Problems of Methods in Parent-Child Research

Even most charitably, research in parent-child relations cannot be viewed as a field in which methodology is exemplary and in which evidence is firm and consistent. But even most critically or despairingly, this field cannot be dismissed as unimportant in behavioral or developmental theory. Despite or because of these facts, how parents bring up their children and how parental characteristics are infused into child personality are questions that continue to inspire research. The motivations for this paper grow out of the feeling that critical and positive stocktaking can contribute toward better research on socialization processes. Two kinds of reviews need to go hand in hand. One is an evaluation of substantive findings on relations between maternal behavior and child characteristics, an accounting of relations that have stood up with impressive consistency. Unfortunately, there has been no recent general review of the extensive body of data on childrearing and associated child characteristics, the kind of review that is concerned with the extent to which relations have been replicated and discrepant reports reconciled. A second kind of review is needed which concerns itself with the methods by which data on childrearing are obtained and the nature of inferences about these data. This discussion is concerned primarily with the latter issue-how research on socialization is done. Its starting point is the tremendous discrepancy which exists between the sophistication of theory and the sophistication of methods as they meet in research on parent-child relations. Hypotheses from psychoanalytic, learning, and cognitive theories are abundant and provocative regarding parental influences upon the child, but their fruits are in great measure lost on findings that rest on extremely shaky sources of evidence.