| A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z | AA | AB | AC | AD | AE | AF | AG | AH | AI | AJ | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | Master Tracker — Family Violence & Housing Insecurity Review | |||||||||||||||||||||||||||||||||||
2 | Each row = one article. Fill in fields from the Data Extraction sheet. | |||||||||||||||||||||||||||||||||||
3 | Study ID | Title | First Author | Year | Journal | Type of Study | Research Design | Analysis Method(s) | Sample Size | Sample Type | Recruitment Method | Geographic Location | Adult Age | Adult Gender/Sex | Adult Race/Ethnicity | Child Age | Child Gender/Sex | Child Race/Ethnicity | SES/Socioeconomic Details | IPV Measured? | IPV Tool(s) | IPV Inclusion Criteria | Type of IPV | Child Maltreatment Measured? | Maltreatment Tool(s) | Type of Maltreatment | Housing Instability Type | Housing Instability Tool(s) | Housing instability inclusion criteria | Main Findings | Sig. Findings | Non-Sig. Findings | Intervention? | Intervention Name | Stated Limitations | Notes |
4 | Associations Between Intimate Partner Violence and Increased Economic Insecurity Among Women and Transgender Adults During the COVID-19 Pandemic | Fedina, Lisa | 2023 | Psychology of Violence | Quantitative; cross-sectional survey | Online survey measuring IPV and economic insecurity during stay-at-home orders and 3 months prior | Chi-square tests and binary logistic regression | 1,169 complete responses | General population sample of women and transgender/nonbinary individuals | Qualtrics Panels using online quota sampling | Michigan (Midwestern U.S.) | Age groups: 18–34 (42.3%), 34–44 (20.1%), and 44+ (37.6%). Mean not provided. | White (66%), Black (15.8%), Latinx (5%), Asian (5.6%), Multiracial (4.8%), and Other (2.5%) | NR | NR | Income: <$39,999 (36.9%), 40k–79,999 (35.5%), >$80k (23.6%) | Yes | 20 dichotomous behavioral items modified from the NISVS and Danger Assessment | Self-report endorsement of physical, sexual, psychological, or technology-based items | Physical, sexual, psychological, and technology-facilitated abuse | No | NA | NA | Housing insecurity (failure to pay full rent or mortgage because of inability to afford it) | Single self-report yes/no item | Any month since COVID-19 where the full amount was not paid | IPV victimization during stay-at-home orders was associated with three times higher odds of housing insecurity. Multiracial and pregnant individuals were at elevated risk | IPV during stay-at-home orders significantly predicted housing insecurity (AOR = 3.12, 95% CI [1.81, 5.63], p < .001) | IPV immediately prior to the pandemic was not significantly associated with housing insecurity in the adjusted multivariate model (AOR = 1.56, 95% CI [0.89, 2.74]) | No | NA | Cross-sectional data prevents causal investigation; susceptibility to recall bias; limited to English-speaking populations | ||||
5 | An Exploratory Analysis of Domestic and Intimate Partner Violence Victimization Among Persons Experiencing Eviction | Benson-Goldsmith, Michaela E. | 2025 | Violence Against Women | Quantitative; observational, retrospective analysis of administrative court records | Retrospective analysis of court interaction life histories (2017–2019) for a random sample of evicted adults | t-tests, chi-square tests, bivariate logistic regression, and multivariate binomial logistic regression | 306 evicted adults | Adults who experienced a formal court-ordered eviction | Public records request for all eviction cases in Douglas County, Nebraska; a simple random sample was then extracted | Douglas County, Nebraska; Midwestern U.S | Mean: 36 years (SD 12.59); age was missing for 33.01% of the sample | Black (44.12%), White (43.46%), and Other race/ethnicity (12.42%) | NR | NR | Identified as low-income/experiencing poverty due to being in the "poverty trap" of formal eviction | Yes | A history of DV/IPV victimization variable was created for each subject as a dichotomous value (1= yes, 0= no). Subjects were coded 1 for having a history of victimization if they had at least one civil court case classified as a protection order case prior to their eviction case and in which the subject was listed as petitioner or plaintiff. | At least one civil court case classified as a protection order case prior to the primary eviction | IPV as defined by Nebraska law via formal protection orders | No | NA | NA | Eviction, Previous evictions, reason for eviction | State administrative court records (Nebraska.gov judicial case system); dichotomous variable was made for prior civil court contacts related to prior evictions. | Subjects who underwent a formal, court-ordered removal (dismissed cases were excluded) | 22.5% of evicted adults had a history of DV/IPV victimization. 83.8% of these survivors experienced their first DV/IPV court contact before their first eviction | DV/IPV survivors were significantly more likely to have previous eviction cases (59.4% vs. 40.9%; χ 2 =7.38, p<.01) | Previous eviction case was not significantly associated with a court history of DV/IPV victimization (OR: 0.73 [0.36, 1.50], p not reported) | No | NA | Single U.S. county; excludes "informally" evicted persons; measuring IPV via formal court contacts omits unreported abuse; missing demographic data in administrative records | ||||
6 | Housing Insecurity Among College Students: Associations with Interpersonal Violence Victimization | Johnson, Laura | 2026 | Journal of Interpersonal Violence | Quantitative; cross-sectional campus climate survey | Rotating survey modules assign participants to either a sexual violence (SV) or dating violence (DV) module | Bivariate and multivariate logistic regression | SV module (n = 2,672); DV module (n = 1,645) | Community sample of college students | Email invitation to all students; tiered incentive structure (raffles) | Large, urban public university in the Northeast | Mean age: 24.2–24.5 years | SV sample: White (30.4%), Latine/Hispanic (25.9%), Asian/Asian American (20.9%), Black (17.2%), and Other (5.7%); DV sample: | NR | NR | 73.4%–76.6% of students received financial aid | Yes (Dating violence) | Modified Measure of Adolescent Relationship Harassment and Abuse (MARSHA) | Students romantically involved since coming to college | Physical and psychological dating violence | No | NA | NA | Housing insecurity (eviction, thrown out, shelter, abandoned building/car, or unsure where to sleep) | five items adapted from Goldrick-Rab et al. (2017) | Any occurrence in the past 12 months | DV survivors (physical or psychological) have four times greater odds of housing insecurity than students who have not experienced DV while in college. | Physical DV predicted housing insecurity (OR = 3.62, 95% CI [2.43, 5.40], p < .001). Psychological DV predicted housing insecurity (OR = 3.72, 95% CI [2.50, 5.53], p < .001) | None. | No | NA | Low response rate to recruitment email (11.1%); randomization prevented looking at SV/DV overlap; cross-sectional design | ||||
7 | Childcare Subsidies and Reduced Intimate Partner Violence Victimization: An Unintended Benefit | Marçal, Katherine | 2026 | Journal of Interpersonal Violence | Quantitative; longitudinal birth cohort study (secondary analysis) | Data from the Future of Families and Child Wellbeing Study (FFCW), Year 3 interview | Confirmatory Factor Analysis (CFA) and Path analysis with latent variables | 1,987 mothers | Low-income mothers who were income-eligible for subsidies and in an intimate relationship | Longitudinal birth cohort sampled from hospitals in large cities in the United States | 20 large American cities | Mean: 24.05 years (SD 0.12) at the time of the child's birth | Black (48.11%), Hispanic/Latina (34.73%), White (14.29%), and Other (2.87%) | Approximately 3 years old | NR | Income-eligible for federal childcare subsidies (low-income) | Yes | 6 Likert-type items assessing frequency of current abusive behaviors | Currently in an intimate relationship at the year 3 interview | Emotional abuse, physical/sexual violence, and coercive control | No | NA | NA | Housing insecurity (missed payments, eviction, doubling up, or literal homelessness) | Summed score of five dichotomous adverse housing experiences in the past year | Any reported adverse housing event due to inability to afford costs | Housing insecurity was significantly associated with increased IPV victimization (β=.15,p<.001) | Housing insecurity was significantly associated with increased IPV victimization (β=.15,p<.001) | None. | Yes | Child Care Development Fund (Government childcare subsidies) | Oversampled unmarried parents; cross-sectional design for the Year 3 analysis; reliance on self-reported data | ||||
8 | Prevalence and Correlates of Violence Experienced by Trans Women | Jackson, Akira | 2022 | Journal of Women's Health | Quantitative; respondent-driven sampling (RDS) study | Part of the National HIV Behavioral Surveillance (NHBS) using interviewer-administered surveys | prevalence ratio (PR) tests | 201 trans women | Transgender women residing in San Francisco | Respondent-driven sampling (RDS) with 15 initial seeds | San Francisco, California | Age groups: 18–29 (13.9%), 30–39 (20.9%), 40–49 (25.9%), and 50+ (39.3%) | Hispanic/Latinx (37.3%), Black (20.9%), White (17.9%), Multiple/Other (14.9%), and Asian/PI (9.0%) | NA | NA | 57.7% earned below the federal poverty level | Yes | Interviewer-administered survey assessing the past 12 months | Any occurrence in the past 12 months | Intimate partner violence | No | NA | NA | Unstable housing (SRO hotels, emergency shelters, transitional housing, or outdoors) | Interviewer-administered survey on current and past year housing status | Classification of currently homeless or living in temporary accommodations | Unstable housing was associated with higher IPV (PR = 4.40, 95% CI [1.38, 14.01], p = 0.012). Current homelessness was associated with higher IPV (PR = 2.38, 95% CI [1.25, 4.55], p = 0.009) | None. | No | NA | RDS tends to reach those with lower income; social desirability bias in interviewer-administered surveys; causality cannot be established | |||||
9 | https://doi.org/10.3390/ children10040707 | Father–Mother Co-Involvement in Child Maltreatment: Associations of Prior Perpetration, Parental Substance Use, Parental Medical Conditions, Inadequate Housing, and Intimate Partner Violence with Different Maltreatment Types | Lee, Joyce Y. | 2023 | Children | Quantitative cross-sectional analysis of national child welfare administrative data | Retrospective observational study using secondary administrative data (NCANDS Child File (FY 2018) for 11 states providing complete risk factor data) | Logistic regression analysis using maximum likelihood estimation with robust standard errors | 6996 triads (co-perpretating mother, co-perpetrating father, and victimized child) | Child welfare-involved families with substantiated maltreatment leading to foster care entry | Case-level records extracted from the federal NCANDS database | 11 U.S. states (AR, DE, FL, IN, MI, MN, MS, NE, SD, TX, and UT) | Mothers: 30.14 years (SD 7.07); Fathers: 33.79 years (SD 8.89) | Over half of the couples were Non-Hispanic/Latinx White (58%), followed by Other race (18%), Non-Hispanic/Latinx Black (14%), and Hispanic/Latinx (9%) | Mean: 2.96 years (SD 4.14) | Female (48.21%) and Male (51.79%) | 52.04% experienced economic insecurity (financial problems or public assistance receipt) | Yes | IPV was one binary variable (yes/no) for any abusive or coercive act by one parent on another | Recorded in state child protective service response files for substantiated cases | Defined within NCANDS as any abusive, violent, coercive, forceful, or threatening act or word inflicted by one parent on another | Yes; both prior perpetration of child maltreatment (IV) and type of child maltreatment (DV were measured | Prior perpretation of child maltreatment: Two binary variables (yes/no)—one for mothers and another for fathers; Type of child maltreatment: Binary variables (yes/no) were created for each of the four child maltreatment types. | Prior perpetration of child maltreatment: According to NCANDS, a prior abuser or perpetrator is defined as an individual who caused or knowingly allowed child maltreatment to occur with a previous determination (i.e., occurred before the current child maltreatment reportdisposition date) in the state’s information system of substantiated or indicated child maltreatment; Type of child maltreatment: (1) physical abuse; (2) neglect; (3) sexual abuse; (4) emotional abuse | one binary variable (yes/no) | Defined within NCANDS as having housing facilities that are substandard, overcrowded, unsafe, or otherwise inadequate for living with the child (e.g., homelessness) | Facilities substandard or inadequate for living with the child | Inadequate housing was associated with higher odds of neglect but lower odds of physical abuse | Descriptive statistics: Intimate partner violence significantly positively associated with inadequate housing (.08, p < .001). Prior history of child maltretment significantly negatively associated with inadequate housing (-0.04, p < .01). Child physical abuse (0.06, p< .001) and child sexual abuse (-0.04, p < .01) significantly associated with inadequate housing. Logistic Regressions: Inadequate housing significantly associated with .37 lower odds of father-mother co-involved child physical abuse (AOR: 0.63 [0.48,0.81], p < .001) and 2.29 greater odds of father-mother co-involved child neglect (AOR: 2.29 [1.44,3.64], p < .001) | Descriptive statistics: Neglect (0.00, p not reported) and emotional abuse (0.02, p not reported) not significantly associated with inadequate housing. Logistic Regressions: Inadequate housing not significantly associated with father-mother co-involved sexual abuse (AOR: 0.56 [0.25,1.26], p = 0.162) or emotional abuse (AOR: 0.75 [0.39,1.43], p =0.377) | No | NA | Cross-sectional, so inability to make causal conclusions. Analytic sample only includes biological parents. Unable to examine specific maternal and paternal risk factors contributing to co-involvement in child maltreatment. Unable to include measures that capture social context as a potential risk factor. | |||
10 | Sexual Relationship Violence Among Young, Partnered Sexual Minority Men in the United States | Stephenson, Rob | 2023 | Journal of Interpersonal Violence | Quantitative; Cross-sectional | Secondary analysis of baseline data from a pilot RCT (We Prevent) | Multivariable logistic regression; Independent samples t-tests | 318 young sexual minority men | Young sexual minority men (YSMM) in relationships with men | Social media advertisements (Facebook, Instagram, Jack’d) | United States | Mean 20.50 (SD 2.37), range 15–25 | 70.4% White, 12.3% Multiracial, 7.9% Black | NA | NA | NA | 28.9% food insecurity; 7.2% unstable housing | Yes | Abbreviated IPV-GBM Scale, adapted from the Conflict Tactics Scale | Self-report of any experience in their current partnership | Physical, emotional, monitoring/controlling, and sexual | No | NA | NA | Unstable housing (street, car, shelter) | Self-report question on housing status in past 30 days | Dichotomous yes/no based on any unstable housing | Unstable housing was the strongest predictor of sexual IPV and any IPV. | Unstable housing was significantly associated with sexual IPV: aOR = 7.22, 95% CI: 1.66, 31.34, p<0.01. Unstable housing was significantly associated with any IPV: aOR = 6.06, 95% CI: 1.92, 19.08, p<0.01. | None. | Yes | We Prevent | Underreporting risk; convenience sampling (online only); recall bias; non-dyadic | |||
11 | https://doi.org/10.1177/08862605231179720 | Violence Victimization, Homelessness, and Severe Mental Illness Among People Who Use Opioids in Three US Cities | Hong, Chenglin | 2023 | Journal of Interpersonal Violence | Quantitative; Cross-sectional | Tablet-based survey administered in-person at syringe service programs (SSPs) | Multivariable logistic regression | 1,127 people who use opioids | People who use opioids | Recruitment at SSP sites in three cities | Los Angeles, Las Vegas, and Atlanta | Mean 41.2 (SD 11.8), range 19–68 | 62.6% Male; 36.0% Female; 1.4% Other | 42.4% White, 20.6% Hispanic, 20.1% Black | NA | NA | NA | 68.3% income ≤ $20,000; 54.2% unemployed | Yes | IPV was measured by asking, “Have you been the victim of any forms of IPV (e.g. violence committed by a spouse, partner, boy/girlfriend, etc.) against you?” | Reported "yes" to having been a victim of violence by a spouse/partner | Intimate partner violence (as part of broader victimization assessment) | No | NA | NA | Homelessness or housing insecurity | Self-report (living on street, car, or shelter); Participants were asked, “Are you currently homeless or housing insecure (living on the street, in a car, at a homeless shelter)?” | Categorized as "currently homeless/housing insecure" | Recent IPV victimization alone showed non-significant differences by housing status in bivariate chi-square tests (p=.313) | None. | Recent IPV victimization alone showed non-significant differences by housing status in bivariate chi-square tests (p=.313) | No | NA | Reliance on three U.S. cities; cross-sectional (no causality); self-report bias | |
12 | https://doi.org/10.1177/08862605221140038 | Prevalence and Correlates of Non-Dating Sexual Violence, Sexual Dating Violence, and Physical Dating Violence Victimization among US High School Students during the COVID-19 Pandemic: Adolescent Behaviors and Experiences Survey, United States, 2021 | Krause, Kathleen H. | 2023 | Journal of Interpersonal Violence | Quantitative; Cross-sectional; Nationally representative | Adolescent Behaviors and Experiences Survey (ABES) - web-based survey conducted by CDC | Multivariable logistic regression | 7,705 high school students | U.S. students in grades 9 to 12 | Three-stage cluster sampling approach | United States (National) | NA | NA | NA | NR | 51.9% Male / 48.1% Female | CHECK | 28.5% reported parent job loss; 23.8% experienced hunger during the pandemic | Yes (Dating violence) | 3 survey items on forced sexual acts and intentional physical harm by a dating partner | Self-report of ≥ 1 experience in the past 12 months | Sexual dating violence and physical dating violence victimization | Yes | ABES survey items on emotional and physical abuse by a parent or other adult in the home; During the COVID-19 pandemic, how often did a parent or other adult in your home swear at you, insult you, or put you down? During the COVID-19 pandemic, how often did a parent or other adult in your home hit, beat, kick, or physically hurt you in any way? | Emotional and physical abuse by a parent | Homelessness (where student slept in past 30 days) | ABES survey items; "During the past 30 days, where did you usually sleep?" | Self-report of sleeping in friends' homes, shelters, motels, cars, or other public places | Homelessness was significantly associated with being a victim of physical dating violence among male students | Males: Homelessness significantly increased the likelihood of being a victim of physical dating violence (OR = 7.22, 95% CI: 2.31, 22.52, p<0.05) | Males: Homelessness did not significantly affect the likelihood of being a victim of sexual dating violence (OR = 4.88, 95% CI: 0.77, 30.80, p not reported). Females: Homelessness did not significantly affect the likelihood of being a victim of sexual (OR = 0.92, 95% CI: 0.27, 3.10, p not reported) or dating violence (OR = 1.15, 95% CI: 0.40, 3.27, p not reported) . | No | NA | Cross-sectional design (no causality); outcomes not mutually exclusive; snap-shot of a 6-month period; recall bias | |
13 | https://doi.org/10.1177/21501319221113543 | Clustering of Social Determinants of Health Among Patients | Schiltz, Nicholas K. | 2022 | Journal of Primary Care & Community Health | Quantitative; Cross-sectional | Secondary analysis of health system social needs screening data | Association rule mining (machine learning) and log-binomial regression | 23,161 patients | Safety-net health system patients | Systematic screening (in-person, telephone, or online) | Cleveland, Ohio | Mean 51.0 (SD 18.2) | 69.1% female | 66.1% White, 25.4% Black, 6.1% Hispanic | NA | NA | NA | 31.9% reported income < $25,000 | Yes | Validated survey questions from standard screening tools (Chagin et al., 2021) | Pre-defined criteria for being at risk for each social determinant were used to categorize questionnaire responses. For example, patients who answered often or sometimes to either of the 2 food security questions were categorized as being food insecure. | General Intimate partner violence | No | NA | NA | Inability to pay for housing or utilities | Validated survey questions from standard screening tools (Chagin et al., 2021) | Pre-defined criteria for being at risk for each social determinant were used to categorize questionnaire responses. For example, patients who answered often or sometimes to either of the 2 food security questions were categorized as being food insecure. | IPV did not meet the 2% prevalence threshold to cluster directly with housing instability dyads | None. | IPV did not meet the 2% prevalence threshold to cluster directly with housing instability dyads | No | NA | Single health system focus, reliance on self-report, and pandemic-influenced social determinants | |
14 | https://doi.org/10.1111/phn.12819 | The relationship between housing instability and intimate partner violence: A retrospective study | Wilson, Patty R. | 2021 | Public Health Nursing | Quantitative; Retrospective secondary data analysis; Longitudinal | Fragile Families and Child Well-Being Study data; baseline, 1-year, 3-year, and 5-year follow-ups | Logistic regression; Chi-square tests | 4,000 women | Mothers from large-scale U.S. dataset | Recruited in hospitals after birth | 20 U.S. cities | 36.2% aged 20–24; 22.8% over 30 | 100% female | Black 50.8%, Hispanic 27.1%, White 22.2%. | NA | NA | NA | Low income; 69% earned < $10,000/year; 38.4% < High school education | Yes | Binary variable constructed from FFCWB items | Categorized as "IPV over time" if reported at baseline, Year 1, and Year 5 | General Intimate partner violence | No | NA | NA | Moving 2+ times, shelter stays, doubling up for financial reasons, eviction | Binary variable constructed from FFCWB items | Reported housing status at Year 5 | Women experiencing housing instability were significantly more likely to experience IPV over time. | Housing instability associated with IPV over time: OR = 2.36, 95% CI: 1.22–4.54, p<0.05. Women with early HI (year 1and year 3 only) reported a significant relationship between HI andIPV over time when individual factors were included in the model(OR 2.23, 95% CI [1.21–4.13]) but not when the situational and socialstructural factors were included.An association between HI and IPV over time was found in Blackwomen when situational factors (OR 0.62, 95% CI [0.44–0.87]) wasincluded in the model. However, an association was not indicatedbetween HI and IPV over time when social structural factors (OR0.66, 95% CI [0.44–1.00]) were included in the model. The follow-ing women reported a relationship between HI and IPV over time:had less than a high school education (OR 1.64, 95% CI [1.15–2.36]),did not receive support from others (OR 1.92, 95% CI [1.12–3.30]),had economic hardship (OR 2.04, 95% CI [1.46–2.87]), and did notreceive formal housing assistance (OR 0.66, 95% CI [0.45–0.97]).Additionally, an association between HI and IPV over time was foundin women who were involved with the father of the baby if they livedtogether (OR 0.23, 95% CI [0.17–0.32]). | The interaction term between housing status and race/ethnicitywas not significant (Table 3). This suggests that HI does not have astronger association with IPV over time in Black women comparedto White women. | No | NA | Secondary data analysis (limited variables); didn't examine health outcomes; self-report bias | Reports of IPV and HI together decreased slightly over time from Year 1 to Year 5 |
15 | https://doi.org/10.1007/s10896-025-00846-9 | Pathways to Safety and Housing Stability Among Intimate Partner Violence Survivors Receiving Supportive Housing Services: A Mixed Methods Analysis | Bevilacqua, Kristin G. | 2025 | Journal of Family Violence | Mixed-methods analysis. It utilized a prospective, quasi-experimental design for the quantitative portion and semi-structured interviews for the qualitative portion | prospective evaluation with survey data collected at baseline, 3-month, and 6-month intervals | linear and logistic mixed effects models (stratified by housing program, co-parenting status, and homicide risk) and cross-lagged panel structural equation modeling (SEM) | 70 participants | Intimate partner violence (IPV) survivors receiving supportive housing services (either Rapid Rehousing or Transitional Housing) | Clients of House of Ruth Maryland (HRM), a comprehensive IPV service provider, were invited to participate | Baltimore, Maryland (urban) | Mean 33.11 years (SD = 7.25) | 100% female | 77.1% Black, African American, or African; 8.6% Multiracial; 5.7% Hispanic or Latino; 4.3% Other; 2.9% White; 1.4% Asian | NA | NA | NA | largely low-income; 63.9% reported a 2018 household income between $0 and 20,000;60.9% reported any employment in the 30 days prior to baseline | Yes | short form of the Revised Conflict Tactics Scale (CTS) | Homicide risk was measured using the 20-item Danger Assessment | Recent physical and sexual IPV; homicide risk | No | NA | NA | living in an undesired location, difficulty paying for housing, trouble obtaining housing, trouble with a landlord, and threats or notices of eviction | 10-item Housing Instability Index (with 9 items utilized in the final score) | experiencing homelessness due to IPV and were receiving supportive housing services | Housing support services simultaneously improved housing stability and reduced IPV among survivors by the 6-month follow-up. Economic dependence on an abusive partner remains a significant risk factor for IPV; baseline housing instability was a significant predictor of economic dependence at 6 months | Significant reductions at 6 months were found for IPV (RRH: -2.90, p < 0.001; TH: -2.68, p = 0.016) and housing instability (RRH: -0.73, p = 0.01; TH: -1.89, p = 0.01) | The study did not find a significant longitudinal (cross-lagged) relationship between IPV and housing stability in the SEM model. Reductions in housing instability were not statistically significant for those with variable baseline homicide risk (p = 0.16) and were borderline for those with severe/extreme homicide risk (p = 0.055) | Yes | Transitional Housing and Rapid Rehousing | Lack of an unexposed control arm, small sample size (especially for TH and stratified analyses), relatively short follow-up duration (6 months), geographic limitation to Baltimore, and potential impact of the COVID-19 pandemic on generalizability | Qualitative data revealed that housing acts as a "pathway to safety" by providing physical and emotional distance from the abuser and allowing survivors to manage contact (e.g., for co-parenting) without compromising their safety |
16 | https://doi.org/10.1016/j.ypmed.2021.106950 | Affordable housing through the low-income housing tax credit program and intimate partner violence-related homicide | Austin, Anna E. | 2022 | Preventive Medicine | Quantitative, observational state-level longitudinal study | state-level ecological study using surveillance data from 2005 to 2016; longitudinal approach by analyzing 131 state-year observations across 13 states | generalized linear regression models with generalized estimating equations (GEE) to account for repeated measures within states over time; negative binomial distribution to calculate rate ratios (RRs) and adjusted for multiple time-varying state-level covariates | 131 state-year observations from 13 states; 6,351 IPV-related homicides | State-level population data derived from the National Violent Death Reporting System (NVDRS) | existing public and surveillance databases (NVDRS, HUD LIHTC Property Data, Eviction Lab data) | 13 U.S. states (Colorado, Georgia, Kentucky, Massachusetts, Maryland, New Jersey, New Mexico, North Carolina, Oklahoma, Oregon, South Carolina, Virginia, and Wisconsin) | 66.5% of homicide victims were between the ages of 15 and 44 | 65.2% of homicide victims were women | The study adjusted for state-level percentages of race and ethnicity as covariates | NA | NA | NA | unemployment rate: (means of 6.0%–6.6%); median household income: (mean of ~52,181–54,902) | Yes, IPV-related homicide | Data extracted from the National Violent Death Reporting System (NVDRS) | Homicides were included if an NVDRS abstractor determined the death resulted from conflict or violence between current or former intimate partners | IPV-related homicide (including deaths of non-intimate partners, such as children, if the incident originated in an IPV conflict) | No | NA | NA | lack of affordable housing availability and eviction rates | HUD LIHTC Property Data (affordable units placed in service) and Eviction Lab data (court-ordered evictions) | LIHTC availability was categorized as above or below the median (≥30 vs. <30 units per 100,000 population); Eviction rates were categorized as above or below the median (≥3500 vs. <3500 per 100,000 renter population) | Greater availability of affordable housing (≥30 LIHTC units per 100,000 population) was associated with a lower rate of IPV-related homicide at the state level. The association between affordable housing and lower homicide rates was more pronounced in states with higher eviction rates | In state-years with ≥30 LIHTC units, the rate of IPV-related homicides was 12% lower than in state-years with <30 units (RR = 0.89; 95% CI: 0.81, 0.98). In states with ≥3500 evictions per 100,00 renter population, the rate of IPV-related homicides was 20% lower (RR = 0.83; 95% CI: 0.74, 0.93), highlighting the value of affordable housing units. | In states with <3500 evictions, the relationship between IPV-related homicides and availability of affordable housing units was not significant (RR = 0.91; 95% CI: 0.81, 1.03) | Yes | Low-Income Housing Tax Credit (LIHTC) | cannot prove individual-level associations due to ecological nature of the study; small number of homicides prevent smaller geographic analysis (e.g., county-level); potential for unmeasured confounding; lack of data on whether victims actually resided in LIHTC units | |
17 | https://doi.org/10.1177/08862605211043588 | Intimate Partner Violence Exposure and Adolescent Mental Health Outcomes: The Mediating Role of Housing Insecurity | Marçal, Katherine E. | 2022 | Journal of Interpersonal Violence | Quantitative, longitudinal study | Data from the Fragile Families and Child Well-being Study, specifically from the Years 5, 9, and 15 interviews. It tested prospective, direct, and indirect pathways over a 10-year period | Structural equation modeling (SEM) with latent variables, including confirmatory factor analysis (CFA) and path analysis | 2425 mothers | Socioeconomically diverse mothers who maintained at least partial custody of the study's focal child | Families were recruited in hospitals at the time of the child's birth in 20 large U.S. cities, with an intentional oversample of births to unmarried parents | 20 large American cities across the United States (urban) | NR | 100% female | 51.22% Black, 23.55% Hispanic, 22.19% White, and 3.05% Other | Approximately 5, 9, and 15 years old at the time of the three data collection waves used | 51.88% Male and 48.12% Female | NR | largely low-income; average household income at Year 5 was $38,055 | Yes | Self-report of 11 items covering frequency of physical, psychological, and sexual violence and coercive control (at Year 5) | Experiences of current IPV were assessed via self-report during the Year 5 interview | Physical, psychological, and sexual violence, and coercive control | No | NA | NA | missed rent/mortgage payments, missed utility payments, doubling up (moving in with others), eviction, and literal homelessness | Five dichotomous self-report items (at Year 9) | Participants reported whether they had experienced any of the five indicators in the past 12 months | Childhood exposure to maternal IPV at age 5 was significantly associated with increased housing insecurity at age 9. Housing insecurity mediated the link between IPV exposure and adolescent depression. | IPV (at Year 5) was significantly associated with housing insecurity(at Year 9) (β=0.18,p<0.001). There is an indirect effect of IPV (at year 5) on adolescent depression (at year 15) mediated by housing insecurity (at year 9) (β=0.02,p<0.05). | The indirect effect of IPV on adolescent anxiety via housing insecurity was not statistically significant (β=0.015) | No | NA | urban areas, so may not generalize to rural settings; self-reported data (subject to social desirability bias); exclusion of maternal mental health as a covariate; limited range of adolescent outcomes | |
18 | https://doi.org/10.1016/j.amepre.2018.01.020 | Recent Intimate Partner Violence and Housing Instability Among Women Veterans | Montgomery, Ann Elizabeth | 2018 | American Journal of Preventive Medicine | Quantitative, observational, retrospective analysis of electronic medical record (EMR) data | Retrospective analysis using data extracted from the U.S. Department of Veterans Affairs (VA) Corporate Data Warehouse | Logistic regressions | 8,427 women Veterans | Women Veteran outpatients seeking care at Veterans Health Administration (VHA) facilities | The study utilized administrative data from VHA electronic medical records for patients screened for past-year IPV at 13 VA facilities between April 2014 and April 2016 | 13 VHA facilities across the United States | 18-30(16.9%), 31-40 (26.8%), 41-50 (23.1%), 51+ (33.2%) | 100% female | 52.4% White; 38.9% Black or African American; 3.5% Other or multiple; 5.3% Missing. Ethnicity: 5.6% Hispanic or Latina; 90.1% Not Hispanic or Latina | NA | NA | NA | 68.4% received compensation for a service-connected disability | Yes | Extended-Hurt, Insult, Threaten, Scream (E-HITS) scale | A positive screen was defined as a total score of ≥7 on the E-HITS scale, indicating past-year experience | Physical (being hurt), psychological/emotional (insulted, talked down to, screamed, or cursed at), threats of harm, and sexual violence (forced sexual activities) | No | NA | NA | Homelessness and risk of homelessness, including literal homelessness, imminent risk of instability, and use of transitional or supportive housing programs | ICD-9 (v60) and ICD-10 (z59) diagnostic codes; VHA Homeless Program clinic stop codes (e.g., HUD-VASH, HCHV); and the Homelessness Screening Clinical Reminder (HSCR) | A composite measure defined as having any diagnostic code for housing instability, use of a VHA Homeless Program, or a positive response to the HSCR (indicating homelessness or risk within the past/next 60 days) | IPV was positively associated with every indicator of housing instability measured | Controlling for age and race, screening positive for IPV increased the odds of housing instability by nearly 3 times (AOR = 2.74, 95% CI = 2.26, 3.32). IPV survivors were 4 times as likely to screen positive for homelessness on the HSCR (AOR = 4.05, 95% CI = 2.55, 6.44) and twice as likely to be in the HUD-VASH permanent supportive housing program (AOR = 2.22, 95% CI = 1.70, 2.90). Among women with past-year IPV, housing instability was significantly more likely for those who identified as Black or African American (AOR = 1.84, 95% CI = 1.20, 2.82), screened positive for Military Sexual Trauma (AOR = 1.94, 95% CI = 1.30, 2.90), had a mental health diagnosis (AOR = 2.21, 95% CI = 1.39, 3.49), had a substance use disorder (AOR = 6.96, 95% CI = 3.79, 12.76). Being married (AOR = 0.32, 95% CI = 0.20, 0.50) and receiving service-connected disability compensation (AOR = 0.56, 95% CI = 0.36, 0.88) were associated with significantly lower odds of housing instability among IPV survivors. | Among women Veterans who screened positive for IPV, serving in the OEF/OIF era was not significantly associated with housing instability (AOR = 0.84, 95% CI = 0.53, 1.32). While those aged 41–50 had higher odds of instability than the ≥50 group, the 18–30 (AOR = 1.51, 95% CI = 0.80, 2.83) and 31–40 (AOR = 1.29, 95% CI = 0.73, 2.27) age groups did not reach statistical significance in the adjusted model. | No | NA | Data relies on clinical disclosure, so IPV may be undercounted; housing data is limited to VHA sources and does not capture community-based services or unreported instability; E-HITS lacks nuance regarding the context or severity of the IPV experience; findings may not be generalizable to women Veterans not seeking VHA care | |
19 | https://doi.org/10.1016/j.socscimed.2022.115030 | Eviction, intimate partner violence and HIV: Expanding concepts and assessing the pathways through which sexual partnership dynamics impact health | Groves, Allison K. | 2022 | Social Science & Medicine | Quantitative, longitudinal cohort study | prospective design at baseline and a 6-month follow-up | Bootstrapped path analyses (for mediation); Chi-square and t-tests | 282 adults | Low-income residents; sample deliberately structured to include individuals with a history of criminal legal involvement | flyers, outreach to local service providers, community meetings, and snowball sampling | New Haven, Connecticut (urban) | 44.9 years (SD = 11.5) | 64.5% Male and 35.5% Female | 64.9% Black, 25.5% White, and 9.6% Other | NA | NA | NA | mean monthly income was $1,070; eligibility required receiving public assistance, being homeless, or living in a low-income census tract | Yes | A modified WHO Conflict Tactics Scale consisting of 10 items assessing physical and sexual violence | Self-report of at least one act of violence across sexual partnerships in the six months prior to the follow-up interview | Physical and sexual violence (both victimization and perpretation) | No | NA | NA | Landlord-related forced moves (legal evictions, informal forced moves due to rent hikes, foreclosure, or landlord-initiated pressure) | Binary self-report measure at baseline. Participants were asked if they had been legally evicted in the past two years or if their last move was forced by a landlord for specific reasons. | Experience of a legal eviction or a landlord-related forced move within the two years prior to baseline | Individuals who experienced these forced moves were significantly more likely to subsequently report both IPV victimization and perpetration. IPV perpetration significantly mediated the relationship between forced moves and HIV sexual risk. IPV victimization was a marginal mediator of the relationship between forced moves and HIV sexual risk. | Forced move was significanly associated to IPV victimization (β=0.19, p=.02) and perpretation (β=0.25,p=.003). IPV perpetration significantly mediated the relationship between forced moves and HIV sexual risk (β=0.16,p=.02). | IPV victimization was a marginal mediator of the relationship between forced moves and HIV sexual risk (β=0.09,p=.06). No direct effect was found between landlord-related forced moves and HIV sexual risk in either the victimization model (β=0.01,p=.87) or the perpetration model (β=−0.05,p=.55) | No | NA | Use of a convenience sample (limiting generalizability), lack of statistical power for racial/gender sub-group analyses, and a broad reference period (2 years) for forced moves that may have led to an underestimation of the effects | |
20 | https:doi.org/10.1007/s10461-022-03705-6 | Abuse in the Continua: HIV Prevention and Care Outcomes and Syndemic Conditions Associated with Intimate Partner Violence Among Black Gay and Bisexual Men in the Southern United States | Wirtz, Andrea | 2022 | AIDS and Behavior | Quantitative; Cross-sectional surveys | Two independent rounds of cross-sectional surveys conducted in two cities | chi-squared and t-tests, Poisson regression models with robust variance estimation to calculate prevalence ratios (PrR), and multivariable regression | 629 men | Black gay and bisexual men | Multi-pronged convenience sampling including referrals from community organizations, peer referrals (some incentivized), flyers at social/health venues, social media outreach, and presence at LGBTQ events | Baltimore, Maryland, and Jackson, Mississippi | Median age = 29 years (IQR: 25–38) | Male (assigned male sex-at-birth and identifying as men) | 96.7% Black/African American only; 3.3% multi-racial (including Black/African American) | NA | NA | NA | High levels of poverty; 62.2% of the total sample reported income below the federal poverty line . 43.5% were employed full-time, while 21.8% were unemployed | Yes | Three self-report items adapted from prior research with gay and bisexual men assessing lifetime physical, sexual, and psychological IPV victimization (Finneran & Stephenson, 2013; Newcomb & Mustanski, 2016; Stephenson & Finneran, 2013). For each positive response to a lifetime IPV measure, a question was displayed to ask if that form of violence had occurred within the last 12 months. A composite dichotomous variable was created for lifetime IPV based on at least one positive response to any of the three lifetime measures. This process was repeated for past-year IPV victimization. | Self-report of any lifetime or past-year (recent) experiences based on the three items | Physical, sexual, and psychological (lifetime and within past 12 months) | No | NA | NA | Recent housing instability (within the past 12 months) | A single categorical demographic question (question not specified) | Self-report of instability within the last 12 months | Recent housing instability was independently associated with a significantly higher prevalence of recent IPV victimization. | Recent housing instability was independently associated with a significantly higher prevalence of recent IPV victimization (adjPrR = 1.36; 95% CI [1.02, 1.82]; p=0.039). 32.7% of participants experiencing recent IPV reported housing instability (p<.001) | None. | Yes | ACCELERATE! | Convenience sampling (not representative); lack of data on polyvictimization, perpetration, or childhood experiences; small multiracial sample size; and the predominantly cross-sectional nature of the data, which prevents establishing causality | |
21 | https://doi.org/10.1016/j.amepre.2017.03.021 | Economic Insecurity and Intimate Partner and Sexual Violence Victimization | Breiding, Matthew J. | 2017 | American Journal of Preventive Medicine | Quantitative ; nationally representative cross-sectional survey | Analysis of data from the 2010 National Intimate Partner and Sexual Violence Survey (NISVS); dual-frame random-digit-dial telephone survey design to assess victimization and economic factors experienced in the preceding 12 months | Multivariate logistic regression modeling, weighted analyses, chi-square tests | 16,507 adults | Nationally representative sample of the non-institutionalized U.S. adult population | Random-digit-dial telephone survey including both landline and cell phones | United States | 18+; mean and range NP | 55% Women, 45% Men | NR | NA | NA | NA | NR | Yes | Behaviorally specific questions from the 2010 National Intimate Partner and Sexual Violence Survey (NISVS) | Self-reported experience of specific violent or aggressive acts by an intimate partner (spouse, boyfriend/girlfriend, or dating partner) within the 12 months preceding the interview | Physical violence, stalking, psychological aggression, control of reproductive or sexual health, and contact sexual violence (SV) | No | NA | NA | Housing insecurity, defined specifically as worry or stress about having enough money to pay rent or mortgage | A single question from the Social Context module of the Behavioral Risk Factor Surveillance System (BRFSS) | Respondents were designated as insecure if they reported being worried/stressed about rent/mortgage "sometimes," "usually," or "always" (versus "never" or "rarely") in the past 12 months | Robust and significant associations were found between housing insecurity and all forms of IPV for both women and men. Women and men reporting recent housing insecurity were significantly more likely to experience physical violence, contact sexual violence, psychological aggression, and reproductive control compared to those who were housing secure. The relationship remained significant even after controlling for family income, age, race, education, and marital status. | Women (Housing Insecurity vs. IPV): Women with housing insecurity had significantly higher 12-month prevalence of physical violence (7.3% vs. 3.0%), contact SV (3.4% vs. 1.2%), stalking (4.8% vs. 1.1%), psychological aggression (20.2% vs. 8.6%), and reproductive control (2.4% vs. 0.9%). Men with housing insecurity had significantly higher 12-month prevalence of physical violence (7.3% vs. 3.0%), contact SV (3.4% vs. 1.1%), psychological aggression (25.9% vs. 12.8%), and reproductive control (2.8% vs. 0.8). | For men, stalking was not significantly associated with recent housing insecurity in the bivariate analysis. | No | NA | cross-sectional, so no directionality of association; low response rate (33.6%), so potential nonresponse bias; self-report data subject to recall bias and telescoping; housing insecurity measure may capture psychological distress rather than just objective financial hardship | Reproductive control definition: refusal to use a condom; for women, when a partner tried to get her pregnant when she did not want to become pregnant; for men, when a partner tried to get pregnant when he did not want her to become pregnant |
22 | https:doi.org/10.1186/s40621-016-0069-4 | Associations of financial stressors and physical intimate partner violence perpetration | Schwab-Reese, Laura M. | 2016 | Injury Epidemiology | Quantitative; secondary analysis of longitudinal survey data (Add Health), but only Wave IV | This was a secondary analysis of Wave IV data from The National Longitudinal Study of Adolescent to Adult Health (Add Health). While the parent study is prospective longitudinal, this analysis is primarily concurrent/retrospective. | Logistic regression, multinomial logit regression, chi square tests | 11,499 participants | Community-based; representative sample of the United States population of young adults (originally recruited as adolescents) | Systematic sampling and stratification (from the original Add Health study design) | United States | Range: 24 - 32 at time of Wave IV interview | NR | NR | NA | NA | NA | Significantly more women (27.7 %) than men (22.9 %) experienced at least one financial stressor | Yes | Self-report questions asking about the frequency of perpetration in the last 12 months | Participants must have been in a romantic relationship in the 12 months prior to the interview | Physical IPV (categorized as threats/minor physical IPV, severe physical IPV, and physical IPV resulting in injury) | No | NA | NA | Housing nonpayment (rent or mortgage) and eviction | Self-report dichotomous questions regarding the past 12 months | Experiences occurring within the 12 months prior to the interview | Housing nonpayment was associated with increased odds of perpetrating all three forms of physical IPV (minor, severe, and causing injury) in adjusted analyses. Eviction was specifically associated with increased odds of severe physical IPV perpetration but not the other two categories when modeled separately. | Housing nonpayment was significantly associated with increased odds of perpetrating all three measured forms of physical intimate partner violence (IPV), including threats/minor physical IPV (AOR = 1.68; 95% CI: 1.33–2.12), severe physical IPV (AOR = 1.74; 95% CI: 1.38–2.20), and IPV resulting in injury (AOR = 1.78; 95% CI: 1.21–2.59). When examining specific patterns of co-occurring violence, housing nonpayment also significantly predicted the combination of threats/minor and severe physical IPV (AOR = 1.95; 95% CI: 1.40–2.71) as well as the presence of all three forms of IPV (AOR = 1.95; 95% CI: 1.61–2.91). Experiencing eviction significantly increased the odds of perpetration of severe physical IPV specifically (AOR = 2.39; 95% CI: 1.35–4.23) and was associated with the pattern of both threats/minor and severe physical IPV (AOR = 3.72; 95% CI: 1.83–7.57) | Housing nonpayment was not significantly associated with perpetrating only threats/minor physical IPV (AOR = 1.35; 95% CI: 0.97–1.89) or only severe physical IPV (AOR = 1.43; 95% CI: 0.84–2.43) when compared to no perpetration in multinomial models. When modeled individually for eviction, there were no significant associations found with the perpetration of threats/minor physical IPV (AOR = 1.68; 95% CI: 0.98–2.88) or IPV resulting in injury (AOR = 0.61; 95% CI: 0.19–2.02). In multinomial models, eviction was not a significant predictor for only threats/minor physical IPV (AOR = 0.86; 95% CI: 0.35–2.10), only severe physical IPV (AOR = 2.23; 95% CI: 0.42–11.75), or the co-occurrence of all three forms of IPV (AOR = 0.84; 95% CI: 0.26–2.69). | No | NA | perpetrator self-reports may be biased; individual level analyses, limiting temporal inferences about whether stressors preceded the IPV; no examination of co-occurrence of stressor patterns | |
23 | https://doi.org/10.1093/aje/kwaf003 | Physical Intimate Partner Violence Among Women Experiencing Homelessness Before and During Pregnancy | Sakai-Bizmark, Rie | 2025 | American Journal of Epidemiology | Quantitative; observational, population-based study using survey data (PRAMS) | secondary analysis of cross-sectional survey data from the Pregnancy Risk Assessment Monitoring System (PRAMS). Data were collected via questionnaires mailed to mothers 2 to 4 months after a live birth delivery. | Multivariable logistic regression; multinomial regression | 186,891 respondents,representing an estimated 11,489,161 women after survey weights | Community-based population consisting of postpartum women who recently had a live birth | systematic random sample of birth certificates, with oversampling of high-risk subpopulations | United States (47 participating states in 2020) | ≤17 (0.8%), 25-29 (29.9-30.0%) and 30-34 (18.3-30.1%) | 100% female | Non-Hispanic White (44.8–62.2%), Non-Hispanic Black (13.1–32.4%), Hispanic (14.6–15.9%), Non-Hispanic Asian (1.0–4.7%), and Other (4.0–7.3%) | NA | NA | NA | NA | Yes | Self-report via PRAMS survey questions asking if anyone "pushed, hit, slapped, kicked, choked, or physically hurt" the participant in the 12 months before or during pregnancy | Self-reported experience of physical IPV during the 12 months prior to or during the most recent pregnancy | Physical IPV | No | NA | NA | Homelessness (including sleeping outside, in a car, or in a shelter) | Self-report response to the statement: "I was homeless or had to sleep outside in a car, or in a shelter" | Homelessness experienced during the 12 months before the new baby was born | The study found that 27.1% of women experiencing homelessness reported physical IPV, a significantly higher rate than the 3.4% reported by domiciled women. Physical IPV was associated with several adverse health outcomes among domiciled women, such as postpartum depression (aOR = 1.83) and preterm birth (aOR = 1.29), but these associations were significantly smaller among women experiencing homelessness. | 27.1% of women experiencing homelessness reported physical IPV, a significantly higher rate than the 3.4% reported by domiciled women. For housed women, the experience of physical IPV was significantly associated with higher odds of postpartum depression (aOR = 1.83; 95% CI: 1.62–2.08) and preterm birth (aOR = 1.29; 95% CI: 1.08–1.53), along with lower odds of attending a postpartum checkup (aOR = 0.56; 95% CI: 0.49–0.64). When looking at the moderating effect of housing, the interaction terms revealed that the associations between IPV and these outcomes were significantly weaker for women experiencing homelessness regarding postpartum depression (aOR interaction = 0.74; 95% CI: 0.58–0.94), preterm birth (aOR interaction = 0.68; 95% CI: 0.49–0.93), and postpartum checkup attendance (aOR interaction = 1.32; 95% CI: 1.01–1.74). | Homelessness did not significantly moderate the relationship between physical IPV and low birthweight (aOR interaction = 1.03; 95% CI: 0.81–1.30), small for gestational age (aOR interaction = 0.97; 95% CI: 0.72–1.31), breastfeeding for less than 8 weeks (aOR interaction = 0.95; 95% CI: 0.74–1.23), insufficient weight gain (aOR interaction = 1.21; 95% CI: 0.91–1.60) or excessive weight gain (aOR interaction = 1.00; 95% CI: 0.74–1.37). | No | NA | unable to determine temporal sequence of IPV and homelessness; self-report may lead to bias; no psychological or sexual abuse measured; survey relied on physical addresses and phones, which may underrepresent the most vulnerable women experiencing homelessness, potentially leading to selection bias | |
24 | https://doi.org/10.1177/08862605241263588 | Predicting Intimate Partner Violence Perpetration Among Young Adults Experiencing Homelessness in Seven US Cities Using Interpretable Machine Learning | Um, Mee Young | 2025 | Journal of Interpersonal Violence | Quantitative; supervised machine learning study using an interpretable ensemble learning approach (gradient boosting classifier) | Cross-sectional survey (Homeless Youth Risk and Resilience Survey - HYRRS) | Gradient boosting classifier with five-fold cross-validation; SHapley Additive exPlana-tions (SHAP) values were used to rank predictor importance and direction | 1,426 participants | Young adults experiencing homelessness recruited from agencies | Purposive sampling from collaborating host agencies (drop-in centers, emergency shelters, and permanent supportive housing) | Seven U.S. cities: Denver, CO; Houston, TX; Los Angeles, CA; New York, NY; Phoenix, AZ; San Jose, CA; and St. Louis, MO | Mean 20.88 (SD = 2.09); Range 18 to 26 years | Cisgender male (58.54%), Cisgender female (33.94%), and Transgender, genderqueer, or other (7.52%) | African American (37.29%), White (18.96%), Latino/a (17.35%), Mixed race/ethnicity (16.22%), and Other (10.18%, including American Indian and Asian or Pacific Islander) | NA | NA | NA | 31.83% were currently employed; 69.19% had at least a high school diploma or GED; 30.68% had a lifetime homeless duration over 2 years | Yes | Self-report via a single question asking if the participant had ever "hit, pushed, shoved, punched, kicked or physically hurt" a romantic/intimate partner on purpose since becoming homeless; IPV victimization item not reported | Affirmative response to physical IPV perpetration (outcome variable) and IPV victimization (predictor) during the period of homelessness | Physical IPV perpetration and victimization | No | NA | NA | Homelessness or being unstably housed (e.g., sleeping in a shelter, on the streets, or temporarily with friends/family) | Eligibility screener for housing status; no specific item reported | Participants must have spent the previous night in a shelter, on the streets, or in a temporary/unstable arrangement | One-fifth (20.01%) of young adults reported perpetrating physical IPV while homeless, while over one-quarter (26.24%) reported IPV victimization while homeless. The most critical factor in predicting IPV perpetration was having experienced IPV victimization while homeless (prediction probability = 37.54%). | In the machine learning model, which achieved a five-fold cross-validation accuracy of 0.81, IPV victimization while homeless was the strongest predictor of perpetration (prediction probability = 37.54%). Sexual victimization while homeless (0.79%) was also a predictor of IPV perpretation. | None. | No | NA | self-report, so risk of social desirability bias; cross-sectional, so no causal inferences; did not explicitly model bidirectional IPV or other forms of violence (sexual, psychological); sample was limited to English speakers recruited from specific agencies | Witnessing parental IPV was related to IPV perpetration. |
25 | https://doi.org/10.1037/vio0000467 | Intimate Partner Violence During COVID-19: Demographic and Economic Correlates of Survivor Experiences | Wood, Leila | 2023 | Psychology of Violence | Quantitative, cross-sectional survey | Participants were surveyed in December 2020 regarding their experiences in the previous 12 months and specifically during the COVID-19 pandemic | Analysis of variance (ANOVA), t tests, chi-square tests, and logistic regression | 374 participants | Recent survivors of IPV | Recruitment in partnership with over 20 social service agencies and churches (including medical clinics and food pantries) via email, print materials, online client portals, and social media | Harris County (Houston), Texas | Range 18 - 46+; the largest group was 26–35 (38.2%), followed by 18–25 (30.5%), 36–45 (19.3%), and 46+ (12.0%) | 77.3% Female, 19.8% Male, 2.9% Other (transgender, nonbinary, unknown) | 39.8% Hispanic, 28.3% Non-Hispanic Black, 22.5% Non-Hispanic White, 9.4% Other/multiracial | NA | NA | NA | 67.9% had monthly incomes of $1,000 or less; 75.9% had a history of lifetime homelessness | Yes | Modified version of the Partner Violence Scale (for eligibility); Conflict Tactics Scale Revised–Short Version (CTS2) (physical, psychological, sexual IPV); Scale of Economic Abuse (2 items); seven-item scale by Kennedy et al. (coercive control); and single items for reproductive coercion and coerced labor | Endorsement of at least one of two behaviorally specific past-year IPV victimization questions (threats of hurt or being pushed/grabbed/shaken) | Physical, psychological, sexual, economic, reproductive coercion, coerced labor, and coercive control | No | NA | NA | Lifetime homelessness (frequency) and homelessness since the COVID-19 pandemic began | Study-developed items; not speccifically reported | NA | Pandemic-related homelessness was significantly associated with self-reported increases in IPV victimization. Survivors of homelessness during COVID-19 were nearly twice as likely to report worsening violence | Homelessness during the COVID-19 pandemic was significantly associated with self-reported increases in IPV victimization, with survivors experiencing homelessness being nearly twice as likely to report worsening violence (OR = 1.94, 95% CI [1.27, 2.98], p < .01) | None. | No | NA | Convenience sample from one city, so limited generalizability; cross-sectional design, so cannot establish temporal relationships or long-term impacts; use of limited items for some constructs like economic IPV | |
26 | https://doi.org/10.1016/j.jacr.2021.03.006 | Recognizing Isolated Ulnar Fractures as Potential Markers for Intimate Partner Violence | Khurana, Bharti | 2021 | Journal of the American College of Radiology | Quantitative, retrospective, observational cohort study | Electronic health record (EHR) query and clinical chart review of patients at three level I trauma centers | Comparative analysis, Relative Risk (RR) calculation, and median/interquartile range comparisons using p-values | 62 women | Women aged 18 to 50 with isolated ulnar fractures | Retrospective EHR query | Northeast United States (three large, urban academic medical centers) | Mean 31 years; range 18 - 50 | 100% female | CHECK | NA | NA | NA | NA | Yes | Clinical chart review (confirmed through self-disclosure or clinical suspicion) and historical imaging analysis (radiology-based prediction) | Group stratification based on disclosure (confirmed), clinical suspicion (possible), or lack of evidence (not suspected/not IPV) | Physical (direct assault resulting in fracture) | No | NA | NA | Homelessness (lack of domicile) | Clinical chart review | NR | Confirmed IPV cases were significantly associated with homelessness. | Confirmed IPV cases were associated with homelessness (46% vs. 0%; p < .001). | None. | No | NA | Only 22.5% of patients (14 of 62) had formal documentation of IPV evaluation in their clinical charts at the time of injury. | |
27 | https://doi.org/10.1177/0886260519876024 | Exploring Multiple Forms of Intimate Partner Violence in a Gender and Racially/Ethnically Diverse Sample of Transgender Adults | King, Wesley M. | 2021 | Journal of Interpersonal Violence | Quantitative; secondary analysis of cross-sectional survey data | Secondary analysis of the 2015 U.S. Transgender Survey (USTS), which used purposive sampling for a one-time online survey | Univariable and bivariable distributions (weighted proportions) and multivariable logistic regression models | 23,999 adult transgender participants | Transgender adults living in the United States who had ever been in a romantic or sexual relationship | Purposive sampling via a network of community-based organizations serving transgender individuals | United States (all 50 states) | 18-24 (40.0%), 25-44 (44.2%), 45-64 (13.8%), 65+ (2.1%) | 31.2% transgender men, 34.2% transgender women, 27.5% assigned-female-at-birth (AFAB) nonbinary, and 7.1% assigned-male-at-birth (AMAB) nonbinary | 31.2% transgender men, 34.2% transgender women, 27.5% assigned-female-at-birth (AFAB) nonbinary, and 7.1% assigned-male-at-birth (AMAB) nonbinary | NA | NA | NA | 67.7% had an annual household income of 49,999 orless | Yes | 24 questions categorized into subtypes: physical IPV (9 items), psychological IPV (10 items), stalking (1 item), forced sex (1 item), and trans-related IPV (3 items) | Lifetime self-report by participants who had ever been in a romantic or sexual relationship | Physical, psychological, stalking, forced sex, and trans-related (e.g., being told they weren't a "real" man/woman or having hormones withheld) | No | NA | NA | Lifetime homelessness | Single item asking if the participant had ever been homeless, including staying in a shelter, living on the street/car, or temporary housing due to inability to afford housing | Any lifetime occurrence of the self-reported conditions | Lifetime homelessness was consistently and significantly associated with increased odds of experiencing all five subtypes of IPV | Lifetime homelessness was associated with more than twice the odds of all IPV subtypes in multivariable models: Physical IPV: OR = 2.5, 95% CI [2.3, 2.7, Psychological IPV: OR = 2.4, 95% CI [2.2, 2.6], Forced Sex: OR = 2.3, 95% CI [2.1, 2.6], Stalking: OR = 2.2, 95% CI [1.9, 2.4], Trans-Related IPV: OR = 2.1, 95% CI [1.9, 2.3]. Bivariate analysis showed significantly higher prevalence for those with homelessness vs. those without (χ 2p<.001 for all): Psychological (58.7% vs 33.5%), Physical (56.5% vs 31.2%), Stalking (27.7% vs 13.3%), Forced Sex (31.6% vs 16.4%), and Trans-Related (42.6% vs 24.2%) | None. | No | NA | cross-sectional, so unable to make causal conclusions; measures lifetime exposure,not recency/frequency; self-report bias; potential for the sample to have higher SES than the general transgender population | |
28 | https://doi.org/10.1177/08862605221078805. | Differential Experiences of Sexual, Physical, and Emotional Intimate Partner Violence Among Transgender and Gender Diverse Adults | Kattari, Shanna K. | 2022 | Journal of Interpersonal Violence | Quantitative, cross-sectional study using survey data | Secondary analysis of the 2018 Michigan Trans Health Survey | chi-square tests of independence, and multivariate logistic regressions | 557 transgender and gender diverse adults | Transgender and gender diverse adults living in Michigan | Convenience and snowball sampling conducted both online (social media, listservs) and in-person (Pride events, community organizations) | Michigan, USA (69.7% urban/suburban; 30.3% rural/small city/frontier) | Mean 28.61 (SD: 10.27) | 28.4% Transmasculine, 26.8% Nonbinary, 22.3% Additional genders, and 19.4% Transfeminine | 76.5% White, 6.6% Black/African American, 5.0% Multiracial/Biracial, 4.3% Latinx/Chicanx/Hispanic, and 3.1% Additional races . | NA | NA | NA | Over a third of the sample (39.0%) had ever experienced homelessness | Yes | Survey items created by an IPV scholar and workshopped by community partners; respondents were asked about behaviorally specific incidents | Lifetime self-report of specific victimization experiences by an intimate partner | Physical violence, sexual violence (forced sex), and emotional/psychological violence (specifically threats of being outed and gender belittlement) | No | NA | NA | Lifetime homelessness | A single-item re-coded into a dichotomous yes/no variable regarding whether the participant had ever experienced homelessness | Self-reported lifetime occurrence of homelessness | Experiences of homelessness were significantly related to a higher likelihood of experiencing every form of IPV measured. | Survivors with a history of homelessness were more than five times as likely to have experienced physical IPV (OR = 5.52, 95% CI [3.65, 8.34], p < .001). History of homelessness was associated with nearly three times the odds of forced sex by a partner (OR = 2.92, 95% CI [1.96, 4.33], p < .001), three times the odds of outing threats by a partner (OR = 3.15, 95% CI [1.90, 5.22], p < .001), andmore than three times the odds of having one's gender belittled by a partner (OR = 3.28, 95% CI [2.13, 5.05], p < .001). Chi-square tests showed that homelessness history was significantly associated with a higher lifetime prevalence of physical violence (56.2% vs. 43.8%), forced sex (54.8% vs. 29.0%), threats of being outed (34.1% vs. 10.9%), and gender belittlement (43.8% vs. 16.5%) when compared to those who had never experienced homelessness (p < .001 for all). | None. | No | NA | Cross-sectional data prevents causal relationships; small sample size for certain sub-categories; a sample that was primarily White (76.5%); an incident-based operationalization of IPV rather than a pattern-based one | |
29 | https://doi.org/10.1097/QAI.0b013e3181561f7f | Intimate partner violence perpetration against main female partners among HIV-positive male injection drug users | Frye, Victoria | 2007 | Journal of Acquired Immune Deficiency Syndromes | Quantitative; cross-sectional analysis of baseline data from a randomized trial | Post hoc analysis of baseline data from the INSPIRE study, a multisite randomized trial. Data were collected via Audio Computer-Assisted Structured Interviewing (A-CASI) to enhance reporting of sensitive behaviors | Chi-square tests, t-tests, and multivariate logistic regression | 317 men | HIV-positive male current injection drug users (IDUs) who reported having at least one main female partner | Recruited from clinic and community sources | Four U.S. cities: Baltimore, Miami, New York City, and San Francisco. | Mean 43.3 years (SD = 6.3) | 100% male | 63.7% African American; 20.1% Hispanic; 9.4% White; 6.8% Other | NA | NA | NA | 49% earned <$5,000 per year; 39.8% had less than a high school education; and less than 10% held a full-time paid job | Yes | modified version of the Conflict Tactics Scale (CTS) | Self-reported use and/or threat of physical or sexual IPV against a main female partner in the past 12 months | Physical (e.g., hitting, throwing things, using a knife or gun) and sexual (using force to make a partner have sex) | Yes | Two self-report questions asking if the respondent was ever "beaten, physically attacked or physically abused" or "sexually attacked, raped or sexually abused" as a child | Physical abuse and sexual abuse | Homelessness | Self-report (Yes/No) regarding status in the past year and at the time of the interview | Identification as currently homeless or having been homeless in the year prior to the interview | Homelessness was a significant predictor of IPV perpetration. | homelessness in the past year was significantly associated with a higher likelihood of IPV perpetration (AOR = 2.11, 95% CI = 1.20–3.71, p < 0.01); 36.4% of those who perpetrated IPV were homeless in the past year compared to only 17.8% of those who did not perpetrate IPV (p < 0.01) | No statistically significant association was found between having a childhood history of abuse and adult IPV perpetration (42.7% for perpetrators vs. 38.9% for non-perpetrators, p = 0.56) | Yes | INSPIRE | cross-sectional, so unable to make causal conclusions; convenience sampling, so may not be generalizable; self-report, so potential underreporting; measure of IPV not include psychological or financial abuse; did not differentiate between offensive and defensive violence (self-defense) | |
30 | https://doi.org/10.1086/726284 | Victimization and Mental Health Symptoms Among Recently Incarcerated Women: The Roles of Homelessness Prior to Incarceration and Community-Based Social Support | Fedock, Gina | 2025 | Journal of the Society for Social Work and Research | Quantitative, observational, cross-sectional survey | A cross-sectional survey was administered to women who had recently entered prison (within the past three years) | Structural equation modeling (SEM), confirmatory factor analysis (CFA) | 487 incarcerated women | Recently incarcerated women | The research team visited dayrooms in a large, multisecurity level prison | Midwestern state prison | Mean 36.18 years (SD = 10.25); Range 18–70 years | 100% female | 56.5% white, 31.0% Black or African American, and 11.8% another race or multiracial | NA | NA | NA | before prison, 35.5% had less than 12 years of education, 58.5% had experienced homelessness in the year prior to incarceration, and 41.1% reported being unable to pay rent or mortgage | Yes | 12 items from the Conflict Tactics Scale (CTS) and the Revised Conflict Tactics Scale (CTS2) | Self-report of physical and/or sexual victimization during adulthood | Physical violence (severe and minor subscales), verbal aggression, and sexual victimization | No | NA | NA | Homelessness, eviction, being forced to move in with family or friends (doubling up), living in a shelter/transitional housing, living in a hotel, sleeping in a car, and sleeping in an abandoned building | A 7-item assessment based on HUD research (e.g., Gubits et al., 2016) where participants provided yes/no responses to specific housing experiences in the year prior to incarceration | answering "yes" to at least one of the seven housing experiences | IPV victimization was positively and significantly associated with preprison homelessness. | IPV victimization was positively and significantly associated with preprison homelessness (b = 0.17, p < .01). Homelessness partially mediated the association between IPV and depressive (b=0.421, p<.01), anxiety (b=0.439, p<.01), and PTSD (b=0.319, p<.01) symptoms. | None. | No | NA | Cross-sectional data, so no causality or temporal order of events; Lack of measures for specific forms of oppression and exclusion (e.g., racism, transphobia, ableism); Information on service engagement was limited primarily to housing shelters; social support measure was not standardized or specifically tailored to the conditions of incarceration; Statistical significance reported does not automatically reflect practical or clinical significance | |
31 | https://doi.org/10.1136/oemed-2020-106487 | Incidence and predictors of violence from clients, intimate partners and police in a prospective US-based cohort of women in sex work | Decker, Michele | 2021 | Occupational and Environmental Medicine | Quantitative, observational, longitudinal, prospective cohort study | A prospective 12-month study with five assessments conducted at 3-month intervals (baseline, 3, 6, 9, and 12 months) | Poisson regression models with robust variance; forward and backward stepwise selection was used for multivariable models | 250 enrolled at baseline; 160 completed the 12-month follow-up; 103 completed all five visits | Street-based cisgender female sex workers | recruited from 14 urban zones with high concentrations of street-based sex work; screening and data collection occurred on a mobile van | Baltimore, Maryland (urban) | Mean 35.7 years; Range 18–61 years | 100% cisgender women | 66% Non-Hispanic white, 22.8% Non-Hispanic Black, and 10.8% Hispanic or other | NA | NA | NA | racial/ethnic disparities, economic blight, substance abuse, HIV and violence. | Yes | Adapted Revised Conflict Tactic Scale (CTS2) | Self-report of physical or sexual violence from non-paying intimate partners | Physical violence (hit, punched, slapped, threatened with weapon) and sexual violence (forced or pressured vaginal/anal sex) | Yes | Behavioral assessment of physical (hit, slapped, punched) or sexual (forced/coerced touching or sex) abuse before age 18. | Physical and sexual abuse | Homelessness | A time-varying binary (yes/no) self-report assessed at 3-month intervals | Participants were identified as homeless if they reported being so within the 3-month recall window | Homelessness was a strong, independent predictor of incident IPV. | IPV and Homelessness: Adjusted Incidence Rate Ratio (aIRR) = 2.0; 95% CI [1.3, 2.9]; p<.001; | None. | No | NA | 12-month duration may not capture long-term dynamics; Incidence rates were underestimated because they were capped at one event per 3-month reporting period; Women with greater structural vulnerabilities (homelessness, daily drug use) were more likely to be lost to follow-up (retention bias); non-probability sample from a single urban setting may not represent all sex workers, especially those working indoors or online | |
32 | https://doi.org/10.1080/10926771.2024.2418546 | Experiences of Dating Violence and Protective Factors Among Adolescents in Vulnerable Contexts | Rutman, Shira P. | 2024 | Journal of Aggression, Maltreatment & Trauma | Quantitative, observational, cross-sectional | cross-sectional analysis using anonymous, voluntary baseline survey data from participants in a statewide education program | chi-square tests, multivariable logistic regression | 12633 adolescents; 1,951 adolescents in vulnerable contexts and 10,682 adolescents living at home | Adolescents in vulnerable contexts (foster care, juvenile justice facilities, and experiencing homelessness/unstable housing) compared to those living at home | recruited from those enrolled in the California Personal Responsibility Education Program (CA PREP), which serves youth in communities facing health inequities | 20 counties in California | NA | NA | NA | Range: 10–21 years | Vulnerable population: 65.6% male, 34.4% female; Home population: 51.6% male, 48.4% female | Total sample was predominantly Hispanic (76.4% of those at home; 59.0%–63.4% of those in vulnerable contexts) . Other groups included Non-Hispanic Black (14.4% in vulnerable contexts vs. 5% at home), Non-Hispanic White, and Non-Hispanic other/multiple races | Program eligibility was based on local indicators of health inequities, including concentrated poverty, birth rates, and racially isolated areas. | Yes (Dating violence) | measures were drawn from the Youth Risk Behavior Survey and coded as one or more experiences of each type: “Physical dating violence,” “Sexual dating violence,” “Any physical and/or sexual dating violence,” and “Both physical and sexual dating violence” (Clayton et al., 2023). | Self-report of physical or sexual dating violence experienced during the past 12 months among participants who reported dating during that period | Physical dating violence (being hit, slammed, or injured with a weapon) and sexual dating violence (being forced to do sexual things) | No, but foster care involvement was used as one of the definitions for "vulnerable context" | NA | NA | Homelessness/unstable housing, defined as couch surfing, living on the street, in a car, or in a shelter | Combined data from youth self-report (survey) and agency reports (program setting data, such as receiving services in a shelter or transitional housing) | Self-reported the status or received program services in a shelter or transitional housing setting . | Adolescents in vulnerable contexts overall reported significantly higher percentages of past-year ADV (21.1%) compared to those living at home (15%) | Adolescents experiencing homelessness have 56% higher odds of experiencing any dating violence compared to those living at home (aOR=1.56; 95% CI [1.08,2.27]; p<.05). Homelessness is associated with double the odds of experiencing physical dating violence (aOR=2.00; 95% CI [1.34,2.93]; p<.05). | The adjusted odds ratio for sexual dating violence among adolescents experiencing homelessness is 1.46 (95% CI [0.95,2.23]) but was no longer statistically significant after adjusting for protective factors. | Yes | CA PREP | self-report, so risk of underestimation of dating violence due to stigma; cross-sectionality preventing determining direction of the relationship between context and dating violence; survey asked about ADV in the past year but protective factors in the past three months; lack of detailed information on the duration or lifelong experience of context/violence; gender identity/sexual orientation measures did not reflect current best practices | |
33 | https://doi.org/10.1007/s10896-025-00817-0 | Advancing the Study of Adverse Adult Experiences: A Validation Study of the Adult Experiences Survey | Mersky, Joshua P. | 2025 | Journal of Family Violence | Quantitative; validation study using baseline survey data from an observational, prospective investigation; this specific analysis is cross-sectional | This analysis examines baseline survey data collected between October 2021 and May 2023 from participants in the Strong and Stable Families Study | multivariate analysis of individual heterogeneity and discriminatory accuracy (MAIHDA), bivariate correlation analyses, multiple linear regression, and multigroup confirmatory factor analysis (CFA) with alignment optimization | 2,004 adults | Mixed; 647 primary caregivers receiving services from Family Resource Centers (FRCs) and 1,489 primary caregivers randomly sampled from the general population of families with children | FRC participants were recruited from 18 centers in Wisconsin. The general population sample was recruited via an address-based sampling frame from a commercial vendor (surveys distributed by mail and followed by reminder phone calls) | Wisconsin | Mean: 40 years (SD = 8.2); range 18-79 | 85.8% female; 14.2% male/other (9 participants identifying as "another gender" were omitted from inferential analyses due to small sample size) | 81.0% White; 5.6% Black; 5.5% Hispanic; 4.1% American Indian or Alaska Native; 3.8% Other | NA | NA | NA | 28.1% low income (<50,000) | Yes | Adult Experiences Survey (AES), which includes 10 core adversities | Self-report; items were dichotomized where responses of "once" or "more than once" were coded as having experienced the adversity. | Physical or emotional abuse and sexual abuse by a current or former romantic partner or spouse | No | NA | NA | Homelessness | Adult Experiences Survey (AES); homelessness was measured on a 5-point Likert scale (never to very often) and then dichotomized for analysis | Self-report via dichotomized survey item | Physical or emotional IPV and sexual IPV were both found to be positively associated with homelessness | Correlation between physical or emotional IPV and homelessness: r = 0.29, p < 0.01; Correlation between sexual IPV and homelessness: r = 0.23, p < 0.01 | None. | No | NA | cross-sectional design prevents causal conclusions; potential for unmeasured confounding (such as prior ACEs; data relies on brief self-report assessments; and the sample is not fully representative of the U.S. population | |
34 | https://doi.org/10.1177/1077801216664426 | Sheltering for Safety in Community Women With Divorce Histories | Warnecke, Ashlee J. | 2017 | Violence Against Women | Quantitative, cross-sectional community-based study | Retrospective, cross-sectional design where women reported on lifetime intimate partner abuse and sheltering for safety across up to four past intimate relationships from which they had divorced or separated . | Kruskal–Wallis tests for continuous variables (due to skewed distributions); chi-square tests and Fisher’s exact test for categorical variables; and Bonferroni corrections for follow-up comparison | 197 women | Community sample of midlife women with histories of divorce or separation from a stressful relationship | Recruited from the community for a broader investigation into IPV and aging; eligibility was determined via telephone interviews | Louisville, Kentucky | Median age = 55.22 years (IQR = 5.75) | 100% female | 79.19% Caucasian/White; 18.27% African American/Black; 0.51% American Indian/Alaskan Native; 2.03% Other | NA | NA | NA | 59.28% of the sample had an annual household income under $40,000; 50.26% had completed a college degree or higher | Yes | Psychological Maltreatment of Women Inventory–Short Form (coercive control); the Revised Conflict Tactics Scale (severe physical assault, sexual coercion, and injury); and supplemental questions regarding head injury | Self-report; participants had to have a history of separation or divorce from a cohabiting male partner but no experience of IPV in the past year | Coercive control, severe physical violence, severe sexual coercion, and injury (general and head injury) | No | NA | NA | Sheltering for safety, categorized as formal sheltering (agency shelters) and informal sheltering (staying with friends or family) | Self-report items where women indicated if they had ever left a relationship to seek safe shelter in the community, with friends/family, or both. | Self-reported lifetime history of seeking safety through sheltering | Women with any sheltering history (formal or informal) reported significantly higher levels of IPV severity (coercive control, severe violence, and injury) than women who never sheltered. Women who sheltered formally did not differ significantly from those who sheltered informally in terms of abuse severity | Association between sheltering history and coercive control: χ 2 (2,N=197)=31.899,p<.0001; Association between sheltering history and severe violence exposure: χ2(2,N=196)=60.69,p<.0001; Association between sheltering history and general injury: χ2 (2,N=197)=49.98,p<.0001; Association between sheltering history and head injury: Fisher’s exact p<.0001,ϕ=.52; Both formal and informal shelterers reported significantly more severe violence than non-shelterers (p<.0001,r=.53 and r=.52 respectively) | No significant difference between formal vs. informal sheltering for coercive control: p=.30,r(106)=.10, severe violence: p=.04,r(106)=.20 (did not meet the Bonferroni-adjusted alpha of .016), general injury: p=.11,r(106)=.16, or head injury (p=.18,r(106)=.17) | No | NA | retrospective reporting may be subject to memory fallibility; IPV was measured globally across all past relationships rather than for one specific relationship; the study lacked a direct comparison group recruited from an agency shelter, and it did not comprehensively assess all forms of formal resource use | |
35 | https://doi.org/10.1002/jcop.20374 | INTIMATE PARTNER VIOLENCE RELATIONSHIP DISSOLUTION AMONG COUPLES WITH CHILDREN: THE COUNTERINTUITIVE ROLE OF "LAW AND ORDER" NEIGHBORHOODS | Emery, Clifton R. | 2010 | Journal of Community Psychology | Quantitative; longitudinal analysis of representative survey data | Prospective, longitudinal study using Waves 1 and 2 of the Project on Human Development in Chicago Neighborhoods (PHDCN). | Generalized Boosting Model (GBM) and logistic regression | 555 cases | Community sample; a representative random probability sample of children and their primary caregivers | Three-stage cross-sectional stratified cluster sample of neighborhoods | Chicago, Illinois | Mean age = 33.76 years (SD = 7.13) at Wave 1 | 88.3% mothers; 6% fathers; 1.5% grandmothers; 4.2% other relationship to child | 46.85% Hispanic; 32.43% African American; 10.18% Caucasian; 9.91% Other race | NR | NR | NR | Mean annual household income = $26,662 (SD = $18,321); 25.77% had some high school education; 11.35% had a high school diploma; 35.50% had some post-high school education; 4.32% had a college degree or more | Yes | Conflict Tactics Scale (CTS) 1 | Self-report; participants must have experienced IPV (minor or severe physical violence) in the last year at Wave 1 | Minor violence (pushed, grabbed, slapped) and severe violence (kicked, bit, hit, beat up, choked, or threatened with a weapon) | No | NA | NA | Residential instability (measured as a neighborhood-level characteristic rather than an individual household status) | Calculated using 1990 census data aggregated to the neighborhood level, focusing on factors like homeownership and residential stability | Neighborhood-level factor included in the model | neighborhood-level residential instability was not significantly associated with the dissolution of IPV relationships | None. | Neighborhood residential instability was not a significant predictor of IPV relationship dissolution (eB = 0.55, 95% CI: 0.16–1.96) | No | NA | secondary data not specifically designed to study relationship dissolution; study lacked a measure of which partner decided to end the relationship; potential statistical power problems for the null findings; results are limited to Chicago and couples with children; no controls for individual or neighborhood selection effects | |
36 | https://doi.org/10.1089/jwh.2024.0038 | Antepartum Intimate Partner Violence: Development of a Risk Prediction Model | Prakash, Jaya | 2024 | Journal of Women's Health | Quantitative; a retrospective case-control chart review study | Retrospective case-control study. It compared pregnant patients who self-reported IPV (cases) with age-matched pregnant patients who did not (controls) using electronic medical record (EMR) data | Chi-squared tests, Fisher’s exact tests, stepwise logistic regression, N-fold cross-validation | 214 pregnant patients | Pregnant patients at a tertiary care academic center | Cases identified via referral to an IPV support/services program or provider documentation of IPV concerns | Boston, Massachusetts | Case patients mean age = 27.2 years (SD = 6.4); Control patients mean age = 27.6 years (SD = 5.5) | 100% female | Cases: 38% Black, 17.6% White, 11.1% Hispanic, 1.9% Asian, 31.5% None of the above. Controls: 35.8% Black, 20.8% White, 0% Hispanic, 17.0% Asian, 26.4% None of the above | NA | NA | NA | 44.4% of cases had unstable housing compared to 13.2% of controls | Yes | EMR documentation of provider-noted IPV concerns and/or participant response to the standardized screening question: "Do you feel safe in your relationship?" | Self-report or provider documentation in the EMR during the indexed pregnancy | Physical, emotional, and sexual | No | NA | NA | Unstable housing | EMR review of free-text and coded notes written by Social Work and Obstetrics & Gynecology providers | Documented status of unstable housing in clinical notes | Unstable housing was identified as one of six significant predictors of antepartum IPV. In the final risk prediction model, housing instability and depression were the strongest predictors, each receiving the highest point score weighting. | Bivariate association between unstable housing and antepartum IPV: OR = 5.26, 95% CI: 2.67–10.36, p < 0.001. Predictive value of housing instability in the logistic regression model: OR = 4.12, 95% CI: 1.92–8.84, p < 0.001 (Point Score = 14) | None. | No | NA | Relatively small cohort; findings were from a single institution, so limited generalizability; data may have missed care received at other facilities; potential for information bias in clinical notes | |
37 | https://doi.org/10.1177/00333549241278631 | Physical Intimate Partner Violence and Increased Partner Aggression During Pregnancy During the COVID-19 Pandemic: Results From the Pregnancy Risk Assessment Monitoring System | D'Angelo, Denise V. | 2025 | Public Health Reports | Quantitative, observational, population-based surveillance | Cross-sectional survey utilizing the Pregnancy Risk Assessment Monitoring System (PRAMS). Data collected retrospectively via a questionnaire supplement 2 to 6 months after a live birth, focusing on experiences that occurred during the most recent pregnancy. | Weighted prevalence estimates, Wald χ2 tests, adjusted prevalence ratios (APRs) using multivariable logistic regression | 14,154 respondents | Population-based sample of individuals with a recent live birth | Monthly stratified random sampling from birth certificate records; respondents were contacted via mail with telephone follow-up for nonrespondents | 29 US jurisdictions (Alaska, Arizona, Arkansas, Connecticut, Delaware, District of Columbia, Florida, Georgia, Illinois, Iowa, Louisiana, Maryland, Massachusetts, Michigan, Missouri, Nebraska, New Jersey, New York City, North Dakota, Oregon, Pennsylvania, Puerto Rico, South Dakota, Tennessee, Utah, Vermont, Virginia, West Virginia, and Wyoming) | Mean NR; Range from <20 to ≥35 years | 100% individuals with a recent live birth | Hispanic, non-Hispanic White, Black, Asian/Pacific Islander, American Indian/Alaska Native, and multiple races/other single race | NA | NA | NA | NR | Yes | PRAMS core survey and the Maternal COVID-19 Experiences supplemental questionnaire | Self-report of being physically hurt by a husband or partner/ex-partner during pregnancy, or report of increased aggression by partner perceived to be due to the COVID-19 pandemic | Physical IPV (pushed, hit, slapped, kicked, choked, or physically hurt) and increased physical, sexual, or emotional aggression | No | NA | NA | Moving or relocating and becoming homeless due to the COVID-19 pandemic | Self-report via the PRAMS Maternal COVID-19 Experiences supplemental questionnaire | Responding "yes" to having to move/relocate or becoming homeless due to the COVID-19 pandemic | Housing instability was identified as being strongly associated with increased risk for violence during pregnancy | Any housing stressor and Physical IPV: APR = 2.33 (95% CI, 1.41-3.84); Any housing stressor and Increased Aggression: APR = 2.83 (95% CI, 2.03-3.94); Became homeless and Physical IPV: APR = 4.76 (95% CI, 2.69-8.41); Became homeless and Increased Aggression: APR = 5.24 (95% CI, 3.24-8.49); Had to move and Physical IPV: APR = 1.89 (95% CI, 1.12-3.20); Had to move and Increased Aggression: APR = 2.55 (95% CI, 1.82-3.58) | None. | No | NA | Findings may not be generalizable outside of the 29 jurisdictions; study only includes individuals with a live birth (excludes stillbirths); self-report, so potential recall and social desirability bias; exact timing of violence relative to pandemic onset was unknown; unable to disaggregate types of increased aggression | |
38 | Intimate violence in the lives of homeless and poor housed women: prevalence and patterns in an ethnically diverse sample | Browne, Angela | 1997 | American Journal of Orthopsychiatry | Quantitative, observational, case-control design | Cross-sectional analysis of baseline data from an ongoing longitudinal case-comparison study | Chi-square and Fisher’s exact tests, logistic regression modeling | 436 mothers | Low-income female heads of households | Homeless mothers were randomly enrolled from nine emergency/transitional shelters and two motels. Housed mothers were randomly selected from women visiting the Department of Public Welfare office to discuss welfare (AFDC) eligibility. | Worcester, Massachusetts | Mean: 27.3 years; range: 15 - 30+ | 100% female | 39.0% White (Anglo), 36.5% Puerto Rican, 16.5% Black, and 8.0% Other | NA | NA | NA | Extreme poverty; mean annual income was $8,982; all participants were receiving Aid to Families with Dependent Children | Yes | physical aggression scale of the Conflict Tactics Scales (CTS) | Self-report of severe physical violence victimization since age 17 or since the respondent was "on her own" | Severe physical violence (e.g., kicked, bit, hit with fist/object, beaten up, choked, threatened with weapon) and verbal threats of harm | Yes | Conflict Tactics Scales and specific adapted questions for sexual molestation | Severe physical violence by a childhood caretaker (before age 18) and childhood sexual molestation (advances or unwanted experiences before age 18) | Homelessness (residing in a shelter) | Enrollment criteria based on current residential status (shelter vs. community housing) | Homelessness defined as residing in a shelter for at least seven consecutive days. Housed participants had to have no history of homelessness | Homelessness increased the risk of needing medical treatment for IPV and any lifetime violence by intimates. | Homelessness and needing medical treatment for IPV: Adjusted Odds Ratio (AOR) = 1.82 (95% CI, 1.11–2.98); Homelessness and any violence by intimates across the lifespan: Significant association (87.6% for homeless vs. 78.6% for housed, p = .01) | Homelessness and Severe Physical Violence by Any Caretaker: AOR = 1.13 (95% CI, 0.75–1.71); Homelessness and Sexual Molestation (Childhood): AOR = 0.99 (95% CI, 0.66–1.49); Homelessness and IPV by Current/Most Recent Partner: AOR = 0.91 (95% CI, 0.59–1.40); specific p-values not reported | No | NA | limited generalizability (specific to Worcester and small number of Black individuals; some types of violence, such as sibling violence or abuse by non-intimate adult relatives, were not assessed | ||
39 | https://doi.org/10.1002/jcop.22978 | Patterns and predictors of material hardship among poor families with children | Marçal, Katherine E. | 2023 | Journal of Community Psychology | Quantitative; longitudinal and observational | secondary data from the Fragile Families and Child Well-being Study, an ongoing prospective longitudinal survey; used data from mothers at Year 3 and Year 5 | Latent Class Analysis (LCA), multinomial logistic regression, and Multiple Imputation with Chained Equations (MICE) | 1,576 mothers | Low-income families with young children; mothers with earned household incomes below the federal poverty level who had at least partial custody of their children | Recruited in hospitals after birth, with an intentional oversample of unmarried parents | 20 large American cities | Mean at child's birth: 23.90 (SD = 5.60) | 100% female | 57.5% Black, 30.3% Hispanic, 9.7% White, and 2.5% Other | NA | NA | NA | All participants were living below the federal poverty level | Yes | seven ordinal items adapted from Lloyd (1997) | Self-report scores about current relationship. Higher scores indicated greater levels of IPV. | violent, coercive, or psychologically harmful behaviors (e.g., insulting/criticizing, preventing visits with friends/family, preventing work/school, withholding money, slapping, kicking, hitting, and sexual violence) | No | NA | NA | Homelessness, eviction, doubling up due to inability to afford housing, and missing a rent or mortgage payment | Survey on Income and Program Participation (SIPP) and the New York City Social Indicators Survey (SIS) | Self-report by mothers regarding their experiences over the past year at the Year 5 interview | IPV victimization at Year 3 increased the likelihood of a family being in the "Housing Insecure" class at Year 5 compared to the "Cost-Burdened but Managing" class | IPV victimization predicted membership in the "Housing Insecure" class relative to the "Cost-Burdened but Managing" class: OR = 1.107, 95% CI [1.029, 1.190], p < 0.01 | None. | No | NA | household expenses like childcare or transportation were excluded; did not account for enrollment in assistance programs; self-reports, so potential social desirability bias; not generalizable to suburban or rural populations | IPV victimization did not significantly distinguish between families who became "Housing Insecure" versus those who became "Food Insecure" (OR = 1.010, 95% CI [0.915, 1.116]) |
40 | https://doi.org/10.1007/s10834-015-9465-0 | The roles of instability and housing in low-income families’ residential mobility | Kull, Melissa A. | 2016 | Journal of Family and Economic Issues | Quantitative; longitudinal and observational | Three waves of data from the Three City Study, a longitudinal multi-method study over a 6-year period. | Multilevel mixed negative binomial regression | 2393 women | Low-income families (household income less than 200% of the federal poverty line) with a female caregiver and at least one child who is 0-4 or 10-14 | stratified random sampling procedure in moderate-poverty and high-poverty neighborhoods | Boston, MA, Chicago, IL, San Antonio, TX | Mean: 33.71 years (SD = 9.71) | 100% female | 53% Hispanic, 40% Black, and 6% White/other | NA | NA | NA | All families had incomes less than 200% of the federal poverty line | Yes | Conflict Tactics Scale (CTS). More specifically, 13 physical violence items (e.g., being threatened, shoved, or slapped) | An overall dichotomous indicator was created capturing any experiences of IPV in the 12 months prior to the interview | Physical IPV | Yes | Maternal self-report regarding their own history of abuse (no specific items listed) | Physical or sexual abuse | Residential mobility (frequency of moves) | Mothers reported at Waves 2 and 3 the number of times they had moved since the previous wave | Count of moves between waves | Experiences of intimate partner violence (IPV) were significantly associated with a higher rate of residential moves. Mothers' own experiences of childhood domestic violence also predicted a higher rate of mobility above and beyond recent IPV. | Recent IPV predicted a higher rate of residential moves: IRR = 1.22, p < 0.01 (Model 2, controlling for family characteristics) and IRR = 1.20, p < 0.01 (Full Model 4). This indicates a 20–22% increase in the rate of mobility for those experiencing IPV. Childhood domestic violence (maternal history) was a significant predictor of a higher rate of residential moves: IRR = 1.26, p < 0.01 (Model 1) and IRR = 1.17, p < 0.01 (Full Model 4). This corresponds to a 17–26% higher rate of mobility. | None. | No | NA | 6-year snapshot may miss influential earlier life course experience; cross-sectional, so no causality; data could not distinguish where families moved (e.g., within vs. between neighborhoods); findings are specific to urban low-income families at the height of the housing bubble | |
41 | https://doi.org/10.1080/01488376.2015.101866 | Physical and psychological aggression toward a child among homeless, doubled-up, and other low-income families | Park, Jung Min | 2015 | Journal of Social Service Research | Quantitative; longitudinal and observational | Data from three waves of the Fragile Families and Child Well-Being Study (FFS), specifically when the children were 3, 5, and 9 years old. It is a prospective cohort study | Logistic regression with generalized estimating equations | 2332 mothers | Low-income families living at or below the federal poverty threshold in at least one wave | Recruited in hospitals after birth, with an intentional oversample of unmarried mothers | 20 U.S. cities | Mean at child's birth: 23.90 (SD = 5.60) | 100% female | 56% Black, 30% Hispanic, and 15% Immigrant | NA | NA | NA | all incomes at or below the federal poverty threshold | Yes | Mothers were asked if they were slapped or hit by the child's father or seriously hurt in a fight with him in the previous 12 months | Self-report of any of the physical violence items | Physical IPV | Yes | Parent–Child Conflict Tactics Scales (PC-CTS) | Physical aggression (shaking, hitting with object, spanking, slapping, pinching) and psychological aggression (shouting, swearing, name-calling, threatening to send away) towards their child | Homelessness (shelter, street, or temporary housing) and doubling up (living with relatives/friends due to economic hardship) | Maternal self-report at each interview wave | Homelessness: stayed in a shelter/place not meant for housing in the past 12 months. Doubling up: moved in with others because of financial problems in the previous 12 months | Homelessness was significantly associated with a higher frequency of physical and psychological aggression toward a child compared to housed low-income families. Doubling up was not significantly associated with adverse parenting behaviors once other demographic and stress factors were controlled for in the multivariate model. Mothers with a homeless episode were more than twice as likely to experience domestic violence (22%) compared to the low-income housed group (10%) | Homelessness predicted a high level of physical aggression toward the child: OR = 1.86, p < 0.01. Mothers with a homeless episode were more than twice as likely to experience domestic violence (22%) compared to the low-income housed group (10%) (p < .001) | Doubling up did not significantly predict physical (OR = 1.10) or psychological (OR = 0.86) aggression towards child in the multivariate model. | No | NA | self-reports, so risk of social desirability bias; lacked data on the severity and frequency of housing instability episodes; could not determine if homeless families differed in parenting prior to becoming homeless | |
42 | https://doi.org/10.1007/s11524-016-0049-6 | Early Life Psychosocial Stressors and Housing Instability among Young Sexual Minority Men: the P18 Cohort Study | Krause, Kristen D. | 2016 | Journal of Urban Health: Bulletin of the New York Academy of Medicine | Quantitative; cross-sectional analysis of baseline data from a longitudinal cohort study | Data from the baseline visit of the Project 18 Study (P18), a cohort study | chi-square tests, multivariable logistic regression | 598 young sexual minority men | young sexual minority men | community outreach (flyers, Internet advertisements) and venue-based methods (community centers, college campuses, bars/clubs) | New York City | Range: 18-19 years old | 100% born biologically male | 38.3% Hispanic/Latino, 29.9% White non-Hispanic, 14.2% Black non-Hispanic, 5.2% Asian non-Hispanic, and 12.5% Mixed Race/Other non-Hispanic | NA | NA | NA | Perceived familial socioeconomic status was 32.2% lower class, 37.4% middle class, and 30.5% upper class | No | NA | NA | NA | Yes | Self-report items regarding Adverse Childhood Experiences (ACEs) | Lack of basic needs (neglect of food, hygiene, or clothing), physical abuse (kicking, hitting, or punching by a parent/caregiver), and sexual abuse by a parent/caregiver | Sleeping in unstable housing (e.g., street, park, shelter, abandoned building) and living in unstable housing (e.g., temporary/transitional housing, jail, drug treatment, halfway house, or temporarily with family/friends) | ACASI asking participants about their housing situations in the 6 months preceding the assessment | Reporting any of the specified unstable sleeping or living situations in the 6 months prior to assessment | Childhood neglect (lack of basic needs) significantly increased the odds of currently sleeping in unstable housing. Childhood physical abuse was significantly associated with currently living in unstable housing. | Young men who experienced neglect between 1 and 10 times during their childhood had significantly higher odds of sleeping in unstable housing (AOR = 2.52, 95% CI [1.11, 5.72], p < 0.05), and these odds increased to over five times for those who experienced neglect more than 10 times (AOR = 5.47, 95% CI [1.82, 16.45], p < 0.05). Participants who reported being physically abused between 1 and 10 times as children were significantly more likely to be currently living in unstable housing (AOR = 1.92, 95% CI [1.10, 3.36], p < 0.05), while the odds for those who experienced physical abuse more than 10 times were more than three and a half times higher (AOR = 3.51, 95% CI [1.85, 6.09], p < 0.001). Physical abuse was correlated with both sleeping (p=0.013) and living (p<0.001) unstably; sexual abuse was correlated with living unstably (p=0.007). | Childhood Physical Abuse did not significantly predict sleeping in unstable housing in the multivariable model. Childhood Sexual Abuse was not a significant predictor of living in unstable housing in the multivariable model. | No | NA | cross-sectional, so no causality; non-probability recruitment methods may limit generalizability; self-report, so sucject to social desirability bias; lack of measures addressing the specific reasons why participants became unstably housed | |
43 | https://doi.org/10.1016/j.chiabu.2015.06.002 | Housing mobility and cognitive development: Change in verbal and nonverbal abilities | Fowler, Patrick J. | 2015 | Child Abuse & Neglect | Quantitative, longitudinal, observational study | Data from the National Survey of Child and Adolescent Well-Being; tracked families over 36 months with assessments at baseline, 18-month follow-up, and 36-month follow-up | Latent Growth Modeling | 2442 families | nationally representative sample of families under investigation for child abuse and neglect who remained together following the investigation | two-stage stratified probabilistic sampling design was used; first, the U.S. was divided into nine strata (eight states with the largest child welfare caseloads and one stratum for the remaining 42 states/DC). Second, families were randomly selected from monthly investigation lists generated by 92 primary sampling units | United States | Mean: 34.47 years (SD = 8.32) | 89% female | 47% White, 27% African American, 18% Hispanic, and 8% other ethnicities | Mean: 8.95 (SD = 3.26) at baseline; range 4 - 16 years | NR | 47% White, 27% African American, 18% Hispanic, and 8% other ethnicities | 56% of families earned below $20,000 annually, and one-third were below the federal poverty level | No | NA | NA | NA | Yes | caregiver self-reports using the Parent-Child Conflict Tactics Scale (CTS-PC) and caseworker-reported primary reasons for investigation (official allegations) | Physical abuse, sexual abuse, emotional abuse, and neglect | Housing mobility (defined by the number of residential moves) | Baseline: Caregiver report of the number of moves in the last 12 months; Follow-ups: Tracking address changes (fuzzy matching algorithm) between interviews; Early mobility: Caregiver report of moves during the child's first 12 months of life | Self-report of mobility | Housing mobility was significantly positively correlated with caregiver-reported neglect and being investigated for neglect and physical abuse. | Housing mobility was significantly positively correlated with caregiver-reported neglect (r=.083,p<.01) and being investigated for neglect (r=.071,p<.01) and physical abuse (r=−.079,p<.01). | There was no statistically significant correlation identified between housing mobility and caregiver self-reported physical abuse (r = .024) and investigations for sexual abuse (r = -.030) or other types of abuse (r=.034) | No | NA | observational design limits ability to account for selection processes, retrospective reports may elad to memory and recall biases; gap in housing histories between infancy and initial investigation; study lacked data on infants and toddlers; findings may not be generalizable beyond families investigated for child abuse and neglect | |
44 | https://doi.org/10.1016/j.chiabu.2022.10569 | Domains of housing insecurity: Associations with child maltreatment risk | Marçal, Katherine E. | 2022 | Child Abuse & Neglect | Quantitative, longitudinal, observational study | Data from the Fragile Families and Child Well-being Study, focused on mothers who retained custody when children were aged 5 (predictor wave) and 9 (outcome wave) | Confirmatory Factor Analysis, Path Analysis | 1,804 mothers | Low-income, predominantly nonwhite, urban mothers | Recruited in hospitals after birth using stratified clustered sampling, with an intentional oversample of unmarried mothers | 20 U.S. cities | Mean: 24.99 years (SD = 5.96) | 100% female | 53.63% Black, 23.90% Hispanic, 19.63% White, and 3% other ethnicities | NA | NA | NA | mean household income of $36,035 at Year 5 | No | NA | NA | NA | Yes | Mother self-reports using the Parent-Child Conflict Tactics Scale (CTS-PC) | Psychological abuse, physical abuse, and neglect | Housing quality: physical environment of the home (e.g., adequate lighting, a safe interior for children, absence of broken windows); Housing stability: ability to remain consistently housed in a single location and avoid displacement (e.g., doubling up, experiencing homelessness); housing affordability: ability to pay for housing costs | Quality: Interviewer observation using nine items from the HOME Inventory (e.g., broken windows, lighting, safety); Stability: Mother self-reports on eviction, doubling up, or homelessness in the past year; Affordability: Mother self-reports on missing rent/mortgage payments, missing utility payments, or borrowing money to pay bills | Self-report of housing insecurity | Housing affordability was significantly associated with reduced psychological and physical abuse, but not neglect. Neither housing quality nor housing stability had a significant association with any of the three maltreatment types when controlling for other household factors. | Affordability and psychological abuse: β=−1.109,p<0.05,95% CI [−2.092,−0.063]; Affordability and physical abuse: β=−0.660,p<0.05,95% CI [−1.519,0.052] (While the p-value was <.05, the bootstrapped confidence interval crossed zero; the authors reported this as significant) | Quality and Psychological Abuse: β=−0.087,95% CI [−0.426,0.119]; Stability and psychological abuse: β=1.026,95% CI [−0.230,1.673]; Quality and physical abuse: β=0.010,95% CI [−0.165,0.166]; stability and physical abuse: β=0.497,95% CI [−0.434,1.235; affordability and neglect: β=−0.746,95% CI [−1.699,−0.081]; quality and neglect: β=0.043,95% CI [−0.248,0.223] stability and neglect: β=0.527,95% CI [−0.225,1.593] | No | NA | limited to large U.S. cities, so may not be generalizable; mothers without custody at Years 5 or 9 (including those whose children were removed by child welfare) were excluded, so most severe maltreatment cases may have been omitted; data collected before COVID-19 pandemic | |
45 | https://doi.org/10.1080/10522158.2018.1469563 | The Impact of Housing Instability on Child Maltreatment: A Causal Investigation | Marçal, Katherine E. | 2018 | Journal of Family Social Work | Quantitative, longitudinal, observational study using causal inference techniques (Propensity Score Analysis) | Data from the Fragile Families and Child Well-being Study, focusing on the baseline, 5-year, and 9-year follow-up interviews; two propensity score analysis approaches (greedy matching and propensity score weighting) were used to try to establish a causal explanation between housing instability at Year 5 and maltreatment at Year 9 | Propensity Score Analysis (Greedy Matching and Propensity Score Weighting), Logistic Regression (to generate propensity scores), and Ordinary Least Squares (OLS) multivariate linear regression | 2,284 mothers in the original sample. The post-greedy matching analytic sample included 848 mothers (424 matched pairs) | Low-income, predominantly nonwhite, urban mothers | Recruited in hospitals after birth using stratified clustered sampling, with an intentional oversample of unmarried mothers | 20 U.S. cities | Uncorrected sample, mean age of mothers who did not experience housing instability was 25.29 years (SD=6.11), while those who did was 23.89 years (SD=5.05) at the time of the child's birth | 100% female | 49.6% Black, with other mothers identified as White, Hispanic, or "Other" | NA | NA | NA | 38.2% lived below the federal poverty level at Year 5, and 34.8% at Year 9 | Yes | dichotomous indicator of whether mothers reported having been hit, kicked, or slapped by the child’s father or a current partner in the past year | Self-report of specific physical acts (hitting, kicking, slapping) | Physical violence (hitting, kicking, slapping) | Yes | Mother self-reports using a modified version of the Parent-Child Conflict Tactics Scale (CTSPC) | Physical abuse, psychological aggression (e.g., cursing, shouting, name-calling), and neglect | Homelessness, eviction, "doubling up" with others, and difficulty affording rent, mortgage, or utility payments | Mother self-reports at the Year 5 interview | Self-report of at least one of five housing-related risks (difficulty affording rent/utilities, eviction, doubling up, or homelessness) | Housing instability was found to lead to a small, statistically significant increase in child maltreatment behaviors. Domestic violence was found to be a significant co-occurring risk factor that was much more common among families experiencing housing instability compared to those who were stably housed | Housing instability and child maltreatment (matched sample): b=0.65,SE=0.20,p<.01. Housing instability and child maltreatment (weighted sample): b=0.80,SE=0.17,p<.001. Domestic Violence and Housing Instability (bivariate/demographics): A significant difference was found between the groups before matching (χ2=26.51, p<.001). Domestic Violence and Housing Instability (unweighted regression): b=1.27,SE=0.26,p<.001. | Domestic Violence and Housing Instability (post-propensity score weighting): After applying weights to achieve sample balance, the relationship was no longer statistically significant (b=0.14,SE=0.32). | No | NA | Greedy matching significantly reduced sample size, potentially limitng generalizability; sample exclusively urban, so may not generalize to rural communities; self-report may lead social desirabolity bias; no data on other caregivers besides mother | |
46 | https://doi.org/10.1016/j.chiabu.2022.105792 | The mediating role of housing insecurity in the relationship between neighborhood social cohesion and child maltreatment among low-income urban families | Maguire-Jack, Kathryn | 2022 | Child Abuse & Neglect | Quantitative, longitudinal, observational study | Data from the Fragile Families and Child Well-being Study, focusing on Waves 3, 4, and 5; neighborhood social cohesion when children were age 3, housing insecurity at age 5, and child abuse behaviors at age 9 | Structural Equation Modeling, Confirmatory Factor Analysis, Path Analysis | 3,123 mothers | Low-income, predominantly nonwhite, urban mothers | Recruited in hospitals after birth using stratified clustered sampling, with an intentional oversample of unmarried mothers | 20 U.S. cities | mean at the time of the child's birth: 25.20 years (SD=6.02) | 100% female | 50.34% Black, 24.18% Hispanic, 22.00% White, and 3.49% some other race | NA | NA | NA | average household income was $36,222 at Year 5 | No | NA | NA | NA | Yes | Mother self-reports using the Parent-Child Conflict Tactics Scales (CTSPC) | Psychological abuse and physical abuse | Housing insecurity (missing rent/mortgage payments, eviction, missing utility payments, electricity being turned off, doubling up, and homelessness) | Mother self-reports on six dichotomous items regarding their experiences in the past year | Self-report of housing insecurity | Housing insecurity was significantly associated with higher levels of psychological and physical abuse. Housing insecurity partially mediated the relationship between neighborhood social cohesion at age 3 and physical and psychological abuse at age 9. | Housing Insecurity and Psychological Abuse: β=0.168,p<.001,95% CI [0.108,0.214]. Housing Insecurity and Physical Abuse: β=0.148,p<.001,95% CI [0.075,0.206]. Indirect Effect (Cohesion → Housing → Psychological Abuse): β=−0.012,p<.05,95% CI [−0.025,−0.002]. Indirect Effect (Cohesion → Housing → Physical Abuse): β=−0.011,p<.05,95% CI [−0.022,−0.002]. | None. | No | NA | urban areas, so may not generalize to rural settings; self-reported data (subject to social desirability bias); study did not specifically model residential moves parents may have made; no standard definition or measurement tool for housing insecurity used | |
47 | https://doi.org/10.1016/j.chiabu.2019.02.005 | Identifying underlying constructs of childhood adversity in a low-income population | Cohen-Cline, Hannah | 2019 | Child Abuse & Neglect | Quantitative; cross-sectional survey analysis | Cross-sectional study using a retrospective life-course survey where participants recalled experiences from two age ranges: childhood (6–12 years) and adolescence (13–18 years) | Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA), utilizing tetrachoric correlations and oblique (ObliMin) rotations | 1,686 respondents for the 6–12 age range and 1,613 for the 13–18 age range | Low-income, adult Medicaid members | A representative sample was selected from eligible Medicaid members and mailed surveys with a $5 incentive; follow-up included additional mailings, telephone calls, and door-to-door visits for a subset | Portland, Oregon | 30 and under (14.7%), 31–40 (19.1%), 41–50 (19.1%), 51–60 (28.3%), and over 60 (18.3%); range: 18 - 65 years | 63% female | White (65%), Black (14%), Hispanic (8%), and Other (21%) | NA | NA | NA | The population was low-income (Medicaid-enrolled) | No (only witnessing IPV between caregivers) | NA | NA | NA | Yes | Self-report survey questions adapted from standard ACE measures for abuse (verbal, physical, sexual) and neglect | Verbal/emotional abuse, physical abuse, sexual abuse, and neglect | Homelessness | Self-report survey question: "Were there times when your family (or just you, if you were on your own) had no regular place to live – like living on the street or in a car – for at least a few weeks?" | Binary self-report (Yes/No) for experiences lasting "at least a few weeks" | In both age groups, housing instability and witnessing violence between caregivers (IPV) loaded onto the same underlying construct, which the authors labeled "Instability." Neglect (feeling unprotected) was also found to associate with the Instability factor in the final CFA models for both age ranges. | In the 6–12 age group, witnessing violence between caregivers and housing instability both showed significant factor loadings on the Instability factor, at 0.710 and 0.878, respectively. In the 13–18 age group, witnessing violence between caregivers and housing instability had significant loadings on the Instability factor at 0.615 and 0.848, respectively. Additionally, the correlation between the "Support" factor (which includes emotional neglect) and the "Stability" factor (which includes IPV and housing instability) was significant at 0.454 for ages 6–12 and 0.371 for ages 13–18 | Specific forms of child maltreatment—including verbal/emotional, physical, and sexual abuse—did not load onto the same construct as housing instability, as their factor loadings were below the 0.5 threshold and they did not form their own distinct factor in this population. | No | NA | Potential recall or response bias; the survey may have omitted other important adversities; the study did not measure the frequency or intensity of experiences; and the results may not be generalizable due to a low response rate (26%) and potential lack of representativeness | |
48 | https://doi.org/10.1016/j.chiabu.2013.10.022 | Fatal and non-fatal child maltreatment in the US: an analysis of child, caregiver, and service utilization with the National Child Abuse and Neglect Data Set | Douglas, Emily M. | 2014 | Child Abuse & Neglect | Quantitative; national comparison study using secondary analysis of administrative data | Cross-sectional analysis of the 2010 National Child Abuse and Neglect Data Set (NCANDS) child file | chi-square tests, t-tests, and Šidák correction method | 202465 children | Children and families involved with child welfare services | Data was obtained from the NCANDS child file, which aggregates child-specific data from investigated reports to state child welfare agencies | United States | For perpetrators, the mean age was 29.50 years for fatal cases and 32.76 years for non-fatal cases | 40.2 % male | NA | Fatal victims: 47.7% were under age 1 and 79.4% were under age 4 | 60.1% of victims were male | For fatal cases, the sample was 66.8% White and 25.0% "Ethnicity" | 29.5% overall received public assistance (TANF), and 30.1% reported financial problems | Yes | Binary reporting (Yes/No) of "presence of domestic violence in the home" within the NCANDS administrative records | noted in the state child protective service agency investigation | Domestic violence | Yes | Substantiated reports of abuse or neglect within the NCANDS dataset | Neglect, physical abuse, sexual abuse, emotional abuse | Inadequate housing | Binary reporting (Yes/No) within the caretaker social/behavioral characteristics section of the NCANDS data | "Inadequate housing" as identified during a child welfare investigation | Families of children who were fatally maltreated experienced more financial and housing instability than families of non-fatally maltreated children | inadequate housing was significantly more prevalent in fatal maltreatment cases (28.7%) than in non-fatal cases (24.9%) (p = .009) | None. | No | NA | The study used archival data with limited categorical measures; the NCANDS dataset is not a statistically representative census of all US fatalities; state and county data were masked, preventing regional analysis; and the analysis was limited to the bivariate level, so relationships might change in a multivariate model | |
49 | https://doi.org/10.1007/s00127-012-0551-4 | Conduct disorder behaviors, childhood family instability, and childhood abuse as predictors of severity of adult homelessness among American veterans | Tsai, Jack | 2013 | Social Psychiatry and Psychiatric Epidemiology | Quantitative; longitudinal secondary analysis | Longitudinal analysis of a combined national observational and experimental dataset; childhood factors were assessed retrospectively at baseline, with a follow-up assessment conducted one year after enrollment in supported housing | Multiple regression analyses (conducted in blocks to control for sociodemographics and mental health); Pearson correlations | 1,161 participants | Homeless veterans with a psychiatric and/or substance use disorder | Veterans were referred by clinicians in specialized VA homeless programs and entered the HUD-VASH program across 19 sites | United States | Mean: 42.97 (SD = 7.99) | 94.92% male | Black (47.55%), White (44.53%), Hispanic (5.08%), and Asian/Indian/Alaskan (1.81%) | NA | NA | NA | All were homeless at entry | No | NA | NA | NA | Yes | single dichotomous (Yes/No) self-report item | Physical, sexual, or emotional abuse before the age of 18 | Adult homelessness (lifetime duration, number of episodes, and past 3-month severity); childhood family instability (moving often, parents missing) | Structured clinical interviews at baseline and every 3 months; duration/episodes reported in years/times | participants had to be living in a shelter or on the street for at least 30 days | At baseline, both childhood abuse and family instability were weakly associated with a higher number of lifetime homeless episodes. However, one year after receiving supported housing and case management (HUD-VASH), childhood problems were not predictive of housing outcomes. | Before entry into the program, childhood abuse was significantly associated with the number of lifetime homeless episodes (β = 0.10, p < 0.01). Childhood family instability was significantly associated with a higher number of lifetime homeless episodes (β = 0.12, p < 0.01) | Childhood abuse was not significantly associated with the total duration (years) of lifetime homelessness (β = 0.04) or the number of nights homeless in the past month (β = -0.03) at baseline. One year after enrollment in HUD-VASH, childhood abuse remained not predictive of nights homeless (β = 0.07) or nights housed (β = -0.06). Childhood family instability was also not significantly associated with the duration of lifetime homelessness (β = 0.06) or nights homeless in the past month (β = -0.00) at baseline. | Yes | HUD-VASH | retrospective self-report, so subbject to recall bais; childhood abuse was measured by a single item; older measures were used; sample size of women was too small for separate statistical analysis | |
50 | https://doi.org/10.1042/peds.2024-068174 | State-Based Eviction Moratoria and Child Maltreatment During the COVID-19 Pandemic | Zhang, Liwei | 2025 | Pediatrics | Quantitative; quasi-experimental difference-in-differences analysis | Difference-in-differences strategy comparing 185 "always treated" counties (states with continuous moratoria from mid-March 2020 to mid-August 2021) to 133 "never treated" counties (states with no moratoria) | Difference-in-differences analyses using linear regression models with state and time fixed effects; event study models were used to test the parallel trends assumption | 318 counties | Counties with at least 1,000 child maltreatment reports | Secondary analysis of administrative data from the National Child Abuse and Neglect Data System (NCANDS) and the COVID-19 Eviction Moratoria and Housing Policy Database (EMHPD) | 17 US states | NR | NR | NR | NA | NA | NA | Focused on financially struggling and low-income families at risk of eviction | No | NA | NA | NA | Yes | Administrative data from NCANDS measuring biweekly counts of screened-in CPS reports | Physical abuse, sexual abuse, and neglect | Eviction | Policy data from the EMHPD, indicating whether a state-level moratorium was in place during a biweekly period | Continuous implementation of a state-based eviction moratorium | Eviction moratoria were significantly associated with reductions in reports of physical abuse, sexual abuse, and neglect. | Eviction moratoria were significantly associated with reduced biweekly reports of physical abuse (b=−0.073; 95% CI, -0.119 to -0.027; p<0.01), representing a 16.04% reduction. Sexual abuse reports were also significantly reduced (b=−0.034; 95% CI, -0.051 to -0.018; p<0.001), a 21.12% reduction. Neglect reports showed a significant reduction (b=−0.217; 95% CI, -0.346 to -0.088; p<0.01), representing a 12.17% reduction. | the effect of eviction moratoria on the overall child maltreatment report rate was not statistically significant (b=−0.074; 95% CI, -0.271 to 0.123) | Yes | State-Based Eviction Moratoria | Reliance on official CPS reports which may not capture all actual maltreatment; potential influence of other unmeasured COVID-19 federal/state policies (e.g., tax credits); reporting bias due to school instructional modes (virtual vs. in-person); the possibility that moratoria delayed rather than fully prevented evictions and subsequent maltreatment | |
51 | https://doi.org/10.3390/ijerph14101197 | Predicting Child Protective Services (CPS) Involvement among Low-Income U.S. Families with Young Children Receiving Nutritional Assistance | Slack, Kristen S. | 2017 | International Journal of Environmental Research and Public Health | Quantitative; exploratory study using a linked survey and administrative dataset | Longitudinal; survey data (Family Support Study) from late 2010/early 2011 was linked to 18 months of prospective administrative data on child welfare involvement | Logistic regression, Chi-square tests, and one-way ANOVAs | 1,065 completed surveys | Low-income families with young children receiving WIC benefits | Survey packets were distributed by staff at 22 WIC offices across Wisconsin and through a home visiting program in Milwaukee | Wisconsin | 36.6% of parents were under 25 years old | NR | Non-Hispanic White/Other (65.4%), Non-Hispanic Black (10.6%), and Hispanic (19.4%) | NA | NA | NA | All participants were below 185% of the federal poverty line. The average annual income for working groups was less than $18,000. | No | NA | NA | NA | Yes | Administrative records of investigated child maltreatment reports | All types of maltreatment leading to a CPS investigation | Housing moves | Self-report survey question measuring more than one housing move in the past year | Binary indicator for experiencing more than one move | Housing instability was a significant predictor of CPS involvement. Families that experienced more than one move in the past year were significantly more likely to be investigated for child maltreatment. | Housing instability was significantly associated with an increased risk of a CPS investigation (OR = 2.02, p < 0.05). The "Work and Welfare" group (which had a lower CPS risk than the welfare-only group) was the most likely to experience housing instability (24.3%) compared to other groups (p < 0.01). | None. | No | NA | Lacked information on spouse/partner income or informal supports; relied on investigated reports which may under-identify maltreatment; the response rate was moderate (49.5%); findings were correlational, so no causality | |
52 | https://doi.org/10.1177/10778012231159418 | Resilience and Challenges Among Women With Histories of Addiction and Victimization Newly Admitted to a Sober Living Home: A Brief Report | Edwards, Katie M. | 2024 | Violence Against Wome | Quantitative brief report; cross-sectional | Cross-sectional design utilizing baseline data collected within two weeks of participants' arrival at a sober living home | Pearson’s r | 44 women | Women with histories of substance use disorders (SUDs) and domestic and sexual violence (DSV) victimization newly admitted to a sober living home | Purposive sampling over a 28-month period; the house manager provided recruitment letters to newly admitted residents who then contacted program staff if they wished to participate | Phoenix, Arizona | Mean: 40.3 years (SD = 12.53); Range: 22–67 | 100% female | 90.1% White non-Hispanic | NA | NA | NA | 77.3% reported a household income under $20,000 and 63.6% were unemployed | Yes | single dichotomous item was used to assess the presence of DSV victimization within the past 6 months | women with histories of DSV victimization recently admitted to a specialized sober living home | Domestic and sexual violence (DSV) victimization | No | NA | NA | Housing instability (e.g., not having enough money for rent and landlords threatening eviction) | Housing Instability Index, a 10-item scale developed for survivors of domestic violence | Being in a shelter | Recent DSV victimization (past 6 months) was found to be unrelated to housing instability | None. | Recent DSV victimization (past 6 months) was found to be unrelated to housing instability (r = .22) | Yes | SEED | small and non-diverse sample, so may not be generalizable; cross-sectional, so unable to assess changes over time; reliance on self-report, so subject to bias | |
53 | https://doi.org/10.1016/j.chiabu.2023.106438 | Material hardship in the postpartum year: Links to child maltreatment | Marçal, Katherine E. | 2023 | Child Abuse & Neglect | Quantitative, observational, longitudinal study | Data from the Fragile Families and Child Well-being Study, focusing on baseline, year 1, and year 3 waves | Latent Class Analysis, multinomial regression, and the Bolck et al. (BCH) approach for distal outcomes | 3,966 mothers | Low-income, predominantly nonwhite, urban mothers | Recruited in hospitals after birth using stratified clustered sampling, with an intentional oversample of unmarried mothers | 20 U.S. cities | Mean: 25.21 years (SD = 6.05) | 100% female | 7.93% Black, 25.95% Hispanic, 22.44% White, and 3.68% Other | NA | NA | NA | 34.80% of households were living below the federal poverty level | No | NA | NA | NA | Yes | Parent-Child Conflict Tactics Scale (CTSPC) | Physical assault (e.g., spanking, hitting, shaking), psychological aggression (e.g., shouting, swearing, threats), and neglect (e.g., leaving child home alone, inability to provide food or medical care) toward their children | Missed rent/mortgage payments, doubling up (moving in with others), eviction, and homelessness (spending a night in a shelter or on the street) | Eight dichotomous items derived from the Survey on Income and Program Participation (SIPP) and the New York City Social Indicators Survey (SIS) | Self-reported experience of the hardship in the past 12 months "because there was not enough money" | Difficulty maintaining stable and affordable housing in the postpartum year is associated with an elevated risk of maltreating their children. Mothers categorized as "Severely Housing Insecure" exhibited the highest levels of physical assault, psychological aggression, and neglect towards their children compared to "Stable" mothers | Mothers in the "Severely Housing Insecure" class (high probability of doubling up, eviction, and homelessness) were significantly more likely to engage in physical assault (χ2 =16.25,p<0.001), psychological aggression (χ2=11.15,p<0.01), and neglect (χ2 =17.55,p<0.001) compared to "Stable" mothers. Mothers in the "Cost-Burdened" class (high level of missed rent and utility payments) were significantly more likely to engage in psychological aggression (χ2 =5.04,p<0.05) and neglect (χ2=4.46,p<0.05) than "Stable" mothers. | No significant differences were found in physical assault between "Cost-Burdened" and "Stable" mothers (χ2 =3.21). The "Housing Insecure with Hunger" class did not significantly differ from the "Stable" class in physical assault (χ2 =0.54), psychological aggression (χ2 =2.56), or neglect (χ2 =1.24). | No | NA | subtypes of material hardship were assessed dichotomously, which does not capture severity or frequenc; urban population, so may not be generalizable to suburban or rural populations | |
54 | https://doi.org/10.1016/j.chiabu.2020.104472 | Variability and stability in child maltreatment risk across time and space and its association with neighborhood social & housing vulnerability in New Mexico: A bayesian space-time model | Barboza-Salerno, Gia Elise | 2020 | Child Abuse & Neglect | Quantitative, observational, longitudinal study | A Bayesian space-time model tracking child maltreatment substantiations at the census tract level over a 9-year period (2007–2015) | Bayesian hierarchical space-time modeling using Integrated Nested Laplace Approximation | 499 census tracts | Statewide community population (all children in New Mexico) | secondary administrative data from the Protective Services Division of New Mexico’s Department of Child Youth & Families | New Mexico, USA | NR | NR | Only neighborhood-level minority percentage (Mean = 58.9%) and Diversity Index (Mean = 63.4) were reported | NR | NR | NR | 32% of children lived in households with high housing cost burden, 36% had parents lacking secure employment, and 24% lived in high-poverty areas; mean poverty rate of 20.8% and a mean unemployment rate of 9.8% | No | NA | NA | NA | Yes | Administrative counts of substantiated child abuse and neglect reports | Physical abuse and neglect | Eviction rate and rent burden (defined as spending 35% or more of income on rent) | Court records for evictions from the Princeton Eviction Lab; rent burden data from the American Community Survey (ACS) | Ecological measures at the census tract leve | Higher neighborhood eviction rates and rent burden were significantly associated with an increased risk of substantiated child maltreatment. Census tracts with increasing trends in child maltreatment over the 9-year period were more likely to have higher eviction rates. | Every one-unit increase in the neighborhood eviction rate was associated with an 8.7% increase in the risk of child abuse and neglect (Incidence Density Ratio [IDR]=1.088,95% CrI [1.01,1.112]). A one standard deviation increase in the eviction rate increased the relative risk of maltreatment by 27.2%. Rent burden was significantly associated with an increased risk of substantiated child maltreatment (IDR=1.11,95% CrI [1.01,1.13]). Census tracts exhibiting a positive (increasing) differential time trend in child maltreatment were found to have significantly higher eviction rates (p<0.001). | None. | No | NA | ecological design prevents evaluating individual or family-level factors; potential bias in administrative data (e.g., increased visibility of poor families to reporters); study was unable to disaggregate findings by specific child maltreatment subtypes | |
55 | https://doi.org/10.1016/j.chiabu.2020.104740 | A spatiotemporal analysis of the impact of COVID-19 on child abuse and neglect in the city of Los Angeles, California | Barboza, Gia E. | 2021 | Child Abuse & Neglect | Quantitative, observational, longitudinal spatiotemporal analysis | A longitudinal spatiotemporal analysis covering a total of 362 days, split into two 181-day periods: a baseline period (July 24, 2019, to January 20, 2020) and a pandemic period (January 21, 2020, to July 19, 2020) | Negative binomial regression, Spatiotemporal pattern mining (Emerging Hot Spot Analysis using the Getis-Ord Gi* statistic), and the Mann-Kendall trend test | 1,012 census tracts in the city of Los Angeles; 1,209 incident points were aggregated into the analysis. The individual incident count was 661 pre-COVID and 614 during the COVID-19 period. | Community-level (census tracts) and reported crime incidents | Secondary administrative data obtained from the City of Los Angeles Open Data portal, specifically crimes reported to the Los Angeles Police Department | Los Angeles, California | NR | NR | NR | NR | Pre-COVID: Male 49.8%, Female 50.2%. COVID-19 period: Male 50.7%, Female 49.3% | Pre-COVID: 65.1% Hispanic, 23.0% Black, 7.6% White, 1.2% Asian, 3.2% Other. COVID-19 period: 59.9% Hispanic, 26.6% Black, 7.7% White, 1.1% Asian, 4.6% Other | Neighborhood-level variables included: severe rental burden (Mean = 0.421), poverty (Mean = 0.278), chronic absenteeism (Mean = 0.142), labor force participation (Mean = 0.620), and the California Strong Start Index (Mean = 7.940). | No | NA | NA | NA | Yes | Geographic locations of crimes reported to the LAPD, filtered for codes indicative of physical child abuse (California Penal Code 273d) and neglect (California Penal Code 270) | Physical child abuse (simple and aggravated assault) and neglect | Severe rental burden, defined as households spending more than 50% of their monthly income on rent and utilities | American Community Survey (ACS) 5-year estimates | Neighborhood-level percentage of severely rent-burdened households within each census tract | While there was a statistically significant 8% decline in overall reports of child abuse and neglect (CAN) across the city during the pandemic, severe rental burden was significantly associated with increasing trends of CAN in specific "hot spots." Areas with high levels of structural risk, including housing insecurity, co-occurred with emerging and intensifying hot spots of CAN during the COVID-19 period. | For every one-unit increase in the percentage of severely burdened households, the incident rate of CAN increased by 83.1% (β=1.29, exp(β)=3.66, p<0.001). Severe rental burden had the largest predicted effect on CAN counts in intensifying hot spots compared to other neighborhood patterns. Areas identified as intensifying hot spots had an incident rate 2.80 times higher than consecutive cold spots (p<0.001) | None. | No | NA | study used police incident data rather than child protective services data; it only analyzed physical abuse and neglect, omitting sexual or emotional abuse; the results are highly dependent on the chosen spatial and temporal scales; analysis did not fully account for seasonal effects in the spatiotemporal models | |
56 | https://doi.org/10.1177/10775595231218177 | Affordable Housing and Neighborhood Child Maltreatment Reports | Littleton, Tenesha | 2024 | Child Maltreatment | Quantitative, cross-sectional ecological study | A cross-sectional ecological design using 2,341 census tracts as the primary unit of analysis to explore the relationships between subsidized housing distribution, neighborhood housing costs, and child maltreatment reports | Bayesian conditionally autoregressive (CAR) spatial models and bivariate analysis using Kendall’s tau-b | 2,341 census tracts | Neighborhood-level data (census tracts) from a large metropolitan county | Secondary administrative data obtained from the California Child Welfare Indicators Project, the U.S. Department of Housing and Urban Development, and the U.S. Census Bureau | Los Angeles, California | NR | NR | nearly 50% Latinx, 9% African American/Black, 15.6% Asian, and approximately 70% White | NR | NR | NR | Median household income was $76,367, with 14% of residents living in poverty | No | NA | NA | NA | Yes | Administrative data on all allegations of suspected child maltreatment (screened-in, substantiated, or unsubstantiated) from the California Child Welfare Indicators Project | General child maltreatment (abuse and neglect) | Rent burden (average percentage of income paid in rent), and the supply of Housing Choice Voucher (HCV) and Low-Income Housing Tax Credit (LIHTC) units | U.S. Census Bureau’s American Community Survey (ACS) 5-year estimates (2011–2015), HUD datasets for HCV, and California Tax Credit Allocation Committee (CTCAC) records for LIHTC | Neighborhood-level rates per 100 low-income renter households or average percentage of household income spent on rent | Neighborhoods with higher numbers of subsidized rental units (HCV and LIHTC) experienced higher numbers of child maltreatment reports. Conversely, neighborhoods where residents paid a greater share of their income in rent (typically higher-opportunity areas) had fewer reports. | Neighborhoods in which households paid a greater share of their income in rent were significantly associated with fewer reports of child maltreatment (Relative Risk [RR] = 0.9968, 95% CI [0.9942, 0.9995]) . A higher rate of Housing Choice Voucher (HCV) properties per 100 low-income renters was significantly associated with a higher risk of child maltreatment reports (RR = 1.0020, 95% CI [1.0005, 1.0034]). A higher rate of Low-Income Housing Tax Credit (LIHTC) rental properties was significantly associated with a higher risk of child maltreatment reports (RR = 1.0012, 95% CI [1.0004, 1.0018]). In bivariate analyses, child maltreatment reports were positively correlated with neighborhood housing cost (τ b =.296,p<.01), the supply of LIHTC properties (τ b =.215,p<.01), and the supply of HCV properties (τ b =.325,p<.01) | Residential instability (percentage of individuals who moved in the last 12 months) was not significantly associated with child maltreatment reports in the multivariate model (RR = 1.0002, 95% CI [0.9973, 1.0032]) | No | NA | cross-sectional, so unable to make causal conclusions; ecological unit of analysis masks individual or family-level differences; potential for selection bias in neighborhoods; data (from 2015) may not reflect recent shifts in housing programs | |
57 | https://doi.org/10.1177/10775595221125917 | Are Home Evictions Associated with Child Welfare System Involvement? Empirical Evidence from National Eviction Records and Child Protective Services Data | Tang, Shichao | 2024 | Child Maltreatment | Quantitative, observational, longitudinal study using county-level data | A longitudinal analysis at the county-year level using administrative datasets covering 2000–2016 for foster care entries and 2010–2016 for child abuse and neglect reports | Fixed-effect linear regression models (controlling for county and year effects) | 2,759 unique counties were included in our county-level analysis using NCANDS data and 2,802 unique counties were included using AFCARS data | County-level population (ecological) based on national administrative records | Secondary administrative data obtained from the National Child Abuse and Neglect Data System (NCANDS), the Adoption and Foster Care Analysis and Reporting System (AFCARS), and the Princeton Eviction Lab | approximately 87–89% of all U.S. counties | NR | NR | NR | NR | NR | NR | poverty rate: Mean ≈ 12%; rent burden: Mean ≈ 27–28%; and unemployment rate: Mean ≈ 6.6–7.3% | No | NA | NA | NA | Yes | Administrative records of screened-in CAN reports (NCANDS) and foster care entries (AFCARS) | General child maltreatment (abuse and neglect) | Home eviction rate (judgments) and home eviction filing rate | National county-level records from the Princeton Eviction Lab | Number of evictions or filings per 100 renter-occupied homes within a county | County-level increases in evictions are significantly associated with higher rates of both screened-in child maltreatment reports and foster care entries. On average, one additional eviction per 100 renter-occupied homes in a county is associated with a 1.3% increase in CAN reports and a 1.6% increase in foster care entries. | General population: One additional eviction per 100 renter-occupied homes was associated with a 1.3% increase in screened-in CAN reports (95%CI[0.6%,2.0%]) and a 1.6% increase in foster care entries (95%CI[0.6%,2.5%]). Every one additional eviction filing per 100 renter-occupied homes was associated with a 0.6% increase in foster care entries (95%CI[0.1%,1.1%]); By Race (CAN Reports): Every one additional eviction per 100 renter-occupied homes was associated with a 1.2% increase in reports for White children (95%CI[0.01%,2.4%]) and a 1.8% increase for Black children (95%CI[0.5%,3.2%]); By Race (Foster Care): Eviction rates were associated with an 8.1% increase in foster care entries for Hispanic children (95%CI[4.9%,11.3%]) and a 5.4% increase for Black children (95%CI[2.5%,8.3%]); | Eviction filings had no statistically significant relationship with screened-in CAN reports for the general population. There was no significant association between eviction or filing rates and CAN reports for Hispanic children. The relationship between eviction rates and foster care entries was not statistically significant for White children | No | NA | cross-sectional, so cannot draw causal conclusions; cannot disentangle whether evictions cause actual maltreatment or simply increase family visibility to reporters; data does not capture "informal" evictions occurring outside of the court system | |
58 | https://doi.org/10.1007/s10461-019-02753-9 | Young Transgender Women of Color: Homelessness, Poverty, Childhood Sexual Abuse and Implications for HIV Care | Eastwood, Elizabeth A. | 2021 | AIDS and Behavior | quantitative, cross-sectional study | cross-sectional analysis using baseline survey data from a multi-site demonstration project (Special Project of National Significance) | t-tests, Chi-square, odds ratios, multivariable logistic regression | 102 transgender women | HIV-positive young transgender women of color | recruited as they entered nine urban project sites or through community outreach | New York City, Chicago, San Francisco Bay area, and Los Angeles | Mean: 22.30 years (SD = 1.62); Range: 18-24 years | 100% transgender women | 58.8% Non-Hispanic Black, 28.4% Hispanic/Latina, and 12.7% Other/multiracial | NA | NA | NA | 92.9% lived below the 2013 Federal Poverty Level | No | NA | NA | NA | Yes | six items about unwanted touching, kissing, genital contacts, and sexual acts before the age of 18 | Childhood sexual abuse | Recent homelessness (past 6 months) and housing transience (two or more moves in the last 6 months) | Self-reported binary questions (yes/no) about homelessness and the number of housing moves in the last 6 months | Having been homeless at any time in the past 6 months | In bivariate analyses, any childhood sexual abuse was strongly associated with recent homelessness (OR = 5.94). However, in the multivariable logistic regression model, childhood sexual abuse was not a significant predictor of homelessness when other factors like food insecurity and viral suppression were included. | In bivariate analyses, any childhood sexual abuse was significantly associated with homelessness (OR = 5.94, 95% CI 2.01–17.49, p = .001). More specifically, Being touched in an intimate way (OR = 4.14, 95% CI 1.47–11.65, p = .005); Being forced/coerced to have genital sex (OR = 4.53, 95% CI 1.49–13.81, p = .006); Being forced/coerced to perform oral sex (OR = 3.44, 95% CI 1.24–9.58, p = .015); Being forced/coerced to kiss someone in a sexual way (OR = 3.25, 95% CI 1.12–9.43, p = .026) | In the multivariable regression model, childhood sexual abuse was not significantly associated with homelessness (Adjusted OR = 0.47, 95% CI 0.10–2.29, p = .353). In bivariate analysis, housing transience (2+ moves) was not significantly related to homelessness (OR = 2.57, 95% CI 0.69–9.53, p = .146). The specific childhood experience of someone rubbing their genitals against the participant was not significantly associated with homelessness (OR = 1.43, 95% CI 0.52–3.90, p = .486). Being forced/coerced to touch another person in an intimate way was not significant (OR = 1.79, 95% CI 0.67–4.82, p = .243) | No | NA | Cross-sectional data prevents causal investigation; susceptibility to social desirability bias or participants' incomplete understanding of clinical values; potential for under-enrolling the most disenfranchised individuals | When the researchers moved to the multivariable logistic regression (the final model), they did not enter every individual action separately. Instead, they used the single summary variable for "Childhood sexual abuse" to represent the collective history of those experiences. |
59 | https://doi.org/10.1177/1077801203253402 | Domestic Violence and Housing Problems: A Contextual Analysis of Women's Help-Seeking, Received Informal Support, and Formal System Response | Baker, Charlene K. | 2003 | Violence Against Women | quantitative, cross-sectional study | A cross-sectional analysis using a sample of 110 women recruited from three types of agencies (welfare, shelters, and criminal justice) to examine housing experiences after separating from an abusive partner | bivariate correlations, MANOVA, hierarchical multiple regression, and logistic regression | 110 women | Women who had experienced domestic violence and were separated from their partners for at least two weeks | Convenience and purposive sampling were used, including screening women in welfare waiting areas, calling women with criminal cases pending against partners, and approaching women residing in shelters | Atlanta, Georgia | Mean: 30.2 years (SD = 8.1); Range: 19-53 years | 100% female | 82% African American, 5% White, 5% Latina, 1% Asian American, 1% Native American, and 6% other | NA | NA | NA | 54% reported a monthly income of $1,000 or less; 66% were currently employed | Yes | Revised Conflict Tactics Scale (CTS2) | Participants must have experienced physical or sexual violence, including threats of harm, during the separation period | Psychological, physical (minor and severe), and sexual violence | No | NA | NA | Recent homelessness (e.g., staying with friends/family, shelters, motels, cars, or parks for at least 7 days) and housing problems (e.g., late rent, credit problems, threatened eviction, skipping meals to pay rent, or selling belongings to pay rent) | Housing Problems Index (HPI) | For homelessness, the criterion was being without a home for at least 7 days following separation | Higher violence severity was a predictor of hosing problems. | Greater violence severity was significantly associated with a higher number of housing problems (β=.48,p<.05) | Violence severity did not significantly predict homelessness (OR=0.97,p>.05). | No | NA | Cross-sectional data prevents causal investigation; potential sampling bias since many women could not be reached due to disconnected phones; tecruitment from shelters may have inflated the overall rate of homelessness in the sample; study also relied on a newly developed, non-standardized measurement index | |
60 | https://doi.org/10.1176/appi.ajp.20230649 | Pathways to Homelessness: Childhood Maltreatment and Psychiatric Symptoms Increase Risk of Homelessness | Widom, Cathy Spatz | 2024 | American Journal of Psychiatry | quantitative, longitudinal study | A prospective cohort design where children with documented histories of maltreatment (ages 0–11) and a matched control group were followed into middle adulthood. Homelessness was assessed at four waves (mean ages 29, 39, 41, and 47) | structural equation modeling, logistic regression, t-tests, and bootstrapping | 1,196 participants were interviewed at the first wave | A community sample consisting of a maltreated group and a demographically matched control group | Cases of abuse and neglect were obtained through official court records from 1967–1971 in a Midwestern metropolitan area; controls were matched based on hospital of birth and neighborhood schools | Metro area in the Midwest of the US | mean at the first interview: 29.2 years (range 18–40); mean at the final interview: 47.0 years | 48.7% female | 62.9% White (non-Hispanic), 34.9% Black, and 3.8% Hispanic | NA | NA | NA | 60.0% completed high school, 54.9% held unskilled or semiskilled jobs, and only 13.7% held professional or semiprofessional positions | No | NA | NA | NA | Yes | Documentation was obtained from official court records of substantiated cases of abuse or neglect | Physical abuse (e.g., bruises, fractures), sexual abuse (e.g., fondling, rape), and neglect (extreme failure to provide food, clothing, shelter, or medical attention) | Ever homeless (no regular place to live for at least a month) and past-year homelessness | Self-report binary questions during interviews | Having no regular place to live for at least a month | Childhood maltreatment significantly predicted homelessness across all time points from young to middle adulthood. Psychiatric symptoms (specifically depression, PTSD, and antisocial personality disorder) were identified as significant pathways through which childhood maltreatment increases the risk of future homelessness | Individuals with documented childhood maltreatment were more likely to have ever been homeless by age 29 than the control group (25.6% vs. 12.3%, AOR = 2.54, 95% CI = 1.86–3.50, p < 0.001). Childhood maltreatment significantly predicted past-year homelessness at age 29 (AOR = 2.09, 95% CI = 1.31–3.43, p = 0.002). Childhood maltreatment predicted past-year homelessness at age 39 (AOR = 2.13, 95% CI = 1.12–4.28, p = 0.03), age 41 (AOR = 2.60, 95% CI = 1.26–5.91), and age 47 (AOR = 2.90, 95% CI = 1.26–7.51), even after controlling for earlier homelessness. Significant indirect paths from childhood maltreatment to future homelessness were found through symptoms of depression (β = 0.01, p = 0.04), PTSD (β = 0.02, p = 0.02), and antisocial personality disorder (β = 0.03, p = 0.01). | In the mediation models, alcohol use symptoms (β = 0.01, p = 0.09) and drug use symptoms (β = 0.01, p = 0.08) were not significant indirect pathways between childhood maltreatment and future homelessness. | No | NA | Findings may not be generalizable to unreported cases or cases from higher socioeconomic status families; study used a specific cohort from the Midwest growing up in the 1960s/70s, which may not reflect current service landscapes; psychotic disorders were not assessed; homelessness measures relied on self-reports, which can be subjective; low fit indices in SEM models suggest other unmeasured factors contribute to homelessness | |
61 | https://doi.org/10.1007/s10995-018-02698-w | Homelessness in Childhood and Adverse Childhood Experiences (ACEs) | Radcliff, Elizabeth | 2019 | Maternal and Child Health Journal | quantitative, cross-sectional, descriptive study | a cross-sectional analysis using 2016 data from the South Carolina Behavioral Risk Factor Surveillance System (BRFSS) survey | Chi-square tests and applied survey sampling weight | 7,490 respondents | community sample of non-institutionalized adults | random digit dialing of landlines and cell phones | South Carolina, US | Age was categorized in decades: 18–29, 30–39, 40–49, 50–59, 60–69, and 70–79. 53.5% of the weighted sample was under the age of 40 | 52.1% female and 47.9% male | 68.6% non-Hispanic White, 24.2% non-Hispanic Black, 4.3% Hispanic, and 2.9% other race/ethnicity | NA | NA | NA | 41.9% had a high school diploma/GED or less; 48.2% had an annual income of less than $50,000 | No (only witnessing IPV between caregivers) | NA | NA | NA | Yes | 11-question ACE module adapted from the CDC-Kaiser ACE study | Physical abuse, emotional abuse, and sexual abuse | Childhood homelessness, defined as a family being unable to afford a place to live | A single self-report question: "How often were you homeless when you were growing up?" | Participants who responded "rarely, sometimes, often, or very often" were counted as having experienced childhood homelessness | dults who reported homelessness in childhood experienced significantly greater exposure to higher numbers and types of ACEs compared to those who were never homeless | Those with childhood homelessness reported significantly higher rates of physical abuse (43.9% vs. 12.2%, p<.0001), emotional abuse (66.4% vs. 30.2%, p<.0001), and forced sex (22.4% vs. 4.5%, p<.0001). | None. | No | NA | study relied on retrospective self-reports, which may lead to under-reporting; cross-sectional design prevents determining causality or the timing of events (e.g., whether abuse preceded homelessness); definition of "homeless" in the survey question may have been interpreted by some as general housing insecurity; the survey excluded institutionalized adults, such as those who are incarcerated | Adults who were homeless in childhood were significantly more likely to have witnessed household violence (IPV) than those who were not homeless (58.3% vs. 17.3%, p<.0001) |
62 | https://doi.org/10.1037/a0017212 | Parent-adolescent violence and later behavioral health problems among homeless and housed youth | Haber, Mason G. | 2009 | American Journal of Orthopsychiatry | Quantitative, prospective longitudinal study | prospective cohort design with data collection at baseline, 1.5-year follow-up, and 4.5-year follow-up | Exploratory Factor Analysis, hierarchical multiple regression, and t-tests | 375 youth (234 homeless, 141 housed) | Adolescents aged 13–17 from a Midwestern metropolitan area; the homeless sample was recruited from youth shelters | Homeless youth were recruited from all existing shelters in a five-county metropolitan area. Housed youth were matched to the homeless sample and recruited via acquaintance nominations or directly from the same neighborhoods | Metro area in the Midwest of the US | NA | NA | NA | Baseline age range: 13 - 17 years | 245 females and 130 males | 191 non-Latino White and 184 Black adolescents | NA | No | NA | NA | NA | Yes | Revised Conflict Tactics Scale (CTS2) | Physical abuse (mild and severe), physical injury, and psychological aggression (verbal and nonverbal) | Homelessness (shelter residence) | Residence in a youth shelterelter | Residence in a youth shelter | Homeless youth reported significantly higher rates of all types of parent-adolescent violence at baseline compared to matched housed youth. While violence predicted later behavioral health problems, the Risk Amplification Model was not supported, since homelessness did not strengthen the link between violence and poor outcomes compared to the housed group. | At baseline, homeless youth reported significantly more parent violence (t=−6.17,p<.001) and psychological violence (t=−2.34,p<.05) than housed youth. At 1.5-year follow-up, parent violence significantly predicted mental health symptoms (GSI) but only for the housed group (β=−.374,p<.05). | The Risk Amplification Model was not supported; homelessness did not strengthen the link between family violence and outcomes as hypothesized. | No | NA | Reliance on single-informant self-reports from adolescents; substantial attrition occurred at the 1.5-year (n=240) and 4.5-year (n=294) follow-ups; the sample's predominance of first-time shelter residents may not generalize to youth with more chronic street histories | |
63 | Foster care children and family homelessness | Zlotnick, Cheryl | 1998 | American Journal of Public Health | Quantitative; cross-sectional | cross-sectional analysis using data from an ongoing longitudinal study | chi-square tests, logistic regression | 195 children | Child welfare-involved (newly placed foster children) | Children were randomly selected from a northern California county social service agency list of newly placed foster children between 1993 and 1996 | northern California county | NA | NA | NA | All children were under 4 years of age. The sample was split between those <1 year (1.5%), 1–1.9 years (37.4%), 2–2.9 years (30.3%), and 3–3.9 years (30.8%). | 52.8% male and 47.2% female | 66.2% Black, 16.4% White, 5.1% Latino, and 12.3% Other | 35.7% of children in the county lived in families living on 185% or less of the federal poverty level | No | NA | NA | NA | Yes | Abstracts of county social services records and Child Protective Services intake forms | Neglect (including parental substance abuse), abandonment, sibling abuse, physical abuse, and sexual abuse | Homelessness or "no stable abode" | Review of CPS intake forms or court reports | Parents were included if the CPS intake form or court report described them as "homeless" or as having "no stable abode" | Almost half of the sample (48.7%) had birth parents who had experienced homelessness. Neglect, often secondary to substance abuse, was the most common reason for placement (98.5% of the total sample). | A significantly higher proportion of children whose birth parents had no history of homelessness were in foster care because of sexual abuse (8.0% vs. 0% for those with a history of homelessness, P<.001). Foster children whose birth parents had been homeless were significantly more likely to have siblings who also had been placed in foster care compared to those whose parents had no history of homelessness (52.6% vs. 36.0%, P<.05). In a logistic regression model, the history of a foster child’s sibling also being placed in foster care significantly predicted parental homelessness (OR = 1.99; 95% CI = 1.09, 3.64; P=.025). | There were no statistically significant differences between children of homeless vs. non-homeless parents regarding placement for neglect (99.0% vs. 98.0%), abandonment (52.6% vs. 43.0%), parental incarceration (10.5% vs. 13.0%), or physical abuse (4.2% vs. 3.0%). | No | NA | sample was limited to children under 4 from a single county system, which may not be generalizable; the determination of homelessness relied solely on court or CPS documentation, which may have underestimated the true prevalence of parental homelessness if episodes were not recorded | ||
64 | Relative contributions of parent substance use and childhood maltreatment to chronic homelessness, depression, and substance abuse problems among homeless women: mediating roles of self-esteem and abuse in adulthood | Stein, Judith A. | 2002 | Child Abuse & Neglect | Quantitative; cross-sectional; latent variable methodology | Cross-sectional analysis of baseline data from a larger longitudinal health promotion study | Structural Equation Modeling, Confirmatory Factor Analysis, and path analysis | 581 women | Homeless women residing in shelters or sober living centers | Recruited through directors of 35 homeless shelters and sober living recovery programs | Los Angeles, California | Mean: 33.5 years; Range: 6 to 50 years | 100% female | 54% African-American, 23% Latina, 22% White | NA | NA | NA | Participants were currently homeless; mean years of education = 11.1 years (median = 12 years) | Yes | single self-report item asking whether abuse occurred immediately before the participant’s latest homelessness episode | Self-report | Physical IPV | Yes | multiple-indicator latent variable indicated by self-reported sexual, physical, and verbal abuse | Sexual abuse (molested, raped), physical abuse (kicked, slapped, or hit with an object), and verbal/emotional abuse (ignored, yelled at, put down) | Chronic homelessness | latent variable indicated by (1) the self-reported number of times ever homeless and (2) total time spent homeless in months and years | Living in a shelter or sober living recovery program for at least one week and uncertain of residence in the next 60 days | Childhood abuse directly predicted later chronic homelessness and recent physical abuse. Recent physical abuse also directly predicted chronic homelessness. Childhood abuse had significant indirect effects on chronic homelessness mediated through later adult physical abuse. | Childhood abuse significantly directly predicted chronic homelessness (path coefficient = .25, p<.001). Recent physical abuse (IPV) significantly predicted chronic homelessness (path coefficient = .25, p<.001). Childhood abuse had a significant indirect effect on chronic homelessness that was mediated through recent physical abuse (p<.001). In bivariate correlations, childhood abuse was significantly correlated with chronic homelessness (r=.36,p≤.001. Bivariate correlations also showed that recent physical abuse was significantly correlated with chronic homelessness (r=.36,p≤.001) | Self-esteem was not a significant mediator between childhood abuse and chronic homelessness because the bivariate correlation between self-esteem and chronic homelessness was small and not statistically significant (r=−.06). While parent substance abuse was significantly correlated with chronic homelessness in the bivariate analysis (r=.22,p≤.01), it was not a significant predictor of chronic homelessness in the final path model. | No | NA | sample was primarily African-American and Latina women in Los Angeles, which may limit generalizability; the study relied on retrospective self-reports, which may be subject to recall bias or repression of traumatic events; the childhood measurement items were noted as "simplistic"; the cross-sectional design prevents confirmation of causal directionality | ||
65 | https://doi.org/10.1002/ajcp.12203 | Child Maltreatment, Problem Behaviors, and Neighborhood Attainment | Chauhan, Preeti | 2017 | American Journal of Community Psychology | Quantitative; Prospective cohort design | Prospective cohort design using three time points: Initial case identification (1967–1971), young adulthood assessment of problem behaviors at mean age 29 (1989–1995), and middle adulthood assessment of neighborhood characteristics at mean age 40 (2000–2002) | chi-square, t-tests, structural equation modeling, full information maximum likelihood estimates | Initial sample N=1,575 (n=908 maltreated; n=667 matched controls). Multivariate models utilized n=1,124 participants from the first interview | Child welfare-involved (children with documented histories of abuse and neglect) and a matched control group | The maltreated group consisted of substantiated cases of childhood abuse or neglect processed in a Midwestern metropolitan area's courts between 1967 and 1971. The control group was matched based on age, sex, race/ethnicity, and approximate family social class during the same period | Metro area in the Midwest of the US | Mean for the first interview: 29.23 (SD = 3.78) (problem behaviors); mean for the second interview 39.52 (SD = 3.53) (neighborhood outcomes) | 51% female at the first interview | 61% non-Hispanic White, 33% non-Hispanic Black, 4% Hispanic, and 2% other. Multivariate models focused on White (n=735) and Black (n=389) participants only | NA | NA | NA | Predominantly from the lower end of the socioeconomic spectrum due to court involvement | No | NA | NA | NA | Yes | Review of official court records during years 1967 to 1971 | Physical abuse, sexual abuse, and neglect | Homelessness | Two self-report items asking if participants had (a) been homeless in the past 12 months and (b) ever been homeless | Self-report of homelessness | Childhood maltreatment significantly predicted a higher risk of homelessness in young adulthood compared to matched controls. Homelessness (as part of a higher-order "problem behaviors" factor) partially mediated the relationship between childhood maltreatment and residing in less desirable neighborhoods (high poverty, high disorder, low social cohesion) in middle adulthood. | In bivariate analyses, maltreated children were at significantly greater risk for homelessness compared to matched controls (P<.05). In the path analysis for the overall sample, the direct relationship between child maltreatment and the higher-order problem behavior factor (which included homelessness) was statistically significant. Indirect pathways showed that childhood maltreatment significantly predicted residence in neighborhoods characterized by physical disorder, low social cohesion, and economic disadvantage via problem behaviors, including homelessness. For Black children, there were significant indirect paths from childhood maltreatment through problem behaviors (including homelessness) to residence in neighborhoods with greater physical disorder and less social cohesion. For White children, the indirect path from maltreatment through problem behaviors (including homelessness) was significant for physical disorder, social cohesion, and economic disadvantage. | In bivariate analysis for the White sample specifically, maltreated White children did not differ significantly from White controls in the extent to which they were at increased risk for homelessness in the past year. For Black children, the indirect path from maltreatment through problem behaviors (including homelessness) was not significant for neighborhood economic disadvantage or social control. The indirect path to neighborhood social control was not significant for White children. | No | NA | Maltreatment cases were limited to those processed through courts, which may represent more serious cases and may not generalize to unreported abuse; sample primarily reflects the lower end of the socioeconomic spectrum; findings are based on a specific Midwestern metropolitan area and may not generalize to other geographic or cultural contexts | |
66 | https://doi.org/10.1016/j.jad.2025.03.063 | Sex differences in the relationship between childhood sexual abuse and adult homelessness among underserved bipolar outpatients | Mulligan, Daniel J. | 2025 | Journal of Affective Disorders | Quantitative; Cross-sectional; Observational | Cross-sectional design utilizing self-report data collected during a single clinic visit | Moderated logistic regression, binary logistic regression, and bivariate correlations | 201 outpatients | Underserved bipolar outpatients | Participants were recruited from an outpatient behavioral health clinic affiliated with a large urban public hospital that primarily serves under-resourced populations | Metro area in the Southwestern US | Mean age: 39.4 years (SD = 11.3); Range: 19–71 years | 69.2% female and 30.8% male | 74.6% Black/African American, 18.4% non-Hispanic White, 2.0% Hispanic/Latino, and 5.0% Other or Multiracial | NA | NA | NA | 35.3% had a high school degree or GED as their highest education; 76.6% were unemployed; 76.1% had a monthly income below $1000 | No | NA | NA | NA | Yes | Childhood Sexual Abuse (CSA) scale of the Childhood Trauma Questionnaire—Short Form (CTQ-SF) | Childhood sexual abuse | Adult homelessness | Self-report items where participants indicated if they considered themselves homeless and described where they slept at night | Self-report of homelessness | Childhood sexual abuse was significantly and positively associated with adult homelessness for females, but this relationship was not significant for males | Sex moderated the relationship between childhood sexual abuse (CSA) and adult homelessness (Wald χ2(1)=6.53,P<.05 in Model 1; P<.01 in Model 2). For females, childhood sexual abuse severity was a significant positive predictor of homelessness (OR = 1.09, 95% CI: [1.03, 1.16], P<.01 in Model 1; OR = 1.08, 95% CI: [1.02, 1.15], P<.01 in Model 2). Females with "severe to extreme" CSA were 2.2 times more likely to be homeless compared to those with "none/minimal" CSA. | For males, the relationship between childhood sexual abuse and adult homelessness was not statistically significant (b=−0.05,P=.29, OR = 0.95, 95% CI: [0.88, 1.04]). The bivariate correlation between CSA and housing status for the total sample was not statistically significant (r=0.10,P>.05). | No | NA | Non-representative sample from a single hospital system, so the internal and external validity were limited; self-report data, which is subject to social desirability and recall biases; cross-sectional design prevents any causal interpretations of the relationship between childhood sexual abuse and housing outcomes | |
67 | https://doi.org/10.1097/NMD.0000000000001730 | Correlates of Past Childhood Homelessness in a Nationally Representative Sample of US Adults | Bommersbach, Tanner J. | 2024 | The Journal of Nervous and Mental Disease | Quantitative; Retrospective cross-sectional survey | Retrospective analysis using data from the National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III), comparing a three-level hierarchy of homelessness: never homeless, homeless only as children (before age 15), and homeless as both children and adults | chi-square and Student t tests), weighted multivariable-adjusted logistic regression, and Risk Ratios | 36,154 respondents | Nationally representative sample of noninstitutionalized US adults | Multistage probability sampling | United States (National) | Mean ages: Never homeless (46.6), Child homelessness only (41.3), Both child and adult homelessness (43.5), Adult homelessness only (44.6) | NA | NA | NA | Adulthood income <$20,000 (Child only: 26.0%; Both: 48.7%); Adulthood college completion (Child only: 7.7%; Both: 4.8%) | Yes | One self-report item within a 7-item violence measure: "ever getting into a fight that came to swapping blows with someone like a husband, wife, boyfriend, or girlfriend" | Self-reported occurrence since age 15 | Physical IPV | Yes | 29-item questionnaire (AUDADIS-5) adapted from the Adverse Childhood Experiences (ACE) study | Childhood neglect (9 items) and childhood sexual abuse (4 items) | Childhood homelessness (before age 15) and adult homelessness (after age 15) | Self-report of having "no regular place to live—like living on the street or in a car" lasting at least 1 month | Episode lasting at least 1 month before/after age 15 | Childhood homelessness is robustly associated with severe childhood adversities, such as neglect and sexual abuse. Individuals with childhood homelessness were nearly 50 times more likely to experience adult homelessness. | Individuals with childhood homelessness were 46.9 times more likely to also experience adult homelessness (95% CI 32.3–68.1). In multivariate analysis, childhood neglect independently predicted childhood homelessness (OR = 1.08, 95% CI: 1.04, 1.12, p<.01). In bivariate analyses, childhood neglect (Cohen d = 1.78) and sexual abuse (Cohen d = 1.21) were significantly higher in individuals who were homeless as children compared to those never homeless (p<.001). | None. | No. | NA | retrospective design introduces potential recall bias; definition of homelessness focused on "unsheltered" status, which may underestimate prevalence by missing sheltered or "doubled up" populations; the use of age 15 as a cutoff may misclassify some childhood experiences as adult ones | |||
68 | https://doi.org/10.1177/0886260510365868 | Childhood victimization and crime victimization | McIntyre, Jared Kean | 2011 | Journal of Interpersonal Violence | quantitative, observational, and prospective cohort design study | prospective cohort design where children with documented histories of abuse and neglect were matched with a control group. Data collection occurred in phases: archival record review (1967–1971), a second phase of interviews (1989–1995), and a third phase of interviews (2000–2002) in middle adulthood | Logistic and ordinary least square (OLS) regressions and cross-tabulations | 892 individuals (497 in the abused/neglected group and 395 in the matched control group) | individuals with court-substantiated cases of child abuse and neglect and matched controls who were skewed toward the lower end of the socioeconomic spectrum | Maltreatment cases were identified through archival records of county juvenile and adult criminal courts from 1967–1971. Comparison group members were matched based on county birth records or school records from the same time period | Metro area in the Midwest of the US | Mean: 39.5 years (SD = 3.53); range: 30 -47 years | 51.2% female | 60.9% White, 32.5% Black, and 4.4% Hispanic | 11 years of age or younger at the time of the documented abuse or neglect incident | children matched by sex | children matched by race | primarily from lower socioeconomic backgrounds; 31.3% had less than a high school education | No | NA | NA | NA | Yes | official archival records of substantiated court cases | Physical abuse, sexual abuse, and neglect | Homelessness | Self-report items from the Diagnostic Interview Schedule (DIS-III-R) antisocial personality disorder module | Participants reported a "period of homelessness in the past 12 months" or "a period when you had no regular place to live, for at least a month or so" | Individuals with a history of childhood abuse and/or neglect were significantly more likely to experience homelessness compared to the control group. Furthermore, homelessness was found to be a partial mediator between childhood maltreatment and subsequent physical crime victimization for women and for the Black and Hispanic subgroups. | Childhood abuse and neglect was a significant predictor of homelessness (OR = 2.60, 95% CI = 1.81–3.72, p < .001). In the female subgroup, homelessness was identified as a partial mediator for physical crime victimization. When homelessness was included in the model, the odds ratio for childhood maltreatment remained significant but was attenuated (OR = 3.38, 95% CI = 2.09–5.44, p < .001). For participants in the Black and Hispanic subgroups, homelessness was found to mediate the relationship between childhood maltreatment and physical crime victimization. When homelessness was controlled for, the relationship between childhood maltreatment and physical crime victimization was attenuated to non-significance (OR = 1.58, 95% CI = 0.89–2.81, p > .05). | Homelessness was not a mediator for White participants because it was not a significant predictor of physical crime victimization for them (OR = 1.17, 95% CI = 0.96–1.42, p > .05). | No | NA | study relied on official court records which represent only the most extreme cases; the sample was skewed toward lower socioeconomic status, limiting generalizability to middle or upper-class families; the extent of unreported abuse in the control group was unknown; the findings are specific to a cohort from the Midwest in the late 1960s/early 1970s | |
69 | https://doi.org/10.1111/j.1525-1446.2004.21103.x | Factors influencing homelessness in women | Anderson, Debra Gay | 2004 | Public Health Nursing | quantitative, descriptive correlational design study | retrospective self-report design where participants were interviewed and completed standardized questionnaires regarding their childhood experiences (family of origin) and current social support networks | ANCOVA, Pearson's product-moment correlation, chi-square tests, and Fisher's least significant difference (LSD) | 255 women | Mixed community and shelter-involved sample, categorized into three groups: homeless/abused (n=94), never-homeless/abused (n=88), and never-homeless/not-abused (n=73) | shelter settings, group informational sessions, fliers at community colleges, housing complexes, public clinics, and newspaper advertisements | metro area in the Greater Portland Area, US | Mean: 37.1 years (SD = 11.1) | 100% female | 60.3% White, 27.0% African-American, and 11.3% other | NA | NA | NA | high proportion of low-income or marginalized individuals | yes | Interpersonal Relationship Inventory (IPRI) | self-report of domestc violence | General conflict and abuse in current relationships | Yes | Self-report of "childhood trauma" | Physical abuse, sexual abuse, and emotional abuse | Homelessness, including living on the streets, being in a shelter or SRO, or being "doubled up" with others) | Self-report based on inclusion criteria | Participants had to have experienced at least one episode of homelessness lasting at least 1 month but no longer than 1 year | The findings suggest that while abuse alone is not a direct predictor of homelessness, the resulting inability to form and maintain healthy relationships (internal working models) is highly predictive. | The homeless group had significantly more foster family experiences compared to the never-homeless/not-abused group (p < 0.004). | There were no significant differences between the homeless group and the never-homeless/abused group on measures of childhood social support, reciprocity, or conflict | No | NA | NR | |
70 | https://doi.org/10.1016/j.jad.2020.06.039 | The role of psychiatric symptoms and environmental vulnerability factors in explaining the relationship between child maltreatment and suicidality: A prospective investigation | Widom, Cathy Spatz | 2020 | Journal of Affective Disorders | quantitative, observational, and prospective cohort design study | prospective cohort design initiated in 1986, following a group of children with court-documented maltreatment and a matched control group into middle adulthood | Structural equation models, t-tests, ANOVAs, and chi-square tests | 890 individuals (495 in the maltreated group and 395 in the matched control group) | Individuals with court-substantiated cases of child maltreatment (ages 0–11) and matched controls from similar socio-economic backgrounds | Maltreatment cases were identified through county juvenile and adult criminal court records from 1967–1971; controls were matched via hospital birth records or elementary school records | Metro area in the Midwest of the US | At the time of the homelessness assessment (Interview 1), the mean was 29.0 years. At the time of the suicidality assessment (Interview 2), the mean was 39.0 years (SD = 3.49), and range 30 - 47 years. | 48.9% male and 51.1% female | 60.8% White, non-Hispanic, and 32.6% Black, non-Hispanic and other races | 11 years of age or younger (range 0–11) at the time of the documented maltreatment | matched by sex | matched by race | predominantly from the lower end of the socioeconomic spectrum | No | NA | NA | NA | Yes | Documented archival court records of substantiated abuse and neglect | Physical abuse, sexual abuse, and neglect | Homelessness | Self-report asking participants if they had ever been homeless for "at least a month or so" | self-reported period of homelessness lasting at least one month | Childhood maltreatment was significantly associated with a higher likelihood of homelessness in young adulthood. In turn, homelessness was found to be a significant mediator in the relationship between childhood maltreatment and adult suicide attempts. While many individuals reported suicidal thoughts, those who experienced both child maltreatment and homelessness were significantly more likely to actually attempt suicide. | Individuals with documented histories of childhood maltreatment were significantly more likely to report having experienced homelessness (26.7%) compared to the matched control group (12.7%, p < .001). There was a significant indirect path from childhood maltreatment to adult suicide attempts through the mediator of homelessness (indirect effect = 0.10, p = 0.003). In the full structural equation model, homelessness remained a significant mediator (indirect effect = 0.10, p = 0.002), explaining approximately 18.87% of the total effect of childhood maltreatment on suicide attempts. Individuals who had made a suicide attempt reported a significantly higher rate of homelessness (38.8%) compared to those who had suicidal thoughts only (22.6%) and those with no thoughts or attempts (14.9%, p < .001) | non-significant trend for homelessness to predict suicidal thoughts as an outcome of childhood maltreatment (indirect effect = 0.03, p = 0.07) | No | NA | findings based on court-substantiated cases which may not generalize to middle-class or non-reported cases; the Midwestern cohort from the late 1960s and 1970s may not perfectly reflect modern society; the study lacked information on emotional maltreatment and family history of suicide; the fit indices for the path models were not high, suggesting other mechanisms may be involved | |
71 | https://doi.org/10.1002/ajcp.12146 | Permanent Housing for Child Welfare-Involved Families: Impact on Child Maltreatment Overview | Fowler, Patrick J. | 2017 | American Journal of Community Psychology | Quantitative, longitudinal, and a Randomized Controlled Trial | A longitudinal RCT that followed families for 2.5 years. Families were randomly assigned to either the intervention (Family Unification Program plus Housing Advocacy Program) or the control group (Housing Advocacy Program alone). Data collection occurred at five time points: baseline, and 6, 12, 18, and 30 months following referral | Multilevel latent growth models, Negative binomial regressions, intent-to-treat analysis approach | 150 families containing 372 children | Homeless and child welfare-involved families. Specifically, these were intact families under investigation for child abuse and neglect whose inadequate housing was identified as a threat to child out-of-home placement | Child welfare caseworkers identified and referred eligible families to the program | metro Chicago, Illinois | mean: 32.0 years (SD = 8.5) in the intervention group and 31.2 years (SD = 7.6) in the control group | 92.0% female in the intervention group; 94.7% female in the control group | intervention group: 68.0% African American, 21.3% Latino, 9.3% White, and 1.3% other. control group: 65.3% African American, 20.0% Latino, 13.3% White, and 1.3% other | mean: 6.1-6.2 years (SD = 5.1); Range: 0 - 15 years | NR | NR | all families had annual gross incomes below 30% of the area median income | No | NA | NA | NA | Yes | Parent–Child Conflict Tactics Scales Second Edition (CTSPC-2) | Psychological aggression, physical assault (minor acts), and neglect | Homelessness and inadequate housing | Professional identification by child welfare caseworkers and caregiver self-report of "living on the streets or in shelters" | Families whose inadequate housing "threatened child out-of-home placement" and who met federal income and criminal history eligibility for housing subsidies | The study found marginal treatment differences in the change of caregiver-reported child maltreatment (specifically minor physical assault and neglect) over 2.5 years between those who received permanent housing and those who received advocacy alone. Psychological aggression remained high across both groups regardless of housing status. Additionally, access to permanent housing did not yield greater benefits for families identified as "higher risk" (e.g., those with substance abuse or depression) | For the treatment group, families that were literally homeless (living on streets/in shelters) at baseline reported lower initial psychological aggression compared to those who were inadequately housed but not literally homeless (b=−0.66, p<.05. For the treatment group, families that were literally homeless at baseline exhibited a significantly steeper increase in the frequency of physical assault over the 30-month follow-up period (b=0.29, p<.05). Families referred for the permanent housing intervention (FUP+HAP) showed significant overall declines in minor physical assault (b=−0.36, p<.01) and neglect (b=−0.36, p<.05) over time. | There were no statistically significant differences between the intervention and control groups regarding the rate of change for psychological aggression (p=.38), physical assault (p=.07), or neglect (p=.07), though assault and neglect "approached significance." Homelessness at baseline was not a significant predictor of initial levels of neglect (b=−1.40, p>.05) or physical assault (b=−0.43, p>.05) for the treatment group. In the control group (HAP only), homelessness at baseline did not significantly predict initial levels or changes in any of the three maltreatment domains (p>.05). | Yes | Family Unification Program | Potential underreporting of maltreatment due to social desirability or fear of child welfare involvement; a small sample size that was only powered to detect moderate program effects; a focus only on primary caregivers; lack of generalizability beyond a single city | |
72 | https://doi.org/10.1016/j.chiabu.2018.01.029 | Beyond residential mobility: A broader conceptualization of instability and its impact on victimization risk among children | Merrick, Melissa T. | 2018 | Child Abuse & Neglect | Quantitative and cross-sectional | pooled cross-sectional data from three waves (2008, 2011, and 2014) of the National Surveys of Children’s Exposure to Violence (NatSCEV) | Logistic regression models, t-tests, ANOVA, and Scheffe post-hoc tests, Likelihood-ratio tests | 12,935 cases | nationally representative sample of youth | a mix of random digit dialing (RDD) and targeted oversampling of households with children, cell phone-only households, and underrepresented racial groups | United States (national) | NR | NR | NR | Mean: 8.6 years; range: one month - 17 years | 51.2% male and 48.8% female | 56.8% White, non-Hispanic; 14.5% Black, non-Hispanic; 20.6% Hispanic (any race); and 7.8% Other, non-Hispanic | average SES value was standardized at -0.07 (range -2.5 to 1.8), based on a composite of parent's education and household income | No (only witnessing IPV between caregivers) | NA | NA | NA | Yes | Juvenile Victimization Questionnaire (JVQ) | Physical abuse (hitting, beating, kicking), emotional abuse (calling names, mean things), and neglect (lack of food, medical care, or safe shelter) | Homelessness (living on the street or in a shelter) and excessive residential mobility (moving more than once every 3.33 years) | Self-report or caretaker interview regarding lifetime experiences (e.g.," How many times has your [child] moved households since he/she was born? (Response divided by child’s age is greater than 0.3, or more than once every 3.3 years and made into a dichotomous variable)) | Having lived on the street or in a shelter at any time in life; or moving households at a rate greater than 0.3 moves per year of life | Excessive residential mobility alone was a significant predictor of most victimizations, but almost all these associations lost significance once other factors like homelessness, child removal, and parental incarceration were included. The study found that accumulated instability (the LDF Index) is a much stronger predictor of victimization than moving alone. | The LDF Index (including homelessness) was a significant predictor for maltreatment (AOR = 1.54, 95% CI = 1.39–1.71, p < .001). Homelessness significantly overlapped with excessive residential mobility, with 64.7% of homeless children also meeting criteria for excessive moves. | When the LDF index was controlled for, excessive residential mobility was no longer a significant predictor of maltreatment (AOR = 1.11, 95% CI = 0.85–1.43, p > .05). | No | NA | analysis of lifetime factors may underestimate the impact of immediate risks from recent instability; unable to disentangle single events (like a natural disaster) that simultaneously trigger multiple destabilizing factors | |
73 | Adverse childhood experiences: are they risk factors for adult homelessness? | Herman, Daniel B. | 1997 | American Journal of Public Health | Quantitative, observational, and retrospective | A 1994 telephone follow-up study of a 1990 nationwide random-digit-dial survey. It compared a national probability sample of formerly homeless persons with a comparison group of never-homeless persons regarding their retrospective reports of childhood experience. | Univariate analyses and multivariate logistic regression | 487 (92 formerly homeless respondents and 395 never-homeless control subjects) | Community (national probability sample of US household members) | Respondents were recruited from a previous nationwide random-digit-dial telephone survey of 1507 adult residents. The follow-up targeted all 169 respondents who previously reported homelessness and a purposely selected sample of never-homeless controls | United States (National) | 18–24 (4.3%), 25–54 (67.6%), 55–64 (11.6%), and 65+ (16.5%) | 56.8% Female, 43.2% Male | 85.4% White, 8.5% African American, 3.3% Hispanic, and 2.8% Other | NA | NA | NA | NR | No | NA | NA | NA | Yes | Seven-item condensed version of the Childhood Experience of Care and Abuse Interview (CECA) was used to measure "lack of care"; Physical and sexual abuse were measured using single-item questions | Physical abuse (severely beaten and badly bruised/injured), sexual abuse (molested, abused, or raped), and lack of care (parental antipathy and indifference) | Adult lifetime literal homelessness | Self-report interview | having had to sleep overnight in a shelter, an abandoned building, a vehicle, or out in the open because the respondent did not have a place to live | Lack of parental care and physical abuse during childhood are powerful risk factors for adult homelessness. While sexual abuse showed a nonsignificant trend toward increasing risk, the combination of lack of care and either type of abuse dramatically increased the likelihood of subsequent homelessness | Lack of parental care was strongly associated with adult homelessness (unadjusted OR = 12.7, 95% CI = 5.0, 31.9; adjusted OR = 13.1, 95% CI = 4.7, 36.2). Childhood physical abuse significantly increased the risk of homelessness (unadjusted OR = 15.8, 95% CI = 5.9, 42.1; adjusted OR = 18.0, 95% CI = 5.4, 59.5). The combination of lack of care and either type of abuse (physical or sexual) resulted in a dramatically increased risk of homelessness (unadjusted OR = 26.0, 95% CI = 9.2, 73.6; adjusted OR = 38.9, 95% CI = 11.3, 133.6) | Childhood sexual abuse was associated with a nonsignificant trend toward homelessness in both univariate (unadjusted OR = 1.7, 95% CI = 0.6, 4.6) and multivariate analyses (adjusted OR = 1.6, 95% CI = 0.5, 5.2). | No | NA | Recall bias is a threat as the data on childhood risks and homelessness were collected retrospectively; Measures of physical and sexual abuse were limited to single-item questions; The sampling of currently housed individuals might underrepresent those with chronic, long-term homelessness who never reacquired stable housing; The observational design makes it difficult to make definitive causal claims; Measures of social support were restricted to current indicators, limiting the ability to formally test it as a mediating variable | ||
74 | https://doi.org/10.1111/j.1553-2712.2001.tb00209.x | Domestic abuse in the emergency department: can a risk profile be defined? | Zachary, Mary J. | 2001 | Academic Emergency Medicine | Quantitative, cross-sectional survey | cross-sectional survey conducted over a six-week period in the summer of 1995, with data collection occurring 24 hours a day, seven days a week | Cross-tabular, bivariate analysis and multivariate logistic regression | 611 women | Women presenting to an urban, public, academic-affiliated emergency department (ED) primarily serving a low-income group | Medical student interviewers approached adult female patients waiting to see a provider in the acute or urgent care areas of the ED | New York City, New York | Range: 16 - 100 years | 100% female | 45.2% Hispanic, 30.3% African American, and 24.4% White | NA | NA | NA | 57.9% were unemployed, and 20.5% had not completed high school | Yes | A semistructured questionnaire with screening questions adapted from American Medical Association guidelines. Severity was categorized using the Conflict Tactics Scale, Couple Form R (CTS). | Self-report in response to two direct questions: "Have you ever been emotionally or physically abused by your partner or someone important to you?" and "Are you currently being emotionally or physically abused by your partner or someone important to you?" | Emotional, physical, and sexual abuse, and threats of physical or sexual abuse | No | NA | NA | Homelessness (current status at the time of ED visit) | Self-report as part of demographic information collected via interview | Self-identification as homeless within the residence demographic category | Lifetime domestic abuse was significantly associated with homelessness, though recent domestic abuse (within the past 12 months) was not | In the bivariate analysis, homeless women were significantly more likely to have experienced lifetime domestic abuse (OR 9.5, 95% CI = 3.9 to 23.2. In the multivariate logistic regression, homelessness remained strongly and significantly associated with lifetime domestic abuse (OR 5.8, 95% CI = 2.2 to 15.0). | Recent domestic abuse (within the preceding 12 months) was not significantly associated with homelessness in the bivariate analysis (OR 1.0, 95% CI = 0.3 to 3.6) | No | NA | Convenience sampling rather than consecutive enrollment may have introduced selection bias; The study excluded non-English-speaking women, which limited findings regarding the immigrant population; Screening questions were not validated, and the study relied on the patient's own assessment of what constitutes "abuse"; Small sample sizes in some clinical categories may have affected the results | |
75 | Comparison of psychosocial and behavioral profiles of victimized and nonvictimized homeless women and their intimate partners | Nyamathi, Adeline | 2001 | Research in Nursing & Health | Quantitative, quasi-experimental, and cross-sectional | cross-sectional analysis of baseline data from a larger longitudinal health promotion study | Chi-square tests, t-tests, Wilcoxon two-sample tests, and multivariate logistic regression analysis | 1014 total (507 homeless women and their 507 intimate male partners) | Homeless women and their partners | Flyers and presentations at 36 shelters, and outreach by research staff to women in their "places of abode," such as streets and alleys | Los Angeles, California | Women: Mean 35 years (SD 8.6); range 18–50 years. Partners: Mean 37 years (SD 8.7), range 18 years or older | primary subjects: 100% female; intimate sexual partners: 100% male | Women: 60% African American, 17% White, 13% low-acculturated Latina, 9% high-acculturated Latina, and 1% mixed/other. Partners: 66% African American, 24% Latino, and 12.7% White | NA | NA | NA | Extremely impoverished and homeless. Mean education was 11.1 years for women and 11.5 years for partners. | Yes | Two yes/no items asking if the woman had been physically or sexually assaulted since living on her own; affirmative responses were followed by questions about the perpetrator's relationship to the woman (e.g., intimate partner) | Self-report of physical or sexual assault since living on one's own | Physical assault and sexual assault | Yes | Two single yes/no items asking "Were you ever sexually abused as a child?" and "Were you ever physically abused as a child?" | Physical abuse and sexual abuse | Literal homelessness; categorized as chronic (homeless for at least one year) and episodic (homeless more than once) | Structured questionnaire eliciting length of time and number of times homeless | Spending at least one night in a shelter or a place not intended for human habitation, or being uncertain of a residence in the next 60 days | While victimized women and their partners showed significant bivariate correlations with chronic homelessness, these housing factors did not remain significant in the final multivariate models when controlling for other variables. | Bivariate analysis showed that victimized women were significantly more likely to be long-term (chronic) homeless (57.4% vs 46.5%, p<.05) and to have had multiple episodes of homelessness (63.2% vs 50.0%, p<.01) than nonvictimized women. The male partners of victimized women were also significantly more likely to be chronically homeless (58.8% vs 48.7%, p<.05) compared to the partners of nonvictimized women. | In the multivariate logistic regression model for women, chronic homelessness (Adjusted OR 0.94, 95% CI = 0.58, 1.52, p=.788) and having been homeless more than once (Adjusted OR 1.08, 95% CI = 0.67, 1.73, p=.757) were not significant independent predictors of adult victimization. For male partners, the odds of being the partner of a victimized woman were not significantly impacted by their own chronic homelessness (Adjusted OR 1.16, 95% CI = 0.74, 1.83, p=.515) or multiple episodes of homelessness (Adjusted OR 1.13, 95% CI = 0.72, 1.83, p=.592) in the multivariate model. | No | NA | Potential underreporting of abuse due to the sensitive nature of the topic. Use of single-item questions and labeling (e.g., "abuse," "rape") may have limited disclosure. The cross-sectional design prevents determining the temporal order or causality between variables like substance use, mental health, and victimization. It was unknown if the current intimate partners were involved in the reported past partner abuse | Child maltreatment questions were not specific about maltreatment occurring by a caregiver | |
76 | https://doi.org/10.1007/s10964-008-9331-6 | A prospective examination of the path from child abuse and neglect to illicit drug use in middle adulthood: the potential mediating role of four risk factors | Wilson, Helen W. | 2008 | Journal of Youth and Adolescence | Quantitative, prospective, and longitudinal | A prospective cohort design following children with documented cases of maltreatment (ages 0–11) from 1967–1971. Mediators were assessed in young adulthood (approx. age 29, 1989–1995) and the primary outcome was assessed in middle adulthood (approx. age 41, 2000–2002). | Confirmatory factor analysis and latent variable structural equation modeling | Original N=1,575; Young adulthood interview N=1,196; Middle adulthood interview N=896 | Child welfare-involved (substantiated cases of maltreatment) and a matched comparison group | Abused and neglected children were identified through official court records (juvenile/family or adult criminal courts). Control subjects were matched on age, sex, race/ethnicity, and approximate family social class | Midwestern United States | Young adulthood: mean was 29.2 years, and range was 19 - 40.7 years; Middle adulthood: mean was 39.5 years, and range was 30 - 47 years. | Young adulthood: 48.7% female and 51.3% male; middle adulthood: 51.0% female and 49.0% male | Young adulthood: 61.5% White, 32.5% Black, and 6.0% Hispanic and other; middle adulthood: 60.8% White, 32.6% Black, and 6.6% Hispanic and other | Initial cases were children 11 years or less at time of incident | 49.0% female and 51.0% male | 66.1% White, 32.5% Black, and 1.5% Hispanic and other | skewed toward the lower end of the socioeconomic spectrum | No | NA | NA | NA | Yes | Review of official records from 1967–1971 | Physical abuse, sexual abuse, and neglect | Homelessness in young adulthood | Self-report of having experienced a period of homelessness in the past 12 months or having ever had no regular place to live for at least a month or so | Affirmative response to either of the two self-report items | For women, a "problem behavior syndrome" (comprised of homelessness, prostitution, crime, and school problems) fully mediated the path between childhood maltreatment and middle-adulthood illicit drug use. For men, while childhood maltreatment predicted the young adult risk factors (including homelessness), these factors did not predict drug use in middle adulthood | For women, childhood abuse and neglect (CAN) was significantly associated with homelessness in young adulthood (β=.34,p<.001). For men, CAN was also significantly associated with homelessness in young adulthood (β=.32,p<.01). In single-factor models for women, homelessness significantly mediated the relationship between CAN and illicit drug use (CAN to homelessness β=.30,p<.001; homelessness to drug use β=.18,p<.05). | When all four mediators were examined together in the full structural model, no individual path from homelessness to drug use remained significant for either gender, as the effects were overlapping. | No | NA | Documented cases most likely represent the most extreme instances and do not capture cases that did not come to the attention of authorities; The results from the Midwest in the late 1960s/70s may not generalize to all cases or regions; The sample is skewed toward lower SES, potentially limiting generalizability to middle-class populations; The study did not control for all potential factors influencing the effects of childhood victimization | |
77 | Determinants of behavior in homeless and low-income housed preschool children | Bassuk, Ellen L. | 1997 | Pediatrics | Quantitative, observational, and used an unmatched case-control design | cross-sectional study that compared sheltered homeless families with low-income housed families who had never been homeless | T-tests, Wilcoxon rank-sum tests, chi-square tests, Pearson product-moment correlations, and multiple linear regression | 167 preschool children (77 homeless and 90 low-income housed) | Sheltered homeless families and low-income housed familie | Homeless families were recruited from all nine emergency shelters and transitional housing facilities in Worcester, plus two welfare hotels. Housed families were recruited from female-headed families receiving AFDC (Aid to Families With Dependent Children) at the local Department of Public Welfare office | Worcester, Massachusetts | mean: 27 years | 100% female | 35.1% to 36.7% were Puerto Rican | Mean: 4 years (2 years, 6 months - 5 years 11 months) | 60% boys and 40% girls | White (35.1% homeless vs. 41.1% housed), Black (14.3% homeless vs. 13.3% housed), Puerto Rican (35.1% homeless vs. 36.7% housed), and Other (15.6% homeless vs. 8.9% housed) | All families lived below the poverty level. Housed families received AFDC | Yes | Conflict Tactics Scale (CTS) | Mothers were asked about physical and sexual victimization by intimates and nonintimates across their lifespan | Physical and sexual assault by intimate partners and nonintimates | Yes | Maternal report of physical or sexual abuse, and official records of care and protection investigations (51A) or foster care placements | Sexual abuse, physical abuse, and neglect | Literal homelessness (living in shelters/hotels) and residential instability (number of moves) | modified Personal History Form | Homeless: In a shelter or hotel for at least seven days. Housed: Receiving AFDC and never having been homeless | Homeless preschoolers were significantly more likely to have undergone a neglect investigation and been placed in foster care | Homeless preschoolers were significantly more likely to have undergone a neglect investigation (40.3% vs. 21.2%, p=.007) and been placed in foster care (19.5% vs. 7.8%, p=.03) | There were no significant differences between homeless and housed children in maternal reports of child sexual abuse (5.3% vs. 4.5%, p=.82) or child physical abuse (10.5% vs. 6.8%, p=.40) | No | NA | The sample size was small, and the data may have missed information as fathers were not interviewed directly; cross-sectional design prevents determining causal directions; CBCL lacks normative data for specific minority groups, such as Puerto Rican Hispanics | ||
78 | https://doi.org/10.1002/ajcp.12136 | The Impact of Family Unification Housing Vouchers on Child Welfare Outcomes | Pergamit, Michael | 2017 | American Journal of Community Psychology | Quantitative and quasi-experimental; waitlist comparison design | Multi-site quasi-experiment that tracked families referred to the Family Unification Program against a comparison group of eligible families on a waitlist. Outcomes were measured through 18 months post-referral using child welfare administrative data. It also used propensity score matching as an additional check for reunification outcomes in one site. | Probit regressions, Cox-proportional hazards, and instrumental variables for Complier Adjusted Causal Effect | 326 children in Portland and 502 children in San Diego | Child welfare-involved families, divided into preservation cases (intact families) and reunification cases (families with children in out-of-home care) | Caseworker referral of families determined eligible under FUP criteria (inadequate housing as a primary factor in imminent placement or delaying discharge) | Portland, Oregon and San Diego, California | Portland mean: 32 years; San Diego mean: 31 years | 79% female in Portland, 88% female in San Diego | Portland household heads were 94% minority. San Diego household heads were approximately 79% minority | Preservation mean ages: Portland 6.95/6.78 (waitlist/treatment); San Diego 6.24/6.66. Reunification mean ages: Portland 7.72/7.09; San Diego 6.92/6.55 | Preservation: 56%/59% female (Portland) and 49%/51% female (San Diego). Reunification: 37%/51% female (Portland) and 43%/42% female (San Diego). | Portland preservation: waitlist was 42% White, 14% Black, 26% Hispanic, and 16% Native American San Diego preservation, waitlist was 18% White, 38% Hispanic, and 40% Black; in Portland reunification cases, the waitlist was 56% White, 12% Black, 9% Hispanic, and 21% Native American. In San Diego, it was 18% White, 43% Black, and 32% Hispanic | Extremely poor | Yes | Official FUP referral forms and child welfare administrative data | Being displaced by domestic violence at the time of referral | Displacement by domestic violence and emotional abuse (listed as a case open reason in San Diego) | Yes | Official substantiated reports from child welfare administrative data | Neglect, physical abuse, sexual abuse, and emotional abuse | Homelessness (literally homeless), overcrowded housing, dilapidated housing, and being at risk of losing housing | FUP referral forms and child welfare administrative data | housing that was substandard, overcrowded, unsafe, or homelessness | Provision of FUP vouchers had little impact on preventing child removal from the home but showed a positive impact on reunification (specifically in Portland) and significantly speeded up child welfare case closure. Impacts on new reports of abuse and neglect were mixed but generally pointed toward a reduction, except for an unexpected increase in reports for reunification cases in Portland. | In Portland, preservation cases were over 40% more likely to close if referred to FUP (ITT marginal effect 0.415, p<.01) and over 50% more likely if they successfully received a voucher (CACE 0.514, p<.01). FUP preservation cases in Portland closed roughly five times as fast as comparison cases (Hazard ratio 5.237, p<.01). FUP reduced the probability of a new report of maltreatment for preservation cases in San Diego by approximately 19.3% (ITT marginal effect -0.193, p<.05). In Portland preservation cases, FUP was associated with a significant reduction in substantiated re-reports of maltreatment (CACE marginal effect -0.256, p<.05). Using PSM, FUP families in Portland were 29.9% more likely to reunify than comparison families (p<.01). Reunification cases in Portland were 28.4% more likely to close under ITT (p<.01) and 38.2% more likely under CACE (p<.01). Reunification cases in Portland showed a significant increase in any new re-report (ITT marginal effect 0.574, p<.01) and substantiated re-reports (ITT marginal effect 0.248, p<.05) | In both sites, FUP showed no significant impact on preventing out-of-home placements for preservation cases. There was no significant impact of FUP on the probability of case closure for preservation cases in San Diego. The impact of FUP on the probability of reunification in San Diego was not statistically significant. In San Diego, there were no significant impacts of FUP on closing reunification cases or speeding up their closure | Yes | Family Unification Program | The waitlist design may have been affected by changes in referral behavior; selection bias may have occurred as early referrals came from existing caseloads; the study had low power due to small sample sizes; and the results from only two sites have limited generalizability | |
79 | https://doi.org/10.1177/1077559516679513 | Rethinking the Measurement of Adversity | Mersky, Joshua P. | 2016 | Child Maltreatment | quantitative, observational secondary analysis of data | study utilized a retrospective assessment of childhood experiences via a secondary analysis of data collected at enrollment in home visiting services. It also included a test-retest reliability component with a convenience subsample. | bivariate and multivariate regressions, Exploratory Factor Analysis, test-retest reliability assessments using Kappa coefficients and Intraclass Correlation Coefficients | 1,241 women. A subsample of n = 83 was used for test-retest reliability | A diverse sample of low-income women who received home visiting services | convenience sample of clients already enrolled in the Family Foundations Home Visiting (FFHV) program | Wisconsin, US | Mean age: 24.2 years (SD = 5.7); Range: 16–46 years | 100% female | 33.2% Caucasian (White, non-Hispanic), 27.4% African American (Black, non-Hispanic), 8.0% American Indian, 8.9% "other," and 22.6% Hispanic | NA | NA | NA | The sample was high-risk/low-income; over 98% had household pretax incomes at or below 200% of the federal poverty threshold or were eligible for federal means-tested benefits | No (only witnessing IPV between caregivers) | NA | NA | NA | Yes | Childhood Experiences Survey (CES), using BRFSS items supplemented with additional neglect questions | Physical abuse, sexual abuse, emotional abuse, physical neglect, and emotional neglect | Homelessness | A single CES item: "How often were you homeless when you were growing up?" (Rated on a 5-point Likert scale: never, rarely, sometimes, often, or very often) | Self-report of the frequency of childhood homelessness | Childhood homelessness was significantly and positively correlated with every form of family violence measured, including physical abuse, sexual abuse, emotional abuse, physical neglect, emotional neglect, and household domestic violence. | Childhood homelessness was significantly and positively associated with all forms of family violence (p < .01), including physical abuse (r = .22), sexual abuse (r = .21), emotional abuse (r = .24), physical neglect (r = .24), emotional neglect (r = .25), and household domestic violence (r = .23). | None. | No | NA | retrospective, self-reported data, which may be subject to memory bias; temporal order of events cannot be guaranteed, so potential for reverse causality; results may have limited generalizability as the sample consisted only of low-income women in home visiting services; omitted variable bias was a concern as the measurement models were not exhaustive | |
80 | https://doi.org/10.1002/jcop.20169 | Homelessness, violence exposure, and school participation among urban adolescent mothers | Kennedy, Angie C. | 2007 | Journal of Community Psychology | quantitative, exploratory cross-sectional survey | cross-sectional design with a purposive sampling approach | MANOVA, logistic regression, and hierarchical regression | 120 mothers | Poor, urban adolescent mothers | Purposive recruitment from an alternative high school, an adolescent health center, and a youth homeless shelter/street outreach program | large midwestern city, USA | Range : 16-20 years | 100% female | 54% Latina, 33% African American, and 13% White/Mixed ethnicity | NA | NA | NA | 87% reported receiving some type of public assistance | Yes | 12-item physical assault subscale of the Revised Conflict Tactics Scale (CTS2) | Self-report of both current (past 12 months) and lifetime ("ever") physical violence | Physical IPV | Yes | physical assault subscale of the CTS2 | physical abuse (e.g., being slapped, beaten up, or threatened with a weapon) by a parent or adult caregiver | Homelessness (current and historical) | Self-report via two questions: “Are you homeless right now?” and “If you are not homeless right now, have you ever been homeless?” | Participants were categorized as having "ever experienced homelessness" if they answered yes to either current or historical status | Participants who had been homeless reported significantly higher rates of exposure to all types of violence (witnessing parental violence, physical abuse, and partner violence). Violence exposure, moderated by social support, was a significant contributor to the odds of ever being homeless | MANOVA indicated that participants who had ever been homeless reported significantly higher rates of physical abuse by a parent/caregiver (F=30.61,p=.000), and partner violence (F=17.83,p=.000). As partner violence exposure increases, each increase in social support results in a greater decrease in the odds of experiencing homelessness (Exp[B]=.97,p=.03). | In the final logistic regression model, physical abuse by a parent or adult caregiver (PAP) was not a significant independent predictor of ever-homeless status (B=.19,Exp[B]=1.21,p>.05). | No | NA | use of a non-probability/purposive sample; cross-sectional design prevents establishing temporal causality; potential for self-report bias or confusion regarding the definition of homelessness | MANOVA indicated that participants who had ever been homeless reported significantly higher rates of witnessing parental violence (F =15.48,p=.000). Logistic regression showed that each increase in witnessing parental violence increased the odds of ever being homeless by 31% (Exp[B]=1.31,p=.04). |
81 | https://doi.org/10.1080/10911359.2021.2004965 | Race and retention in permanent supportive housing: a secondary analysis of housing outcomes | Twis, Mary | 2021 | Journal of Human Behavior in the Social Environment | quantitative, observational secondary analysis of data | secondary data analysis using HMIS reports spanning 2018–2020 | t-tests, chi-square, logistic regression | 407 participants | Formerly homeless individuals residing in permanent supportive housing (PSH) | Purposive,conveniencesampling; the dataset was obtained from a local Continuum of Care (CoC) Lead Agency's HMIS reports | North Texas, US | Mean: 49.39 (SD = 11.8) | 66% male and 34% female | 49.3% White, 48.8% Black/African-American, 1.0% Native American/Alaska Native, 0.7% Native Hawaiian/Pacific Islander, and 0.5% Asian. 8.4% reported Hispanic ethnicity | NA | NA | NA | The mean monthly income was $979.72 (SD = 543.83) | Yes | Based on HMIS data (typically self-report or caseworker entry) | Identification as a "victim of domestic violence" within the HMIS client records | General "domestic violence" | No | NA | NA | Housing loss (vs. retention) in permanent supportive housing and history of homelessness (number of episodes in the last three years) | HMIS tracking data of housing status and historical episodes | Participants were categorized based on whether they "exited" or "left" PSH during the study period compared to those who retained their housing | Having experienced domestic violence significantly predicted housing loss only for White service recipients. | Among White service recipients, having experienced domestic violence significantly predicted housing loss (OR = 3.73, 95% CI = 1.02–13.65, p = .047). | For nonwhite service recipients, the regression models did not identify any significant predictors of housing loss. | Yes | Permanent Supportive Housing | large amounts of missing data in the HMIS database needed data imputation; Because it relied on secondary data, the study was unable to include the primary voices of people experiencing homelessness; sample was restricted to a single geographic area, so the findings may not be generalizable to the entire United States | |
82 | https://doi.org/10.1016/j.childyouth.2008.01.001 | Where the homeless children and youth come from: A study of the residential origins of the homeless in Miami-Dade County, Florida | Rukmana, Deden | 2008 | Children and Youth Services Review | Quantitative, observational point-in-time survey | cross-sectional spatial analysis using data from an annual point-in-time homelessness survey | Geographic Information Systems, Principal Components Analysis, OLS regression | 1,201 total unduplicated survey respondent. The specific analysis for families with children included 219 homeless families | Homeless families accompanied by children under 18 years of age | Purposive and voluntary sampling was conducted through 50 participating homelessness programs, including outreach, supportive services, emergency shelters, and transitional housing | Miami-Dade County, Florida | Range: 18 - 82 years | NR | NR | Range: 0 - 17 years | 302 girls, 244 boys | NR | NR | Yes | Measured as a neighborhood-level "Risk of domestic violence" factor derived via PCA from census data variables, including the proportion of unemployed females, overall unemployment, and racial composition | ecological risk factor rather than individual screening | Domestic violence | No | NA | NA | Homelessness (residential origins and spatial distribution of homeless children and youth) | Self-report via the survey question: “What was the address of the last house or apartment you lived in?" | Respondents were currently homeless families residing in shelters, transitional housing, or on the streets on the day of the survey | The study found that the residential origins of homeless children are not heavily concentrated in poor neighborhoods but are also located in less poor neighborhoods. Domestic violence was identified as the factor that explains this distribution, as it is not confined strictly to neighborhoods of high poverty. | In the OLS regression analysis (N=219), the "Risk of domestic violence" factor was the strongest predictor of the rate of homelessness among families (β=.580,p<0.01). | None. | No | NA | non-probability, voluntary sample, which may limit generalizability; a relatively large portion of missing prior addresses (171 out of 1,201) | |
83 | https://dx.doi.org/10.1016/j.childyouth.2013.08.012 | Inadequate housing and the child protection system response | Font, Sarah A. | 2013 | Children and Youth Services Review | quantitative, observational, and longitudinal study using a nationally representative sample | The study utilizes the second cohort of the National Survey of Child and Adolescent Well-Being II (NSCAW II). Baseline data were collected between 2008 and 2009, with a follow-up wave occurring 18 months later. | nested logistic regression models | 2,987 cases | Child welfare-involved families | stratified sample of child welfare investigations across 81 counties in the United States | United States (81 counties) | Mean: 32.68 years (SE = 0.27) | NR | 49.85% White, 17.28% Black, 32.87% Other race (including Hispanic) | NA | NA | NA | mean logged household income was 9.85 (SE 0.05); 50.55% were employed | Yes | caseworker's risk assessment | dichotomous indicator was used (1 if the caseworker reported the family was experiencing active domestic violence) | General "domestic violence" | Yes | caseworker reports of investigation allegations and the final determination of substantiation status | Abuse (physical, sexual, or emotional) and neglect (physical, educational, or abandonment) | Doubling up (living with non-immediate family) and recent homelessness (use of emergency housing in the past 12 months) | Caregiver self-reports on a household roster (for doubling up) and caregiver reports of emergency housing use | "Doubled up" meant listing at least two non-immediate family members in the home. "Recent homelessness" was defined as staying in emergency housing at least once in the previous 12 months | Families experiencing homelessness were more likely to be identified as experiencing domestic violence (20.56%) compared to doubled-up families (13.96%) and stably housed families (12.09%). Homeless families were more likely to have allegations of neglect (73.54%) compared to non-homeless families (53.92%). Doubled-up families also had higher neglect allegations (61.98%) than those not doubled up (53.81%). Housing instability alone was not directly associated with substantiation, but it was associated with decreased odds of case closure and increased odds of service receipt. The interaction of homelessness and domestic violence significantly reduced the likelihood of a family receiving services from the child protection syste | Families who experienced homelessness were more likely to be victims of domestic violence (20.56% vs. 12.09%) than stably housed families. For families experiencing both homelessness and domestic violence, the odds of receiving services were significantly decreased compared to other families (p<.05). The joint significance of housing, domestic violence, and their interaction was statistically significant for the likelihood of substantiation (p<.001) and service receipt (p<.05). | The interaction between the housing measures and domestic violence did not show a statistically significant independent association with substantiation odds (Doubling up * DV: OR 0.459; Homelessness * DV: OR 1.555). For the odds of case closure, the interaction between housing instability and domestic violence was not statistically significant (Doubling up * DV: OR 0.311; Homelessness * DV: OR 0.766). For doubled-up families also experiencing domestic violence, the odds of service receipt were not significantly different from doubled-up families without domestic violence (OR 0.780). | No | NA | Lack of data on the duration or timing of homelessness episodes; limited information on the nature of doubling up (e.g., if it led to overcrowding); low statistical power for specific subgroups due to the relatively low incidence of reported housing problems | The study suggests that child welfare workers may view housing instability as an issue of poverty rather than neglect, but the combination of housing issues and other risks (like DV) complicates the system response, highlighting a need for better integration between housing and child protection services. |
84 | https://doi.org/10.1007/s10566-016-9370-1 | A Growth Curve Analysis of Housing Quality among Transition-Aged Foster Youth | Tyrell, Fanita A. | 2017 | Child & Youth Care Forum | quantitative, observational, and longitudinal study | A prospective longitudinal investigation documenting month-by-month changes in housing quality during the first 24 months following emancipation from foster care . Data were collected across three waves (baseline) | Multilevel modeling (Growth Curve Analysis). The study fit unconditional and conditional linear and quadratic growth models | 172 recently emancipated foster youth | Transition-aged foster youth who had "aged out" or "emancipated" from state-sponsored care | recruited via flyers distributed to social service providers, independent living programs, and agencies serving emancipated foster youth | Southern California | Mean: 19.63 years (SD = 1.11); range: 18 - 21 years | 66% female | 84.3% non-White (15.7% European American/White, 23.8% African American/Black, 27.3% Latino American/Hispanic, 0.6% Asian American, and 32.6% multiracial) | NA | NA | NA | 72.7% had completed high school or a GED | No (only witnessing IPV between caregivers) | NA | NA | NA | Yes | Verbal administration of the ETI; two independent raters evaluated each type using McGee and colleagues' (1995) criteria | Physical abuse, sexual abuse, emotional abuse, and neglect | Continuous "housing quality" construct including homelessness, precarious housing (e.g., "couch surfing" or "doubling up"), residential instability (frequency of moves), and financial factors like rent/mortgage costs | Residential history timeline used to code living situations on a 9-point scale (1 = extremely low/homeless to 9 = extremely high/homeowner) | Participants' housing was tracked month-by-month post-emancipation | On average, youth experienced moderate housing quality with modest gains over 24 months, though they struggled to achieve independent housing. Childhood physical abuse, sexual abuse, emotional abuse, and neglect were not significantly related to the housing quality intercept, nor to linear or quadratic change in housing quality. | None. | Childhood physical abuse, sexual abuse, emotional abuse, and neglect were not significantly related to the housing quality intercept, nor to linear or quadratic change in housing quality | No | NA | sample was non-random and limited to Southern California, so may limit generalizability; sample size may not have been sufficient to detect small effects; housing quality coding system was novel and unvalidated; data relied on retrospective reports; reliability for child neglect was relatively low compared to other maltreatment types | Exposure to domestic violence was associated with declining housing quality over time. |
85 | https://doi.org/10.1177/10775595241276416 | Longitudinal Impacts of Housing Cost Burden on Physical and Psychological Aggression of Children From Age 3 to 15 | Marçal, Katherine E. | 2025 | Child Maltreatment | quantitative, observational, and longitudinal study | Data from the Fragile Families and Child Well-being Study. Mothers were interviewed at the child's birth and at 1-, 3-, 5-, 9-, and 15-year intervals. The analysis focused on the waves from Year 3 to Year 15 | Latent growth curve modeling | 1,778 mothers | Mothers who remained the primary caregivers for their children | Recruited in hospitals after birth using stratified clustered sampling, with an intentional oversample of unmarried mothers | 20 large American cities | Mean: 25.32 years (SD = 6.03) | 100% female | 51.35% Black, 22.44% White, 22.67% Hispanic/Latino, and 3.54% Other | NA | NA | NA | largely lower-income | No | NA | NA | NA | Yes | Conflict Tactics Scale, Parent-Child version (CTS-PC) | Physical aggression (hitting or slapping with a bare hand or object) and psychological aggression (yelling, screaming, or cursing) | Housing cost burden (missing a rent or mortgage payment due to inability to pay). The study controlled for eviction, doubling up, and homelessness. | Self-report of whether the mother missed a rent or mortgage payment in the past 12 months | Dichotomous indicator (1 = yes, 0 = no) based on reporting a missed payment | Housing cost burden contributed to significant ongoing risk for parenting aggression. Housing cost burden was associated with psychological aggression within the same time period at ages 3, 5, and 9. Housing cost burden at age 3 had long-lasting (lagged) effects on physical aggression up through age 9. Housing cost burden at age 5 was specifically associated with physical aggression when the child reached age 15 | Housing cost burden at Year 3 was significantly associated with increased physical aggression at age 3 (β=0.41,95% CI [0.195,0.634],p<.001), age 5 (β=0.34,95% CI [0.127,0.559],p<.01), and age 9 (β=0.21,95% CI [0.010,0.415],p<.05). Housing cost burden at Year 3 was significantly associated with concurrent psychological aggression at age 3 (β=0.28,95% CI [0.076,0.481],p<.01). Housing cost burden at Year 5 was significantly associated with concurrent psychological aggression at age 5 (β=0.31,95% CI [−.126,0.492],p<.01) and physical aggression at age 15 (β=0.08,95% CI [0.026,0.129],p<.01). Housing cost burden at Year 9 was significantly associated with concurrent psychological aggression at age 9 (β=0.16,95% CI [0.009,0.305],p<.05). Doubling up at Year 5 was significantly associated with a decrease in physical aggression at age 5 (β=−0.28,95% CI [−0.528,−0.026],p<.05). | Early housing cost burden did not have a statistically significant lagged association with psychological aggression at later waves. Housing cost burden at any measured time point was not significantly associated with psychological aggression at age 15. Experiences of eviction and literal homelessness were not significantly associated with physical or psychological aggression. Housing cost burden at Year 9 was not significantly associated with physical aggression at age 9 (β=0.04,95% CI [−0.122,0.197],p>.05) or age 15 (β=−0.02,95% CI [−0.065,0.023],p>.05). Housing cost burden at Year 15 was not significantly associated with physical aggression (β=0.01,95% CI [−0.034,0.060],p>.05) or psychological aggression (β=0.06,95% CI [−0.016,0.138],p>.05) at age 15. | No | NA | non-representative oversample of unmarried urban mothers may limit generalizability; housing cost burden was measured by a single dichotomous item which may miss some families; reliance on parent self-reports may introduce social desirability bias; the number of aggression items was significantly reduced at the 15-year wave; and the exclusion of children removed from maternal custody may have left out those at the highest risk | |
86 | https://doi.org/10.1177/0886260520943725 | The Effect of Home Foreclosures on Child Maltreatment Rates: A Longitudinal Examination of Neighborhoods in Cleveland, Ohio | Hoyle, Mary Elizabeth | 2022 | Journal of Interpersonal Violence | quantitative, observational, and longitudinal study conducted at the neighborhood (census tract) level | cross-sectional time series design with an autoregressive error structure. The data consists of nine time periods (2006–2015) nested within 175 neighborhoods (census tracts) | cross-sectional time series models with a lagged dependent variable | 1,575 neighborhood-year observations (175 neighborhoods followed over 9 years) | Neighborhoods (census tracts) | synthesized data from the Northeast Ohio and Neighborhood Data for Organizing and the American Community Survey | Cleveland, Ohio | NR | NR | neighborhood racial/ethnic composition mean was approximately 56.6% Black and 8.8% to 9.7% Hispanic | NR | NR | NR | Mean neighborhood poverty rates ranged from 23.6% to 27.1%, and between 44.6% and 53.1% of neighborhoods had above-average unemployment | No | NA | NA | NA | Yes | Substantiated child maltreatment cases from the NEO CANDO System | Allegations of abuse or neglect that were investigated and confirmed by the Department of Children and Family Services | Home foreclosures | Foreclosure filings issued by the Cuyahoga County Common Pleas Court | All instances where a property was placed in foreclosure status, regardless of whether the home was ultimately sold at auction | Neighborhood foreclosure rates are a significant predictor of neighborhood rates of substantiated child maltreatment. This effect was found to be durable over time and persisted even after the housing crisis. The relationship between foreclosures and child maltreatment was direct and not moderated by other neighborhood factors such as poverty, vacancy, or racial composition. | There is a strong, positive link between foreclosure rates and child maltreatment: for every one-unit increase in the foreclosure rate, the maltreatment rate increased by 0.749 (B=0.749,SE=0.313,p<.05). The interaction between the foreclosure rate and the year 2010 was significant and negative (B=−2.284,SE=1.046,p<.05), indicating the effect was less pronounced during that specific year compared to 2007. | Neither esidential instability (% moved) nor vacancy were significantly related to neighborhood child maltreatment rates in the longitudinal models. Interactions between foreclosure rates and other neighborhood structural characteristics (poverty, care burden, unemployment, race, and vacancy) were not statistically significant, indicating the foreclosure-maltreatment relationship is direct. | No | NA | study only included substantiated cases, which may be subject to reporting bias; potential for ecological bias when drawing conclusions about individuals from neighborhood-level data; reliance on 5-year ACS averages rather than true yearly neighborhood data; findings may not be generalizable to other geographic locations; the study did not have data on actual social networks or social supports within the neighborhoods | |
87 | Physically Hazardous Housing and Risk for Child Protective Services Involvement | Hirsch, Bomi Kim | 2015 | Child Welfare | Quantitative, longitudinal observational study | Study utilized longitudinal data from the first wave of the Illinois Families Study (IFS) (1999–2000), which were linked to administrative records on Child Protective Services (CPS) investigations ranging from 1981 through one year following the baseline interview | Hierarchical logistic regression models | 1,363 individuals | Low-income families | stratified sample of Temporary Assistance for Needy Families (TANF) cases in 1998 | Nine counties in Illinois (both urban and rural areas) | Mean: 31.68 (SE: 8.09) | NR | Non-Hispanic Black (77.96%), Hispanic (11.83%), Non-Hispanic White (8.16%), and Other (2.05%) | NR | NR | NR | low-income families receiving TANF; 50.39% were employed at the time of the interview | No | NA | NA | NA | Yes | Administrative records from Child Protective Services | Alleged neglect, alleged physical abuse, and indicated (substantiated) reports of any maltreatment type | Moving history (instability), residential crowding, and hazardous physical housing conditions (e.g., infestations, nonworking utilities, structural deficiencies) | Self-report survey items | "Hazardous housing" was defined as reporting at least one of seven specified hazards; "Moving history" was defined as one or more moves in the past year; "Crowding" was measured as the ratio of people per bedroo | Hazardous physical housing conditions significantly predicted investigated child neglect reports, but were not predictive of physical abuse allegations or indicated (substantiated) reports. Families investigated by CPS were significantly more likely to report hazardous housing, residential crowding, and a history of frequent moves than those without CPS involvement | In bivariate analyses, families with any subsequent CPS investigation were significantly more likely to report hazardous physical housing conditions (44.76% vs. 33.84% in the full sample; p < .05). Bivariate results showed that families with any CPS investigation had significantly higher residential crowding (Mean = 1.96 vs. 1.68; p < .01) and a higher prevalence of moving history (38.18% vs. 24.78%; p < .01). Families with neglect investigations specifically were significantly more likely to report housing hazards (49.31%; p < .05), higher crowding (Mean = 1.99; p < .01), and a moving history (43.80%; p < .01). In multivariate models controlling for demographics, other housing characteristics, and economic well-being, housing hazards significantly increased the odds of any CPS investigation (OR = 1.66; 95% CI: 1.04, 2.65). Hazardous housing retained its statistical significance as a predictor for neglect-related investigations even after controlling for psychosocial well-being. Moving history remained a statistically significant predictor of any CPS investigation in the full multivariate model (OR = 1.66; 95% CI: 1.03, 2.66) | In the full multivariate model, the association between hazardous housing and any CPS investigation was reduced to marginal significance when psychosocial factors were included (OR = 1.59; 95% CI: 0.98, 2.56; p = .058). Hazardous physical housing conditions were not significantly associated with alleged physical abuse reports or indicated (substantiated) CPS reports. Residential crowding was not a statistically significant predictor of any CPS investigation in the multivariate hierarchical models (OR = 1.17; 95% CI: 0.91, 1.51). | No | NA | Results are based on a sample of Illinois welfare participants from the early 2000s and may not be generalizable to current national populations of TANF recipients or the general public; study cannot determine a causal relationship between housing hazards and CPS involvement; results may be subject to omitted variable bias because of observational nature; low occurrence of indicated (substantiated) reports (3%) limits the statistical power to identify associations with that specific outcome | ||
88 | https://doi.org/10.1080/10538712.2011.622354 | The impact of child, family, and child protective services factors on reports of child sexual abuse recurrence | Sinanan, Allison N. | 2011 | Journal of Child Sexual Abuse | Quantitative survival analysis | longitudinal survival analysis with a three-year follow-up period (2002–2004) utilizing administrative data | Univariate survival analysis, Cox proportional hazards regression, partial Gram-Schmidt orthogonalization method | 3,835 substantiated index reports of sexual abuse | Children younger than 18 with a substantiated case of child sexual abuse | Data were extracted from the National Child Abuse and Neglect Data System (NCANDS) for children in seven specific states | Colorado, Maine, Minnesota, Missouri, Texas, Utah, and Wyoming | NR | NR | NR | Children were classified into four groups: 0–2 (6.2%), 3–5 (18%), 6–11 (44%), and 12–17 (31%) | 24% Male, 76% Female | White (62%), Hispanic (25%), and African American (16%) | 15% of the sample families reported financial problems, and 8% reported inadequate housing | Yes | Attempt or existence of harm was recorded via state administrative data in the NCANDS dataset | Administrative records indicating physical, sexual, or psychological harm by one intimate partner to another | Physical, sexual, or psychological harm | Yes | Substantiated reports within the NCANDS administrative data system | Child sexual abuse recurrence; cross-type recurrence including neglect and physical abuse | Inadequate housing (defined as substandard, overcrowded, unsafe, or homelessness) | NCANDS administrative dataset | Families identified in administrative records as living in substandard, overcrowded, unsafe, or otherwise inadequate housing conditions, including homelessness | Inadequate housing was not a statistically significant predictor of child sexual abuse recurrence in this sample. | None. | Inadequate housing was not associated with reports of child sexual abuse recurrence to a statistically significant degree in univariate analysis (Risk Ratio = 1.009, 95% CI: .782–1.300). | No | NA | Administrative data may contain errors, missing data, or inaccuracies compared to case files; difficulty comparing records across states due to differing standards and definitions of abuse; the study only examined recurrence over a three-year period | |
89 | https://doi.org/10.1016/j.drugalcdep.2019.107623 | Opioid prescription rates and risk for substantiated child abuse and neglect: A Bayesian spatiotemporal analysis | Morris, Matthew C. | 2019 | Drug and Alcohol Dependence | Quantitative; Bayesian spatiotemporal analysis | longitudinal study spanning an 11-year period (2006–2016) using annual summary statistics for 95 counties; utilized an autoregressive model to describe time trends | Bayesian spatiotemporal modeling (specifically an autoregressive model) using Markov Chain Monte Carlo (MCMC) techniques | 95 counties | Population-level data for all counties in a single state | Administrative and national databases, such as the KIDS COUNT Data Center, Tennessee Department of Children’s Services, IQVIA Xponent database, and the U.S. Census Bureau | Tennessee | NR | NR | NR | NR | NR | NR | County-level indicator of child poverty rates: Mean = 24.62% to 29.59% | No | NA | NA | NA | Yes | Data on unduplicated counts of substantiated child abuse and neglect cases provided by the Tennessee Department of Children’s Services | Substantiated child abuse and neglect (subtypes such as physical or sexual abuse were not distinguished in the analysis) | vacant housing | 5-year estimates from the U.S. Census American Community Surveys (2006–2011 and 2012–2016) | Percentage of vacant housing units within each county | Vacant housing was positively associated with the risk of substantiated child abuse and neglect | A 10% increase in vacant housing units was associated with a 16% increase in the rates of substantiated child abuse and neglect (Relative Risk [RR] = 1.016; 95% High Density Region [HDR]: 1.005, 1.026). | None. | No | NA | Substantiated case data did not distinguish between subtypes of maltreatment (e.g., neglect vs. physical abuse); Substantiation rates are susceptible to social-ecological factors that influence whether maltreatment is detected; use of aggregate county-level data cannot directly inform inferences regarding individual- or family-level behaviors (ecological fallacy) | |
90 | https://doi.org/10.1016/j.chiabu.2014.05.009 | The effect of material hardship on child protective service involvement | Yang, Mi-Youn | 2015 | Child Abuse & Neglect | Quantitative, prospective longitudinal panel study | Four waves of survey data (1999–2000, 2001, 2002, and 2004) from the Illinois Families Study (IFS), linked to administrative child protection records. It used a prospective design to ensure that material hardship measures preceded CPS outcomes | t-tests, chi-square test, pooled logistic regressions, and caregiver-specific fixed effects logistic regressions to account for unobserved heterogeneity | 1,135 families (contributing 4,237 person-wave observations) . | Low-income families, specifically a sample of Temporary Assistance for Needy Families (TANF) recipients | Stratified random sampling from the 1998 Illinois welfare caseload in nine counties | nine Illinois counties | Mean at Wave 1: 31.51 (SE = 0.25) | 97.30% Female | 79.54% Non-Hispanic Black, 7.24% Non-Hispanic White, and 13.22% Hispanic | NR | NR | NR | Very low-income | Yes (as a control variable) | Self-report of severe physical domestic violence | Respondents reporting being hit, slapped, or kicked by a partner or spouse in the past year | Severe physical domestic violence | Yes | Administrative child protective services records | Investigated reports of neglect, physical abuse, and sexual abuse | Housing hardship | Self-report items capturing difficulty paying rent, doubling up, eviction, or homelessness | Reporting any of the specified housing difficulties in the past 12 months | Housing hardship was the most consistent and strong predictor of all types of CPS investigations. Physical abuse investigations were more responsive to the total level of hardship (multiple co-occurring hardships) rather than any one specific type. | Experiencing housing hardship specifically increased the odds of an investigation for any type of maltreatment by 2.22 times (p < .01) and for neglect by 1.84 times (p < .05). Bivariate correlations showed families with CPS involvement were significantly more likely to report housing hardship (35.96% vs. 25.65%, p < .001). | Physical abuse was not associated with housing hardship when analyzed in fixed effects models. | No | NA | sample is restricted to welfare recipients and may not generalize to the broader low-income population; study relies on investigated reports rather than actual occurrences of maltreatment; potential measurement errors in the material hardship constructs (e.g., failing to capture crowding or housing quality); the model may still be subject to bias from unobserved time-varying characteristics | |
91 | https://doi.org/10.1016/j.chiabu.2017.02.024 | Environmental and individual attributes associated with child maltreatment resulting in hospitalization or death | Thurston, Holly | 2017 | Child Abuse & Neglect | Quantitative; retrospective case-control design | Retrospective case-control study where cases were matched to controls at a 2:1 ratio based on the child's age at the time of the event and the duration between the event and the last CPS investigation | Binary conditional logistic regression, univariate conditional logistic regression, and principal component analysis with varimax rotation to reduce environmental variables into independent components | 702 subjects (234 cases and 468 controls) | Children under age six from families previously investigated by Child Protective Services | study reviewed over 6,000 substantiated investigations in a state child welfare database and cross-referencing Child Death Review Team records; controls were randomly matched from the same administrative database | metro county in California | 27.1 years (SD = 6.5) for cases and 28.5 years (SD = 6.2) for controls | NR | NR | All under 6 years | Cases: 56.4% Male; Controls: 48.3% Male | White Only (34%), Black Only (~30%), Hispanic Only (~15–18%), Asian Only (5–6%), and Other/Unknown (12%) | Proxied by Medi-Cal insurance eligibility, which applied to 81.6% of cases and 86.8% of controls | Yes | Self-disclosure or confirmation via CWS/CMS records, medical providers, law enforcement reports, or local criminal court records | A confirmed physical act of intimate partner violence where the caregiver was either the victim or the perpetrator | Physical acts of intimate partner violence | Yes | Administrative records from the Child Welfare Services/Case Management System and the Child Death Review Team | Serious maltreatment resulting in hospitalization or death, subdivided into abuse (commission) and neglect (omission) | Housing density (multi-family housing complex vs. single-family/duplex), temporary housing (living with others/shelters), and neighborhood vacant housing | Individual data from CWS/CMS and neighborhood-level data from the U.S. Census | "Multi-family housing complex" included apartments or shelters; "Family living with others" included temporary arrangements with friends, family, or in a shelter | Living in a multi-family housing complex significantly increased the odds of a child becoming a case of serious maltreatment compared to living in a house or duplex. High-density housing (multi-family complexes) was associated with doubled odds of being a case in the abuse model, but was not a significant predictor in the neglect model. | In the full multivariate model, living in a multi-family housing complex increased the odds of serious maltreatment by 1.50 times (95% CI: 1.01–2.22, p < .05). In the abuse-only subset, living in a multi-family housing complex was associated with a 2.78-fold increase in the odds of serious abuse (95% CI: 1.19–6.50, p < .01). In bivariate correlations, cases were significantly more likely to live in a multi-family housing complex than controls (55.1% vs. 46.6%, p < .05). | The temporary arrangement of "family living with others" was not significantly associated with serious maltreatment in the multivariate model (AOR = 1.29, 95% CI: 0.88–1.89). In the neglect-only multivariate model, living in a multi-family housing complex was not a statistically significant predictor of maltreatment (AOR = 1.00, 95% CI: 0.61–1.64). | No | NA | The family's residence at the time of the initial CPS investigation may not be the same as where the serious maltreatment event occurred; Medi-Cal status is an imperfect proxy for individual-level poverty; study lacked a measure for neighborhood social cohesion; potential for ascertainment bias, as some controls could be undetected maltreatment cases | |
92 | https://doi.org/10.1016/j.chiabu.2018.07.025 | Understanding trends in neighborhood child maltreatment rates: A three-wave panel study 1990-2010 | Coulton, Claudia J. | 2018 | Child Abuse & Neglect | Quantitative; longitudinal panel study | three-wave longitudinal panel study (1990, 2000, 2010) using census tracts as proxies for neighborhoods | Decade-specific OLS regression, Spatial Error Models (to adjust for spatial autocorrelation), Blinder-Oaxaca (B-O) decomposition of change, and neighborhood-level fixed effects panel models | 475 census tracts | Community-level data representing urban and suburban neighborhoods | administrative child protection records and U.S. Census/American Community Survey databases | Cleveland, Ohio | NR | NR | NR | NR | NR | NR | NR | No | NA | NA | NA | Yes | Administrative records from the Cuyahoga County Department of Children and Family Services | All allegations accepted for investigation | Vacant housing and owner-occupied housing | 1990 and 2000 U.S. Decennial Census and 2008–2012 American Community Survey (ACS) 5-year estimates | Percentages of vacant housing units and owner-occupied housing units within each census tract | Increases in vacant housing was strongly associated with increases in neighborhood child maltreatment report rates. | In the 1990 cross-sectional model, every percentage point increase in neighborhood vacant housing was associated with an increase in child maltreatment report rates (Estimate = 0.36, SE = 0.16, p < .05). In the 2000 cross-sectional model, the association between vacant housing and maltreatment rates became much stronger (Estimate = 2.70, SE = 0.22, p < .001). In the 2010 cross-sectional model, vacant housing remained a strong significant predictor of maltreatment report rates (Estimate = 1.37, SE = 0.14, p < .001). In the neighborhood fixed effects model, an increase in vacant housing within a neighborhood over time was significantly associated with an increase in child maltreatment report rates (Coefficient = 0.78, SE = 0.11, p < .001). In the same fixed effects model, a decrease in owner-occupied housing was significantly associated with an increase in neighborhood child maltreatment (Coefficient = -0.37, SE = 0.14, p < .01). The Blinder-Oaxaca decomposition identified that vacant housing was a primary contributor to maltreatment rate changes between 1990 and 2010 through both changes in its levels (Endowments Estimate = -8.97, SE = 1.19, p < .001) and changes in its predictive strength (Coefficients Estimate = -13.28, SE = 2.96, p < .001). | In the 1990 cross-sectional model, neighborhood owner-occupied housing was not significantly associated with child maltreatment report rates (Estimate = -0.06, SE = 0.04). In the 2000 cross-sectional model, owner-occupied housing showed no significant association with maltreatment rates (Estimate = 0.02, SE = 0.06). In the 2010 cross-sectional model, owner-occupied housing remained non-significant in its association with child maltreatment (Estimate = -0.01, SE = 0.06) | No | NA | potential ecological fallacy, so neighborhood-level associations cannot be directly applied to individual families or children; reliance on administrative agency records may reflect reporting or investigation bias based on neighborhood socio-economic status; census tracts are imperfect proxies for neighborhoods and may not capture the actual lived experiences of residents; study focused on a single metropolitan area, which may limit the generalizability of the findings to other cities | |
93 | https://doi.org/10.1371/journal.pone.0246682 | roles of housing, financial, and food insecurities in understanding the relationship between childhood neglect and violence in adulthood | Schuck, Amie M. | 2021 | PLoS ONE | Quantitative; prospective cohort design | A prospective cohort study where children with court-substantiated neglect (before age 12) and demographically matched controls were followed into adulthood with multiple waves of data collection (1989–1995, 2000–2002, and 2003–2005) and linked to official arrest records through 2012 | Probit structural equation modeling, t-tests, factor analysis | 713 individuals (366 neglected, 347 controls) | Neglected children (court-substantiated) and matched community controls | Cases were identified via 1967–1971 family court records; controls were matched using elementary school and hospital birth records | metro Midwest are of the US | Mean age at the 2003–2005 interview: 41.1 years (SD=3.5) | 51.5% female | 59% White (non-Hispanic), and 41% "All others" (including Black, Hispanic, American Indian, or Pacific Islander) | All children were under age 12 at the time of the neglect experience | matched to adult composition | matched to adult composition | matched o family social class | No | NA | NA | NA | Yes | Family court records from 1967 to 1971 | Neglect (extreme failure to provide adequate food, clothing, shelter, and medical attention) | Adult housing insecurity (inability to afford a place to stay and lack of a working telephone) | Self-report interview questions capturing experiences in the prior 12 months | Self-reporting that they could not afford a place to stay or did not have a working telephone for two weeks or more | Childhood neglect significantly predicted housing insecurity in middle adulthood. Individuals with neglect histories reported significantly higher housing insecurity than matched controls. | Childhood neglect directly predicted adult housing insecurity (Unstandardized Coefficient = 0.12, SE=0.04, p<.05). The indirect path from neglect to violent arrests through housing insecurity was significant (Coefficient = 0.04, SE=0.02, p<.05). Neglect history was significantly associated with higher mean housing insecurity (p=.005) compared to controls. | None. | No | NA | sample was skewed toward lower socioeconomic status, which may limit generalizability to higher-income neglected children; control group might contain "false negatives" (undocumented neglect cases), which would bias results toward the null; study relied on official arrest records, which likely underestimate the true extent of violence; measures for food, housing, and financial insecurities were limited in scope | |
94 | https://doi.org/10.1007/s10461-007-9236-5 | Physical and Sexual Abuse and Unstable Housing among Adolescents with HIV | Eastwood, Elizabeth A. | 2007 | AIDS and Behavior | quantitative, observational, and longitudinal study | Retrospective clinical data analysis measuring housing status at two points in time: prior history (self-report at intake) and current status at the time of program entry | chi-square, odds-ratio, t-tests, stepwise hierarchical logistic regression | 224 adolescents | HIV-infected adolescents | Sequential enrollment of patients at an adolescent medicine clinic in a public hospital between 1991 and 2004 | New York City, New York | NA | NA | NA | Mean age of entry into treatment was 18.0 years (SD=2.5); range: 12–24 | 62.5% Female and 37.5% Male | African American (57.6%), Caribbean Islander (25.0%), Latino (11.6%), White (4.0%), and Other (1.8%) | 68.3% relied on family for financial support and 12.1% received public assistance | No | NA | NA | NA | Yes | Standard clinical intake history (face-to-face interviews) asking for self-reported history of physical and sexual abuse | Physical abuse and sexual abuse | History of unstable housing and current unstable housing | Self-report clinical intake history | Living in a homeless shelter, jail, group home, foster care, or a friend’s home at the time of program entry (current) or in the recent past (prior) | Unstable housing was strongly associated with histories of both physical and sexual abuse. Adolescents with a history of unstable housing were significantly more likely to have experienced physical abuse (62.1% vs. 37.9%, p<.001) and sexual abuse (45.6% vs. 54.4%, p<.05) compared to those with stable histories. | Sexual abuse history was a significant predictor of prior unstable housing (AOR=2.84,95%CI:1.13–7.16,p<.05). In multivariate models, physical abuse (AOR=5.37,95%CI:1.14–25.29,p<.05) and sexual abuse (AOR=6.77,95%CI:1.24–36.89,p<.05) were significant predictors of current unstable housing. | None. | No | NA | Single-site study, so may not be generalizable to other urban cohorts; relied on self-reported data, which may lead to under-reporting of sensitive behaviors; housing was only measured at two points in time, failing to capture the fluidity of housing over a longer duration | the MSM (males who have sex with males) HIV transmission group was found to be protective against current unstable housing in the multivariate model (AOR=0.26,p<.05). |
95 | Multiply homeless families: The insidious impact of violence | Bassuk, Ellen L. | 2001 | Housing Policy Debate | Quantitative; longitudinal case-control study | longitudinal study following families over three points in time (baseline, 12 months, and 24 months) to assess retrospective and prospective homelessness | two-tailed t-tests, Wilcoxon rank-sum, chi-square test, Kaplan-Meier survival analysis, and multivariate logistic regression | 436 families (220 homeless and 216 never-homeless, low-income housed families) | Sheltered homeless mothers and low-income, never-homeless mothers | Homeless families were recruited from all nine family shelters and two welfare hotels in Worcester; never-homeless families were randomly selected from the public assistance caseload | Worcester, Massachusetts | mean age was 25.3 years for first-time homeless mothers and 28.9 years for multiply homeless mothers | 100% female | First-time Homeless: White (28.6%), Black (21.4%), Puerto Rican (41.7%), and Other (8.3%); Multiply Homeless: White (46.2%), Black (26.9%), Puerto Rican (21.2%), and Other (5.8%) | NA | NA | NA | over 92% of both homeless groups were receiving cash assistance at baseline | Yes | Conflict Tactics Scale | Coded positive if the respondent reported experiencing one or more of eight severe physically aggressive acts | Severe physical violence/aggression | Yes (mother's own childhood history) | Conflict Tactics Scale (CTS) for physical violence and self-report items regarding unwanted sexual advances/molestation | Childhood physical abuse and childhood sexual abuse | Literal homelessness, length of stay in shelter, and recidivism (multiple episodes) | Personal History Form and prospective tracking of all residential moves | "Homeless" was defined as sleeping in shelters or welfare hotels; "multiply homeless" was defined as having more than one homeless episode in adulthood | Childhood sexual abuse was identified as a strong independent predictor of repeated episodes of homelessness in adulthood. First-time homeless mothers who experienced recent partner violence after being rehoused were more than three times as likely to experience a second homeless episode. | Women who were sexually molested as children were significantly more likely to be multiply homeless at baseline (Adjusted OR = 2.83, 95% CI: 1.53–7.01, p=.002). Experiencing recent partner violence after being rehoused was a significant predictor of becoming homeless a second time (Adjusted OR = 3.57, 95% CI: 1.327–10.445, p=.015) | In the prospective follow-up model, childhood sexual abuse did not significantly predict a second episode of homelessness (Adjusted OR = 1.114, 95% CI: 0.424–3.149, p=.83). | No | NA | Single-city study may limit generalizability; time differences existed between when the first-time and multiply homeless episodes occurred; the study was unable to determine if those with baseline housing vouchers were actually able to find and use appropriate housing; the 24-month follow-up period may be too short to capture the long-term impact of violence on residential history . | ||
96 | https://doi.org/10.1037/sgd0000730 | Controlling Partner Dynamics in Transgender/Nonbinary Young Adults' Romantic Relationships: Exploring the Roles of Cissexism-Related Beliefs and Material-Need Insecurity | Murchison, Gabriel R. | 2024 | Psychology of Sexual Orientation and Gender Diversity | Quantitative; cross-sectional survey | cross-sectional online survey of romantically partnered U.S. transgender and/or nonbinary young adults | Multivariable generalized linear models, multivariable linear models, and natural effects mediation analysis | 393 participants | romantically partnered U.S. transgender and/or nonbinary (TNB) young adults | Online outreach to campus- and community-based LGBTQ organizations and social media platforms (Facebook, Instagram, Reddit) including paid ads | United States | Mean age: 24.1 years (SD=3.3); range: 18–30 years | Woman/Trans woman (28.3%), Man/Trans man (21.4%), and Nonbinary umbrella (50.3%) | White (42.6%), Black or African American (11.5%), Latine or Hispanic (11.2%), Asian or Asian American (7.7%), and Multiple or additional identities (27.0%) | NA | NA | NA | 34.6% reported food insecurity in the past 12 months; 11.5% reported housing instability in the past 6 months | Yes | 13-item modified version of the Relationship Control subscale of the Sexual Relationships Power Scale (SRPS) | Higher scores on the 1.0–2.9 scale (following score reversal), where higher numbers reflected more disproportionate control exerted by the participant's partner | Controlling behaviors (e.g., control over social activities, clothing, and joint decision-making) | No | NA | NA | housing instability | Single-item self-report survey question | Living in a hotel, boarding house, group home, on the street, or having no fixed address in the past 6 months | Housing instability was significantly and positively associated with controlling partner dynamics. | Housing instability was significantly associated with higher levels of controlling partner dynamics (b=0.67,95%CI:0.36–0.99,p<.001). | none. | No | NA | Cross-sectional design prevents causality or temporal order; potential selection bias because participants were recruited for a study on "relationships," potentially skewing toward those in more stable or satisfying partnerships; reliance on self-reported data may be subject to social desirability bias; the study did not directly assess structural factors like specific state laws or housing policies | |
97 | https://doi.org/10.1177/0886260511423241 | Housing Instability Is as Strong a Predictor of Poor Health Outcomes as Level of Danger in an Abusive Relationship: Findings from the SHARE Study | Rollins, Chiquita | 2012 | Journal of Interpersonal Violence | Quantitative; cross-sectional analysis of baseline data from a longitudinal study | quasi-experimental longitudinal cohort study; the current report utilizes baseline data from women newly enrolled in services | Multiple linear regression and logistic regression | 278 female IPV survivors | IPV survivors who sought services and identified housing stabilization as a primary need | Referrals from four community-based domestic violence and housing programs and self-referral via flyers | Metro Portland area, Oregon | NR | 100% female | White (53.69%), African American (26.62%), American Indian/Alaska Native (7.55%), Asian (1.08%), Native Hawaiian/Other Pacific Islander (1.08%), and Other race (20.86%). 25% identified as Latina | NA | NA | NA | 67.61% earned $1,000 or less per month; 29% were employed; 79.5% had health insurance (primarily Medicaid) | Yes | Danger Assessment (DA) | Self-report of physical and/or sexual violence by an intimate or ex-intimate partner in the previous six months | Physical and/or sexual violence and risk of lethality | No | NA | NA | Difficulty paying rent/utilities, eviction threats/notices, moving more than twice in 6 months, doubling up, and living in unwanted locations | 10-item Housing Instability Index developed for the study | Cumulative count of 10 risk factors in the past 6 months | Each additional housing instability risk factor increased the odds of absence specifically due to IPV by 32%. Housing instability was not a significant predictor of hospital or emergency medical use specifically related to IPV. | Each additional housing instability risk factor increased the odds of absence specifically due to IPV by 32% (OR=1.32,p=.001) | Housing instability was not a significant predictor of hospital or emergency medical use specifically related to IPV (OR=1.11,p=.097). | No | NA | Cross-sectional prevents evaluating cause and effect; sample is limited to survivors who proactively sought services, which may limit generalizability; general health was measured using a single item; the Housing Instability Index was newly developed for this study and requires further validation | |
98 | https://doi.org/10.1177/08862605221078814 | Intimate partner violence and depression of mothers and their adolescent children: The roles of housing insecurity and system contacts | Marçal, Katherine E. | 2022 | Journal of Interpersonal Violence | quantitative, longitudinal study | Data from the Fragile Families and Child Well-being Study. Mothers were interviewed at the child's birth and at 1-, 3-, 5-, 9-, and 15-year intervals. This study specifically utilized data from the Year 5, Year 9, and Year 15 waves | structural equation modeling, confirmatory factor analysis | 2,892 families | families in which the mother retained at least partial custody of the children at the Year 5 and 9 interviews | Recruited in hospitals after birth using stratified clustered sampling, with an intentional oversample of unmarried mothers | 20 large American cities | mean age of mothers at the time of the focal child’s birth: 25.22 years (SD = 6.04) | 100% female | 52.20% Black, 22.92% Hispanic, 21.74% White, and 3.14% other race | Children were approximately 5 years old at the Year 5 interview and 15 years old at the Year 15 interview | 51.33% male and 48.67% female | NR | average household income was $38,162 (SD = $45,905) | Yes | 11 self-reported items regarding how frequently a current or former partner engaged in various abusive behaviors | self-reported experiences of victimization | physical (e.g., hitting, slapping, pushing, throwing things), sexual (e.g., trying to force sex or unwanted sexual acts), and psychological/ emotional violence (e.g., insults, criticism, controlling friends/family, or withholding money) | Yes | Psychological Aggression and Physical Assault subscales of the Parent-Child Conflict Tactics Scale (CTSPC) and maternal self-report on whether they had been investigated by or had children removed by Child Protective Services | physical aggression (e.g., spanking, hitting with an object), psychological aggression (e.g., shouting, insulting), and child welfare investigation or child removal | housing insecurity (skipping rent/mortgage payments, skipping utility payments, eviction, doubling up with others for financial reasons) and homelessness (staying in a shelter, on the street, or in a vehicle) | Mothers self-reported on five dichotomous items regarding their experiences in the past year | Self-reported experience of any of the five indicators during the past year at the Year 5 interview | Mothers’ IPV victimization at Year 5 was directly associated with housing insecurity at Year 5. Furthermore, IPV was directly associated with child welfare contact at Year 9. Housing insecurity at Year 5 was also directly associated with child welfare contact at Year 9. Significant indirect pathways were found from IPV to maternal depression via housing insecurity and subsequent system contacts (both child welfare and criminal justice). | Maternal IPV victimization at Year 5 was directly and positively associated with housing insecurity at Year 5 (standardized coefficient β = 0.18, p < 0.001). Housing insecurity at Year 5 was also directly associated with child welfare contact at Year 9 (β = 0.28, p < 0.001). Additionally, a significant indirect pathway was identified where IPV was linked to increased housing insecurity, which then related directly to increased adolescent depression at Year 15 (β = 0.022, 95% CI [0.005, 0.043], p < 0.05). Another significant indirect pathway showed that IPV was associated with increased housing insecurity, which then related to increased child welfare contact, which finally was associated with increased maternal depression (β = 0.011, 95% CI [0.003, 0.022], p < 0.05) | The indirect pathway from IPV to maternal depression via housing insecurity alone (without the additional mediation of system contacts) was not statistically significant (β = 0.021, 95% CI [−0.004, 0.048]) | No | NA | self-reported data, which may be subject to social desirability bias; oversampling of unmarried parents limits generalizability to higher-income populations; dichotomous indicator for maternal depression was used rather than a continuous measure of severity; the applicability of the CES-D scale for adolescent depression has been questioned for nonwhite samples; IPV was only assessed for mothers who were married or partnered during the study period; measure of housing insecurity might not be comprehensive for youth who spend significant time in other living arrangements | |
99 | https://doi.org/10.1177/08862605221077231 | Racial and gender inequalities in food, housing, and healthcare insecurity associated with intimate partner and sexual violence | Fedina, Lisa | 2022 | Journal of Interpersonal Violence | Quantitative; cross-sectional study | cross-sectional survey design using data from the 2010 National Intimate Partner and Sexual Violence Survey | chi-square, t-tests, and binary logistic regressions | 18,049 adults | nationally representative sample of the U.S. general population | dual-frame (cell-phone and landline), stratified, random digit dial sampling strategy | United States (National) | mean for men: 48.00 years (SD = 17.11) mean for women: 50.00 years (SD = 17.66). Range: 18 to 100 years | 44.7% male, 55.3% female | Men: 80.0% White, 8.6% Black, 8.7% Latino/a, 3.2% Asian/PI, 1.9% AIAN, 3.2% Other, and 3.2% Multiracial; Women: 81.8% White, 8.9% Black, 7.8% Latino/a, 2.1% Asian/PI, 1.5% AIAN, 2.2% Other, and 3.3% Multiracial | NA | NA | NA | 34.0% of men and 27.2% of women reported $75,000 or more, and 5.3% of men and 6.3% of women reported less than $10,000 | Yes | self-report items from the 2010 NISVS regarding victimization in the past 12 months | self-reported past-year exposure to any form of intimate partner violence, sexual violence, or stalking | physical violence (e.g., hit, punched), psychological violence (e.g., called names), control and entrapment (e.g., keeping from friends/family), contact sexual violence (e.g., rape, sexual coercion), and stalking | No | NA | NA | Perceived housing insecurity (related to the stress or worry of having enough money to pay rent or mortgage) | single ordinal 5-point Likert-type indicator: “In the past 12 months, how often would you say you were worried or stressed about having enough money to pay your rent or mortgage?” | Responses of "sometimes," "usually," or "always" were coded as "yes" for housing insecurity | Both men and women with histories of IPV and SV victimization had approximately a two-fold increase in the odds of housing insecurity compared to those without exposure to IPV and SV, even after controlling for income and education. The highest levels of housing insecurity were found among men and women of color exposed to violence, particularly those identifying as Black/African American, Latino/a, and American Indian/Alaska Native. | For women, exposure to IPV and SV was significantly associated with higher odds of housing insecurity (OR = 1.94, 95% CI [1.59, 2.36], p < .001). For men, exposure to IPV and SV was similarly associated with higher odds of housing insecurity (OR = 2.06, 95% CI [1.69, 2.51], p < .001). Violence exposure was significantly associated with housing insecurity for both genders (p < .05) | In the multivariable models, there were no statistically significant differences in housing insecurity odds for Black/African American women (OR = 1.25, 95% CI [0.96, 1.62]) or Latina women (OR = 1.11, 95% CI [0.76, 1.58]) compared to White women when controlling for violence exposure. Latino men did not have significantly different odds of housing insecurity compared to White men when controlling for violence exposure (OR = 1.36, 95% CI [0.97, 1.91]). | No | NA | cross-sectional design, so unable to determine causal relationships; phone-based data collection might underestimate inequalities by excluding those without stable phone access; measures for economic insecurity were perceived and self-reported single items, which may not capture the full nature of the insecurity; the sample was limited to cisgender individuals, and the "Other" race categories were too broad to discern specific ethnic needs; data were collected in 2010, though the authors argue the findings remain highly relevant | |
100 | https://doi.org/10.1080/19419899.2019.1690032 | Associations of physical and sexual violence victimisation, homelessness, and perceptions of safety with suicidality in a community sample of transgender individuals | Drescher, Christopher F. | 2021 | Psychology & Sexuality | quantitative, cross-sectional study | electronic survey design to conduct an LGBTQ health needs assessment | Pearson and point-biserial regressions, hierarchical logistic regression | 70 participants | community sample of self-identifying transgender and gender non-conforming individuals | LGBTQ community events, minority-friendly venues, printed advertisements, and social media | Central Savannah River Area, which spans multiple counties in Georgia and South Carolina | Mean: 29.97 years (SD = 10.56); range: 18 - 65 years | 34.3% trans men, 25.7% trans women, and 40% individuals with a non-binary identity | 77% White and 23% non-White | NA | NA | NA | 59% of the sample had household incomes of less than $30,000 per year, and 51.5% reported being 'occasionally' or 'routinely' unable to purchase sufficient food or other basic necessities | Yes | Single-item self-report questions were used: “Have you ever been hit, kicked, slapped, or hurt by a partner?” and “Has a partner ever forced you into sexual activities against your will?” | Self-report endorsement | Physical violence and sexual violence | No | NA | NA | Homelessness (current and lifetime) | single self-report item: “Are you currently homeless or have you ever been homeless or not had a permanent place to sleep at night?” | self-report endorsement | Partner physical violence history was significantly and positively associated with a history of homelessness. Additionally, sexual violence history was significantly associated with a history of homelessness. | At the bivariate level, partner physical violence history was significantly and positively associated with a history of homelessness (r = .26, p < .05) . Additionally, sexual violence history was significantly associated with a history of homelessness (r = .55, p < .01). | None. | No | NA | self-report data, which is subject to social desirability and stigma; small sample size and lack of racial diversity (predominantly White) limit generalizability; cross-sectional design prevents determining causality; single-item measures may limit reliability | |