ABSTRACTThis article examines some of the assumptions that underlie the use of aggregate data for policy analysis, In particular,itexplores the consequences of the assumption of antagonistic attitudes between “haves” and “have nots” on social welfare policies. Use of public opinion data demonstrates the low salience of regressiveness of taxation and thereby calls into question the validity of a common indice in aggregate analysis. The use of this and related examples is intended to demonstrate the value of supplementing aggregate data analysis of public policy with direct evidence about public attitudes toward those p