The Heterogeneity of Homeless Youth in America: Examining Typologies


Report | October 5, 2011

Files: PDF | 591 KB | 12 pages

Below is an excerpt of this report. To read the full report, please download the file using the link above.


Annual prevalence estimates for homeless youth in the U.S. have ranged as high as 1.6 million among those aged 13-17 (Ringwalt et al., 1998). Robertson and Toro (1999) concluded that youth may be the single age group most at risk of becoming homeless and, yet, this group is the least studied of the three major subgroups among the overall homeless population (i.e., homeless adults, families, and youth). The existing research has documented many of the characteristics of homeless youth and identified a wide range of deficits (see Robertson and Toro, 1999; Toro, Dworsky, and Fowler, 2007). However, studies find rather different profiles of homeless youth, depending on sampling strategies, target age groups, gender balance, measures used, and other methodological factors. For example, it has been noted that studies targeting older youth (sometimes up to age 25), males, and youth from the streets tend to find more problem behaviors, such as substance abuse, mental disorders, risky sexual behavior, and conduct problems (Haber and Toro, 2004; Toro et al., 2007).

A few studies have examined the differences between homeless and housed youth. Homeless youth have less social support than their housed counterparts (Menke, 2000) and experience many hurdles and hardships while in school (Ziesemer, Marcoux, and Marwell, 1994). Furthermore, homeless youth are often victims of various forms of parental maltreatment (Wolfe, Toro, and McCaskill, 1999) and are at an increased risk for various mental disorders, including depression, conduct disorders, and substance abuse (Kennedy, 1991; McCaskill, Toro, and Wolfe, 1998; Unger et al., 1998).

A Promising Three-category Typology: (1) Transient but Connected; (2) High-risk; and (3) Low-Risk

In order to examine the longitudinal impact of an empirically-derived multivariate typology of homeless youth, Braciszewski, Toro, and Jozefowicz-Simbeni (2011b) used a probability sample of 250 initially homeless youth from throughout the Detroit metropolitan area. Youth were recruited from several different agencies providing services to homeless adolescents, including shelters, outpatient and inpatient substance abuse treatment programs, and psychiatric facilities, as well as some street settings. At baseline, the average participant was 15.3 years old (range 13-17). Youth were interviewed again 0.5, 1.0, 2.0, 4.5, 5.5, and 6.5 years after baseline (ages at last follow-up ranged 20-24). Follow-up rates at these time points were 58, 38, 59, 82, 75, and 83 percent, respectively (for further details on the methodology of this research project, see Ahmed, Fowler, and Toro, 2010; Fowler et al., 2008, 2011; Hobden et al., 2011; Tompsett and Toro, 2010; Urberg, Goldstein, and Toro, 2005).

A wide variety of initial characteristics were used to differentiate the sample into subtypes. These included resilience factors (e.g., family cohesion, self-efficacy, employment, school achievement/performance) as well as negative outcomes (e.g., frequent homelessness, sexual abuse, risky sexual behavior, mental health diagnoses/symptoms). Latent class analysis identified a three-class solution that described youth as either (1) transient but connected (n=55), (2) high-risk (n=46), or (3) low-risk (n=149). For the transient but connected youth, mental health and substance use issues were not prominent; however, as the class label suggests, these youth were nonetheless unstable in terms of both housing and school connections. They showed the most extensive histories of homelessness. However, compared to the other two groups, they reported relatively high cohesion in their families and the most sexual partners. High-risk youth were more likely to have dropped out of school, reported more sexual abuse, more sexual partners, and struggled more with depression, conduct, and substance abuse problems. They also showed substantial housing mobility and histories of homelessness. The low-risk group showed low levels of all the problem behaviors mentioned above, as compared to one or both of the other groups. They showed the least extensive histories of homelessness and housing instability. The low-risk group included more males and younger adolescents. Males were also more likely to be classified as high-risk, as were Caucasian youth. Girls were more likely to fall in the transient but connected group.

Housing trajectories differ across groups

Class membership was then used to predict long-term housing trajectories over the 6.5-year time period using hierarchical linear modeling. As expected, low-risk youth experienced the least homelessness over time and were often in secure living environments. Transient but connected individuals continued an alternating pattern of being homeless and housed. Overall, they tended to experience homelessness the most of the three groups, with respite coming only after 5.5 years. High-risk youth showed a trend toward stable housing during mid- to late-adolescence. However, as they entered young adulthood, these youth experienced a spike in homelessness (43 percent experienced some homelessness between the 18-month and 4.5-year follow up), before returning to levels similar to the other classes. Across all three groups, most did eventually find stable housing during the last two follow-up time points (5.5 and 6.5 years).


Taken together, these findings suggest that targeted interventions can be created for homeless youth, given key characteristics found while they are homeless during mid-adolescence (e.g., mental health, substance use, connection to stable schooling). In addition, it is useful to know that many of these youth eventually gain stable housing; thus, even for youth who are experiencing a number of difficulties early on, positive outcomes are often achieved ultimately. Such findings suggest that most homeless youth are “resilient,” at least in terms of their long-term housing outcomes. Similar “positive” findings showing growing housing stability over time were obtained in a recent two-year follow-up of newly homeless youth in Los Angeles and in Melbourne, Australia (Milburn et al., 2007) as well as in longitudinal studies of homeless adults and families (Stojanovic, Weitzman, Shinn, Labay, and Williams, 1999; Toro et al., 1997, 1999).

Gender influence may also be important during this developmental period, as females were more likely to be in the “transient but connected” group (but less likely to be in the other two groups, one doing well initially, one having significant problems in many areas). Such findings are not altogether surprising, given the nature of available services for homeless youth and young adults. Many shelters allow for adolescent females to remain with their families and/or mothers, while male teens are filtered out of all-female facilities. Furthermore, girls in our culture are typically “trained” to be more family-oriented and boys to be more independent. Continued exploration of differential male and female trajectories is warranted in future research, especially with regard to such wide ranging outcomes for males.