There are an estimated 1.6 million runaway and homeless youth in the United States each year, but adolescent homelessness has been relatively understudied compared to adult homelessness [1]. Homeless youth are disproportionately lesbian, gay, bisexual, transgender, or questioning (LGBTQ) [2,3], and the risk of adolescent homelessness tends to increase with age [1]. Prior research has found that male adolescents are more likely to experience homelessness, but have not identified racial/ethnic disparities among this population [4]. Much research has assessed the potential determinants of homelessness (e.g., experiencing child abuse, substance abuse, sexual abuse, etc.) [5], however, aside from Ringwalt and colleagues [4], little literature assesses the correlates and period prevalence of homelessness within a population of homeless and non-homeless adolescents.

Despite heterogeneity in the types of homelessness experienced by this population, there is also very little research assessing which adolescents report what kinds of homelessness experiences. Homeless youth spend the night not only in shelters, but in public places (e.g., on the streets, in abandoned buildings, outdoors, and underground), and/or with strangers [1,4,6,7]. It is important to understand which adolescents experience different types of homelessness, as the risk environment may vary greatly by type. Shelters tend to be relatively protective, whereas spending the night in a public place or with a stranger may expose youth to sexual violence, and/or situations that could be potentially exploitative, including pressures to participate in exchange sex (i.e., trading sex for money, food, drugs, or a place to stay) [4–6,8–11].

Since the 1980s, a large body of work has demonstrated that homeless youth are more likely than housed youth to engage in risky sexual behaviors, such as unprotected and exchange sex [7, 12–17]. However, comparisons of sexual risk behaviors between homeless adolescents and their non-homeless peers are largely made implicitly. When explicit, they are often done with community-based samples [16,18], or accomplished post hoc by contrasting community-based samples of homeless adolescents to probability samples of adolescents in school-based settings, from data sources such as the Centers for Disease Control and Prevention’s (CDC) Youth Risk Behavior Survey (YRBS) [7,12].

Aside from recent findings linking sexual orientation to homelessness, little research exists that provides insight into the demographic characteristics associated with adolescent homelessness. Further, we are aware of no prior research that uses a probability sample among a single population of homeless and non-homeless adolescents to examine the association between sexual risk behavior and homelessness experiences, including the specific type of homelessness experiences (i.e., street, stranger, or shelter stays). Homelessness in this study is defined by one or more nights in the prior year in at least one of several circumstances which constitute homelessness. As such, so-called “runaway” youth who spend shorter periods of time away from home and chronic homeless youth who have extended street stays cannot be disentangled. It is important to understand the extent to which adolescents experience any kind of homelessness in Los Angeles, as the city is one of the epicenters of homelessness in the United States. In the current study, we use a random sample of adolescents in the Los Angeles Unified School District (LAUSD) to assess socio-demographic differences between adolescents who do and do not report homelessness in the past year, and the correlates of experiencing different types of homelessness. We also examine how any experience of homelessness, and the type of homelessness experience, may be associated with sexual risk taking.

Descriptive statistics are presented in Table 1. A majority of participants were Hispanic/Latino (71.5 %). Males and females were approximately evenly distributed in the sample, and 13% of the youth reported their sexual orientation as LGBTQ. Slightly less than one-quarter (23.8%) had experienced some type of homelessness in the past year. More than 10% of the respondents indicated they spent the night in a shelter or in a public place, respectively, and nearly 6% indicated they spent the night with a stranger. About 40% indicated a history of sexual activity and, of these, 61% reported using a condom at last sexual intercourse.

Table 2 presents the multivariate logistic regression models assessing associations with any homeless experience and types of homelessness. Younger age (OR=0.90; 95% CI: 0.81–0.99), male gender (OR=1.94; 95% CI: 1.54–2.46), identifying as LGBTQ (OR=2.18; 95% CI: 1.59–2.99), and identifying as “other race” (OR=2.48; 95% CI: 1.67–3.70) were statistically significantly associated with any experience of homelessness in the past year. Among those who experienced any homelessness, staying with a stranger was more likely among LGBTQ (OR=2.38; 95% CI: 1.32–4.29) and Black/African American youth (OR=3.40; 95% CI: 1.67–6.91). Staying in a shelter was less likely among older students (OR=0.80; 95% CI: 0.67–0.96) and those who identified as LGBTQ (OR=0.44; 95% CI: 0.25–0.78). We found no statistically significant correlates of staying in a public place.

Table 3 presents the multivariate logistic regression models assessing correlates of being sexually active and having had unprotected sex at the last sexual encounter. Among the full sample, older age (OR=1.65; 95% CI: 1.50–1.81), being male (OR=1.62; 95% CI: 1.32–2.00), identifying as LGBTQ (OR=1.73; 95% CI: 1.27–2.36), Black/African American race (OR=1.42; 95% CI: 1.03–1.96), and having had a homelessness experience in the past year (OR=1.57; 95% CI: 1.23–2.01) were all significantly associated with an increased likelihood of being sexually active. Youth of “other races” were less likely (OR=0.59; 95% CI: 0.39–0.99) to have been sexually active compared to Hispanic/Latino youth. Among those who had experienced homelessness in the past year, older age (OR=1.50; 95% CI: 1.24–1.81) and having spent the night with a stranger (OR=3.36; 95% CI: 1.90–5.93) were significant predictors of being sexually active.

Among the full sample, unprotected sex at last sexual encounter was less likely among males (OR=0.57; 95% CI: 0.41–0.80) and more likely among LGBTQ youth (OR=1.87; 95% CI: 1.20–2.89). Among those who experienced homelessness, males were still less likely to have had unprotected sex at last sexual encounter (OR=0.48; 95% CI: 0.24–0.94), while those who stayed with a stranger (OR=3.10; 95% CI: 1.44–6.67) or in a public place (OR=2.61; 95% CI: 1.29–5.27) were more likely to have had unprotected sex at their last sexual encounter when compared to those whose homelessness experience had been in a shelter only.

There are several important findings to emerge from these results. As Ringwalt et al.’s [4] nationally representative study found with data collected 20 years ago, males continue to be more likely to experience homelessness. The racial and ethnic correlates of adolescent homelessness deserve further exploration, potentially in a larger sample of youth, as the current findings that “other races” are associated with increased odds of experiencing homelessness are relatively vague (we cannot parse out which of the “other” categories might be associated with homelessness due to small sample sizes).

As has been demonstrated in recent findings, LGBTQ youth also bear a disproportionate burden of homelessness relative to heterosexual youth [2,3]. Furthermore, it appears that sexual orientation is associated with risky kinds of homelessness experiences, as is Black/African American race. Staying with a stranger was more likely among Black/African American youth and those who identified as LGBTQ, while LGBTQ and older youth were less likely to have stayed in a shelter. Other research has suggested that LGBTQ youth may not feel welcome in shelters that primarily focus their services on heterosexual youth [5, 13]. Staying with a stranger exposes adolescents to greater threats of violence and victimization than shelter stays [4–6,9,11], and staying with a stranger may put young people at particular risk for sexual exploitation. This makes the over-representation of Black/African American and LGBTQ adolescents among those who stay with strangers especially troubling; these groups already experience social marginalization, and the risks associated with stranger-based homelessness experiences may be associated with increased victimization in these already vulnerable populations.

This study adds to prior literature by comparing adolescents’ sexual risk behaviors between those who have experienced homelessness in the past year, versus those who have not, within the same population. In accordance with post hoc comparisons of homeless and housed adolescents [7,12], and community-based samples [16], adolescents in this sample with past-year homelessness experiences are more likely to report lifetime sexual activity. Not using condoms, however, was not associated with reporting homelessness experiences in the prior year, but rather with specific types of homeless experiences. It must be noted that our data did not allow us to assess if this last sexual encounter was during a homelessness episode. These results show a clustering of risk behaviors among Los Angeles adolescents, encompassing certain types of homelessness experiences in the past year and their sex behaviors.

The examination of adolescents who experienced homelessness provides insights into which adolescents experienced what type(s) of homelessness and whether they also reported condom use at their last sexual encounter. As has been seen in community-based samples of homeless youth, adolescents who stayed in a public place were more likely to have reported unprotected sex, relative to those youth who stayed in a shelter [9,10]. Strikingly, staying the night with a stranger was significantly associated with an increase in reports of both lifetime sexual activity and not using a condom at last intercourse, relative to adolescents whose homelessness experience in the past year was in a shelter. Again, we must stress that we do not know if this sexual encounter occurred while staying with a stranger, nor do we know if it involved exchange sex. Despite these limits of the data, what is clear is that the same adolescents who reported staying with a stranger in the past year also disproportionately reported being sexually active and not using a condom during their last sexual encounter.

The potential for associations between stranger stays and sexual exploitation is an area where further research is needed. As a supplemental study, this research was limited to a small number of sex risk questions, and did not assess for exchange sex. Future research should examine the specifics of sexual risk behavior in the context of specific homeless experiences, particularly stranger stays, to better understand whether sex risk occurred during homelessness experiences, and if homelessness is causally linked with sexual risk behavior. A timeline follow back approach may assist in determining correlations between sex risk behaviors and homelessness experiences by identifying specific dates/periods of time when the adolescent was homeless and his/her sexual behaviors.

Other limitations exist in this study. As with any cross-sectional data, these data provide only associations and not causality. Response biases are minimized because this was an anonymous survey, however, social desirability bias among adolescents may also have led to under- or over-reporting of sex behaviors and homelessness experiences. For example, respondents may over-report condom use, as this is a socially desirable behavior, or over-report engagement in sex due to perceived adolescent social norms. There are a few limitations to our sexual risk behavior items. First, we know nothing about the last sexual encounter except for whether or not a condom was used. We do not know if this sex was during a homeless episode or not and we do not know if the partner was a casual partner or a boyfriend/girlfriend. Moreover, our sexual activity variable combines oral, vaginal, and anal intercourse, which limits the results as these sexual activities have varying levels of risk for HIV and other sexually transmitted infections. Homelessness experiences may be under-reported because of social undesirability (though the question specifically did not include the word “homeless” in order to mitigate potential stigma), or because those students who are currently experiencing homelessness are less likely to be in school, and thus were not sampled for the survey. Homelessness experience responses may be over-reported if respondents did not understand the question. For instance, Ringwalt and colleagues [4] found in focus groups after conducting their survey that adolescents had difficulty with the public place, shelter, and stranger response options; for example, “spending a night” in a public place may simply be hanging out with friends late at night and then returning home, while a “shelter” was misconceived by some as squats where their friends met up, and staying with a stranger may truly be a friend of a friend which may not imply exchange sex or the opportunity for sexual coercion[4]. Moreover, these items do not address whether the youth was accompanied by a parent or not during the homelessness episode in question. Many of these items capture “runaway” experiences as well as chronic homelessness [4]. With these items alone, no distinction between the two states can be made and research has shown that most newly homeless youth return home after relatively short stays on the street [21]. These items, however limited, remain the best items used to estimate the prevalence of youth homelessness to date.

These results have important policy and programmatic implications. LBGTQ youth are more likely to spend the night in a public place or with a stranger, perhaps because of perceived and actual stigma associated with their sexual orientation. Because both of these experiences increase risks for violence and victimization (and both of these types of homelessness experiences are associated with unprotected sex in this study), schools should work to address the unique needs of LGBTQ youth who may be experiencing homelessness. Moreover, school-based sexual health programs should recognize the prevalence of sexual risk taking among adolescents who experience homelessness. Addressing the needs of homeless adolescents, especially LGBTQ adolescents, while in school is critical. Discussions of exchange sex should also be included in sexual health education programs. This can be addressed within the context of promoting healthy relationships and power dynamics/balances within relationships.