Reliable and valid longitudinal residential histories are needed to assess interventions to reduce homelessness and increase community tenure. This study examined the test-retest reliability, sensitivity to change, and concurrent validity of the Residential Time-Line Follow-Back (TLFB) Inventory, a method used to record residential histories in the Collaborative Program to Prevent Homelessness (n ϭ 1,381). The Residential TLFB Inventory yielded temporally stable aggregate measures of duration in residential categories, and it revealed significant differences in change over time when contrasting study groups. A comparison of agency and participant data at one site
A R T I C L EThis research was supported by the Center for Mental Health Services (CMHS) and Center for Substance Abuse Treatment (CSAT) of the Substance Abuse and Mental Health Services Administration (SAMHSA) (contract number 280-94-0008). This work is a result of the collaborative effort of the participants of the project's Residential Measures Subcommittee, whose contributions we gratefully acknowledge. The contents of this paper are solely the responsibility of the authors and do not necessarily represent the views of the funding agency. For the SPSS syntax to compute the variables described in this paper, please contact Francine Williams at Policy Research Associates, Inc., Delmar, NY 12054 (phone: 518-439-7415, e-mail: fwilliams@prainc.com Increasing rates of homelessness attest to the continual challenges faced by service providers (National Coalition for the Homeless, 2001; U.S. Conference of Mayors, 2001). Individuals with psychiatric disabilities-an estimated 20-25% of the single adult homeless population, 30-70% of whom have substance use disorders-present agencies with complex needs (Fischer & Breakey, 1991;Koegel, Sullivan, Burnham, Morton, & Wenzel, 1999;Rahav & Link, 1995). Due to the cyclic nature of long-term homelessness often experienced by this dually diagnosed population, longitudinal research is necessary to investigate patterns of homelessness and residential instability and the services that are best suited to address their needs (Hopper, Jost, Hay, Welber, & Haugland, 1997). To this end, it is necessary to obtain reliable and valid data regarding histories of residential stability and homelessness in order to assess the effectiveness of any intervention directed at reducing homelessness.Studies on homelessness have differed, unfortunately, in their methods of measuring the key variables of homelessness and housing stability. Definitions of homelessness and housing stability have varied widely (Abdul Hamid, Wykes, & Stansfeld, 1993;Morrison, 1989), and the methodologies used to obtain data, as well as the samples studied, have been inconsistent (Newman, 2001). Studies of homelessness primarily use self-report and/or observer report to assess participants' homelessness or residential histories. Although several studies use institutional records or residential databases to supplement individual reports, such resources are often no...