Tracking homeless youth over time can suggest both causes of and possibles olutions to the problems they experience. Unfortunately, only a few such studies have been done to date (e.g., Cauce et al., 1994), and few of their findings have yet been published in peer reviewed journals. In part, this paucity of data reflects the fact that data in these longitudinal studies are still being collected. However, it also indicates a general lack of research on homeless youth (as compared to other homeless groups) and a particular lack of longitudinal research on this population.
In any event, we can draw some conclusions based on preliminary results from a study by Toro and his colleagues. A probability sample of 249 homeless youth from throughout the Detroit metropolitan area, plus a matched sample of 149 housed youth, were initially interviewed at ages 13 to 17 and have been followed since at six time-points over a seven-year period. The youth are now aged 20 to 24, and data collection is nearly complete. Most of the adolescents returned fairly quickly to their family of origin. Nearly all (93 percent) of the initially homeless adolescents in the sample were no longer homeless at the 4.5-year follow-up, with one-third living with their parents (33 percent), another third living on their own (34 percent), and still others living with friends or relatives (21 percent). At follow-up, the initially homeless adolescents also reported significantly less conflict with their family and fewer stressful events (Toro & Janisse, 2004). Such trends have also been observed in longitudinal studies of homeless adults (e.g., Toro et al., 1999). People who are sampled because they are currently homeless are often at a particularly low point in their lives. Over time, many exit homelessness and thus appear to function at least somewhat better at follow-up, even though they often are not fully part of “mainstream society” and are likely to be at risk for future homelessness and/or other poor life outcomes.
Ahmed and Toro (2004) used data from the same longitudinal study to examine the relationship between several dimensions of religiosity and substance abuse outcomes over an 18-month follow-up period. Both cross-sectional and longitudinal analyses found that religiosity “buffered” the potentially harmful impact of stress on the outcomes. At the 4.5- year follow-up, greater spirituality protected African American, but not European American, young adults exposed to high levels of community violence or alcohol and drug abuse (Fowler, Ahmed et al., 2006).
Roy and her colleagues in Montreal have been studying various samples of street youth (age 14 to 25) and have followed one sample to observe various health outcomes, including HIV infection (Roy et al., 2003). They have, for example, found high mortality, with an annual death rate of 1 percent (Roy et al., 2004). The most common cause of death, by far, was suicide. This research group is now conducting another longitudinal study with more general purposes. Longitudinal findings will begin to be available in late 2007.
Milburn and her colleagues have followed homeless youth, aged 12 to 20, in Los Angeles (N=498) and in Melbourne, Australia (N=398), over a 12-month period (see Milburn, Rotheram-Borus et al., 2006; Witken et al., 2005). The longitudinal findings are just beginning to be reported in the professional literature (e.g., Milburn, Ayala, et al., 2006; Rosenthal et al., 2007).