because operationalizations of commitment share much of their variance with indices of satisfaction and because shorter term longitudinal designs permit little in the way of predictable change in relationship satisfaction or relationship dissolution. The present study aimed to investigate how commitment and satisfaction combine to affect changes in marriage, using data on commitment, satisfaction, maintenance behaviors, and relationship instability collected from a sample of 172 first-time newlyweds as they negotiated the early high-risk period for relationship change. Following the promising results of Stanley et al. (2006), we used items from the Dedication Commitment subscale of the Commitment Inventory (Stanley & Markman, 1992) to differentiate between the two aspects of commitment that are evident in its definition. We refer to these two aspects as DP and IM. Three sets of hypotheses were tested.
We first examined DP and IM as predictors of changes in relationship satisfaction. Although theoretical interest in commitment typically stems from a desire to explain stay/leave decisions of unhappy spouses, changes in satisfaction merit attention as a lead indicator of eventual relationship dissolution (e.g., Karney & Bradbury, 1997). Although we expected that higher DP and higher IM would predict changes in relationship satisfaction, we also predicted that higher IM would be a significant predictor of relationship satisfaction even after controlling for the effects of DP (Hypothesis 1a). This prediction reflected our more general view that DP functions much like relationship satisfaction, whereas IM is more readily distinguished from relationship satisfaction and provides couples with a second, separate resource for stabilizing their relationship. This same view led us to predict that relationship satisfaction would predict changes in DP more than satisfaction predicts changes in IM (Hypothesis 1b).
Second, we examined DP and IM in relation to marital interaction behavior assessed in the first year of marriage. If the above reasoning was correct, between-subject variability in DP would correlate reliably with observed problem-solving behaviors displayed by spouses, and these associations would fall to nonsignificance after controlling for relationship satisfaction (Hypothesis 2a). Between-subject variability in IM would also correlate with observed problem-solving behaviors, but these associations would remain significant after controlling for satisfaction (Hypothesis 2b). This prediction aided in validating the distinction between DP and IM but was particularly important for testing whether spouses who report inclinations to maintain the relationship do in fact display the behaviors in an interactional context when negotiation and compromise are crucial to the relationship, independent of their relationship satisfaction. This prediction is central to social-psychological analyses of commitment, most of which neglect to observe dyadic interaction directly or differentiate between DP and IM. At the same time, questions about what motivates spouse behavior have long been overlooked in social-learning-based analyses of couple interaction, most of which neglect to go beyond relationship satisfaction or stable personality traits as antecedents of dyadic interaction.
The third and final goal of this article was to investigate the unique contributions of DP and IM to the steps that partners report taking to dissolve the relationship (e.g., thinking about ending the relationship, talking to a friend about ending the relationship, talking to the spouse about ending the relationship) and to whether or not they actually separated or divorced over our 11-year assessment period. The influence of commitment on actual relationship dissolution is likely mediated by cognitive and behavioral steps toward dissolution, and relationship maintenance efforts may be a critical determinant of whether early thoughts about relationship dissolution progress into more specific steps toward separation and finally into leave behaviors (for a short-term study of dating couples that addresses this issue, see VanderDrift, Agnew, & Wilson, 2009). We predicted that any association between DP trajectories and steps toward dissolution would be redundant with relationship satisfaction trajectories (Hypothesis 3a). Conversely, we predicted that IM trajectories would account for unique variance in steps toward dissolution beyond any effects associated with relationship satisfaction (Hypothesis 3b). Next, we expected that variability in DP would not be reliably associated with actual separations and dissolutions beyond relationship satisfaction, whereas IM would be related to separations and dissolutions after controlling for relationship satisfaction (Hypothesis 3c).
Unlike reported steps toward divorce, actual separation is a dyadic outcome, which prompts questions about the interpersonal or couple-level constellation of commitment variables. How might two spouses' IM or DP combine to predict separation? Following Attridge et al. (1995), we reasoned that if commitment operates by counteracting the forces eroding relationships and thereby prevents individuals from taking steps to leave their relationship, then the lower of the two partners' IM scores (i.e., the weak-link partner in the relationship) should more strongly determine the extent to which stability is undermined (Hypothesis 3d).
Participants
The sample consisted of 344 participants from 172 married couples, recruited from marriage licenses filed in Los Angeles County, California, from May 1993 until January 1994. A total of 637 (17.8%) out of 3,606 couples responded to letters inviting them to participate in a longitudinal study of newlywed marriage. All interested responding couples were screened for eligibility by telephone. Eligibility criteria were being in the first marriage; being 18 years of age or older and wives being no older than 35 years; having at least a 10th-grade education; having no children; being able to read, write, and speak English; and having no plans to leave the Los Angeles area. The first 172 couples who kept their laboratory appointment were included in the study.
At Time 1, all couples were newlyweds. On average, wives were 26.0 (SD = 3.4) years of age, and husbands were 27.6 (SD = 3.9) years of age. Participants were Caucasian (66%), Asian American–Pacific Islander (14%), Latino–Chicano (15%), and African American (5%). Couples who did not respond to invitation letters were younger, were less likely to cohabitate before marriage, had less education, and were in lower status jobs than couples who responded (for more details, see Karney et al., 1995).
Procedure
Spouses independently completed questionnaires during an initial 3-hr laboratory session, during which couples were instructed to engage in two standard 10-min discussions of self-identified marital difficulties. Spouses subsequently completed questionnaires at seven more occasions spaced across approximately 6-month intervals over the course of 4 years.
Behavioral assessments
In the lab session, spouses independently identified a source of tension in the marriage and were instructed to discuss their topics in two separate 10-min conversations, with the goal of working toward mutually satisfying solutions. The order of the two discussions was randomly predetermined. Conversations were videotaped, and couples were paid $75 after completion.
Self-report assessments
At the first and the third measurements, questionnaires were administered before and during laboratory sessions. At all other measurements, questionnaires were mailed, and couples were instructed via telephone calls and cover letters. Written and verbal instructions reminded spouses to not consult with one another when responding to the questionnaires. Couples were paid $25 upon completion of the questionnaires.
Measures
Personal commitment
Personal commitment was assessed with items from the dedication commitment scale by Stanley and Markman (1992). Each item was rated on a 7-point scale ranging from 1 = strongly disagree to 7 = strongly agree. We constructed two separate scales reflecting the two features of commitment. The DP scale included four relatively global, future-oriented items (“My marriage is clearly part of my future life plans”; “I want my marriage to stay strong no matter what rough times we may encounter”; “I may not want to be with my partner in a few years from now,” reverse-scored; “My marriage is more important to me than almost anything else in my life”). The IM scale included four items reflecting relatively immediate manifestations of commitment (e.g., “I like to think of my partner and me more in terms of 'us' and 'we' than 'me' and 'him/her'”; “Giving up something for my partner is frequently not worth the trouble,” reverse-scored; “It makes me feel good to sacrifice for my partner”; “My marriage often must take the back seat to other interests of mine,” reverse-scored). Scaled chi-square difference tests on confirmatory factor analyses modeling the interitem covariance at the between- and within-person levels showed that the model with the proposed factors fit the data significantly better than models with only one factor at the between-individual level only, χ2(1) = 9.57, p < .01, or at the within-individual level only, χ2(1) = 4.36, p < .05. We averaged responses to the items to obtain scores for the DP and IM scales. Alphas between persons at each assessment ranged between .60 and .81 (DP: Mdn = .73, IM: Mdn = .70).
Relationship satisfaction
Relationship satisfaction was assessed with the Quality of Marriage Inventory (QMI; Norton, 1983). The QMI provides global evaluation of relationship satisfaction that focuses on perceived marital quality and happiness and does not include items tapping problems or behaviors (e.g., “We have a good marriage”). The items were rated on a 7-point scale (1 = very strong disagreement, 7 = very strong agreement), with overall happiness in the marriage rated on a 10-point scale (1 = very unhappy, 10 = perfectly happy). To reduce overlap with commitment measures, we excluded two items (“My marriage is strong,” “My marriage is stable”). The final scores are thus based on four items. The scale showed good consistency at each measurement, with Cronbach's alphas ranging between .92 and .96.
Problem-solving behavior
Constructive problem-solving behavior in the 10-min problem-solving conversations was assessed by trained coders using the Kategoriensystem für Partnerschaftliche Interaktion (KPI; Hahlweg et al., 1984). The KPI uses 27 codes derived from research on communication skills training and couple therapy. Each speaking turn is coded, and the resulting codes distinguish reliably between distressed and nondistressed couples. Intraclass correlation analyses (all ps < .01) yielded adequate estimates of interobserver agreement, ranging from .62 and .75 for husbands' and wives' negative skills, respectively, to .84 and .90 for husbands' and wives' positive skills, respectively. As our interest was in examining between-person variability in behaviors that would promote relationship maintenance, we computed scores for husbands and wives reflecting the proportion of all coded behaviors that were positive or constructive in tone (e.g., positive solutions, positive feedback and expressions of acceptance, agreement and acceptance of responsibility).
Steps toward separation
Steps toward separation were assessed with items from the Marital Status Inventory (Weiss & Cerreto, 1980), which includes statements reflecting progressively less ambiguous steps toward separation (e.g., “I have occasionally thought of divorce or wished that we were separated, usually after an argument or other incident”; “I have discussed the issue of divorce seriously or at length with my spouse”). Participants indicated whether each statement was true (1) or false (0). We summed responses to seven items from this scale, excluding actual separation or divorce because the latter items reflect couple-level items rather than individual-level items.
Relationship status
Separation and divorce were assessed by tracking participants' intact versus dissolved marital status regularly at the second through eighth assessments, with additional contacts in the 9th through 11th years of marriage. With one exception, all separated couples had eventually filed for divorce; we therefore do not differentiate between separation and divorce. Thirty-seven couples (21.5% of the original 172) divorced in the 11 years of the study; six of these occurred within 2 years of marriage, and 15 occurred within 4 years of marriage.
Analytic Strategy
To take into account nonindependence between spouses' repeated measurements and between husbands' and wives' data, we used multilevel models to test all hypotheses including within-person associations. The models treated the repeated measurements as nested within couples, where husbands' and wives' parameters were estimated simultaneously with a multiple intercept approach, and husbands' and wives' residuals at the measurement level and at the couple level were allowed to correlate (cf. Kenny, Kashy, & Cook, 2006). For ease of presentation, the equations reflecting the models tested are presented including only one set of variables. Between-person associations involving commitment variables and constructive behaviors during the problem-solving discussions were tested using a multivariate regression analysis. Hypotheses concerning steps toward separation were evaluated using a model for zero-inflated count data incorporated in the Mplus software (Muthén & Muthén, 2005), and hypotheses concerning the prediction of divorce were examined using a discrete time hazard model (see Singer & Willett, 2003). For all analyses, commitment and satisfaction scores were standardized (z-scored). Predictors reflecting within-person fluctuations over time were centered at each individual's mean on the respective variable, whereas predictors reflecting between-person differences were centered at the grand mean.
Preliminary Analyses
On average, raw DP and IM scores were high, and the distributions were skewed. We therefore report on normalized (log transformed) scores. Table 1 presents descriptive statistics for all study variables, for husbands and for wives. Over the first 4 years of marriage, the commitment and satisfaction tended to be high, and partners tended to endorse relatively few steps toward ending their relationship. On average, wives scored significantly higher than husbands on IM but were more likely to take steps toward separation. Rates of initially observed constructive behaviors were relatively high and did not differ across husbands and wives.
Commitment, Relationship Satisfaction, Constructive Problem-Solving Behaviors, and Steps Toward Separation: Descriptive Statistics
We used a multivariate multilevel model to examine the covariance of DP, IM, and relationship satisfaction scores at the between-person level (across couples) and at the within-person level (within couples, over time). To this end, we set up an empty model with random variance of husbands' and wives' DP, IM, and relationship satisfaction across couples, and we examined the variance–covariance matrix at the between-person level and at the within-person level to obtain correlation coefficients. All correlations were significant (p < .01). At the between-person level, correlations between commitment components and relationship satisfaction were sizeable. The average between-person correlation of DP and IM was substantial, with r = .717 for husbands and r = .599 for wives. Husbands' and wives' commitment scores were moderately correlated (DP: r = .271, IM: r = .270). Relationship satisfaction levels covaried with DP scores (husbands: r = .717, wives: r = .676) and with IM scores (husbands: r = .650, wives: r = .463), and husbands' and wives relationship satisfaction was substantially correlated (r = .688).
At the within-person level, the correlation between DP and IM was .422 for husbands and .370 for wives. The correlation between relationship satisfaction and DP was .463 for husbands and .548 for wives, and the correlation between relationship satisfaction and IM was .325 for husbands and .336 for wives.
In sum, the measures are performing largely as expected for a sample of newlywed couples. Commitment and relationship satisfaction overlap substantially, yet the overlap between DP and IM is less than perfect. Importantly, the within-person correlations between the two commitment components and satisfaction were moderate, consistent with the assumption that IM, DP, and the QMI reflect distinguishable concepts. Below, we present the results of analyses that tested our four main sets of hypotheses.
How Are Fluctuations in Commitment Related to Changes in Relationship Satisfaction?
Predicting changes in relationship satisfaction
We predicted that fluctuations in IM would be predictive of prospective changes in relationship satisfaction and that the latter effects would remain significant even after controlling for the effects of DP (Hypothesis 1a).
To test this hypothesis, we first set up a model in which residualized increases or decreases in relationship satisfaction were predicted from the DP and IM scores. Equation 1 shows the Level 1 model for the prediction of prospective fluctuations of relationship satisfaction:
RSi+1j is the relationship satisfaction score for a participant of couple j at a subsequent measurement time i + 1. The estimate for β0j is the intercept, β1j captures the stability of relationship satisfaction between time i of the first seven measurements and time i + 1, β2j captures the unique association between the DP commitment score at time i and fluctuations in relationship satisfaction from time i to time i + 1, β3j captures the association between the IM commitment score at time i and fluctuations in relationship satisfaction, and eij is the error term.
The upper section of Table 2 shows the results obtained when predicting prospective fluctuations in relationship satisfaction from DP and IM. Within-person variability in IM covaried significantly with subsequent fluctuations in relationship satisfaction, with higher scores being associated with less decline in relationship satisfaction. For DP, a significant effect was obtained only for wives; a model comparison suggested this effect was significantly greater than the effect for husbands: χ2(1) = 4.8, p < .05. This analysis provides support for Hypothesis 1a, that the IM component of commitment would covary with relationship satisfaction even after controlling for the DP component.
Prediction of Prospective Fluctuations in Relationship Satisfaction and Commitment Components
Predicting prospective fluctuations in commitment
With Hypothesis 1b, we predicted that relationship satisfaction would predict variability in DP to a greater extent than it predicted variability in IM. The equations used to predict fluctuations in commitment from satisfaction are logically equivalent to those shown in Equation 1, above. The lower section of Table 2 shows the results of these analyses. Relationship satisfaction significantly predicted subsequent DP fluctuations for husbands and wives, with lower levels of satisfaction corresponding with declines in DP across the repeated series of seven 6-month assessment lags. A similar effect resulted for husbands', but not wives', IM fluctuations. A model comparison suggested that the effect of husbands differed significantly from that of wives, χ2(1) = 13.1, p < .001. Thus, we supported Hypothesis 1b for wives only; the failure to support this hypothesis for husbands, as reflected by the unexpectedly strong prediction of IM fluctuations by relationship satisfaction, may be rooted in the relatively lesser overlap between DP and relationship satisfaction fluctuations as compared to that of wives (see within-person correlations in the preliminary analyses).
Are the Dimensions of Commitment Differentially Associated With Problem-Solving Behavior?
Our third set of hypotheses involves associations between dimensions of commitment and problem-solving behaviors. We expected that higher levels of DP and IM, as measured at the first two assessments, would be associated with the display of more constructive behaviors during marital conflict relative to all coded behaviors. We predicted that associations between DP levels and observed behavior would become nonsignificant after controlling for relationship satisfaction (Hypothesis 2a), whereas associations between IM and observed behavior would remain significant after controlling for satisfaction (Hypothesis 2b).
To test these hypotheses, we set up a dyadic regression model with correlated residuals to predict participants' constructive problem-solving behaviors from their levels on the commitment features separately and combined, as reflected in
Equation 2:
PBj is the constructive problem-solving behavior of the participant from couple j, β0 captures the average problem-solving score, β1 captures the association between the DP average from the first two assessments and the constructive problem-solving behavior, β2 captures the association between the IM average from the first two assessments and the constructive problem-solving behavior, β3 captures the association between the average relationship satisfaction at the first two assessments and constructive problem-solving behavior, and ej is the error term.
Table 3 summarizes the results of the separate and combined analyses. The results from the separate models, displayed in the upper section of Table 3, show that higher relationship satisfaction and DP commitment were significantly associated with more constructive behaviors during the conflict discussions for husbands and wives. Moreover, wives' IM commitment was also associated with more constructive behaviors. The results of the combined models largely confirm predictions. As shown in the lower section of Table 3, when testing all predictors simultaneously, signification associations between DP commitment and constructive behaviors are no longer significant for husbands or for wives, consistent with Hypothesis 2a. Wives' IM commitment remained significantly associated with problem-solving behaviors after controlling for relationship satisfaction and DP. Indeed, wives' IM scores were the most closely associated to problem-solving behaviors of all predictors, thus supporting Hypothesis 2b. For husbands, however, constructive behaviors were predicted only by relationship satisfaction and not by IM commitment—this effect is significantly different from wives' effect: χ2(1) = 6.9, p < .01—which is inconsistent with Hypothesis 2b.
Associations Between Commitment Components and Constructive Behavior in Conflict Discussions
Although the more encompassing conclusion here is that DP, IM, and relationship satisfaction predict constructive behaviors in different ways for husbands and wives, these results do demonstrate that wives' reported inclinations to maintain the relationship covary reliably with observed behavior, independent of satisfaction. When tested alone, their more global desire to persist in the relationship also covaries with constructive behavior, but, consistent with our view that DP and satisfaction are aligned more closely than IM and relationship satisfaction, this association falls to nonsignificance in the combined model. These results further distinguish IM from DP, and although they have been obtained with cross-sectional data, they provide some support for the idea that the IM dimension of commitment, but not the DP dimension, may motivate actual interactional behaviors in a manner unconfounded by relationship satisfaction.
Does Commitment Predict Relationship Stability?
Steps toward separation
In Hypotheses 3a and 3b, we proposed that a person's variation in relationship satisfaction and in the IM dimension of personal commitment over time would be uniquely associated with steps toward separation or divorce. In contrast, we proposed that variation on the DP dimension of commitment would have no significant effect on steps toward separation beyond the effects of relationship satisfaction. The general model to test this hypothesis is reflected by Equation 3,
SSij represents the score of steps toward separation at time i of the eight measurements for participant j. The estimate for β0j represents the intercept; the estimates for β1j, β1j, and β3j capture the unique associations of relationship satisfaction, DP, and IM features of personal commitment with steps toward separation; and eij is the error term. All predictors were centered at each individual's mean, and thus, the estimates for β1j and β2j and β3j represent within-individual associations.
The distribution of the steps-to-separation scores resembled a zero-inflated Poisson distribution, with a large preponderance of cases where no steps toward separation were reported and a decreasing frequency of advanced steps toward separation. Mplus incorporates procedures for the analysis of this type of data, simultaneously analyzing whether or not a participant scored zero and the extent to which a participant reported higher or lower scores (Muthén & Muthén, 2005). The procedure yields two estimates for each variable that reflect the associations with the two distinct features of the steps-to-separation variable.
Table 4 summarizes the results of these analyses. The results displayed in the top of Table 4 suggest that at times when participants were relatively satisfied, as compared to their average satisfaction, they were less likely to take steps toward separation. Scoring higher or lower on DP at a particular time point, as compared to other time points, was not associated with the likelihood of steps toward separation. Contrary to our assumptions, when wives' IM scores were higher at a particular time point (compared to their average rating), they were more likely to take steps toward separation. This was not true for husbands, gender difference significant: χ2(1) = 6.8, p < .01.
Associations Between Commitment Components and Steps Toward Separation
The bottom section of Table 4 displays the coefficients predicting advances in steps toward separation. Relationship satisfaction was the only significant predictor for husbands, suggesting that when husbands were less satisfied, relative to their average satisfaction, they took more steps toward separation. For women, however, higher relationship satisfaction at a particular point in time did not go along with a progression toward dissolution, and this gender difference was significant, χ2(1) = 9.2, p < .01. Only fluctuations in women's IM component were significantly associated with steps toward separation, suggesting that at times when they reported higher IM than on average, they were less likely to take steps toward separation, gender difference significant: χ2(1) = 4.3, p < .05. Overall, the results provide support for Hypothesis 3a and partial support for Hypothesis 3b.
Long-term prediction of separation and divorce
To examine long-term associations between commitment scores and divorce, we ran discrete time hazard models, following Singer and Willett (2003). This longitudinal approach is based on logistic regression and examines the probability that a couple will divorce during an interval of repeated measurements, given that divorce has not yet occurred. The models thus account for the right-censored nature of the data that is due to relationship dissolution. We used a variable reflecting 1-year intervals as a time variable.
A first step in fitting the model involves specification of the probability function over time. We estimated a linear and quadratic trend over 11 years of marriage, which did not perform significantly worse than a model with a completely general specification for time, χ2(9) = 16.7, p > .05. The model suggested a curvilinear increase of likelihood of divorce (year: B = .850, Wald = 11.069, p < .001, odds ratio [OR] = 2.339; year squared: B = −.061, Wald = 7.048, p < .01, OR = .941), with a peak hazard of .068 during Year 7 and a drop to a probability of divorce of .027 at Year 11. Next, to examine whether IM, but not DP, was associated with divorce and whether the weaker of both partners' IM commitment scores would predict divorce beyond the effects of relationship satisfaction (Hypotheses 3c and 3d), we entered the commitment dimensions and relationship satisfaction as time-invariant predictors of the probability function. We examined whether spouses' early DP and IM levels, as measured during the first two assessments, prior to the occurrence of any divorces, predicted the odds of relationship dissolution. The weak-link effect for commitment was modeled by means of a dyadic variable reflecting the lower of the two partners' IM or DP scores, while adjusting for the main effects of commitment. The coefficient for this weak-link variable thus reflects the extent to which the lower of the spouses' commitment scores is associated with divorce, beyond the effects of each partner's personal commitment level. As shown in Table 5, we ran separate models, first including DP or IM as predictors (Model 1), then adding spouses' early relationship satisfaction (Model 2), before finally adding the term that reflected the weaker of the two spouses' commitment scores (the weak link; Model 3).
Results of Discrete Time Hazard Models Predicting Divorce Over 11 Years of Marriage
Table 5 shows the results for these models. No significant main effects of the commitment variables resulted for DP (men: OR = .839, women: OR = .783) or for IM (men: OR = 1.017, women: OR = .903). When taking into account relationship satisfaction (Model 2), however, wives' higher satisfaction was significantly associated with a lower likelihood of divorce (DP model: OR = .635, IM model: OR = .610). These odds ratios suggest that with each unit increase in relationship satisfaction, one can expect about a 36% (1 − .635) to 39% (1 − .610) decrease in spouses' odds of divorce. Model comparisons suggested that these coefficients differed significantly from the coefficients for men's relationship satisfaction, χ2(1) > 5.8, p < .05. Finally, results for Model 3 suggested no significant effect of the weaker DP score in the couple (OR = 1.006) but a significant effect for the IM weak link (OR = .358), indicating that the risk of divorce increases as a function of the lower of the partners' reported IM. Figure 1 illustrates the effects of IM, displaying the fitted survival function for couples in which both partners reported high IM (0.5 SD above the mean) and for couples in which the husband, the wife or both partners reported low IM (0.5 SD below the mean). This graph reflects the probability that the spouses will stay together beyond the respective year of marriage.
Figure 1. Fitted survival probability over 11 years of marriage as a function of the weak-link spouse's IM commitment. Low IM was defined as 0.5 SD below the sample mean, high IM was defined as 0.5 SD above the sample mean. Estimates are adjusted for marital satisfaction. IM = inclinations to maintain.
Although the persistence of unhappy marriages has long been conceptualized as a reflection of spouses' commitment to the relationship, a high degree of empirical overlap between measures of commitment and relationship satisfaction—particularly in the context of short-term longitudinal studies—has made it difficult to determine whether and how commitment might operate to stabilize distressed couples. Early views of commitment may offer a way around this problem, as they suggest that commitment can be conceived as a desire for the relationship to continue, as distinct from commitment as an inclination to engage in behaviors that would maintain the relationship. We adopted this distinction in the present study, under the assumption that it might prove advantageous for circumventing the high empirical association observed between commitment and relationship satisfaction, while also helping to reconcile theoretical incompatibilities between the investment model and social-learning conceptions of marriage. We assessed relationship satisfaction and both forms of commitment several times over the first 4 years of marriage, and we then used these data to test three sets of hypotheses about commitment and changes in satisfaction, observed marital problem-solving behavior, and 11-year changes in relationship status. All of these hypotheses shared the idea that the DP component of commitment would be largely redundant with relationship satisfaction, whereas the IM component would be a distinct element that would predict relationship-enhancing behavior, slow declines in relationship satisfaction, and reduced steps toward dissolution and actual dissolution, independent of DP.
Results are largely though not entirely consistent with this formulation. Our first prediction was that changes in IM would predict changes in relationship satisfaction, even after controlling for the effects of DP (Hypothesis 1a). The results, shown in Table 2 (top), support Hypothesis 1a and the idea that the IM component benefits relationships even after controlling for the DP component. Thus, while the DP component is not uniquely potent in foreshadowing changes in satisfaction, when commitment is understood specifically as the inclination to maintain the relationship, then relatively committed partners are shown to experience slower declines in satisfaction compared to partners less committed to maintaining the relationship.
Hypothesis 1b extended the idea that DP would overlap more than IM with relationship satisfaction, such that relationship satisfaction was expected to predict changes in DP to a greater extent than satisfaction would predict changes in IM. Wives' results supported this hypothesis. Among husbands, however, relationship satisfaction predicted fluctuations in the DP and IM components reliably and at comparable levels (see Table 2, bottom). This result, indicating that relationship satisfaction may strengthen husbands' inclination to engage in relationship maintenance behaviors beyond the mere desire for the relationship to persist, underscores the possibility that being in a satisfactory relationship may predispose husbands toward pro-relationship behaviors in more immediate ways than those reflected by long-term goals. The possibility that relationship satisfaction not only fuels the desire to stay in the satisfactory relationship but also fosters the inclination to act in ways that benefit the relationship deserves further attention.
Hypothesis 2a predicted that higher scores on the DP component would covary with higher proportions of observed constructive behaviors during problem-solving conversations and that these associations would fall to nonsignificance after controlling for relationship satisfaction. Support for this hypothesis was strong, in that relationship satisfaction did covary with more constructive interactions, and controlling for satisfaction eliminated associations between DP and constructive behavior for husbands and for wives (see Table 3). Hypothesis 2b focused on the IM component and predicted that spouses higher in IM would display more constructive behaviors and that these associations would remain significant after controlling for satisfaction. This prediction was not supported for husbands, as their constructive behaviors were predicted by satisfaction but not by their IM component. However, among wives, when their constructive behavior was examined in relation to satisfaction and their two commitment components, only IM commitment was a reliable predictor. Thus, for wives, IM outperformed DP and satisfaction in the prediction of observed behavior. The failure of the husbands' results to support our hypothesis lies in the lack of an association between the inclination to engage in relationship maintenance behaviors and higher amounts of constructive behaviors displayed in the laboratory interactions (see separate models, Table 3), rather than in the overlap between relationship satisfaction and IM effects. The reasons for the nonsignificant IM–behavior association among husbands remain unclear. Husbands may well fail to translate IM into enacted behavior, but we cannot rule out the possibility that IM fosters maintenance behaviors apart from the relationship problem-solving behaviors sampled here (e.g., prioritizing the relationship in daily decision making).
The third and final set of hypotheses examined links between the two commitment components and two indices of relationship stability. Consistent with Hypothesis 3a, we found no evidence that DP would predict steps toward dissolution independent of relationship satisfaction, regardless of whether we focused on whether or not spouses took steps toward dissolution (see Table 4, top) or on the number of those steps (see Table 4, bottom). Results for the IM component were more complex. When predicting the dichotomous variable of whether or not spouses took any steps to dissolve the relationship, wives were more likely to do so when their scores were higher than at other times. However, at the same time, they took fewer steps overall toward dissolution when the continuous variable was predicted. This set of results provides partial support for Hypothesis 3b (see Table 4) but also raises interesting implications for an understanding of commitment in marriage (see below).
Actual dissolution was the outcome in our final set of analyses. Wives' DP scores did predict dissolution, and this association fell to nonsignificance after controlling for their satisfaction (see Table 5, top), lending partial support to Hypothesis 3c. For the IM component, neither husbands' nor wives' scores predicted dissolution after controlling for wives' satisfaction, but a weak-link interaction emerged whereby the extent to which a partner showed less inclination to maintain the relationship was the critical predictor of risk of marital dissolution (see Table 5, bottom). The comparable interaction for the DP component was not significant. We discuss the implications of these results after first reviewing some of the strengths and limitations of the present study.
Although the findings presented here benefit from the multiple waves of data collected from a relatively large and reasonably diverse sample of couples all in their first marriages and from the use of laboratory-based observation of dyadic interaction, their interpretation is tempered by several important limitations. First, this is not a nationally representative sample of couples, and so, generalizations to the population at large (including remarried couples, older couples, gay and lesbian couples, and couples entering marriage with children) must be made with considerable caution. Second, even within the more limited population of couples that we have sampled, those in our study appear to be at relatively low risk for relationship distress and dissolution. Third, the two commitment concepts emphasized here were assessed by revising an existing measure of commitment. The resulting scales were psychometrically sound, and the findings reported here lend preliminary support to their validity, but additional data are needed to strengthen claims that they capture the DP and IM concepts in a meaningful way. Fourth, although the longitudinal design used here permitted analysis of change in key concepts, the data are correlational and do not permit causal inferences.
Recognizing these limitations, these data do support the assertion that commitment contributes to relationship satisfaction and stability. This latter assertion is prominent in the investment model of course (e.g., Rusbult, 1980; Rusbult et al., 2006), and the present study strengthens this claim by demonstrating that associations between commitment measures and relationship outcomes (i.e., steps toward dissolution and actual stay/leave decisions) remain significant after controlling for relationship satisfaction. The data presented here, which are among the first to document this association, also highlight the possible value of a more specific conceptualization of commitment, in that only the IM component of commitment predicted relationship outcomes. The DP component—that is, the mere desire for the relationship to persist—was not uniquely related to either steps toward dissolution (see Table 4) or to actual relationship dissolution (see Table 5). Overall, the IM component emerged as the element of commitment that was more distinct from marital satisfaction, in terms of concurrent and lagged fluctuations, associations with interaction behaviors, and predictive value for relationship outcomes.
At the same time, the strength and clarity of the links between the IM component and relationship outcomes should not be overstated. IM did not predict husbands' steps to dissolution, and while wives did take fewer steps overall toward dissolution at times when they scored higher in the IM commitment, we also found that at these times, wives were more likely to take any steps toward dissolution when this was quantified as a dichotomous variable. This may be a spurious finding, or it may indicate that wives may be monitoring the state of their relationship more closely when their IM commitment is high and, as a result, may implement maintenance efforts in a more timely and opportune fashion. Doing so would reduce their progression of steps toward dissolution, whereas with less vigilance and lower inclinations to maintain the relationship, they may be slower to respond to problems initially in the relationship, may then find themselves confronted with greater challenges, and in response may be more likely to endorse steps toward dissolution. Thus, high levels of IM may have shorter term costs but longer term benefits, whereas low levels of IM may have shorter term benefits and longer term costs. This interpretation is speculative, but it is consistent with several recent “no pain, no gain” findings suggesting that effective relationship maintenance may require engagement of difficult topics that eventually yields benefits, including relationship stability (e.g., McNulty & Karney, 2004; McNulty, O'Mara, & Karney, 2008; Overall, Fletcher, Simpson, & Sibley, 2009).
The IM component also accounted for unique variance in a dyadic index of relationship stability, divorce or permanent separation, but again not as a simple main effect. From this set of analyses we learned, first, that after controlling for relationship satisfaction, couples in which any one spouse was below average in his or her inclinations to maintain the relationship were at elevated risk of eventual relationship dissolution (see Figure 1). A lack of relationship maintenance efforts on the part of just one partner—for example, neglecting to make sacrifices, raise difficult issues, apologize, inquire about the partner's feelings—may be sufficient to undermine the relationship and bring it to the brink of dissolution. One partner's failure to prioritize the relationship is likely to become increasingly salient to both partners and may be particularly aversive to a committed spouse, thus fueling resentment in these relationships. Figure 1 shows further that the impact of IM becomes increasingly evident even in the early years of marriage.
Examining couples' observed problem-solving behaviors in relation to DP, IM, and relationship satisfaction enabled us to compare predictions from the investment model (which assumes that commitment motivates relationship maintenance, independent of satisfaction) and social-learning conceptions of intimate relationships (which prioritize satisfaction as a correlate of observed behavior, independent of commitment). DP was unrelated to observed behavior after controlling for satisfaction, but at least for wives, IM did correlate with observed behavior, whereas satisfaction did not (see Table 3, bottom). These data are among the first to support this prediction from the investment model with married couples, and they suggest that social-learning models might benefit from recognizing not only that interpersonal processes foreshadow changes in satisfaction but that these processes are a reflection of inclinations to maintain the relationship that are separate from the rewarding or satisfying properties of relationships. Husbands' observed behavior correlated only with their satisfaction, not their IM commitment scores. This could suggest that husbands' reported IM scores bear little relation to how they actually interact, but a more conservative interpretation is that the problem-solving task used here did not provide a good sample of husbands' typical maintenance behaviors; the clear reciprocal associations between IM and satisfaction (see Table 2) supports this more cautious view.
In conclusion, commitment emerged consistently as an important predictor of changes in relationship satisfaction during the first 4 years of marriage and, after controlling for satisfaction, as an important antecedent of relationship dissolution over the first 11 years of marriage. This corroborates key predictions in models of interdependence in close relationships, but the most important contribution of this analysis is the finding that only one aspect of commitment—the inclination to maintain the relationship—accounted for variability in reported steps toward dissolution, in actual dissolution, and in the behaviors wives displayed during problem solving, independent of relationship satisfaction. Future research aimed at refining measures of this concept is likely to prove valuable, as are longitudinal and observational or diary studies that clarify the conditions under which intimate partners are more and less likely to maintain the bond they share.
1 Although data from this sample have been published elsewhere (Johnson et al., 2005; Lavner & Bradbury, 2010; Sullivan, Pasch, Johnson, & Bradbury, 2010), this is the first article from the study reporting commitment data.
2 We based the commitment and relationship satisfaction scores on the first two assessments, as the assessments closest to the behavioral observation, for two reasons. First, this strategy presumably yields more reliable estimates than using only the initial assessment, where newlyweds' variability in commitment scores was limited. Second, this strategy minimizes overlap between commitment and satisfaction scores on the one hand and actual relationship decline over the course of the 4 years on the other hand, which would spuriously inflate associations with behaviors.
Aron, A., Fisher, H. E., & Strong, G. (2006). Romantic love. In A. L.Vangelisti & D.Perlman (Eds.), The Cambridge handbook of personal relationships (pp. 595–614). New York, NY: Cambridge University Press.
Arriaga, X. B., & Agnew, C. R. (2001). Being committed: Affective, cognitive, and conative components of relationship commitment. Personality and Social Psychology Bulletin, 27, 1190–1203. doi:10.1177/0146167201279011
Attridge, M., Berscheid, E., & Simpson, J. A. (1995). Predicting relationship stability from both partners versus one. Journal of Personality and Social Psychology, 69, 254–268. doi:10.1037/0022-3514.69.2.254
Berscheid, E. (1998). A social psychological view of marital dysfunction and stability. In T. N.Bradbury (Ed.), The developmental course of marital dysfunction (pp. 441–460). New York, NY: Cambridge University Press. doi:10.1017/CBO9780511527814.019
Bui, K. T., Peplau, L. A., & Hill, C. T. (1996). Testing the Rusbult model of relationship commitment and stability in a 15-year study of heterosexual couples. Personality and Social Psychology Bulletin, 22, 1244–1257. doi:10.1177/01461672962212005
Drigotas, S. M., Rusbult, C. E., & Verette, J. (1999). Level of commitment, mutuality of commitment, and couple well-being. Personal Relationships, 6, 389–409. doi:10.1111/j.1475-6811.1999.tb00199.x
Finkel, E. J., Rusbult, C. E., Kumashiro, M., & Hannon, P. A. (2002). Dealing with betrayal in close relationships: Does commitment promote forgiveness?Journal of Personality and Social Psychology, 82, 956–974. doi:10.1037/0022-3514.82.6.956
Gottman, J. M. (1979). Marital interaction: Experimental investigations. New York, NY: Academic Press.
Hahlweg, K., Reisner, L., Kohli, G., Vollmer, M., Schindler, L., & Revenstorf, D. (1984). Development and validity of a new system to analyze interpersonal communication: Kategoriensystem für Partnerschaftliche Interaktion. In K.Hahlweg & N. S.Jacobson (Eds.), Marital interaction: Analysis and modification (pp. 182–198). New York, NY: Guilford Press.
Jacobson, N. S., & Margolin, G. (1979). Marital therapy: Strategies based on social learning and behavior exchange principles. New York, NY: Brunner/Mazel.
Johnson, M. D., Cohan, C. L., Davila, J., Lawrence, E., Rogge, R. D., Karney, B. R., & Bradbury, T. N. (2005). Problem-solving skills and affective expressions as predictors of change in marital satisfaction. Journal of Consulting and Clinical Psychology, 73, 15–27. doi:10.1037/0022-006X.73.1.15
Karney, B. R., & Bradbury, T. N. (1995). The longitudinal course of marital quality and stability: A review of theory, methods, and research. Psychological Bulletin, 118, 3–34. doi:10.1037/0033-2909.118.1.3
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. doi:10.1037/0022-3514.72.5.1075
Karney, B. R., Davila, J., Cohan, C. L., Sullivan, K. T., Johnson, M. D., & Bradbury, T. N. (1995). An empirical investigation of sampling strategies in marital research. Journal of Marriage & the Family, 57, 909–920. doi:10.2307/353411
Kelley, H. H. (1983). Love and commitment. In H. H.Kelley, E.Berscheid, C.Christensen, J. H.Harvey, T. L.Huston, G.Levinger, E.McClintock, L. A.Peplau, & D. R.Peterson (Eds.), Close relationships (pp. 265–314). New York, NY: Freeman.
Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic data analysis. New York, NY: Guilford Press.
Lavner, J. A., & Bradbury, T. N. (2010). Patterns of change in marital satisfaction over the newlywed years. Journal of Marriage & the Family, 72, 1171–1187. doi:10.1111/j.1741-3737.2010.00757.x
Le, B., & Agnew, C. R. (2003). Commitment and its theorized determinants: A meta-analysis of the investment model. Personal Relationships, 10, 37–57. doi:10.1111/1475-6811.00035
Le, B., Dove, N. L., Agnew, C. R., Korn, M. S., & Mutso, A. A. (2010). Predicting non-marital romantic relationship dissolution: A meta-analytic synthesis. Personal Relationships, 17, 377–390. doi:10.1111/j.1475-6811.2010.01285.x
Levinger, G. (1965). Marital cohesiveness and dissolution: An integrative review. Journal of Marriage & the Family, 27, 19–28. doi:10.2307/349801
Levinger, G. (1976). A social psychological perspective on marital dissolution. Journal of Social Issues, 32, 21–47. doi:10.1111/j.1540-4560.1976.tb02478.x
Lewis, R. A., & Spanier, G. B. (1979). Theorizing about the quality and stability of marriage. In W. R.Burr, R.Hill, F. I.Nye, & I. L.Reiss (Eds.), Contemporary theories about the family (pp. 268–294). New York, NY: Free Press.
McNulty, J. K., & Karney, B. R. (2004). Positive expectations in the early years of marriage: Should couples expect the best or brace for the worst?Journal of Personality and Social Psychology, 86, 729–743. doi:10.1037/0022-3514.86.5.729
McNulty, J. K., O'Mara, E. M., & Karney, B. R. (2008). Benevolent cognitions as a strategy of relationship maintenance: “Don't sweat the small stuff”…. but it is not all small stuff. Journal of Personality and Social Psychology, 94, 631–646. doi:10.1037/0022-3514.94.4.631
Muthén, B. O., & Muthén, L. K. (2005). Mplus user's guide (3rd ed.). Los Angeles, CA: Muthén & Muthén.
Norton, R. (1983). Measuring marital quality: A critical look at the dependent variable. Journal of Marriage & the Family, 45, 141–151. doi:10.2307/351302
Oriña, M. M., Collins, W. A., Simpson, J. A., Salvatore, J. E., Haydon, K. C., & Kim, J. S. (2011). Developmental and dyadic perspectives on commitment in adult romantic relationships. Psychological Science, 22, 908–915. doi:10.1177/0956797611410573
Overall, N. C., Fletcher, G. J. O., Simpson, J. A., & Sibley, C. G. (2009). Regulating partners in intimate relationships: The costs and benefits of different communication strategies. Journal of Personality and Social Psychology, 96, 620–639. doi:10.1037/a0012961
Rubin, Z. (1970). Measurement of romantic love. Journal of Personality and Social Psychology, 16, 265–273. doi:10.1037/h0029841
Rusbult, C. E. (1980). Commitment and satisfaction in romantic associations: A test of the investment model. Journal of Experimental Social Psychology, 16, 172–186. doi:10.1016/0022-1031(80)90007-4
Rusbult, C. E. (1983). A longitudinal test of the investment model: The development (and deterioration) of satisfaction and commitment in heterosexual involvements. Journal of Personality and Social Psychology, 45, 101–117. doi:10.1037/0022-3514.45.1.101
Rusbult, C. E., Coolsen, M. K., Kirchner, J. L., & Clarke, J. A. (2006). Commitment. In A. L.Vangelisti & D.Perlman (Eds.), The Cambridge handbook of personal relationships (pp. 615–635). New York, NY: Cambridge University Press.
Rusbult, C. E., Martz, J. M., & Agnew, C. R. (1998). The Investment Model Scale: Measuring commitment level, satisfaction level, quality of alternatives, and investment size. Personal Relationships, 5, 357–387. doi:10.1111/j.1475-6811.1998.tb00177.x
Rusbult, C. E., Wieselquist, J., Foster, C. A., & Witcher, B. S. (1999). Commitment and trust in close relationships: An interdependence analysis. In J. M.Adams & W. H.Jones (Eds.), Handbook of interpersonal commitment and relationship stability, (pp. 427–449). Dordrecht, the Netherlands: Kluwer Academic. doi:10.1007/978-1-4615-4773-0_25
Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis: Modeling change and event occurrence. Oxford, England: Oxford University Press.
Stanley, S. M., & Markman, H. J. (1992). Assessing commitment in personal relationships. Journal of Marriage & the Family, 54, 595–608. doi:10.2307/353245
Stanley, S. M., Whitton, S. W., Sadberry, S. L., Clements, M. L., & Markman, H. J. (2006). Sacrifice as a predictor of marital outcomes. Family Process, 45, 289–303. doi:10.1111/j.1545-5300.2006.00171.x
Sullivan, K. T., Pasch, L. A., Johnson, M. D., & Bradbury, T. N. (2010). Social support, problem solving, and the longitudinal course of newlywed marriage. Journal of Personality and Social Psychology, 98, 631–644. doi:10.1037/a0017578
Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of groups. New York, NY: Wiley.
Tran, S., & Simpson, J. A. (2009). Prorelationship maintenance behaviors: The joint roles of attachment and commitment. Journal of Personality and Social Psychology, 97, 685–698. doi:10.1037/a0016418
VanderDrift, L. E., Agnew, C. R., & Wilson, J. E. (2009). Nonmarital romantic relationship commitment and leave behavior: The mediating role of dissolution consideration. Personality and Social Psychology Bulletin, 35, 1220–1232. doi:10.1177/0146167209337543
Van Lange, P. A. M., Rusbult, C. E., Drigotas, S. M., Arriaga, X. B., Witcher, B. S., & Cox, C. L. (1997). Willingness to sacrifice in close relationships. Journal of Personality and Social Psychology, 72, 1373–1395. doi:10.1037/0022-3514.72.6.1373
Waller, W. W., & Hill, R. (1951). The family: A dynamic interpretation. New York, NY: Dryden Press.
Weiss, R. L., & Cerreto, M. C. (1980). The Marital Status Inventory: Development of a measure of dissolution potential. American Journal of Family Therapy, 8, 80–85. doi:10.1080/01926188008250358
Wieselquist, J., Rusbult, C. E., Foster, C. A., & Agnew, C. R. (1999). Commitment, pro-relationship behavior, and trust in close relationships. Journal of Personality and Social Psychology, 77, 942–966. doi:10.1037/0022-3514.77.5.942
Submitted: February 1, 2011 Revised: October 11, 2011 Accepted: October 14, 2011
This publication is protected by US and international copyright laws and its content may not be copied without the copyright holders express written permission except for the print or download capabilities of the retrieval software used for access. This content is intended solely for the use of the individual user. Source: Journal of Personality and Social Psychology. Vol. 102. (4), Apr, 2012 pp. 729-742) Accession Number: 2011-26987-001 Digital Object Identifier: 10.1037/a0026290
