Next Article in Journal
When Help Hurts: Moral Disengagement and the Myth of the Supportive Migrant Network
Previous Article in Journal
Are Undergraduate Students Enrolled in Educational Degrees Concerned About Gender? A Case Study
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Social Sciences
Volume 14
Issue 6
10.3390/socsci14060385
socsci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Changes in Aggressive Behaviors over Time in Children with Adverse Childhood Experiences: Focusing on the Role of School Connectedness

by
Sei-Young Lee
1,* and
Mijin Choi
2
1
Social Work Department, University of Northern Iowa, Cedar Falls, IA 50614, USA
2
School of Social Work, Texas State University, San Marcos, TX 78666, USA
*
Author to whom correspondence should be addressed.
Soc. Sci. 2025, 14(6), 385; https://doi.org/10.3390/socsci14060385
Submission received: 14 April 2025 / Revised: 6 June 2025 / Accepted: 7 June 2025 / Published: 17 June 2025
(This article belongs to the Special Issue The Role of Early Childhood Experiences in Shaping Criminal Offending: Implications for Intervention and Prevention)
Browse Figures
Versions Notes

Abstract

:
Previous research has found the link between adverse childhood experiences (ACEs) and externalizing behaviors later in adolescence and adulthood. However, limited studies have explored longitudinal trajectories of aggressive behaviors affected by ACEs. This study aimed to investigate how aggressive behaviors change over time and compare the trajectories of aggressive behaviors between children with three or fewer (low-risk ACEs) and those with four or more ACEs (high-risk ACEs) with the role of school connectedness on the trajectories of aggressive behaviors over time. The study sample consisted of 4231 children collected by the Future of Families and Child Wellbeing Study, a longitudinal birth cohort study focusing on children in high-risk families across 20 U.S. cities. The mean age of the sample was 15.59 years, with 81.9% of the participants being minorities and 51.9% being boys. The results indicated that children with high-risk ACEs showed twice as high levels of aggressive behaviors as those with low-risk ACEs. School connectedness decreased the starting level and change rate of aggressive behavior for children with high-risk ACEs, while it decreased the starting level of aggressive behavior for those with low-risk ACEs. These findings underscore the protective role of school connectedness on aggressive behavior. Practitioners and policymakers need to make an effort to build safe and supportive social environments for all children, especially for children with high-risk ACEs.

1. Aggressive Behavior and Its Trajectory over Time

Aggression is a common phenomenon that is normally observed during childhood and adolescence (Baker and Jensen 2024). Aggressive behaviors are most often characterized by actions such as forceful hitting, kicking, biting, or pushing that harm or threaten someone physically or psychologically through interpersonal relationships (Eltink et al. 2018; Hay et al. 2021; Ward et al. 2025). Aggression at an early age has been associated with several negative outcomes emerged in childhood and adolescence, such as school dropout (Broidy et al. 2003), substance abuse (Fite et al. 2007), peer rejection, and peer delinquency (Fite et al. 2008), as well as later outcomes in emerging adulthood such as depressive and delinquency symptoms (Cleverly et al. 2012) and unemployment (Kokko and Pulkkinen 2000).
Other studies have also identified the trajectory of aggressive behavior over time. Baker et al. (2019) reported that aggressive behaviors tend to decrease as children age from toddlerhood to adolescence, likely influenced by cognitive development, family variables, and increased socialization, such as the modeling of appropriate behaviors. Similarly, Duggins et al. (2016) reported a decline in aggressive behaviors over two years among students in grades 7 through 10 at the time of baseline. Inconsistently, in a systemic review of research findings published in four specific academic journals, Piquero et al. (2012) documented some longitudinal studies reporting instability of aggression among children and adolescents. Specifically, children with low rates of aggression exhibited increased aggression in adolescence, while children with high rates of aggression showed decreased aggression in adolescence. These inconsistencies in the trajectories of aggression, such as stability or instability, from childhood to adolescence, highlight the need for further research.
Another study examined how values influence aggression with respect to how they play a preventative role against aggressive behaviors and how this relationship between values and aggression changes over time (Benish-Weisman 2015). According to the study, values emerge through interaction with socializing figures, such as parents, and immersion in social contexts, such as schools. The study introduced two dimensions of oppositional value groups in adolescence: self-enhancement values (i.e., power and achievement) vs. self-transcendence values (i.e., universalism and benevolence) and openness-to-change values (i.e., self-direction, stimulation, and hedonism) vs. conservation values (i.e., conformity, tradition, and security). Values close to each other pursue similar motivations and behaviors, while opposing values represent different motivations and behaviors, reflecting either social and psychological congruency or contradictions between values. Regarding aggression, self-enhancement and openness-to-change values are more likely to be related to aggression than those with self-transcendence and conservation values. Benish-Weisman (2015) also documented several previous studies indicating an association between self-transcendence values and a decrease in aggression over time (Aquilar et al. 2018; Benish-Weisman 2015; see Benish-Weisman 2019, p. 263). This association can be intensified by the social environment, including family and schools, which reinforces socially expected behaviors and emphasizes socially focused values, ultimately leading to reciprocal changes in values.

2. Adverse Childhood Experiences and Aggressive Behavior

Adverse childhood experiences (hereafter, ACEs) indicate negative experiences that occurred prior to age 18, which include child maltreatment (i.e., physical, emotional, and sexual abuse and physical and emotional neglect) and family dysfunction (i.e., living with a family member suffering from mental illness, substance use, or incarceration, and parental separation/divorce and intimate partner violence) (DeLisi and Beauregard 2018; Matsuura et al. 2013; McRae et al. 2021; Oei et al. 2023; Stoppelbein et al. 2024). ACEs have been reported to be closely related to antisocial behaviors, criminal behaviors, violent offending, and aggression later in adolescence and adulthood (Aroyewum et al. 2023; Baglivio et al. 2016; DeLisi and Beauregard 2018; Matsuura et al. 2013; Stoppelbein et al. 2024). Aroyewum et al. (2023) added community violence to ACEs, including physical, sexual, and emotional abuse, family dysfunction, and witnessing community and peer violence, which predicted aggressive behaviors such as physical and verbal aggression, anger, and hostility among undergraduate students.
While there has been limited research on the population in childhood and adolescence, ACEs have also been reported to be associated with aggressive behaviors in childhood or adolescence (Allen 2011; O’Connor et al. 2021). A higher number of ACEs was associated with adolescents who had a moderate-to-high probability of endorsing physical, relational, and cyber aggression, as well as victimization, compared to their counterparts categorized as having a low probability of these behaviors (O’Connor et al. 2021). In particular, Greeson et al. (2013) found a significant dose–response relationship between the number of ACEs and behavioral problems in adolescents, as each additional ACE endorsed a significant increase in behavioral problems. Likewise, high-risk ACEs, which indicate having four or more ACEs, are associated with more negative mental health (i.e., depression, anxiety) and behavioral problems (i.e., delinquency, aggression, and violence) compared to low-risk ACEs with three or fewer ACEs (Song 2023). However, limited studies are available comparing the effects of high-risk and low-risk ACEs on aggression with longitudinal data.
McRae et al. (2021) identified the association between ACEs and proactive/reactive aggression among children and early adolescents aged 6 to 14. In particular, the study reported that child maltreatment ACEs—physical, sexual, and emotional abuse and physical or emotional neglect—were associated with reactive aggression, which can be defined as affective or defensive aggression to respond to external threats rather than proactive aggression, which is more goal-oriented and involves planned aggression. Additionally, post-traumatic stress symptoms, such as disturbances in thoughts and behaviors, mediate the association between child maltreatment and reactive aggression. Similarly, Connor et al. (2004) reported that experiences of sexual abuse and abuse perpetrated by an adult are associated with reactive aggression among adolescents referred to psychiatric care.
Little is known about the long-term trajectories of reactive aggression over time, compared to proactive aggression in childhood that is associated with delinquency and violence later in adolescence (Fite et al. 2008). Research focusing on the association between ACEs and aggression during childhood and adolescence has still been limited, with existing studies often examining predominantly male samples or individuals involved in the juvenile justice system or utilizing cross-sectional data (Asscher et al. 2015; Oei et al. 2023; Stoppelbien et al. 2024). This warrants the need for further research, including studies of the general adolescent population, as well as those utilizing longitudinal research designs.

3. Effects of School Connectedness on Aggressive Behaviors with the Social Development Model

The social development model is a useful framework for understanding aggression in childhood and adolescence. This model originates from criminological theory and explains the role of developmental processes in predicting distinct prosocial and antisocial behaviors by identifying biological, psychological, and social factors across multiple social domains (Catalano et al. 1996; Choi et al. 2005). It incorporates the prediction of risk factors that contribute to the development of antisocial behavior, as well as protective factors that mediate or moderate the effects of risk exposure (Catalano et al. 1996). Individuals learn behavior patterns through various social domains, such as family, school, peer groups, and communities, by engaging in opportunities, developing skills, and receiving recognition for social involvement. They adopt prosocial or antisocial beliefs based on the predominant behaviors, norms, and values of those with whom they have formed bonds, and these beliefs, in turn, determine their behavior (Choi et al. 2005). Thus, protective factors in social environments are critical in the change in aggressive behavior over time.
School is one of the important social domains, which provides children and adolescents with opportunities for interacting with others, developing skills, and being socialized, influencing the formation and adoption of their values and behaviors (Choi et al. 2024; Gilligan 2004), School connectedness, which is commonly defined as “subjective feelings of being connected to the schools, including feeling of closeness, inclusiveness, happiness, and safety in school” (Choi et al. 2022, p. 3), has been reported to be associated with increased prosocial behaviors and decreased internalizing and externalizing problems (Bond et al. 2007; Choi et al. 2022; Choi et al. 2024). In particular, several studies have documented that school connectedness is directly or indirectly associated with adolescent aggression (Duggins et al. 2016; Gale and Nepomnyaschy 2024; Lee et al. 2018). For example, Duggins et al. (2016) indicated that school connectedness was negatively associated with aggressive behaviors, presenting with the cross-sectionally collected data. Furthermore, utilizing longitudinal data including school connectedness collected at the age of nine and aggression collected at the age of fifteen, Gale and Nepomnyaschy (2024) demonstrated a negative association between school connectedness and aggressive behaviors reported at age fifteen among Black adolescents, showing the protective role of school connectedness for reducing the level of aggression.
Additionally, the mediating and moderating roles of school connectedness in protecting against aggression were reported. Tian et al. (2019) reported that adolescents’ perceived level of school connectedness moderated the indirect association between parental psychological control and adolescent aggressive behaviors. Similarly, Wang et al. (2023) documented the mediating effects of school connectedness on the increased aggressive behaviors in adolescents at age 15 that is associated with involvement in the child welfare system (CPS contact), suggesting that the school system may help adolescents as a buffer against the negative experiences of child maltreatment. Furthermore, benefiting from a longitudinal study design, Duggins et al. (2016) found that school connectedness influenced the relationship between school bullying victimization and aggression among middle and high school students, potentially contributing to students’ vulnerability over time. However, research examining the role of school connectedness in the trajectory of aggression over time, particularly while considering the effects of adverse childhood experiences (ACEs), remains scarce.

4. The Current Study

The purpose of the current study was to examine and compare the trajectories of aggressive behavior over time between the two groups of children with three or fewer (low-risk) and four or more (high-risk) ACEs and to examine the effect of school connectedness on the trajectory of aggressive behavior, which is a well-known protective factor for aggressive and delinquent behavior. Applying the Social Development Model, the current study hypothesized that (1) the trajectory of aggressive behavior shows a pattern of decrease over time for both groups of children with low- and high-risk ACEs; (2) the initial levels and the change rates of aggressive behavior are different between the two groups of children with low- and high-risk ACEs; and (3) school connectedness significantly affects the initial levels and the change rates of aggressive behavior for both groups of children with low- and high-risk ACEs.

5. Methods

5.1. Sample and Participants

The present study utilized population-based survey data from FFCWS (the Future of Families and Child Wellbeing Study), which is a longitudinal birth cohort study that followed parents and their newborn children across the seven waves in which children’s age—at birth, 1, 3, 5, 9, 15, and 22—from 1998 to 2022. In particular, 3600 unwed couples and 1100 married couples were recruited from 75 hospitals in 20 U.S. cities, utilizing a stratified multistage sampling method from the U.S. cities with a population of 200,000 or more people (Reichman et al. 2001). FFCWS focused on high-risk families in the U.S. Based on population-based sampling, FFCWS recruited a larger proportion of single/unmarried parents (75%), with the majority identifying as minorities: 48% African American and 27% Hispanic. More than 80% of primary caregivers reported having at least a high school diploma (82.4%) at Year 15 and an average household income of USD 31,900 at baseline.
Due to the primary study goals of examining longitudinal changes in aggressive behavior among children, the current study includes three waves of data collected from Years 5, 9, and 15, comprising 4898 children. The final sample size of this study was 4231, which reported ACEs in any of the three waves. The mean age of the adolescents included in the current study at Year 15 was 15.59 years (SD = 0.77), with a range of 14 to 19 years. A majority of them belonged to racial minorities (81.9%), as 49% identified as African Americans and 24.9% as Hispanics, followed by Whites (18.1%), multiracial (5.4%), and others (2.6%). There were slightly more boys than girls (51.9% vs. 48.1%). As this study used secondary data with a sample of 4231 children, post hoc power analysis was conducted. Post hoc power analysis revealed a power of 1.00, with a set of p-values at 0.05 and an effect size ranging from 2.1 to 2.6.

5.2. Measures

5.2.1. Adverse Childhood Experiences

The eight types of ACEs include four types of childhood abuse and neglect (e.g., physical, emotional, and sexual abuse, and physical neglect) and four types of family dysfunction (e.g., parental substance abuse, domestic violence, mental illness, and incarceration). The four types of child maltreatment were collected from subscales of the Conflict Tactics Scale for Parent-Child (CTS-PC; Straus et al. 1998), which is a well-validated measure of child maltreatment. However, sexual abuse reported by Child Protective Services has been collected since Year 5. Thus, this study included three waves of ACEs at Years 5, 9, and 15. Each type of maltreatment was measured with a 7-item scale (e.g., never happened, once, twice, 3–5 times, 6–10 times, 11–20 times, and more than 20 times). For data analysis, all types of ACEs were dichotomized as exposure to a specific ACE (=1) and otherwise (=0).
Sample questions for physical abuse included how many times the parent had “spanked your child on the bottom with your bare hand in the past year” and “slapped the child on hand, arm, or leg,” for emotional abuse, items included “threatened to spank or hit,” and “called the child dumb or lazy, or some other name like that,” and for physical neglect, items “were so drunk/high that you had a problem taking care of your child,” and “not able to make sure child got to a doctor or hospital when needed” were included.
Parental substance use questions concerned whether mothers or fathers reported having four or more drinks and any of five types of drugs (e.g., cocaine and heroin) or misuse of prescription drugs (e.g., sedatives and tranquilizers) in the past 12 months. Parental incarceration was measured by whether the mother, father, or mother’s current partner had spent any time in prison or jail. Sample questions of parental domestic violence were whether a spouse or intimate partner “tried to keep you from seeing or talking to your friends or family” and “tried to make you have sex or do sexual things.” For parental mental illness, questions that measured parental anxiety and depression were used. Parental divorce and separation were created to indicate that the child’s parents were separated or divorced. The reliability of ACEs across three waves was at a moderate level (α = 0.72). Cumulative ACEs were calculated by summing up the ACE score at each wave. Overall, the mean numbers of ACEs from Year 5 to Year 15 decreased over time with a slight increase at Year 9 (Y5: m = 0.72, sd = 1.14; Y9: m = 0.73, sd = 1.06; and Y15: m = 0.64, sd = 0.96). The study participants were grouped by the number of ACEs they experienced at baseline into two groups: the low-risk ACE group for those who experienced three or fewer ACEs (n = 2926) and the high-risk ACE group for those with four or more ACEs (n = 1305).

5.2.2. School Connectedness

School connectedness was adopted from the Panel Study of Income Dynamics Child Development Supplement (PSID-CDS-III, Gale and Nepomnyaschy 2024) to measure the teens’ experiences of inclusiveness, closeness, happiness, and safety at school. It was measured twice at Years 9 and 15. This study used data collected at Year 15. The four items include feelings of closeness, safety, being part of school, and being happy to be at school, with a 5-point scale from 0 = not once in the month to 4 = every day. The reliability was at a moderate level (α = 0.73).

5.2.3. Aggressive Behavior

For aggressive behavior, this study utilized eight questions from the Child Behavior Checklist (CBCL; Achenbach 1992) subscale, which were reported by primary caregivers across the three waves at Years 5, 9, and 15. Questions included those asking if the child is cruel, bullies, or shows meanness to others, destroys things belonging to the family or others, gets in many fights, physically attacks people, and argues a lot. It was measured with a 3-point scale from 1 = Not true to 3 = Often true. Responses were recoded from 0 = not true to 2 = often true for a total score. The reliability of aggressive behavior was moderately high (α = 0.74 at Y5, α = 0.77 at Y9, and α = 0.79 at Y15).

5.2.4. Demographic Information

We account for a set of demographic characteristics that may be related to adolescents’ ACEs and aggressive behavior, which include household poverty level measured at Year 15. Adolescents’ minority status and sex were included for data analysis, which were measured at Year 15.

5.3. Analytic Plan

This study used the first-order latent growth curve models (FLGMs) to examine changes in aggressive behavior across the three waves by the two groups of children with low and high-risk ACEs, since LGM can provide a model for continuous processes of individual development over time using longitudinal data. In this regard, a unique feature of LGM is its ability to estimate the average starting level (intercept) and average rate of change (slope) within individuals and combine this information with the estimation of variance terms, reflecting individual differences in the starting levels and rates of growth (Ferrer et al. 2004). To conduct FLGM tests, the present study used four steps. All data analyses were conducted with IBM SPSS and AMOS version 28.
First, to check for biases in attrition across the three waves of aggressive behavior, this study examined missing patterns of aggressive behavior to identify whether nonresponses across the three waves were selective. Overall, across the three waves, missing responses of aggressive behavior decreased from Year 5 to Year 15 by 32.1% (n = 2929 at Year 5, n = 3222 at Year 9, and n = 3562 at Year 15). For examining missing patterns in the three waves of data, eight possible missing patterns can be identified from the nonmissingness of aggression at all three waves (1,1,1) to the missingness of aggression at all three waves (0, 0, 0). The results from ANOVA tests showed statistically significant missing patterns at waves 2 and 3, which indicate that nonresponses of aggressive behavior may be selective (for Year 1, F (3, 2913) = 1.82, p = 0.14; for Year 9, F (3, 3204) = 3.57, p = 0.015; and for Year 15, F (3, 3526) = 4.07, p = 0.01). The eight missing patterns were tested with key variables of the study model that included age, gender, minority status, poverty, school connectedness, and high-risk ACE groups. Age and high-risk ACE groups were found to be statistically significant (for age, F (5, 3409) = 111.395, p < 0.001; for high-risk ACE group, ꭓ2 = 328.771, df = 7, p < 0.001), which indicates that nonresponses were not random. To adjust for selective nonresponses, multiple imputations were conducted.
Second, group differences in aggressive behavior, school connectedness, and ACEs by gender, minority status, and poverty were examined using independent sample t-tests. This step determines whether there is a difference in aggressive behavior over time, indicating that further data analysis of LGMs is required. Third, repeated measures of ANOVA tests were conducted prior to conducting LGM to identify whether repeated measures of aggressive behavior can be analyzed with LGM. Repeated measures of ANOVA tests indicate whether the change in aggressive behavior over time is significant and whether there is an interaction of the change rates of aggressive behavior between the two groups of adolescents with low- and high-risk ACEs. Fourth, in general, LGM is required to conduct two steps: unconditional and conditional model tests. With respect to the unconditional model, it determines the final change model of aggressive behavior over time, such as no change, linear change, or non-linear change model of aggressive behavior. After confirming the final unconditional change model of aggressive behavior, variances of the intercept and slope of the unconditional change model of aggressive behavior were checked to determine whether the final multivariate LGM test with school connectedness, and other control variables of age, sex, minority status, and poverty could be conducted. This conditional model provides a means to examine whether school connectedness affects the initial level and the change rate of aggressive behavior. Two conditional multivariate LGM tests were tested for the two groups of adolescents with low- and high-risk ACEs to compare the results of the protective role of school connectedness on aggressive behavior over time. The model fit index included RMSEA values no higher than 0.1 and CFI values of 0.9 or higher, indicating a good fit (Kline 1998; Loehlin 1992). The χ2/df is recommended to be less than 3. However, χ2/df is sensitive to the sample size.

6. Results

About a third of children were found to experience four or more cumulative ACEs over 10 years from Year 5 to Year 15 (30.8% for high-risk ACEs). Boys were found to experience significantly more ACEs than girls in all of the three waves with the exception of Year 15 (Y5: t = 2.38, df = 4229, p = 0.02: Y9: t = 3.49, df = 4229, p < 0.001; and Y15: t = 1.9, df = 4229, p = 0.06). Significant racial differences in ACEs were found across three waves (Y5: t = 1.9, df = 3234, p = 0.05; Y9: t = 2.32, df = 3234, p = 0.02; and Y15: t = 2.57, df = 3234, p < 0.01). Minority children reported experiencing significantly higher numbers of ACEs than White children. Moreover, children living below the federal poverty level were found to experience ACEs significantly more than their counterparts across the three waves (Y5: t = 3.69, df = 2241, p < 0.001; Y9: t = 4.56, df = 2610, p < 0.001; and Y15: t = 7.08, df = 2085, p < 0.001). A pairwise t-test showed that changes in ACEs in Years 9 to 15 and Years 5 to 15 were found to be significant (Y9 and 15: t = 5.12, df = 4230, p < 0.001; Y5 and 15: t = 3.99, df = 4230, p < 0.001), which indicated that ACEs significantly decreased over time.
The mean level of aggressive behavior from Years 5 to 15 decreased over time (Y5: m = 2.78, sd = 2.46; Y9: m = 1.91, sd = 2.24; and Y15: m = 1.69, sd = 2.33). Boys were found to be more aggressive than girls in all of the three waves (Y5: t = 3.80, df = 2915, p < 0.001: Y9: t = 6.35, df = 3203, p < 0.001; and Y15: t = 3.84, df = 3526, p < 0.001). No racial/ethnic group differences in aggressive behavior were found at Years 5 and 9, while minority children showed a higher level of aggressive behavior than White children at Year 15 (t = 3.06, df = 956, p = 0.002). Moreover, children living below the federal poverty level were found to be significantly more aggressive than their counterparts across three waves (Y5: t = 3.12, df = 1570, p < 0.001; Y9: t = 3.58, df = 1894, p < 0.001; and Y15: t = 6.14, df = 2120, p < 0.001). Pairwise t-test showed that changes in aggressive behavior were found to be significant across three waves (Y5 and 9: t = 13.67, df = 1401, p < 0.001; Y9 and 15: t = 4.81, df = 1886, p < 0.001; Y5 and 15: t = 16.97, df = 1503, p < 0.001), which indicated that aggressive behavior significantly decreased over time.
Before conducting latent growth curve modeling tests, repeated measures of ANOVAs were conducted to examine whether changes in aggressive behavior over time were significant between and within the two groups of children with low-risk ACEs (0–3 ACEs) and high-risk ACEs (4+ ACEs). As Table 1 shows, the results from repeated measures ANOVA revealed that there was a significant time effect, which shows that changes in aggressive behavior over time were significant (F (1, 2233) = 323.65, p < 0.001). Table 1 also displays the significance of the group effects of aggressive behavior between children in low- and high-risk ACEs, which means that changes in aggressive behavior over time were different between the two groups of adolescents (F (1, 2233) = 289.03, p < 0.001). This can be visualized as shown in Figure 1. The mean levels of aggressive behavior over time for children with high-risk ACEs were approximately twice as high as those with low-risk ACEs, although both displayed a decrease in aggressive behavior over time. Figure 1 also displays the finding that mean trajectories of aggressive behavior for the two groups of children are parallel, which indicated no interaction effect of the trajectories between the two groups of children (F (1, 2233) = 0.06, p = 0.93).
In the process of assessing the best fitting models of trajectories of aggressive behavior for the two groups of children, individual trajectories are examined in Figure 2 and Figure 3. These trajectories indicated that changes in aggressive behavior over time for both children were similar but not linear, while children with high-risk ACEs showed higher levels of aggressive behavior than those with low-risk ACEs, which is consistent with Figure 1.
Likewise, unconditional FLGM tests showed that out of three models (e.g., no change, linear change, and non-linear change models), the non-linear model provided the better fit, as presented in Table 2. The final non-linear change model fits the data (χ2/df = 7.68, CFI = 0.98, and RMSEA = 0.058). In particular, the final non-linear change model for children with high-risk ACEs fitted the data well (χ2/df = 1.42, CFI = 0.99, and RMSEA = 0.03), compared to those with low-risk ACEs (χ2/df = 13.01, CFI = 0.96, and RMSEA = 0.09).
Table 3 presents the variance terms of the intercept and the slope of aggressive behavior. All variances of intercepts and slopes of aggressive behavior were found to be significant for the overall sample and children with high-risk ACEs. These findings indicated substantial individual variabilities in developmental trajectories of aggressive behavior in the overall sample and children with high-risk ACEs. Thus, the intercept and slope variances can be further explained by additional variables with conditional multivariate model tests.
However, the variance of the aggressive behavior slope for children with low-risk ACEs was found to be insignificant, while the variance of the intercept of aggressive behavior was significant. These findings indicate that there was no significant individual variability in changes in aggressive behavior in children with low-risk ACEs, while there was significant variability in the initial levels of aggressive behavior among those. Thus, only the intercept variance can be further explained by additional variables.
Table 4 presents the results of the final multivariate conditional change model of FLGM tests based on the final study model presented in Figure 4, which includes covariates of age, gender, minority status, poverty, and school connectedness. The three final models adequately fit the data. The results indicated that school connectedness significantly affects the initial levels and change rates of aggressive behavior for the overall sample and those with high-risk ACEs. School connectedness also significantly explains the variance of the initial level of aggressive behavior for children with low-risk ACEs. In addition, sex was also found to be significant in the initial levels of aggressive behavior for the three models.

7. Discussion

There is a growing body of evidence showing that ACEs are significantly associated with antisocial behavior, delinquency, aggression, violence, and criminal behaviors in childhood, adolescence, and adulthood (Aroyewum et al. 2023; Baglivio et al. 2016; DeLisi and Beauregard 2018; Matsuura et al. 2013; O’Connor et al. 2021; Stoppelbein et al. 2024). Substantial research also examined protective factors that buffer aggressive behavior, including school connectedness (Bond et al. 2007; Choi et al. 2024; Gale and Nepomnyaschy 2024; Lee et al. 2018). As such, based on the social development model, the current study examined the protective role of school connectedness on the trajectories of aggressive behavior over ten years by comparing two groups of children with low- and high-risk ACEs. Some findings from the current study contribute to the existing literature in several ways.
Regarding ACEs, this study found that children who experienced four or more ACEs (high-risk ACEs) were a third of the study participants (30.8%), which is somewhat higher than those reported by other studies that ranged from 16% to 18.5% (Centers for Disease Control and Prevention 2021; Swedo et al. 2024). This may be due to the characteristics of the children in this study, as FFCWS focused on children in high-risk families with minority single/unmarried parents. This study also found that minority children living in poor families showed higher numbers of ACEs over time than their counterparts. This racial difference in ACEs is consistent with previous study findings that reported the vulnerability of ACEs among children and adolescents in poor families (Swedo et al. 2024). However, this study found that boys showed a higher number of ACEs than girls, which contrasts with findings from nationally representative data (Centers for Disease Control and Prevention 2021; Swedo et al. 2024).
In terms of aggressive behavior, this study found that aggressive behaviors decrease for both children in low- and high-risk ACEs as children age. According to Piquero et al. (2012), although a large proportion of children displayed instability in aggression, children showing low levels of aggression exhibited increased aggression later in adolescence, while those showing high levels of aggression exhibited decreased aggression. However, with longitudinal data, this study found that both children with low- and high-risk ACEs displayed decreased trajectories of aggressive behaviors over time, even though children with high-risk ACEs had about twice as high levels of aggression as those with low-risk ACEs. This finding confirms previous studies that have reported a decrease in aggressive behaviors with age, attributed to cognitive development and increased socialization (Baker et al. 2019; Duggins et al. 2016). This finding also aligns with findings from other studies focusing on neuroscience and aggression. Aggression tends to decrease as children age and their brain structure and cognitive function develop, particularly the changes in the volume of the amygdala and hippocampus (Bos et al. 2018; Pardini et al. 2014; Roberts et al. 2021). The amygdala and hippocampus play key roles in regulating aggressive behavior as the amygdala processes emotional learning and the response to potential threat and danger, and the hippocampus is important in processing memory, cognitions, and learning from consequences, which influence regulating behaviors and negative reactions (LeDoux and Pine 2016; Roberts et al. 2021).
Consistent with existing findings, boys from minority groups living in poor families were more aggressive than their counterparts (Lee et al. 2018; Song 2023). In particular, minority children had no differences in aggressive behavior at ages 5 and 9, while they showed a higher level of aggressive behavior at age 15. This finding may imply that minority children may be aware of challenges in social environments, such as racial discrimination or exposure to violence in their community, that affect their aggressive behaviors later in adolescence, as perceiving racial discrimination was associated with risky behavior and poorer academic achievement (Benner et al. 2018; Carter et al. 2017). In particular, Black and Hispanic adolescents experienced higher levels of violence in their communities, leading to the development of internalizing and externalizing behaviors (Aroyewum et al. 2023; Fowler et al. 2009).
From the final model tests, this study found little variation in the trajectory of aggressive behaviors over time in children with low-risk ACEs, whereas it found significant variation in the trajectory of aggressive behavior for those with high-risk ACEs. For those with low-risk ACEs, their aggressive behavior stably decreased over time, while their initial level of aggressive behaviors was affected by gender and school connectedness. School connectedness significantly decreased their starting level of aggressive behavior. For those with high-risk ACEs, school connectedness significantly decreased both the starting level and trajectories of aggressive behaviors. This finding suggests that school connectedness is crucial for children with high-risk ACEs to buffer their aggressive behaviors, which is consistent with findings from other studies (Crosnoe et al. 2002; Lee et al. 2020; Choi et al. 2024). For example, Lee et al. (2018) examined trajectories of delinquency over time among children in South Korea. Their study reported that the starting level and change rates of school connectedness significantly decreased the starting level and change rate of delinquency over time for children who experienced child abuse and neglect, while school connectedness decreased the starting level of delinquency for those without child maltreatment.
School-related factors, such as school bond, engagement, connectedness, climate, safety, and academic attainment, play a critical protective role. School factors are particularly important for children in high-risk ACEs to deter aggressive behavior. This may be due to their ACEs, as most abusers of ACEs are primary caregivers, leading to emotional disturbance, attachment difficulties, diverse mental health issues (i.e., depression, anxiety, post-traumatic stress disorder), and behavioral problems (Lee et al. 2018; Howe 2009). According to Hirschi’s social bond theory, non-parental figures or alternative figures in other social groups, such as those in school and the community, can play a protective role in preventing children at high risk of ACEs from engaging in delinquency and aggression (Hirschi 2002; Lee et al. 2018). When children experience positive social bonding with alternative parental figures, they can model appropriate behavior and positive bonding experiences (Baker et al. 2019). In particular, school can provide the opportunity to build positive bonding and connectedness with children at high risk of ACEs and can help mitigate the negative experiences they may encounter at home, as they spend most of their time in school.

7.1. Limitations and Recommendations for Future Research

Although this study presents meaningful findings, results should be interpreted with caution due to the study’s limitations. One limitation is the fact that the ACE measure in this study was not a tool created to measure ACEs. We developed an ACE score with questions asking about child maltreatment with the CTS-PC and individual questions about family dysfunctions as described in the Methods section. Therefore, the final ACE score could be different depending on the raters who understood responses in a different way due to missing values. Fortunately, the reliability of ACEs across the three waves was moderately acceptable.
Second, we found that missing values were not random. Although we adjusted them with multiple imputations, the validity of our study findings may be affected by selective attrition and imputation methods. In addition, this study used measures collected from parents’ or primary caregivers’ self-reports for ACEs and aggressive behaviors, which were asked with the same questions across the three waves at Years 5, 9, and 15 to maintain the responses over time. The use of self-reported data from parents and primary caregivers can lead to potential biases due to social desirability and recall bias. However, this intention is to maintain the reliability of longitudinal data as longitudinal data often changes data collection from parents/caregivers when children are young to children when they are able to be interviewed. However, this may indicate that the study findings can be indirect, as the measures were obtained by parents’ assessment of their children.
Last, the generalizability of the study’s findings should be interpreted with caution due to the characteristics of the sample, as FFCWS focused on high-risk families in the U.S. A large proportion of the children and adolescents examined in this study was composed of minorities from single/unmarried parents (75%), which is proportionally higher than the U.S. general population (41.6% by (The U.S. Census 2023)). In addition, this study focused only on the protective role of school connectedness on aggression, while diverse risk and protective factors affecting aggression are in play in a person and their social environments. Thus, it is critical to examine the mechanisms of diverse risk and protective factors on aggression, which enhances understanding of the protective role of school connectedness on aggression.
Given this study’s limitations, future research calls for a more sophisticated model that incorporates a second-order latent growth model (LGM) with measures of aggression, school connectedness, and ACEs, demonstrating strong reliability, validity, and measurement invariance across time. To increase generalizability, future research with an unbiased sample that can represent each country is also called for.

7.2. Practice Implications

Given the study’s limitations, this study also contains meaningful contributions to the literature. First, the focus of this study was to examine the associations among ACEs, aggressive behavior, and school connectedness. As ACEs are highly prevalent in low-income minorities (Swedo et al. 2024), the findings of this study can provide valuable insights for minority children from low-income families. Despite the fact that minority children in low-income families are more vulnerable to experiencing ACEs, school connectedness plays a critical role in protecting them from engaging in aggression and violence.
Aggressive and violent behaviors in adolescence and adulthood can be a part of untreated, comorbid internalizing and externalizing behaviors due to their ACEs (Duprey et al. 2020; van den Heuvel et al. 2025). Children with high-risk ACEs are unfortunate victims while they become offenders when they engage in aggressive and violent behaviors. Interventions and services for their internalizing and externalizing behaviors are urgent before children develop severe aggression and violence.
Substantial research has proven that children can be deterred from aggression and violence when there is an alternative figure who they can rely on regardless of their negative childhood experiences (Hirschi 2002; Lee et al. 2020). The key point is whether children can find and have a trustworthy adult who can care for and support them and can sustain the relationship needed to change their behavior. The significant protective role of school connectedness on aggression implies that they can adopt prosocial values based on the values of those with whom they have bonded in family, school, and community (Benish-Weisman 2015; Choi et al. 2005). School and other community leaders have the potential to collaborate to reduce children’s perceptions of isolation and enhance their prosocial skills, which may serve to interrupt the pathways into juvenile offending for adolescents with ACEs, especially for those with high-risk ACEs.
Children with high-risk ACEs display reactive aggression rather than proactive aggression as a result of reacting to threats that may harm their survival, given that they experience threats (e.g., physical, emotional, and sexual abuse) and deprivation (e.g., physical and emotional neglect) that threaten their survival and ability to thrive (Sheridan and McLaughlin 2014). Many children develop post-traumatic stress symptoms and anxiety through high-risk ACEs that continue into adulthood (Sabella et al. 2024). Their chronic dysregulated mental illness can lead to externalizing behavior (Hammen et al. 2000). Therefore, there is an urgent need to implement “counter-ACEs” interventions that prevent the occurrence of ACEs. Also, mental health treatments that mitigate mental illness affected by high-risk ACEs, such as emotional regulation, mindfulness, and other mental health services, should be provided for children and adolescents. Along with interventions to reduce ACEs and to treat their mental illness, enhancing protective factors such as helping build a positive relationship with an alternative figure in school and community is beneficial to deter aggressive behavior in children and adolescents (Narayan et al. 2018).

8. Conclusions

This study examined and compared longitudinal trajectories of aggressive behavior in children with low and high-risk ACEs. With a perspective of the Social Development Model, this study incorporated school connectedness in the trajectory of aggressive behavior to understand the role of school connectedness on aggression over time. This study found that children with high-risk ACEs showed twice as high levels of aggression over time compared to those with low-risk ACEs. School connectedness decreased the starting level and change rates of aggressive behavior over time for children with high-risk ACEs, while it also decreased the starting level of aggressive behaviors for those with low-risk ACEs. Building supportive environments significantly affects children with high-risk ACEs more than those with low-risk ACEs. Thus, practitioners and policymakers must consider building safe and supportive environments for all children, in particular for children with high-risk ACEs.

Author Contributions

S.-Y.L. conceived the study and both S.-Y.L. and M.C. conceptualized the study to develop the study model. S.-Y.L. obtained the data and conducted data analysis. S.-Y.L. and M.C. wrote the draft and finalized the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

No external funding was used for this study.

Institutional Review Board Statement

This study has been approved as Exempt by the UNI IRB (#IRB-FY25-220).

Informed Consent Statement

As a secondary data analysis, no informed consent can be provided. The original informed consent available at the Future of Families and Child Wellbeing Study (ffcws.princeton.edu).

Data Availability Statement

As a secondary data analysis, we have no right to publish the data. The data can be obtained from the Future of Families and Child Wellbeing Study (ffcws.princeton.edu).

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Achenbach, Thomas M. 1992. Manual for the Child Behavior Checklist/2-3 and 1992 Profile. Burlington: University of Vermont Department of Psychiatry. [Google Scholar]
  2. Aquilar, Serena, Dario Bacchini, and Gaetana Affuso. 2018. Three-year cross-lagged relationships among adolescents’ antisocial behavior, personal values, and judgment of wrongness. Social Development 27: 381–400. [Google Scholar] [CrossRef]
  3. Allen, Brian. 2011. Childhood psychological abuse and adult aggression: The mediating role of self-capacities. Journal of Interpersonal Violence 26: 2093–110. [Google Scholar] [CrossRef] [PubMed]
  4. Aroyewum, Bushura Afolabi, Sunday Oladotun Adeyemo, and Divine Chioma Nnabuko. 2023. Aggressive behavior: Examining the psychological and demographic factors among university students in Nigeria. Cogent Psychology 10: 1–10. [Google Scholar] [CrossRef]
  5. Asscher, Jessica J., Claudia E. Van der Put, and Geert Jan J. M. Stams. 2015. Gender differences in the impact of abuse and neglect victimization on adolescent offending behavior. Journal of Family Violence 30: 215–25. [Google Scholar] [CrossRef]
  6. Baglivio, Michael T., Kevin T. Wolff, Matt DeLisi, Michael G. Vaughn, and Alex R. Piquero. 2016. Effortful control, negative emotionality, and juvenile recidivism: An empirical test of DeLisi and Vaughn’s temperament-based theory of antisocial behavior. Journal of Forensic Psychiatry and Psychology 27: 376–403. [Google Scholar] [CrossRef]
  7. Baker, Erin R., and Cjersti J. Jensen. 2024. Linking teacher-versus child self-report discrepancies in aggression to demographic and cognitive profiles. Early Childhood Education Journal 52: 443–58. [Google Scholar] [CrossRef]
  8. Baker, Erin R., Cjersti J. Jensen, and Marie S. Tisak. 2019. A closer examination of aggressive subtypes in early childhood: Contributions of executive function and single-parent status. Early Child Development and Care 189: 733–46. [Google Scholar] [CrossRef]
  9. Benish-Weisman, Maya. 2015. The interplay between values and aggression in adolescence: A longitudinal study. Developmental Psychology 51: 677–87. [Google Scholar] [CrossRef]
  10. Benish-Weisman, Maya. 2019. What can we learn about aggression from what adolescents consider important in life? The contribution of values theory to aggression research. Child Development Perspectives 13: 260–66. [Google Scholar] [CrossRef]
  11. Benner, Aprile D., Yijie Wang, Yishen Shen, Alaina E. Boyle, Richelle Polk, and Yen-Pi Cheng. 2018. Racial/ethnic discrimination and well-being during adolescence: A meta-analytic review. American Psychologist 73: 855–883. [Google Scholar] [CrossRef]
  12. Bond, Lyndal, Helen Butler, Lyndal Thomas, John Carlin, Sara Glover, Glenn Bowes, and George Patton. 2007. Social and school connectedness in early secondary school as predictors of late teenage substance use, mental health, and academic outcomes. Journal of Adolescent Health 40: 357.e9–357.e18. [Google Scholar] [CrossRef] [PubMed]
  13. Bos, Marieke G. N., Lara M. Wierenga, Neeltje E. Blankenstein, Elisabeth Schreuders, Christian K. Tamnes, and Eveline A. Crone. 2018. Longitudinal structural brain development and externalizing behavior in adolescence. Journal of Child Psychology and Psychiatry 59: 1061–72. [Google Scholar] [CrossRef]
  14. Broidy, Lisa M., Daniel S. Nagin, Richard E. Tremblay, John E. Bates, Bobby Brame, Kenneth A. Dodge, David Fergusson, John L. Horwood, Rolf Loeber, Robert Laird, and et al. 2003. Developmental trajectories of childhood disruptive behaviors and adolescent delinquency: A six-site, cross-national study. Developmental Psychology 39: 222–45. [Google Scholar] [CrossRef]
  15. Carter, Robert T., Michael Y. Lau, Veronica Johnson, and Katherine Kirkinis. 2017. Racial discrimination and health outcomes among racial/ethnic minorities: A meta-analytic review. Journal of Multicultural Counseling & Development 45: 232–59. [Google Scholar] [CrossRef]
  16. Catalano, Richard F., Rick Kosterman, J. David Hawkins, Michael D. Newcomb, and Robert D. Abbott. 1996. Modeling the etiology of adolescent substance abuse: A test of the social development model. Journal of Drug Issues 26: 429–55. [Google Scholar] [CrossRef]
  17. Centers for Disease Control and Prevention. 2021. Adverse Childhood Experiences (ACEs). U.S. Department of Health & Human Services. Available online: https://www.cdc.gov/vitalsigns/aces/index.html (accessed on 6 April 2025).
  18. Choi, Mijin, Jun Sung Hong, Raphael Travis, Jr., and Jangmin Kim. 2022. Effects of school environment on depression among Black and White adolescents. Journal of Community Psychology 51: 1181–1200. [Google Scholar] [CrossRef] [PubMed]
  19. Choi, Mijin, Sei-Young Lee, Jangmin Kim, and Jungup Lee. 2024. The interaction of race/ethnicity and school connectedness in presenting internalizing and externalizing behaviors among adolescents. Children and Youth Services Review 161: 107605. [Google Scholar] [CrossRef]
  20. Choi, Yoonsun, Tracy W. Harachi, Mary Rogers Gillmore, and Richard F. Catalano. 2005. Applicability of the social development model to urban ethnic minority youth: Examining the relationship between external constraints, family socialization, and problem behaviors. Journal of Research on Adolescence 15: 505–34. [Google Scholar] [CrossRef]
  21. Cleverly, Kristin, Peter Szatmari, Tracy Vaillancourt, Michael Boyle, and Ellen Lipman. 2012. Developmental trajectories of physical and indirect aggression from late childhood to adolescence: Sex differences and outcomes in emerging adulthood. Journal of the American Academy of Child & Adolescent Psychiatry 51: 1037–51. [Google Scholar] [CrossRef]
  22. Connor, Daniel F., Jennifer J. Anderson, Ronald J. Steingazd, and Julie A. Cunningham. 2004. Proactive and reactive aggression in referred children and adolescents. American Journal of Orthopsychiatry 74: 129–36. [Google Scholar] [CrossRef]
  23. Crosnoe, Robert, Kristan Glassgow Erickson, and Sanford M. Dornbusch. 2002. Protective function of family relationships and school factors on the deviant behavior of adolescent boys and girls: Reducing the impact of risky friendships. Youth & Society 33: 515–44. [Google Scholar] [CrossRef]
  24. DeLisi, Matt, and Eric Beauregard. 2018. Adverse childhood experiences and criminal extremity: New evidence for sexual homicide. Journal of Forensic Science 63: 484–89. [Google Scholar] [CrossRef] [PubMed]
  25. Duggins, Shaun D., Gabriel P. Kuperminc, Christopher C. Henrich, Ciara Smalls-Glover, and Julia L. Perilla. 2016. Aggression among adolescent victims of school bullying: Protective roles of family and school connectedness. Psychology of Violence 6: 205–12. [Google Scholar] [CrossRef]
  26. Duprey, Erinn Bernstein, Assaf Oshri, and Sihong Liu. 2020. Developmental pathways from child maltreatment to adolescent suicide-related behaviors: The internalizing and externalizing comorbidity hypothesis. Development and Psychopathology 32: 945–59. [Google Scholar] [CrossRef]
  27. Eltink, E. M. A., J. Ten Hoeve, T. De Jongh, G. H. P. Van der Helm, I. B. Wissink, and G. J. J. M. Stams. 2018. Stability and change of adolescents’ aggressive behavior in residential youth care. Child & Youth Care Forum 47: 199–217. [Google Scholar]
  28. Ferrer, Emillo, Fumiaki Hamagami, and John J. McArdle. 2004. Modeling latent growth curves with incomplete data using different types of structural equation modeling and multilevel software. Structural Equation Modeling 11: 452–83. [Google Scholar] [CrossRef]
  29. Fite, Paula J., Craig R. Colder, and John E. Lochman. 2008. The Relation between childhood proactive and reactive aggression and substance use initiation. Journal of Abnormal Child Psychology 36: 261–71. [Google Scholar] [CrossRef] [PubMed]
  30. Fite, Paula J., Craig R. Colder, John E. Lochman, and Karen C. Wells. 2007. Pathways from proactive and reactive aggression to substance use. Psychology of Addictive Behaviors 21: 355–64. [Google Scholar] [CrossRef]
  31. Fowler, Patrick J., Carolyn J. Tompsett, Jordan M. Braciszewski, Angela J. Jacques-Tiura, and Boris B. Baltes. 2009. Community violence: A meta-analysis on the effect of exposure and mental health outcomes of children and adolescents. Development and Psychopathology 21: 227–59. [Google Scholar] [CrossRef]
  32. Gale, Adrian, and Lenna Nepomnyaschy. 2024. School connectedness and mental health among black adolescents. Journal of Youth & Adolescence 53: 1066–77. [Google Scholar] [CrossRef]
  33. Gilligan, Robbie. 2004. Promoting resilience in child and family social work. Social Work Education 23: 93–104. [Google Scholar] [CrossRef]
  34. Greeson, Johanna K. P., Ernestine C. Briggs, Christopher M. Layne, Harolyn M. E. Belcher, Sarah A. Ostrowski, Soeun Kim, Robert C. Lee, Rebecca L. Vivrette, Robert S. Pynoos, and John A. Fairbank. 2013. Traumatic Childhood Experiences in the 21st Century: Broadening and Building on the ACE Studies with Data From the National Child Traumatic Stress Network. Journal of Interpersonal Violence 29: 536–56. [Google Scholar] [CrossRef] [PubMed]
  35. Hammen, Constance, Risha Henry, and Shannon E. Daley. 2000. Depression and sensitization to stressors among young women as a function of childhood adversity. Journal of Consulting and Clinical Psychology 68: 782. [Google Scholar] [CrossRef]
  36. Hay, Dale F., Amy L. Paine, Oliver Perra, Kaye V. Cook, Salim Hashmi, Charlotte Robinson, Victoria Kairis, and Rhiannon Slade. 2021. Prosocial and aggressive behavior: A longitudinal study. Monographs of the Society for Research in Child Development 86: 7–103. [Google Scholar] [CrossRef] [PubMed]
  37. Hirschi, Travis. 2002. Causes of Delinquency. New Brunswick: Transaction Publishers. [Google Scholar]
  38. Howe, David. 2009. A Brief Introduction to Social Work Theory. New York: Palgrave Macmillan. [Google Scholar]
  39. Kline, Rex B. 1998. Principles and Practice of Structural Equation Modeling. New York: The Guilford Press. [Google Scholar]
  40. Kokko, Katja, and Lea Pulkkinen. 2000. Aggression in childhood and long-term unemployment in adulthood: A cycle of maladaptation and some protective factors. Developmental Psychology 36: 463–72. [Google Scholar] [CrossRef]
  41. LeDoux, Joseph E., and Daniel S. Pine. 2016. Using Neuroscience to Help Understand Fear and Anxiety: A Two-System Framework. American Journal Of Psychiatry 173: 1083–93. [Google Scholar] [CrossRef]
  42. Lee, Sei-Young, Margarita Villagrana, and Aweke Tadesse. 2020. Gender-specific Trajectories of Maltreatment, School Engagement, and Delinquency: The Protective Role of School Engagement. Victims & Offenders 16: 587–609. [Google Scholar] [CrossRef]
  43. Lee, Sei-Young, Siyon Rhee, and Margarita Villagrana. 2018. Change in delinquency over time between adolescents with and without maltreatment experiences: Attachment and the school’s role. Children and Youth Services Review 86: 110–19. [Google Scholar] [CrossRef]
  44. Loehlin, John C. 1992. Latent Variable Models: An Introduction to Factor, Path, and Structural Analysis, 2nd ed. Hillsdale: Lawrence Erlbaum Associates, Publishers. [Google Scholar]
  45. Matsuura, Naomi, Toshiaki Hashimoto, and Motomi Toichi. 2013. Associations among adverse childhood experiences, aggression, depression, and self-esteem in serious female juvenile offenders in Japan. The Journal of Forensic Psychiatry & Psychology 24: 111–27. [Google Scholar] [CrossRef]
  46. McRae, Elizabeth M., Laura Stoppelbein, Sarah E. O’Kelley, Paula Fite, and Shana B. Smith. 2021. An examination of post-traumatic stress symptoms and aggression among children with a history of adverse childhood experiences. Journal of Psychopathology & Behavioral Assessment 43: 657–70. [Google Scholar] [CrossRef]
  47. Narayan, Angela J., Luisa M. Rivera, Rosemary E. Bernstein, William W. Harris, and Alicia F. Lieberman. 2018. Positive childhood experiences predict less psychopathology and stress in pregnant women with childhood adversity: A pilot study of the benevolent childhood experiences (BCEs) scale. Child Abuse and Neglect 78: 19–30. [Google Scholar] [CrossRef] [PubMed]
  48. O’Connor, Kelly E., Terri N. Sullivan, Katherine M. Ross, and Khiya J. Marshall. 2021. “Hurt people hurt people”: Relations between adverse experiences and patterns of cyber and in-person aggression and victimization among urban adolescents. Aggressive Behavior 47: 483–92. [Google Scholar] [CrossRef] [PubMed]
  49. Oei, Adam, Dongdong Li, Chi Meng Chu, Irene Ng, Eric Hoo, and Kala Ruby. 2023. Disruptive behaviors, antisocial attitudes, and aggression in young offenders: Comparison of Adverse Childhood Experience (ACE) typologies. Child Abuse & Neglect 141: 1–12. [Google Scholar] [CrossRef]
  50. Pardini, Dustin A., Adrian Raine, Kirk Erickson, and Rolf Loeber. 2014. Lower amygdala volume in men is associated with childhood aggression, early psychopathic traits, and future violence. Biological Psychiatry 75: 73–80. [Google Scholar] [CrossRef]
  51. Piquero, Alex R., Michael L. Carriaga, Brie Diamond, Lila Kazemian, and David P. Farrington. 2012. Stability in aggression revisited. Aggression and Violent Behavior 17: 365–72. [Google Scholar] [CrossRef]
  52. Reichman, Nancy E., Julien O. Teitler, Irwin Garfinkel, and Sara S. McLanahan. 2001. Fragile families: Sample and design. Children and Youth Services Review 23: 303–26. [Google Scholar] [CrossRef]
  53. Roberts, Hannah, Elena Pozzi, Nandita Vijayakumar, Sally Richmond, Katherine Bray, Camille Deane, and Sarah Whittle. 2021. Structural Brain Development and Aggression: A Longitudinal Study in Late Childhood. Cognitive, Affective & Behavioral Neuroscience 21: 401–11. [Google Scholar] [CrossRef]
  54. Sabella, Kathryn, Ian A. Lane, Murron O’Neill, and Natalie Tincknell. 2024. Young adults’ perceptions of adverse childhood experiences as context and causes of mental health conditions: Observations from the United States. Child Protection and Practice 3: 100067. [Google Scholar] [CrossRef]
  55. Sheridan, Margaret A., and Katie A. McLaughlin. 2014. Dimensions of early experience and neural development: Deprivation and threat. Trends in Cognitive Sciences 18: 580–85. [Google Scholar] [CrossRef]
  56. Song, Jihee. 2023. Patterns of Adverse Childhood Experiences and Psychiatric Disorders Among Adolescents with ADHD: A Latent Class Analysis. Child Psychiatry & Human Development, 1–12. [Google Scholar] [CrossRef]
  57. Stoppelbein, Laura, Elizabeth McRae, and Shana Smith. 2024. Exploring the nexus of adverse childhood experiences and aggression in children and adolescents: A scoping review. Trauma, Violence, & Abuse 25: 3346–59. [Google Scholar] [CrossRef]
  58. Straus, Murray A., Sherry L. Hamby, David Finkelhor, Ddavid W. Moore, and Desmond Runyan. 1998. Identification of child maltreatment with the Parent–Child Conflict Tactics Scales: Development and psychometric data for a national sample of American parents. Child Abuse and Neglect 22: 249–70. [Google Scholar] [CrossRef]
  59. Swedo, Elizabeth A., Sanjana Pampati, Kayla N. Anderson, Evelyn Thorne, Izraelle I. McKinnon, Nancy D. Brener, Joi Stinson, Jonetta J. Mpofu, and Phyllis Holditch Niolon. 2024. Adverse Childhood Experiences and Health Conditions and Risk Behaviors Among High School Students—Youth Risk Behavior Survey, United States, 2023. Morbidity and Mortality Weekly Report 73: 39–49. [Google Scholar] [CrossRef]
  60. Tian, Yunlong, Chengfu Yu, Shuang Lin, Junming Lu, Yi Liu, and Wei Zhang. 2019. Parental psychological control and adolescent aggressive behavior: Deviant peer affiliation as a mediator and school connectedness as a moderator. Frontiers in Psychology 10: 358. [Google Scholar] [CrossRef]
  61. The U.S. Census. 2023. Fact Check. Available online: https://www.census.gov/quickfacts/fact/table/US/RHI825223 (accessed on 6 April 2025).
  62. van den Heuvel, Leigh Luella, Ayesha Assim, Milo Koning, Jani Nöthling, and Soraya Seedat. 2025. Childhood maltreatment and internalizing/externalizing disorders in trauma-exposed adolescents: Does posttraumatic stress disorder (PTSD) severity have a mediating role? Development and Psychopathology 37: 55–67. [Google Scholar] [CrossRef]
  63. Wang, Yao, Haksoon Ahn, Roderick A. Rose, and Kimberly Williams. 2023. Effects of school connectedness on the relationship between child maltreatment and child aggressive behavior: A mediation analysis. Child Abuse & Neglect 136: 1–9. [Google Scholar] [CrossRef]
  64. Ward, Kaitlin P., Andrew C. Grogan-Kaylor, Julie Ma, Garrett T. Pace, Shawna. J. Lee, and Pamela. E. Davis-Kean. 2025. Interactions of gender inequality and parental discipline predicting child aggression in low- and middle-income countries. Child Development 96: 7–20. [Google Scholar] [CrossRef]
Socsci 14 00385 g001
Figure 1. Mean trajectories of aggressive behavior by the two groups.
Figure 1. Mean trajectories of aggressive behavior by the two groups.
Socsci 14 00385 g001
Socsci 14 00385 g002
Figure 2. Individual trajectories of aggressive behavior of adolescents with high-risk ACEs.
Figure 2. Individual trajectories of aggressive behavior of adolescents with high-risk ACEs.
Socsci 14 00385 g002
Socsci 14 00385 g003
Figure 3. Individual trajectories of aggressive behavior of adolescents with low-risk ACEs.
Figure 3. Individual trajectories of aggressive behavior of adolescents with low-risk ACEs.
Socsci 14 00385 g003
Socsci 14 00385 g004
Figure 4. Final study model: effects of school connectedness on the initial aggression and the change in aggression.
Figure 4. Final study model: effects of school connectedness on the initial aggression and the change in aggression.
Socsci 14 00385 g004
Table 1. Results from repeated measures ANOVA for aggressive behavior over time.
Table 1. Results from repeated measures ANOVA for aggressive behavior over time.
Time EffectGroup EffectInteraction Effect
Aggressive BehaviorF (1, 2233) = 323.65 ***F (1, 2233) = 289.03 ***F(1, 2233) = 0.07
Note: Significance levels at *** p < 0.001.
Table 2. Results from model fit tests for the hypothetical growth curve model of aggressive behavior over time by the groups.
Table 2. Results from model fit tests for the hypothetical growth curve model of aggressive behavior over time by the groups.
Low-Risk ACEsHigh-Risk ACEsAll
χ2/dfCFIRMSEAχ2/dfCFIRMSEAχ2/dfCFIRMSEA
Aggressive Behavior
    No change20.37/2 ***0.990.05610.26/2 *0.990.05616.85/2 ***0.990.042
    Linear change739.65/2 ***0.470.35550.11/2 ***0.940.136512.82/2 ***0.8080.16
    Non-Linear change26.01/2 ***0.960.0922.85/20.990.01815.37/2 ***0.980.058
Note: Significance levels at * p < 0.05 and *** p < 0.001.
Table 3. Results from unconditional non-linear models of aggressive behavior by groups.
Table 3. Results from unconditional non-linear models of aggressive behavior by groups.
Low-Risk ACEsHigh-Risk ACEs All
Intercept
Mean2.38 (0.31) ***3.53 (0.07) ***2.74 (0.03) ***
Variance1.15 (0.05) ***2.94 (0.18) ***1.99 (0.07) ***
Slope
Mean−1.89 (0.06) ***−1.81 (0.14) ***−1.87 (0.06) ***
Variance0.18 (0.19)2.19 (0.72) **0.86 (0.25) ***
χ2/df26.011.4315.37
CFI0.9640.9990.989
RMSEA0.0920.0180.058
Note: Significance levels at ** p < 0.01; *** p < 0.001.
Table 4. Results from the final latent growth curve models of aggressive behavior affected by school connectedness of the two groups.
Table 4. Results from the final latent growth curve models of aggressive behavior affected by school connectedness of the two groups.
Low-Risk ACEs High-Risk ACEs All
Aggressive BehaviorAggressive BehaviorAggressive Behavior
InterceptSlopeInterceptSlopeInterceptSlope
Age−0.00 (0.06)−0.11 (0.14)−0.04 (0.11)0.04 (0.35)−0.04 (0.05)−0.02 (0.14)
Gender0.10 (0.06) ***−0.10 (0.20)0.13 (0.15) **0.08 (0.46)0.12 (0.06) ***0.08 (0.19)
Minority0.00 (0.09)−0.12(0.29)−0.05 (0.22)0.13 (0.68)−0.00 (0.09)0.02 (0.29)
Poverty−0.06 (0.09)−0.35 (0.28)−0.07 (0.16)0.05 (0.52)−0.09 (0.08) **−0.08 (0.25)
School Connect−0.13 (0.02) ***−0.34 (0.05) *−0.09 (0.03) *−0.25 (0.10) *−0.14 (0.02) ***−0.24 (0.05) **
Model Fit Statisticsχ2/df = 7.72 ***, CFI = 0.93,
RMSEA = 0.05		χ2/df = 3.79 ***, CFI = 0.95,
RMSEA = 0.04		χ2/df = 9.38 ***, CFI = 0.96,
RMSEA = 0.04
Note: Significance levels at * p < 0.05; ** p < 0.01; and *** p < 0.001.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Lee, S.-Y.; Choi, M. Changes in Aggressive Behaviors over Time in Children with Adverse Childhood Experiences: Focusing on the Role of School Connectedness. Soc. Sci. 2025, 14, 385. https://doi.org/10.3390/socsci14060385

AMA Style

Lee S-Y, Choi M. Changes in Aggressive Behaviors over Time in Children with Adverse Childhood Experiences: Focusing on the Role of School Connectedness. Social Sciences. 2025; 14(6):385. https://doi.org/10.3390/socsci14060385

Chicago/Turabian Style

Lee, Sei-Young, and Mijin Choi. 2025. "Changes in Aggressive Behaviors over Time in Children with Adverse Childhood Experiences: Focusing on the Role of School Connectedness" Social Sciences 14, no. 6: 385. https://doi.org/10.3390/socsci14060385

APA Style

Lee, S.-Y., & Choi, M. (2025). Changes in Aggressive Behaviors over Time in Children with Adverse Childhood Experiences: Focusing on the Role of School Connectedness. Social Sciences, 14(6), 385. https://doi.org/10.3390/socsci14060385

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop