| The living arrangements of elderly people in the United States in the early twentieth century are studied with a new data source, the Public Use Sample of the 1910 U.S. Census. The living arrangements of old people in 1910 do not match the stereotypes of extended family households; most households contained no more than one conjugal couple. Unlike today, however, most elderly people lived with at least one child. This research tests hypotheses which are suggested by demographic, economic, and cultural explanations for change in the household structure of the elderly over time. Demographic estimation techniques are used to measure the availability of kin with whom elderly people could reside. Log linear models are estimated to test the importance of economic resources, numbers of children, and race in determining the living arrangements of elderly widows. Numbers of children, among elderly women, and race, among elderly men, are shown to significantly affect the probabilities of living alone and living with children in 1910. Comparing models from 1910 and 1980 suggests that the effect of income on the probability of living alone has changed over the course of the twentieth century in the United States; however, this comparison also suggests that the whole set of relationships which determine living arrangements, including values about privacy, has been altered and that it is impossible to identify one factor as the primary agent of change. |
