Content area
Abstract
Past research on well-being in old age has not produced consistent results regarding the direction of gender differences. Well-being is a multidimensional concept and differences in its definition partially account for the seemingly conflicting findings. This thesis provides a framework for assessing the divergent findings and applies that framework to data from a 1986 survey of the elderly on Java, Indonesia.
Six dimensions of well-being are identified: social, activity, authority, economic, physical, and psychological. Logistic regression is used to measure the gross and net effects of gender on the multiple indicators of each dimension of well-being.
In the bivariate relationships, older women have lower levels of activity, authority, and economic well-being and older men have lower levels of psychological well-being (or life satisfaction). Gender differences in social and physical well-being are not marked. Controlling for possible confounding and mediating factors (age, marital status, household structure, skills) tends to reduce the gender effect, but the same general patterns persist. Persons who are older, unmarried, living alone, and/or unskilled tend to have lower levels of well-being. Women are more likely than men to be members of each of these groups.
Investigation of possible relationships among the dimensions of well-being suggests that: (a) women's lower levels of activity help explain their lower levels of authority and economic well-being; (b) gender differences in current authority status contribute to differences in individual economic resources; (c)women's lesser ability to accumulate economic resources has contributed to their lower current level of authority; and (d) gender differences in related dimensions cannot help explain gender differences in life satisfaction.
The findings reported in the literature review of well-being on Java coupled with the results from the data analysis suggest that gender continues to be an important criterion for social differentiation in old age on Java. The results are consistent with the argument that the gender differences observed in old age likely reflect the continuation of patterns established in younger adulthood.
