The Role of Trauma Type and Age in the Relation Between Trauma Exposure and Intelligence

Abstract

Experiencing potentially traumatic events (PTEs) is associated with deficits in cognitive functioning for youth. Previous research has demonstrated that PTE type may influence the association between PTE experiences and intelligence, such that IQ scores may differ by the type of PTE experienced. Additionally, mixed findings in the literature suggest that these associations might differ by age. The current study examined the association between PTE type and IQ and how age may moderate this association. In a sample of youth in foster care (N = 328, M_age = 13.25), physical, psychological, and sexual abuse and family PTEs were directly associated with verbal and nonverbal IQ. Age significantly moderated the association between IQ and neglect and academic PTEs. Results suggest that broad PTE grouping methods or failing to account for both maltreatment and general PTEs in samples of youth in foster care may conceal differences in how PTEs relate to intellectual functioning.

Keywords

child maltreatment cognitive development measurement child trauma

Exposure to trauma or potentially traumatic events (PTEs) in childhood is not uncommon, with reports from national studies estimating that at least one third of youth in the United States have experienced at least one PTE (Costello, Erkanli, Fairbank, & Angold, 2002; Finkelhor, Turner, Shattuck, & Hamby, 2013). PTEs can include experiences of maltreatment (e.g., physical abuse, sexual abuse, neglect) as well as exposure to general PTEs or fairly nontypical stressors (e.g., natural disaster, car accident, or death of loved one; Costello et al., 2002). Ample literature exists on the negative psychological and physical consequences of experiencing PTEs during childhood and adolescence (for review, see De Young, Kenardy, & Cobham, 2011). As many as one half of youth who experience PTEs display clinical levels of internalizing symptoms (e.g., depression, anxiety; Norman et al., 2012) and/or externalizing problems (Wilson, Dolan, Smith, Casanueva, & Ringeisen, 2012). However, since not all youth who experience PTEs develop negative outcomes, this suggests that the association between PTE exposure and maladjustment is not automatic, and there may be other factors that contribute to the outcomes youth demonstrate.

One likely contributor for how youth adjust to PTE exposure is individual differences in intelligence or problem-solving skills. Intelligence, as measured by IQ (a common proxy for cognitive functioning in youth), is critical for youth development. Several studies have demonstrated a positive association between IQ and development in childhood, such as greater success in school and higher education (Brody, 1997; Tiet et al., 1998), lower frequencies of negative behavior (e.g., delinquency, conduct problems; Murray & Farrington, 2010), and successful social functioning (e.g., increased cooperativeness and self-control among peers; Vanderbilt-Adriance & Shaw, 2008). IQ has also been found to be associated with adjustment and development in PTE-exposed youth more specifically. For example, Saltzman, Weems, and Carrion (2006) found that a high verbal IQ was associated with endorsing fewer symptoms of trauma-related distress. Findings such as these help demonstrate the importance of intelligence, again, measured as IQ in most studies, in healthy development following PTE exposure.

While there is evidence supporting the link between IQ and adjustment, what is less understood are the factors linking PTE and IQ. Research findings provide strong support for an association between IQ and exposure to PTEs, suggesting that youth reporting exposure to PTEs generally tend to score lower on IQ tests, as compared to youth with no history of PTE exposure. For example, comparisons between matched comparison youth with no PTE exposure and youth with histories of PTE (e.g., physical abuse, neglect) showed that nonexposed youth tended to have not only higher IQ scores overall than exposed youth but also higher scores on various subscales of IQ as well as compared to exposed youth (Viezel, Freer, Lowell, & Castillo, 2015). This is true not only for youth exposed to PTEs like maltreatment but non-maltreatment-related exposure as well. In a sample of urban, elementary school students who had a history of exposure to violence (e.g., seeing somebody shot, stabbed, or beat up), exposure was negatively correlated with IQ scores, suggesting that greater exposure to violence was associated with lower IQ scores (Delaney-Black et al., 2002). Additionally, youth with PTE exposure may be more likely to be identified as having an intellectual disability, as compared to youth with no PTE exposure (Vig & Kaminer, 2002).

Although the evidence is fairly compelling that there is a significant relation, mixed findings within the literature suggest that there may be other factors to consider in the relation between PTEs and IQ. For example, in contrast to the previous findings, some studies have reported that there is no significant difference between traumatized and nontraumatized children on measures of intellectual functioning (e.g., Bücker et al., 2012). Closer examination of sample characteristics between studies with mixed findings suggests that two factors that might explain these differences are (a) PTE exposure type and (b) age. PTE exposures are not unidimensional, and additional investigations into PTE type and intelligence suggest the need for more research to better illuminate the relation. Furthermore, given potential age differences regarding PTE response and exposure type, age may also play an important role in the PTE–IQ relation.

Factors Associated With PTE Type and Intelligence

Above and beyond IQ being related to PTE exposure, some researchers suggest that an important component involved in the association between PTE and IQ is the type of PTE youth experience. For example, Kira, Lewandowski, Somers, Yoon, and Chiodo (2012) examined the association between IQ, post-traumatic stress disorder (PTSD) symptoms, and PTE type and found differences between the PTE type experienced and measures of intelligence. The authors’ results suggest that experiences of sexual abuse negatively impacted IQ scores, whereas those who experienced secondary family traumas (e.g., parental war involvement) generally experienced some positive effects on IQ (Kira, Lewandowski, Somers, Yoon, & Chiodo, 2012). These findings are consistent with other reports that suggest different types of PTEs have different kinds of relations with IQ. Viezel, Freer, Lowell, and Castillo (2015) examined the differences in IQ scores between maltreated and nonmaltreated children, and within the same sample, they also compared children with only experiences of neglect to nonmaltreated children. The authors found significant differences between the maltreated group and comparison group, such that the maltreated youth demonstrated lower overall IQ, verbal comprehension, and processing speed scores. When comparing the neglected-only children with the comparison group of nonmaltreated children, the neglected youth scored significantly lower than the comparison group for overall IQ and processing speed, but there was no significant difference for verbal comprehension scores. DePrince, Weinzierl, and Combs (2009) examined different PTEs in relation to IQ and found that family types of PTEs tend to have a greater negative impact on IQ than nonfamily PTEs. These findings suggest that different types of PTEs may have different influences on or relations with intellectual skills. In contrast, however, Crozier and Barth (2005) examined IQ and maltreatment history in school-aged children and found no significant difference between overall IQ scores and maltreatment type.

Taken together, mixed findings within the literature make it difficult to determine whether PTE type is an important factor involved in the association between PTE exposure and IQ. Potential methodological issues with the operationalization of PTE type may help explain some of these mixed findings. Studies that examine IQ and PTE exposure tend to use a general grouping method that combines types of PTE into broad categories that may not adequately separate the different type of PTEs. For example, Viezel et al. (2015) reported differences between PTE type and IQ outcomes, but the researchers in this study examined only two classifications of maltreatment, physical abuse and neglect, while ignoring other forms of abuse such as sexual abuse and psychological abuse. Using broad grouping methods of types could conceal potential differences between PTE type and IQ. Simply focusing on maltreatment PTEs, while ignoring other general types of PTEs, could also possibly explain some of the inconsistent findings. Maltreatment and other general PTEs can be greatly distinct. For example, maltreatment experiences are by definition incidents of interpersonal victimization in which harm to the child is the result of other individuals’ actions such as a parent hitting a child or failing to provide proper medical care for the child. On the other hand, more general PTEs can be the result of nonhuman influence, such as the environment (e.g., natural disaster) or disease (e.g., chronic illness; D’Andrea, Ford, Stolbach, Spinazzola, & van der Kolk, 2012; Finkelhor, 2008). Additionally, maltreatment events tend to be experiences that are direct (e.g., raped or forced to watch someone perform sexual acts), whereas other general PTEs can include events that result from indirect harm exposure (e.g., witnessing a car accident; Elklit & Schouwenaars, 2016).

It is important to note that youth may have exposure to more than one event, thus research that simplifies exposure to yes/no and to only one or a few types (e.g., car accident or not, maltreated or not) of PTEs may fail to recognize the heterogeneity of exposure in their samples (e.g., Finkelhor et al., 2013). Moreover, research suggests that when the full range or accumulation of PTEs is considered, there is a fairly stable dose–response relation such that the more PTEs one experiences, the greater the risk for pathology or other negative outcomes, often referred to as the cumulative risk hypothesis (Raviv, Taussig, Culhane, & Garrido, 2010). Thus, focusing on just one type of PTE can exclude other important past experiences, which may contribute to current functioning. For example, Costello, Erkanli, Fairbank, and Angold (2002) reported a reciprocal relationship between high-magnitude (e.g., sexual abuse or death of loved one) and low-magnitude traumatic events (e.g., parental arrest or reduction in standard of living), such that experiencing either one was associated with a twice as likely chance for experiencing the other. Further, certain PTEs, at both high and low magnitudes, tend to co-occur at a greater frequency, which could also conceal certain nuances between each PTE type if not properly accounting for each type of PTE. For example, when considering maltreatment types, neglect and psychological abuse often tend to co-occur at a higher percentage with other maltreatment types (e.g., McGuire & Jackson, 2018). Thus, when examining the impact of PTEs on intellectual abilities, it may be necessary to take into consideration the influence of as many types or kinds of exposures as possible to better explain the range of behavioral and emotional outcomes in youth.

Age, PTE Exposure, and Intelligence

In addition to potential influences based on differences between PTE types, another possible explanation for conflicting evidence among studies investigating IQ and PTE type could be the range of age of youth included in research samples. The impact from PTE exposure may be influenced by developmental differences between school-aged, middle childhood, and adolescence-aged youth. This includes age-related factors associated with PTE exposure, such as response and understanding of the traumatic event and coping strategies (Vernberg, 1999). For example, younger children may not be able to encode as many details of a PTE, thus limiting their knowledge and memory of the event (Salmon & Bryant, 2002). In some instances, younger children may not even appraise an event as traumatic, as compared to older youth. In turn, differences such as these between younger and older youth may exacerbate, either directly or indirectly, potential negative outcomes or protect a child from experiencing negative outcomes associated with a PTE. For example, this might include age-related differences in the development of psychopathology, which might interfere with intellectual skills as a secondary consequence (De Bellis, Woolley, & Hooper, 2013; Schoedl et al., 2010). For example, Kira et al. (2012), who reported differences in IQ outcomes for different types of PTE exposure, examined data from an adolescent population. Crozier and Barth (2005), who report no differences between IQ and PTE type, used data from a younger sample of school-aged children.

Additionally, another reason to consider age, especially when investigating type of PTE exposure, is that differences in type of PTE exposure are also associated with age, such that youth may be more likely to experience a certain type of PTE at a particular age. For example, Jones and McCurdy (1992) found that adolescents tend to experience higher rates of emotional maltreatment, whereas infants tend to experience higher rates of physical neglect. Differences between age and type of PTE have also been demonstrated for nonmaltreatment PTEs, such that youth may be more likely to be exposed to community violence or witnessing traumatic events at certain ages (e.g., Richters & Martinez, 1993). Thus, if certain types of PTEs are more prominent at certain ages, studies with restricted age ranges may not accurately capture certain PTEs.

Current Study

Although comparisons across studies provide some initial evidence, methodological differences between studies make it difficult to know with certainty whether differences do exist. The current study attempted to address this gap in the literature by examining the association between a wide variety of PTEs and intellectual functioning in youth at different ages to examine whether age moderates these associations. It was hypothesized that (1) different types of PTEs would have different or unique relations with IQ scores, and (2) the association between PTE type and IQ would be moderated by the youth’s age such that effect of PTE on IQ would differ across age. Given the wide variety of PTE types and associations being tested, the hypotheses are largely exploratory and not focused on a specific type of PTE.

Method

Participants

Participants were 328 foster care youth with an average age of 13.25 (SD = 2.93) and their primary caregiver. The sample consisted of mostly males (53.1%), and the largest represented ethnic group was African American (50.6%), followed by White or Caucasian (33.6%), multiracial (9.3%), and Hispanic/Latino (4.3%). 51.5% of primary caregivers were foster mothers or fathers, 28.1% were primary caretaker staff at residential facilities, 12.7% were biological relatives, and 7.8% listed themselves as others (e.g., transitional living program staff). Participants were recruited as part of a larger project, Studying Pathways to Adjustment and Resilience in Kids (SPARK). Youth were excluded from the SPARK project if they were non-native English speakers, had a history of psychotic symptoms, or were diagnosed with a disability such as Autism Spectrum Disorder or other type of developmental delay. Further, all youth had to be in foster care for at least 30 days. Given the demands of the data collection procedures, the norms under which the measures were evaluated, and the nature of questions being asked, these exclusion criteria were established to ensure youth could properly and accurately complete data collection. Participants’ caregivers were screened for exclusion criteria via phone or in person at recruitment and their first session. Further, youth with any identified disability were excluded from the recruitment information provided through the Division of Family Services (e.g., contact information for these youth were not given to the project staff).

Measures

Intelligence

Intelligence was measured using the Kaufman Brief Intelligence Test, Second Edition (KBIT-2; Kaufman & Kaufman, 2004). The KBIT-2 is an individualized test of verbal and nonverbal intellectual ability for individuals 4–90 years of age (Kaufman & Kaufman, 2004). Two scores were derived from the KBIT-2 and used in the analyses: verbal and nonverbal IQ scores. The reliability estimates for the KBIT-2 are high with mean α coefficients above .80 across each subscale (Kaufman & Kaufman, 2004). The KBIT-2 has strong validity coefficients relative to full-length IQ tests. For example, the KBIT-2 has shown strong correlations with Full Scale IQ scores from the Wechsler Intelligence Scale for Children, Fourth Edition (r > .76; Bain & Jaspers, 2010). These findings suggest that the KBIT-2, although brief, is a good representation of actual intelligence, as measured by full-length intelligence tests (Kaufman & Kaufman, 2004). For a more comprehensive review, see Bain and Jaspers (2010).

Maltreatment history

Participants’ history of maltreatment was obtained through youth self-report of lifetime maltreatment exposure. The questions used in the current study were adapted from the Modified Maltreatment Classification System (MMCS; English & LONGSCAN Investigators, 1997). The MMCS was developed based on the widely used Maltreatment Classification System (MCS; Barnett, Manly, & Cicchetti, 1993), a coding system used for analyzing records about youth’s maltreatment history from the District of Family Services. Participants responded to questions across four types of maltreatment (physical abuse, 19 items; sexual abuse, 12 items; psychological abuse, 15 items; and neglect, 22 items). Maltreatment questions asked youth to indicate on a scale from 1 (never) to 5 (almost always) how often they experienced a certain type of abuse (e.g., “In your lifetime, about how often did someone kick or punch you?”). If youth answered with any number between 2 and 5, this was counted as an exposure to maltreatment. The total number of items endorsed was then summed for each type of maltreatment. This type of method has shown high levels of fidelity in accessing maltreatment experiences and been used successfully in previous studies (e.g., Jackson et al., 2016).

PTE or non-maltreatment-related PTEs

General PTE history was assessed using a modified version of the Life Events Checklist (LEC; Johnson & McCutcheon, 1980). The LEC is a self-report measure of 46 age-appropriate PTEs. Participants were asked to report whether any of the events (e.g., death of a loved one) had occurred in their lifetime. The events included in the LEC were categorized into four domains, family events (e.g., “In your lifetime, how often have any of your family members had a serious injury/illness?”), friend events (e.g., “In your lifetime, how often have you seen any of your friends get hurt?”), academic events (e.g., “In your lifetime, how often have you felt that one of your teachers/coaches were disappointed in you for not doing good enough?”), and personal events (e.g., “In your lifetime, how often have you been in an accident or natural disaster?”). A total sum of PTEs in each category was calculated by summing together all events endorsed within that category.

Demographics

Demographics were collected from youths’ caregivers using self-report measures. Caregivers were asked to report on the child’s age, ethnicity, socioeconomic status, and placement type for sample description. Age of the youth was included in the analyses.

Procedure

Institutional review from the university and the state’s Department of Social Services Review Board approved all methods and procedures prior to data collection. For the SPARK project, data were collected across three time points (3 months apart), but only Time 1 data are included in the present study. Data collection was conducted by clinical child psychology graduate students. Consent for foster youth participation was obtained from Division of Social Services (DSS) and the Chief Judge of the Circuit Court since youth were legally in state custody. In addition to this form of consent, youth were provided with an assent form prior to data collection. Caregivers also provided consent.

After reviewing and obtaining consent, participants (youth and their caregivers) completed the study measures via an audio-computer-assisted self-interview (ACASI) on a laptop computer. This method of data collection allows for participant confidentiality and autonomy when completing the study measures (Jackson, Gabrielli, Tunno, & Hambrick, 2012). For each measure, the questions were presented to the participants on the computer screen and read aloud through headphones, and participants were asked to type in their answers on the computer. Measures completed by youth with the ACASI system included the maltreatment questionnaire and the LEC. Caregivers also completed a demographics measure using the ACASI system. Additional measures not used in the current study were included in the questionnaire battery. The KBIT-2 was administered to youth by a graduate-level student in a face-to-face interview, which occurred at different times for youth either before, after, or during breaks when completing the ACASI. In total, data collection took approximately 3 hr to complete. Youth and their caregivers were given regular breaks to ensure they were able to stay focused while completing study measures.

After participants completed all of the study measures, they completed an extensive debriefing process with the research staff. The purpose of this process was to assess for any lasting negative effects of the data collection experience, as well as to address any potential issues flagged by the ACASI system (e.g., suicidal ideation). Participants were compensated for their participation. See Jackson et al. (2012) for a more detailed report of the SPARK project and data collection methods.

Data Analysis

Path analysis procedures were conducted using structural equation modeling with a maximum likelihood estimator (MLR) with robust standard errors in the R software package 3.4.0 (R Core Team, 2014). MLR estimation was used to calculate robust standard errors and parameter estimates because it accounts for multivariate non-normality in the variables’ distribution. To test the study hypotheses that different types of PTEs are related to different IQ scores and that the associations between IQ and PTE types were moderated by age, number of PTE types (physical abuse, sexual abuse, psychological abuse, neglect, academic PTEs, family PTEs, friend PTEs, and personal PTEs) were used as predictor variables and IQ subscale scores (nonverbal and verbal) were included as outcome variables. To test the moderation of age, a Variable ×Age interaction term was added in the model for each type of PTE as predictors of verbal and nonverbal IQ. Age was treated as a continuous variable. All predictor variables were mean-centered prior to data analysis and the creation of interaction terms, as suggested improved interpretability (Kline, 2015).

Pathway coefficients used to examine the association between each PTE type, age, age and PTE interactions, and the IQ outcomes were estimated using bootstrap bias-corrected confidence intervals (CIs; MacKinnon, Lockwood, & Williams, 2004). Calculating coefficients using this method produces lower Type I error rates, while also allowing for sufficient statistical power. As a measure of model fit, the χ² test statistic, the root mean square residual error of approximation (RMSEA), the standardized root mean squared residual (SRMR), the comparative fit index (CFI), and the Tucker–Lewis Index (TLI) were examined as fit indices. Values less than 0.05 for the RMSEA, 0.08 for the SRMR were used as cutoffs for good model fit (Browne & Cudeck, 1993; Hu & Bentler, 1999). Additionally, values greater than 0.95 were used as cutoffs for model fit as indicated by the CFI and TLI test statistics (Hu & Bentler, 1999). A fully saturated model was tested first for the model with all parameters freely estimated. After estimating the saturated model to test the study hypotheses, a series of hierarchical nested structural models were evaluated to create a more parsimonious model by trimming the fully saturated model (Anderson & Gerbing, 1988). This included fixing the nonsignificant pathway coefficients in the saturated model to zero and then comparing the newly specified model with the saturated model using χ² difference tests (Chou, Bentler, & Satorra, 1991).

Results

Descriptive Statistics

All continuous variables were examined for potential skewness or kurtosis because analyses in the current study require that variables included in the models be normally distributed to minimize potential inaccuracy and biases. No variables indicated being non-normally distributed as determined by all skew indices being less than 3 and all kurtosis indices less than 10 (Kline, 2015). The results from the descriptive analysis for the variables of interest are presented in Table 1. In the current sample, psychological abuse was the most common type of maltreatment endorsed, followed by physical abuse, neglect, and sexual abuse. For other general PTE types, family PTEs was most experienced in the current sample. The second most endorsed PTE type was friend PTEs, followed by academic and personal PTEs. The standardized norms for nonverbal and verbal IQ on the KBIT are 100. In the current sample, the average nonverbal and verbal IQ scores were in the average and low average range, respectively.

Table 1.

Means, Standard Deviations, and Ranges for Study Participants.

Study Variables	Mean or percentage	SD	Min	Max
Demographics
Age	13.38	3.08	8	21
Gender (male)	53.1%
PTE types
Psychological abuse	6.55	5.00	0	15
Physical abuse	4.21	3.30	0	16
Neglect	2.88	4.14	0	22
Sexual abuse	1.68	2.71	0	12
Family traumas	9.31	2.26	0	11
Friend traumas	8.66	2.84	0	11
Academic traumas	6.07	2.64	0	10
Personal traumas	4.84	1.76	0	7
Intelligence scores
Verbal IQ	87.66	12.04	59	125
Nonverbal IQ	93.12	15.69	40	132

Path Analysis Results

Results from the path analysis can be found in Table 2. A full, saturated model was calculated by allowing all of the possible parameters to be freely estimated. Although the model fit indices cannot be interpreted because the model is saturated, the pathway coefficient can be examined and used for creating a more parsimonious model. Model trimming was conducted by systematically deleting paths between PTE type and IQ subtype based on nonsignificanct pathway coefficients in the saturated model. After each sequential fixing of the path coefficients, the new model was then compared to the saturated model and the previously trimmed model using a χ² difference test.

Table 2.

Path Analysis Results.

Fit Variables	Full model		Constrained model
χ²	0		4.22
CFI	NA		1
TLI	NA		1.12
RMSEA	NA		<0.01
SRMR	NA		0.01
Predictor Variables	Verbal IQ	Nonverbal IQ	Verbal IQ	Nonverbal IQ
Physical abuse	0.87 (0.25)* [0.42, 1.32]	0.91 (0.36)* [0.20, 1.55]	0.86 (0.23)* [0.41, 1.34]	0.88 (0.34)* [0.20, 1.47]
Sexual abuse	−0.78 (0.29)* [−1.33, −0.17]	−0.30 (0.33) [−0.99, 0.35]	−0.66 (0.27)* [−1.22, −0.12]
Psychological abuse	−0.33 (0.16)* [−0.63, 0.00]	−0.39 (0.19)* [−0.75, 0.01]	−0.35 (0.15)* [−0.65, −0.07]	−0.46 (0.18)* [−0.83, −0.09]
Neglect	−0.01 (0.16) [−0.32, 0.26]	−0.07 (0.17) [−0.40, 0.28]		−0.05 (0.17) [−0.34, 0.36]
Family PTEs	1.82 (0.34)* [1.22, 2.50]	1.36 (0.47)* [0.49, 2.44]	1.76 (0.29)* [1.22, 2.36]	1.17 (0.44)* [0.33, 2.05]
Friend PTEs	−0.06 (0.32) [−0.71, 0.57]	−0.23 (0.41) [−1.08, 0.59]
Academic PTEs	−0.36 (0.33) [−1.03, 0.29]	−0.20 (0.43) [−1.14, 0.63]	−0.27 (0.26) [−0.77, 0.28]
Personal PTEs	0.02 (0.50) [−1.07, 0.92]	1.45 (0.70) [0.00, 2.86]		1.25 (0.52)* [0.22, 2.29]
Age	−0.70 (0.24)* [−1.20, −0.20]	−0.82 (0.30)* [−1.42, −0.18]	−0.73 (0.23)* [−1.22, −0.29]	−0.92 (0.28)* [−1.52, −0.36]
Age × Physical Abuse	−0.05 (0.09) [−0.23, 0.13]	0.11 (0.13) [−0.14, 0.38]	−0.06 (0.08) [−0.23, 0.10]	0.11 (0.12) [−0.14, 0.34]
Age × Sexual Abuse	0.13 (0.11) [−0.11, 0.37]	−0.04 (0.12) [−2.42, 0.27]	0.14 (0.10) [−0.08, 0.34]
Age × Psychological Abuse	0.11 (0.06) [−0.00, 0.21]	0.03 (0.06) [−0.09, 0.16]	0.10 (0.05) [0.00, 0.21]	0.03 (0.07) [−0.09, 0.17]
Age × Neglect	−0.03 (0.05) [−0.12, 0.07]	0.16 (0.05)* [0.04, 0.26]		0.17 (0.05)* [0.05, 0.27]
Age × Family PTEs	−0.01 (0.11) [−0.20, 0.23]	0.02 (0.15) [−0.25, 0.30]	−0.02 (0.09) [−0.18, 0.18]	0.06 (0.13) [−0.18, 0.33]
Age × Friend PTEs	−0.04 (0.11) [−0.25, 0.19]	0.10 (0.13) [−0.19, 0.34]
Age × Academic PTEs	0.21 (0.13) [−0.03, 0.46]	−0.01 (0.15) [−0.30, 0.27]	0.19 (0.09)* [0.01, 0.38]
Age × Personal PTEs	−0.02 (0.17) [−0.38, 0.28]	0.12 (0.23) [−0.31, 0.58]		0.17 (0.16) [−0.16, 0.50]

Note. SDs are listed in the parentheses and 95% CI in the brackets of each regression coefficient. CFI = comparative fit index; TLI = Tucker–Lewis Index; RMSEA = root-mean-square residual error of approximation; PTE = potentially traumatic event.

*p < .05.

As theory and empirical evidence suggest, all pathways should be included in the model; thus, only those pathways with nonsignificant main effects and interaction effects were removed from the model. In the final model, the paths between verbal IQ and neglect, academic PTEs, personal PTEs, and the corresponding interaction terms were constrained to zero. Moreover, the paths between nonverbal IQ and sexual abuse, friend PTEs, academic PTEs, and the corresponding interaction terms were constrained to zero. Removal of these paths did not result in a model with poorer fit as indicated by a nonsignificant χ² difference test, χ²(12) = 4.22, p = .98). This indicates that a more parsimonious model was created. Additionally, the newly estimated model had good fit as indicated by the fit statistics (CFI = 1.00; TLI = 1.12; RMSEA = 0.00; SRMR = 0.01).

Multiple main and interaction effects were observed in the tested model. There was a main effect of age on both verbal (B = −0.73, p < .01, 95% CI [−1.22, −0.29]) and nonverbal (B = −0.92, p < .01, 95% CI = [−1.52, −0.36]) IQ scores. These main effects were qualified by interaction effects (see below). Additionally, there was a main effect on both verbal and nonverbal IQ scores for physical abuse (verbal IQ: B = 0.86, p < .01, 95% CI = [0.41, 1.34]; nonverbal IQ: B = 0.88, p < .01, 95% CI = [0.20, 1.47]), psychological abuse (verbal IQ: B = −0.35, p = .02, 95% CI = [−0.65, −0.07]; nonverbal IQ: B = −0.46, p = .01, 95% CI = [−0.83, −0.09]), and family PTEs (verbal IQ: B = 1.76, p < .01, 95% CI = [1.22, 2.36]; nonverbal IQ: B = 1.17, p < .01, 95% CI = [0.33, 2.05]). These results suggest that on average, a greater number of physical abuse events and family PTEs were associated with higher verbal and nonverbal IQ scores, whereas a greater number of psychological abuse incidents were associated with lower verbal and nonverbal IQ scores.

For verbal IQ, there was a significant main effect for sexual abuse (B = −0.66, p = .01, 95% CI = [−1.22, −0.12]), indicating that a greater number of sexual abuse incidents are associated with a lower verbal IQ score. Moreover, for nonverbal IQ, there was a significant main effect for personal PTEs (B = 1.25, p = .02, 95% CI = [0.22, 2.29]). This suggests that personal PTEs are associated with higher nonverbal IQ scores.

The main effect for age and verbal IQ was qualified by a significant interaction between academic PTEs and age (B = 0.19, p = .03, 95% CI = [0.01, 0.38]). Slope coefficients were tested at the mean age (13.38) and ± 1SD (10.30 and 16.46, respectively) and ± 2 SDs from the mean age (7.23 and 19.54, respectively) to test age across a range of age groups (school-aged, adolescence, young adult). Significance for each slope was calculated using a Wald z significance test (Kline, 2015; Preacher, Curran, & Bauer, 2006). The simple slope analysis revealed that on average, academic PTEs were associated with a lower verbal IQ score for youth aged 7.23 (B = −1.43, p = .01) and 10.30 (B = −0.85, p = .01). All other slope coefficients were nonsignificant. Additionally, there was a significant interaction for neglect and age on nonverbal IQ (B = 0.17, p < .01, 95% CI = [0.05, 0.27]). Similar to the above interaction for academic PTEs and age, simple slope analyses were conducted to test the significance of change among age groups at the mean and ± 1SD and ± 2SDs from the mean. The simple slope analysis revealed significant slopes for youth at ages 7.23 years (B = −1.09, p = .02) and 19.54 years (B = 0.99, p = .02). Results suggest that on average, experiences of neglect for younger children tended to be associated with lower nonverbal IQ scores, whereas experiences of neglect in older children were associated with higher nonverbal IQ scores. All other slope coefficients were nonsignificant.

Discussion

The purpose of the current study was to provide a better understanding of the association between different types of PTEs and intelligence in youth. Furthermore, the current study also sought to test whether or not the association between PTE type and intelligence was dependent on age. To address gaps in the literature, all types of maltreatment and general PTE were simultaneously examined to better represent the whole experience of the youth and to what extent types of events potentially influence measures of verbal and nonverbal intelligence. Results suggest that not all types of PTEs and maltreatment have a similar effect on intelligence (both nonverbal and verbal IQ). Furthermore, not all types of PTEs may influence intelligence similarly across childhood.

As hypothesized, significant differences emerged between PTE type and verbal and nonverbal intelligence. Physical abuse, psychological abuse, and family PTEs were significantly associated with both nonverbal and verbal IQ scores. Sexual abuse was only significantly associated with verbal IQ, and personal PTEs was only associated with nonverbal IQ. The findings that psychological abuse was negatively associated with both verbal and nonverbal IQ and that sexual abuse was negatively associated with verbal IQ are in line with previous research (e.g., Gould et al., 2012; Maguire et al., 2015). Interestingly, psychological abuse was the only form of PTE that was negatively associated with both verbal and nonverbal IQ. Even in the current study, when taking into account many other types of maltreatment and PTEs in the same analysis, psychological abuse still demonstrated a significant negative association with both forms of IQ. These results speak to the need for increased attention and more research on psychological abuse, especially as it pertains to cognitive development. This type of maltreatment is fairly new in the child maltreatment literature, as compared to the other main forms of maltreatment, despite evidence that suggests it is just as common and just as harmful as other abuse types (White, English, Thompson, & Roberts, 2016). Moreover, psychological abuse is also highly comorbid with other types of maltreatment and PTEs (Kaplan, Pelcovitz, & Labruna, 1999). This suggests that future research on PTE and intellectual functioning should attempt to obtain measures of psychological abuse experience when testing PTE types and IQ.

Perhaps most surprising were the positive associations between IQ and physical abuse, family PTEs, and personal PTEs. Experiences of PTEs can have profound effects on youths’ cognitive, behavioral, and social functioning (Rouse & Fantuzzo, 2009), all of which are related to intellectual functioning. Given the type of negative outcomes often expected to occur for youth following exposure to PTEs, one would not intuitively think youth would score higher IQs with more exposure to these events. However, this is not the first study to find a positive association between PTEs and IQ specifically. For example, Kira et al. (2012) reported in their analysis of IQ, trauma, and PTSD that some types of PTEs (e.g., secondary traumas such as witnessing or hearing about other’s traumatic events) were positively associated with IQ. In the current study, this may have been the case as well for family PTEs since these are a form of secondary traumas where the child does not directly experience the event. Other studies have reported similar relations between IQ and PTEs (e.g., Harpur, Polek, & van Harmelen, 2015).

Kira et al. (2012) suggest that a potential explanation for findings that PTEs are positively associated with IQ comes from PTSD symptoms associated with the events. Evidence for this claim was supported by findings that showed PTSD components of emotional detachment and dissociation mediated positive association between PTEs and IQ (Kira et al., 2012). The authors believed that symptoms such as emotional numbness and avoidance may decrease the overall impact of the PTEs on the child, which may limit the impact of the event on intellectual functioning. Furthermore, related studies examining behavioral and neural functioning suggest that dissociative symptoms, such as those associated with PTSD, are associated with heightened levels of attention and memory abilities (de Ruiter, Elzinga, & Phaf, 2006). Thus, certain PTSD symptoms, which may be the result of experiencing PTEs such as physical abuse, family, or personal PTEs, may actually be a factor buffering the impact of these events on intelligence. When examining the association between intelligence and PTEs in future studies, researchers should work to include measures of PTSD symptoms.

Additionally, Harpur, Polek, and van Harmelen (2015) suggested that another explanation for findings of positive associations between PTEs and IQ may be external factors that occur following PTE exposure, such as placement in foster care. Children who are placed in a more safe and enriching environment following abuse may have a better opportunity for positive cognitive development than in their previous home or environment. Overall, given that multiple studies have demonstrated a positive association between PTEs and IQ, future studies should work to identify potential mechanisms at work that may be influencing this connection. Although the present study did not examine the role of foster care in youth functioning, investigation of the quality of the child’s environment could be an important area for future research. Overall, the conclusion proffered from these results is not that experiencing PTEs is a positive thing but rather that the relation among certain variables with PTEs is complex. Additional research is needed to better illuminate the impact of PTE on cognitive skills and explore in greater detail why some youth experience positive outcomes and others negative outcomes.

Furthermore, the results of the current study help inform the field about the association between neglect and intellectual functioning. Although in contrast to many reports that neglected children experience greater cognitive deficits as compared to other types of maltreatment (e.g., physical abuse; Hildyard & Wolfe, 2002), neglect was not independently associated with nonverbal or verbal IQ. It may be the case that further examination into specific types of neglect is needed. Previous studies have demonstrated that different types of neglect (e.g., physical neglect, emotional neglect) are associated with different outcome trajectories such as academic and behavioral functioning (Hildyard & Wolfe, 2002). For example, physical neglect, or the failure to provide basic needs or protect a child, has often been associated with lower cognitive and intellectual functioning, as compared to emotional neglect or the inattention to a child’s emotional well-being (Polonko, 2006). Given that neglect, perhaps compared to other forms of abuse, has several distinct subtypes, it is possible that overall measures of neglect are not adequate to capture how neglect is associated with IQ scores. It is possible that future research may benefit from separating neglect into its meaningful subgroups to determine whether operationalizing educational neglect, physical neglect, or emotional neglect as distinct experiences is more meaningful for IQ function than an overall neglect score.

Instead of a direct association between neglect and IQ, the impact of neglect may depend on the age of the child. In the current study, younger children with more neglect experiences tended to receive lower IQ scores, whereas in young adults, more neglect experiences tended to receive higher IQ scores. This also appears to be the case for academic PTEs as results suggest that younger, school-aged children tend to receive lower verbal IQ scores as experiences of academic PTEs increase. There may be particular periods of childhood and certain characteristics of these developmental periods where youth are more vulnerable to experiences of neglect and academic PTEs. For example, younger children with PTE experiences may be more likely to develop mental health problems that can interfere with cognitive functioning (e.g., depression; Schoedl et al., 2010). It could also be that younger children experience more physiological changes that influence cognitive abilities, whereas older children have already reached a more mature developmental period where the impact on physiological development and cognitive functioning may not be as much, or they have had time to improve their functioning or develop methods for compensating for deficits in functioning. For example, earlier maltreatment has been reported to be associated with smaller brain size development (De Bellis et al., 1999). Another explanation may be that the influence of PTE on intelligence could attenuate or demonstrate a ceiling effect. Evidence from related research on child adjustment has provided evidence of a quadratic or nonlinear relation between PTEs and adjustment as number of PTEs increases (Horan & Widom, 2015). Overall, while some argue that experiencing PTEs at an earlier age may buffer the impact of PTEs on functioning because younger children may not have matured cognitive abilities (Maccoby, 1983), this notion is not supported in the current study. Instead, findings appear to provide evidence in favor of other theories and empirical findings that suggest younger children are more at risk than their older peers of demonstrating more indication of negative adjustment as a result of PTE experiences such as neglect (Kaplow & Widom, 2007).

Taken together, the differential results for different types of PTE exposure and IQ support the notion that generalizing all types of PTEs and maltreatment into one “score” could potentially produce misleading claims about the association between PTEs and/or maltreatment and outcomes of interest (Lau et al., 2005). It is suggested that maltreatment and PTE types be categorized and individually analyzed if a study population contains youth with different experiences of PTEs. The practice of grouping youth with all types of PTEs into one category, then comparing that group to a control group with no history of PTEs, is not uncommon (Jackson, Gabrielli, Fleming, Tunno, & Makanui, 2014). However, as the current study demonstrates, grouping methods such as these may be problematic because they may hide differences in not only what types of events are important for a certain outcome but also the direction of influence. As demonstrated in the current study, some PTEs were positively associated with IQ, whereas others were negatively associated with IQ. If these PTEs were combined together into one PTE exposure variable, it is likely no association would be detectable.

Moreover, results also revealed that both maltreatment and more general PTEs are both positively and negatively associated with verbal and nonverbal IQ scores. As previously mentioned, much of the past research on intellectual functioning and PTEs has failed to account for both experiences of maltreatment and general PTEs (e.g., Crozier & Barth, 2005). Experiences of maltreatment and general PTEs could be considered two distinct categories of PTEs (D’Andrea et al., 2012; Finkelhor, 2008), with different kinds of influence perhaps on cognitive functioning as the data in the present study suggest. Moreover, these experiences are not monolithic within types (e.g., different kinds of physical abuse or general PTEs). This demonstrates the need for inclusion of a focus on types in research on PTEs and outcomes of interest and more specifically intellectual functioning. Additionally, studies examining type of PTE should take into consideration the strong correlation between PTE types (Herrenkohl & Herrenkohl, 2009). Experiencing multiple types of PTEs is not uncommon, with the majority of youth experiencing more than one type (Finkelhor et al., 2013). This suggests that studies using models that test different types of PTEs should use data analysis techniques that account for the association between the PTE types, as was done in the present study.

Limitations

Although the results of the present study provide the field with several important new findings, they are not without limitations. One, the present study did not consider the severity of PTE exposure. Research has demonstrated that above and beyond just the kind of PTE, but how “bad” it was is likely important for predicting adjustment for youth exposed to PTEs (English et al., 2005; Jackson et al., 2014). Other characteristics of PTEs beyond type of PTE have been largely ignored in the intelligence literature, and most of the literature on PTEs and intelligence have focused on type (Kira et al., 2012; Viezel et al., 2015). Including additional information about youths’ PTE experiences, such as severity or proximity to the event, may help better explain some of the significant and nonsignificant findings. It is suggested that future researchers complete a more in-depth analysis of PTE characteristics and their association with intelligence, outside of examining PTE type exclusively. For example, future research could utilize methods such as latent class or profile analysis or expanded hierarchical analysis to classify certain groups of youth with PTE experiences based on the type, frequency, and severity of their PTE experiences (see Lau et al., 2005; Rivera, Fincham, & Bray, 2018, for more information on these techniques) or include other PTE characteristics as covariates.

Two, the current study did not include mental health symptoms in the data analysis model. As described previously, mental health symptoms, including both internalizing and externalizing, may potentially explain the association between certain types of PTEs and intelligence (e.g., PTSD; Kira et al., 2012). The type of PTE may be related to the type of mental health symptom experienced by youth, which in turn may influence intelligence differently. For example, internalizing symptoms have been shown to be more strongly associated with physical abuse and sexual abuse, as opposed to incidents of neglect (Pears, Kim, & Fisher, 2008). The development of internalizing symptoms could then contribute to intellectual and cognitive performance in youth (Masson, East-Richard, & Cellard, 2015; Rapport, Denny, Chung, & Hustace, 2001). For example, Kira et al. (2012) found that PTSD symptoms mediated the association between PTEs and IQ. Future research on intelligence and PTEs should take into account mental health symptoms by including this information in their models.

Despite these limitations, the current study is an important empirical step forward in understanding how PTEs may affect intelligence in youth. Results reaffirm previous findings that the association between IQ and different types of PTEs is diverse, as well as add new information on how the association between IQ and PTEs may depend on age. There is a need to consider how developmental factors among different age groups may contribute to outcomes of interest. The current study also highlights that the connection between PTEs and cognitive functioning in particular is not straightforward. According to results in the current study, not all types of PTEs may contribute to intellectual functioning in youth, and it may be the case that some types of PTEs could be associated with higher IQ scores. Moreover, the current study reiterates that researchers should make an effort to not categorize youth into general PTE-exposed or non-PTE-exposed groups but instead examine how different PTE types may be associated with an outcome of interest. Future research that takes into consideration these aspects, such as in the current study, will help increase understanding in the study of PTEs.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported in part by the National Institutes of Mental Health [RO1 Grant MH079252-03, 2011].

ORCID iD

Austen McGuire

References

Anderson

J. C.

Gerbing

D. W.

(1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103, 411–423.

Bain

S. K.

Jaspers

K. E.

(2010). Review of Kaufman brief intelligence test, second edition. Journal of Psychoeducational Assessment, 28, 167–174.

Barnett

Manly

J. T.

Cicchetti

(1993). Defining child maltreatment: The interface between policy and research. Child abuse, Child Development, & Social Policy, 8, 7–73.

Brody

(1997). Intelligence, schooling, and society. American Psychologist, 52, 1046–1050.

Browne

M. W.

Cudeck

(1993). Alternative ways of assessing model fit. Sage Focus Editions, 154, 136–162.

Bücker

Kapczinski

Post

Ceresér

K. M.

Szobot

Yatham

L. N.

Kauer-Sant’Anna

. (2012). Cognitive impairment in school-aged children with early trauma. Comprehensive Psychiatry, 53, 758–764.

Chou

C. P.

Bentler

P. M.

Satorra

(1991). Scaled test statistics and robust standard errors for non-normal data in covariance structure analysis: A Monte Carlo study. British Journal of Mathematical and Statistical Psychology, 44, 347–357.

Costello

E. J.

Erkanli

Fairbank

J. A.

Angold

(2002). The prevalence of potentially traumatic events in childhood and adolescence. Journal of Traumatic Stress, 15, 99–112.

Crozier

J. C.

Barth

R. P.

(2005). Cognitive and academic functioning in maltreated children. Children & Schools, 27, 197–206.

10.

D’Andrea

Ford

Stolbach

Spinazzola

van der Kolk

B. A.

(2012). Understanding interpersonal trauma in children: Why we need a developmentally appropriate trauma diagnosis. American Journal of Orthopsychiatry, 82, 187–200.

11.

De Bellis

M. D.

Keshavan

M. S.

Clark

D. B.

Casey

B. J.

Giedd

J. N.

Boring

A. M.

… Ryan

N. D

. (1999). Developmental traumatology part II: Brain development. Biological Psychiatry, 45, 1271–1284.

12.

De Bellis

M. D.

Woolley

D. P.

Hooper

S. R.

(2013). Neuropsychological findings in pediatric maltreatment: Relationship of PTSD, dissociative symptoms, and abuse/neglect indices to neurocognitive outcomes. Child Maltreatment, 18, 171–183.

13.

de Ruiter

M. B.

Elzinga

B. M.

Phaf

R. H.

(2006). Dissociation: Cognitive capacity or dysfunction? Journal of Trauma & Dissociation, 7, 115–134.

14.

De Young

A. C.

Kenardy

J. A.

Cobham

V. E.

(2011). Trauma in early childhood: A neglected population. Clinical Child and Family Psychology Review, 14, 231–250.

15.

Delaney-Black

Covington

Ondersma

S. J.

Nordstrom-Klee

Templin

Ager

… Sokol

R. J

. (2002). Violence exposure, trauma, and IQ and/or reading deficits among urban children. Archives of Pediatrics & Adolescent Medicine, 156, 280–285.

16.

DePrince

A. P.

Weinzierl

K. M.

Combs

M. D.

(2009). Executive function performance and trauma exposure in a community sample of children. Child Abuse & Neglect, 33, 353–361.

17.

Elklit

Schouwenaars

(2016). The relationship between child maltreatment and exposure to traumatic events during later adolescence and young adulthood. Scandinavian Journal of Child and Adolescent Psychiatry and Psychology, 4, 115–122.

18.

English

D. J.

The LONGSCAN Investigators. (1997). Modified Maltreatment Classification System (MMCS) Unpublished Manual. Retrieved from http://www.iprc.unc.edu/longscan/pages/mmcs/LONGSCAN%20MMCS%20Coding.pdf

19.

English

D. J.

Upadhyaya

M. P.

Litrownik

A. J.

Marshall

J. M.

Runyan

D. K.

Graham

J. C.

Dubowitz

(2005). Maltreatment’s wake: The relationship of maltreatment dimensions to child outcomes. Child Abuse & Neglect, 29, 597–619.

20.

Finkelhor

(2008). Childhood victimization. New York, NY: Oxford University Press.

21.

Finkelhor

Turner

H. A.

Shattuck

Hamby

S. L.

(2013). Violence, crime, and abuse exposure in a national sample of children and youth: An update. JAMA Pediatrics, 167, 614–621.

22.

Gould

Clarke

Heim

Harvey

P. D.

Majer

Nemeroff

C. B.

(2012). The effects of child abuse and neglect on cognitive functioning in adulthood. Journal of Psychiatric Research, 46, 500–506.

23.

Harpur

L. J.

Polek

van Harmelen

A. L.

(2015). The role of timing of maltreatment and child intelligence in pathways to low symptoms of depression and anxiety in adolescence. Child Abuse & Neglect, 47, 24–37.

24.

Herrenkohl

R. C.

Herrenkohl

T. I.

(2009). Assessing a child’s experience of multiple maltreatment types: Some unfinished business. Journal of Family Violence, 24, 485–496.

25.

Hildyard

K. L.

Wolfe

D. A.

(2002). Child neglect: Developmental issues and outcomes. Child Abuse & Neglect, 26, 679–695.

26.

Horan

J. M.

Widom

C. S.

(2015). Cumulative childhood risk and adult functioning in abused and neglected children grown up. Development and Psychopathology, 27, 927–941.

27.

L. T.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1–55.

28.

Jackson

Cushing

C. C.

Gabrielli

Fleming

O’Connor

B. M.

Huffhines

(2016). Child maltreatment, trauma, and physical health outcomes: The role of abuse type and placement moves on health conditions and service use for youth in foster care. Journal of Pediatric Psychology, 41, 28–36.

29.

Jackson

Gabrielli

Fleming

Tunno

A. M.

Makanui

P. K.

(2014). Untangling the relative contribution of maltreatment severity and frequency to type of behavioral outcome in foster youth. Child Abuse & Neglect, 38, 1147–1159.

30.

Jackson

Gabrielli

Tunno

A. M.

Hambrick

E. P.

(2012). Strategies for longitudinal research with youth in foster care: A demonstration of methods, barriers, and innovations. Children and Youth Services Review, 34, 1208–1213.

31.

Johnson

J. H.

McCutcheon

S. M.

(1980). Assessing life stress in older children and adolescents: Preliminary findings with the Life Events Checklist. Stress and Anxiety, 7, 111–125.

32.

Jones

E. D.

McCurdy

(1992). The links between types of maltreatment and demographic characteristics of children. Child Abuse & Neglect, 16, 201–215.

33.

Kaplan

S. J.

Pelcovitz

Labruna

(1999). Child and adolescent abuse and neglect research: A review of the past 10 years. Part I: Physical and emotional abuse and neglect. Journal of the American Academy of Child & Adolescent Psychiatry, 38, 1214–1222.

34.

Kaplow

J. B.

Widom

C. S.

(2007). Age of onset of child maltreatment predicts long-term mental health outcomes. Journal of Abnormal Psychology, 116, 176–187.

35.

Kaufman

A. S.

Kaufman

N. L.

(2004). Kaufman brief intelligence test, Second edition. Bloomington, MN: Pearson.

36.

Kira

Lewandowski

Somers

C. L.

Yoon

J. S.

Chiodo

(2012). The effects of trauma types, cumulative trauma, and PTSD on IQ in two highly traumatized adolescent groups. Psychological Trauma: Theory, Research, Practice, and Policy, 4, 128–139.

37.

Kline

R. B.

(2015). Principles and practice of structural equation modeling. New York, NY: Guilford Press.

38.

Lau

A. S.

Leeb

R. T.

English

Graham

J. C.

Briggs

E. C.

Brody

K. E.

Marshall

J. M.

(2005). What’s in a name? A comparison of methods for classifying predominant type of maltreatment. Child Abuse & Neglect, 29, 533–551.

39.

Maccoby

E. E.

(1983). Social-emotional development and response to stressors. In Rutter

Garmezy

(Eds.), Stress, coping, and development in children (pp. 217–234). New York, NY: McGraw-Hill.

40.

MacKinnon

D. P.

Lockwood

C. M.

Williams

(2004). Confidence limits for the indirect effect: Distribution of the product and resampling methods. Multivariate Behavioral Research, 39, 99–128.

41.

Maguire

S. A.

Williams

Naughton

A. M.

Cowley

L. E.

Tempest

Mann

M. K.

… Kemp

A. M

. (2015). A systematic review of the emotional, behavioural and cognitive features exhibited by school-aged children experiencing neglect or emotional abuse. Child: Care, Health and Development, 41, 641–653.

42.

Masson

East-Richard

Cellard

(2015). A meta-analysis on the impact of psychiatric disorders and maltreatment on cognition. Neuropsychology, 30, 143–156.

43.

McGuire

Jackson

(2018). Dimensions of maltreatment and academic outcomes for youth in foster care. Child Abuse & Neglect, 84, 82–94.

44.

Murray

Farrington

D. P.

(2010). Risk factors for conduct disorder and delinquency: Key findings from longitudinal studies. The Canadian Journal of Psychiatry, 55, 633–642.

45.

Norman

R. E.

Byambaa

Butchart

Scott

Vos

(2012). The long-term health consequences of child physical abuse, emotional abuse, and neglect: A systematic review and meta-analysis. PLOS Medicine, e1001349. doi:10.1371/journal.pmed.1001349

46.

Pears

K. C.

Kim

H. K.

Fisher

P. A.

(2008). Psychosocial and cognitive functioning of children with specific profiles of maltreatment. Child Abuse & Neglect, 32, 958–971.

47.

Polonko

K. A.

(2006). Exploring assumptions about child neglect in relation to the broader field of child maltreatment. Journal of Health and Human Services Administration, 29, 260–284.

48.

Preacher

K. J.

Curran

P. J.

Bauer

D. J.

(2006). Computational tools for probing interactions in multiple linear regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics, 31, 437–448.

49.

Rapport

M. D.

Denney

C. B.

Chung

K. M.

Hustace

(2001). Internalizing behavior problems and scholastic achievement in children: Cognitive and behavioral pathways as mediators of outcome. Journal of Clinical Child Psychology, 30, 536–551.

50.

Raviv

Taussig

H. N.

Culhane

S. E.

Garrido

E. F.

(2010). Cumulative risk exposure and mental health symptoms among maltreated youth placed in out-of-home care. Child Abuse & Neglect, 34, 742–751.

51.

Richters

J. E.

Martinez

(1993). The NIMH community violence project: I. Children as victims of and witnesses to violence. Psychiatry: Interpersonal and Biological Processes, 56, 7–21.

52.

Rivera

P. M.

Fincham

F. D.

Bray

B. C.

(2018). Latent classes of maltreatment: A systematic review and critique. Child Maltreatment, 23, 3–24.

53.

Rouse

H. L.

Fantuzzo

J. W.

(2009). Multiple risks and educational well-being: A population-based investigation of threats to early school success. Early Childhood Research Quarterly, 24, 1–14.

54.

Salmon

Bryant

R. A.

(2002). Posttraumatic stress disorder in children: The influence of developmental factors. Clinical Psychology Review, 22, 163–188.

55.

Saltzman

K. M.

Weems

C. F.

Carrion

V. G.

(2006). IQ and posttraumatic stress symptoms in children exposed to interpersonal violence. Child Psychiatry and Human Development, 36, 261–272.

56.

Schoedl

A. F.

Costa

M. C. P.

Mari

J. J.

Mello

M. F.

Tyrka

A. R.

Carpenter

L. L.

Price

L. H.

(2010). The clinical correlates of reported childhood sexual abuse: An association between age at trauma onset and severity of depression and PTSD in adults. Journal of Child Sexual Abuse, 19, 156–170.

57.

Team

R. C.

(2014). R: A language and environment for statistical computing [Internet]. 2013. Disponible sur: http://www.R-project.org.

58.

Tiet

Q. Q.

Bird

H. R.

Davies

Hoven

Cohen

Jensen

P. S.

Goodman

(1998). Adverse life events and resilience. Journal of the American Academy of Child & Adolescent Psychiatry, 37, 1191–1200.

59.

Vanderbilt-Adriance

Shaw

D. S.

(2008). Protective factors and the development of resilience in the context of neighborhood disadvantage. Journal of Abnormal Child Psychology, 36, 887–901.

60.

Vernberg

E. M.

(1999). Children’s responses to disaster: Family and systems approaches. In Gist

Lubin

(Eds.), Response to disaster: Psychosocial, community, and ecological approaches (pp. 193–210). Philadelphia, PA, US: Brunner/Mazel.

61.

Viezel

K. D.

Freer

B. D.

Lowell

Castillo

J. A.

(2015). Cognitive abilities of maltreated children. Psychology in the Schools, 52, 92–106.

62.

Vig

Kaminer

(2002). Maltreatment and developmental disabilities in children. Journal of Developmental and Physical Disabilities, 14, 371–386.

63.

White

C. R.

English

Thompson

Roberts

Y. H.

(2016). Youth self-report of emotional maltreatment: Concordance with official reports and relation to outcomes. Children and Youth Services Review, 62, 111–121.

64.

Wilson

Dolan

Smith

Casanueva

Ringeisen

(2012). NSCAW child wellbeing spotlight: Adolescents with a history of maltreatment have unique service needs that may affect their transition to adulthood (OPRE Report #2012-49). Washington, DC: Office of Planning, Research and Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services.