Dimensions of homelessness.

Investigations of homelessness have been hampered by the lack of operational definitions sensitive enough to achieve subgroup differentiation and simple enough to permit replication. As a consequence, programming and policy development have often proceeded based on varying assessments of the composition, size, and needs of the homeless population. This paper describes the empirical use of duration of homelessness and dwelling place as elements of an operational definition of homelessness. The approach reflects a conceptualization of homelessness as a continuous variable that can be described by coordinates of time and place. A screening instrument that quantified the homeless experience was developed and evaluated in conjunction with a federally funded demonstration project for homeless substance-abusing men and women. Eight hundred and thirty-nine men and women from six public detoxification centers were screened over a two-year period that began in August 1988. Respondents were asked eight questions to assess duration (time) and location (place) of homelessness before they entered the detoxification center. A simple index was constructed retrospectively and found to differentiate the sample into homeless and near-homeless subgroups. Between-group differences were statistically significant in demographics, presenting problems, and probability for successful intervention. These data paralleled previously reported differences between homeless subgroups and support the concurrent validity of the index. Cronbach's alpha (.72) showed the index to be moderately reliable. Differentiation of homeless persons into meaningful subgroups appears possible and programmatically recommended. Homelessness is not a unitary phenomenon, and it is unlikely to respond to therapeutic interventions that fail to consider individual differences.