In this article attention is paid to longevity and some predictors of longevity in 455 respondents in a 12-year longitudinal research that started in 1974 on relocation of older people. Participants in this research project stemmed from five groups, which can be distinguished by the degree of diminished autonomy of (desired) housing facility. The more autonomous the housing facility, the lower the mortality rate appeared to be. In this article a new measure for the respondents' longevity is presented, the 'overlevingsscore' (OS; 'survival score'). The people living autonomously in 1974 appeared to become a little older than indicated by their life expectancy in 1974. Those who lived in an institution at that time, however, lived shorter. Next, we tried to predict individual OSs by means of some demographic and psychosocial variables. A positive subjective health and an engaged attitude towards life in old age appear to indicate a long life. Frequency of relocation after 1974 did not influence the OS.