Predicting Change in Marital Satisfaction from Husbands' and Wives' Conflict Resolution Styles

Abstract

Identifying what specific conflict resolution styles are linked to change in marital satisfaction is important because managing conflict is one of the central tasks of maintaining a marriage (Gottman, 1994) and because declines in marital satisfaction herald a series of processes indicative of a deteriorating marriage (Gottman & Levenson, 1992). Findings regarding the concurrent link between conflict resolution styles and marital satisfaction have been consistent in indicating that each spouse's marital satisfaction is positively related to the frequency with which each spouse uses constructive strategies to resolve conflict (such as agreement, compromise, and humor) and negatively related both to the frequency with which each spouse uses destructive strategies to resolve conflict (such as conflict engagement, withdrawal, and defensiveness) and to the joint frequency with which the wife uses conflict engagement and the husband uses withdrawal (the demand-withdraw pattern). This pattern of findings has been obtained regardless of whether conflict resolution styles were assessed by means of brief behavioral observations (Gottman & Krokoff, 1989; Heavey, Layne, & Christensen, 1993; Noller, Feeney, Bonnell, & Callan, 1994) or by means of self-report and partner-report questionnaire data (Christensen, 1988; Heavey et al., 1993; Huston & Vangelisti, 1991; Noller et al., 1994). Given that the concurrent link between conflict resolution styles and marital satisfaction is well-established, researchers have used longitudinal data to assess the plausibility of two causal relations involving conflict resolution styles and marital satisfaction. The first causal relation--the assumption that the use of certain conflict resolution styles causes marital satisfaction--is based on interdependence theory (Rusbult, 1983) which posits that perceived rewards to a relationship (such as the frequent use of constructive conflict resolution strategies) and perceived costs to the relationship (such as the frequent experience of negative conflict resolution styles) determine satisfaction with the relationship. The longitudinal data consistent with this causal relation would indicate that the frequency with which certain conflict resolution styles are used at Time 1 predicts change in marital satisfaction. The second causal relation--the assumption that the level of marital satisfaction is causally related to the frequency with which certain conflict resolution styles are used--is based on self-fulfilling prophecy theory (e.g., Snyder, Tanke, & Berscheid, 1977) which posits that one's attitude (e.g., level of satisfaction with the marriage) provides a psychological environment that elicits behavior (e.g., conflict resolution styles) that reinforces and is consistent with the initial attitude. The longitudinal data consistent with this causal relation would indicate that the level of marital satisfaction at Time 1 predicts the degree of change in the frequency with which certain conflict resolution styles are used. Unfortunately, longitudinal findings relevant to the plausibility of either causal relation have been inconsistent (Gottman & Krokoff, 1989; Heavey et al., 1993; Huston & Vangelisti, 1991; Noller et al., 1994). In part, this could be due to four methodological and two conceptual limitations of these studies. Regarding the methodological limitations, first, because Gottman and Krokoff (1989) and Heavey et al. (1993) used measures of marital satisfaction that also tapped frequency of conflict and disagreement, the correlations from these measures may have been inflated (see review by Fincham & Bradbury, 1987). Second, only Huston and Vangelisti (1991) reported that marital satisfaction scores changed appreciably over the time interval studied, raising the possibility that, in the other studies, a restriction of range accounted for nonsignificant findings. Third, because the first assessments made by Huston and Vangelisti (1991) and Noller et al. …