Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Aug 1.
Published in final edited form as: Child Abuse Negl. 2024 Jul 24;154:106953. doi: 10.1016/j.chiabu.2024.106953

Profiles of Socioemotional Functioning in Children With and Without CPS-Subtantiated Maltreatment: Associations with Child Maltreatment and Dating Violence

Victoria M Atzl 2, Justin Russotti 1, Catherine Cerulli 1,2, Dante Cicchetti 1,3, Elizabeth D Handley 1
PMCID: PMC11325266  NIHMSID: NIHMS2014497  PMID: 39053219

Abstract

Objective:

Person-centered approaches are essential for characterizing heterogeneity in child development as it relates to child maltreatment (CM) and dating violence. The present study had two aims: 1) identify person-centered patterns of childhood socioemotional functioning, 2) examine whether patterns of child socioemotional functioning mediate the association between CM and dating violence.

Participants and Setting:

Wave 1 comprised N = 680 children ages 10–12 years with and without experiences of CPS-substantiated CM facing socio-economic challenge. Wave 2 included N = 407 emerging adults ages 18–24 years old.

Methods:

Children participated in a summer camp research program at Wave 1 and a follow up interview at Wave 2. Participant CM history and socioemotional functioning was assessed at Wave 1. Exposure to dating violence was assessed at Wave 2. A latent profile analysis identified patterns of socioemotional functioning. Then regression analyses examined associations of socioemotional functioning with CM and dating violence.

Results:

Three profiles of child socioemotional functioning were identified (well-regulated/low distress, high externalizing/high aggression, high internalizing). CM was significantly associated with membership in the high externalizing/high aggression class. Patterns of child socioemotional functioning did not mediate the association between CM and dating violence, although number of subtypes of CM had a significant positive direct effect on dating violence.

Conclusions:

Results underscore the multidimensional nature of socioemotional functioning and the predictive power of number of subtypes of CM on dating violence. Results can be harnessed by clinicians and policy makers to identify those at risk and interrupt cycles of violence.

Keywords: child maltreatment, latent profile analysis, internalizing, externalizing, emotion regulation, dating violence

Child maltreatment (CM) including physical abuse, sexual abuse, emotional abuse, physical neglect, and emotional neglect, is one of the most detrimental early experiences a child can have, significantly increasing risk for a wide variety of negative health outcomes across the lifespan (Cicchetti & Toth, 2005). One domain in which CM can have a significant negative impact is on the development of healthy interpersonal relationships (Cicchetti & Toth, 2005). Children who have been maltreated are more likely to develop regulatory, emotional and behavioral challenges which in turn lead to disrupted functioning in relationships in later developmental stages, including dating violence (Cicchetti & Toth, 2016; Handley et al., 2019; Handley et al., 2021; Masten & Cicchetti, 2010). Dating violence is a prevalent public health concern defined as physical, sexual, or psychological violence between individuals in a romantic or dating relationship (Center for Disease Control, 2021). The association between CM and dating violence has been established, including that polyvictimization, the experience of multiple types of victimization, is associated with increased risk of dating violence (Cuevas et al., 2020). Less is understood, however, about the potential combined effects of the varied regulatory, behavioral and emotional challenges that can result from CM and in turn be related to dating violence (Cascardi & Jouriles, 2018).

Profiles of Socioemotional Functioning

Socioemotional functioning involves foundational skills in identifying, modulating and communicating emotions in socially appropriate ways and developing appropriate, supportive relationships with close others (Darling-Churchill & Lippman, 2016). This includes specific skills such as emotion regulation, executive functioning, and prosocial behavior (Darling-Churchill & Lippman, 2016). Competence in these areas allows children to develop positive self-esteem, supportive peer relationships, academic success, etc. while delayed or disrupted development of these skills can lead to psychopathology, aggressive behaviors and peer rejection, among other negative outcomes (Kennedy et al., 2022; Ma et al., 2022). Emotion regulation, executive functioning skills, and internalizing and externalizing symptoms have been demonstrated empirically to impact one another (Dvir et al., 2014). For example, there is much interplay between emotion regulation and executive functioning skills, both involving abilities to integrate information, make decisions, and respond behaviorally, and both have been implicated in the development of internalizing and externalizing disorders (Dvir et al., 2014). The interrelated nature of these socioemotional capacities has led to an increase in person-centered literature focused on understanding the synergistic and interactive associations between domains of socioemotional functioning in children and adolescents (outside of the context of CM or dating violence; Kennedy et al., 2022; Ma et al., 2022). These studies indicate that while some children or adolescents demonstrate the same level of functioning across domains of socioemotional functioning others exhibit unique patterns of differences in domains of socioemotional functioning, suggesting a person-centered approach is well-suited to deepen our understanding of the synergistic associations of socioemotional functioning domains with CM and dating violence (Kennedy et al., 2022; Ma et al., 2022).

CM & Socioemotional Functioning

Some of the most widely researched consequences of CM are psychopathology, aggression and poor self-regulation, which are influential components of children’s adjustment and adaptation (Cicchetti & Toth, 2005; Cicchetti & Toth, 2016; Parker et al., 2015). Prospective research has established that CM significantly increases short- and long-term risk for internalizing and externalizing symptoms and related psychopathology including anxiety, depression, oppositional defiant disorder, and aggression (Baldwin et al., 2023; Cicchetti & Toth, 2016; Handley et al., 2019; Handley et al., 2021). Research has also focused on CM’s negative influence on self-regulation, demonstrating that exposure to CM leads to increased emotional lability, compromised regulatory ability, and deficits in emotional clarity, all components of emotion regulation (Cicchetti & Toth, 2016; Dvir et al., 2014; Kim & Cicchetti, 2010) as well as attention and executive functioning deficits including poor cognitive flexibility, inhibitory control and working memory (Cowell et al., 2015). Furthermore, polyvictimization has found been found to be associated increased internalizing and externalizing symptomology compared to children who experienced a single type of victimization (Finkelhor et al., 2007).

Socioemotional Functioning & Dating Violence

Dating violence and related phenomena such as intimate partner violence (IPV) are understood to be a multi-factorial phenomenon predicted by many different factors at different ecological levels (Dutton, 1995). Using Dutton’s nested ecological theory on partner violence as a frame, CM represents a disruption to the microsystem (settings the individual participates in directly), which can then negatively impact several different ontogenetic factors (i.e., individual level factors), including psychopathology, regulatory capacity or aggressive behavior (Dutton, 1995). Ontogenetic factors that follow CM have been studied empirically as precursors to dating violence (Vagi et al., 2013). Evidence indicates that externalizing problems, internalizing problems, aggressive behaviors, impulse control and emotion regulation difficulties are significantly associated with dating violence victimization and perpetration (Claussen et al., 2022; Handley et al., 2019; Handley et al., 2021; Spencer et al., 2019). Given the varied factors at just one ecological level (i.e., ontogenetic) it is essential to examine these factors in an integrative way to understand how they may predict dating violence.

Pathways from CM to Dating Violence

Cross sectional and longitudinal process models have identified several individual-level factors that mediate the link between CM and dating violence including internalizing problems, externalizing problems and emotion regulation challenges (e.g., Cascardi & Jouriles, 2018; Gratz et al., 2009). Executive functioning skill deficits have been demonstrated as a correlate of aggression in romantic relationships (e.g., Horne et al., 2020), and are theoretically posited to increase risk of dating violence perpetration but have rarely been studied in process models of CM and dating violence (Jouriles et al., 2012). Few of these process model studies have utilized prospective longitudinal designs. Those that have provide evidence for psychopathology, anger and aggression as mediators of the link between self-reported CM and dating violence (Cascardi, 2016; Fitzgerald, 2021; Narayan et al., 2013). Moreover, two studies that utilized prospective longitudinal designs and CPS record-based maltreatment data indicated that antisocial behavior and aggression in childhood mediated the association between CM and negative romantic interactions in emerging adulthood (Handley et al., 2019; Handley et al, 2021).

While much attention has been paid to understanding individual mediators of the link between CM and dating violence, less research has examined the impact of multiple mediators at the same time (Cascardi & Jouriles, 2018; Hèbert et al., 2021). Studies that have explored the impact of multiple mediators at once often do so using a variable-centered and/or sequential mediation framework with cross-sectional data (e.g., Kendra et al., 2012; Simons et al., 2014; Taft et al., 2010). For example, two studies have shown that PTSD mediated the association between CM and anger regulation, which then mediated the link between PTSD and emotional dating violence perpetration (Kendra et al., 2012; Taft et al., 2010). These studies make valuable contributions to the literature, but are unable to capture the likely integrative and bidirectional impact of the explanatory factors on the link between CM and dating violence.

The Current Study

Utilizing a multi-informant, prospective, longitudinal approach, the current study examined two aims in a primarily Black sample of children with and without CPS-substantiated CM histories facing socio-economic challenges: 1) identify person-centered patterns of child socioemotional functioning, 2) examine whether patterns of child socioemotional functioning mediate the association between CM and dating violence. Since person-centered approaches are inherently exploratory and classes are unknown, it is hypothesized that distinct patterns of socioemotional functioning will emerge that will have unique associations with dating violence and CM. Although we are unaware of previous studies exploring the particular combination of socioemotional indicators presented herein, findings from previous studies suggest that a mixture of classes representing poor functioning across domains, positive functioning across domains, and 2–3 distinct patterns of mixed functioning across socioemotional domains will emerge. Based on previous literature, it was also hypothesized that patterns of socioemotional functioning would mediate the association between CM and dating violence such that CM would predict more disrupted patterns of socioemotional functioning which would in turn predict increased likelihood of experiencing dating violence.

Methods

Participants

Participants included in the current study (N = 407) engaged in two waves of data collection: the first, a summer camp research program during childhood, followed by a research interview in emerging adulthood. Wave 1 comprised N = 680 children with (n = 360) and without (n = 320) CPS-substantiated CM and facing socio-economic challenges (Mage = 11.28, SDage = .97; 50.0% male; 71.6% Black, 11.8% White, 12.6% Hispanic and 4.0% biracial/other race). Children were recruited through a Department of Human Services (DHS) liaison who reviewed Child Protective Services (CPS) reports to identify children with a history of CM or whose family had a history of maltreatment. Families who received Temporary Assistance for Needy Families and did not have maltreatment histories substantiated by CPS were also recruited by the DHS liaison to obtain a socioeconomically similar sample of children without histories of CPS-substantiated CM. To confirm that the comparison group had not experienced any CPS-substantiated CM, comprehensive DHS record searches were conducted. Mothers who were recruited for both the CPS-involved and comparison groups were interviewed by trained research assistants using the Maternal Child Maltreatment Interview to further verify a lack of CPS involvement for children included in the comparison group (Cicchetti, Toth & Manly, 2003).

About nine years after Wave 1, participants were re-contacted to participate in a follow-up research interview. Comprehensive recruitment strategies including public internet searches, records of last known addresses, contact information from medical records and neighborhood canvassing were utilized to locate participants. Due to attrition, this resulted in N = 427 emerging adult participants at Wave 2. The current study includes 407 of the 427 because 20 participants were missing data on parameters of CM. At Wave 2, emerging adults had an average age of 19.65 years (SDage = 1.15, Range: 18–23 years), were about evenly distributed by gender (52.1% female), and majority Black (77.6% Black, 16.5% White, 5.9 % Other; 18.9% Hispanic/Latino). Just over half of the emerging adults included in the current study (52.3%) had a history of CPS-substantiated CM. Chi square analyses indicated that participants who completed Wave 2 data collection were not significantly different from participants who did not complete Wave 2 on participant age, gender, race/ethnicity or family income. Furthermore, there were no significant differences between Wave 2 completers and non-completers on internalizing symptoms, externalizing symptoms, peer- and counselor-reported aggression and the conflict effect score from the Attention Network Test (ANT). There was, however, a statistically significant difference in number of CM subtypes, such that non-completers experienced fewer types of CM than completers, t (405) = 2.55, p < .01).

Procedure

Children participated in a week-long summer camp research program at Wave 1 that included traditional camp activities and research assessments conducted by trained research assistants. Child participants were each part of a group of eight same-aged and same-sex peers (four with CPS-substantiated CM histories and four without) led by three trained research assistants who were blinded to children’s CPS-involvement status and study hypotheses. Parents provided consent for their child’s participation and for review of DHS records pertaining to the family. Children provided assent before participating. Children completed self-report measures about their emotions, experiences and behavior, were administered neuropsychological tasks, and rated their peers’ behavior. Counselors also rated child participants in their group on socioemotional functioning. For a comprehensive description of research summer camp procedures, please see Cicchetti & Manly (1990). For Wave 2, emerging adult participants were interviewed in private rooms by trained research assistants after providing consent to participate. Research assistants were blinded to each participant’s CM status and study hypotheses.

Measures

Wave 1 Measures

CM.

The Maltreatment Classification System (MCS; Barnett, Manly & Cicchetti, 1993), a standardized, comprehensive coding system of CM dimensions, was utilized to rate children’s official CPS records from birth to Wave 1 (age 10–12). Raters trained to reliability in the MCS coded records for the presence of four subtypes of CM: physical abuse, emotional abuse (including witnessing IPV), sexual abuse and neglect. Dichotomous (0 = no, 1 = yes) variables representing the presence of each subtype of CM were then summed to create a continuous variable representing the total number of subtypes of CM experienced from birth to Wave 1.

Child Behavioral Checklist (CBCL).

At the end of the week-long summer camp, counselors completed the Teacher Report Form (TRF) of the CBCL, a reliable and valid measure of child internalizing and externalizing behaviors (Achenbach, 1991). Each child was independently rated by two counselors on 113 items of child behavior (e.g., “Defiant, talks back to staff,” “Complains of loneliness”). Counselors rated how true each statement was of the child on a 3-point scale (0 = “not true,” 1 = sometimes or somewhat true,” and 2 = “very or often true”). The ratings from each counselor were then averaged. Counselors had good reliability [average ICC for externalizing subscale = .83; average reliability for internalizing subscale (k) = .68]. For the current study, the internalizing problems subscale and externalizing problems subscale raw scores were utilized in latent profile analysis as indicators of socioemotional functioning.

Revised Child Manifest Anxiety Scale (RCMAS).

The RCMAS is a self-report measure of child anxiety that has been well-validated and found to have good psychometric properties (Reynolds & Richmond, 1997; Murris et al., 2002). Children were asked to respond yes or no about whether they experienced 37 different symptoms of anxiety (e.g., “I worry a lot of the time”). A total raw score was calculated for each child to indicate total anxiety (α = .83). Total anxiety was utilized in latent profile analysis as an indicator of socioemotional functioning.

Pittsburgh Youth Survey.

The Pittsburgh Youth Survey (Loeber et al., 1998) is a self-report assessment of deviant behaviors that has been shown to have strong predictive and convergent validity compared to records of delinquency (Farrington et al., 1996). Child participants were asked (yes/no) if they have ever engaged in 32 different behaviors in their lifetime, including aggression towards others, destruction of property, rule-breaking and substance use. One item (“Have you ever sniffed glue?”) was excluded from the total score because of extremely low endorsement. Affirmative responses were summed to create a total conduct score (α = .84) which was then utilized in latent profile analysis as an indicator of socioemotional functioning.

Attention Network Test (ANT).

The child version of the ANT is a computer task that assesses attention network functioning (Rueda et al., 2004). Children were presented with a single fish or a row of five fish and asked to indicate with a right or left mouse click which way the central fish on the screen was facing. The task included both congruent (all fish facing the same direction) and incongruent (central fish facing a different direction than the other fish) trials. The ANT produced three subscales representing different attentional network capabilities: alerting, orienting and conflict. The conflict score (the difference in median reaction time between congruent and incongruent trials) was used in latent profile analysis as an indicator of socioemotional functioning. Higher conflict scores represent greater difficulty responding flexibly to incongruent visual information.

Emotion Regulation Checklist (ERC).

The ERC is a valid, 24-item measure of child adaptive and maladaptive emotion regulation that was completed by counselors about child participants (Shields & Cicchetti, 1998). The Emotion Regulation subscale, consisting of eight items related to abilities in managing negative emotions, navigating transitions and modulating one’s arousal (e.g., “Responds positively to neutral or friendly overtures by peers”) was used in latent profile analysis as an indicator of socioemotional functioning. Two camp counselors rated each child on how frequently they exhibited these behaviors using a 4-point Likert scale (1 = “never”, 2 = “sometimes,” 3 = “often,” 4 = “always”) and their scores were averaged to create a single Emotion Regulation subscale score per child (average ICC = .72). The Emotion Regulation subscale was reverse coded so that higher scores reflect higher levels of emotion regulation problems and will hereafter be referred to as emotion dysregulation.

Peer Ratings.

On the final day of camp, each child completed a sociometric measure on their peers (Bukowski et al., 2000). Children rated the other children in their group on multiple behavioral domains using a 3-point scale (0 = “not true,” 1 = “sort of true,” 2 = “very true”). The item “starts fights, says mean things, pushes or hits others” was used in the current study. Ratings from peers were averaged to create a single mean score for each child which was then utilized in latent profile analysis as an indicator of socioemotional functioning.

Mt. Hope Bully-Victim Questionnaire (VV-R).

The VV-R is a 10-item measure of bullying and victimization that was completed by counselors about child participants (Shields & Cicchetti, 2001). Five items pertained to overt aggressive behavior towards peers (e.g., “Is physically aggressive towards peers who are weaker or more vulnerable”). Two counselors rated each child on the frequency with which they displayed bullying behaviors using a 4-point Likert scale (1 = never, 2 = sometimes, 3 = often, 4 = always). An average score of counselor reports was created for each child and then utilized in latent profile analysis as an indicator of socioemotional functioning. Counselors demonstrated good reliability, with an ICC of .75.

Wave 2 Measures

Dating Violence.

At Wave 2, emerging adult participants were asked to respond yes or no to the question “Have you ever experienced or perpetrated domestic violence?” This resulted in a dichotomous variable representing the presence of lifetime dating violence victimization and/or perpetration, which was used as the outcome variable in analyses. About a fifth of the sample at Wave 2 (21.3%, N = 91) reported experiencing dating violence in their lifetime.

Data Analytic Plan

Latent Profile Analysis

Using Mplus Version 8.8 (Muthen & Muthen, 1998–2018) and maximum likelihood robust (MLR) estimation, a latent profile analysis (LPA) of Wave 1 measures of symptomology, self-regulation and aggression was conducted. SPSS Statistics 29 was utilized for bivariate analyses and calculating descriptive information and correlations among variables of interest. The following statistical fit indices were used to inform optimal profile selection for the LPA: lower Akaike Information Criterion (AIC; Akaike, 1987), lower Bayesian Information Criterion (BIC; Schwarz, 1978), lower sample size adjusted BIC (ssBIC; Sclove, 1987), higher entropy values, a significant bootstrap likelihood ratio test (BLRT; Collins & Lanza, 2010), a significant Lo-Mendell-Rubin Adjusted Likelihood Ratio Test (LMR test; Vuong, 1989) and a significant Vuong-Lo-Mendell-Rubin Likelihood Ratio Test (VLMR test; Vuong, 1989). The best, most parsimonious model was chosen based on model fit, empirical identification of the classes and theoretical rationale (Dziak et al., 2017). Further, class membership probability was considered when choosing the best class solution with a preference for class membership probabilities greater than 10% to enhance generalizability and replicability. Maximum likelihood robust estimation was used to accommodate non-normally distributed continuous indicators when conducting the LPA.

Mediational Analysis

Once the best fitting class solution was selected, a structural equation model (SEM) was conducted in MPlus using WLSMV estimation to test whether child socioemotional functioning mediated the association between CPS-substantiated CM and dating violence. Class assignment was saved to be used as a variable in subsequent analyses so predictors and consequences of socioemotional functioning could be assessed without altering the latent profile solution (Asparouhov & Muthén, 2014). Most likely class membership was coded into a set of binary dummy variables. The well-regulated/low distress class was not included in the model so it could serve as the reference group. Because raw scores that hadn’t been normed were used as indicators of child socioemotional functioning, Age at Wave 1 was included in the model. Gender and age at Wave 2 were included as predictors of dating violence in the model because of their significant associations with dating violence in bivariate analyses [χ2(1, N = 41) = 7.89, p < .05, t(418) = 2.78, p < .05, respectively]. Missing data ranged from 0% to 5.2% across study variables. Deviant behaviors assessed on the Pittsburgh Youth Survey (PYS) was the only variable that had greater than 5% missing data (5.2%).

Results

Rates of CM and Dating Violence in the Current Sample

About half of the sample experienced at least one subtype of CM (52.3%). Experience of CPS-substantiated CM in this study was defined by the number of different subtypes a child experienced from birth to Wave 1 (ages 10–12). Among children who experienced CPS-substantiated CM, 42.3% experienced a single subtype of CM (N = 90), 40.4% experienced two subtypes of CM (N = 86), 15.5% experienced three subtypes of CM (N = 33) and 1.9% (N = 4) experienced four subtypes of CM.

Latent Profile Analysis

Latent profile analysis was conducted for 1–5 classes (see Table 1 for bivariate correlations between latent profile indicators). The three-class solution was chosen based on model fit, interpretability of the classes and identification of class solutions (see Table 2). The three-class solution had good comparative fit. While the BLRT was significant across class solutions, the three-class solution had significant LMR and VLMR tests, while the four- and five-class solutions did not. The three-class solution was further deemed optimal because of the acceptability of the smallest class size and interpretability of the classes in contrast to the four- and five-class solutions. To define the classes, between group differences were compared for each indicator across classes (see Table 3). There were significant differences across classes on indicators of internalizing symptoms, externalizing symptoms, aggressive behavior and emotion dysregulation (detailed below) but there were no significant differences across classes on cognitive flexibility, measured by the ANT conflict effect score.

Table 1.

Bivariate correlations among latent profile indicators and predictor and outcome variables

Indicator 1 2 3 4 5 6 7 8 9 10
1. Number of Maltx Subtypes --
2. Dating Violence (0/1) .17*** --
3. Anxiety (Child-Report) .08 .11* --
4. Internalizing Problems (Counselor-Report) .03 −.12* .12* --
5. Externalizing (Counselor-Report) .17*** −.01 .04 .17*** --
6. Deviance (Child-Report) .11* .07 .23*** −.04 .28*** --
7. Aggression (Counselor-Reported) .16*** −.00 −.03 .00 .79*** .24*** --
8. Aggression (Peer-Reported) .15** .01 .00 .00 .68*** .22*** .64*** --
9. Emot. Dysreg. (Counselor-Reported) .08 −.11* .07 .51*** .39*** .10 .39*** .30*** --
10. ANT Conflict Effect Score .06 .00 .00 .07 .13** −.02 .01 .11* .07 --
Open in a new tab

Note. Emot. = Emotion; Dysreg. = Dysregulation; Maltx = Maltreatment; ANT = Attention Network Test

*

p < .05,

**

p < .01,

***

p < .001

Table 2.

Latent class analysis fit statistics for 1–5 classes

Class Log Likelihood AIC BIC ssBIC BLRT (p value) VLMR (p value) LMR (p value) Entropy % smallest class
1 −8,484.93 17,001.86 17,066.77 17,016.00
2 −8,087.04 16,224.08 16,325.50 16,246.17 <.001 <.001 <.001 .94 19.44
3 −7,958.01 15,984.01 16,121.94 16,014.05 <.001 <.01 <.01 .94 9.13
4 −7,833.76 15,753.52 15,927.96 15,791.51 <.001 >.05 >.05 .94 6.79
5 −7,767.86 15,639.71 15,850.66 15,685.65 <.001 >.05 >.05 .94 7.26
Open in a new tab

Note. AIC = Akaike Information Criteria; BIC = Bayesian Information Criteria; ssBIC = sample size adjusted BIC; BLRT = Bootstrap Likelihood Ratio Test; VLMR = Vuong-Lo-Mendell-Rubin Likelihood Ratio Test; LMR = Lo-Mendell-Rubin Adjusted Likelihood Ration Test

Table 3.

Between group differences in indicators for 3-class model and overall sample

Latent Classs
Indicator Class 1: Well-Regulated/Low Distress M(SD) Class 2: High Externalizing/High Aggression M(SD) Class 3: High Internalizing M(SD) Overall Sample M(SD) Welch’s F d
RCMAS Anxiety 9.33(6.24)a 9.44(6.51)ab 12.28(7.36)b 9.62 (6.44) 3.72*
CBCL Internalizing 2.44(2.50)a 4.38(3.70)b 15.44(4.79)c 4.00 (4.78) 143.42***
PYS Delinquent Bxs 4.19(4.00)a 6.82(4.70)b 4.00(3.49)a 4.67 (4.22) 13.02***
CBCL Externalizing 3.41(4.12)a 20.87(8.08)b 6.64(6.00)c 7.07 (8.62) 179.33***
VV-R Victimizer 1.17(.28)a 2.29(.55)b 1.14(.19)a 1.38 (.56) 161.90***
Peer Rating (fights) .19(.27)a .95(.47)b .17(.29)a .33 (.44) 98.89***
ERC Emot Dysreg 1.73(.38)a 2.21(.38)b 2.40(.51)b 1.89 (.46) 72.94***
ANT Conflict Effect 66.98(57.51)a 83.31(62.10)a 81.13(64.54)a 71.34 (59.32) 2.92
N = 306
71.7%		 N = 82
19.2%		 N=39
9.1%
Open in a new tab

Note.

*

p < .05,

**

p < .01,

***

p < .001;

Class 1 = Low Distress/Well Regulated; Class 2 = High Externalizing/High Aggression, Class 3 = High Internalizing

Bxs = behaviors; Emot. = Emotion; Dysreg = Dysregulation

abc

Classes do not share subscripts differ by p < .05

d

Welch’s F test and Games-Howell post-hoc comparisons were used for all variables (except RCMAS Anxiety, PYS Delinquent Bxs & ANT Conflict Effect) because the homogeneity of variance assumption was not met.

Class 1: Well-Regulated/ Low Distress

The well-regulated/low distress class comprised 71.7% of the sample (N = 306). The average latent class probability for most likely latent class membership for this class was 0.98. This class was characterized by significantly lower average scores on emotion dysregulation compared to the two other classes. This class also exhibited significantly lower levels of counselor-reported internalizing and externalizing symptoms compared to the other two classes. Finally, aggression scores, as reported by counselors and peers, and self-reported delinquent behaviors were significantly lower than the high externalizing/high aggression class (Class 2).

Class 2: High Externalizing/ High Aggression

Class 2, the high externalizing/high aggression class, comprised 19.2% of the current sample (N = 82). The average latent class probability for most likely latent class membership for this class was 0.96. The primary distinguishing feature of this class was extremely high aggression and externalizing symptoms, as evidenced by an average counselor-reported externalizing score one standard deviation above the mean. Counselor-reported externalizing symptoms and self-reported delinquent behaviors were both significantly higher for this class than the other two classes. This was also true for peer-reported and counselor-reported aggression. Further, counselor-reported and peer-reported aggression were each one standard deviation above the overall sample mean. This class also evidenced significantly higher counselor-reported internalizing symptoms than the well-regulated/low distress class. Finally, this class exhibited above average mean levels of emotion dysregulation which were significantly higher than the well-regulated/low distress class but not significantly higher than Class 3 (high internalizing).

Class 3: High Internalizing

Class 3, the high internalizing class, comprised 9.1% of the current sample (N = 39). The average latent class probability for most likely latent class membership for this class was 0.95. This class’s primary distinguishing feature was above-average internalizing symptoms, as evidenced by counselor-reported internalizing symptoms more than two standard deviations above the mean that were also significantly higher than the other two classes. Members of this class self-reported internalizing symptoms that were significantly higher than the well-regulated/low distress class. This class also exhibited levels of emotion dysregulation one standard deviation above the mean, which were significantly higher than the well-regulated/low distress class but not significantly higher than the high externalizing/high aggression class.

Mediation Analysis

An SEM (see Figure 1) was conducted to examine the second aim of the study. The model had good fit [χ2 (9, N = 407) = 7.08, p = .63, CFI = 1.00, RMSEA = 0.00, SRMR = .04]. The results indicated there was a significant direct effect of number of subtypes of CM on the likelihood of being in the high externalizing/high aggression class (vs. the well-regulated/low distress class; β = .25, p < .001) but not on the likelihood of being in the high internalizing class (vs. the well-regulated/low distress class; β = −.18, p > .05). Age at Wave 1 was significantly related to membership in the high internalizing class (vs. the well-regulated/low distress class; β = −.18, p < .05) but not the high externalizing/high aggression class (vs. the well-regulated/low distress class; β = .08, p > .05). Number of subtypes of CM also had a significant direct effect on likelihood of experiencing dating violence such that greater number of subtypes experienced predict a higher likelihood of experiencing dating violence (β = .29, p < .01). Age at Wave 2 also was significantly related to likelihood of dating violence (β = .17, p < .05), as was gender (β = .19, p < .01; females were more likely to experience dating violence).

Figure 1.

Figure 1.

Open in a new tab

Mediational Model

Note. * p < .05, ** p < .01, *** p < .001; Standardized path coefficients are presented. Dashed lines represent non-significant paths. For the model shown, the reference group was the Well-Regulated/Low distress class.

Neither membership in the high internalizing class (vs. the well-regulated/low distress class; β = −.28, p > .05) nor membership in the high externalizing/high aggression class (vs. the well-regulated/low distress class; β = −.13, p > .05) had a significant main effect on the likelihood of experiencing dating violence. Furthermore, class membership was not found to mediate the association between CM and dating violence as evidenced by non-significant specific indirect effects of number of CM subtypes on dating violence through membership in the high internalizing class (β = −.03, p > .05) and through membership in the high externalizing/high aggression class (β = −.03, p > .05), as well as a nonsignificant total indirect effect of number of subtypes of CM on dating violence (β = −.06, p > .05). The lack of mediation was further confirmed through examination of 95% asymmetric confidence intervals. Confidence intervals that include zero are not statistically significant. The asymmetric confidence intervals confirmed that neither membership in the high internalizing class (LL = −.15, UL = .04) nor the high externalizing/high aggression class (LL = −.16, UL = .08) mediated the association between number of subtypes of CM and dating violence.

Discussion

The current study utilized a person-centered approach to examine whether patterns of socioemotional functioning mediate the association between CM and dating violence. This study advances the field of CM research by utilizing person-centered methodology and a multi-informant approach to deepen our understanding of the synergistic relationships of domains of socioemotional functioning in children with and without histories of CPS-substantiated CM and their associations with dating violence. Further, this study adds to a very small literature demonstrating the association between variation in CPS-substantiated CM and dating violence using a prospective longitudinal approach with record-based CM data in a primarily Black sample.

Aim1: Latent Profiles of Socioemotional Functioning in Children with and without CPS-Substantiated CM Histories

The LPA approach allowed for the simultaneous examination of several indices of development across multiple reporters which elucidated different patterns of socioemotional functioning. For example, while emotion dysregulation distinguished the well-regulated/low distress class from the other two classes, both the high externalizing/high aggression class and the high internalizing class had comparable levels of emotion dysregulation, underscoring the necessity of a multidimensional approach to understand differences in child socioemotional functioning. Further, the ANT conflict effect score, an indicator of cognitive flexibility, did not significantly differ across classes, suggesting that in this sample, other factors better distinguished children from one another on socioemotional functioning.

The current study broadly aligns with the extant literature using person-centered approaches to capture patterns of developmental capacities following CM which demonstrate broadly similar patterns in which one of the identified classes represents individuals that, on average, are functioning well and could be described as resilient, while the other identified classes have different patterns of challenges and capacities, exemplifying the developmental psychopathology tenet of multifinality (Cicchetti & Rogosch, 1996; Martinez-Torteya et al., 2017; Russotti et al., 2020). The current study’s findings highlight that groupings of socioemotional functioning in children with and without histories of CPS-substantiated CM are not as simple as “low, medium, high” severity but vary based on patterns of developmental factors across domains of functioning. This is exemplified by the fact that the high externalizing/high aggression class, in addition to substantial levels of externalizing challenges, also exhibited significantly higher counselor-reported internalizing symptoms than the well/regulated low distress class, which may point to the emergence of internalizing symptoms in this class resulting from the negative consequences of behavioral challenges and conflict with peers (Nobile et al., 2013).

Aim 2: Mediation Analysis

The results indicated that patterns of child socioemotional functioning did not significantly mediate the association between CPS-substantiated CM and dating violence. Further, patterns of socioemotional functioning were not significantly directly associated with the likelihood of experiencing dating violence, while CM was. The lack of mediation and direct association between patterns of child socioemotional functioning and dating violence may be understood in terms of the background-situational theory of dating violence, which posits that distal background factors (e.g., psychopathology, CM) create vulnerability for dating violence, while the addition of more proximal situational factors (e.g., current stress, alcohol use) and proximal relationship factors (e.g., negative interactions) increase the likelihood for specific instances of dating violence to occur (Collibee & Furman, 2016; Riggs & O’Leary, 1989). While CM may be a potent enough background risk factor to predict dating violence on its own, patterns of child socioemotional functioning may represent a background risk factor that increases vulnerability for dating violence only when coupled with a more proximal risk factor. Moreover, children’s socioemotional capacities at 10–12 years old (age at Wave 1 participation) could have changed as they developed into emerging adults (age at Wave 2 participation, 18–22 years old); socioemotional functioning measured more proximally to late adolescence/emerging adulthood might be related to dating violence. It’s also possible that there are other explanatory variables that better explain the association between variation in CM and dating violence, such as acceptance of violence in relationships or insecure attachment (Cascardi & Jouriles, 2018). While these variables have been established as correlates and in some cases mediators of dating violence individually, future research is necessary to understand how they might interact with other explanatory factors to portend risk for dating violence (Cascardi & Jouriles, 2018).

Results indicated that as the number of CM subtypes increased, the likelihood of being in the high externalizing/high aggression class (vs. the well-regulated/low distress class) significantly increased while the likelihood of being in the well-regulated/low distress class (vs. the high externalizing/high aggression class) significantly decreased. This finding aligns with polyvictimization literature which indicates that more types of victimization lead to greater externalizing challenges (e.g., Finkelhor et al., 2007). It was unexpected that number of CM subtypes did not significantly predict the likelihood of being in the high internalizing class, given previous literature demonstrating a robust association between CM and internalizing symptoms (Cicchetti & Toth, 2005). Since the high externalizing/high aggression class also experienced substantial internalizing symptoms, it could be that variation in CM was more strongly related to the co-occurrence of symptomology than internalizing symptoms alone (Duprey et al., 2020).

Although CM was associated with decreased likelihood of being in the well-regulated/low distress class, almost half of children in the well-regulated class experienced at least one subtype of CM, underscoring that resilience is possible in the context of CM. This aligns with previous literature demonstrating that children who experience CM can achieve successful adaptation in a variety of domains (Cicchetti, 2013; Masten 2012). In the case of the current study, the well-regulated/low distress class exhibited less externalizing and internalizing symptomology, less aggressive behavior and better emotion regulation than the other two classes. Furthermore, this finding underscores the importance of investigating variation in outcomes within maltreated populations by probing factors related to the CM experience (e.g., number of subtypes, timing, etc.) that may influence development. If this study had utilized a binary (yes/no) CM variable, the differential associations between CM variation and child socioemotional functioning may have been lost.

The current study found a significant main effect of number of subtypes of CM on likelihood of experiencing dating violence. This finding deepens the previous literature on the association between CM and dating violence by focusing on the prospective association of variation in CM and its influence on dating violence (Goncy et al., 2021). Theoretically, experiencing more subtypes of CM may represent a greater depth of betrayal from trusted adults and greater number of experiences in which a child is unsafe or fearful, both of which may increase the likelihood and intensity of subsequent developmental cascades of maladaptation that increase the risk for dating violence (Masten & Cicchetti, 2010).

Strengths and Limitations

This study adds to the literature on CM, socioemotional development and dating violence by utilizing a person-centered, multi-informant approach to deepen our understanding of the heterogeneous nature of child socioemotional functioning and its associations with CM and dating violence. Strengths of this study include using latent profile analysis to identify patterns of socioemotional functioning in children with and without CPS-substantiated CM histories which illuminated different patterns of challenges and capacities. Further, this study utilized a prospective, longitudinal, multi-informant design in a predominately Black sample facing socio-economic challenges, as well as CM data substantiated through CPS records. The multi-informant approach of this study was particularly robust as it allowed for an examination of socioemotional functioning from the perspective of the children themselves, their skills in a computer task, their peers’ observations and their counselors’ reports.

This study also had limitations. After completing the LPA, children were assigned class membership to examine patterns of socioemotional functioning as a predictor of dating violence, a strategy that can introduce statistical bias. However, this risk was diminished by the high entropy levels in the LPA solution (Clogg, 1995). Further, this approach had the advantage of allowing for the addition of gender, age and CM as covariates in the second model. Additionally, the single yes/no question used to measure lifetime involvement in dating violence prevented a more nuanced examination of the impact of CM and socioemotional functioning on dating violence perpetration vs. dating violence victimization. However, a single dichotomous question allowed participants to avoid identifying as a victim or perpetrator, a potential barrier to reporting and an especially difficult journey for victims who believe they are perpetrators engaging in bidirectional violence when they defend themselves (Cort et al., 2010). The question also used the term “domestic violence” without providing a specific definition, which although broadly understood to mean violence between romantic or dating partners (Center for Disease Control, 2021), may have been understood by some participants to apply to all violence experienced in the home. Finally, the use of CPS record-based CM has pros and cons. While CPS record-based CM has strengths in that it is prospectively reported and confirmed by multiple sources, it can also lead to underestimation of CM because it doesn’t capture CM experiences that don’t rise to the level of CPS investigation or involvement, and/or CM experiences that are hidden from CPS or CPS does not discover. Operationalizing CM as number of subtypes of CM experienced does not capture chronicity (i.e., number of CPS reports), which may also negatively influence socioemotional functioning and increase the likelihood of domestic violence. This is an important area for future research.

Conclusions

In conclusion, this study aimed to characterize profiles of child socioemotional functioning in children with and without histories of CPS-substantiated CM to examine whether profile membership mediated the association between CM and dating violence. The three profiles identified (well-regulated/low distress, high externalizing/high aggression and high internalizing) underscore the multidimensional and interactive nature of socioemotional functioning. The findings also highlight resilience in the context of CM, as almost half of the well-regulated/low distress group experienced at least one subtype of CM. Additionally, the findings highlight the specific impact of experiencing multiple subtypes of CM on externalizing problems and aggressive behaviors, underscoring the importance of assessing for CM history and broader trauma history when providing clinical services to children with this pattern of challenges. Patterns of child socioemotional functioning did not predict dating violence in adolescence/emerging adulthood when accounting for number of subtypes of CM, highlighting the predictive power of variation in CM on the probability of experiencing dating violence. Further research, clinical work and policy advances that incorporate nuanced measurement of dating violence contexts to identify youth most at risk for being in dating violence-involved relationships is essential so that cycles of violence and maladaptation can be interrupted.

Acknowledgements

We are grateful to the National Institute on Drug Abuse (R01-DA01774 to D.C.), National Institute of Mental Health (R01-MH083979 to D.C.) and P50-HD096698 to D.C.) for their support of this work. Thank you to the individuals who participated in the research.

Funding

sources were not responsible for any aspect of the study design; or data collection, analysis, or interpretation.

Footnotes

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

Declarations of interest: none.

References

  1. Achenbach TM (1991). Manual for the teacher’s report form and 1991 class. University of Vermont, Department Psychiatry. [Google Scholar]
  2. Akaike H (1987). Factor analysis and AIC. Psychometrika, 52, 371–386. [Google Scholar]
  3. Asparouhov T, & Muthén B (2014). Auxiliary variables in mixture modeling: Three-step approaches using M plus. Structural equation modeling: A Multidisciplinary Journal, 21(3), 329–341. [Google Scholar]
  4. Baldwin JR, Wang B, Karwatowska L, Schoeler T, Tsaligopoulou A, Munafò MR, & Pingault JB (2023). Childhood maltreatment and mental health problems: A systematic review and meta-analysis of quasi-experimental studies. American Journal of Psychiatry, 180(2), 117–126. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Barnett D, Manly JT, & Cicchetti D (1993). Defining child maltreatment: The interface between policy and research. In Cicchetti D & Toth S (Eds.), Child abuse, child development, and social policy (pp. 7–73). Ablex. [Google Scholar]
  6. Bukowski WM, Sippola L, Hoza B, & Newcomb AF (2000). Pages from a sociometric notebook: An analysis of nomination and rating scale measures of acceptance, rejection, and social preference. New Directions for Child and Adolescent Development, 2000(88), 11–26. 10.1002/cd.23220008804 [DOI] [PubMed] [Google Scholar]
  7. Cascardi M (2016). From violence in the home to physical dating violence victimization: The mediating role of psychological distress in a prospective study of female adolescents. Journal of Youth and Adolescence, 45, 777–792. [DOI] [PubMed] [Google Scholar]
  8. Cascardi M, & Jouriles EN (2018). Mechanisms underlying the association of exposure to family of origin violence and adolescent dating violence. In Adolescent Dating Violence (pp. 159–188). Academic Press. [Google Scholar]
  9. Centers for Disease Control and Prevention. (2021). Youth Risk Behavior Survey Data. Available at: www.cdc.gov/yrbs.
  10. Cener for Disease Control and Prevention (2021). #TakeaStand Against Domestic Violence. https://www.cdc.gov/violenceprevention/featuredtopics/intimate-partner-violence.html#:~:text=Intimate%20partner%20violence%2C%20also%20known,does%20not%20require%20sexual%20intimacy.
  11. Cicchetti D (2013). Annual research review: Resilient functioning in maltreated children–past, present, and future perspectives. Journal of Child Psychology and Psychiatry, 54(4), 402–422. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cicchetti D, & Manly JT (1990). A personal perspective on conducting research with maltreating families: Problems and solutions. In Brody GH & Sigel IE (Eds.), Methods of family research: Biographies of research projects, Vol. 2 Clinical populations (pp. 87–133). Lawrence Erlbaum Associates. [Google Scholar]
  13. Cicchetti D, & Rogosch FA (1996). Equifinality and multifinality in developmental psychopathology. Development and Psychopathology, 8(4), 597–600. [Google Scholar]
  14. Cicchetti D, & Toth SL (2005). Child maltreatment. Annu. Rev. Clin. Psychol, 1, 409–438. [DOI] [PubMed] [Google Scholar]
  15. Cicchetti D, & Toth SL (2016). Child maltreatment and developmental psychopathology: A multilevel perspective. Developmental Psychopathology, 3, 457–512. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Cicchetti D, Toth SL, & Manly JT (2003). Maternal maltreatment classification interview. Rochester, NY: Unpublished manuscript, Mt. Hope Family Center. [Google Scholar]
  17. Claussen C, Matejko E, & Exner-Cortens D (2022). Exploring risk and protective factors for adolescent dating violence across the social-ecological model: A systematic scoping review of reviews. Frontiers in Psychiatry, 13, 933433. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Clogg CC (1995). Latent class models: recent developments and prospects for the future. Handbook of Statistical Modeling in the Social Sciences, 311–359. [Google Scholar]
  19. Collibee C, & Furman W (2016). Chronic and acute relational risk factors for dating aggression in adolescence and young adulthood. Journal of Youth and Adolescence, 45, 763–776. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Collins LM, & Lanza ST (2010). Latent class analysis with covariates. In Latent class and latent transition analysis: With applications in the social, behavioral, and health sciences. [Google Scholar]
  21. Cort NA, Cerulli C, & He H (2010). Investigating health disparities and disproportionality in child maltreatment reporting: 2002–2006. Journal of Public Health Management and Practice: JPHMP, 16(4), 329. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Cowell RA, Cicchetti D, Rogosch FA, & Toth SL (2015). Childhood maltreatment and its effect on neurocognitive functioning: Timing and chronicity matter. Development and Psychopathology, 27(2), 521–533. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Darling-Churchill KE, & Lippman L (2016). Early childhood social and emotional development: Advancing the field of measurement. Journal of Applied Developmental Psychology, 45, 1–7. [Google Scholar]
  24. Duprey EB, Oshri A, & Liu S (2020). Developmental pathways from child maltreatment to adolescent suicide-related behaviors: The internalizing and externalizing comorbidity hypothesis. Development and Psychopathology, 32(3), 945–959. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Dutton DG (1995). The domestic assault of women: Psychological and criminal justice perspectives. UBC press. [Google Scholar]
  26. Dvir Y, Ford JD, Hill M, & Frazier JA (2014). Childhood maltreatment, emotional dysregulation, and psychiatric comorbidities. Harvard Review of Psychiatry, 22(3), 149. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Dziak D, & Jachimczyk B & Kulesza WJ (2017). IoT-based information system for healthcare application: Design methodology approach. Applied Sciences, 7(6), 596. [Google Scholar]
  28. Farrington DP, Barnes GC, & Lambert S (1996). The concentration of offending in families. Legal and Criminological Psychology, 1(1), 47–63. 10.1111/j.2044-8333.1996.tb00306.x [DOI] [Google Scholar]
  29. Finkelhor D, Ormrod RK, & Turner HA (2007). Poly-victimization: A neglected component in child victimization. Child Abuse & Neglect, 31(1), 7–26. [DOI] [PubMed] [Google Scholar]
  30. Fitzgerald M (2021). Developmental pathways from childhood maltreatment to young adult romantic relationship functioning. Journal of Trauma & Dissociation, 22(5), 581–597. [DOI] [PubMed] [Google Scholar]
  31. Goncy EA, Basting EJ, & Dunn CB (2021). A meta-analysis linking parent-to-child aggression and dating abuse during adolescence and young adulthood. Trauma, Violence, & Abuse, 22(5), 1248–1261. [DOI] [PubMed] [Google Scholar]
  32. Gratz KL, Paulson A, Jakupcak M, & Tull MT (2009). Exploring the relationship between childhood maltreatment and intimate partner abuse: Gender differences in the mediating role of emotion dysregulation. Violence and Victims, 24(1), 68–82. [DOI] [PubMed] [Google Scholar]
  33. Handley ED, Russotti J, Rogosch FA, & Cicchetti D (2019). Developmental cascades from child maltreatment to negative friend and romantic interactions in emerging adulthood. Development and Psychopathology, 31(5), 1649–1659. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Handley ED, Russotti J, Warmingham JM, Rogosch FA, Todd Manly J, & Cicchetti D (2021). Patterns of child maltreatment and the development of conflictual emerging adult romantic relationships: An examination of mechanisms and gender moderation. Child Maltreatment, 26(4), 387–397. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Hébert M, Lapierre A, MacIntosh HB, & Ménard AD (2021). A review of mediators in the association between child sexual abuse and revictimization in romantic relationships. Journal of Child Sexual Abuse, 30(4), 385–406. [DOI] [PubMed] [Google Scholar]
  36. Horne K, Henshall K, & Golden C (2020). Intimate partner violence and deficits in executive function. Aggression and Violent Behavior, 54, 101412. [Google Scholar]
  37. Jouriles EN, McDonald R, Mueller V, & Grych JH (2012). Youth experiences of family violence and teen dating violence perpetration: Cognitive and emotional mediators. Clinical Child and Family Psychology review, 15, 58–68. [DOI] [PubMed] [Google Scholar]
  38. Kendra R, Bell KM, & Guimond JM (2012). The impact of child abuse history, PTSD symptoms, and anger arousal on dating violence perpetration among college women. Journal of Family Violence, 27, 165–175. [Google Scholar]
  39. Kennedy SM, Tonarely NA, Halliday E, & Ehrenreich-May J (2022). A person-centered approach to understanding heterogeneity of youth receiving transdiagnostic treatment for emotional disorders. Journal of Consulting and Clinical Psychology, 90(3), 234. [DOI] [PubMed] [Google Scholar]
  40. Kim J, & Cicchetti D (2010). Longitudinal pathways linking child maltreatment, emotion regulation, peer relations, and psychopathology. Journal of Child Psychology and Psychiatry, 51(6), 706–716. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Lanza ST, & Cooper BR (2016). Latent class analysis for developmental research. Child Development Perspectives, 10(1), 59–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Loeber R, Farrington DP, Stouthamer-Loeber M, & Van Kammen WB (1998). Antisocial behavior and mental health problems: Explanatory factors in childhood and adolescence. Psychology Press. [Google Scholar]
  43. Ma TL, Zarrett N, Puente K, Liu Y, Vandell DL, Simpkins SD, & Yu MVB (2022). Longitudinal Links Between Profiles of Social Emotional Behaviors in Childhood and Functioning in Early Adolescence. The Journal of Early Adolescence, 42(6), 765–792. [Google Scholar]
  44. Martinez-Torteya C, Miller-Graff LE, Howell KH, & Figge C (2017). Profiles of adaptation among child victims of suspected maltreatment. Journal of Clinical Child & Adolescent Psychology, 46(6), 840–847. [DOI] [PubMed] [Google Scholar]
  45. Masten AS (2012). Resilience in individual development: Successful adaptation despite risk and adversity. In Educational resilience in inner-city America (pp. 3–25). Routledge. [Google Scholar]
  46. Masten AS, & Cicchetti D (2010). Developmental cascades. Development and Psychopathology, 22(3), 491–495. [DOI] [PubMed] [Google Scholar]
  47. Muthen LK, & Muthen BO (1998–2018). Mplus user’s guide.
  48. Narayan AJ, Englund MM, & Egeland B (2013). Developmental timing and continuity of exposure to interparental violence and externalizing behavior as prospective predictors of dating violence. Development and Psychopathology, 25(4pt1), 973–990. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Nobile M, Colombo P, Bellina M, Molteni M, Simone D, Nardocci F, … & Battaglia M (2013). Psychopathology and adversities from early-to late-adolescence: a general population follow-up study with the CBCL DSM-Oriented Scales. Epidemiology and Psychiatric Sciences, 22(1), 63–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Parker JG, Rubin KH, Erath SA, Wojslawowicz JC, & Buskirk AA (2015). Peer relationships, child development, and adjustment: A developmental psychopathology perspective. Developmental Psychopathology: Volume One: Theory and Method, 419–493. [Google Scholar]
  51. Reynolds CR, & Richmond BO (1997). What I think and feel: A revised measure of children’s manifest anxiety. Journal of Abnormal Child Psychology, 25, 15–20. doi: 10.1023/A:1025751206600 [DOI] [PubMed] [Google Scholar]
  52. Riggs DS, & O’Leary KD (1989). The development of a model of courtship aggression. Violence in dating relationships: Emerging Social Issues, 53–71. [Google Scholar]
  53. Rueda MR, Fan J, McCandliss BD, Halparin JD, Gruber DB, Lercari LP, & Posner MI (2004). Development of attentional networks in childhood. Neuropsychologia, 42(8), 1029–1040. 10.1016/j.neuropsychologia.2003.12.012 [DOI] [PubMed] [Google Scholar]
  54. Russotti J, Warmingham JM, Handley ED, Rogosch FA, & Cicchetti D (2020). Characterizing competence among a high-risk sample of emerging adults: Prospective predictions and biological considerations. Development and Psychopathology, 32(5), 1937–1953. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Schwarz G (1978). Estimating the dimension of a model. The Annals of Statistics, 6, 461–464. [Google Scholar]
  56. Sclove SL (1987). Application of model-selection criteria to some problems in multivariate analysis. Psychometrika, 52, 333–343. [Google Scholar]
  57. Shields A, & Cicchetti D (1998). Reactive aggression among maltreated children: The contribution of attention and emotion regulation. Journal of Clinical Child Psychology, 27, 381–395. doi: 10.1207/s15374424jccp2704_2 [DOI] [PubMed] [Google Scholar]
  58. Shields A, & Cicchetti D. (2001). Parental maltreatment and emotion dysregulation as risk factors for bullying and victimization in middle childhood. Journal of Clinical Child Psychology, 30,349–363. [DOI] [PubMed] [Google Scholar]
  59. Simons LG, Simons RL, Landor AM, Bryant CM, & Beach SR (2014). Factors linking childhood experiences to adult romantic relationships among African Americans. Journal of Family Psychology, 28(3), 368. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Spencer CM, Stith SM, & Cafferky B (2019). Risk markers for physical intimate partner violence victimization: A meta-analysis. Aggression and Violent Behavior, 44, 8–17. [Google Scholar]
  61. Taft CT, Schumm J, Orazem RJ, Meis L, & Pinto LA (2010). Examining the link between posttraumatic stress disorder symptoms and dating aggression perpetration. Violence and Victims, 25(4), 456–469. [DOI] [PubMed] [Google Scholar]
  62. Vagi KJ, Rothman EF, Latzman NE, Tharp AT, Hall DM, & Breiding MJ (2013). Beyond correlates: A review of risk and protective factors for adolescent dating violence perpetration. Journal of Youth and Adolescence, 42, 633–649. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Vuong Q (1989). Likelihood ratio tests for model selection and non-nested hypotheses. Econometrica, 57, 307 – 333. [Google Scholar]
  64. Warmingham JM, Handley ED, Rogosch FA, Manly JT, & Cicchetti D (2019). Identifying maltreatment subgroups with patterns of maltreatment subtype and chronicity: A latent class analysis approach. Child Abuse & Neglect, 87, 28–39. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES