The Empirical Classification of Formal Groups

predicting differential or involvement with this role. 5. The unimportant differences found within the upper, middle, and lower SES categories suggest that finer SES discriminations are of little value in studies of childrearing attitudes. However, further work involving comparisons of high, moderate, and low status-aspiration groups within each major SES category still seems desirable; refinements such as those suggested by degree of status congruence or crystallization or identification should also help to break down the gross concept of social class frequently employed in research on child rearing. 6. Class differences in the type and frequency of child punishment used by parents have been reasonably well established in earlier studies. Our finding that class differences in rejection of certain aspects of the maternal role also exist suggests that such rejection may be associated with frequency and extent of child punishment. This proposition could be tested by comparing the distribution of these factors in a single sample of mothers. Greater rejection of the role should lead to greater punishment of children, but the opposite might be predicted for rejecting mothers who remain keenly aware of the loving mother ideal and are thus unable to accept their own feeling of resentment toward child-rearing tasks. 7. The implications of the attitude patterns found for the family as an institution and for the individuals involved deserve to be explored further in both family and stratification research.'9