Fauber and Long's (1991) overview of research on family therapy with children is a valuable integration of the literatures on the family correlates of and treatments for childhood disorders. Several concerns apply to some of the inferences they draw from risk research, however. Their assertion that various sources of family distress have effects that are mediated primarily through parenting is questionable, as is their suggestion that parenting therefore is the appropriate focus of family treatment. The conceptual issues of reductionism, linearity, holism, and change in defining causality are discussed in questioning these conclusions about etiology and treatment. Other empirical and methodological issues are raised briefly, particularly as they relate to statistical models of direct and indirect influences and to the body of correlational and analogue research on how parental conflict influences children.