“…Consequently, many analyses of this subject rely upon census data which are constrained by the number of living arrangement features covered and by the fact that the data are generally based on household heads. Where original data on living arrangements are collected, analyses tend to focus on either specific settings such as age-segregated housing (McAuley and Offerle, 1981;Devlin, 1980;Smith and Lipman, 1972), retirement villages (Jacobs, 1975;Martin, 1973), or particular groups of the elderly such as widows (Fengler and Danigelis, 1982) women (Hunter, et al, 1979), the poor (Winiecke, 1973), or professionals (Murikami and Pynoos, 1981). A few studies have considered a comparison of elderly persons with various housing circumstances and background characteristics, but these are often restricted to non-probability samples in selected housing situations (e.g.…”