1.2.1 Distal Factor: Socioeconomic Status and Health

The increasing number of literatures shows clearly that socioeconomic status（SES）and health are strongly related, in both industrialized and developing countries, in both welfare states and liberal democracies. The issue of whether SES affects health or vice versa has been controversial. Two theories were extracted from these disputes: “social causation”and“health selection”. The social causation theory claims that health is related to socially determined structural factors such as SES. The health selection theory suggests that SES is affected by health, and that the healthy people move up the class hierarchy while the less healthy people move down. Given that the research population is composed of elderly adults, this study uses the social causation theory;that is, SES impacts an individual's health. A social gradient in health can be identified in both western countries, and eastern countries:people with high SES are more likely to have better health as assessed by self-rated health（SRH）, functional status, or mortality, and regardless of whether SES is measured by levels of income, years of education or occupational class.

In addition, SES has accumulative effects, which means that socio-economic differences in health escalate with an increase in age. However, several studies identified that SES differences in health expand through late middle-age and decline thereafter. Declining health ine qualities in later life have been attributed to selective mortality, social sector services targeting older adults, and cohort effects. However, some studies did not draw a consistent conclusion. Lamper suggested that small socioeconomic differences in functional aspects of health up to the age of 90 years, were followed by significant differences in those aged 90 years and over in a research population aged 70 years and over. A Germany study has shown that socioeconomic differences were significant among those aged 70～79 and disappeared after 80 years of age. Another Germany study revealed that only a slight age variation existed in the association between SES and health among individuals aged 60 and above. In Japan, there is limited knowledge about how the effects of SES on mortality interact with age and gender. Liang and his colleagues pointed out that there is no significant educational difference of mortality among the 70～79 age group. In contrast, a Mexican study analyzed by Smith and Goldman using a nationally representative sample of older adults, indicated no significant age variation in the effects of education and wealth on SRH and physical functioning.

Gender differences also emerged in the relationship between SES and health. In Japan a cross-sectional study was carried out among 9650 participants aged between 47 and 77 to identify gender differences of the impacts of income on health. Males with a low household income were more likely to report poor or fair health but not females. Another Japanese study, conducted by Liang et al. in 2002, found an opposite association to western countries with an educational crossover observed among elderly men. This association may be due to gender and SES differences in the causes of death, morbidity, and health behavior. Fukuda and his colleagues found that the relationship between mortality and SES（including income and education）was stronger in men than in women. In line with this gender difference, Smith and Goldman also claimed an SES-related difference in health was smaller in older women than men. While Bassuk, Berkman and Amick recognized education, household income and occupational prestige were generally associated with lower mortality for men, this was true only for women regarding income among elderly residents in four US communities（East Boston, Massachusetts；New Haven, Connecticut; east-central Iowa; and the Piedmont region of North Carolina）. Prus and Gee believed that the relationship between income and health is only significant in older women aged≥65 years, based on data from the 1994 to 1995 National Population Health Survey in Canada.

Compared with studies in western countries, the research on the relationship between SES and health is very limited in Asian countries, let alone among elderly people; furthermore, little is known about how the SES-health link differed by age and gender. Therefore, consistent results have not yet to be drawn.