Problems of children living in homeless shelters may diminish over time. Buckner, Bassuk, Weinreb, and Brooks (1999) found in a cross-sectional analysis that children’s psychiatric symptoms peaked after about four months in shelter; thereafter, children seemed to adapt to the shelter environment. One to two years after living in shelter, the initial effects of homelessness on children's internalizing problems that were noted by Buckner and colleagues (1999) had diminished (Buckner, Beardslee, & Bassuk, 2004). Longitudinally, they observed exposure to violence, which can impact homeless and housed children alike, to be a more potent predictor of children's mental health problems over time than homelessness. And in a longitudinal study, Shinn and colleagues (2008) found only modest differences across all domains between housed poor children and homeless children who remained with their families five years after shelter entry, by which time most families were re-housed. Differences between groups were more strongly related to current stressors than to past homelessness. Using a city’s Board of Education records for children in the same sample, Rafferty and colleagues (2004) found that the about-to-be homeless children did not differ from continuously housed children in achievement before the homeless children entered shelter. Rather, the homeless children’s performance dropped while they were in shelter and partially recovered after they were re-housed, so that they no longer differed significantly from the other poor children. Over the five years following shelter entry, the homeless youth had more school mobility and greater grade retention (repeating a grade) than poor, continually housed youth.
The eight sites of the Homeless Families Program funded by the Substance Abuse and Mental Health Services Administration (SAMHSA) enrolled over 1,100 children ages 2–16 years while they were living in shelter and followed them (and their mothers) for 15 months. Families were eligible if mothers had a mental illness or substance problem. Children's behavior problems were measured repeatedly over four time points. A fairly similar pattern was observed in which behavior problems were initially elevated when children were assessed in a shelter environment but then moderated in severity over time as families became re-housed (Buckner, Weinreb, Rog, Holupka, & Samuels, in press). Five subgroups of children were identified by these investigators. The trajectory of behavior problems over time looked virtually identical across these groups (some initial elevation followed by improvement). What was different was the initial level of behavior problems these different groups had (ranging from virtually none to severe). Such findings are consistent with the Huntington, Buckner, and Bassuk (2008) finding that homeless children are not one homogenous group. Finally, Shinn, Samuels, and Fischer (under review) randomized homeless families in which the mother had a mental illness or substance problem to a Family Critical Time Intervention (described below) or to usual care, and followed them over time. The most dramatic finding was an improvement in children’s mental health and school outcomes in both experimental and control groups from 3 months to 24 months following shelter entry. Improvements were observed across all age groups from preschool to adolescence and in both mother and child reports.