It is estimated that physical violence may occur in one half of all American marriages, with often devastating consequences, including murder, suicide, psychological difficulties, and physical injuries. Despite the magnitude of this health problem little empirical data have been gathered on the characteristics of abusive relationships or on the development of spouse abuse. In order to obtain such information, 240 couples will be assessed, half recruited from a random sample of the community and the other half form a hospital providing therapy. Equal numbers of couples will be selected for each of three groups: abusive couples; nonabusive but maritally distressed couples and nonabusive, nondistressed couples. In addition to these three cells, we will also be examining couples who are abusive but report nondistressed relationships. Although we are skeptical about the existence of such couples, recent studies have argued that such couples do exist. In order to understand how abusive couples might nevertheless define their relationships as happy ones, we are going to identify such couples and study them. Thus, all in all we have a 2x2 design, defined by the factors of abuse status (abusive or nonabusive) and marital satisfaction level (currently distressed or nondistressed). The major goal of the study is to examine cognitive, physiological, and interaction al variables as possible antecedents and consequences of spouse abuse. Among these variables, primary interest will be focused on: Communication during a marital interaction task; casual attributions offered by victims for their partner's abusive behavior and by perpetrators for their own abusive behavior; and autonomic nervous system arousal (as well as self-reported affective responding) of subjects during marital interactions. Violent couples will be compared to nonviolent couples on a number of dimensions. The dimensions will also be used to predict the later occurrence of abuse. Finally, changes in the pattern (e.g., self-esteem, attitudes towards women, violence in family of origin, depression, psychopathology) will be examined longitudinally, to tease out the temporal relationship between these variables and marital violence.
