Childhood externalizing behaviors

Childhood externalizing behaviors include negative behaviors, such as aggression, defiance, inattention, hyperactivity, disruption, hostility, and poor impulse control, stemming from a lack of behavior regulation [6,7]. These behaviors are some of the most common reasons for referral to child and adolescent mental health services [8,9]. Addressing externalizing behavior in childhood is a public health priority not only due to its prevalence, but also due to its complex comorbidities and associated outcomes later in life [7]. Children with early onset disruptive behaviors tend to experience more detrimental outcomes in their adulthood including higher rates of antisocial behavior, mental health problems, and healthcare services utilization [10]. Externalizing behavior is relatively stable, resistant to treatment [7], and is more common in boys than in girls across developmental stages [11,12]. Although less common, girls with disruptive behaviors are at an increased risk for depression, self-harm, post-traumatic stress disorder, substance abuse, interpersonal aggression, early and risky sexual behavior, and lower academic achievement [10].

Neighborhood collective efficacy

Despite the increases in the prevalence of child and adolescent mental health and behavioral disorders in the US, the role of larger structural and societal influences on externalizing behaviors is scarcely examined in the current literature despite an increased focus on examining the role of social determinants of health (SDOH) as drivers of health outcomes in recent years. Comprising of both built environment (e.g., green spaces and zoning ordinance) and social environment (e.g., neighborhood social processes) [13,14], neighborhood environment is a potent SDOH that affects health outcomes in children spanning chronic illnesses (i.e., asthma and obesity), life expectancy [15–18] and psychological health [15,19–21].

In this sense, neighborhood collective efficacy (NCE) is an essential component of the neighborhood social environment that has not always been included in existing literature. A scoping review of neighborhood effects on early child development identified that most of the studies conceptualized neighborhood effects as neighborhood poverty (or socio-economic status), describing the association between neighborhood poverty and mental health outcomes in children [22]. However, NCE--defined as “social cohesion among neighbors combined with their willingness to intervene on behalf of the common good” [23]--is seldom included in the construct of neighborhood. NCE is a structural social-interactive mechanism in which neighborhoods with high levels of collective efficacy possess high social cohesion amongst its residents who are willing to fight to protect communal facilities and access to healthcare, maintain the neighborhoods against neglect, destruction, or abandonment [22–24]. Furthermore, NCE encompasses the sense of community held by community members, which captures residents’ sense of membership, influence, fulfillment of needs, and shared emotional connection to their neighborhood [25]. NCE has shown to be significant predictor of neighborhood crime and youth delinquency [23,26]. Based on these early findings, NCE has been further applied to various child health outcomes such as childhood obesity and adolescent mental health [27,28]. A recent systematic review evaluated the interaction between individual racial/ethnic identity and neighborhood context in differentiating developmental trajectories of conduct problems, and found that 2 studies (out of 18) revealed significant effects of NCE, such that high collective efficacy was associated with a decrease in delinquency from ages 12–15 [29]. However, to our knowledge, this is one of the first systematic reviews to evaluate the existing evidence specifically linking NCE and child externalizing behaviors across developmental trajectories.

Given that NCE is a significant, yet sometimes overlooked aspect of the neighborhood environment that may have implications for externalizing behaviors in children and adolescents, the goal of this manuscript is to conduct a systematic review to examine the role that NCE plays in child and adolescent externalizing behavior outcomes to inform practice and policy decisions. The overarching research question for this review was: What is the relationship between NCE and child externalizing behaviors across developmental trajectories? Given the early literature, we hypothesized that neighborhoods with higher NCE will be associated with decreased prevalence and/or severity of child externalizing behaviors throughout children’s developmental trajectories.

Study inclusion and exclusion criteria

This study followed a set of predetermined inclusion criteria. Criteria for inclusion were: 1) use of quantitative design to assess the relationship between NCE and child/adolescent behavioral outcomes; 2) publication in a peer-reviewed journal; 3) sample that included children and adolescents (defined as under 21 years of age); the 4) measurement of NCE using the measures, or a variation of the measure, created by Sampson and colleagues, which is the only validated assessment that includes social control, social cohesion/trust, [23] and 5) measurement of externalizing behaviors using a reliable and validated measure. No time or language restrictions were applied for a comprehensive search. The search was conducted in November 2024. Several exclusion criteria were applied to this study to ensure a high-level systematic review. Meeting abstracts, dissertations, pre-prints, editorials, and commentaries were excluded. Experimental studies were excluded given that they do not address the standard condition of the neighborhood and children’s externalizing behaviors in their natural state as desired.

Search strategy

We identified manuscripts via database searches through PubMed, PsycINFO, and CINAHL. The search string is presented in a supplemental file (S1 File). Identified publications were managed using Covidence and Zotero softwares.

Study selection and data extraction

From employing the predetermined search strategy described above, we identified 294 manuscripts after 149 duplicates were removed. J.L. and H.P. independently screened the titles and abstracts of the manuscripts for relevancy guided by the pre-determined inclusion and exclusion criteria (Fig 1). After title and abstract review, 164 manuscripts were removed as they did not adequately answer the research question. Then, J.L. and H.S. conducted full-text reviews of the remaining 130 manuscripts with documented reasons for either inclusion or exclusion. Of the 130 manuscripts, 113 studies were excluded after full text review. All papers identified in the search and the rationale for exclusion are outlined in S1 Table. Furthermore, the reasons for exclusion are also detailed in Fig 1, which shows that the most common reason for exclusion at this stage was due to not addressing the outcome variable of interest (N = 83) and not utilizing NCE as the predictor variable and externalizing behavior as the outcome variable (N = 24). Notably, we identified one study that was published in Japanese. The full text was verified by the second author of this study who is fluent in Japanese. This study was ultimately excluded due to the age of the participants. Finally, 17 articles met the criteria for final extraction. J.L. extracted the 17 studies onto a template which included author name, publication year, purpose, sample, setting, measure of NCE and its psychometric properties, measure of externalizing behavior and its psychometrics, type of analysis, covariates, and results (Table 2). H.S. reviewed the extraction chart for accuracy and thoroughness.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. The National Heart, Lung, and Blood Institute (NHBLI) Quality Assessment Tool for Observational Cohort and Cross-sectional studies.

https://doi.org/10.1371/journal.pone.0337512.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Extraction Chart: Summary of studies included in the review (n = 17, sample size = 28,957).

https://doi.org/10.1371/journal.pone.0337512.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA flow chart diagram.

https://doi.org/10.1371/journal.pone.0337512.g001

Quality assessment

We assessed study quality using the guidelines from the National Heart, Lung, and Blood Institute (NHBLI) Quality Assessment Tool for Observational Cohort and Cross-sectional studies (Table 1) [31].

Study characteristics

Among the 17 studies included in this review and in the extraction chart (Table 2), the sample sizes ranged from 243 to 4,898 with an average sample size of 1,738 (SD = 1,465). Most samples (N = 16) were comprised of caregiver-child (or children) dyads, with one study which included only adolescents without their caregivers. Most studies utilized participants from the US (13 out of 17) in addition to studies from the United Kingdom, Japan, Sierra Leone, and Vietnam. Fifteen out of 17 studies were published within the past 10 years (9 were published within the past 5 years, as of 2024). There was no missing data in the included studies.

Study designs included 12 longitudinal studies and 5 cross-sectional studies. All studies used Neighborhood Collective Efficacy scale created by Sampson and colleagues to quantify NCE, which is comprised of two subscales: informal social control and social cohesion/trust. The informal social control subscale includes 5 Likert-type items that ask respondents about the likelihood that their neighbors could be counted on to intervene in various ways for the greater good of the neighborhood. The social cohesion and trust subscale also includes 5 Likert-type items in which respondents are asked about how strongly they agree that the neighbors are willing to help each other out and can be trusted [23]. Most studies (13 out of 17) used the CBCL to measure externalizing behavior. The rest of the manuscripts used Oxford Measure of Psychosocial Adjustment (OMPA) [32], Social Skills Rating System (SSRS) [33], Youth Self Report (YSR) [34], and questions adopted from the National Longitudinal Study of Adolescent Health [35]. While similar with CBCL, YSL is designed to be answered by adolescents rather than their parents or caregivers [36]. Finally, it is important to note while all studies studied externalizing behavior outcomes, they used varying terms to describe the same concept (i.e., “antisocial behavior”, “delinquency”) [37].

Data synthesis

Most of the included studies (Number (N) =15 of 17) found a significant relationship between NCE and externalizing behaviors (Table 2). We grouped key findings from selected studies, into the following themes: 1) Externalizing behavior outcomes by developmental stages (early childhood, middle childhood, and adolescence); 2) Contextual factors (moderating, mediating) that contribute to the strength and directionality of the relationship between NCE and externalizing behaviors.

1. 1) Developmental stages

We define early childhood as ages 12 months to 5 years 11 months and 30 days, middle childhood as 5 years to 11 years 11 months and 30 days, and adolescence as 11 years to 20 years 11 months and 30 days. We followed the age windows created by Environment Influences on Child Health Outcomes (ECHO) Program by the National Institutes of Health (NIH) [38]. According to ECHO, the life stages intentionally overlap given that there is age-variability in normal maturation and development. If the included studies were comprised of participants who overlapped between age groups, we categorized the studies by the age group that included the majority of participants.

Early childhood (12 months to 5 years 11 months 30 days)

In this systematic review, all 5 studies that examined the relationship between NCE and externalizing behaviors in early childhood yielded significant associations [39–43]. The included studies in this age group were each evaluated for its quality of evidence. While there some differences across the 5 studies in whether NCE was measured prior to CBCL and whether NCE was assessed more than once over time, all studies were deemed to be of high-quality evidence in their clear statement of objectives, defined study populations, and sufficient timeframe to observe an association between NCE and externalizing behaviors. Notably, all studies in this age group were derived from the Future of Families and Child Wellbeing Study (FFCWS; formerly Fragile Families), which is a prospective birth cohort of 4,898 children born between 1998 and 2000 and their parents, spanning 20 large, urban cities in the US. Specifically, the studies examined children at 3 and 5 years of age (represented by Wave 3 and 4 from the FFCWS cohort, respectively). Given the strength of the FFCWS cohort, some studies examined the relationship longitudinally. For instance, Ma and Grogan-Kaylor examined the association of NCE with longitudinal patterns of early externalizing behavior problems (N = 3,705) [41]. The longitudinal multilevel model showed that the average level of externalizing behavior significantly declined from age 3–5 (β = –0.069, p < .001), and NCE had a noticeable inverse association with differences in externalizing behavior at mean age, net of covariates (β = –0.04, p < .001); however, some studies examined the relationship between NCE and externalizing behavior cross-sectionally. For instance, Ma et al. examined direct associations between mothers’ perception of NCE and children’s behavior (N = 2,388) assessed at Wave 4 [42]. The study found that perceived NCE and externalizing behavior problems at age 5 demonstrated small but significant inverse relationship (r = −.14, p < .001). In addition, higher values of NCE were associated with lower externalizing behavior (β = −0.01, p = .04). Finally, Pei et al. aimed to capture the specific pathways from neighborhood structural factors and process factors to early childhood externalizing symptoms (N = 2,722) [43]. NCE was collected at Wave 3 and early childhood behavioral problems were measured at Wave 4. The analysis showed that social cohesion had both direct and indirect effects on early childhood behavioral problems (β = −0.11, p = .001) while social control did not have any significant effects on early childhood behavioral problems. In summary, all studies included in this age group were of high quality evidence and demonstrated significant relationship between NCE and child externalizing behaviors, both cross-sectionally and longitudinally.

Middle childhood (5 years to 11 years 11 months and 30 days)

Six studies [33,37,44–47] focused on the relationship between NCE and externalizing behavior in middle childhood. Three of the studies in this category used longitudinal approaches [37,44,45] while the rest [33,46,47] (n = 3) employed cross-sectional approaches. The study population in this developmental stage spanned the U.K., Japan, US, and Vietnam. Most studies described significant relationship between NCE and child externalizing behavior in middle childhood, while one study did not show significant relationship [45]. The included studies in middle childhood were also each evaluated for its quality of evidence. While there some differences across the studies in the rate of participation and whether NCE was assessed more than once over time, all studies were deemed to be of high-quality evidence in their inclusion of clear statement of objectives, defined study populations, and the inclusion of key covariates.

A significant longitudinal relationship between NCE and child externalizing behavior was shown in Ichikawa et al. in which they examined the longitudinal association between change in NCE and change in children’s psychosocial development in Japanese children (Age range: 4−17 at Wave 1; N = 918) and found that social cohesion was associated with a decrease in child total problem score (β = −0.22; 95% CI: −0.37 to −0.001; d = −0.03), while informal social control was associated with lower externalizing problems (β = −0.16; 95% CI: −0.30 to −0.03; d = −0.02) [44]. A significant cross-sectional relationship between NCE and child externalizing behavior was shown in the study by Moren-Cross et al., in which the relationship between maternal subjective neighborhood attributes and child behavioral problems was examined in third grade children (N = 576) [33]. Specifically, they found that neighborhoods lacking in NCE were associated with children with elevated problem behaviors (β = 0.094, p < .05). Finally, Browning et al. investigated whether NCE had direct effects on mental health outcomes and whether NCE buffers children against mental health consequence of life-threatening community violence [45]. They found that NCE did not show main effects on externalizing behaviors in both boys and girls, but it functioned as a buffer against mental health consequences of girls’ exposure to life-threatening violence. Overall, the studies included in this age group were of high quality evidence and mostly demonstrated a significant relationship between NCE and externalizing behaviors in middle childhood.

Adolescence (11 years to 20 years 11 months and 30 days)

Finally, 6 studies [32,34,35,48–50] examined the relationship between NCE and externalizing behavior in adolescence. Four studies examined the relationship longitudinally [32,35,48,49], while two did so cross-sectionally [34,50]. The adolescent population also varied greatly spanning former child soldier adolescents in Sierra Leon [32], African American adolescents in the US [34,50], and youth from the FFCWS cohort [35,48,49]. Of the 6 studies, 5 demonstrated a significant relationship between NCE and externalizing behaviors in adolescence, while one study did not show a significant relationship. The included studies in adolescence were also each evaluated for its quality of evidence. While there also differences across the studies in the rate of participation and whether NCE was measured prior to externalizing behaviors, similar to other developmental stages, all studies were deemed to be of high-quality evidence in other critical areas including clear statement of objectives, defined study populations, inclusion of key covariates, and clearly defined outcome variables that were applied consistently across all participants.

Several studies have highlighted the significant longitudinal relationship between NCE and externalizing behaviors in adolescence. Pei explored this relationship using data from the FFCWS cohort. The study examined NCE collected from caregivers in early childhood (at age 3) and externalizing behavior outcomes measured at ages 3,5,9, and 15 using parallel-process growth curve modeling, and found that children living in neighborhoods with a higher level of social cohesion had lower initial levels of externalizing symptoms (β = −0.14, 95% CI = −0.19, −0.09, p < .001), which was eventually positively associated with a steady decrease in externalizing symptoms into adolescence (β = 0.06, 95% CI = 0.01, 0.12, p = .03) [49]. Hardi and colleagues, on the other hand, aimed to test the cumulative and specificity effects of adversity across various developmental periods to predict internalizing and externalizing behaviors in youth. While low NCE at age 9 (r² = 0.049%) and cumulative effect (r² = 0.083%) did not significantly predict externalizing behaviors, the cumulative effect of all types of adversity (including low NCE) significantly predicted youth externalizing behaviors (r² = 1.16%, p < .001) [35]. While both studies used data from the FFCWS cohort, how they operationalized externalizing behaviors, their sample sizes, study aims, and methodological approaches differed significantly. Despite these differences, both studies demonstrate the longitudinal relationship between NCE and externalizing behaviors in adolescence. Conversely, Betancourt and colleagues examined community-level characteristics in relation to externalizing behaviors among former child soldiers in post-conflict Sierra Leone longitudinally (N = 243; Mean age = 16.6) [32]. The study did not yield any significant relationship between NCE and externalizing behavior (β = 0.01, p > .05).

Finally, Sterrett-Hong and colleagues examined the influence of neighborhood conditions on behavioral difficulties among low-income African American adolescents cross-sectionally (N = 524, Mean age = 15.89, SD = 1.42). Notably, this was the only study to use both adolescent-reported and parent-reported outcomes, and found that social cohesion was significantly related to adolescent-reported hyperactivity and inattention (β = −0.09, p < .001), and higher social cohesion showed significantly lower levels of adolescent-reported conduct problems (β = −0.12, p < .001) [50]. Interestingly, the same study found that parent-reported measure of NCE was not a significant predictor for hyperactivity/inattention and conduct problems in adolescence. In summary, the studies included in this age group were of high quality evidence, and mostly revealed a significant relationship between NCE and externalizing behaviors in adolescence.

1. 2) Contextual factors

There were several contextual factors highlighted in the studies pertaining to the relationship between NCE and externalizing behavior, including corporal punishment, race and ethnicity, adverse childhood experiences (ACEs), and parenting stress.

Corporal punishment

There were 4 studies that examined the effect of corporal punishment in the context of NCE and externalizing behaviors [39–42]. Castillo and colleagues examined whether the relationship between NCE and child externalizing behavior was mediated by mothers’ use of corporal punishment and found a significantly mediated relationship (β = −0.11, p < .01) [37]. Ma and Klein found that, in a random intercept model exploring simultaneous effects of perceived NCE and maternal corporal punishment on externalizing behaviors on 5-year-olds, higher NCE was associated with lower externalizing behaviors (β = −0.01, p = .040) and increased frequency of corporal punishment predicted higher levels of externalizing behaviors [42]. On the contrary, 2 studies found was that maternal spanking was positively associated with behavior problems but did not find that the associations of low NCE with externalizing problems were mediated by maternal spanking [40,41]. While these studies demonstrated the direct effect of corporal punishment on child externalizing behavior problems, we found limited evidence for the mediating effect between NCE and behavior problems in children.

Race and ethnicity

Ma and colleagues examined whether race and ethnicity moderate the association between neighborhood and early behavior problems [42]. They found that Black and Hispanic mothers reported lower levels of NCE compared to their White counterparts. They also found that perceived NCE and externalizing problems at age 5 demonstrated small but significant inverse relationships. To examine the moderating effect of race/ethnicity on this relationship, they examined the simple NCE slopes for each racial and ethnic groups. Results indicated that race and ethnicity (categorized as White, Black, and Hispanic) did not moderate the association between NCE and externalizing behavior in this study.

Adverse childhood experiences (ACEs) and parenting stress

Finally, Wang and colleagues examined the longitudinal associations between NCE and later adolescent outcomes [48]. Specifically, they examined if ACEs and parenting stress mediate the association. The study found that NCE was significantly related to mothers’ parenting stress (β = −0.26, p < .001), which was positively related to adolescent’s delinquency and behavior problems (β = 0.06, p < .05; β = 0.14, p < .001, respectively). In addition, lower levels of NCE were associated with more ACEs (β = −0.07, p < .01), which were associated with increased levels of delinquency and behavioral problems (β = 0.08, p < 0.001; β = 0.15, p < .001, respectively).

Clinical implications

There are many persistent challenges in addressing the current childhood mental and behavioral health crisis in the US. For one, shortage of child psychiatrists in the US is well documented [63]. This is especially concerning given that the shortage tends to be concentrated in rural and high-poverty areas in the US. Adding to the concern, Hoffmann and colleagues found that areas with mental health workforce shortages were associated with an increased youth suicide rate [63]. Given the demonstrated mental health workforce shortage and other related barriers, mental health providers may not have the means and the resources to address how neighborhood-level barriers, such as NCE, may be play a role in child behavioral health concerns. However, given the growing evidence that structural level characteristics affect child behavioral health outcomes, child mental and behavioral health providers should consider these as potential targets for interventions. Indeed, there has been increasing recognition amongst clinicians to focus on primary prevention of pediatric mental health crisis by engaging community stakeholders including the school systems and community mental health therapists [64].

In addition to the issues related to the mental health workforce, there has been limited evidence for interventions developed and tested to improve neighborhood collective efficacy in disadvantaged communities. One example of such intervention was conducted by Ohmer (2016) in which a pilot community-based intervention to facilitate collective efficacy among youth and adult residents was administered in the southeastern United States. The intervention included community organizing and mobilization, a training program to facilitate collective efficacy, and a community-based project developed by participants to address a youth violence prevention issue, and it resulted in increased levels of collective efficacy reported by the participants [65]. Building on the promising effects of interventions targeting NCE in disadvantaged communities, we recommend fostering collaborations among clinicians, policymakers, researchers, and community representatives. Such partnerships are essential for adapting and scaling these interventions to better address the behavioral and mental health needs of children and youth in the community.

Finally, this systematic review revealed the strongest evidence between NCE and child externalizing behaviors in early childhood. Therefore, researchers, clinicians, and child mental health workers should focus on this important early developmental period as a window of opportunity to find solutions to close the gap on health inequities later in life [56].

Limitations

Due to limited number of studies to date examining the relationship between NCE and child externalizing behavior outcomes, this systematic review included only 17 studies, 5 of which employed cross-sectional approaches. Furthermore, amongst the longitudinal studies (12 studies), 2 of them examined the variables of interest cross-sectionally. The heterogeneity of the study designs and the diverse measures used to operationalize externalizing behaviors prevented the possibility of conducting meta-analysis.

While most studies utilized CBCL to measure the degree of externalizing behaviors, 4 studies used different measures (YSR, SSRS, OMPA, and questions adopted from the National Longitudinal Study of Adolescent Health) which could have affected the outcome analyses given the inherent differences in the measures in quantifying externalizing behaviors.

Of the included studies, 8 employed samples from the FFCWS cohort, which includes racially diverse sample of low-income families, with racial and ethnic minorities and unmarried mothers being overrepresented [64]. It should be noted that the overlap in the study population could have been a possible source of bias in the study outcome.

Additionally, internalizing and externalizing behaviors are typically correlated with one another, and thus, can make it challenging to draw the conclusions that NCE is associated only with externalizing behavior outcomes without controlling for the associations with internalizing behaviors [66].

The systematic review included studies published throughout the world, but most studies (n = 13) resulted from the US. Literature shows that compared to other high-income countries, the US represents one of the highest geographic health disparities alongside one of the worst other health-related outcomes and inequalities in developed nations [67,68]. Given that there was scant representation of other continents and countries in this systematic review, there could have been more pronounced bias towards finding significant relationship between NCE and externalizing behaviors given the over-representation of participants from the US. Given these limitations, the findings may not be as generalizable outside of the US.