Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Jan 1.
Published in final edited form as: J Public Child Welf. 2019 Mar 15;14(2):192–208. doi: 10.1080/15548732.2019.1590288

Predictors and correlates of unstable housing experiences among a child welfare-involved sample

Janet U Schneiderman 1, Andrea K Kennedy 2, Theresa A Granger 3, Sonya Negriff 4
PMCID: PMC7539746  NIHMSID: NIHMS1523311  PMID: 33041723

Abstract

The study examined whether youth demographics, family factors, and maltreatment type were related to unstable housing and whether unstable housing predicted delinquency and marijuana use. Participants included 216 child welfare-affiliated adolescents (mean age = 18.2 years). Youth with more lifetime residences were more likely to experience unstable housing although Latino youth (compared to White, Black, or multiethnic/biracial) were less likely to experience unstable housing. Unstable housing was associated with subsequent delinquency. Caregiver type (parent vs. relative/unrelated caregiver) was not related to unstable housing, thus homelessness prevention programs should include youth who remain with their parents and those with non-parent caregivers.

Keywords: unstable housing, homelessness, adolescent, child welfare, delinquency, juvenile

Introduction

The period of adolescence and the transition to adulthood can be difficult for child welfare-affiliated youth, as this population has a high prevalence of mental health problems accompanied by increasing underutilization of mental health services as youth age into adulthood (Ringeisen, Casanueva, Urato, & Stambaugh, 2009). Contributing to these difficulties is the high likelihood of unstable housing experiences for this population (Kushel, Yen, Gee, & Courtney, 2007). Unstable housing includes homelessness and other nonstable residences, such as living in cars (Corneil et al., 2006). Most research on child welfare-affiliated youth, i.e. children who have had an open case in the child welfare department, has focused on unstable housing following foster care, and there is call for research on unstable housing to include youth in child welfare who remain with their parents (Ryan, Perron & Huang, 2016). The present study examined the demographic, family, and maltreatment characteristics that were related to an unstable housing experience for child welfare-affiliated youth who remained with their parents as well as those who had a non-parent caregiver. Additionally, this study examined the relationship between unstable housing and subsequent delinquency and marijuana use in late adolescence.

Family poverty is at the root of unstable housing in families involved with child welfare as rent is usually the largest expenditure in household budgets and families struggle to meet the need for stable housing (Cunningham, Pergamit, Baum, & Luna, 2015). Families in general in the United States are getting poorer, and homelessness is associated with poverty even more now than in the 1980’s (Grant, 2013). Thus, the effect of poverty on child welfare-involved families and their children is imperative to consider in order to understand how these children end-up having unstable housing experiences.

Additionally, for child-welfare involved youth, the type of maltreatment they have experienced has been found to be related to variability in the experience of unstable housing. In a small qualitative study, sexual abuse was identified by 39% and physical abuse was identified by 50% of the homeless youth (Ferguson, 2009). In a larger sample of 303 homeless youth ages 12–20, more than one-third identified physical violence as the reason for leaving home (Mallett & Rosenthal, 2009). The relationship of sexual and physical abuse to youth homelessness was also found in an earlier quantitative study conducted in Seattle, Washington, where one-half of the youth reported being physically abused and almost one-third experienced sexual abuse (Tyler & Cauce, 2002). For former foster youth, unstable housing experiences were associated with a history of physical and sexual abuse (Fowler, Toro & Miles, 2009). Sexual abuse and physical abuse are the maltreatment types most often associated with unstable housing, although one study found neglect was associated with unstable housing in young adults (Shelton, Taylor, Bonner, & van den Bree, 2009).

One of the negative outcomes associated with unstable housing among child-welfare affiliated youth is engagement in delinquent behavior (Dworsky & Courtney, 2009). The stress of being in an unstable housing experience can lead youth to adopt “survival behavior” which includes meeting the basic needs of food and clothing; youth sometimes engage in delinquent behaviors to meet these needs (Thompson, 2017). Additionally, unstable housing as well as delinquency can differ by the sex of the youth. In a longitudinal study of youth who aged out of foster care, being male and more engagement in delinquency behaviors were related to experiencing homelessness (Dworsky, Napolitano, & Courtney, 2013). Similarly, male former foster youth reported more homelessness and unstable housing than their female counterparts (Kushel et al., 2007). Delinquent behaviors that begin in childhood are more frequent in males than females. This trend of sex differences in the onset of delinquency continues through adolescence, although the difference between sexes is smaller in adolescence than in childhood (Moffitt & Caspi, 2001). Delinquency often co-occurs with substance use and both risk behaviors are outcomes related to child welfare involvement and experiences of unstable housing. In Canada, a population-based study found that youth who experienced unstable housing had higher rates of both problematic substances use and violence concerns than youth who were stably-housed (Smith, Hawke, Chaim & Henderson, 2017). For adolescents who attended a HIV clinic, those who used drugs were more likely to use marijuana than other hard drugs, and those with a history of unstable housing were more likely to use marijuana compared to youth without a history of unstable housing (Eastwood & Birnbaum, 2007). Therefore, we examined whether unstable housing predicted the most common untoward outcomes of delinquency and marijuana use in our sample of child welfare-affiliated youth.

In thinking about how unstable housing contributes to risk for adolescents and young adults, the concept of housing careers for youth has been proposed (Collins & Curtis, 2011). A housing career refers to the sequence of dwellings a person or family occupies and how this process of housing affects outcomes for the family members. Frequent moves resulting in a high number of residences (housing instability) has some risk for adolescents, but negative housing situations or having unstable housing, such a living in jails or being homeless, most often result in negative outcomes (Collins & Curtis, 2011; Frederick, Chwalek, Hughes, Karabanow, & Kidd, 2014). Thus, housing instability and unstable housing are different concepts; housing instability is solely the total count of residences, whereas unstable housing is the lack of quality in the type of housing. Unstable housing, child welfare involvement, delinquency, and marijuana use in adolescence all appear to be interrelated and these relationships have been shown to differ based on individual and family factors.

Most research on unstable housing in child welfare-affiliated youth is focused on how the foster care experience is related to an adolescent’s subsequent homelessness, or how a family’s unstable housing is a factor in causing child welfare involvement (Dworsky, 2014; Fowler, Marcal, Zhang, Day, & Landsverk, 2017). There is very little research on unstable housing in child welfare-affiliated adolescents who have not experienced foster care. Ryan, Perron, and Huang (2016) noted that many of the risk factors for unstable housing are similar for child welfare-involved youth who remain in their home of origin compared to those who enter foster care and called for more research on unstable housing to be conducted on youth with a child welfare history who remain with their parents.

Current Study

Addressing this gap, the sample for this study included both youth who ever had a nonparent caregiver and those who remained with their parents. This study examined: (a) whether youth demographics (age, sex, ethnicity/race), family factors (living with a nonparent caregiver, number of lifetime residences, family income), and maltreatment type were related to an incidence of unstable housing; and (b) whether unstable housing predicted subsequent delinquency and marijuana use after controlling for relevant confounding variables. A better understanding of risk factors for unstable housing can help social services, health, and child welfare professionals prevent unstable housing experiences during childhood and adolescence and possibly curtail subsequent risk behaviors.

Methods

Participants

Data came from an ongoing longitudinal study examining the effects of child welfare-documented maltreatment on adolescent development. At Time 1 (T1; 2002–2005), the enrolled maltreated sample included 303 adolescents aged 9–13 years old. Of these youth, 73% completed the T4 assessment (n = 222), an average of 7.2 years after T1 (2009–2012). The current study included those who completed a household stability questionnaire at Time 4 (T4; n = 219); however, three participants were removed from the final sample due to their ability to recall less than 1 year of lifetime residence history or not being able to account for residence history prior to age 7. Number of lifetime residences was one of the covariates in this study, thus we could only utilize participants who could recall their whole housing history to be able to total their number of residences. The final sample was 216 maltreated participants. A comparison sample was obtained as part of the larger study but was not used for the current analyses.

Participants were selected from a large city in California. The inclusion criteria were: (a) a new substantiated referral to the child welfare department (CWD) during the preceding month for any type of maltreatment; (b) child age of 9–12 years (some children turned 13 years old between when they enrolled and were interviewed); (c) child identified as Latino, Black, or White (non-Latino); and (d) child resided in one of 10 zip codes in a designated county at the time of referral to the CWD. White (non Latino), Black, and Latino are the major race/ethnicity categories involved in the CWD where this study took place and the original study chose to focus on only those race/ethnicities in order to have the power to test race/ethnic differences (Los Angeles County Children and Family Services, 2019). With the approval of the CWD, the county juvenile court, and the institutional review board of the affiliated university, all CWD-referred potential participants’ caregivers identified by the CWD were contacted by mail. Of the families referred by CWD, 77% of the caregivers agreed to participate.

Procedures

Assessments were conducted at an urban research university. Institutional approvals were obtained from the university and juvenile court. After assent for adolescents ˂ 18 years old, and consent for the adolescents ≥ 18 years old and caregivers were obtained, the caregivers and adolescents completed questionnaires and tasks during a 4-hour protocol. The protocol for data collection was similar for each of the four time points. Both the child and caregiver were paid for their participation according to the National Institutes of Health standard compensation rate for healthy volunteers.

Measures

Demographics and caregiver type.

Caregivers provided information at T1 about the child’s birth date, sex, ethnicity, and race. Birthdate was used to calculate the youth’s age at the T4 assessment. Although the CWD identified the child’s ethnicity/race as Latino, Black, or White (non-Latino), caregivers included multi-ethnic or bi-racial as a category. Caregiver type for the sample was dichotomized as youth who reported having a birth parent or adoptive parent as their primary caregiver at all four time points versus youth who reported living with a nonparent caregiver (relative or unrelated caregiver) at any time point. We did not have foster care status, therefore we could only categorize caregiver type. At T4, 15.74% (n = 34) of the maltreated youth were not living with any caregiver. When no caregiver was listed at T4, we used caregiver information from Times 1, 2, and 3. Family income was reported by caregivers at all four time points. These categories include $14,999 and under, $15,000-$29,999, $30,000-$59,999, and $60,000 and above, and categories were assigned a number (1–4) in ascending order. Income category was averaged across all time points. If no caregiver was listed at T4, this time point was excluded from the averaged income. Employment information was only collected from the caregiver who attended the assessment, thus we did not have a complete picture of employment of all caregivers in the household and employment status was not included.

Maltreatment Classification.

Research assistants abstracted information from child welfare case records to classify the types of maltreatment experienced (see Negriff, Schneiderman, & Trickett [2016] for details of the record abstraction). Categories included neglect (n = 169), emotional abuse (n = 118), physical abuse (n = 113), and sexual abuse (n = 46). Definitions were derived from the Department of Children and Family Services (n.d.). Four separate maltreatment variables indicated the presence or absence of that particular type of maltreatment (not mutually exclusive) and 76% experienced more than one type. The maltreatment classification was determined at T1 after abstraction of the case files at T1 and previous to T1. Maltreatment reports subsequent to T1 were not collected. Maltreatment type was dichotomized into those who had experienced physical and/or sexual abuse (with or without neglect or emotional maltreatment) and those who had experienced neglect and/or emotional maltreatment but not physical or sexual abuse. We chose this dichotomization because physical abuse and sexual abuse were the maltreatment types most associated with experiencing unstable housing.

Measures in the following sections came from data collected at T4.

Residence number and type.

Youth completed a household stability questionnaire designed for this study to measure lifetime residence history only at T4. In this interviewer-administered survey, youth were asked to list their first residence and sequentially account for all residences until the present. Based on this self-report, the number of residences was totaled by counting all residences prior to the year before T4 (e.g., if the adolescent was 18 years old at T4, only residences from birth to age 17 were included). The household stability questionnaire also asked youth to list the type of housing for each residence (e.g., house, car, etc.). Type of housing was classified as either stable or unstable housing. For this analysis, we used the categories of unstable housing that policy-makers and researchers have used. (Corneil et al., 2006; Cox, Henwood, Wenzel, & Rice, 2016). Unstable housing categories in this study included homeless, car, transition housing for foster youth, juvenile hall, jail, prison, camp, youth authority, group home, shelter, church, and motel. Only incidents of unstable housing identified prior to the year before T4 data collection were included (e.g., if the adolescent was 18 years old at T4, only unstable housing incidents from birth to age 17 were included). Stable housing categories included home, apartment, mobile home, military housing, job corps dorm, and college residence. Two variables were created for unstable housing: 1) number of incidents of unstable housing and 2) at least one incident of unstable housing.

Delinquency.

Youth completed the Adolescent Delinquency Questionnaire (adapted from Huizinga & Elliott, 1986) via computer to protect participant confidentiality. The present study combined two scales from the original questionnaire: person offenses (7 items, e.g., “attacked someone with a weapon with the idea of seriously hurting them,” α = .74) and property offenses (15 items, e.g., “damaged or destroyed someone else’s property on purpose,” α = .92). All questions had six possible answers: 0, 1, 2, 3, 4, and 5 or more times during the previous 12 months. Scores had a possible range of 0 to 110 on the combined delinquency scale.

Marijuana use.

Adolescents also reported their own substance use on the Adolescent Delinquency Questionnaire. Marijuana use was taken from one question that asked about the frequency of marijuana or hashish use during the previous 12 months. Potential responses included 0, 1, 2, 3, 4, and 5 or more times.

Analyses

All analyses were conducted using SAS 9.4. Both unstable housing variables were tested for all analyses. The first variable (i.e., number of incidents of unstable housing) was not significant in our models, and thus, the dichotomous variable (at least one incident of unstable housing) was used for all reported analyses. Prior to modeling, we tested the correlations between of all study variables.

Logistic regression was used to predict the odds of having an incident of unstable housing 1 year prior to T4. Predictors included demographics (age, sex, being Latino), type of caregiver, average family income category, physical and/or sexual abuse, and number of residences. Race/ethnicity was dichotomized into each racial/ethnic group reported (Black, Latino, White and multiethnic/biracial) versus remaining youth for all analyses. Ultimately, being Latino was the only significant predictor of unstable housing, and thus we used this dichotomization for all analyses.

Linear regression was used to examine the housing predictors of delinquency and marijuana use, separately. We tested two housing predictors, an incident of unstable housing and number of residences. Since the measurement of delinquency behaviors included the year before Time 4, we limited the housing predictors (an incident of unstable housing and number of residences) to a year prior to data collection at T4. Covariates included age, sex, being Latino, type of caregiver, average family income category, and a history of physical and/or sexual abuse.

Results

Sample characteristics

Table 1 reports demographic information for the participants and means of the outcome variables for the total sample. The final sample was predominantly Latino or Black (77.3%), with a fairly equal representation of sex (male = 47.7%). The average age was slightly more than 18 years old (M = 18.24, SD = 1.34). The average number of residences was 5.6 (SD = 3.41, range = 1–17), and 17.1% of the sample had experienced unstable housing at least once. The breakdown for incidents of unstable housing was as follows: group homes (37.9%), juvenile hall (20.7%), shelter (17.2%), motel (9.2%), homeless (4.6%), car (4.6%), youth authority (3.4%), and transition housing for foster youth (2.3%). The correlations between all study variables are shown in Table 2.

Table 1.

Characteristics of the sample (N = 216)

Variable
Age (years) M (SD) 18.24 (1.34)
Sex n (%)
 Female 113 (52.3)
 Male 103 (47.7)
Race and ethnicity n (%)
 Black 92 (42.6)
 Latino 75 (34.7)
 White 20 (9.3)
 Mixed or biracial 29 (13.4)
Caregiver n (%)
 Biological or adoptive 113 (52.3)
 Non-parent 103 (47.7)
Average family income n (%)
 $14,999 and under 48 (22.3)
 $15,000-$29,999 116 (54.0)
 $30,000-$59,999 37 (17.2)
 $60,000 and above 14 (6.5)
Physical and/or sexual abusea n (%) 126 (58.3)
Number of residences M (SD) 5.6 (3.41)
Unstable housing statusb n (%) 37 (17.1)
Number of incidents of unstable housingc M (SD) 2.2 (2.25)
Person and property offense delinquencyd M (SD) 6.1 (12.41)
Marijuana usee M (SD) 1.8 (2.19)
Open in a new tab
a

0 = neglect or emotional maltreatment but not physical or sexual abuse, 1 = physical and/or sexual abuse with or without neglect or emotional maltreatment.

b

0 = no incidents of unstable housing, 1 = at least one incident of unstable housing.

c

Among those with an incident of unstable housing.

d

Response options were 0, 1, 2, 3, 4, and 5 or more times during previous 12 months; 22 items; score range of 0–110.

e

Response options were 0, 1, 2, 3, 4, and 5 or more times during previous 12 months; one item; score range of 0–5.

Table 2.

Pearson correlation coefficients (N = 216).

  1 2 3 4  5 6 7 9 9 10
1. Age ---
2. Sexa 0.05 ---
3. Latinob 0.07 0.13 ---
4. Caregiverc 0.03 0.00 −0.03 ---
5. Average family income categoryd −0.03 0.06 −0.19** 0.34*** ---
6. Physical and/or sexual abusee 0.03 −0.04 −0.05 −0.02 0.14* ---
7. Number of residences 0.18** 0.03 0.08 0.14* 0.13 0.09 ---
8. Unstable housing statusf 0.14* −0.08 −0.13 0.06 0.09 0.11 0.47*** ---
9. Person and property offense delinquencyg 0.06 −0.20** 0.01 −0.02 −0.13 0.13 0.07 0.19** ---
10. Marijuana useh 0.14* −0.13 0.00 0.00 −0.01 0.21** 0.14* 0.09 0.34*** ---
Open in a new tab
a

0 =girl, 1 = boy

b

0 = race other than Latino, 1 = Latino.

c

0 = birth parent or adoptive parent, 1 = unrelated or related caregiver.

d

Categories include $14,999 and under, $15,000-$29,999, $30,000-$59,999, and $60,000 and above.

e

0 = neglect or emotional maltreatment but not physical or sexual abuse, 1 = physical and/or sexual abuse with or without neglect or emotional maltreatment.

f

0 = no incidents of unstable housing, 1 = at least one incident of unstable housing.

g

Response options were 0, 1, 2, 3, 4, and 5 or more times during previous 12 months; 22 items; score range of 0–110.

h

Response options were 0, 1, 2, 3, 4, and 5 or more times during previous 12 months; one item; score range of 0–5.

*

p < .05.

**

p < .01.

***

p < .001

Model predicting unstable housing

For the logistic regression predicting any incident of unstable housing, significant predictors were being Latino (OR = 0.29; CI = 0.10, 0.83) and number of residences (OR = 1.45; CI = 1.27, 1.65) (see Table 3). Latinos were less likely to have an incident of unstable housing compared to non-Latinos. When ethnicities other than Latino were included in the model, none were statistically significant. In terms of the number of residences, every additional residence increased the odds of having an incident of unstable housing by 1.5 times.

Table 3.

Logistic regression predicting unstable housing statusa (N = 216).

Predictor b p-value SE OR CI
Demographic Characteristics
 Age 0.13 0.430 0.16 1.13 0.83–1.55
 Sexb 0.47 0.295 0.45 1.60 0.66–3.84
 Latinoc −1.26 0.021 0.54 0.29 0.10–0.83
Family Characteristics
 Number of residences 0.37 <0.001 0.07 1.45 1.27–1.65
 Caregiverd −0.20 0.672 0.47 0.82 0.32–2.07
 Average family income categorye 0.07 0.814 0.32 1.08 0.58–2.00
Physical and/or sexual abusef 0.42 0.268 0.45 1.52 0.63–3.70
Open in a new tab
a

0 = no incidents of unstable housing, 1 = at least one incident of unstable housing.

b

0 =girl, 1 = boy.

c

0 = race other than Latino, 1 = Latino.

d

0 = birth parent or adoptive parent, 1 = unrelated or related caregiver.

e

Categories include $14,999 and under, $15,000-$29,999, $30,000-$59,999, and $60,000 and above.

f

0 = neglect or emotional maltreatment but not physical or sexual abuse, 1 = physical and/or sexual abuse with or without neglect or emotional maltreatment.

Models predicting delinquency and marijuana use

The results for the final model predicting delinquency and marijuana use separately are presented in Table 4. One or more incidents of unstable housing predicted higher levels of delinquency (β = 0.19, t = 2.45, p = 0.015). Of the covariates, being male significantly predicted higher levels of delinquency (β = 0.19, t = 2.79, p = 0.006) and lower average family income predicted higher levels of delinquency (β = −0.16, t = −2.15, p = 0.033).

Table 4.

Linear regression predicting delinquency and marijuana use (N = 216).

  Delinquency Marijuana Use
  b p-value SE β b p-value SE β
 Age 0.37 0.557 0.63 0.04 0.21 0.062 0.11 0.13
 Sexa 4.63 0.006** 1.66 0.19 0.60 0.043* 0.30 0.14
 Latinob 0.84 0.642 1.79 0.03 0.03 0.930 0.32 0.01
 Caregiverc 0.74 0.671 1.75 0.03 0.02 0.960 0.31 0.00
 Average family income categoryd −2.63 0.033* 1.22 −0.16 −0.12 0.576 0.22 −0.04
 Physical and/or sexual abusee 2.99 0.078 1.69 0.12 0.84 0.006** 0.30 0.19
 Number of residences −0.05 0.858 0.28 −0.01 0.07 0.176 0.05 0.11
 Unstable housing statusf 6.18 0.015* 2.53 0.19 0.06 0.897 0.45 0.01
R2 0.112 0.093
Open in a new tab
a

0 =girl, 1 = boy

b

0 = race other than Latino, 1 = Latino.

c

0 = birth parent or adoptive parent, 1 = unrelated or related caregiver.

d

Categories include $14,999 and under, $15,000-$29,999, $30,000-$59,999, and $60,000 and above.

e

0 = neglect or emotional maltreatment but not physical or sexual abuse, 1 = physical and/or sexual abuse with or without neglect or emotional maltreatment.

f

0 = no incidents of unstable housing, 1 = at least one incident of unstable housing.

*

p < .05.

**

p < .01.

The results for the final model predicting marijuana use are also presented in Table 4. An incident of unstable housing did not predict marijuana use. Two of the covariates were related to marijuana use: physical and/or sexual abuse predicted more frequent marijuana use (β = 0.19, t = 2.86, p = 0.006) and male sex predicted more marijuana use (β = 0.14, t = 2.04, p = 0.043).

Discussion

This study examined whether demographic, family, and maltreatment characteristics were associated with an unstable housing experience for a child welfare-affiliated population and whether an incident of unstable housing predicted delinquency or marijuana use. We found that youth who had more lifetime residences and were not Latino were more likely to have an incident of unstable housing. A previous analysis which included both child welfare-affiliated and comparison adolescents found that having more lifetime residences was associated with person offense delinquency (Schneiderman, Kennedy, Negriff, Jones, & Trickett, 2016), whereas in the current analysis we found that living in an unstable residence was also an important predictor of delinquency for child welfare-affiliated adolescents when controlling for number of residences. For this analysis, we only used the child welfare-affiliated sample because the comparison sample did not have enough variability in unstable housing incidents. The number of incidents of unstable housing did not increase the chances of an adolescent reporting delinquency, but ever having an unstable housing incident did. Unstable housing did not predict marijuana use. This study illustrates the importance of preventing one or more unstable housing experience for child welfare-affiliated youth during their childhood or adolescence since an incident of unstable housing predicted delinquent behaviors.

Latino youth were less likely to have an incident of unstable housing. Latino youth might have had larger families, lived in more multigenerational houses, and had more extended family members who provided housing (Landale, Oropesa, & Bradatan, 2006). The average size of a Latino family (3.87 people) is larger than the national average (3.19 people; Population Research Institute, 2015). Swartz (2009) concluded that Black and Latino families are more likely to provide practical support and housing support than White families, although in this study Latino youth were the only group less likely to have an incident of unstable housing. It could be that many of the Latino youth in this study came from immigrant families. In 2014, California (the location of this study) had 5.5 million foreign-born immigrants from Latin America (Migration Policy Institute, 2016). Immigrant families tend to pool resources to make sure that their families have social and economic support (Swartz, 2009). Another possible explanation for the decreased likelihood of Latino youth to report unstable housing is the concept of familism, which is common in Latino families. Familism can be viewed as a protective factor and includes the idea that family cohesion, extended family networks, and high social support may mitigate against untoward outcomes for their youth, including unstable housing (Landale et al., 2006).

Not surprisingly, in this study youth with more lifetime residences were more likely to have an unstable housing experience. The relationship between more residences and unstable housing was also found in a study of women with recent criminal justice history where multiple residences were more common among those in unstable housing (80%) than those in stable housing (44%; Weir et al., 2007). When conceptualized broadly, unstable housing begins with high residential instability (moving residences) and financial distress which can lead to unstable home environments (i.e. precarious housing) such as doubling-up, living in hotels, and ultimately homelessness (Clark, 2013; Dickson-Gomez et al., 2009). Even for youth in foster care, moving to a neighborhood where there is more residential instability increases the risk of delinquency (Huang, Ryan, & Rhoden, 2016).

Most of the research about child welfare involvement and unstable housing is specifically about foster care (Dworsky et al., 2013; Huang et al., 2016). Unfortunately, the foster care status of the study population was not known, but we did have information about types of caregivers. As expected, our findings indicated that having ever lived with a nonparent caregiver did not predict an incident of unstable housing compared with always having lived with birth parents or adoptive parents. Therefore, caregiver type was not a correlate of unstable housing and youth who ever lived with relatives or foster unrelated caregivers were no more likely to have an incident of unstable housing than youth who always lived with their parents.

Unstable housing predicted delinquency but not marijuana use. The positive association between unstable housing and delinquency is supported by other studies (Dworsky et al., 2013; Lee & Villagrana, 2015). Thus, a housing career that includes even one unstable residence during childhood or adolescence is a predictor of delinquency even when controlling for number of residences. Our unstable housing category included jail and youth authority facilities, and 12 of our participants experienced at least one incident of incarceration in addition to other forms of unstable housing. In order to test whether including incarcerated participants in the unstable housing category affected the delinquency reports, we performed a post-hoc analysis examining if incarceration impacted delinquency. No statistically significant difference occurred in delinquency scores between those who were ever incarcerated and those who were never incarcerated. All of the families in this study lived in similar low-income communities, thus community influence rather than housing influence might have been important in marijuana use. For example, researchers have found that youth living in deteriorating neighborhoods were more likely to use marijuana two years after high school than adolescents living in always-good neighborhoods (Furr-Holden et al, 2011).

Limitations

This study has restricted generalizability because the study population was primarily minority and urban. The adolescents self-reported both their number of lifetime residences and the type of housing, but their recollections may have not been completely accurate. The adolescents’ recollections of time spent in each residence were fairly precise for recent residences, thus we were able to determine which residence changes and unstable housing experiences occurred within the last year. Recollections of time spent in each residence prior to the last year were imperfect, therefore we could not determine the length of time spent in each type of residence. Participants did not indicate whether their unstable housing experience occurred with or without their family. Adolescents in this study did not report their sexual orientation, and thus we could not explore whether a known factor for experiencing unstable housing, LGBTQ status (Mountz, 2011), was a risk factor. Social expectations might have affected the adolescents’ reports of delinquency behaviors and marijuana use. Although information about other substances was included in the survey, the frequencies of use of substances other than marijuana were not high enough to be tested. Adolescents reported the type of caregiver they were living with but did not report whether they were in formal foster care. The inability to determine the foster care status of participants made it more difficult to compare our findings to studies that examined unstable housing after or between foster care placements. Additionally, we did not have human subjects’ approval to abstract child welfare records during the course of the study, thus we only had maltreatment data up until their enrollment in the study (T1).

Conclusions

This study found that even one incident of unstable housing during adolescence was associated with subsequent delinquency in child welfare-affiliated youth. Unstable housing, including homelessness, during childhood and adolescence is receiving more attention nationally. Housing is one of the social determinants of health and supporting a secure home environment is one strategy to ensure better health and educational opportunities for youth (Wahowiak, 2016). Obtaining a housing history for children who are part of the child welfare system could provide important information for the development of interventions for these populations (Fowler et al., 2017). The federal government has developed a voucher system to support families with rental assistance in the private market, although only one fourth of eligible families are receiving vouchers (Fischer, 2015). The Office of the Assistant Secretary for Planning and Evaluation in the U.S. Department of Health and Human Services commissioned an evaluation of 14 family homelessness prevention and treatment programs that combined human services with housing supports such as housing vouchers (Cortes, Dunton, Henry, Rolston, & Khadduri, 2012). In the evaluation, no programs reported that they targeted youth or families with high residential mobility or coordinated with CWDs. Some youth in this study who had an unstable housing experience might not benefit from family programs because they were already not living with their families. Some of the most promising models for decreasing youth homelessness are rapid rehousing programs for youth and young adults, which include moving youth quickly into supportive housing, providing financial assistance, and offering developmentally appropriate case management (U.S. Department of Housing and Urban Development, 2012). These programs also need to target youth with high residential mobility and be connected to CWDs. It is especially important that housing programs target both child welfare-affiliated maltreated youth who remain at home with their parents as well as those who have non-parent caregivers. More research is needed to identify what types of events within the youths’ home environment preceded an unstable housing experience. In addition, comparisons between unstable housing experiences in youth with a foster care experience and maltreated youth who remained in their home of origin are needed.

Acknowledgments:

The authors want to acknowledge the National Institutes of Health for three grants that supported this research: K01- HD069457 (PI: Negriff); RO1-HD39129 (PI: Trickett); R01- DA02456902 (PI: Trickett). The authors would also like to acknowledge Penelope K. Trickett, PhD for her conceptualization of this study.

We acknowledge that this submission has not been published previously and is not being considered for publication elsewhere.

Footnotes

Disclosures: None

Contributor Information

Janet U. Schneiderman, Suzanne Dworak-Peck School of Social Work, University of Southern California, Department of Children, Youth, and Families and Department of Nursing, 669 West 34th Street, Los Angeles, CA, 90089, 213-821-1338.

Andrea K. Kennedy, Email: andrea.kennedy@usc.edu, Suzanne Dworak-Peck School of Social Work, University of Southern California, Department of Children, Youth, and Families.

Theresa A. Granger, Email: tgranger@usc.edu, Suzanne Dworak-Peck School of Social Work, University of Southern California, Department of Nursing.

Sonya Negriff, Email: negriff@usc.edu, Suzanne Dworak-Peck School of Social Work, University of Southern California, Department of Children, Youth, and Families.

References

  1. Clark A (2013). Is High Residential Mobility Indicative of Housing Instability? Retrieved from Population Studies Center at The University of Michigan: file:///C:/Users/juschnei/Downloads/PAA2016_submissiondraft_092515_AnneClark%20(1).pdf [Google Scholar]
  2. Collins ME, & Curtis M (2011). Conceptualizing housing careers for vulnerable youth: Implications for research and policy. American Journal of Orthopsychiatry, 81, 390–400. doi: 10.1111/j.1939-0025.2011.01107.x [DOI] [PubMed] [Google Scholar]
  3. Corneil TA, Kuyper LM, Shoveller J, Hogg RS, Li K, Spittal PM, … Wood E (2006). Unstable housing, associated risk behaviour, and increased risk for HIV infection among injection drug users. Health & Place, 12, 79–85. doi: 10.1016/j.healthplace.2004.10.004 [DOI] [PubMed] [Google Scholar]
  4. Cox R, Henwood B, Wenzel SL, & Rice E (2016). Roadmap to a unified measure of housing insecurity University of Southern California Center for Economic and Social Research-Schaeffer Working Paper No. 2016–013 Retrieved from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2817626 [Google Scholar]
  5. Cortes A, Dunton L, Henry M, Rolston H, & Khadduri J (2012). Linking human services and housing assistance for homeless families and families at risk of homelessness Retrieved from https://aspe.hhs.gov/sites/default/files/pdf/76456/index.pdf [Google Scholar]
  6. Cunningham M, Pergamit M, Baum A, & Luna J (2015). Helping families involved in the child welfare system achieve housing stability Washington, DC: The Urban Institute; Retrieved from https://www.urban.org/sites/default/files/publication/41621/2000105-Helping-Families-Involved-in-the-Child-Welfare-System-Achieve-Housing-Stability.pdf [Google Scholar]
  7. Department of Children and Family Services. (n.d.) DCFS Glossary Retrieved from http://www.lacdcfs.org/aboutus/dcfsglossary/pagec.htm [Google Scholar]
  8. Dickson-Gomez J, Hilario H, Convey M, Corbett AM, Weeks M, & Martinez M (2009). The relationship between housing status and HIV risk among active drug users: a qualitative analysis. Substance Use & Misuse, 44, 139–162. doi: 10.1080/10826080802344823 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Dworsky A (2014). Families at the nexus of housing and child welfare Washington, DC: First Focus; Retrieved from http://www.chhs.ca.gov/Child%20Welfare/PASS-Families-at-the-Nexus-of-Housing-and-Child-Welfare%20(1).pdf [Google Scholar]
  10. Dworsky A, & Courtney ME (2009). Homelessness and the transition from foster care to adulthood. Child Welfare, 88(4), 23. [PubMed] [Google Scholar]
  11. Dworsky A, Napolitano L, & Courtney M (2013). Homelessness during the transition from foster care to adulthood. American Journal of Public Health, 103, S318–S323. doi: 10.2105/AJPH.2013.301455 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Eastwood EA, & Birnbaum JM (2007). Physical and sexual abuse and unstable housing among adolescents with HIV. AIDS and Behavior, 11(2), 116. [DOI] [PubMed] [Google Scholar]
  13. Ferguson KM (2009). Exploring family environment characteristics and multiple abuse experiences among homeless youth. Journal of Interpersonal Violence, 24, 1875–1891. [DOI] [PubMed] [Google Scholar]
  14. Fischer W (2015). Research shows housing vouchers reduce hardship and provide platform for long-term gains among children Retrieved from Center on Budget and Policy Priorities website: http://www.cbpp.org/sites/default/files/atoms/files/3-10-14hous.pdf [Google Scholar]
  15. Fowler PJ, Marcal KE, Zhang J, Day O, & Landsverk J (2017). Homelessness and aging out of foster care: A national comparison of child welfare-involved adolescents. Children and Youth Services Review, 77, 27–33. doi: 10.1016/j.childyouth.2017.03.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fowler PJ, Toro PA, & Miles BW (2009). Pathways to and from homelessness and associated psychosocial outcomes among adolescents leaving the foster care system. American Journal of Public Health, 99(8), 1453–1458. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Frederick TJ, Chwalek M, Hughes J, Karabanow J, & Kidd S (2014). How stable is stable? Defining and measuring housing stability. Journal of Community Psychology, 42(8), 964–979. doi: 10.1002/jcop.21665 [DOI] [Google Scholar]
  18. Furr-Holden CDM, Lee MH, Milam AJ, Johnson RM, Lee KS, & Ialongo NS (2011). The growth of neighborhood disorder and marijuana use among urban adolescents: A case for policy and environmental interventions. Journal of Studies on Alcohol and Drugs, 72(3), 371–379. doi: 10.15288/jsad.2011.72.371 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Grant R, Gracy D, Goldsmith G, Shapiro A, & Redlener IE (2013). Twenty-five years of child and family homelessness: where are we now? American Journal of Public Health, 103(S2), e1–e10. doi: 10.2105/AJPH.2013.301618 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Huang H, Ryan JP, & Rhoden M-A (2016). Foster care, geographic neighborhood change, and the risk of delinquency. Children and Youth Services Review, 65, 32–41. doi: 10.1016/j.childyouth.2016.03.019 [DOI] [Google Scholar]
  21. Huizinga D, & Elliott DS (1986). Reassessing the reliability and validity of self-report delinquency measures. Journal of Quantitative Criminology, 2, 293–327. doi: 10.1007/BF01064258 [DOI] [Google Scholar]
  22. Kushel MB, Yen IH, Gee L, & Courtney ME (2007). Homelessness and health care access after emancipation: Results from the Midwest Evaluation of Adult Functioning of Former Foster Youth. Archives of Pediatrics & Adolescent Medicine, 161, 986–993. doi: 10.1001/archpedi.161.10.986 [DOI] [PubMed] [Google Scholar]
  23. Landale NS, Oropesa RS, & Bradatan C (2006). Hispanic families in the United States: Family structure and process in an era of family change. In Tienda M & Mitchell F (Eds.), Hispanics and the future of America (pp. 138–178). Washington, DC: National Academies Press. [PubMed] [Google Scholar]
  24. Lee S-Y, & Villagrana M (2015). Differences in risk and protective factors between crossover and non-crossover youth in juvenile justice. Children and Youth Services Review, 58, 18–27. doi: 10.1016/j.childyouth.2015.09.001 [DOI] [Google Scholar]
  25. Los Angeles County Department of Children and Family Services (2019). Fact sheet. Retrieved from C:\Users\juschnei\Downloads\DCFS_Factsheet_Summary_FY_CY.mht [Google Scholar]
  26. Mallett S, & Rosenthal D (2009). Physically violent mothers are a reason for young people’s leaving home. Journal of Interpersonal Violence, 24, 1165–1174. [DOI] [PubMed] [Google Scholar]
  27. Migration Policy Institute. (2016). State immigration data profiles Retrieved from http://www.migrationpolicy.org/programs/data-hub/state-immigration-data-profiles
  28. Moffitt TE, & Caspi A (2001). Childhood predictors differentiate life-course persistent and adolescence-limited antisocial pathways among males and females. Development and Psychopathology, 13(2), 355–375. [DOI] [PubMed] [Google Scholar]
  29. Mountz S (2011). Revolving doors: LGBTQ youth at the interface of the child welfare and juvenile justice systems. LGBTQ Policy Journal at the Harvard Kennedy School, 2011 Edition. [Google Scholar]
  30. Negriff S, Schneiderman JU & Trickett PK (2016). Concordance between self-reported childhood maltreatment versus case record reviews for child welfare-affiliated adolescents: Prevalence rates and associations with outcomes. Child Maltreatment 1–11. doi: 10.1177/1077559516674596 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Population Research Institute. (2015). Hispanic family size in USA shrinking Retrieved from https://www.pop.org/content/hispanic-family-size-usa-shrinking [Google Scholar]
  32. Ringeisen H, Casanueva CE, Urato M, & Stambaugh LF (2009). Mental health service use during the transition to adulthood for adolescents reported to the child welfare system. Psychiatric Services, 60(8), 1084–1091. [DOI] [PubMed] [Google Scholar]
  33. Ryan JP, Perron BE, & Huang H (2016). Child welfare and the transition to adulthood: Investigating placement status and subsequent arrests. Journal of Youth and Adolescence, 45, 172–182. doi: 10.1007/s10964-015-0305-1 [DOI] [PubMed] [Google Scholar]
  34. Schneiderman JU, Kennedy AK, Negriff S, Jones J, & Trickett PK (2016). Maltreated and comparison adolescents’ recollections of lifetime residences: Relationship to delinquency and marijuana use. Journal of Child and Family Studies 25(12), 3481–3487. doi: 10.1007/s10826-016-0506-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Shelton KH, Taylor PJ, Bonner A, & van den Bree M (2009). Risk factors for homelessness: Evidence from a population-based study. Psychiatric Services, 60, 465–472. doi: 10.1176/ps.2009.60.4.465 [DOI] [PubMed] [Google Scholar]
  36. Smith T, Hawke L, Chaim G, & Henderson J (2017). Housing instability and concurrent substance use and mental health concerns: An examination of Canadian youth. Journal of the Canadian Academy of Child and Adolescent Psychiatry, 26, 214–223. [PMC free article] [PubMed] [Google Scholar]
  37. Swartz TT (2009). Intergenerational family relations in adulthood: Patterns, variations, and implications in the contemporary United States. Annual Review of Sociology, 35, 191–212. doi: 10.1146/annurev.soc.34.040507.134615 [DOI] [Google Scholar]
  38. Thompson J (2017). Urban youth homelessness: The role that past trauma has upon housing instability Retrieved from https://sophia.stkate.edu/cgi/viewcontent.cgi?referer=https://scholar-google-com.libproxy1.usc.edu/&httpsredir=1&article=1802&context=msw_papers [Google Scholar]
  39. Tyler KA, & Cauce AM (2002). Perpetrators of early physical and sexual abuse among homeless and runaway adolescents. Child Abuse & Neglect, 26, 1261–1274. doi:10.1016/S0145-2134(02)00413-1 [DOI] [PubMed] [Google Scholar]
  40. U.S. Department of Housing and Urban Development. (2012). Rapid rehousing models for homeless youth Retrieved from https://www.hudexchange.info/homelessness-assistance/resources-for-homeless-youth/rrh-models-for-homeless-youth/ [Google Scholar]
  41. Wahowiak L (2016). Healthy, safe housing linked to healthier, longer lives: Housing a social determinant of health. Nation’s Health, 46(7), 1–19. Retrieved from http://thenationshealth.aphapublications.org/content/46/7/1.3.full [Google Scholar]
  42. Weir BW, Bard RS, O’Brien K, Casciato CJ, & Stark MJ (2007). Uncovering patterns of HIV risk through multiple housing measures. AIDS and Behavior, 11(2), 31–44. doi: 10.1007/s10461-007-9284 [DOI] [PubMed] [Google Scholar]

RESOURCES