Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Oct 1.
Published in final edited form as: Early Child Res Q. 2023 Aug 12;65:407–416. doi: 10.1016/j.ecresq.2023.08.001

Profiles and Predictors of Neurodevelopmental Functioning among Young Children Experiencing Family Homelessness

Janette E Herbers 1, Carmela J DeCandia 2, Katherine T Volk 3, George J Unick 4
PMCID: PMC10449385  NIHMSID: NIHMS1924429  PMID: 37635734

Abstract

In the context of family homelessness, children experience acute adversities related to loss of housing and residential mobility compounded with more chronic, poverty-related adversities and stressors. Among children in families experiencing homelessness, variability in experiences and outcomes warrant person-centered approaches to better delineate patterns of risk and resilience. Using latent profile analysis as a person-centered approach, we identified five distinct profiles of neurodevelopmental functioning within a sample of 231 children (ages 3–5 years old) staying in emergency homeless shelters with their families. Latent profiles were informed by indicators from parent-reported items for ten different domains of neurodevelopmental functioning. We examined whether demographic and ecological factors including age, ethnicity, adverse childhood experiences, parent mental health, and overreactive parenting would predict profile membership. Overall, half of the children in the sample demonstrated a profile of resilient functioning across developmental domains. Profiles of maladaptive functioning differed in areas of strength and challenge, with a small percentage of children showing poor functioning across all domains. Children whose parents had more mental health problems or overreactive parenting were significantly more likely to show profiles of poor functioning than to show resilient functioning. Implications for future research, practice, and policy are discussed.

Keywords: homelessness, adversity, resilience, person-centered methods, neurodevelopment

Children who experience family homelessness have more problems with social, behavioral, and emotional functioning as well as worse school engagement and academic achievement on average compared to their non-homeless peers (Clark et al., 2019; Haskett & Armstrong, 2019; Herbers et al., 2020). Best estimates from multisite efforts to screen for housing instability and homelessness at pediatric emergency department visits and hospital intake suggest that homelessness may affect roughly 12% of U.S. children (Sandel et al., 2018). In the context of family homelessness, children experience acute adversities related to loss of housing and residential mobility compounded with more chronic, poverty-related adversities and stressors (Cutuli & Herbers, 2014). Together, these acute and chronic adversities can threaten healthy child development in numerous ways.

By definition, children experiencing homelessness “lack a fixed, regular, and adequate place to sleep at night” (NCHE, 2021). In the United States, the primary reasons for family homelessness are poverty, lack of affordable housing, and residential instability (Gubits et al., 2018; Marçal et al., 2021). When families experience homelessness, they must negotiate acute risks related to loss of housing alongside more chronic risks of poverty, like underemployment and very low income. Children experiencing homelessness have more barriers to health care access and significantly more food insecurity compared to children who are stably housed (Volk et al., 2023). Acute risks of homeless episodes can include intimate partner violence, loss of employment, eviction, exposure to natural disaster, and family separations. Children meeting the definition of homelessness can vary considerably in which of these acute and chronic adversities they experience and how their parents or other primary caregivers are functioning to support them.

Caregivers of children who experience family homelessness are predominantly young, single mothers with limited educational backgrounds (Bassuk et al., 2020; Krahn et al., 2018). Marginalized ethnic minority groups are overrepresented, particularly African American and Latin American individuals. Compared to stably housed caregivers, caregivers of families experiencing homelessness have higher rates of depression and other mental health problems (Bassuk et al., 2015; Corman et al., 2016). For some children, their caregivers identify less social support and more relationship conflict (Lucke et al., 2021; Marra et al., 2009). Longer stays in shelter are associated with more caregiver stress and mental health concerns for both caregivers and children (Marçal et al., 2021). Among women with similar homeless episode durations, mothers show higher rates of depression and posttraumatic stress than non-mothers (Krahn et al., 2018). Mothers in shelters often describe feeling disempowerment in the parenting role, which challenges self-efficacy and increases stress (Bradley et al., 2018; Vrabic et al., 2022). The numerous stressors caregivers experience likely contribute to poor parenting practices, with higher rates of ineffective discipline and physical and psychological aggression observed among families experiencing homelessness compared to stably housed families (Holtrop et al., 2017; Park et al., 2015).

The risks associated with homelessness appear early, with elevated rates of health and developmental delays among infants, toddlers, and preschool-aged children that predict difficulties with later functioning (Fanning, 2021; Haskett et al., 2016; Herbers et al., 2020). Beginning prenatally, children’s development is shaped by experience and environmental contexts such that adverse conditions of poverty, chronic stress, and psychosocial adversity can interfere with the formation and refinement of fundamental neural, physiological, and behavioral systems (McEwen, 2017; Shonkoff & Garner, 2012). In a meta-analysis of children’s mental health, it was estimated that up to 20 percent of preschoolers experiencing family homelessness meet criteria for a mental health condition (Bassuk et al., 2015), and up to 25% demonstrate a developmental delay by age five (Haskett et al., 2016). Other studies have demonstrated elevated rates of parent- and teacher-reported behavior problems, poor social skills, and social-emotional difficulties (Brumley et al 2015; Haskett et al., 2016; Rybski & Israel, 2019). Children who experience family homelessness are more likely to struggle with school readiness skills, early reading, and later achievement in both math and reading across their schooling (Herbers et al., 2012; Manfra et al., 2019; Masten et al., 2015).

Despite these risks, many children who experience family homelessness show resilience, succeeding in school and in terms of social and emotional wellbeing (Herbers et al., 2020; Masten et al., 2015). Resilience refers to the capacity of a dynamic system, such as an individual child or a family, to adapt effectively to disturbances that threaten the viability, function, or development of that system (Masten & Palmer, 2019). With such variability in experiences related to and co-occurring with homelessness, it is not surprising that outcomes associated with homelessness also vary. Problems tend to aggregate such that problems in one domain of functioning like academic achievement are likely to be accompanied by problems in other domains, like social-emotional development (Herbers et al., 2020; Obradović et al, 2010; DeCandia et al., 2022). Knowing which children are affected in which ways demands a more person-centered approach to best target services to their specific family contexts and developmental needs (Huntington et al., 2008).

Like studies from the broader field of risk and resilience, most findings from studies with children who experience homelessness are based upon variable-centered approaches (Bassuk et al., 2015; Evans et al., 2013; Herbers et al., 2019; Jobe-Shields et al., 2015). In variable-centered approaches, researchers focus on quantitative differences in individual variables as the outcomes of interest. These approaches assume that statistical relations among predictors and outcomes are constant for all children within the sample. This assumption helps to test unique effects of certain characteristics on certain problems, but it can obscure meaningful differences in the functioning of the whole persons involved. In contrast, person-centered approaches consider profiles of a range of relevant variables, aiming to identify meaningful subgroups with common patterns of these variables within broader populations (Laursen & Hoff, 2006; von Eye et al., 2015).

When seeking to better understand how different children may adapt in contexts of family homelessness, a person-centered approach to functioning can reveal differences among individual children and families that would not be apparent by focusing on specific neurodevelopmental outcome variables such as gross motor development or language skills in isolation (Bassuk et al., 2015; Haskett et al., 2021; Herbers et al., 2019). For example, when examining social-emotional learning (SEL) among preschool children, Denham and colleagues (2012) identified different groups of children ages 3–4 years in both private and publicly funded (Head Start) daycare settings who clustered together in the ability to learn and participation in the classroom. These groups demonstrated meaningful differences such that the children who fell in the “SEL-Risk” group (42.9%) demonstrated lower self-regulation skills and emotion knowledge than those in other groups, with higher angry response patterns. This person-centered approach allowed for the identification of the specific needs of a select group of children and targeted SEL interventions for them specifically.

There are only a few examples of person-centered approaches to assess functioning across domains for children experiencing homelessness (see Haskett et al., 2021; Herbers et al., 2020; Huntington et al., 2008; Obradović, 2010). Among children experiencing housing instability in Head Start, which are publicly funded early learning programs for low-income children ages 3–5, teacher-child relationship quality predicted profiles of high functioning in social skills, hyperactivity, and both internalizing and externalizing behavior (Haskett et al., 2021). Obradović (2010) examined how effortful control, a construct related to emotional and cognitive self-regulation, predicted likelihood of resilience across academic, social, and behavioral domains of functioning in kindergarten and first grade for children recruited from family shelters. Herbers et al. (2020) used prospective, longitudinal data to examine mental health profiles at age 15 to show that youth with a history of homelessness during childhood were less likely than stably housed youth to show a profile of resilience across self- and parent-reported internalizing, externalizing, and substance use problems. In each of these studies, one subgroup of children experiencing homelessness could be differentiated by resilient functioning across domains while other subgroups showed variability in domains of strength and weakness. More person-centered work has been recommended to better delineate and predict these patterns of individual differences in child functioning in the context of family homelessness (Bassuk et al., 2015; Huntington et al., 2008).

The Current Study

In the current study, we employed latent profile analysis as a person-centered approach to address research questions about different profiles of neurodevelopmental functioning among young children (ages 3–5) staying in emergency homeless shelters with their families. First, we were interested in understanding whether and how many distinct subgroups, or unique profiles, of neurodevelopmental functioning would be apparent based on indicators from ten areas of neurodevelopmental functioning from the Neurodevelopmental Ecological Screening Tool (NEST), a measure designed and validated for use in shelters and other high-risk settings (DeCandia et al., 2021). We used a model comparison approach to determine whether the sample was best characterized by two profiles – generally high scores versus generally low scores – or whether the sample would reveal more distinctions in latent patterns of neurodevelopmental functioning evidenced by more than two unique subgroups. Our second research question considered whether demographic and ecological factors from NEST would predict profile membership. We expected to find at least three distinct profiles of neurodevelopmental functioning, with one profile representing resilient functioning across all domains, one representing problems across domains, and at least one profile showing a mix of strengths and weaknesses in neurodevelopmental domains. When predicting profile membership from demographic and ecological factors, we expected to find that membership in the resilient profile would be more likely for children whose parents reported fewer mental health problems, less overreactive parenting, and fewer adverse childhood experiences.

Method

Participants and Procedures

Participants were 231 children aged 3–5 years and their parents who were staying in family homeless shelters either currently or within the past 30 days. Participants were recruited from nine organizations across the United States, from seven states representing the Northeast, Southeast, Midwest, and Western regions of the country. To be eligible, families had to be currently homeless or housed after homelessness within the last 30 days, have basic proficiency in spoken English, and have a child ages 3–5 years. Sample characteristics are reported in Table 1 and were consistent with typical rates among families staying in homeless shelters in the United States. Most (78%) reported their ethnicities as nonwhite (43% Black/African American, 31% Hispanic/Latinx, 4% Native American, Hawaiian, or self-identified as “other”), and 93% of caregivers being the children’s biological mothers. Caregivers ranged in age from 18 to 60 years (M = 30.65, SD = 6.86). Children in the sample were 48.4% female with an average age of 4.37 years (SD = 0.86).

Table 1.

Family demographics.

Caregiver Child
Race/Ethnicity % %
 African American 50 43
 White 21 16
 Hispanic/Latinx 20 31
 Native American/Hawaiian/Other 8 8
Caregiver relationship to child
 Mother 93
 Father 5
 Other 2
Age (years)
 M (SD) 30.65 (6.86) 4.37 (0.86)
% %
 3 years 41%
 4 years 31%
 5 years 28%
 17–25 years 22%
 26–39 years 68%
 40–60 years 10%
Open in a new tab

NEST is a web-based screening tool consisting of 105 multiple-choice questions across three domains (neurodevelopmental, caregiver, and environment). It is administered via computer by paraprofessionals (e.g., case managers), sitting side-by-side with caregivers. For this study, case managers’ training included a description of NEST and why it is important to the field, how to administer NEST (including sample scripts and practice scenarios), and a discussion of logistics surrounding data collection. After completing informed consent with a trained member of the study staff, the shelter case managers guided the child’s primary caregiver through NEST’s multiple-choice questions. This process took place in a private space at each program’s facilities and lasted approximately 30 minutes. Children did not answer any questions directly; all data were based on caregiver report in response to NEST. The research described in this paper was approved by the Heartland Institutional Review Board (IRB), including all forms and procedures.

Measures

All study measures were drawn from NEST, which was developed as an ecologically based, parent-report, web-based screener to identify a child’s level of neurodevelopmental functioning and psychosocial risk (DeCandia et al., 2020; 2021; 2022). NEST screens child risk and protective factors across three domains: neurodevelopment, caregiver functioning, and environment. Results of the validation work indicated that NEST can identify a young child’s level of developmental risk with 100% sensitivity and 65% specificity as compared to established measures in the field (DeCandia et al., 2021).

Neurodevelopmental Functioning.

NEST assesses five areas of neurodevelopmental functioning: Motor Skills, Functional Communication, Adaptability and Coping, Executive Functioning, and Social-Emotional skills. There are a total of 10 indicators used to assess the five areas: gross motor (large muscle activities like walking, running, and climbing, 5 items), fine motor (finely tuned movements like grasping objects, 5 items), communication (conveying thoughts and feelings with words to get needs met, 7 items), coping with everyday challenges (effectively managing common disappointments, frustrations, and new situations, 3 items), attention (encoding information and sustaining focus, 3 items), problem solving (finding ways to achieve simple goals like finding a desired object, 5 items), cognitive processing (comprehending information, 3 items), social skills (skills such as demonstrating concern for others and managing conflicts with peers, 4 items), imaginative play (engaging in symbolic or pretend play, 4 items), and emotion regulation (managing emotional reactions and calming down independently or with help from others, 7 items). Internal reliabilities for the ten individual scales ranged from α = .69–.83. Raw scores from items for each indicator were converted to 3-point risk scores indicating level of risk for the domain with 0 = low risk, 1 = medium risk, 2 = high risk. Detailed information about the scoring and risk scale development is available elsewhere (DeCandia et al., 2021). We utilized the risk scores for the ten individual scales as indicators for a latent profile analysis to inform identification of latent profiles of neurodevelopmental functioning.

Caregiver Mental Health.

Mental health of the child’s primary caregiver was assessed with two brief measures embedded in NEST: the Patient Health Questionnaire (PHQ-9: Kroenke et al., 2001) and the Abbreviated PTSD Checklist-Civilian version (PCL-C: Lang et al., 2012). The PHQ-9 asks how much an individual has been bothered by depression symptoms over the last two weeks, such as “having little interest or pleasure in doing things,” “trouble falling or staying asleep or sleeping too much,” and “trouble concentrating on things.” Internal reliability for the sample was α = .89. The Abbreviated PCL-C includes six items assessing symptoms of posttraumatic stress by asking how often in the past month the participant has been bothered by, for example, “repeated, disturbing memories, thoughts, or images of stressful experience from the past” or “avoiding activities or situations because they reminded you of a stressful experience from the past.” Internal reliability for the sample was α = .94. Responses to interval scales for items were averaged within measures. Total scores for each measure (r = .72) were then converted to z-scorers and combined to form a single composite of caregiver mental health, with higher scores indicating more mental health problems.

Overreactive Parenting.

For the ecological assessment of parenting quality, NEST includes an abbreviated 5-item screener of overreactive parenting adapted from the Arnold Parenting Scale (Arnold et al., 1993). The 5-item screener previously had been validated with mothers from low-income, racially diverse neighborhoods (Reitman et al., 2001). The measure asks individuals to rate their parenting on a continuum by filling in dots corresponding to a 7-point scale. Sample items include: “In the past two months when my child misbehaves, I raise my voice or yell (7) to “I speak to my child calmly” (1), and “When my child misbehaves, I usually get into a long argument with my child” (7) to “I don’t get into an argument” (1). Internal reliability within the sample was strong, α = .75.

Adverse Childhood Experiences.

Also embedded in NEST as part of the environmental domain is a measure of adverse childhood experiences of the child based on those used by Blodget (2012). Caregivers responded to a list of nine different adverse experiences such as “witnessed domestic violence” and “death of a parent or primary caregiver,” indicating whether their child had ever experienced each item in their life. Sums of endorsed items were recoded to the following: 0 = no ACEs, 1 = 1 or 2 ACEs, 2 = 3–5 ACEs, and 3 = 6 or more ACEs reported in the child’s lifetime. The mean ACE for our sample was 2.57, and 45% had scores of 2 or 3, indicating three or more total ACEs.

Plan for Analysis

We utilized latent profile analysis to examine whether there were meaningful subgroups, or latent profiles, of children within the sample based on their scores on the ten items comprising the neurodevelopmental domain of NEST. Missing data were accounted for using full information maximum likelihood (Collins & Lanza, 2010). Using MPlus version 8.3 (Muthen & Muthen, 1998), we included ten indicators of gross motor, fine motor, communication, everyday challenges, attention, problem solving, cognitive processing, social skills, play, and emotional regulation to inform the models with increasing numbers of latent profiles. We then compared the different model solutions using Akaike’s information criteria (AIC; Akaike, 1987), a sample-size adjusted Bayes Information Criterion (aBIC), the bootstrapped likelihood ratio test (BLRT; McLachlan & Peel, 2000), and the interpretability of the solution based on theory (Collins & Lanza, 2010; Masyn, 2013). Lower values of AIC and BIC indicate a better balance of model fit and parsimony while the BLRT compares each model solution to the model solution with one fewer profile. We also examined posterior probabilities of assignment and overall entropy, which indicate the degree of differentiation among profiles for each solution.

Once we identified the model solution with the best balance of model fit, parsimony, and interpretability, we used Asparouhov and Muthen’s (2014) three-step approach for auxiliary variables in mixture modeling to test whether other study variables of demographics, adversity, and caregiver functioning predicted membership in the different latent profiles. The three-step approach with Mplus software involves assigning participants to their most likely latent profile based on the posterior probabilities of the accepted model, accounting for error arising from classification uncertainty. The covariates are then used to predict the categorical variable of profile membership in a multinomial logistic regression model. We examined results of the multinomial regression model with each profile as a referent to understand how profiles might be differentiated from each other using simultaneous predictors of child gender, child age in months, African American/Black ethnicity, Hispanic/Latinx ethnicity, ACEs, caregiver mental health risk, and overreactive parenting risk.

Results

Profiles of Neurodevelopmental Risk

Based on all the information in our model comparison approach, we considered the solution with five latent profiles to have the best balance of fit, parsimony, and interpretability (see Table 2). The 5-profile solution had the lowest values for AIC and sample-adjusted BIC and a significant result for the BLRT. The posterior probabilities of assignment ranged from .95 to .99, and overall entropy of .97 indicated excellent class separation.

Table 2.

Model comparison statistics with accepted model highlighted in gray.

# profiles # parameters estimated AIC aBIC Entropy BLRT BLRT p
1 20 4788.061 4791.913 - - -
2 31 4254.417 4260.387 0.941 −2374.03 <.0001
3 42 4086.802 4094.891 0.969 −2096.21 <.0001
4 53 3994.012 4004.219 0.955 −2001.4 <.0001
5 64 3915.881 3928.206 0.972 −1941.19 <.0001
6 75 3957.881 3974.251 0.975 87.181 <.0001
Open in a new tab

Means and standard errors of the ten neurodevelopmental domain indicators for each latent profile are displayed in Figure 1. The largest profile (A) included 56.6% of the sample, with average scores below 0.5 across all ten indicators of neurodevelopmental risk. We gave this profile the descriptive label of “resilient functioning.” Representing 18.4% of the sample, Profile B had scores below 0.5 only in domains of challenges, attention, social skills, and emotion management and a score above 1.0 on cognitive processing. We labeled this profile “cognitive challenges only.” The third largest profile (C) represented 12.7% of the sample and had scores above 1.0 on domains of gross motor, fine motor, communication, challenges, attention, cognitive processing, and play with no domain scores below 0.5. It had the highest scores compared to all other profiles on domains of gross motor, fine motor, communication, cognitive processing, and play. We labeled this profile “multi-domain with motor challenges.” The fourth profile (D), represented 8.1% of the sample, had scores above 1 on domains of fine motor, everyday challenges, social skills, and emotion management and a score below 0.5 only on cognitive processing. We labeled this profile “social-emotional challenges.” The final profile (E) represented only 4.2% of the sample, with scores above 1.0 on challenges, attention, problem-solving, cognitive processing, social skills, play, and emotion management. Profile E was labeled “multi-domain with social challenges,” with no domain scores below 0.5 and higher scores than all other profiles on domains of attention, problem-solving, cognitive processing, social skills, and emotion management.

Figure 1.

Figure 1

Open in a new tab

Average scores for neurodevelopmental risk domains by latent profiles

Predictors of Profile Membership

A summary of the five profiles with their labels is presented in Table 3. Mean differences in scores for study covariates by latent neurodevelopmental profile are displayed in Figure 2. Next, we describe results of the multinomial regression model assessing which of the covariates would significantly predict membership in the different profiles when accounting for all other covariates (see Table 4).

Table 3.

Summary of five latent profiles of neurodevelopmental risk.

Profile Domains at Risk (M > 1.00) Domains of Resilience (M < 0.50, *M < 0.75) Predictive Factors
A (56.6%): Resilient Functioning None All Reference group
B (18.4%): Cognitive challenges only Cognitive Processing Gross motor, Everyday challenges, Attention, Problem solving, Social skills, Emotion regulation Younger age, Caregiver mental health
C (12.7%): Multi-domain with motor challenges Gross motor, Fine motor, Communication, Everyday challenges, Attention, Problem-solving, Cognitive processing, Play *Social skills Male gender, Hispanic, Caregiver mental health, Overreactive parenting
D (8.1%): Social-emotional challenges Fine motor, Everyday challenges, Social skills, Emotion regulation *Problem solving, Cognitive processing Male gender, African American, Hispanic, Caregiver mental health
E (4.3%): Multi-domain with social challenges Fine motor, Communication, Everyday challenges, Attention, Problem-solving, Cognitive processing, Play *Gross motor Male gender, Caregiver mental health, Overreactive parenting
Open in a new tab

Figure 2.

Figure 2

Open in a new tab

Average z-scores (top) and percentages (bottom) for descriptive statistics and covariates by latent profile. Error bars for z-scores represent standard errors of the mean. Dotted lines for percentages represent percentages for the overall sample.

Table 4.

Odds ratios for Profile A (Resilient) compared to the other four profiles

A vs. B A vs. C A vs. D A vs. E
Child age 2.13* 1.91 2.15 1.95
Child gender (female) 1.56 4.05* 6.88** 16.3**
African American 1.42 1.55 0.21** 0.51
Hispanic/Latinx 2.33 0.37** 0.09** 0.81
Child ACEs 1.34 0.88 0.76 1.28
Caregiver Mental Health 0.63* 0.54** 0.35** 0.30**
Overreactive Parenting 0.90 0.55** 0.86 0.51*
Open in a new tab
**

p < .01,

*

p < .05

With the largest latent profile (Profile A, resilient functioning) as a reference group, results of the multinomial regression model indicated that older child age significantly predicted lower likelihood of being classified in Profile B (OR = 0.47, p < .001), Profile C (OR = 0.47, p = .001), and profile D (OR = 0.52, p = .003) compared to profile A, and that being a girl significantly predicted lower likelihood of being in Profile C (OR = 0.15, p < .001), Profile D (OR = 0.25, p < .001), or Profile E (OR = 0.25, p < .001) compared to A. Hispanic/Latinx ethnicity significantly predicted greater likelihood of being classified in Profile C than Profile A (OR = 10.63, p = .038). Higher scores for caregiver mental health risk significantly predicted greater likelihood of being classified in Profile C (OR = 2.84, p = .006), Profile D (OR = 1.86, p = .010), and Profile E (OR = 3.34, p = .005) compared to Profile A. Higher scores for overreactive parenting risk significantly predicted greater likelihood of being classified in Profile D (OR = 1.84, p = .033). There were two effects at the nonsignificant trend level, with higher scores for caregiver mental health risk predicting greater likelihood of being in Profile B (OR = 1.59, p = .067), and overreactive parenting predicting greater likelihood of being in Profile E (OR = 1.97, p = .091) compared to Profile A.

With Profile B as the reference group, results of the model indicated that being a girl significantly predicted lower likelihood of being classified in Profile C (OR = 0.23, p < .001), Profile D (OR = 0.38, p = .009), and Profile E (OR = 0.09, p < .001) compared to Profile B. Caregiver mental health risk significantly predicted lower likelihood of being in Profile A (OR = 0.63, p = .020).

With Profile C as the reference group, Hispanic/Latinx ethnicity predicted lower likelihood of being classified in Profile B (OR = .04, p < .001), Profile D (OR = 0.26, p = .011), or Profile E (OR = 0.12, p < .001). Black/African American ethnicity predicted lower likelihood of being classified Profile B (OR = 0.15, p < .001) and Profile D (OR = 0.14, p < .001). Adverse childhood experiences predicted lower likelihood of being classified in Profile B compared to Profile C (OR = 0.57, p < .037). Caregiver mental health risk significantly predicted lower likelihood of being in Profile A (OR = 0.35, p < .001) and at trend level for Profile B (OR = 0.65, p = .059) compared to Profile C.

With Profile D as the reference group, Hispanic/Latinx ethnicity predicted lower likelihood of being classified in Profile A (OR = 0.37, p = .005) or Profile B (OR = 0.16, p < .001). Caregiver mental health risk also predicted a lower likelihood of being in Profile A (OR = 0.35, p < .001) and at trend level for Profile B (OR = 0.56, p = .05) compared to Profile D.

Finally, with Profile E as the reference group, caregiver mental health risk predicted a lower likelihood of being classified in Profile A (OR = 0.30, p < .001) and Profile B (OR = 0.48, p = .019). Overreactive parenting risk also predicted a lower likelihood of being classified in Profile A (OR = 0.51, p = .016) and at trend level for Profile B (OR = 0.57, p = .084) compared to Profile E.

Discussion

Children who stay with their families in emergency homeless shelters face numerous threats to their healthy development. Although these children all have experiences of poverty and unstable housing in common, they differ in their other experiences of adversity and their access to strengths and resilience factors that may mitigate impacts of risk. Investigating individual differences and patterns within the heterogeneity of this group has potential not only to better understand processes of risk and resilience, but also to inform intervention efforts that seek to reduce risk and promote positive functioning among children experiencing family homelessness. In this study, we identified five unique profiles across domains of neurodevelopmental functioning in children ages 3–5 years experiencing homelessness. We considered how family and ecological factors predicted the likelihood of children demonstrating profiles of resilient or maladaptive functioning.

Results of our latent profile analysis indicated five distinct profiles at varying levels of neurodevelopmental risk (Table 3). Over half of the children in the sample showed a profile of resilient neurodevelopmental functioning, with scores below the at-risk range across domains of motor, communication, attention, problem solving, cognitive processing, social skills, and emotion regulation. An additional 18% of children, who were on average younger than those in the resilient functioning profile, showed a profile with only one at-risk score (in cognitive processing). The difference in these groups may be attributable to age more than other factors, as NEST has been shown to have less reliability for the 3 to 3.5 year old group (DeCandia et al., 2021). In addition to their younger age, there may be other confounding factors not assessed such as schooling or medical issues. For this group of children, assessment by a early childhood educator or child development expert could be beneficial to ascertain if the remedy is maturation and time, or if early intervention is warranted. Together, these two profiles of generally good functioning accounted for a strong majority of the children in the sample, highlighting how resilience is common among children at risk due to homelessness. As expected, profiles of resilience were significantly more likely for children with fewer adverse childhood experiences, children whose caregivers reported fewer symptoms of mental health problems, and children whose parents reported less overreactive parenting. Girls were also more likely to show profiles of resilience as compared to boys in this age group.

The evidence of resilience among many young children experiencing homelessness does not negate the potent threats of deep poverty and housing instability (Bassuk et al., 2020; Haskett & Armstrong, 2019). Researchers, providers, and policy makers should not be tempted to conclude from this work that family homelessness is not a serious problem, but rather should recognize how thoughtful prevention and intervention efforts can leverage resilience factors to support all families who face these challenging circumstances. Resilience in many families highlights how caregiver health and positive parenting can protect many children. This means that efforts should strongly emphasize supporting the wellbeing of caregivers as the linchpins of young children’s resilience. This can take many forms, including provision of physical and financial resources as well as trauma-informed approaches that empower caregivers with voice and choice rather than presuming deficiencies or proscribing one-size-fits-all solutions to families with unique needs (DeCandia et al., 2023; Vrabic et al., 2020).

Among the children who did not show resilience, we identified three distinct profiles of problematic functioning across neurodevelopmental domains. All three of the maladaptive profiles demonstrated elevated risk scores in fine motor, everyday challenges, and emotion regulation compared to the resilient group. The most severe, and the least common, was Profile E, characterized by all domains being at-risk with the highest being in domains of attention problems, problem solving, cognitive processing, and social skills. All these skills are developing in early childhood as the brain is growing in volume, neural networks are forming, and cognitive and social capacities are being learned (Shonkoff & Garner, 2012). This lays the very foundation for resilient functioning. Our results indicate that there is a small group of young children experiencing homelessness that are at very high risk for later neurodevelopmental and psychosocial delays and impairments without intervention. Compared to those in the resilient functioning profile, these children had significantly higher scores for caregiver mental health problems and overreactive parenting. This confirms our hypotheses and is consistent with previous research (Herbers et al., 2014; Holtrop et al., 2019; Labella et al., 2019; Zhang et al., 2020). For this high-risk group, caregivers as well as the children require intervention, and a two generation strategy is warranted.

In another maladaptive profile (D), the highest scores involved elevations in everyday challenges and social skills with relative strengths in cognitive processing and problem-solving. Compared to children in the resilient profile, being in Profile D was predicted by higher scores for caregiver mental health problems and overreactive parenting. Social emotional skills and self-regulatory capacities are a hallmark of resilient functioning throughout the lifespan (Masten, 2018), the seeds of which are planted in infancy and sprout during the early childhood years. Preschool-age children typically develop the capacity to be soothed and calmed by an adult, take turns, follow directives, and get along with others at an age-appropriate level (Denham et al., 2012; Obradović et al., 2010). Difficulties in this arena may lead to later mental health problems, often seen as an outgrowth of self-regulatory deficits in the early years that were not addressed (Shonkoff & Garner, 2012). In contrast, another of the maladaptive profiles (C) showed the highest scores spanning more domains including gross and fine motor problems, cognitive processing, and play along with mild elevations in attention and problem-solving. While both profiles of C and D showed mixed domains of risk and resilience, children in profile C appeared to demonstrate a more global impact to their development. African American and Hispanic/Latinx children were overrepresented in this group compared to the resilient profile, suggesting that cultural factors, and the experiences of discrimination that accompany minority group experience, may be exerting additional pressures on child outcomes (DeCandia et al., 2022).

While some factors clearly predicted profile membership (e.g., caregiver mental health), others did not. However, considering mean differences in predictor variables between profiles can help to describe how children with different profiles may present. Such differences are descriptive of the different profiles based on group averages but were not predictive beyond other covariates in our regression model. For example, children in Profile C had the highest average score for ACEs, and this average was significantly higher than the average ACEs score in Profile B. Also, resilient children in Profile A were older on average than children in profiles B, C, and D; and the highest risk children in Profile E had the fewest girls compared to other profiles, had the largest proportion of African American children with no White children, and had the smallest proportion of Hispanic/Latinx children compared to the other three profiles. The differences in ethnicity across profile membership should be considered because there may be culturally relevant issues in the assessment of ecological risks and/or domains of children’s neurodevelopment. As an example, Hispanic/Latinx ethnicity was especially prevalent and predictive for Profile C and Profile D, which both had noteworthy elevations in gross and fine motor problems compared to the other profiles. It is worth considering whether there may be culturally relevant differences in caregiver’s perceptions of motor development.

Caregiver mental health problems appeared to be the variable most predictive of children’s profile membership, with significant differences for all other profiles compared to the resilient profile, Profile A. Importantly, caregiver functioning cannot be separated from parenting. In this study, overreactive parenting also uniquely distinguished the low risk resilient from the high-risk maladaptive profiles. Less overreactive parenting predicted profile membership in the resilient functioning group (A), and more overreactive parenting predicted membership in the group with multidomain risks with social challenges, group (E). This finding is consistent with the literature on risk and resilience in children experiencing homelessness as well as children experiencing risk and adversity more generally: more responsive, less overreactive parenting is associated with better child adjustment, and conversely, harsh and negative parenting is associated with some of the worst outcomes (Bornstein, 2015).

The wellbeing of young children is closely linked with the wellbeing of their caregivers, who represent their most proximal context for access to physical and financial resources as well as for nurturing relationships and care (Masten & Palmer, 2019). The importance of caregiver functioning did not vary substantially between the different maladaptive profiles, representing specific developmental issues children present. Therefore, efforts to support healthy child development in contexts of family homelessness should emphasize caregiver and family functioning in concert with assessment and intervention for the whole range of developmental concerns, neurocognitive, social, emotional, and motor as all are relevant in early childhood.

Considering the recent attention to ACEs in the lives of young children experiencing homelessness, it is important to recognize that caregiver mental health problems and overreactive parenting were more predictive of maladaptive profile membership than were children’s ACE scores. These findings do not negate the important of recognizing the role of trauma in young children’s lives, but they do underscore the point that adverse experiences themselves are less informative than indicators of how children and families are coping and adapting in the face of threat. It is also important to note that we did not measure caregiver ACEs, which would likely be associated with caregiver functioning and could predict child functioning consistent with intergenerational transmission of trauma (Cutuli et al., 2017; Narayan et al., 2021). Stress reactivity, emotion regulation, and cognitive processing are all subject to adverse impacts leading to maladaptive developmental trajectories (McEwen, 2017). While ACEs, for the parent or child, are a means to quantify risk exposure, better identifying early adversity and the mechanistic links to impairment for the child is essential to understand how risk and resilience operate in development (Shonkoff, 2016).

Beyond the United States where the current sample was obtained, tools like NEST could be invaluable for efficient screening of children’s neurodevelopmental functioning in socioecological context. Global tools focusing on child development are generally lacking in their ability to identify those at high risk for long term developmental and mental health impacts (Faruk et al., 2020). Research and intervention in low- and middle-income countries in particular could benefit from NEST and similar screening tools that could be easily accessible and adaptable to local contexts for understanding how children’s neurodevelopmental functioning is embedded in their families and broader ecological systems.

Strengths and Weaknesses

With a person-centered approach, the present study yielded nuanced information about how different children in families experiencing homelessness tend to present either resilience or patterns of maladaptive functioning across developmental domains. Strengths of the study include an approach to assessment that was well-suited to the high-risk context of homeless shelters and has been validated with evidence of good psychometric properties (DeCandia et al., 2021). NEST screens for indicators of child wellbeing across functional domains, familial, and ecological factors to place functioning within developmental context. The tool also includes a consistent method of scoring each domain with theoretically informed cut-offs to indicate neurodevelopmental risk, lending itself to straightforward interpretation. The latent profile solution was clear, with evidence of a best-fitting model and good entropy indicating strong separation of the different profiles.

Because NEST was designed to be both effective and practical as a screening tool in low-resourced settings, it relies exclusively on parent-report to assess child functioning, caregiver mental health, parenting, and other ecological factors. Although there are noteworthy strengths to parent-report approaches, including the unique wealth of knowledge caregivers possess about their children, there are also inherent weaknesses. Shared method variance is likely to explain some of the association between indicators of child functioning and other factors, and socially desirable responding on the part of parents can introduce bias. Because socially desirable responding may be most relevant to the parent report of caregiver overreactive parenting, it would be beneficial to replicate these results in future studies with observational measures of parenting. For NEST scoring, it is also noteworthy that some age and gender effects were predictive of different profile membership, which could indicate that future research with larger samples could allow for norms accounting for age and gender differences to be refined to further strengthen the psychometric properties of the measure.

Finally, the current study utilized correlational, cross-sectional data, with children’s neurodevelopment, caregiver functioning, and other factors all measured at a single point in time. We cannot infer any directional or causal effects from the significant associations in this study. Based on theory, we suspect that problems with caregiver mental health and overreactive parenting lead to children’s neurodevelopmental challenges, but it is also possible that caregivers of children with more challenges develop more mental health problems and engage in more overactive parenting as a result of their children’s behaviors and needs. Future research could build upon these findings with longitudinal data to investigate both the direction of effects and whether associations are maintained or change over developmental time. Such efforts could also consider aspects of homeless experiences, such as time spent in shelter and whether and when more stable housing is achieved, to better describe child and family functioning in the unique context of homelessness.

Conclusion

There are important individual differences among children who experience similar adversities of extreme poverty and homelessness. Age, gender, and ethnicity are just some ways to classify children into groups. The results of this study indicate that distinct profiles exist among children experiencing homeless such that we must look to how children are developing, across domains, and what factors are associated with those outcomes, to best understand what each child needs. The person-centered approach highlighted something we already knew: many children experiencing homelessness are resilient in the face of such tremendous adversity. These NEST profiles go further, however, describing how poor functioning can present differently across developmental domains. This more nuanced approach has implications for research, practice, and policy. As we are better able to identify which children are affected in what ways, we are better able to provide targeted services that recognize each child’s unique developmental trajectory to make the best use of limited resources for serving families in need.

Highlights.

  • Children experiencing family homelessness demonstrated five distinct profiles of neurodevelopmental functioning, with over half demonstrating resilience.

  • Children from ethnic minority groups of African American or Hispanic/Latinx were less likely to demonstrate profiles of resilience.

  • Profiles of at-risk functioning were associated with parent mental health problems and overreactive parenting.

Funding Sources:

This study was funded by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD (Grant #1 R44 HD088291-01). The NICHD approved the study design during the grant review process but was not involved in the decision to submit this paper for publication.

We acknowledge all programs and families who participated in this study and shared their experiences so that they might support research to benefit other parents and young children experiencing homelessness.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Disclosure Statement: The authors declare that they have no conflicts of interest.

Contributor Information

Janette E. Herbers, Villanova University, Department of Psychological and Brain Sciences, Villanova, PA, USA

Carmela J. DeCandia, Artemis Associates, Watertown, MA, USA

Katherine T. Volk, C4 Innovations, Needham, MA, USA.

George J. Unick, University of Maryland, School of Social Work, Baltimore, MD, USA

References

  1. Akaike H (1987). Factor analysis and AIC. In Selected papers of Hirotugu Akaike (pp. 371–386). Springer, New York, NY. [Google Scholar]
  2. Asparouhov T, & Muthén B (2014). Auxiliary variables in mixture modeling: Using the BCH method in Mplus to estimate a distal outcome model and an arbitrary secondary model. Mplus Web Notes, 21(2), 1–22. [Google Scholar]
  3. Bassuk EL, Hart JA, & Donovan E (2020). Resetting policies to end family homelessness. Annual Review of Public Health, 41, 247–263. 10.1146/annurev-publhealth-040119-094256 [DOI] [PubMed] [Google Scholar]
  4. Bassuk EL, Richard MK, & Tsertsvadze A (2015). The prevalence of mental illness in homeless children: A systematic review and meta-analysis. Journal of the American Academy of Child & Adolescent Psychiatry, 54(2), 86–96. [DOI] [PubMed] [Google Scholar]
  5. Bornstein MH (2015). Children’s parents. In Bornstein MH, Leventhal T, & Lerner RM (Eds.), Handbook of child psychology and developmental science: Ecological settings and processes (pp. 55–132). John Wiley & Sons, Inc.. [Google Scholar]
  6. Bradley C, McGowan J, & Michelson D (2018). How does homelessness affect parenting behaviour? A systematic critical review and thematic synthesis of qualitative research. Clinical child and family psychology review, 21(1), 94–108. 10.1007/s10567-017-0244-3 [DOI] [PubMed] [Google Scholar]
  7. Brumley B, Fantuzzo J, Perlman S, & Zager ML (2015). The unique relations between early homelessness and educational well-being: An empirical test of the continuum of risk hypothesis. Children and Youth Services Review, 48, 31–37. 10.1016/j.childyouth.2014.11.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Clark R, Weinreb L, Flahive J, & Seifert R (2019). Infants exposed to homelessness: Health, health care use, and health spending from birth to age six. Health Affairs, 38(5), 721–728. 10.1377/hlthaff.2019.00090 [DOI] [PubMed] [Google Scholar]
  9. Collins LM, & Lanza ST (2010). Latent class and latent transition analysis for the social, behavioral, and health sciences. Hoboken, NJ: Wiley. [Google Scholar]
  10. Corman H, Curtis M, Noonan K, & Reichman N (2016). Maternal depression as a risk factor for children’s inadequate housing conditions. Social Science & Medicine, 149, 76–83. 10.1016/j.socscimed.2015.11.054 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Cutuli JJ, & Herbers JE (2014). Promoting resilience for children who experience family homelessness: Opportunities to encourage developmental competence. Cityscape, 16, 113–140. [Google Scholar]
  12. Cutuli JJ, Montgomery AE, Evans-Chase M, & Culhane DP (2017). Childhood adversity, adult homelessness and the intergenerational transmission of risk: A population-representative study of individuals in households with children. Child & Family Social Work, 22(1), 116–125. 10.1111/cfs.12207 [DOI] [Google Scholar]
  13. DeCandia CJ, Herbers JE, Unick GJ, & Volk KT (2023). Parent characteristics associated with neurodevelopmental competence for young children experiencing family homelessness. Journal of Child and Family Studies. Advance online publication. 10.1007/s10826-023-02566-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. DeCandia CJ, Unick GJ, & Volk KT (2021). Validating the neurodevelopmental domain of neurodevelopmental ecological screening tool: An ecological screener for use with 3–5 year-old children impacted by poverty and homelessness. Journal of Psychoeducational Assessment, 39(5), 531–548. 10.1177/07342829211005575 [DOI] [Google Scholar]
  15. DeCandia CJ, Volk KT, & Unick GJ (2022). Evolving our understanding: Housing instability as an ACE for young children. Adversity and Resilience Science, 3(4), 365–380. 10.1007/s42844-022-00080-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. DeCandia CJ, Volk KT, Unick GJ, Rose L, & Donegan W (2020). Developing a screening tool for young children using an ecological framework. Infants and Young Children, 33(4), 237–258. 10.1097/IYC.0000000000000173 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Denham SA, Bassett H, Mincic M, Kalb S, Way E, Wyatt T, & Segal Y (2012). Social–emotional learning profiles of preschoolers’ early school success: A person-centered approach. Learning and Individual Differences, 22(2), 178–189. 10.1016/j.lindif.2011.05.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Evans GW, Li D, & Whipple SS (2013). Cumulative risk and child development. Psychological Bulletin, 139, 1342–1396. 10.1037/a0031808 [DOI] [PubMed] [Google Scholar]
  19. Fanning K (2021). What about the babies? A critical review of infants’ and toddlers’ absence in homelessness scholarship. Infant Behavior and Development, 64, 101625. 10.1016/j.infbeh.2021.101625 [DOI] [PubMed] [Google Scholar]
  20. Faruk T, King C, Muhit M, Islam MK, Jahan I, Baset KU, Badawi N, & Khandaker G (2020). Screening tools for early identification of children with developmental delay in low- and middle-income countries: A systematic review. BMJ Open, 10(11), e038182. 10.1136/bmjopen-2020-038182 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Gubits D, Shinn M, Wood M, Brown SR, Dastrup SR, & Bell SH (2018). What interventions work best for families who experience homelessness? Impact estimates from the Family Options Study. Journal of Policy Analysis and Management, 37(4), 835–866. 10.1002/pam.22071 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Haskett ME, & Armstrong JM (2019). The experience of family homelessness. In Fiese BH, Celano M, Deater-Deckard K, Jouriles EN, & Whisman MA (Eds.), APA handbooks in psychology®. APA handbook of contemporary family psychology: Applications and broad impact of family psychology (pp. 523–538). American Psychological Association. 10.1037/0000100-032 [DOI] [Google Scholar]
  23. Haskett ME, Armstrong JM, & Tisdale J (2016). Developmental status and social–emotional functioning of young children experiencing homelessness. Early Childhood Education Journal, 44(2), 119–125. 10.1007/s10643-015-0691-8 [DOI] [Google Scholar]
  24. Haskett ME, Neal SC, & Norwalk KE (2021). Exploring Variability in Social and Behavioral Adjustment Among Children in Head Start Experiencing Homelessness. Journal of Education for Students Placed at Risk (JESPAR), 26(4), 337–351. 10.1080/10824669.2020.1869554 [DOI] [Google Scholar]
  25. Herbers JE, Cutuli JJ, Jacobs EL, Tabachnick AR, & Kichline T (2019). Early childhood risk and later adaptation: A person-centered approach using latent profiles. Journal of Applied Developmental Psychology, 62, 66–76. 10.1016/j.appdev.2019.01.003 [DOI] [Google Scholar]
  26. Herbers JE, Cutuli JJ, Keane JN, & Leonard JA (2020). Childhood homelessness, resilience, and adolescent mental health: A prospective, person-centered approach. Psychology in the Schools, 52, 1830–1844. 10.1002/pits.22331 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Herbers JE, Cutuli JJ, Supkoff L, Heistad D, Chan C-K, Hinz E, & Masten AS (2012). Early reading skills and academic achievement trajectories of students facing poverty, homelessness, and high residential mobility. Educational Researcher, 41(9), 366–374. 10.3102/0013189X12445320 [DOI] [Google Scholar]
  28. Herbers JE, Cutuli JJ, Supkoff LM, Narayan AJ, & Masten AS (2014). Parenting and coregulation: Adaptive systems for competence in children experiencing homelessness. American Journal of Orthopsychiatry, 84, 420–430. 10.1037/h0099843 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Holtrop K, Piehler TF, Gewirtz AH, & August GJ (2017). Observed parenting in families exposed to homelessness: Child and parent characteristics as predictors of response to the Early Risers intervention. In Child and Family Well-Being and Homelessness (pp. 27–48). Springer, Cham. [Google Scholar]
  30. Huntington N, Buckner JC, & Bassuk EL (2008). Adaptation in homeless children: An empirical examination using cluster analysis. American Behavioral Scientist, 51, 737–755. 10.1177/0002764207311985 [DOI] [Google Scholar]
  31. Jobe-Shields L, Andrews AR, Parra GR, & Williams NA (2015). Person-centered approaches to understanding early family risk. Journal of Family Theory & Review, 7, 432–451. 10.1111/jftr.12118 [DOI] [Google Scholar]
  32. Krahn J, Caine V, Chaw-Kant J, & Singh AE (2018). Housing interventions for homeless, pregnant/parenting women with addictions: A systematic review. Journal of Social Distress and the Homeless, 27(1), 75–88. 10.1080/10530789.2018.1442186 [DOI] [Google Scholar]
  33. Labella M, Narayan A, McCormick C, Desjardins C, & Masten A (2019). Risk and adversity, parenting quality, and children’s social-emotional adjustment in families experiencing homelessness. Child Development, 90(1), 227–244. 10.1111/cdev.12894 [DOI] [PubMed] [Google Scholar]
  34. Laursen BP, & Hoff E (2006). Person-centered and variable-centered approaches to longitudinal data. Merrill-Palmer Quarterly, 52, 377–389. 10.1353/mpq.2006.0029 [DOI] [Google Scholar]
  35. Lucke CM, Narayan AJ, Monn AR, & Masten AS (2021). Social support and parenting among mothers experiencing homelessness. Parenting. Advance online publication. 10.1080/15295192.2021.1882830 [DOI] [Google Scholar]
  36. Manfra L (2019). Impact of homelessness on school readiness skills and early academic achievement: A systematic review of the literature. Early Childhood Education Journal, 47(2), 239–249. 10.1007/s10643-018-0918-6 [DOI] [Google Scholar]
  37. Marra JV, Lin HJ, McCarthy E, Ford J, Rodis E, & Frisman LK (2009). Effects of social support and conflict on parenting among homeless mothers. American Journal of Orthopsychiatry, 79(3), 348–356. [DOI] [PubMed] [Google Scholar]
  38. Marçal KE, Fowler PJ, Hovmand PS, & Cohen J (2021). Understanding mechanisms driving family homeless shelter use and child mental health. Journal of Social Service Research, 47(4), 473–485. 10.1080/01488376.2020.1831681 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Masten AS, Fiat AE, Labella MH, & Strack RA (2015). Educating homeless and highly mobile students: Implications of research on risk and resilience. School Psychology Review, 44, 315–330. 10.17105/spr-15-0068.1 [DOI] [Google Scholar]
  40. Masten AS, & Palmer AR (2019). Parenting to promote resilience in children. In Bornstein MH (Eds.), Handbook of parenting (3rd ed., Vol. 5, pp. 156–188). Routledge. [Google Scholar]
  41. Masyn KE (2013). Latent class analysis and finite mixture modeling. In Little T (Ed.), The Oxford handbook of quantitative methods (Vol. 2, pp. 551–611). Oxford, UK: Oxford University Press. [Google Scholar]
  42. McEwen BS (2017). Neurobiological and systemic effects of chronic stress. Chronic Stress, 1, 1–11. 10.1177/2470547017692328 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. McLachlan G, & Peel D (2000). Finite mixture models. New York, NY: Wiley. [Google Scholar]
  44. Muthén LK, & Muthén BO (2012). Mplus statistical analysis with latent variables: Users guide (7th ed.). Los Angeles, CA: Muthén & Muthén. [Google Scholar]
  45. Narayan AJ, Lieberman AF, & Masten AS (2021). Intergenerational transmission and prevention of adverse childhood experiences (ACEs). Clinical Psychology Review, 85, 101997. 10.1016/j.cpr.2021.101997 [DOI] [PubMed] [Google Scholar]
  46. National Center for Homeless Education (NCHE). (2021). Federal data summary, school years 2016–17 through 2018–19: Education for homeless children and youth. https://nche.ed.gov/wp-content/uploads/2021/04/Federal-Data-Summary-SY-16.17-to-18.19-Final.pdf
  47. Obradović J (2010). Effortful control and adaptive functioning of homeless children: Variable-focused and person-focused analyses. Journal of Applied Developmental Psychology, 31, 109–117. 10.1016/j.appdev.2009.09.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Sandel M, Sheward R, Ettinger de Cuba E, Coleman S, Frank D, Chilton M, et al. (2018). Unstable housing and caregiver and child health in renter families. Pediatrics, 141(2), e20172199. 10.1542/peds.2017-2199 [DOI] [PubMed] [Google Scholar]
  49. Volk KT, DeCandia CJ, & Unick GJ (2023). The full picture: Incorporating ecological factors when conducting developmental screening with young children experiencing homelessness. Child Indicators Research, 16(1), 87–108. 10.1007/s12187-022-09978-6 [DOI] [Google Scholar]
  50. von Eye A, Bergman LR, & Hsieh C (2015). Person-oriented methodological approaches. In Overton WF & Molenaar PC (Eds.), Theory and method: Vol. 1. Handbook of child psychology and developmental science (7th ed., pp. 789–842). Hoboken, NJ: Wiley. [Google Scholar]
  51. Zhang J, Lee SK, Piehler TF, Gewirtz AH, & August GJ (2020). Bidirectional relations between parenting practices and child externalizing behaviors in formerly homeless families: A random-intercept cross-lagged panel analysis. Parenting, 20(3), 177–199. 10.1080/15295192.2019.1694833 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES