Parental perceived immigration threat and children’s mental health, self-regulation and executive functioning in Pre-Kindergarten

Abstract

Many children in immigrant households endure unique stressors shaped by national, state and local immigration policies and enforcement activity in the United States. Qualitative studies find that during times of heightened immigration enforcement, children as young as three years of age show signs of behavioral distress related to national anti-immigrant sentiment and the possibility of losing a parent. Using multiple sources of data from 168 racially and ethnically diverse families of children in pre-Kindergarten, the present study examined variability in perceived levels of immigration enforcement threat by parental immigrant status and ethnicity. This study examined associations between immigration enforcement threat and child mental health, self-regulation, and executive functioning and whether parent immigrant status or child gender moderate these associations. We found substantial variability in perceived immigration threat, with immigrant parents and Latinx parents reporting significantly greater levels of immigration threat compared to non-immigrant parents and non-Latinx parents. Immigration enforcement threat was associated with greater child separation anxiety and overanxious behaviors, and lower self-regulation among boys and girls and among children of immigrant and US-born parents. In contrast to our hypothesis, immigration enforcement threat was associated with higher self-regulation according to independent assessor ratings. Educators and healthcare providers working with young children from immigrant and Latinx households should be aware of the disproportionate stress experienced by immigrant and Latinx families due to a xenophobic sociopolitical climate marked by heightened immigration enforcement threat and racist, anti-immigrant rhetoric.

One out of every four children under the age of 18 years in the United States (U.S.) has an immigrant parent (Batalova et al., 2020). Children in immigrant households endure unique stressors shaped by national, state and local immigration policies and enforcement activity that threaten family unity and children’s sense of safety (Barajas-Gonzalez et al., 2018; Barajas-Gonzalez et al., 2021; Dreby, 2015; Gonzales, 2016; Zayas & Cook Heffron, 2016). A growing literature demonstrates a compelling link between adolescents’ worry and fear about the immigration climate and their well-being. For example, self-reported worry about the immigration climate and uncertainty about family safety have been associated with anxiety, depression, and suicidal ideation among U.S. born adolescents in immigrant families (Eskenazi et al., 2019; Gulbas et al., 2015; Roche et al., 2021). Qualitative studies with Latinx adolescents indicate difficulty focusing in school due to immigration-related fear and concerns regarding their safety and that of their family (Wray-Lake et al., 2018). Less is known about the association between immigration-related fear and young children’s well-being. A better understanding of the impact of the U.S. immigration climate on young children’s well-being - conceptualized here as their mental health, self-regulation and executive functioning - is necessary to inform health promoting policies and practices. This is especially urgent for young children given the importance of mental health, executive functioning and self-regulation for learning and healthy development (Center on the Developing Child, 2010; Blair & Raver, 2015), the cumulative nature of learning (Shuell, 1986; Ansari & Gottfried, 2021), and the harmful sequelae of untreated childhood distress (National Scientific Council on the Developing Child, 2020; Shonkoff et al., 2021).

Restrictive immigration policies, anti-immigrant rhetoric, and immigration enforcement represent a pervasive threat for children in immigrant households

Over the past several decades, restrictive immigration policies, anti-immigrant legislation, and increased immigration enforcement activity in targeted communities have produced a “culture of fear” among immigrant families in the U.S. (Vargas et al., 2017; p. 460). Scholars trace current anti-immigrant and racist attitudes toward immigrants to the 1960s (e.g., Hart-Cellar Act (1965)), a period with limits on the number of people that could enter the U.S. from the Western hemisphere (Walsdorf et al., 2019). This legislation resulted in mass “illegal” migration as it rendered thousands of immigrants from Latin America with familial and economic ties to the U.S. without means of obtaining timely authorized legal status (Massey & Pren, 2012). Since 1965, politicians and journalists have increasingly described immigration from Latin America as a threat to the U.S. (Chavez, 2008; Chavez et al., 2019; Massey & Pren, 2012).

A series of exclusionary immigration laws passed over the past three decades has resulted in a significant increase in deportations and policing by immigration agents in the U.S. interior (Hagan et al., 2010). The 1996 Illegal Immigrant Responsibility and Immigrant Reform Act (IIRIRA) and Anti-Terrorism Effective Death Penalty Act (AEDPA) expanded the range of offenses for which a person could be deported, permitted retroactive deportation for previous offenses, and strengthened the cooperation between local and state enforcement to enforce immigration laws (Kanstroom, 2007). These laws made it easier to arrest, detain and deport non-citizens (Hagan et al., 2010). The passage of the USA Patriot Act (2001) expanded the categories of immigrants eligible for deportation by increasing authority to deport groups of immigrants perceived as threats to national security (Haganet et al., 2010). Two years later, the Department of Homeland Security (DHS) created Immigration and Custom Enforcement (ICE), with the mission to detain and deport “criminal and fugitive” non-citizens at the border and within the interior of the U.S. (Kanstroom, 2007). The Secure Communities Strategy (2009) formed a partnership and database sharing agreement between the Departments of Justice and Homeland Security to deport “criminal aliens.” Reflecting the racialization of immigration, Latinxs were disproportionately affected by immigration enforcement, with 93% of Latinxs arrested through Secure Communities although they only comprised 77% of the undocumented population in the U.S. at the time (Kohli et al., 2011). Of those arrested through Secure Communities, 70% had no criminal history and nearly 40% had a U.S. citizen spouse or child (Kohli et al., 2011). Since 2009, over half a million parents of U.S. citizen children have been deported (DHS, 2020; MPI, 2016).

Multiple restrictive immigration policy changes during the 45^th presidency (Pierce & Bolter, 2020) including a January 2017 executive order to prioritize all unauthorized immigrants for deportation, increased funding for immigration enforcement in the U.S. interior, attempts to rescind the Deferred Action for Childhood Arrivals (DACA) program, the separation of children from their parents at the Southern border under “zero tolerance”, and changes to the public charge rule heightened immigration enforcement fear in the U.S. (Capps et al, 2018; Nienhusser & Oshio, 2018). These changes in immigration policy and enforcement practices, coupled with disparaging language from politicians and widespread media coverage of immigrants being detained and deported, are linked to significant distress in immigrant households (Abu-Ras et al., 2018; Eskenazi et al., 2019; Roche et al., 2018; Roche et al., 2021).

A growing body of literature indicates that immigration laws, policies and enforcement practices “spillover” to other groups, including U.S.-born Latinxs and other U.S. born residents (Aranda et al., 2014; Krieger et al.,2018; Novak et al., 2017). For example, a cohort study of 397 U.S.-born Latinx adolescent children of immigrant parents found that worry about family separation because of restrictive immigration policies was associated with adolescents’ anxiety, sleep problems, and blood pressure in the year after the 2016 presidential election. Relative to adolescents with lower levels of worry, U.S. born adolescents reporting “high” levels of worry about family separation due to deportation had higher self- and parent-reported anxiety, and worse sleep quality (Eskenazi et al., 2019).

Increasingly restrictive immigration policies, greater enforcement in the U.S. interior, and national anti-immigrant sentiment can function as pervasive threat for immigrant households (and potentially for non-immigrant households), contributing to chronic uncertainty about family unity, and by extension, a child’s sense of safety (Barajas-Gonzalez et al., 2018; Dreby, 2012; Gulbas et al., 2016). Parent and child care provider reports of child behaviors suggest that during heightened times of immigration enforcement, children as young as three years of age are adversely affected by national anti-immigrant sentiment and the possibility of losing a parent (Cervantes et al., 2018). Qualitative data collected in 2017 with more than 150 early childhood educators and parents across multiple states - specifically about the impact of the immigration climate on young children under 8 years old - identifies increased aggression, anxiety and withdrawal in young children due to the immigration climate (Cervantes et al., 2018). Importantly, Cervantes and colleagues (2018) found that children whose parents are U.S. citizens also experience fear of losing a parent, suggesting that immigration policies can negatively affect young children even though, or perhaps because, they do not understand the details of the policies.

Threat in early childhood and implications for mental health, self-regulation and executive functioning

Children’s experiences of their environment as unsafe or threatening elicit a stress response which can disrupt information processing and alter the ways they selectively engage with - or disengage from - their environments (Aneshensel & Sucoff, 1996; McCoy et al., 2015; Ross & Mirowsky, 2001). Threat triggers a set of stress responses that includes the activation of neuroendocrine, neurocognitive, and attentional systems (Pechtel & Pizzagalli, 2011), which can alter cognitive and emotional functioning through the development of biased and less accurate attentional and appraisal processes of social and emotional stimuli (Dodge, 2006). For example, increased monitoring of the environment for both real and perceived threats, as evidenced by faster attending to negative stimuli among children living in communities characterized by high levels of crime (McCoy et al., 2015), can take attention away from benign stimuli (Posner & Rothbart, 2000). These patterns of social cognition constitute processes that are associated with behaviors such as early vigilance and behavioral suppression (Fowler et al., 2009; Henrich & Shahar, 2013). While heightened vigilance to threat is adaptive in a potentially dangerous situation, prolonged activation of the threat response may translate into emotional and behavioral problems (Wilson et al., 2009). Indeed, early childhood experiences of harm or threat of harm are associated with mental health problems in adolescence (Miller et al., 2021).

This pattern of attention deployment may also disrupt the development of executive function and self-regulation if it results in habitually more reactive, less prototypically well-regulated, responses to the environment (Blair & McKinnon, 2013; Blair & Raver, 2012). Self-regulation is “the primarily volitional management of arousal or activity in attention, emotion, and stress response systems in ways that facilitate the use of executive function abilities in the service of goal-directed actions” (Ursache et al., 2012, p.123). As such, executive functions are “cognitive aspects of self-regulation that are engaged for the purposes of the effortful processing of information and intentional top down control of behavior” (Blair & Ursache, 2011). Threat activates a neurobiological system that supports monitoring the environment to identify both real and perceived threats at the cost of taking attention away from other, non-threatening stimuli (Posner & Rothbart, 2000). For example, children living in communities characterized by high levels of crime exhibit a more vigilant profile of attention deployment in which they are, on average, faster to attend to negative stimuli than are children from lower crime communities (McCoy et al., 2015). This pattern of directing attention toward potential threat and away from other aspects of the environment may have costs for the use and development of self-regulation. Over time, the consistent experience of threat could adaptively lead to a habitually more reactive and less self-regulated response to the environment. Consistent with developmental psychobiological and experiential canalization models (Blair & McKinnon, 2013), impairments in attention and impulse control, two important aspects of self-regulation, have also been found among children living in proximity to a recent homicide (Sharkey et al., 2012). Moreover, recent violent crime in children's residential communities has been shown to lead to faster but less accurate task performance on the dot probe, a task measuring threat bias (children were approximately 9 years old; McCoy, et al., 2015). No studies, however, have examined the role of immigration enforcement threat in association with children’s self-regulation or executive functioning.

The present study

In a sample of children enrolled in public pre-Kindergarten (Pre-K) programs in early childhood centers in historically disinvested neighborhoods in New York City (NYC), this study aims to: (1) describe the level of immigration-related threat experienced by immigrant and U.S. born parents; (2) examine differences in levels of immigration-related threat by parental immigrant status and by ethnicity; and (3) understand whether immigration-related threat is associated with child mental health problems (e.g., anger, sadness, overanxious fear, separation anxiety), self-regulation (e.g., emotion regulation and attention), and executive function (e.g., effortful cognitive processing such as inhibitory control and set shifting). We hypothesized that parent report of immigration enforcement threat would be associated with higher rates of mental health problems, lower levels of self-regulation, and lower performance on executive function tasks among children.

Additionally, we examined whether associations between immigration-related threat and child outcomes are moderated by parent immigrant status or child gender. We test for moderation by parent immigrant status given the possibility that households where at least one parent is an immigrant may experience the immigration climate more acutely than non-immigrant households (Asad, 2020; Roche et al., 2018). We also test for moderation by child gender, as some studies examining the impact of immigration worry and distress have found a moderated effect, where the association between immigration-related worry and some indicators of distress are significant among boys, but not girls (Eskenazi et al., 2019; Santos et al., 2018). For example, a study of 689 Latinx middle schoolers in Arizona revealed a significant association between awareness of SB1070 (a law mandating that local police verify the immigration status of any person they reasonably suspect of being in the country without proper documentation) and lower levels of self-regulation in the classroom among boys but not girls (Santos et al., 2018).

Data and Methods

Sample

Participants were families who were part of a study of programmatic supports for early education centers with pre-K programs in historically disinvested neighborhoods. Nineteen centers in NYC were involved in the study. Centers were eligible for the study if they served high proportions of students living in poverty (defined by the school district, in terms of the proportion of students from census tracts with concentrated poverty, relative to all pre-K programs district-wide, and the number of students living in temporary housing), and if they had at least two pre-K classrooms. Based on the aims of the broader implementation-effectiveness study, pre-K leaders were invited to participate in the study if the leaders’ expressed interest in programmatic supports related to family engagement and social-emotional learning and willingness to be randomly assigned to implement a center-based program for families. This study utilizes baseline data (prior to implementation of programmatic supports) collected from parents, teachers and children between October, 2018 and April, 2019.

Families were eligible for the study if their child attended pre-K in one of the 19 study centers and if one parent/caregiver spoke English or Spanish (> 90% of these centers had families who spoke at least one of these languages). We conducted all research activities in the parent’s preferred language; 44% of caregivers indicated a preference for communication to be in Spanish (and were contacted by bilingual research assistants). Surveys, administered over the phone, lasted approximately one hour; parents received a $25 gift card for their participation. The Institutional Review Boards of the University and School District approved all study activities (#s17-01812). The phased informed consent procedure invited parents to consent to their child’s participation in study assessments and teacher ratings of children’s behavior (n=313), and inquired about their interest in participating in telephone surveys (n=302). We were able to reach, consent, and complete phone surveys with 182 parents (81 did not answer their phone or did not respond at the scheduled time of interview despite multiple attempts, 9 had disconnected or incorrect phone numbers, and 30 declined participation). The current study includes 168 families (92%) with complete survey data on key study variables (i.e., child outcomes and the Perceived Immigration Policies Effects (PIPES, Ayón, 2017)). The mean age of the children included in this study was 4.44 years (53.34 months). According to parents, 93 (56%) of the children were boys and 148 (95%) were U.S. born; gender was not provided for three children and birth location was not provided for 12 children. A total of 118 (70%) children had parents who identified as Latinx. Of the Latinx parents who provided their national heritage (n=114, 96%), 39 (34%) were from the Caribbean (28 from Dominican Republic, 11 from Puerto Rico); 37 (32%) were from Mexico; 19 (17%) were from Central America (11 from Guatemala, 3 from Central America with no specific country provided, 2 from El Salvador; 1 each from Honduras, Nicaragua and Panama); 11 (10%) were from South America (9 from Ecuador and 1 each from Peru and Chile) and 8 (7%) parents indicated that they were two or more races or nationalities. A total of 23 (14%) parents identified as non-Latinx Black, 13 (8%) identified as non-Latinx White, 9 (5%) identified as non-Latinx Asian, and 5 (3%) identified as other race/ethnicity. Seventy (42%) parents indicated they were immigrants: 46 (39%) of Latinx respondents, 3 (13%) of non-Latinx Black respondents, 8 (67%) of non-Latinx White respondents, 8 (89%) of non-Latinx Asian respondents, and 5 (100%) of those identifying their race/ethnicity as “other.” One hundred children (61%) had parents who had a high school (or less) level of education; information was not available for six children. Of the 99 (59%) children whose parents provided family income, 63 (64%) reported an income below 200% of the federal poverty level.

Procedures

The study used a phased consent approach: first, the study team invited families to participate via school-based assessments including direct assessments of the child, teacher report on children’s experiences in pre-K and review of administrative records; second, the study team invited parents to participate in phone surveys about themselves and their child. Parents provided written consent for school-based assessments, administrative data and teacher report on children’s behaviors; parents gave verbal consent for the phone survey. The study provided all consent forms and recruitment materials in English and Spanish, and the study team included bilingual Spanish speakers for recruitment and phone interviews.

The current study includes data from multiple sources. Parents reported on immigration threat, demographic information, and children’s mental health through phone surveys, and teachers reported on children’s behavior in the classroom through paper surveys. Children participated in direct assessments of executive function, and research staff administering the assessments completed ratings of child behavior during the assessment.

Measures

All measures were in English and Spanish, using the developer’s Spanish version when available. For measures without the developer’s translation or in which the developer’s translation contained inaccuracies as determined by the bilingual study team, the study team translated and back translated the measures, and each measure was verified by a working group of four native Spanish speakers. We used a consensus approach for translation of words or phrases with several options for translation; the lead author made final decisions as needed. We present scale reliability statistics below for the whole sample and by language.

Immigration Enforcement Threat

The Perceived Immigration Policies Effects Scale (PIPES, Ayón, 2017), developed originally in Spanish, measures the impact of immigration policies on parents. The current study used the threat to family subscale (5 items, overall α = .90, English α = .84, Spanish α = .87; “did you fear that you or a family member would be reported to immigration officials?”) and children’s vulnerability (hereafter referred to as “threat to child”; 5 items, overall α = .80, English α = .82, Spanish α = .80; “have your children been stressed about family member being deported or detained?”). To increase relevance for non-immigrant families in our study, we added two items to the Threat to Family subscale to expand the construct to include friends (e.g. “did you fear that your friends would be reported to immigration officials?”). Caregivers respond on a five-point scale ranging from 1 to 5 (Never, Rarely, Sometimes, Very Often, Always). A higher average score is indicative of higher levels of parental perceived immigration enforcement threat.

Child Mental Health

Parents reported on four scales from the The National Institutes of Health (NIH) Toolbox Emotion Battery (anger, fear-overanxious, fear-separation anxiety, sadness) for children aged 3-7 (Gershon et al., 2013). The NIH Toolbox Emotion Battery was developed and validated by a large team of measurement experts across multiple institutions as part of the NIH Blueprint for Neuroscience Research Validity. The NIH Toolbox measures are available in both English and Spanish. The anger scale (9 items, overall α = .80, English α = .79, Spanish α = .83 based on our sample) measures hostility and expression of frustration (e.g., “has temper tantrums or a hot temper”; the sadness scale (7 items, overall α = .67; English α = .52, Spanish α = .73) measures low levels of positive affect and cognitive indicators of depression (e.g. “cries a lot”); the fear-over anxious scale (6 items, overall α = .66, English α = .65, Spanish α = .69) assesses fear, worry and hyperarousal (e.g., “worries about things in the future”); and the fear-separation anxiety scale (7 items, overall α = .73, English α = .64; Spanish α = .76) assesses fear of being separated from home and from loved ones (e.g., “is afraid of being away from home”). Both fear subscales measure symptoms of anxiety reflective of autonomic arousal and perceptions of threat. Parents respond to symptom frequency using a three-point scale (Never/Not True, Sometimes True, Often/Very True). We used the standardized scale scores from each of the four subscales; intercorrelations ranged from r= .26 to .56. Internal consistency has been previously reported in both the English and Spanish versions of the parent-report for children aged 3-7, with alphas ranging from .77 (sadness) to .85 (anger) in the English version (Salsman et al., 2013) and .66 (fear-overanxious) to .71 (fear-separation anxiety; Fox et al., 2020). There is some initial evidence of concurrent validity in parent-report of children aged 3-7 (Salsman et al., 2013).

Child Self-Regulation

Parents and Teachers reported on Emotion Regulation with the Emotion Regulation subscale of the Social Competence Scale (Parent and Teacher versions; CPPRG, 1992; Jones et al., 2015). The parent version includes six items (e.g., “your child can accept things not going his/her way”; “your child controls his/her temper when there is a disagreement”) with high internal reliability in both English and Spanish in our sample (overall α = .85, English α = .79, Spanish α = .90). The teacher version includes ten items (e.g., “this child easily becomes angry with me” and “it is easy to be in tune with what this child is feeling”), with high internal reliability in both English and Spanish in our sample (overall α = .95, English α = .96, Spanish α = .91). Teachers and parents rated the extent to which each item is characteristic of the child on a 5-point scale (Not at all, A little, Moderately Well, Well, Very well). The measure has high discriminant validity, with significantly higher levels of social competence in normative samples compared to higher risk samples. Construct validity has been documented on the parent version of the Social Competence Scale with preschool-aged children; the scale was relatively stable over 24 months and was correlated with measures of social competence and parent ratings of emotion regulation (Corrigan, 2002; Gouley et al.,2008). The concordance for parent and teacher rating of social competence was low (r=.17), as is often the case with parent and teacher rating of child behavior as they are rating behavior in different settings (Achenbach et al., 1987; Fält et al., 2018; Klein et al., 2019).

Attention/Impulse Control was assessed using the Preschool Self-Regulation Assessment (PSRA) Assessor Report. The PSRA is a 28-item measure that assesses the child’s level of attention, emotion and behavior on a 4-item scale ranging from Never to Always (Raver et al., 2009; Smith-Donald et al., 2007). Trained research staff completed the PSRA immediately after they administered a battery of direct child assessments. The PSRA was developed with Black and Latinx families in preschool, and demonstrates adequate construct and concurrent validity (Raver et al., 2011; Smith-Donald et al., 2007). The PSRA includes two scales: attention/impulse control and positive emotion. The attention/impulse control subscale includes 18 items (e.g., “thinks and plans before beginning each task”; “has difficulty waiting between tasks”; “daydreams, has trouble focusing”. The attention/impulse control subscale has good reliability (α = .93; inter-rater reliability ICC =.86) (Smith-Donald et al., 2007).

Attention problems.

We assessed attention problems with the attention problems subscale from The Pediatric Symptom Checklist (PSC). The PSC is a reliable and valid measure for assessing children’s externalizing, internalizing, and attention problems based on parent report (Gardner et al., 2007; Gardner et al., 1999; Murphy et al., 2016). For this study, the sum of the attention problem scale was used (4 items, overall α = .73, English α = .79, Spanish α = .62). The scale has been validated for use in in primary care and community settings for children ages 4 to 16 years old from diverse backgrounds. Parents rated how well each statement describes their child on a three-point scale (Never (0), Sometimes (1), Often (2)). The PSC has shown high rates of sensitivity and specificity in different pediatric settings (Jellinek et al., 1988).

Child Executive Functioning

Hearts and Flowers.

The Hearts and Flowers (Davidson et al., 2006) task assesses children’s executive function skills by requiring attention shifting, working memory and inhibitory control. We administered the task to children individually on a tablet in English or Spanish. In the task, a red heart or flower appears on the left or right side of the screen (2500ms) and children are asked to respond to the stimuli by tapping a button on the left or right side of the screen, following a set of rules. The task includes three sections of increasing difficulty: hearts, flowers, and hearts and flowers mixed. On the hearts (congruent) block, the rule is to press the button that is on the same side as the heart (12 Trials). On the flowers block, the rule is to press the button that is on the opposite side of the flower (12 Trials). Finally, on the 32 mixed trials, both hearts (press on same side) and flowers (press on opposite side) rules are presented and children must follow the pre-established rules for each. The Hearts and Flowers has been validated for use on tablets (Obradović et al., 2018). Preliminary analyses of children’s performance on this task suggested that the mixed block of trials may have been too difficult, thus, in line with prior work, performance on the task is scored as percent accuracy on the flowers block, (Diamond & Barnett). Higher accuracy indicates better executive function. The Hearts and Flowers has demonstrated sensitivity to intervention programs (Diamond et al., 2007; Lakes et al., 2013) and changes in accuracy and reaction times by age (Davidson et al., 2006).

Dimensional Change Card Sort (DCCS).

The Dimensional Change Card Sort (Gershon et al., 2013; Zelazo et al., 2013) primarily requires the attention shifting aspect of executive function. Children are presented with a series of cards depicting images that vary along two dimensions: color and shape (e.g., boats and rabbits, colored white and yellow). Children are first asked to sort the cards by one dimension (i.e., color) for 6 trials. To receive a passing score for this block, children must correctly sort at least 5 out of 6 cards. In the second block of trials, children are then asked to sort the cards by the other dimension (i.e., shape) for 6 trials. If children correctly sort at least 5 out of 6 cards in order to pass this block of trials, they then move on to the third block of ‘border’ trials. In this third block of 12 trials, when the card has a border around it, children are asked to sort the card by the color dimension, but when the card does not have a border around it, children are asked to sort the card by the shape dimension. Children who correctly sort at least 9 of 12 cards are deemed to pass the third block of the task. Children receive a score of 0 if they did not pass the first block of trials, a score of 1 if they passed only the first block of trials, a score of 2 if they passed the second block of trials, and a score of 3 if they passed the border trials (Zelazo, 2006). Prior studies have shown moderate to excellent test-retest reliability for the DCCS (Beck et al., 2011; Fuhs & Days, 2014). Researchers have also demonstrated the validity of the DCCS with children ages 3–5 years (Beck et al., 2011) as well as its convergent validity (Carlson et al., 2016).

Pencil Tap.

The Pencil Tap task (Diamond & Taylor, 1996) assesses children’s executive function by primarily requiring inhibitory control and working memory. Children are presented with 16 trials in which they are asked to tap on a table twice with a pencil when the research assistant taps once, and once when the experimenter taps twice. The task is scored as the percentage of trials on which children respond correctly. This task has demonstrated robust psychometric properties, with reliability coefficients of alpha = .93 in our Pre-K sample. Good internal consistency has been previously reported ranging from .82 to .87 in pre-school to .75 to .84 in 5 years olds/kindergarten (Blair & Razza, 2007; Bierman et al, 2008).

Analysis

To describe the level of immigration-related threat experienced by families (aim 1) and by immigrant status and ethnicity (aim 2), we conducted a series of descriptive analyses and regression analyses (unadjusted and adjusted for confounders). Given that the Children’s Vulnerability subscale includes questions about the child’s behavior due to immigration threat, subsequent analyses of associations with child outcomes were limited to the Threat to Family/Friends subscale (hereafter referred to as the “threat to family” subscale for brevity).

To examine the association between threat to family and child outcomes (aim 3), we carried out multivariate analysis of variance because we considered 11 child outcomes. The multivariate analysis allows assessing the association between threat to family and multiple child outcomes simultaneously, which avoids the risk of Type I error due to multiple testing (for multiple child outcomes). We grouped child outcomes into three categories and conducted three multivariate analyses. One analysis for mental health related outcomes (including sadness, fear-overanxious, fear-separation anxiety, and anger outcomes; correlations among these four measures range from r= .34 to .47), one for self-regulation related child outcomes (including parent and teacher-rated emotion regulation, assessor-rated attention/impulse control problem, parent-rated attention problems; correlations among these four measures range from r= .02 to .35), and the other for executive function related outcomes (including Hearts & Flowers, pencil tap, and DCCS; correlations among these three measures range from r= .24 to .39). We used SAS Proc MIXED linear mixed effects models to examine the association (Diggle et al., 2005). In the multivariate analysis, we created a long data file that included a categorical indicator for the child outcome domain (e.g., a 4-level indicator for child mental health outcomes). Child outcome scores were modeled as a function of threat to family (continuous score) and the indicator of child outcome domain, adjusted for parent education, parent immigrant status, parent ethnicity, and child gender. We allowed residuals between domains to correlate. To account for potential correlations among outcomes of children, we included random effects for schools in the model.

Of the 165 students whose parents provided PIPES data and child gender data, 93% had parent ratings, 80% had child test data and 56% had teacher ratings. Incomplete executive function data was largely due to technical issues, and incomplete data for teacher ratings was largely due to teacher non-response. To account for partial missing data, we applied multiple imputation to impute missing values, an appropriate method to address data missingness (Madley-Dowd et al., 2019). We generated 10 imputed data sets for subsequent analysis (Little & Rubin, 2002). We used SAS® software, version 9.3 of the SAS System (SAS Institute Inc. 2011) Proc MI and Proc MIANALYZE to impute, merge, and analyze data.

To study whether associations between threat to family and child outcomes differed by child gender and parent immigrant status, we built on the main effect model (using imputed data), and added a moderator, moderator-by-PIPES, moderator-by-domain, and moderator-by-PIPES-by-domain interaction in the main effect model. We tested each moderator separately.

Results

Level of immigration threat experienced in pre-K households.

Consistent with the literature on the spillover effects of immigration policies and enforcement practices, both immigrant and non-immigrant parents reported worry about immigration threat. As shown in Table 1, one-third to one-half of pre-K households (31 - 56%) reported experiencing immigration threat to family at least “sometimes”. The proportion of families reporting immigration threat to child was lower, with 5 - 12% of all respondents reporting concerns about the impact of the immigration climate on their pre-K child at least “sometimes”. Parents who experienced higher levels of immigration threat to family tended to experience higher levels of concerns regarding immigration threat to child (r = .50).

Table 1.

Perceived Threat to Family and Friends, and Children’s Vulnerability

	Total	Immigrant Status		Ethnicity
PIPES Item Response (Binary; 1=experience fear/worry sometime or more than sometime, 0=never or rarely experience fear/worry)	Total Sample (n=168)	Immigrant (n=70)	Non- Immigrant (n=98)	Latinx (n=118)	Non Latinx (n=50)
	%	%	%	%	%
Threat to Family & Friends
Worry about the impact of immigration policies on family (1= ≥ sometime)	43.3	53.9	35.1*	50.0	26.0**
Worry about the impact of immigration policies on friends	55.9	61.4	51.6	62.9	42.0*
Fear you or a family member would be reported to immigration officials	30.9	37.1	25.8	40.7	8.0***
Fear friends would be reported to immigration officials	39.5	50.0	31.3*	47.9	20.0***
Worry family separation due to deportation	41.9	49.3	36.1	53.0	16.0***
Children’s Vulnerability (Threat to child)
Any of your children been stressed about family members being deported or detained?	9.1	16.2	4.2**	12.2	2.0*
Any of your children felt unsafe due to immigration policies?	10.3	16.2	6.3*	13.9	2.0*
Been concerned that your children were having emotional problems due to immigration policies	9.0	14.7	5.2*	12.1	2.0*
Any of your children feared authorities due to immigration policies	4.8	7.4	3.1	6.0	2.0
Any of your children had difficulties focusing in school due to immigration policies	11.7	22.1	4.3***	16.8	0.0***
Mean Scale Score (Likert scale 1-5)	M(SD)	M(SD)	M(SD)	M(SD)	M(SD)
Threat to Family/Friends (Mean of items 1-5)	2.34 (1.21)	2.63 (1.28)	2.11 (1.11)**	2.62 (1.26)	1.67 (.78)***
Children’s Vulnerability (Mean of items 6-10)	1.31 (.69)	1.53 (.90)	1.15 (.45)***	1.41 (.79)	1.07 (.27)***
Correlation between Threat to Family/Friends & Children’s Vulnerability	.50***	.49***	.50***	.48***	.29*

Note. Cronbach’s Alpha reliability is .90 for both the Threat to Family & Friends scale and the Children’s Vulnerability scale. Group comparisons were based on chi-square for the binary items, and non-adjusted regression analysis for the continuous mean scale scores. Similar findings emerged for the adjusted analysis (adjusting for child gender and parental education).

^*p <.05

^**p <.01

^***p <.001

Level of Immigration Threat by Immigrant Status and Ethnicity

There were mean-level (unadjusted) differences between immigrant and ethnicity groups on immigration threat to family and threat to child (Table 1). Immigrant respondents reported more threat to family and more threat to child than non-immigrant respondents did. Latinx respondents reported more threat to family and more threat to child than non-Latinx respondents did. Group differences remained significant after adjusting for potential confounders (parental education and child gender). Examination of item-level differences illustrate the unique experiences of immigrant and Latinx families (see Table 1). For example, whereas only 4% of non-immigrants indicated concerns about children being stressed about family members being deported or detained, this concern was endorsed by 16% of immigrant respondents. Similarly, 2% of non-Latinx respondents indicated that their children felt unsafe due to immigration policies in comparison to 14% of Latinx respondents. Whereas 8% of non-Latinx respondents feared that they or a family member would be reported to immigration officials, 40% of Latinx respondents feared that they or a family member would be reported to immigration officials.

In order to explore immigrant-by-ethnicity group patterns, we graphed item level responses (see Figure 1) and mean levels of immigration threat to family and threat to child in our sample (see Figure 2). Figure 2 shows mean level of immigration threat to family, and threat to child scales for four non-overlapping groups: Latinx immigrant (n =46); Latinx non-immigrant (n = 72); immigrant, non Latinx (n=24) and non-immigrant non-Latinx (n=26). Visual inspection of findings for both scales show the highest level of threat for the Latinx immigrant group, and a stepwise pattern for the threat to family and friends scale.

Figure 1. Perceived Threat to Family/Friends, and Children’s Vulnerability: By Immigrant Status among Latinx and non-Latinx Families.

Note. Among Latinx families, 46 were immigrants and 72 were non-immigrants. Among non-Latinx families, 24 were immigrants and 26 were non-immigrants. Figure shows percentage indicating “at least sometimes”. The first five statements make up the children’s vulnerability (ie. threat to child) subscale. The last five items make up the threat to family/friends subscale.

Figure 2. Mean levels of perceived immigration threat for four immigrant status and ethnicity subgroups.

Association between Immigration Threat and Child Outcomes

We carried out three multivariate analyses to examine the association between immigration threat to family and child outcomes assessed with multiple measures (mental health outcomes, child self-regulation outcomes, and executive function outcomes). The magnitude of association between immigration threat to family and child mental health and child self-regulation outcomes differed by child outcome measures (threat-by-outcome domain interaction terms were significant, p < .05).

Post hoc univariate analyses of the four measures of parent-rated mental health found significant associations with fear- over-anxiousness B(SE)= 1.91 (.76), p < .01 and fear-separation anxiety B(SE)= 2.97 (1.92), p<001, with higher threat associate with more fear in both cases. There was a trend for the association between threat and child sadness (p = .053), and no association with child anger.

Post hoc univariate analyses of the measures of self-regulation outcomes found significant associations with parent rated attention and assessor rated attention. For attention, higher threat to family was associated with lower parent rated attention B(SE)= −4.25 (1.77), p < .05, but higher assessor rated impulse control B(SE) 3.10 (1.41), p < .05 (see Table 2).

Table 2.

Association Between Immigration Threat and Child Mental Health and Self-regulation

	Mental Health				Self-Regulation
Main Effect	Anger	Sadness	Overanxious	Separation Fear	Emotion Regulation (P)	Emotion Regulation (T)	Attention (P)	Attention/ Impulse Control (A)
	B (SE)	B (SE)	B (SE)	B (SE)	B (SE)	B (SE)	B (SE)	B (SE)
PIPES-Total	1.28 (0.78)	1.58 (0.85)+	1.91 (0.76)**	2.97 (0.92)***	-2.25 (1.53)	2.87 (2.18)	-4.25 (1.77)*	3.10 (1.41)*

Note. Table presents the results of post-hoc analyses, which followed the multivariate analyses. Consistent with the multivariate models, all analyses control for five covariates: (ethnicity: non-Latinx Black, 0/1 dummy code, Latinx, 0/1 dummy code), first generation immigrant (0/1), parental education (1=post high school/HS, 0=HS or less) and child sex (1=girl, 0=boy). (P) = parent report, (A)= assessor report, (T) = teacher report.

There were no significant associations between immigration threat to family and executive function.

Moderation by child gender and immigrant status

Across all domains, we found no significant moderation by child gender or respondent immigrant status (all p-values ≥.05).

Discussion

The present study adds to our understanding of immigration-related threat experienced by immigrant and US-born families of Latinx and non-Latinx ethnicity with children enrolled in pre-K. This is the first study to examine the extent to which immigration enforcement threat is associated with mental health, self-regulation and executive function of children in pre-K. We find significant variability in parental perceived immigration threat, with immigrant parents reporting significantly greater levels of immigration threat to family and to child compared to U.S.-born parents. We also find significantly greater levels of immigration threat endorsed by Latinx parents relative to non-Latinx parents. The higher level of perceived threat among immigrants is unsurprising, given multiple restrictive immigration policy changes since 2017, the 45^th President’s focus on immigration enforcement (Pierce & Bolter, 2020) and use of anti-immigrant rhetoric (Canizales & Vallejo, 2021; Ndulue, 2019; Newman et al., 2021). The higher rates of threat endorsed by Latinx parents (versus non-Latinx parents) may be reflective of immigration enforcement activity and rhetoric as Latinxs have been disproportionately profiled by immigration enforcement officers (Golash-Boza & Hondagneu-Sotelo, 2013; Menjívar et al., 2018). Moreover, appealing to white nationalist racism (Canizales & Vallejo, 2021; Newman et al., 2021), the 45^th President announced his presidential bid in 2016 by singling out Latinxs with denigrating rhetoric about immigration and building a wall at the U.S. - Mexican border (Nacos et al., 2020; Washington Post, 2015). These actions increased unease in many Latinx households (Hswen et al 2020; Lopez et al., 2018; Roche et al., 2018).

In examining whether immigration threat to family is associated with child well-being, we find that greater threat to family was associated with child separation anxiety and over-anxious behaviors. The association with two aspects of anxiety in young children is consistent with studies documenting an association between worry about family unity and safety due to the immigration climate and adolescent anxiety (Capps et al., 2020; Eskenazi et al., 2019; Roche et al., 2021) and extends this association to an earlier developmental period. Prior studies have shown that fear of separation from a primary caregiver can result in children having difficulty concentrating and the development of low self-esteem (Allen et al., 2015; Gulbas et al., 2016, Zayas et al., 2015). Children experiencing threat of separation and environmental stressors may be more prone to psychosocial challenges resulting in feelings of lack of control, which can contribute to the early development of childhood anxiety (Chorpita & Barlow, 1998). Findings from this study add to a growing literature indicating a significant link between immigration enforcement threat and anxiety in youth.

Consistent with hypotheses, parents who experience higher levels of immigration threat to family report lower levels of self-regulation in their young children. Specifically, threat was associated with more attention problems. This finding is consistent with literature on the role of threat and violence in shaping children’s self-regulation (Blair & McKinnon, 2013; Blair & Raver, 2012; McCoy et al., 2015). The developmental psychobiological and experiential canalization models of self-regulation postulate that because of the experience of threat, children adaptively expend more resources in monitoring their environments for potential threats, leaving fewer resources and opportunities for the development of self-regulation skills. Chronic activation of neuroendocrine, neurocognitive and attentional systems may shape the development of self-regulation by altering stress response systems and patterned behaviors in a way that is more reactive rather than consciously controlled or prototypically well-regulated (Blair & McKinnon, 2013). In line with this hypothesis, additional work with a subsample of 63 children in the current study who also completed the dot probe task (an assessment of selective attention to emotional stimuli) demonstrated a trend for immigration threat to be related to child hypervigilance (Barajas-Gonzalez et al., 2020).

The positive association between threat to family and assessor report of child self-regulation was not hypothesized, but not entirely surprising given the specific type of threat faced by children and the strategies that parents may be teaching their young children to protect the family. These results may reflect that children feel differently – or that they learn to behave differently – in and outside of their home (Horner et al., 2014; Suárez-Orozco et al., 2011; Zayas, 2015). Interviews with undocumented immigrant mothers indicate that they socialize their children to “stay under the radar” (Zayas, 2015) as a form of protection, “Well, I tell them, ‘try not to get involved in any problem and also always be aware of what is happening around you,’ because as I said last time I am very afraid of what may happen in the schools or anywhere, there’s a lot of evil now and … and I think it’s because of this man [45th President] that he’s been in charge of sowing a lot of hatred against us. And unfortunately, because of our appearance, we tend to be the target of many perverse people.” Rendón García, p. 294). Thus, children may feel they need to be hyper-compliant outside the home so as to not call attention to themselves – and by extension, to their family members (Horner et al., 2014; Suárez-Orozco et al., 2011; Zayas, 2015). This may be particularly true when children are asked to interact with an unfamiliar adult as was the case during the direct assessment upon which assessors’ ratings of attention/impulse control were based. Furthermore, extrapolating from studies of adults which have shown that self-presentation uses self-regulatory resources and makes subsequent self-regulation more difficult (Pachankis, 2007; Vohs et al., 2005), children’s self-regulatory abilities may be vulnerable to depletion such that when children must actively present themselves as highly self-regulated (i.e. impulse control) during the school day, they have fewer resources to continue to self-regulate in the face of stress at home.

Contrary to our hypothesis, we did not find an association between immigration threat and executive function. Our results are consistent, however, with past empirical work which found no relation of community violence to performance on direct assessments of executive function in preschoolers (Sharkey et al., 2012). One possibility is that concurrent associations are not detectable in early childhood, but the influence of immigration threat will exert its impact on executive function over time, especially if threat persists (Barajas-Gonzalez et al., 2021). Indeed, in line with the model of experiential canalization, chronic exposure to threat can heighten reactivity in stress response systems which impairs executive function (Blair & Raver, 2013) and the fundamental brain architecture built in the early years of life (Shonkoff et al., 2012).

Strengths and Limitations

The present study has several strengths, including a diverse sample of Latinx and non-Latinx immigrant and U.S. born families, the use of multiple reporters and sources of data to assess indicators of child well-being, a comprehensive battery to assess self-regulation and executive functioning in pre-K, and the use of a validated measure to capture a more detailed understanding of parents’ fears and worries regarding immigration policies and enforcement threat (Ayón, 2017). Nonetheless, the cross-sectional nature of this study is a limitation; this study is correlational and causal inferences cannot be made. Future studies that examine developmental processes over time are needed to better understand the developmental implications of immigration threat in early childhood on mental health, self-regulation and executive function. It is possible that school climate, and quality of relationships with teachers in particular, may exacerbate or ameliorate the impact of immigration threat on child outcomes, as school climate has been found to moderate the association between neighborhood violent crime and children’s test scores (Laurito et al., 2019). Future studies are encouraged to include indicators of school climate in their understanding of the association between immigration enforcement threat and child well-being.

Another limitation is our reliance solely on parent report for our assessment of child mental health symptoms. It is possible that parent ratings of children's mental health are influenced by parents' own well-being. Future studies with multiple informants on child internalizing and externalizing behaviors are recommended.

Although we observed consistent patterns of perceived immigration threat across four groups (i.e., Latinx immigrant, Latinx non-immigrant, non-Latinx immigrant, non-Latinx, non-immigrant), with Latinx immigrant families demonstrating the highest level of perceived threat, the number of families in some subgroups was too small to draw any firm conclusions. Future studies with larger samples are needed. The small number of non-Latinx Black and Asian participants in this study precluded our ability to examine the impact of immigration enforcement threat within these groups. Future research with larger samples of non-Latinx Black and Asian families is recommended given the disproportionate detention and deportation of Black immigrants (Ibrahim, 2020; Palmer, 2017) and the association between immigration enforcement threat and mental health in South Asian immigrant families (Barajas-Gonzalez, Huang et al., 2021). Future research with larger samples of Latinx families is also needed to examine the impact of immigration enforcement threat among different Latinx subgroups, as data indicates that Mexican and Central American communities are especially impacted by racialized immigration policies and enforcement (Gómez Cervantes, 2019; Menjívar et al., 2018).

This study took place in NYC, which has considerable racial, ethnic and linguistic diversity and a long history with immigrant populations. Moreover, NYC has several policies in place to support immigrant communities, such as limiting how city law enforcement and correction facilities can work with ICE and prohibiting when ICE and non-local law enforcement can enter schools. As such, study findings may underestimate the implications for children in communities where police officers perform immigration law enforcement (Dreby, 2015; Vargas et al., 2017).

Nonetheless, the present study is significant in bringing attention to the association between immigration policies and enforcement practices and young children’s well-being during times of heightened enforcement. In doing so, this study expands the literature on the influence of context on children’s mental health, self-regulation and executive function, beyond the context of poverty (Raver & Blair, 2020), to examine an additional context shaped by restrictive immigration policies, which function as a form of structural racism (Gee & Ford, 2011; Marks et al., 2018). This context, immigration enforcement threat, disproportionately impacts children in immigrant and Latinx households. Consistent with the literature on the spillover effects of immigration policies and enforcement practices, we find immigration enforcement threat is adversely associated with the well-being of U.S. citizen children. Findings implicate immigration enforcement threat as a contributor to young children’s anxiety and attention regulation in pre-K.

Educators, social workers and mental health practitioners working with immigrant and Latinx students in disinvested neighborhoods should be aware of the disproportionate stress experienced by immigrant and Latinx families due to a xenophobic, sociopolitical climate, restrictive changes in immigration policies, national anti-immigrant rhetoric and increased immigration enforcement activity in the U.S. interior that disproportionately targets communities of color. Children in pre-K may have unmet mental health needs, given the manifestation of immigration related fears as internalizing (e.g. anxiety) behaviors rather than externalizing behaviors (e.g. anger), which may go undetected in schools (Eklund et al., 2020). Educators and practitioners working with children from immigrant and Latinx households are encouraged to use trauma-informed approaches to strengthen supports for children’s social-emotional skills to cope with stressors beyond their control, minimize exposure to unnecessary stressors in the classroom, and build relationships that may serve to bolster families in some small way as they continue to weather a restrictive immigration climate (Miller et al. 2019). This means listening deeply and affirming children’s feelings, including their fears; teaching children simple feeling words and how to recognize feelings in their minds and bodies; and extending basic social emotional lessons to create culturally-affirming opportunities for children to share their experiences through stories, art and play (Communities for Just Schools, 2020). Further, creating safe, nurturing and predictable classrooms is important for all young children, and serves an essential protective function for children experiencing threat by soothing the physiological stress response that can otherwise impede children’s learning and development (Ursache et al., 2012). When teachers engage in humanizing practice (Doucet, 2017) children feel seen, heard and cared for; they believe that they and their families matter; and they develop a sense of self and belonging in community that may mitigate the impact of trauma that face outside of school (Adair & Colegrove, 2021). Education stakeholders are encouraged to prioritize funding for more school-based mental health professionals and supports for these professionals (Barajas-Gonzalez, 2019, 2021). We urge policymakers to support healthier immigration policies that center children’s well-being.

Public Policy Relevance:

Findings indicate that immigrant and Latinx families experience disproportionate worry related to the racialized sociopolitical climate, changes in immigration policies, national anti-immigrant rhetoric and increased immigration enforcement activity in the U.S. interior. Children in pre-Kindergarten may have unmet mental health needs given the manifestation of immigration related fears as internalizing behaviors, rather than externalizing behaviors, which may go undetected in schools.