In&odudion lessness6,9; other studies have linked lifetime victimization to similar conditions.26 In this paper we examine the degree to which psychiatric and substance problems and victimization place poor, NewYork City families at elevated risk of requiring emergency housing, and we document the prevalence of such problems among poor families who are housed and those who are homeless. Much of the previous research in this area has focused on documenting the prevalence of mental problems among homeless persons. Although in several of these studies relatively high rates of disorder were found among the people in the homeless samples, these rates were not sufficiently high to support the idea that psychiatric and substance-abuse problems are the principal causes of homelessness.'12 Moreover, there is reason to believe that psychiatric and substance-abuse problems are less of an issue for the growing numbers ofhomeless families than they are for single adults.2-5 For example, rates of past psychiatric hospitalization for homeless women with children have ranged from 8 to 14%,6-9 whereas hospitalization rates for samples comprised primarily of single homeless adults have ranged from 10 to 35%.7,10-20 With some notable exceptions,21-23 many of these earlier studies of homeless people have been limited by a lack of comparison groups, an inability to employ random sampling, and a use of cross-sectional samples of currently homeless people rather than of new entrants into homelessness.Although the main focuses of our study included the roles ofpsychiatric history and substance abuse in homelessness, we also considered victimization