Introduction

Social scientists have underscored the role of the family in the origins of child behavior problems, suggesting the importance of lack of parental involvement or overt neglect (McCord, 1991), inadequate supervision (Patterson, DeBaryshe, & Ramsey, 1989), parental rejection or hostility (Rohner, 1980), aversive discipline such as nagging or scolding (Patterson, 1986), and parent–child separation (Bowlby, 1944). A strict, punitive style, nagging and scolding, or erratic behavior have been pinpointed as probably the most relevant aspects, particularly in studies of aggression, delinquency, and other externalizing behavior problems (Burt, 1925; Loeber & Stouthamer-Loeber, 1986; McCord, 1988; Patterson, 1986).

While it is compelling to hypothesize that a child's experiences within the family are relevant to the development of delinquency and antisocial behavior, there is increasing interest in a broader focus, incorporating the contextual framework in which a family is placed (Melton, 1992; Sampson, 1992; Tonry, Ohlin, & Farrington, 1991). To study the influence of community variation on family characteristics, samples need to be drawn from a number of different communities and in sufficient numbers that there is individual variation within and between communities. However, there are methodological issues in studying large samples.

Much of the key research on the effect of coercive interactions within the family has been derived from relatively small-scale investigations using detailed observations of parent–child interactions (Dowdney Skuse, Rutter, & Mrazek, 1985; Patterson, 1982).