In this paper, we review problems of epidemiologic method in surveys that reported on mental illness in homeless adults in the United States in the 1980s (Arce, et al., 1983; Vergare & Arce, 1986; Bassuk, et al., 1984; Bassuk, et ah, 1986; Crystal, et ah, 1986; Fischer, et ah, 1986; Koegel, et al., 1988; Kroll, et ah, 1986; Morse & Calsyn, 1986; Rossi, et al., 1987; Roth & Bean, 1986; Snow, et al., 1986; Struening, 1986; Susser, et al., 1989a). These surveys can be construed, in epidemiologic terms, as ~’case surveys” of homelessness, since most gathered data on homeless persons without selecting a second group specifically for comparison. Surveys included in our review meet two minimal criteria:
Journal