Dianne T. Thompson
U.S. Department of Justice

FULL TEXT: Adobe Acrobat (*.pdf, 395 KB)

This study uses new methods and approaches that augment findings from previously completed research on length of stay in assisted housing programs. This study differs from previous research in six areas. First, most previous research used mean and median calculations for a single program year; the present study, however, evaluates data from an 8-year period of time—1995 to 2002. Second, due largely to data limitations, the prior research has generally focused on currently assisted households that continue to receive housing assistance. This new research includes data for households that have exited the programs (former households) to gain a broader perspective on housing tenure. Third, this study identifies multiple program participants, or mixed households, that moved between public housing and housing voucher programs across the eight-year study period. Using an 8-year data file allows for this type of identification. Fourth, this study separately identifies the length of stay for participants with very short durations (less than six months). The existence of this group might reflect an administrative data collection problem, or it may suggest some other phenomenon among assisted housing recipients worthy of further investigation. Fifth, this study systematically identifies data gaps, logical inconsistencies, and out-of-range data in the file using a data quality process that goes beyond what has been done in past work. Last, and perhaps most importantly, this study presents tenure estimates for assisted households based on median survival time that may be more realistic than calculations that rely solely on mean and median summary statistics. Estimates based on the life-table method produces statistics, including the median survival time, that account for the current cases in making predictions about housing tenure.