The pattern of age differences in voluntary association memberships found in previous studies is consistently curvilinear But association membership levels are related to income and education and older persons have lower levels of income and education than younger persons. Data from the 1972 CPS American National Election Study (N =2,597) and the 1974 NORC General Social Survey (N = 1,352) are analyzed by Multiple Classification Analysis in order to examine the pattern of age differences in voluntary association memberships after removing the effects of income and education to see whether the curvilinear pattern can be explained by socioeconomic variation among different age strata. The resulting patterns show increasing levels of membership through the age range 35–44 and then either generally stable or increasing levels through the age range 75 +. Similar findings are in evidence for both males and females, although males have higher association membership levels at all ages. The characteristically lower association membership levels found among the older age strata, therefore, appear to be attributable to the manner in which their socioeconomic characteristics differ from other age strata.