Maternal Childhood Maltreatment and Offspring Emotional and Behavioral Problems

Abstract

This study examined hostility and harsh discipline of both mothers and fathers as potential mechanisms explaining the association between a maternal maltreatment history and her offspring’s internalizing and externalizing problems. Prospective data from fetal life to age 6 were collected from a total of 4,438 families participating in the Generation R Study. Maternal maltreatment was assessed during pregnancy using a self-administered questionnaire. Mothers and fathers each reported on their psychological distress and harsh discipline when the child was 3 years. Children’s internalizing and externalizing problems were assessed by parental reports and child interview at age 6. Findings from structural equation modeling showed that the association between a maternal maltreatment history and her offspring’s externalizing problems was explained by maternal hostility and harsh discipline and, at least partially, also by paternal hostility and harsh discipline. Child interview data provided support for both these indirect paths, with associations largely similar to those observed for parent reports.

Keywords

child maltreatment fathers parenting child psychopathology cohort studies

The experience of maltreatment during childhood often has a negative impact on the individual’s functioning at multiple levels (Browne & Finkelhor, 1986; Cicchetti & Toth, 1995; DiLillo, 2001; Pears & Capaldi, 2001). Childhood maltreatment has been associated with substantially increased risks for adult hostility, affective symptoms (Browne & Finkelhor, 1986; Cicchetti & Toth, 1995), and impaired interpersonal functioning (DiLillo, 2001; Pears & Capaldi, 2001). For example, it has been found that mothers with a history of childhood maltreatment reported less confidence and emotional control in their role as parents and were more likely to endorse harsh disciplinary practices to their own children (DiLillo, 2001; Pears & Capaldi, 2001). According to recent findings, such repercussions of childhood maltreatment may extend beyond its immediate victims into succeeding generations (Collishaw, Dunn, O’Connor, & Golding, 2007; Lang, Gartstein, Rodgers, & Lebeck, 2010; Roberts, O’Connor, Dunn, & Golding, 2004).

Collishaw, Dunn, O’Connor, and Golding (2007) showed that children of mothers with a history of childhood maltreatment were prone to worsening emotional and behavioral problems from age 4 to 7, and this association was partially accounted for by maternal affective symptoms and hostile behavior. Similarly, Roberts, O’Connor, Dunn, and Golding (2004) showed that children of mothers with a history of childhood maltreatment were more likely than children of nonmaltreated mothers to have adjustment problems at age 4, and this association was mediated by maternal psychopathology symptoms including depression and anxiety. Children of maltreated mothers may be exposed to an accumulation of stressors that may be the consequence of their mothers’ maltreatment history and may independently affect their development. In addition to maternal affective symptoms and hostility, other contextual factors such as domestic violence, whether witnessed or directly experienced by the offspring, and mothers’ low social support may be hypothesized to operate on the association between a maternal maltreatment history and her offspring’s adjustment problems (Conger, Schofield, Neppl, & Merrick, 2013; Jaffee et al., 2013; Kitzmann, Gaylord, Holt, & Kenny, 2003).

There is mixed support for the cycle of violence hypothesis, suggesting that individuals who experienced childhood maltreatment are at elevated risk of becoming harsh parents (Thornberry, Knight, & Lovegrove, 2012). It has been demonstrated that this intergenerational continuity in harsh parenting is largely indirect, through victims’ hostile or aggressive behavior, rather than direct (Neppl, Conger, Scaramella, & Ontai, 2009). Parental harsh discipline, in turn, is among the most robust risk factors for child aggressive, delinquent, and antisocial problems—supposedly externalizing problems but has also been associated with child depressive symptoms—an internalizing problem (Chang, Schwartz, Dodge, & McBride-Chang, 2003; Gershoff, 2002; Maughan, Pickles, & Quinton, 1998; Thornberry et al., 2012; Vostanis et al., 2006). An important implication of these findings is that targeting mothers with maltreatment histories may be effective in reducing risks for psychopathology in succeeding generations. Nevertheless, the cycle of maltreatment is not inevitable and is found to be broken in most mother–child dyads (Thornberry et al., 2012). Several important questions remain about how a mother’s maltreatment history is associated with her offspring’s maladjustment.

A major gap in the literature is that previous studies addressing this association have focused primarily on maternal characteristics as mediating factors and thus have placed little emphasis on paternal characteristics. This is surprising, given the high levels of assortative mating for both affective and behavioral problems, potentially leading to high levels of similarity between partners (Krueger, Moffitt, Caspi, Bleske, & Silva, 1998; Mathews & Reus, 2001; Rhule-Louie & McMahon, 2007). In addition, there is accumulating evidence suggesting that women who experienced maltreatment in their family of origin are more likely than nonmaltreated women to have husbands or other adult partners who engage in hostile or aggressive behavior (Barnes, Noll, Putnam, & Trickett, 2009; Browne & Finkelhor, 1986; Testa, VanZile-Tamsen, & Livingston, 2005). Thus, even when mothers who experienced childhood maltreatment are not themselves harsh parents, they may expose their children to other perpetrators of violence or maltreatment (Jaffee et al., 2013). Fathers’ hostile or aggressive behaviors, being putative correlates of a maternal past history of maltreatment, have been associated with harsh parenting and subsequent offspring adjustment problems (Blazei, Iacono, & McGue, 2008; Harold et al., 2011).

Using data from a large population-based prospective cohort, the current study is one of the first to prospectively examine both maternal and paternal mechanisms linking a maternal history of childhood maltreatment to her offspring’s internalizing and externalizing problems. Based on the assortative mating theory, we first hypothesized that a maternal history of maltreatment is associated with maternal affective symptoms and hostility but also with paternal affective symptoms and hostility. Our second hypothesis was that parental hostility is associated with parental harsh discipline, which ultimately is associated with child internalizing and externalizing problems. More specifically, we hypothesized that the association between a maternal maltreatment history and her offspring’s outcomes would be mediated by maternal hostility and harsh discipline but also by paternal hostility and harsh discipline. We further expected that associations of parental affective symptoms with child internalizing or externalizing problems are not independent from the other variables. No data on domestic partner violence were available to further expand this multiple mediator model. However, we examined the potential contribution to this model of global family dysfunction, such as the family members’ inability to solve problems, to communicate in a clear and open fashion, to meet responsibilities, or to support each other.

In the present study, efforts were undertaken to meet rigorous standards for mediation and to minimize bias due to shared method or shared source variance. That is, the putative mediators and child outcomes were rated by different informants on separate occasions over a period of several years. More specifically, mothers and fathers each rated their psychological distress and harsh discipline when the child was 3. Most importantly, children’s internalizing and externalizing problems were measured with two different methods, including parent report using a behavior checklist (Child Behavior Checklist [CBCL]; Achenbach & Rescorla, 2000) and child self-report in a puppet interview (Berkeley Puppet Interview [BPI]; Ablow & Measelle, 2003; Ablow et al., 1999).

These measures of child internalizing and externalizing problems have shown to be only moderately correlated, and there is evidence suggesting that children’s ratings provide unique information to predict child problem behavior beyond that provided by parental ratings (Arseneault, Kim-Cohen, Taylor, Caspi, & Moffitt, 2005). Against this background, we decided not to combine the parental and child ratings into internalizing and externalizing constructs but to conduct two analyses with the two different source measures of the outcomes. Consistency of findings across multiple informants and methods would suggest that associations of a maternal maltreatment history and parental harsh discipline with the child outcomes are not attributable to shared method or source variance. This particularly holds when reports are based on different methods (questionnaire vs. interview) and measures are known to be only moderately correlated with one another, as was the case here. Furthermore, we included a measure of prior behavior problems as a covariate to take into account potential reciprocal effects between mediators (e.g., harsh discipline) and child outcomes. We expected that the paths of interest linking parental harsh discipline to child internalizing or externalizing problems would be consistent across parent and child reports. We further expected that these paths would be independent of prior behavior problems.

Method

Design and Study Population

This study was embedded in Generation R, a population-based cohort from fetal life onward (Jaddoe et al., 2012). Pregnant women living in the study area in Rotterdam, the Netherlands, with an expected delivery date between April 2002 and January 2006 were invited to participate. In the Generation R Study, 6,543 mothers had completed a questionnaire during pregnancy assessing past physical, emotional, and sexual maltreatment. We excluded all twins, leaving 6,384 children and their families eligible for follow-up. A total of 4,489 caregivers (70.3%), reported on child emotional and behavioral problems when the child was on average 6 years of age. We excluded 51 children aged ≥8 years, leaving a sample of 4,438 children (69.5% of the 6,384 eligible children) for analyses.

Of the 6-year-old children in this sample, 49.5% were boys and 59.8% were first born. Of their mothers (mean age = 30.97, standard deviation [SD] = 4.63), 24.9% were of non-Western national origin (10.0% Mediterranean, 8.3% Caribbean, and 6.6% other non-Western) and 9.1% were single. Of the mothers, 11.8% had a low income and 17.0% were lower educated (vs. 28.9% intermediate and 54.2% high). A total of 2,696 fathers (99.2% biological fathers) participated in this study. Of the participating biological and nonbiological fathers, 97.1% and 87.5% were living in the home with the mother and the 3-year-old child, respectively. Of the nonbiological fathers, 85.7% were involved in child rearing on a daily basis (>1 hr a day).

Procedure

Written informed consent was obtained from all participants. The Medical Ethical Committee of the Erasmus University Medical Center, Rotterdam, approved the study. Information on the timing of assessments is presented subsequently. As mentioned previously, all 4,438 children in our sample had data on a maternal history of maltreatment and the child outcomes as reported by parents. Information on a maternal history of maltreatment was obtained using a self-administered questionnaire at, on average, 20.6 (SD = 1.2) weeks of gestation. At the same assessment wave, pregnant women (n = 4,388) and their partners (n = 3,324) each rated their family’s overall functioning. Children’s internalizing and externalizing problems were examined using parent report at two assessment waves, when the child was 1.5 years (mean = 1.53, SD = 0.08) and 6 years (mean = 6.02, SD = 0.39) of age. At the age of 6, we additionally assessed internalizing and externalizing problems using child self-report in a puppet interview (n = 3,914; mean age = 6.09, SD = 0.40). In our sample of 4,438 children, a total of 3,212 mothers and 2,696 fathers each rated their harsh discipline when the child was 3 years (mean = 3.05, SD = 0.10 and mean = 3.06, SD = 0.11, respectively). At the same assessment wave, mothers and fathers each rated their psychological distress. All of the above variables were used continuously in our analyses and are presented in more detail subsequently. Additional information on cutoff scores for being in the problematic range is provided where relevant.

Measures

Maternal history of childhood maltreatment

We examined a maternal history of maltreatment via a latent construct. Indicators of this maltreatment construct included all five subscales of the 34-item short version of the Childhood Trauma Questionnaire (CTQ; Bernstein et al., 1994) collected at 20 weeks of gestation: physical abuse (5 items, α = .84), emotional abuse (5 items, α = .86), sexual abuse (6 items, α = .82), emotional neglect (10 items, α = .89), and physical neglect (8 items, α = .57). The CTQ is a validated self-report questionnaire which is commonly used in retrospective studies of childhood maltreatment. Mothers rated CTQ items on a 5-point Likert-type scale (1 = never true to 5 = very often true) to indicate the extent to which each item was true for them before age 18. The CTQ has demonstrated good test–retest reliability (Bernstein, Ahluvalia, Pogge, & Handelsman, 1997; Bernstein et al., 1994). The 28-item CTQ version (each of the above subscales is measured using 5 items; Bernstein et al., 2003) includes cutoffs for moderate to severe exposure to each form of maltreatment that we adapted to the 34-item version for descriptive purposes only; emotional abuse (score ≥ 13), physical abuse (score ≥ 10), sexual abuse (score ≥ 10), emotional neglect (score ≥ 30), and physical neglect (score ≥ 16).

Child behavioral and emotional problems

CBCL

Parents, mostly mothers (93%), filled out the CBCL/1.5–5 (Achenbach & Rescorla, 2000) at multiple assessment waves. In this study, the CBCL broadband scales internalizing and externalizing at age 6 years were included as outcomes. The CBCL total problems scale at age 1.5 years was included as a covariate. The CBCL includes 99 items on which parents rate the extent to which each statement describes their child “now or within the past 2 months” on a 3-point scale (0 = not true, 1 = somewhat or sometimes true, or 2 = very true or often true). The psychometric properties of the CBCL are well established (Achenbach & Rescorla, 2000). For the child internalizing and externalizing broadband scales, raw scores, and not T-scores, were used in regression analyses because of their greater variability. The CBCL internalizing and externalizing problem scores may be categorized as being in the borderline range based on the 83rd percentile of a Dutch norm group (Tick, van der Ende, Koot, & Verhulst, 2007).

The CBCL preschool form (age 1.5–5) as well as the CBCL school form (age 6–18) could have been applied to children in the age range of 5–7 years. Because we anticipated the majority of children to be younger than 6 years of age at the time of assessment, and assessment waves relied on sending the parents of children in this narrow age range the same questionnaire, we decided to use the CBCL preschool form for all children. As anticipated, the majority of children in this sample (59.4%) were 5 years of age at the time of assessment, and thus their ages fell within the normative range for the CBCL preschool form. Only 37.6% and 3% of the children were 6 or 7 years of age, respectively. A check of internal consistency across these age groups showed that Cronbach’s α coefficients were as high in 6- or 7-year-olds as in 5-year-olds (α = .86 for internalizing and α = .90 for externalizing in both age groups).

BPI

The BPI (Ablow & Measelle, 2003; Ablow et al., 1999), a validated semi-structured interactive interview technique to obtain self-reports from young children, was administered to 6-year-old children. In the Generation R Study, we used and validated a 50-item interview comprising two Social Scales assessing Peer Relations and six Symptomatology Scales assessing Depression, Separation Anxiety, Overanxious, Oppositional Defiant, Overt Hostility, and Conduct Problems (Ringoot et al., 2013). The first three Symptomatology Scales target internalizing problems, while the last three target externalizing problems (Ablow & Measelle, 2003).

In line with the procedures outlined by Measelle, Ablow, Cowan, and Cowan (1998), the interviews were taped for scoring by coders. A total of 37 research assistants were trained for coding BPI interviews. However, the vast majority of interviews was coded by only 12 of them. To assess reliability during the coding process, each of the assistants double coded 10% of their total number of coded interviews, with average intraclass correlation coefficients ranging from .96 (Overanxious, Overt Hostility, and Conduct Problems) to .98 (Depression and Separation Anxiety; Ringoot et al., 2013). The BPI has shown an adequate factor structure, acceptable internal consistencies (e.g., α = .71 for the internalizing and α = .77 for the externalizing broadband scale in the current study), and validity as indexed by associations with sociodemographic factors (Ringoot et al., 2013).

Potential mediators

Mothers and fathers each reported on their psychological distress and disciplinary practices when the child was 3 years of age. Parental psychological distress was measured using the Dutch version of the Brief Symptom Inventory (BSI; De Beurs, 2004; Derogatis, 1993), a validated 53-item self-report questionnaire, which is widely used in clinical and research settings. Mothers and fathers rated the extent to which each item described their feelings in the past week according to a 5-point scale ranging from 0 (not at all) to 4 (extremely). From this questionnaire, the 5-item subscale on hostility (α = .61 and .56 for mothers and fathers, respectively) was used. According to an addendum to the BSI manual (De Beurs, 2009), scores >0.60 indicate high hostility. An affective symptoms subscale was created by averaging the 6-item subscale on anxiety (α = .74 and .63) and the 6-item subscale on depression (α = .83 and .77).

Disciplinary styles of both parents were assessed using an adapted version of the Parent–Child Conflict Tactics Scale (CTS-PC; Straus, Hamby, Finkelhor, Moore, & Runyan, 1998) when the child was 3 years. Parents rated their use of discipline during the past 2 weeks on a 6-point scale ranging from “never” to “five times or more.” The categories “twice,” “3 times,” “4 times,” and “5 times” were combined due to low prevalence rates. This resulted in a 3-point scale (0 = never, 1 = once, 2 = twice or more). In the Generation R Study, factor analysis of this adapted CTS-PC has identified a harsh discipline construct consisting of 6 items (see Jansen et al., 2012), including “I shook my child,” “I shouted or screamed angrily at my child,” “I called my child names,” “I threatened to give a slap but I did not do it,” “I angrily pinched my child’s arm,” and “I called my child stupid or lazy or something like that” (α = .63 for mothers and α = .57 for fathers). The modest Cronbach’s αs reflect the small number of items and low base rates in a population-based setting. A harsh discipline sum score was calculated by adding the 6 items.

Family functioning was assessed using the General Functioning scale of the McMaster Family Assessment Device (Byles, Byrne, Boyle, & Offord, 1988), a validated 12-item self-report measure of family health and pathology, at 20 weeks of gestation. Pregnant women and their partners were each asked to rate how well each item described their family on a 4-point scale ranging from strongly disagree to strongly agree. Sample items include “We cannot get along well with each other” and “There are a great many unpleasant, painful feelings in the family.” The item scores were summed and divided by 12, yielding a family functioning sum score ranging from 1 to 4, with higher scores representing higher levels of family dysfunction (α = .90).

Covariates

Information on different indicators of socioeconomic status (SES) and sociodemographic characteristics were obtained by questionnaire during pregnancy. These socioeconomic and sociodemographic variables, which may have partially contributed to nonresponse, are described in more detail subsequently. We further included as covariates child gender and age at the assessment of outcome. As mentioned previously, we also included prior ratings of child problems to take into account potential reciprocal effects.

Family income was defined by the total net monthly income of the household and was dichotomized into above (≥€1200) versus below low-income threshold (<€1200 [equals ∼US$2,190, inflation correction of 2014]). Mothers’ highest attained educational level was classified into low (primary school or lower vocational training), middle (intermediate vocational training), and high (higher vocational training or university; Statistics Netherlands, 2004a). Maternal national origin, maternal age at intake, marital status (married or cohabiting vs. single), and parity (previous pregnancies: 0 vs. ≥1) were also considered as covariates. Maternal national origin was based on the country of birth of the mothers’ parents (Statistics Netherlands, 2004b) and was classified into Western (Europe, North America, Oceania, or Japan), Mediterranean (Morocco and Turkey), Caribbean (Surinam and Dutch Antilles), and other non-Western.

Data Analyses

The data were first explored using Statistical Package for the Social Sciences (SPSS) Version 20.0 (IBM Corporation). Bivariate correlations were conducted to explore the relations among the study variables. The primary analyses were conducted in Mplus version 6.1 (Muthén & Muthén, 1998–2010). Structural equation modeling (SEM) was employed to test the extent to which a maternal history of childhood maltreatment is associated with her child’s internalizing and externalizing problem scores at the age of 6 years, and the extent to which these associations are mediated by maternal or paternal hostility and harsh discipline.

Covariates were entered as predictors of all endogenous variables in the model (parental psychological distress, harsh discipline, and child outcomes). In addition, our overall measure of prior emotional and behavioral symptoms was entered as a predictor of parental harsh discipline and child outcomes to take into account potential reciprocal effects. Furthermore, our model took into account possible covariance among harsh discipline from mothers and fathers, psychological distress from mothers and fathers, and also among the two child outcome variables. To enhance the readability, paths of the covariates and the covariances among the variables are not shown in the figures. Furthermore, only paths that were significant at the p < .05 level are presented in the figures.

We used a robust maximum likelihood estimator, which provides test statistics robust to nonnormality. In order to clarify the findings from SEM, we conducted additional tests of specific indirect effects using the model constraint option in Mplus. These specific indirect effects were derived from both a single sample method and a bootstrap method using 5,000 samples (Preacher & Hayes, 2008). Because χ² values are sensitive to sample size, we used the root mean square error of approximation (RMSEA) and the comparative fit index (CFI) as our main indices of model fit (Browne & Cudeck, 1993; Hu & Bentler, 1999). For the RMSEA, values of .05 or lower indicate close fit, the range of .05 to .08 is interpreted as reasonable fit, the range of .08 to .10 as marginal fit, and values greater than .10 as unacceptable fit. For the CFI, values greater than .90 generally indicate a reasonably good fit.

Missing values on covariates and the putative mediators were imputed using the Markov Chain Monte Carlo multiple imputation technique with Predictive Mean Matching for continuous variables in SPSS. We generated 20 data sets with 10 iterations each. Parental psychopathology variables taken from the first wave of data collection were included as auxiliary variables to increase the precision of estimates. These 20 data sets were analyzed in Mplus and parameter estimates were averaged over the set of analyses. In our analysis with children as informants on internalizing and externalizing problems, missing data on the child outcomes (12%) were imputed to obtain comparable samples.

Response Analysis

Nonrespondents at the 6-year assessment (n = 1,946; 6,384–4,438) were more often of non-Western origin, 54.3% versus 24.9%, χ²(1) = 494.88, p < .001, n = 6,171, and were more often lower educated, 38.8% versus 17.0%, χ²(1) = 333.23, p < .001, n = 6,093, than respondents (n = 4,438). Nonrespondents more often had a low income than respondents, 36.5% versus 11.8%, χ²(1) = 432.75, p < .001, n = 5,471 due to missing data at baseline.

Results

Sample characteristics are presented in Table 1. All predictor, mediator, and outcome variables were used continuously in our analyses. However, we also present prevalences of moderate to severe maltreatment, high hostility, and borderline internalizing and externalizing problems for descriptive purposes. Of the mothers, 5.7%, 7.7%, and 5.8% had experienced moderate to severe physical, emotional, and sexual maltreatment, respectively. Furthermore, 15.3% and 2.2% had experienced moderate to severe emotional and physical neglect, respectively. Of the mothers and fathers, respectively, 4.6% and 3.9% met criteria for high hostility. Of the children, 10.7% had internalizing and 6.6% had externalizing problems in the borderline range.

Table 1.

Sample Characteristics.

	M	SD	Interquartile Range
Childhood maltreatment history
Emotional abuse, score	6.98	3.36	5.00–7.50
Physical abuse, score	5.98	2.34	5.00–6.00
Sexual abuse, score	6.66	2.23	6.00–6.00
Emotional neglect, score	20.58	8.48	14.00–25.00
Physical neglect, score	9.06	2.11	8.00–9.00
Child outcome variables
CBCL Internalizing, score	5.79	5.72	2.00–8.00
CBCL Externalizing, score	7.27	6.45	2.00–10.00
BPI Internalizing, score	58.83	12.39	49.00–66.27
BPI Externalizing, score	52.00	10.66	45.00–56.00
Mediators
Mother
Disciplinary style, score	2.08	1.84	1.00–3.00
Affective symptoms, score	0.15	0.28	0.00–0.17
Hostility, score	0.18	0.26	0.00–0.20
Family dysfunction, score	1.53	0.46	1.17–1.83
Father
Disciplinary style, score	1.78	1.79	0.00–2.00
Affective symptoms, score	0.12	0.22	0.00–0.17
Hostility, score	0.16	0.24	0.00–0.20
Family dysfunction, score	1.51	0.40	1.17–1.83

Note. n = 4,438. SD = standard deviation; CBCL = Child Behavior Checklist; BPI = Berkeley Puppet Interview.

The Measurement Model

Before testing our structural model, we performed confirmatory factor analysis (CFA) to establish the validity of our proposed maltreatment latent construct. CFA showed a good fit to the data for this measurement model, χ²(5) = 77.59; RMSEA = .057; CFI = .97, n = 4,438. Furthermore, factor loadings for the maltreatment indicators were reasonably strong and statistically significant, with highest loadings for emotional and physical abuse and lowest loadings for emotional and physical neglect (all ps < .001; see Figure 1). Thus, CFA indicated that it was acceptable to employ the proposed latent maltreatment construct in the remaining analyses.

Figure 1.

Multiple mediation model of a maternal history of childhood maltreatment and parents’ report of child internalizing and externalizing problems. Note. CBCL = Child Behavior Checklist. Values are standardized coefficient estimates. Paths with regression coefficients significant at the p < .05 (*), p < .01 (**), or p < .001 (***) level are highlighted by solid lines. Covariates included prior child behavior problems, the child’s age at the assessment of outcome and gender, parity, mother’s age at intake, national origin of mother, marital status, low family income (≤€1,200 [equals ∼US$2,190, inflation correction of 2014]), and education of mother. In order to improve readability, paths for these covariates and also for covariances among the variables are not shown in the figure.

Model Testing

Figure 1 shows the standardized path coefficients of our model predicting parent reports of offspring internalizing and externalizing problems. As expected, parent and child reports of internalizing and externalizing problems were only moderately correlated (see Table 2). This justifies our decision not to aggregate the maternal and child ratings but to conduct our analysis twice with the two different source measures of internalizing and externalizing problems. The results of our separate analysis predicting children’s self-reports of internalizing and externalizing problems are shown in Figure 2. Neither mother report nor father report of family dysfunction was associated with these child-reported outcomes independently from the other variables. Hence, this scale was not added to the multiple mediator model.

Table 2.

Bivariate Correlations for Study Variables.

	1	2	3	4	5	6	7	8	9	10	11
1. Maternal maltreatment	—
2. CBCL internalizing	.15	—
3. CBCL externalizing	.12	.71	—
4. BPI internalizing	.05	.13	.09	—
5. BPI externalizing	.01	.07	.16	.29	—
6. Maternal discipline	.13	.17	.28	.10	.12	—
7. Maternal affective	.22	.24	.21	.09	.04	.22	—
8. Maternal hostility	.19	.24	.24	.08	.03	.31	.64	—
9. Paternal discipline	.07	.13	.23	.05	.14	.41	.12	.16	—
10. Paternal affective	.14	.18	.16	.10	.05	.11	.28	.24	.18	—
11. Paternal hostility	.12	.16	.17	.09	.06	.15	.24	.28	.27	.58	—

Note. CBCL = Child Behavior Checklist; BPI = Berkeley Puppet Interview. Estimates highlighted in boldface were statistically significant at the p < .05 level.

Figure 2.

Multiple mediation model of a maternal history of childhood maltreatment and child self-reported internalizing and externalizing problems. Note. BPI = Berkeley Puppet Interview. Values are standardized coefficient estimates. Paths with regression coefficients significant at the p < .05 (*), p < .01 (**), or p < .001 (***) level are highlighted by solid lines. Covariates included prior child behavior problems, the child’s age at the assessment of outcome and gender, parity, mother’s age at intake, national origin of mother, marital status, low family income (≤€1,200 [equals ∼US$2,190, inflation correction of 2014]), and education of mother. In order to improve readability, paths for these covariates and also for covariances among the variables are not shown in the figure.

SEM showed a good fit to the data presented in Figure 1, χ²(105) = 1,185.93; RMSEA = .048; CFI = .92, n = 4,438, and Figure 2, χ²(105) = 1,174.62, RMSEA = .048; CFI = .91, n = 4,438. Although a maternal maltreatment history did not have a significant direct effect on either parent or child reports of externalizing problems, there were multiple indirect paths to both parent and child reports of externalizing problem scores, which are presented in detail subsequently. Coefficients of all paths linking study variables, regardless of statistical significance, are shown in Table 3.

Table 3.

Path Coefficients for the Associations of a Maternal Maltreatment History, Putative Mediators, and Child Outcomes.

	B (SE)	β	p Value
Putative mediators
Maternal hostility ON
Maternal maltreatment	0.02 (0.00)	.18	<.001
Maternal affective ON
Maternal maltreatment	0.02 (0.00)	.19	<.001
Paternal hostility ON
Maternal maltreatment	0.01 (0.00)	.10	.001
Paternal affective ON
Maternal maltreatment	0.01 (0.00)	.12	<.001
Maternal harsh discipline ON
Maternal maltreatment	0.02 (0.01)	.04	.084
Maternal hostility	1.53 (0.17)	.23	<.001
Maternal affective	−0.04 (0.17)	−.01	.826
Paternal harsh discipline ON
Maternal maltreatment	0.01 (0.02)	.02	.409
Paternal hostility	1.53 (0.20)	.22	<.001
Paternal affective	0.18 (0.21)	.02	.398
Child outcomes
Parent-reported Internalizing ON
Maternal maltreatment	0.09 (0.04)	.04	.020
Maternal hostility	1.74 (0.64)	.09	.006
Maternal affective	1.24 (0.57)	.07	.029
Maternal harsh discipline	0.05 (0.07)	.02	.466
Paternal hostility	0.25 (0.54)	.01	.644
Paternal affective	1.26 (0.67)	.05	.058
Paternal harsh discipline	0.11 (0.08)	.03	.165
Parent-reported Externalizing ON
Maternal maltreatment	0.04 (0.04)	.02	.355
Maternal hostility	1.91 (0.60)	.08	.001
Maternal affective	0.58 (0.58)	.03	.314
Maternal harsh discipline	0.39 (0.07)	.12	<.001
Paternal hostility	0.34 (0.58)	.01	.558
Paternal affective	1.03 (0.65)	.04	.112
Paternal harsh discipline	0.31 (0.08)	.09	<.001
Child-reported Internalizing ON
Maternal maltreatment	−0.02 (0.07)	−.00	.820
Maternal hostility	1.04 (1.22)	.02	.396
Maternal affective	0.13 (1.01)	.00	.897
Maternal harsh discipline	0.22 (0.14)	.03	.117
Paternal hostility	1.50 (1.34)	.03	.264
Paternal affective	1.26 (1.59)	.02	.427
Paternal harsh discipline	0.11 (0.17)	.02	.511
Child-reported Externalizing ON
Maternal maltreatment	−0.05 (0.06)	−.01	.470
Maternal hostility	−1.01 (1.03)	−.03	.326
Maternal affective	0.51 (0.85)	.02	.549
Maternal harsh discipline	0.39 (0.13)	.07	.004
Paternal hostility	0.18 (1.23)	.00	.883
Paternal affective	0.67 (1.29)	.02	.605
Paternal harsh discipline	0.47 (0.15)	.08	.002

Note. Covariates included prior child behavior problems, the child’s age at the assessment of outcome and gender, parity, mother’s age at intake, national origin of mother, marital status, family income, and education of mother.

Figures 1 and 2 both illustrate that having a history of childhood maltreatment placed mothers at greater risk for hostility (β = .18, p < .001) and affective symptoms (β = .19, p < .001). Further, consistent with our first hypothesis, mothers with a childhood history of maltreatment were more likely to partner with a man who also had higher levels of hostility (β = .10, p = .001) and affective symptoms (β = .12, p < .001). A mother’s maltreatment history was unrelated to her own harsh discipline and to her partner’s harsh discipline after adjusting for prior levels of child problem behavior and other covariates. However, we confirmed our second hypothesis in that hostile mothers or fathers were more likely than nonhostile mothers or fathers to endorse harsh disciplinary practices to their children (β = .23, p < .001 for mothers; β = .22, p < .001 for fathers), which ultimately predicted higher offspring externalizing problems. More specifically, mothers’ harsh discipline (β = .12, p < .001; β = .07, p = .004) and fathers’ harsh discipline (β = .09, p < .001; β = .08, p = .002) each independently increased the risk of parent-reported as well as child-reported externalizing symptoms 3 years later (see Figures 1 and 2). Interestingly, the pathway of interest linking parental harsh discipline to child externalizing problems was consistent across informants. However, mothers’ affective symptoms (β = .07, p = .029) were independently associated with parent reports but not child-self reports of internalizing problems. Similarly, maternal hostility was independently associated with child internalizing (β = .09, p = .006) and externalizing (β = .08, p = .001) problems in the analysis using parent reports but not in the analysis using child self-reports.

Specific indirect effects derived from a single sample method and a bootstrap method using 5,000 samples (results from the single sample method are shown here) each confirmed our hypothesis that a maternal maltreatment history has indirect associations with both parental and child reports of externalizing problems through maternal hostility and maternal harsh discipline (b = 0.010, p < .001; b = 0.010, p = .014) and through fathers’ hostility and fathers’ harsh discipline (b = 0.004, p = .004; b = 0.007, p = .023).

Additional Model Testing

In a first additional analysis, we excluded children of single mothers (9.1%). In a second additional analysis, we excluded children of non-Western mothers (24.9%). Consistency of findings would suggest that effects were not driven by single-parent families or by non-Western families. Indeed, both additional analyses showed only small changes in effect sizes (data not shown). In addition, very similar mediating pathways were observed.

Discussion

Findings from this prospective population-based study support previous research suggesting that the negative consequences of maltreatment may extend beyond its immediate victims into succeeding generations (Collishaw et al., 2007; Roberts et al., 2004). The current study is one of the first to show that the link between a maternal maltreatment history and her offspring’s externalizing problems may be explained by maternal hostility and harsh discipline and, at least partially, also by paternal hostility and harsh discipline. The findings of this study concur with those of other investigators who found that the intergenerational continuity in harsh parenting is largely indirect, through victims’ hostile behavior (Neppl et al., 2009). Interestingly, a maternal maltreatment history was associated not only with maternal hostility and affective symptoms but also with paternal hostility and affective symptoms. The individual domains of hostility and affective symptoms were correlated between the parents. These results are consistent with other studies in suggesting that assortative mating, indexed by the correlation between parents, may exist for both emotional and behavioral problems (Krueger et al., 1998; Mathews & Reus, 2001; Rhule-Louie & McMahon, 2007).

Child interview data provided support for the two hypothesized pathways linking a maternal maltreatment history to her offspring’s externalizing problem scores through maternal and paternal hostility and harsh discipline, with findings largely similar to those observed for parents’ report on a behavior checklist. Parental hostility and harsh discipline appeared to contribute to children’s externalizing problems above and beyond other risk factors including parental affective symptoms and family dysfunction. Some of the other paths, for instance, those linking maternal affective symptoms to offspring internalizing problems, were not supported by the child interview data. Similarly, a maternal maltreatment history did not have a significant overall or direct effect on internalizing or externalizing problems as reported by the child. However, the presence of a significant overall effect is not always necessary for mediation to occur (MacKinnon, Krull, & Lockwood, 2000). When studying a mediation model in which the predictor and outcome are not temporally proximal, as is the case here with a maternal history of childhood maltreatment and offspring adjustment problems at age 6, one may be able to detect mediation but not an overall or direct effect (Shrout & Bolger, 2002).

The previous findings suggest that although the possibility of shared method variance inflating some of the associations cannot be excluded, it probably does not account for the pathways of interest linking maternal or paternal harsh discipline to offspring outcomes. It is also unlikely that associations between a maternal maltreatment history and father reports of psychological distress are solely attributable to shared method variance. In addition, given that maternal maltreatment histories, potential mediators and child outcomes were assessed on separate occasions with a considerable time interval between them, it is unlikely that the associations detected are an artifact of shared variance due to current affective state. Furthermore, we showed that associations of parental harsh discipline with offspring externalizing problems were independent of prior levels of child problem behavior. In addition, the consistency of the study findings across samples including or excluding single mothers and non-Western mothers indicates robust findings.

In our multiple mediator model, maternal and paternal harsh discipline appeared to be associated primarily with child externalizing problems and had less impact on internalizing problems. There is an abundant literature demonstrating that parental harshness is associated with a range of externalizing problems including aggressive, antisocial, as well as delinquent behaviors (Gershoff, 2002; Thornberry et al., 2012). In addition, parental harshness has been associated with internalizing problems and depressive symptoms in particular (Gershoff, 2002; Thornberry et al., 2012). Future research in this field may focus on a more specific dimension rather than on the broad emotional dimension of internalizing problems that was used in the current study. In other words, future research may investigate whether mediation by parental hostility and harsh discipline occurs for specific internalizing problems, particularly depression.

In this study, we used a population-based sample and not a clinical sample where participants who are substantially impaired or suffer severe distress following childhood maltreatment are strongly represented. Population-based studies have strong potential for obtaining representative samples of families and their children and provide results that may be generalizable to the broader population. For example, the mean scores on the hostility scales for mothers and fathers in this study were within 1 SD reported for other general population samples (De Beurs, 2009). However, although large and sociodemographically diverse, the Generation R Study was not designed to be a nationally representative of all children and their families. Despite the caution that must be placed on comparing rates of maltreatment and child behavior across studies that differ in sample composition or methodology, the patterns of findings were largely comparable to those obtained in other population-based studies (Collishaw et al., 2007; Roberts et al., 2004).

As other researchers conducting population-based studies (Roberts et al., 2004), we found that the standardized effect sizes of the observed associations were small in magnitude. Because of the complex multifactorial nature of childhood maltreatment and psychological distress, the impact of single determinants is deemed to be relatively weak. In addition, the fact that a maternal maltreatment history, potential mediators, and offspring psychopathology were all assessed on separate occasions over a period of several years and that multiple informants and methods were used (i.e., mother report and father report and child interview) implies that rigorous standards for mediation were met. As such, although small in magnitude, any observed association is potentially important in understanding how a maternal maltreatment history is associated with her offspring’s psychopathology.

Despite its strengths, including the use of a large population-based sample, a prospective design, and multiple informant ratings, the current study must be interpreted within the context of several limitations. First, a maternal history of childhood maltreatment was determined using a validated but retrospective questionnaire, the CTQ. Retrospective reports of child maltreatment may be influenced by recall bias (Brewin, Andrews, & Gotlib, 1993). Studies have shown that this recall bias primarily involves underreporting—that is, individuals who experienced childhood maltreatment fail to report this maltreatment in adulthood (Fergusson, Horwood, & Woodward, 2000; Hardt & Rutter, 2004). Underreporting of maltreatment in this study would have led to an underestimation of the prevalence and correlates of maltreatment. Of particular concern is the possibility that, compared with nondepressed individuals, depressed individuals may be more likely to recall negative experiences such as maltreatment. Under these circumstances, the association between a mother’s maltreatment history and adult psychological distress would have been overestimated. Although maternal retrospective reports of her maltreatment history could not be validated against official records, we were able to demonstrate their associations with both maternal and paternal reports of emotional status.

Another limitation is that, like other cohort studies, the Generation R Study is prone to selective dropout. Our response analysis showed that selection occurred toward high functioning families with higher SES. We included as covariates those variables that were likely to have contributed to nonresponse, including family income, education, and national origin. Although it is certain that selective dropout has an impact on statistical power, a recent study and simulations on the Avon Longitudinal Study of Parents and Children sample showed that this does not necessarily affect the validity of regression models with regard to disruptive behavior (Wolke et al., 2009). Furthermore, there may be additional variables not included in this study, such as domestic partner violence, that mediate the association of a maternal maltreatment history and her offspring’s externalizing problems. In addition, children may share genetic liability to dysfunction with their parents. For example, associations of parental hostility or harshness with offspring externalizing problems may be due to shared genes.

Finally, it is a limitation of the current study that no data were obtained on whether mothers’ maltreatment experiences were intra- or extrafamilial, or to whom mothers’ own hostility was directed. Evidence suggests that maltreatment by relatives is sometimes, but not always, associated with greater trauma than maltreatment by nonrelatives (for a review, see Browne & Finkelhor, 1986). Against this background, mothers with maltreatment experiences involving relatives may show poorer mental health and parenting than mothers with experiences involving any other type of perpetrator. Furthermore, associations may have been stronger had all mothers directed their hostility to their children. Although we had no direct measure of this, we were able to show a significant association between a mother’s hostility and her harshness specifically directed toward her child.

Data on paternal histories of childhood maltreatment were not available in the current study, and future research may investigate whether child outcomes differ depending on whether the father, mother, or both were victims of maltreatment. In the absence of data on mating processes, we can only speculate that assortative mating processes explain these observed links between a maternal maltreatment history and psychological distress in later family life. Future research may explain why or how mothers with a history of childhood maltreatment come to affiliate with hostile men. Although partner similarity is often thought to arise from partner selection processes or assortative mating, it may also result from the subsequent influence partners have on each other’s behavior.

From this study, several recommendations can be made for intervention and prevention. The construct of harsh discipline as used in this study is not completely the same as child maltreatment, although overlap certainly exists. The line between harsh discipline and maltreatment or abuse is complex and tends to be ill defined (Coleman, Dodge, & Campbell, 2010). If anything, the use of harsh disciplinary strategies may predispose parents to abusive behavior (Jackson et al., 1999). In the current study, we had no means of testing whether the children in our sample were maltreated through, for instance, the use of official records. However, instances of yelling, pinching, and threatening appeared to play a key role in the association between a maternal maltreatment history and her offspring’s externalizing problems, suggesting that children at risk due to their mothers’ experiences of childhood maltreatment may be identified as in need of services before their parents’ disciplinary strategies become so severe as to require more intensive intervention.

Policies to improve the psychological distress of mothers but also the psychological distress of fathers (even if not qualifying for a clinical diagnosis) may help interrupt, at least partially, the tendency of a maternal maltreatment history to affect the succeeding generation. However, we showed that the associations of parental hostility with child externalizing problems were not direct but rather indirect through parental harsh discipline. Parent training is widely considered an effective intervention for child internalizing and externalizing problems. Several evidence-based parent training programs exist, among which self-administered written programs or bibliotherapy for parents of 3- to 6-year-old children (Forehand, Merchant, Long, & Garai, 2010). Despite the evidence supporting the important role of fathers in child development, fathers continue to be underrepresented in such parent training programs (Bagner & Eyberg, 2003; Lundahl, Tollefson, Risser, & Lovejoy, 2008). In a meta-analysis combining several estimations on father involvement in parent training, it was concluded that the inclusion of fathers increases the likelihood of more positive outcomes including child behavior and parenting practices (Lundahl et al., 2008). Furthermore, father involvement may serve to maintain the treatment gains in the family (Bagner & Eyberg, 2003). By showing that the negative consequences of maltreatment may extend beyond its immediate victims into succeeding generations through maternal hostility and harsh discipline as well as through paternal hostility and harsh discipline, the findings of the current study suggest that in the future, more attention should be directed to fathers.

Footnotes

Acknowledgments

The Generation R Study is conducted by the Erasmus Medical Center Rotterdam in close collaboration with the Faculty of Social Sciences of the Erasmus University Rotterdam, the Municipal Health Service Rotterdam area, the Rotterdam Homecare Foundation, and the “Stichting Trombosedienst & Artsen-laboratorium Rijnmond” (STAR) in Rotterdam. We gratefully acknowledge the contribution of participating children and their parents, general practitioners, midwives, and pharmacies in Rotterdam. The general design of the Generation R Study is made possible by the Erasmus Medical Center Rotterdam, the Erasmus University Rotterdam, and the Netherlands Organization for Health Research and Development (ZonMw, grant no. 10.000.1003).

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Frank C. Verhulst is the contributing editor of the Achenbach System of Empirically Based Assessment, from which he receives remuneration. The other authors declare that they have no conflict of interest.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by an additional grant from the Netherlands Organization for Health Research and Development (ZonMw, grant no. 15.700.1004).

References

Ablow

J. C.

Measelle

J. R.

(Eds.). (2003). Manual for the Berkeley puppet interview: Symptomatology, social, and academic modules (BPI 1.0). MacArthur Foundation Research Network Psychopathology and Development (David J. Kupfer, Chair): The MacArthur Working Group on Outcome Assessment: University of Pittsburgh.

Ablow

J. C.

Measelle

J. R.

Kraemer

H. C.

Harrington

Luby

Smider

Kupfer

D. J.

(1999). The MacArthur three-city outcome study: Evaluating multi-informant measures of young children’s symptomatology. Journal of the American Academy of Child and Adolescent Psychiatry, 38, 1580–1590.

Achenbach

T. M.

Rescorla

L. A.

(2000). Manual for ASEBA preschool forms & profiles. Burlington: University of Vermont Center for Children, Youth, & Families.

Arseneault

Kim-Cohen

Taylor

Caspi

Moffitt

T. E.

(2005). Psychometric evaluation of 5- and 7-year-old children’s self-reports of conduct problems. Journal of Abnormal Child Psychology, 33, 537–550.

Bagner

D. M.

Eyberg

S. M.

(2003). Father involvement in parent training: When does it matter? Journal of Clinical Child and Adolescent Psychology, 32, 599–605.

Barnes

J. E.

Noll

J. G.

Putnam

F. W.

Trickett

P. K.

(2009). Sexual and physical revictimization among victims of severe childhood sexual abuse. Child Abuse & Neglect, 33, 412–420.

Bernstein

D. P.

Ahluvalia

Pogge

Handelsman

(1997). Validity of the childhood trauma questionnaire in an adolescent psychiatric population. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 340–348.

Bernstein

D. P.

Fink

Handelsman

Foote

Lovejoy

Wenzel

Ruggiero

(1994). Initial reliability and validity of a new retrospective measure of child abuse and neglect. The American Journal of Psychiatry, 151, 1132–1136.

Bernstein

D. P.

Stein

J. A.

Newcomb

M. D.

Walker

Pogge

Ahluvalia

Zule

(2003). Development and validation of a brief screening version of the childhood trauma questionnaire. Child Abuse & Neglect, 27, 169–190.

10.

Blazei

R. W.

Iacono

W. G.

McGue

(2008). Father-child transmission of antisocial behavior: The moderating role of father’s presence in the home. Journal of the American Academy of Child and Adolescent Psychiatry, 47, 406–415.

11.

Brewin

C. R.

Andrews

Gotlib

I. H.

(1993). Psychopathology and early experience: A reappraisal of retrospective reports. Psychological Bulletin, 113, 82–98.

12.

Browne

Finkelhor

(1986). Impact of child sexual abuse: A review of the research. Psychological Bulletin, 99, 66–77.

13.

Browne

M. W.

Cudeck

(1993). Alternative ways of assessing model fit. In Bollen

K. A.

Long

J. S.

(Eds.), Testing structural equation models (pp. 136–162). Newbury Park, CA: Sage.

14.

Byles

Byrne

Boyle

M. H.

Offord

D. R.

(1988). Ontario child health study: Reliability and validity of the general functioning subscale of the McMaster family assessment device. Family Process, 27, 97–104.

15.

Chang

Schwartz

Dodge

K. A.

McBride-Chang

(2003). Harsh parenting in relation to child emotion regulation and aggression. Journal of Family Psychology, 17, 598–606.

16.

Cicchetti

Toth

S. L.

(1995). A developmental psychopathology perspective on child abuse and neglect. Journal of the American Academy of Child and Adolescent Psychiatry, 34, 541–565.

17.

Coleman

D. L.

Dodge

K. A.

Campbell

S. K.

(2010). Where and how to draw the line between reasonable corporal punishment and abuse. Law and Contemporary Problems, 73, 107–166.

18.

Collishaw

Dunn

O’Connor

T. G.

Golding

(2007). Maternal childhood abuse and offspring adjustment over time. Development and Psychopathology, 19, 367–383.

19.

Conger

R. D.

Schofield

T. J.

Neppl

T. K.

Merrick

M. T.

(2013). Disrupting intergenerational continuity in harsh and abusive parenting: The importance of a nurturing relationship with a romantic partner. Journal of Adolescent Health, 53, S11–S17.

20.

De Beurs

(2004). Brief symptom inventory-manual. Leiden, The Netherlands: Pits.

21.

De Beurs

(2009). Brief symptom inventory-manual addendum. Leiden, The Netherlands: Pits.

22.

Derogatis

L. R.

(1993). Brief Symptom Inventory (BSI): Administration, scoring and procedures-manual (3rd. ed.). Minneapolis, MN: National Computer Systems.

23.

DiLillo

(2001). Interpersonal functioning among women reporting a history of childhood sexual abuse: Empirical findings and methodological issues. Clinical Psychology Review, 21, 553–576.

24.

Fergusson

D. M.

Horwood

L. J.

Woodward

L. J.

(2000). The stability of child abuse reports: A longitudinal study of the reporting behaviour of young adults. Psychological Medicine, 30, 529–544.

25.

Forehand

R. L.

Merchant

M. J.

Long

Garai

(2010). An examination of parenting the strong-willed child as bibliotherapy for parents. Behavior Modification, 34, 57–76.

26.

Gershoff

E. T.

(2002). Corporal punishment by parents and associated child behaviors and experiences: A meta-analytic and theoretical review. Psychological Bulletin, 128, 539–579.

27.

Hardt

Rutter

(2004). Validity of adult retrospective reports of adverse childhood experiences: Review of the evidence. Journal of Child Psychology and Psychiatry, 45, 260–273.

28.

Harold

G. T.

Rice

Hay

D. F.

Boivin

van den Bree

Thapar

(2011). Familial transmission of depression and antisocial behavior symptoms: Disentangling the contribution of inherited and environmental factors and testing the mediating role of parenting. Psychological Medicine, 41, 1175–1185.

29.

Bentler

P. M.

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

30.

Jackson

Thompson

R. A.

Christiansen

E. H.

Colman

R. A.

Wyatt

Buckendahl

C. W.

Peterson

(1999). Predicting abuse-prone parental attitudes and discipline practices in a nationally representative sample. Child Abuse & Neglect, 23, 15–29.

31.

Jaddoe

V. W.

van Duijn

C. M.

Franco

O. H.

van der Heijden

A. J.

van IJzendoorn

M. H.

de Jongste

J. C.,

Hofman

(2012). The generation R study: Design and cohort update 2012. European Journal of Epidemiology, 27, 739–756.

32.

Jaffee

S. R.

Bowes

Ouellet-Morin

Fisher

H. L.

Moffitt

T. E.

Merrick

M. T.

Arseneault

(2013). Safe, stable, nurturing relationships break the intergenerational cycle of abuse: A prospective nationally representative cohort of children in the United Kingdom. Journal of Adolescent Health, 53, S4–S10.

33.

Jansen

P. W.

Raat

Mackenbach

J. P.

Hofman

Jaddoe

V. W.

Bakermans-Kranenburg

M. J.

Tiemeier

(2012). Early determinants of maternal and paternal harsh discipline: The Generation R Study. Family Relations, 61, 253–270.

34.

Kitzmann

K. M.

Gaylord

N. K.

Holt

A. R.

Kenny

E. D.

(2003). Child witness to domestic violence: A meta-analytic review. Journal of Consulting and Clinical Psychology, 71, 339–352.

35.

Krueger

R. F.

Moffitt

T. E.

Caspi

Bleske

Silva

P. A.

(1998). Assortative mating for antisocial behavior: Developmental and methodological implications. Behavior Genetics, 28, 173–186.

36.

Lang

A. J.

Gartstein

M. A.

Rodgers

C. S.

Lebeck

M. M.

(2010). The impact of maternal childhood abuse on parenting and infant temperament. Journal of Child and Adolescent Psychiatric Nursing, 23, 100–110.

37.

Lundahl

B. W.

Tollefson

Risser

Lovejoy

M. C.

(2008). A meta-analysis of father involvement in parent training. Research on Social Work Practice, 18, 97–106.

38.

MacKinnon

D. P.

Krull

J. L.

Lockwood

C. M.

(2000). Equivalence of the mediation, confounding, and suppression effect. Prevention Science, 1, 173–181.

39.

Mathews

C. A.

Reus

V. I.

(2001). Assortative mating in the affective disorders: A systematic review and meta-analysis. Comprehensive Psychiatry, 42, 257–262.

40.

Maughan

Pickles

Quinton

(1998). Parental hostility, childhood behavior, and adult social functioning. In McCord

(Ed.), Coercion and punishment in long-term perspectives (pp. 34–58). New York, NY: Cambridge University Press.

41.

Measelle

J. R.

Ablow

J. C.

Cowan

P. A.

Cowan

C. P.

(1998). Assessing young children’s views of their academic, social, and emotional lives: An evaluation of the self-perception scales of the Berkeley Puppet Interview. Child Development, 69, 1556–1576.

42.

Muthén

L. K.

Muthén

B. O.

(1998-2010). Mplus user’s guide (6th ed.). Los Angeles, CA: Author.

43.

Neppl

T. K.

Conger

R. D.

Scaramella

L. V.

Ontai

L. L.

(2009). Intergenerational continuity in parenting behavior: Mediating pathways and child effects. Developmental Psychology, 45, 1241–1256.

44.

Pears

K. C.

Capaldi

D. M.

(2001). Intergenerational transmission of abuse: A two-generational prospective study of an at-risk sample. Child Abuse & Neglect, 25, 1439–1461.

45.

Preacher

K. J.

Hayes

A. F.

(2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40, 879–891.

46.

Rhule-Louie

D. M.

McMahon

R. J.

(2007). Problem behavior and romantic relationships: Assortative mating, behavior contagion, and desistance. Clinical Child and Family Psychology Review, 10, 53–100.

47.

Ringoot

A. P.

Jansen

P. W.

Steenweg-de Graaff

Measelle

J. R.

van der Ende

Raat

Tiemeier

(2013). Young children’s self-reported emotional, behavioral and peer problems: The Berkeley Puppet Interview. Psychological Assessment, 25, 1273–1285.

48.

Roberts

O’Connor

Dunn

Golding

(2004). The effects of child sexual abuse in later family life; Mental health, parenting and adjustment of offspring. Child Abuse & Neglect, 28, 525–545.

49.

Shrout

P. E.

Bolger

(2002). Mediation in experimental and nonexperimental studies: New procedures and recommendations. Psychological Methods, 7, 422–445.

50.

Statistics Netherlands. (2004a). Standaard Onderwijsindeling 2003 [Standard Classification of Education 2003]. Voorburg/Heerlen, The Netherlands: Author.

51.

Statistics Netherlands. (2004b). Allochtonen in Nederland 2004 [Migrants in the Netherlands 2004]. Voorburg/Heerlen, The Netherlands: Author.

52.

Straus

M. A.

Hamby

S. L.

Finkelhor

Moore

D. W.

Runyan

(1998). Identification of child maltreatment with the parent-child conflict tactics scales: Development and psychometric data for a national sample of American parents. Child Abuse & Neglect, 22, 249–270.

53.

Testa

VanZile-Tamsen

Livingston

J. A.

(2005). Childhood sexual abuse, relationship satisfaction, and sexual risk taking in a community sample of women. Journal of Consulting and Clinical Psychology, 73, 1116–1124.

54.

Thornberry

T. P.

Knight

K. E.

Lovegrove

P. J.

(2012). Does maltreatment beget maltreatment? A systematic review of the intergenerational literature. Trauma, Violence, & Abuse, 13, 135–152.

55.

Tick

N. T.

van der Ende

Koot

H. M.

Verhulst

F. C.

(2007). 14-year changes in emotional and behavioral problems of very young Dutch children. Journal of the American Academy of Child and Adolescent Psychiatry, 46, 1333–1340.

56.

Vostanis

Graves

Meltzer

Goodman

Jenkins

Brugha

(2006). Relationship between parental psychopathology, parenting strategies and child mental health—Findings from the GB national study. Social Psychiatry and Psychiatric Epidemiology, 41, 509–514.

57.

Wolke

Waylen

Samara

Steer

Goodman

Ford

Lamberts

(2009). Selective drop-out in longitudinal studies and non-biased prediction of behaviour disorders. The British Journal of Psychiatry, 195, 249–256.