Family Correlates of Parent-Child Attitude Congruence: Is it Time to Throw in the Towel?

Abstract

Connell (1972), in an exhaustive review of decades of research on parent-child similarity on political and social attitudes, found that, when methodologically weak studies were eliminated, parent-child correlations were uniformly low. Given this situation, he suggested that the inconclusive nature of research on family factors influencing parent-child attirude similarity (Sigel, 1970) may be due to the fact that the family has not been an important source of attitude-socialization. Therefore, it would not be surprising that family variables fail to predict parent-child attirude congruence. Data collected from a random sample of 66 college students (32 males and 34 females) and their families gave an opportunity to explore this hypothesis. Both parents and students were given (separately) a questionnaire containing 16 standardized instruments measuring political, social, and psychological attitudes (including Economic Conservatism, Wotk Security, Machiavellianism, Value Achievement, Sexual Permissiveness, etc.). Openended and structured questions were utilized to constluct 7 indices of family emotional climate and interaction: permissiveness, parental warmth, parental understanding, family conflict, quality of communication, frequency of communication, family solidarity; several of these measures have been used with other samples (cf. Thomas, 1971). Scores for 4 parent-child dyads (mo-son, fa-son, mo-daugh, fa-daugh) were analyzed separately, giving 448 univariate correlations (7 predictor variables x 16 measures of parent-child attitude similarity X 4 dyadic groups). Only 66 of the 448 correlations were significantly greater than zero (14% for parent-son dyads and 15% for parentdaughter dyads). Even more striking was the fact that the significant correlations were more or less randomly distributed across the variables and were on the whole quite low (none exceeded .50, and most were in the 'teens and twenties). As a measure of over-all family influence, the 7 family variables were utilized in a multivariate analysis with each of the dependent variables. Only 6 of the 64 coefficients (9%) were significantly greater than zero. As with the univariate analysis, the pattern of significant correlations appeared to be completely random. In no case were more than 2 of the 16 multiple correlations significant for any parent-child dyad, and the particular attitudes significantly usociated with family variables differed for each parent-child dyad. The present study is limited by the fact that the family variables, although utilized in previous research, were not standardized and possess only face validity. It is obviously premature to call for throwing in the towel in the search for family predictors of parentchild attitude congruence. But it appears that at this time the burden of proof falls to those who claim such an association exists. The present study gives no basis for rejecting