Child Psychiatry Hum Dev DOI 10.1007/s10578-015-0535-1

ORIGINAL ARTICLE

Effects of Callous–Unemotional Traits on the Association Between Parenting and Child Conduct Problems Kathleen I. Crum • Daniel A. Waschbusch Daniel M. Bagner • Stefany Coxe



Ó Springer Science+Business Media New York 2015

Abstract The current study investigated whether and how callous–unemotional traits (CU) moderated the association between specific parenting practices and child conduct problems (CP) with a special consideration of informant patterns. Associations between CU, ODD and CD symptom severity, and the parenting practices of deficient monitoring, positive involvement, and negative/ineffective discipline were examined across parent and teacher reports on 851 elementary-school students. Relative to children with low CU, in children with high CU: (1) positive parenting was associated with lower CD, but increased ODD; (2) negative/ineffective discipline was associated with increased ODD; (3) deficient monitoring was associated with increased CD. Results were not robust across informants. These findings suggest that in the context of CU, the associations between parenting and CP differ based on parenting characteristics, CP dimensions, and informant, and that families may benefit from treatment targeting specific parenting practices based on CP symptom profiles. Keywords Child mental health  Callous–unemotional traits  Conduct problems  Parenting

K. I. Crum (&)  D. A. Waschbusch  D. M. Bagner  S. Coxe Department of Psychology, Florida International University, 11200 SW 8th St, AHC 1, Room 140, Miami, FL 33199, USA e-mail: [email protected] Present Address: D. A. Waschbusch Penn State Hershey Medical Center, Pennsylvania State University, Hershey, PA, USA

Introduction Comprising one of the most common mental health referrals, child conduct problems (CP)—including oppositional defiant disorder (ODD) and/or conduct disorder (CD)— place an enormous burden on affected children and families [1]. As demonstrated by decades of research, parenting and family characteristics are key factors in shaping the development of CP in children [e.g., 2, 3], and thus represent important avenues for prevention and intervention. Harsh and inconsistent discipline, corporal punishment, poor parental monitoring, rejection, coercion, and low parental responsiveness are associated with increased child CP severity [4, 5]. However, research also has demonstrated that children with CP are a heterogeneous rather than homogeneous group [6], and this finding extends to research on the association between parenting and CP. CU Traits May Explain Differential Responses to Parenting Among Children with CP One factor that has proven particularly important with regard to explaining differences in response to parenting among children with CP is callous–unemotional (CU) traits. CU traits refer to features traditionally associated with psychopathy, including an affective style that is characterized by a lack of empathy or caring for others, as well as a lack of ‘‘conscience’’ or guilt following transgressions [7]. Many studies have demonstrated that children with both CP and CU traits (CPCU) differ in important ways from children with CP without CU traits, including on measures of antisocial behavior [8], emotional processing [9–11], and treatment response [12]. Children with CPCU exhibit especially stable and severe antisocial behavior [13, 14], including greater proactive and reactive

123

Child Psychiatry Hum Dev

aggression than children with CP alone [8, 15]. Furthermore, there is a body of evidence showing that CU traits may moderate the association between negative and positive parenting practices, such as harsh discipline and parental warmth, and child CP [16]. However, the literature on parenting, CP, and CU has thus far generated unclear conclusions. In examining the role of parenting for children with CPCU, a handful of studies suggest that the association between parenting and CP differs depending on the level of CU traits. For example, harsh punishment and low parental warmth [17], inconsistent discipline [18], and overall ineffective parenting as measured using a composite score [19, 20] are associated with greater externalizing behavior, CP severity, and aggression—but only in children with low CU levels. Similarly, other research has shown that positive parenting techniques such as reinforcement for appropriate behavior are associated with less severe CD only in children with low CU levels [21], while parental warmth and responsiveness is associated with less severe CP only in children with high CU levels [22–25]. In contrast, other studies suggest that CU traits do not moderate the association between parenting and CP. For example, the associations between CP and corporal punishment [21], monitoring and problem behavior [26], positive behavioral support and problem behavior [27], and harsh discipline and disruptive behavior problems [28] were not moderated by CU traits or CU-related behavior. Given these conflicting findings, it remains unclear whether CU traits moderate the association between parenting practices and child CP. Methodological Issues May Explain Mixed Evidence Regarding CU, CP, and Parenting As discussed in a review on the role of parenting in the antisocial behavior displayed by children with varying CU trait levels [16], these mixed findings may be accounted for, at least in part, by methodological discrepancies. Shared method variance due to reliance on parent report— including parent reports of both parenting practices and child emotional and behavioral characteristics—as well as varied definitions of parenting styles and practices, complicate interpretation of previous findings. Our own examination of the extant literature on this topic revealed at least three important methodological factors that have varied across past studies, and that may partially account for these inconsistent findings. Differences Across Types of Parenting First, research conducted to date either combines different types of negative parenting behavior into a single composite

123

within a sample (e.g., an ineffective parenting composite consisting of poor monitoring, inconsistent discipline, and corporal punishment), or focuses on only a single dimension of negative parenting behavior within a sample (e.g., harsh discipline). Although these studies are informative, they do not provide information about the potentially divergent impact of multiple negative parenting qualities within their sample. There is reason to believe that distinguishing between types of negative parenting is important. Frick et al. [29] noted that specific parenting practices contributed uniquely to youth behavior problems within young, middle, and adolescent age groups, though the interaction between parenting and CU was not fully accounted for in this study. For example, poor monitoring accounted for 20 % of the variance in CD in middle childhood, while corporal punishment accounted for 44 % of this variance. These results suggest the need to investigate the individual contributions of qualitatively different negative parenting practices, and their interactions with CU, rather than combining these practices analytically. Specifically, deficient monitoring is critically important for CP prevention in at-risk children [30], and may be especially significant to the CP of children who also show high levels of CU traits. In particular, deficient monitoring plays a critical role in linking CU traits to serious behavior problems, given its relationship to both indirect violence exposure (e.g., witnessing acts of violence against others) and delinquent peer affiliation—important mediators between CU and delinquency [31–33], as well as wellestablished risk factors for CP [34]. To date, only one study has examined the relation between parental monitoring, CP, and CU, which found that the link between monitoring and adolescent problem behavior remained stable across youth CU trait levels [26]. While existing literature suggests the moderating effect of CU on the parenting-CP association may differ depending on the specific aspect of negative parenting examined, little research has addressed these questions directly. Differences Across Dimensions of Child CP Just as the moderating effects of CU may differ across aspects of parenting, differences may also emerge across aspects of CP. Existing research examining the impact of CU on the association between parenting and CP has typically not accounted for differences between ODD and CD, despite evidence for the importance of differentiating between these forms of psychopathology. Most studies in this area have either collapsed across disorders to form CP or externalizing behavior composites, or have focused solely on one aspect of behavior problems (e.g., ODD or CD). However, many studies have shown that the course

Child Psychiatry Hum Dev

and prognosis of ODD and CD differ [35, 36], and that ODD is more than a stepping stone to CD. Indeed, population-based studies reveal a substantial percentage of children with CD do not meet ODD criteria [37]. Further, ODD symptoms robustly predict later delinquency beyond the influence of co-occurring CD symptoms [38]. Of relevance to the relation between CP dimensions and parenting, ODD symptoms show a transactional relationship with timid parental discipline over time, while CD symptoms predict decreased parental supervision, indicating differential relations between parenting strategies and these aspects of problem behavior [39]. These findings suggest the importance of differentiating between ODD and CD symptoms when investigating parenting and CU. Indeed, Falk and Lee [21] found that positive parenting was unrelated to CD at high CU trait levels, but inversely related to CD at low to moderate CU trait levels. In contrast, CU traits did not moderate the relation between positive parenting and ODD. To date, this is the only study to account for differences between ODD and CD while assessing the impact of CU on associations between parenting and CP.

Differences Across Informants Finally, differences may also exist across informants. Previous research examining whether CU moderates the association between parenting and CP has either relied solely on parent report, or combined parent and teacher report [16]. There are several reasons to examine parent and teacher reports of CP separately when assessing the impact of CU on their association with parenting. On a broad level, differences often exist in both the way children behave across settings and the way informants perceive behavior [40, 41], and these differences are revealed by examining data as a function of informant. Of specific relevance to the current study, it may be the case that the role of CU in moderating the CP-parenting association differs depending on the informant. To date, only two studies have investigated this possibility. As previously described, Falk and Lee [21] found differences as a function of informant, reporting interaction effects between CU and positive parenting for parent-rated, but not teacherrated, CD. In contrast, Oxford et al. [20] reported that CU moderated the relationship between ineffective parenting and teacher-rated—but not parent-rated—externalizing behavior, such that ineffective parenting was positively associated with teacher-rated behavior problems for children with high, but not low, CU trait levels. These studies make it clear that there may be important informant-based differences when examining CP, CU, and parenting, but the fact that their results were inconsistent suggests a need for additional research.

The Role of ADHD in the Relationship Between Parenting, CP, and CU In the present study, one possible confounding factor is attention-deficit/hyperactivity disorder (ADHD). It is crucial to account for ADHD symptoms, as ADHD may play a critical role in understanding CP, CU and parenting. First, ADHD and CP frequently co-occur and do so disproportionately. Specifically, about one-half of children with ADHD have CP, whereas the overwhelming majority of children with CP have ADHD [42, 43]. Second, compelling evidence suggests that the co-morbidity of ADHD and CP results in greater cognitive and social impairment compared to either ADHD or CP alone [44]. Notably, this impairment extends to the parent–child relationship [45], as comorbid CP exacerbates the conflicted parent–child interactions and strained family functioning experienced by parents of children with ADHD [46]. With specific regard to CU, one study thus far accounted for ADHD while examining the relations between CP, CU, and parenting [21]. These authors found that ADHD diagnostic status was consistently and positively associated with caregiver-reported CP, as well as positively correlated with both CU and negative parenting. Clearly, research suggests that ADHD is significantly associated with parenting, CP, and CU, and it is therefore important to account for ADHD when evaluating relationships between these constructs. Current Study Aims and Hypotheses The purpose of this study was to further examine whether CU traits moderate the association between CP in children and the type of parenting they experience. This study built on existing literature by: (a) separately examining distinguishable dimensions of parenting, namely positive parental involvement, negative/ineffective discipline, and deficient monitoring; (b) separately examining distinguishable dimensions of CP, namely ODD and CD symptom severity; and (c) examining these associations separately for different informants (parent and teacher) of CP. To address our third aim, we will use hierarchical linear modeling (HLM) to regress parent-rated predictors (e.g., CU and parenting) on parent-rated and teacher-rated CP outcomes collected as part of a larger school-based study. HLM is an expansion of linear regression that allows researchers to examine the effects of variables at multiple levels. In this study, the two levels of analysis were the child level (level 1) and the classroom or teacher level (level 2), given the nested (children within classrooms) nature of the data. Based on previous literature, we have several hypotheses for each parenting dimension as follows:

123

Child Psychiatry Hum Dev

Positive Parenting In light of Falk and Lee’s [21] study showing that the relation between positive parenting and CP was moderated by CU traits only for parent-rated CD, we hypothesized that: (1a) CU traits will moderate the relation between positive parental involvement and CD, but not ODD, such that the relation between positive parenting and CD will be stronger when CU levels are low; and (1b) There will be a moderating effect of CU on the relation between parenting and CP when CP is reported by the parent, but not when CP is reported by the teacher. Negative/Ineffective Discipline In light of previous research [8, 28] showing that the importance of harsh discipline/corporal punishment is consistent across various facets of child disruptive behavior— including more serious behavior problems, such as aggression and CD, and less serious behavior problems, such as ODD—and across CU levels and informants, we hypothesized that: (2a) The link between negative/ineffective discipline and both ODD and CD severity will remain stable across varying CU levels; and (2b) This effect will be robust across informants.

in age from 5 to 12 years (M = 8.13, SD = 1.93) and included 809 boys (52 %) and 745 girls (48 %). Estimated Hollingshead [48] four-factor socioeconomic status (SES) scores ranged from 14 to 66 (M = 36, SD = 13), which indicates the sample was mostly in the middle class. Ethnic and racial information of the participants was not collected (at the request of the participating school district) but the schools served communities that were over 95 % Caucasian. Of the 1,555 students who participated in the intervention, 704 did not return parent report forms and therefore no information was available on this portion of the sample regarding parent or family characteristics, such as demographic information and parent-rated study variables (e.g., parent ratings of child behavior and parenting practices). At least partial information on parent and family characteristics was available for 851 participants. Procedure

Methods

All procedures used in this study were approved by a university-based Institutional Review Board and by the participating schools and school district. All data reported in this study were collected prior to the start of the intervention. Participating schools were recruited by contacting principals of all elementary schools in the district and giving them information about the intervention project. Principals then met with their staff and subsequently contacted the project coordinator if their school wished to participate. Parent ratings were collected by sending them home with children, along with a cover letter and a return envelope. The cover letter explained the purpose of the project and allowed collection of passive consent. Follow up notes were sent home to parents who did not immediately return ratings, and the same information was conveyed to parents in several public meetings held in each school. Approximately one-half (54.7 % of the sample) of parents returned completed ratings. Teacher ratings were completed approximately 4–6 weeks after the start of the school year, prior to initiating the school-wide intervention. Teachers (n = 66) were given an in-service day if they agreed to complete ratings on all students in their classrooms, and all teachers elected to do so.

Participants

Measures

Participants were students in one of seven elementary schools in a single school district of eastern Canada, all of whom received an intervention program designed to prevent and treat disruptive behavior and related problems using behavior therapy delivered at universal, targeted, and clinical levels [47]. The participating school district included 58 elementary schools which served approximately 13,000 elementary school students. The participants ranged

Assessment of Disruptive Symptoms-DSM-IV Version (ADS-IV)

Deficient Monitoring In light of previous research [26] linking both ODD and CD to parental knowledge and control of youth behavior across CU levels, as well research [31–34] linking monitoring in particular to the CD behavior of children with CU, we hypothesize that: (3a) CU will moderate the relation between deficient monitoring and CD, but not ODD, in that the association between monitoring and CD will be stronger when CU levels are high; and (3b) We refrain from making hypotheses regarding informants, due to the exploratory nature of these analyses.

123

The ADS-IV consists of 29 (for teachers) or 38 (for parents) items designed to measure symptoms of attentiondeficit/hyperactivity disorder (ADHD) and oppositional defiant disorder (ODD) in children [49]. The ADS-IV was completed by parents [n = 857 (ADHD and ODD)] and

Child Psychiatry Hum Dev

teachers [n = 1,550 (ADHD), n = 1,548 (ODD)]. The majority of items on the ADS-IV consist of DSM-IV symptoms of ADHD and ODD rated on Likert scales that range from 0 to 4, with lower ratings indicating the child exhibits the symptom less than other children, and higher ratings indicating the child exhibits the symptom more than other children. Items were averaged to compute an ADHD score (Cronbach’s alpha: Teacher = 0.98; Parent = 0.95) and an ODD score (Cronbach’s alpha: Teacher = 0.98; Parent = 0.94). The psychometric properties of the ADSIV have been supported [49]. Conduct Disorder Rating Scale-DSM-IV Version (CDRSIV) The CDRS-IV consists of 13 (for teachers) or 19 (for parents) items that are designed to measure symptoms of conduct disorder (CD) in children [50]. The CDRS was completed by parents (n = 855) and teachers (n = 1,550). The majority of items on the CDRS-IV consist of DSM-IV symptoms of CD rated on Likert scales that range from 0 to 4, with lower ratings indicating the child has never exhibited the symptom, and higher ratings indicating that the child has frequently exhibited the symptom. Items were averaged to compute a CD score (Cronbach’s alpha: Teacher = 0.67; Parent = 0.74). The psychometric properties of the CDRSIV have been supported [50]. Nova Scotia Modified IOWA Connors (NSIC) The NSIC consists of 35 items designed to measure several aspects of disruptive and aggressive behavior in children [15], including: inattentive–impulsive–overactive behavior, oppositional–defiant behavior, reactive aggression, proactive aggression, relationship aggression, holding a grudge, peer problems, and callous–unemotional traits. Items on the NSIC were drawn from several sources [51–56]. The NSIC was completed by both parents and teachers, but only parent ratings were used in this study (n = 853). The majority of items on the NSIC are rated using Likert scales that range from 0 to 3, with lower scores indicating the child does not at all exhibit the behavior, and higher scores indicating the child exhibits the behavior very much. Of relevance to this study is the callous–unemotional (CU) scale, which consists of the following items: seems to enjoy being mean; is cold and uncaring; lacks remorse for misbehavior. Parent ratings on these items were averaged to compute a CU score (Cronbach’s alpha = 0.76). Although the CU scale from the antisocial process screening device (APSD) is now a widely-used and well-established measure of CU [57], it was not available at the time this data was collected. The NSIC CU scale has been used in previous research on CU traits [58, 59], and the psychometric

properties of this new scale have been examined in other samples. Specifically, the inter-rater (parent–teacher) reliability in a clinical sample of children with CP (n = 148) [60] was significant (r = 0.40) at a value identical to the APSD CU scale parent–teacher correlation [61], and greater than the average parent–teacher correlation for psychopathology ratings [40]. Further, the scale’s validity has been supported by a significant correlation with the APSD CU scale in both the clinic sample described above (r = 0.60) and in a sample of 208 elementary-school children (r = 0.50) [62], as well as a significant positive— yet moderate—correlation with oppositional-defiant behavior (r = 0.55) and peer problems (r = 0.42) in a clinic sample [63]. Alabama Parenting Questionnaire (APQ) The APQ is a 42-item questionnaire designed to measure aspects of parenting shown to be significantly associated with CP in children [64]. The APQ was completed as a self-report measure by parents (n = 856). The APQ consists of items that ask about specific parenting practices rated on Likert scales that range from 1 (‘‘Never’’) to 5 (‘‘Always’’). At the request of the participating school board, three items from the original APQ were dropped, including two items that assessed corporal punishment and one item that assessed parental monitoring and supervision. Following previous research [4, 65], items were summed to compute three scores: parental positive involvement (Cronbach’s alpha = 0.85), negative/ineffective discipline (Cronbach’s alpha = 0.71), and deficient monitoring (Cronbach’s alpha = 0.60). The reliability and validity of the APQ have been established in several studies [29, 66]. Data Analytic Plan Missing Values Analysis As previously described, only a sub-sample of participants completed parent reports. Though teacher reports on study variables (except parenting) were available, only parentrated predictor variables were used to allow comparison of teacher- and parent-rated CP outcomes with other factors held constant. Due to the extensive use of parent-report variables in these models (both predictors and outcomes) and the extensive missingness (nearly 50 %), we were unable to use traditional missing data methods such as multiple imputation to account for missingness on the parent variables. Analysis of only the sub-group who returned ratings raises questions about the generalizability of the data. To better understand this issue, we conducted a missing values analysis [67] to determine if the individuals included in our analyses were different from the

123

Child Psychiatry Hum Dev

individuals who were excluded. This analysis indicated that the children in families with missing parent reports had significantly higher teacher-reported ADHD scores on the Assessment of Disruptive Symptoms (included mean = 1.72, excluded mean = 1.81, t = -2.2, p \ 0.05), but did not differ on demographic information (age, sex) or on other rating scale measures, including measures relevant to this study (ODD, CD, CU). As parents were the sole reporters of parenting practices, it was not possible to evaluate differences in parenting practices between missing and non-missing groups. However, we noted a varied distribution of subscale scores on the parenting measure, which included high, low, and average scores, thereby suggesting that the sample of children whose parents returned ratings was not likely biased against the extreme values on this measure. The small effect size of the difference in ADHD scores (r2 = 0.003) gives us additional confidence in the applicability of our findings across students whose parents did and did not return ratings. Hierarchical Linear Models For each parenting factor—including positive involvement, negative/ineffective discipline, and deficient monitoring— a hierarchical linear model (HLM) [68] was used to examine the relations among CU traits, parenting behavior, CD, and ODD. All variables in this study were collected at the child level, but the clustering of children within classrooms means that children in the same classroom may be more correlated with one another than with two randomly selected children, particularly on teacher-reported variables. HLM allows the partitioning of variance into a portion attributable to differences among children (level 1 variance) and a portion attributable to differences among classrooms or teachers (level 2 variance). All analyses were conducted in Mplus 6.12 [69] using the MLR estimator to provide robust standard error estimates. For the 851 children from 66 classrooms who were included in our analysis, less than 1 % of values were missing for each outcome variable. Full information maximum likelihood estimation was used for all parameter estimates to account for missing values on the outcome variables. As evidence suggests ADHD affects adult caregivers’ perceptions of child behavior [45], ADHD symptom severity was included as a covariate in each model. Additionally, age and gender have been shown to impact the nature and severity of CP [43], and thus were controlled for in each model. Prior to assessing the models of interest, we calculated intraclass correlations (ICCs) for all outcome variables. The ICC represents the proportion of variance in the outcome that is due to differences at level 2. In this study, the ICC is the proportion of variability in the outcomes that is due to differences between classrooms/teachers. For CD,

123

the ICCs were 0.022 for parent report and 0.063 for teacher report; 2.2 % of the variability in parent report of CD and 6.3 % of the variability in teacher report of CD were due to differences between classrooms/teachers. For ODD, the ICCs were 0.010 for parent report and 0.351 for teacher report; 1 % of the variability in parent report of ODD and 35.1 % of the variability in teacher report of ODD were due to differences between classrooms/teachers. These values result in design effects of 1.26, 1.75, 1.12, and 5.17, respectively, indicating that in non-hierarchical models, variances and standard errors would be underestimated as much as 5.17 times. The primary focus of this study was the child level (level 1) rather than the classroom or teacher level (level 2). The observed ICCs and design effects in the sample indicate that there is substantial variability between classrooms. In order to obtain regression coefficients that reflect associations at the child level and that are not biased by mean differences between classrooms, all predictors (except gender) were centered at the classroom level (centering within group) [70].

Results Overview Table 1 shows the means, variances, and correlations for all variables. The upper portion of the table presents variance and zero-order correlation values within each classroom. The lower portion of the table presents mean and standard error values that are aggregated across the 66 classrooms (the within-group centering process produces mean values of 0 for all predictors within each classroom; aggregated means were formed by calculating variable means within each classroom, then computing the weighted average across the 66 classrooms). Interaction terms are omitted from this table for ease of interpretation. For each of the three parenting factors, a single hierarchical structural equation model was conducted in which all outcome variables were predicted by all covariates and predictors (including interaction terms). Parent and teacher reports of ODD and CD were outcomes in all models. Child age, child gender, and child ADHD score were included as covariates on all outcomes in all models. In each model, CU traits, parenting behavior (positive involvement, ineffective discipline, or deficient monitoring), and the interaction of CU and parenting were entered as predictors of outcomes. All predictors were measured at the child level (level 1); there were no predictors measured at the classroom or teacher level (level 2). Measures of model fit indicated good fit at the child level using standard cut-offs of RMSEA \ 0.05, CFI [ 0.95, and

Child Psychiatry Hum Dev Table 1 Within-classroom means, variances, and correlations 1

2

3

1. Parent report of CD

0.05

2. Teacher report of CD

0.340

0.025

3. Parent report of ODD

0.550

0.193

4

5

6

7

8

0.313

0.402

0.246

0.571

5. Age (years)

0.069

0.104

0.027

0.129

-0.105

-0.113

-0.108

-0.165

-0.031

7. ADHD

0.449

0.232

0.690

0.351

0.031

-0.116

0.386

8. CU traits

0.606

0.182

0.450

0.210

0.024

-0.093

0.381

-0.127

-0.039

-0.165

-0.103

-0.011

-0.023

-0.181

9. Positive involvement

10

11

0.652

4. Teacher report of ODD 6. Gender (female = 1)

9

0.19 0.25 0.09 -0.149

42.342

10. Negative/ineffective discipline

0.232

0.043

0.353

0.147

-0.005

0.004

0.250

0.237

-0.212

15.627

11. Deficient monitoring Means

0.299 0.125

0.158 0.052

0.249 1.72

0.189 1.39

0.072 7.95

-0.042 0.48

0.225 1.78

0.240 0.11

-0.181 67.34

0.321 15.67

6.478 9.32

Standard errors

0.05

0.03

0.21

0.15

0.09

0.11

0.17

0.07

1.84

1.13

0.71

Variances are shown on the diagonal; correlations are shown off the diagonal. These values are the same across all 66 clusters (classrooms). Means and standard errors are aggregated and based on the 66 classroom means for each variable

Table 2 Within-group regression coefficients Parenting model Positive involvement

Negative/ineffective discipline

Deficient monitoring

Outcome

Age

Parent CD

0.025

Teacher CD

0.034**

Parent ODD

0.007

Teacher ODD

Gender

ADHD

CU

Parenting

CU X parenting

Level 1 variance

Level 2 variance 0.000

0.001

0.087***

-0.099

-0.002*

0.008

0.026***

-0.020

0.045***

-0.032

0.000

0.001

0.023***

0.001

0.042

0.782***

-0.449

-0.006*

0.016***

0.304***

0.005

0.199**

-0.165**

0.354***

0.165

-0.005

0.001

0.482***

0.325***

Parent CD

0.028

-0.007

0.089***

0.190

0.002

0.010

0.027***

0.001

Teacher CD

0.033*

-0.021

0.045***

0.153*

0.000

-0.005

0.023***

0.001

Parent ODD Teacher ODD

0.018 0.199**

0.016 -0.168**

0.759*** 0.346***

-0.052 0.284

0.029*** 0.012

0.032** -0.005

0.291*** 0.481**

0.008 0.323***

Parent CD

0.021

-0.006

0.083***

0.131

0.020**

0.026***

0.001

0.032*

-0.022*

0.042***

-0.029

0.004

0.007

0.023***

0.001

0.031

0.780***

0.192

0.015

0.032

0.304***

0.011

-0.161**

0.344***

0.434

0.033**

-0.023

0.477***

0.323***

Teacher CD Parent ODD Teacher ODD

-0.001 0.188**

0.007**

CU is callous–unemotional traits. Parent CD is parent report of child CD, Teacher CD is teacher report of child CD, etc * p \ 0.05, ** p \ 0.01, *** p \ 0.001. Age and gender were included as covariates in all models

SRMR \ 0.05 [71]. Fit indices for each model are reported in their respective sections. Slightly higher SRMR (between) values may indicate misfit at the classroom level of the models, though there is relatively little research into acceptable values of level 2 fit indices [72]. Table 2 shows the within-group regression coefficients for the three tested models. Main effects of CU and parenting are described only when the interaction between CU and parenting was not significant. Main effects and interactions are evaluated at alpha = 0.05 using Wald z tests— asymptotic or approximate significance tests provided by Mplus. For significant interactions, effects are described as

simple slopes of the effect of parenting on ODD or CD symptom severity at varying levels of CU traits, using the method of Preacher et al. [73]. Table 2 also shows the residual variance of ODD and CD after taking predictors into account. There was significant level 1 variance in parent and teacher reports of ODD and CD after taking level 1 predictors into account, indicating that significant variability remained to be explained. Likewise, there was significant level 2 variance for teacher report of ODD after taking level 2 predictors into account, indicating that significant variability in teacher reports of ODD remained to be explained by teacher/classroom level factors.

123

Child Psychiatry Hum Dev

Effects of CU and Parenting on CP Positive Involvement Model fit indices suggested a good fit to the data at the child level, RMSEA = 0.013, CFI = 0.999, SRMR (within) = 0.002, SRMR (between) = 0.174. Positive involvement significantly predicted parent report of CD (b = 0.002, p \ 0.05) after controlling for CU and the interaction of CU and positive involvement. In contrast, neither positive involvement nor its interaction with CU predicted teacher report of CD. CU traits moderated the effect of positive involvement on parent report of ODD (b = 0.016, p \ 0.001). As shown in Fig. 1, for children with lower CU levels, there was a negative relation between positive involvement and parent-rated ODD symptom severity. For children with higher CU levels, there was a positive relation between positive involvement and parent-rated ODD symptom severity. In contrast, neither positive involvement nor its interaction with CU predicted teacher report of ODD symptom severity.

Fig. 2 Relation between negative/ineffective discipline and parent report of ODD at different levels of CU traits. Significant positive relationship for all values of CU. For CU = 0, b = 0.029, p \ 0.001. For CU = 1, b = 0.061, p \ 0.001. For CU = 2, b = 0.093, p \ 0.001

report of ODD, but this association was stronger for children with higher levels of CU traits. In contrast, neither negative/ineffective discipline nor its interaction with CU predicted teacher report of ODD.

Negative/Ineffective Discipline

Deficient Monitoring

Model fit indices suggested a good fit to the data at the child level, RMSEA = 0.047, CFI = 0.991, SRMR (within) = 0.005, SRMR (between) = 0.133. Neither negative/ineffective discipline nor its interaction with CU predicted parent report of CD. In contrast, CU traits significantly predicted teacher report of CD (b = 0.153, p \ 0.05); higher CU was associated with higher teacher report of CD. CU traits moderated the effect of negative/ ineffective discipline on parent report of ODD (b = 0.032, p \ 0.01). As shown in Fig. 2, higher negative/ineffective discipline was generally associated with higher parent

Model fit indices suggested a good fit to the data at the child level, RMSEA = 0.054, CFI = 0.988, SRMR (within) = 0.003, SRMR (between) = 0.147. Neither deficient monitoring nor its interaction with CU predicted parent report of ODD. However, deficient monitoring significantly predicted teacher report of ODD (b = 0.033, p \ 0.01); higher deficient monitoring was associated with higher teacher report of ODD. CU traits moderated the effect of deficient monitoring on parent report of CD (b = 0.020, p \ 0.01). As shown in Fig. 3, higher deficient monitoring scores were generally associated with higher parent report of CD, but this association was stronger for children with higher values of CU traits. In contrast, neither deficient monitoring nor its interaction with CU predicted teacher report of CD.

Discussion

Fig. 1 Relation between positive involvement and parent report of ODD at different levels of CU traits. Significant negative relationship for CU \ 0.1174 (0.0074 points below the aggregated mean of 0.11); significant positive relationship for CU [ 1.0896 (0.9796 points above the aggregated mean of 0.11). For CU = 0, b = -0.006, p \ 0.05. For CU = 1, b = 0.010, p \ 0.05. For CU = 2, b = 0.026, p \ 0.01

123

The present study was designed to assess the influence of CU traits on the association between parenting characteristics and child CP severity in a community sample, while addressing several methodological factors that have accounted for discrepancies in previous literature. As discussed below, results showed that the associations between CP, CU, and parenting were complex and differed as a function of all three factors— type of parenting, type of CP, and informant. The results of this study are broadly consistent with many other studies in demonstrating a significant association between parenting and CP in children [74]. That is,

Child Psychiatry Hum Dev

Fig. 3 Relation between deficient monitoring and parent report of CD at different levels of CU traits. Significant positive relationship for CU [ 0.0563 (0.0537 points below the aggregated mean of 0.11). For CU = 0, b = 0.007, p \ 0.05. For CU = 1, b = 0.027, p \ 0.001. For CU = 2, b = 0.047, p \ 0.001

negative parenting practices (i.e., low positive parental involvement, negative/ineffective discipline, and deficient monitoring) were generally associated with higher CP symptom severity. These findings mirror studies conducted since at least the 1960s in demonstrating that the type of parenting children experience has a significant impact on their level of antisocial behavior in well-specified ways [75]. Results are also broadly consistent with more recent research that suggests the association between parenting and children’s antisocial behavior may be moderated by CU traits [16], but that methodological disparities across studies play a role in the mixed results observed thus far in the literature. In particular, the results of this study suggest that although parenting is generally associated with children’s CP, the nature and presence of this association is influenced by the level of CU traits in the child. However, in agreement with Waller et al. [16], this conclusion (a) does not imply that CP in children showing significant levels of CU traits is immune to the influence of parenting, and (b) must be tempered to account for additional factors that influence these relationships. Importantly, our findings add potential caveats to this broad conclusion. First, teacher report of CP was generally not associated with parenting. Only one of six associations between parenting and teacher-reported CP was significant, and there were no significant CU by parenting interactions when examining teacher-reported CP as an outcome. In contrast, four of the six possible associations between parenting and parent-reported CP were significant, and three of the six CU by parenting interactions were significant. Consistent with our results, Falk and Lee [21] reported that neither main effects nor interactions of parenting (positive or negative) and CU predicted teacher report of CP (either ODD or CD), even though the authors found associations with parent report of CP. In contrast, Oxford et al. [20] found that teacher report of CP, defined using a broad-band

measure of externalizing behavior problems, was associated with the interaction between CU and ineffective parenting. Interestingly, the measure of ineffective parenting used by Oxford et al. [20] included the same deficient monitoring items (along with others) used in our study, and an interaction was found when measuring CU on a continuous, rather than dichotomous, scale. Thus, results of this study are consistent with the Falk and Lee [21] study in suggesting that the association between parenting, CU, and their interaction is more robust with regard to parent- than teacherreported CP, and our findings are consistent with Oxford et al. [20] in suggesting that deficient monitoring may play a role in understanding teacher report of CP. As with any predictor/outcome data obtained from a single informant, it is possible that robust results obtained from analyses using both parent-rated predictors and outcomes may be attributed, at least in part, to shared method variance. However, literature on informant perspectives suggests the potential for these discrepancies to reflect informative, important differences, rather than measurement error [76]. In the current study, this may apply to differences in behavioral contingencies between settings, as well as differential effects of parenting on behavior exhibited in the home versus school setting. Further investigation is required to assess whether these discrepancies are related to differential long-term outcomes, as well as agreement between reported and observed behavioral differences across informant [76]. Second, results of our study suggested the association between CU and parenting may differ depending on the specific type of CP examined. Judging by our findings, it appears that some parenting factors may influence ODD symptoms more than CD symptoms in children with CU traits, while others may influence CD more than ODD. For example, while the relation between negative/ineffective discipline and CP was stronger for ODD among children with high levels of CU, the relation between this parenting factor and CD remained the same across CU levels. Further, past research has demonstrated that (1) the association between poor parental monitoring and antisocial behavior is stronger for adolescents than for younger children [29, 30], and (2) the severity of antisocial behavior tends to increase with age [30]. Taken together, these findings suggest that deficient monitoring is more likely to be associated with severe forms of antisocial behavior, as reported in the current study. More striking in terms of the contrast between ODD and CD were findings concerning the role of positive parental involvement. Observed associations between CD and positive parental involvement—regardless of CU trait level—are consistent with past research and theory [77] in showing that positive parenting practices are generally associated with less severe antisocial behavior in children.

123

Child Psychiatry Hum Dev

In contrast, CU traits moderated the association between positive parental involvement and parent report of ODD. This finding is inconsistent with results of a previous study [21] which reported that CU traits moderated the association between positive parenting and parent-reported CD, but not parent-reported ODD. Although the pattern of results differed, with one study reporting moderating effects for ODD and another reporting moderating effects for CD, they are consistent in demonstrating that it is important to distinguish between ODD and CD when examining associations between parenting and CU traits. Interestingly, CU moderated the relation between positive parental involvement and parent-reported ODD in the opposite direction than we predicted (see Fig. 1). These results are consistent with Falk and Lee [21], who also reported that more positive parenting was significantly associated with lower parent-reported CP only for children with low CU traits, whereas positive parenting was associated with higher parent-reported CP for children with high CU traits—although the effect was of trending significance in Falk and Lee’s study. Our findings should also be considered in light of research that shows high CU traits are associated with more insecure attachment to parents early in life [78] and decreased parental involvement and parental monitoring [26, 79] over the course of development. Taken together, these outcomes seem to paint a picture in which parents feel less attached to offspring with high CU traits, feel that their attempts to become positively involved with their children are rebuffed or met with CP, and gradually become more withdrawn from them. It is important to note, however, that not all research is consistent with this picture. For example, some research has reported that positive parenting factors (e.g., parental warmth) are associated with decreases in children’s antisocial behavior among children with high CU traits [22]. Better understanding of these seemingly inconsistent findings through additional research would be highly beneficial. Finally, it is worth noting that the most robust association that emerged in this research was the association between ADHD and CP. Whether CP was defined using parent or teacher as informant, or using ODD or CD symptom severity, it was always the case that higher ADHD was associated with more severe CP even after controlling for parenting, CU, and their interaction. These results are consistent with past studies demonstrating the importance of accounting for ADHD when examining CP [44], as well as examining CU [61, 80], and show that research neglecting to account for ADHD when examining these constructs is at risk of testing a mis-specified model. Because these data were collected from a sample that largely consisted of typically developing children recruited from elementary school settings, these results are also

123

consistent with studies suggesting that ADHD accounts for a large share of disruptiveness in classroom settings—even more so than ODD and CD [57, 81]—and that treatment targeting ADHD is likely to be an important component of any effective treatment for CP [82]. It is important to acknowledge several limitations to the current study. First, sample data were obtained using notes sent home to parents of children in elementary schools, and parents that responded may be different than those that did not respond, thereby reducing generalizability of the findings. Indeed, missing variable analyses found that children in families with missing parent reports had significantly higher teacher-reported ADHD scores. Though ADHD severity was included as a covariate, ADHD may be related to perceptions of CP [45, 83]. However, the difference between children whose parents did and did not return ratings was very small in magnitude (explaining substantially less than one percent of variance), suggesting it may not be meaningful. Further, as parents were the sole reporters of parenting practices, it was not possible to evaluate differences in parenting practices between missing and non-missing groups. However, we noted a varied distribution of subscale scores on the APQ, suggesting that the sample of children whose parents returned ratings was not likely biased against the extreme values on this measure. Second, due to the community nature of the sample, variability in CP and CU was limited, with mostly low levels observed. Given that the vast majority of studies on parenting, CP, and CU have used samples of clinically-referred (3), aggressive (3), or at-risk (4) children, it is somewhat difficult to directly compare our results to previous studies, although we believe this study provides valuable complementary information to those studies. Third, though the brief measure used to assess CU traits was appropriate for this study, in which parents and teachers rated a large number of children on several variables, and has demonstrated reliability and validity, replication using a widespread measure of CU would be advantageous. However, more commonly used measures of CU were not available when data were collected for the present study. Additionally, as noted in the measures section, information on corporal punishment, as well as several deficient monitoring items, was not collected at the request of the participating school board. Further, the threefactor APQ used in the current study does not encompass all potentially relevant facets of parenting, such as parental warmth. Future studies would do well to supplement parent and teacher report with observational data, as well as administer the full versions of established measures of CU traits and parenting. Fourth, although racial and ethnic information was not collected at the request of the participating school district, the schools involved in the current study served a primarily Caucasian community

Child Psychiatry Hum Dev

[84]. The homogeneity of this sample may limit generalization of findings, creating a need for future studies to investigate associations between CP, CU, and parenting in a more ethnically diverse sample. Finally, causality and directionality of the associations observed in this study cannot be inferred due to the cross-sectional nature of the data. Assessment of developmental trajectories and transactional relationships between these constructs must be conducted in future research, as there is evidence for the possibility of child effects on parenting over time [26, 79, 85]. Despite these limitations, the current study had several important strengths. We addressed methodological limitations of research on the effects of parenting in children with CPCU by accounting for multiple parenting facets, CP dimensions, and informants within the study sample. While most of the cross-sectional studies in this area tended to have small sample sizes, our sample included over eight hundred children. Though the community-based nature of this sample was noted as a limitation above, it also allows generalization of results beyond information gathered from a clinic or institution. The use of sophisticated SEM techniques allowed informal cross-informant comparison, which prevented the loss of scale- and item-level information that accompanies combining across parent and teacher ratings of child behavior. Collectively, our findings suggest the importance of parenting practices to CP in a population of children whose behavior problems’ sensitivity to contextual influences has previously been a point of contention. Given that specific CP symptom profiles of children with CU are not affected by all parenting practices in the same manner, tailoring interventions to specific behavioral difficulties—as well as specific deficits in the parent–child relationship—may lead to more effective intervention for CP. Indeed, children showing high levels of CU traits may benefit from individualized intervention [86]. Our findings help clarify which pieces of interventions must be tailored, and according to which of the child and family’s needs, harnessing valuable clinical information. For example, parenttraining programs specifically targeted towards reinforcing the use of neutral, consistent discipline may be chosen for children with CU and a strong ODD symptom profile. In contrast, for children with CU and a strong CD symptom profile, parent-training programs targeting parental supervision of child behavior may be most effective. Future research should focus on identifying the longterm effects of parenting on ODD and CD symptoms in children with CU traits, with the goals of improving treatment development and selection, as well as informing developmental theory on parent–child relationships in this population. In this way, subsequent research could inform prevention and early intervention efforts based on the

nature and setting of behavioral impairment. Intervention designs informed by these basic-science studies, and based on specific CP-symptom and CU profiles, should be tested to assess their ability to improve child and family outcomes. A final recommendation for follow-up research concerns variability in the child and classroom levels of our analyses. Results suggest that the prediction of CP outcomes in our models would be improved by accounting for other factors in addition to examined aspects of parenting, CU, and their interaction. Emotional processing abnormalities, for example, constitute within-child factors that, in theory, increase risk for antisocial behavior and may explain additional variance in models of CP development [9–11]. Further research that examines these factors would improve models of CP development. Overall, the current study answers several important questions in the literature on the impact of parenting in children with varying levels of CU traits, and highlights the contribution of various parenting practices to specific aspects of CP behavior in the context of CU traits. Our results suggest that consideration of clinical profiles, including problem behavior type, CU traits, and behavior differences across settings, is important to case conceptualization and intervention development. Future research in this area may provide families with interventions tailored to their specific needs.

Summary The current study investigated whether and how CU traits moderated the relation between specific parenting practices and child ODD and CD as a function of informant in a community sample. Results showed that type of parenting, type of behavior problem, and the setting in which behavior occurs are important factors to consider when examining the complex relations between parenting, CP, and CU. No significant CU by parenting practice interactions were found when examining teacher report of CP as an outcome. In contrast, three of six possible interactions were significant when examining parent report of CP as an outcome. Specifically, negative/ineffective discipline and deficient monitoring were especially influential to ODD and CD severity (respectively) in children with high CU trait levels. Additionally, CU traits moderated the relation between positive parental involvement and ODD in an intriguing and unexpected way, such that greater positive involvement was associated with greater ODD symptom severity in children with high CU trait levels. Together, these findings suggest that while the nature of the relationship between parenting and CP is indeed influenced by CU traits, children with high levels of these traits are not immune to the influence of parenting. In fact, some

123

Child Psychiatry Hum Dev

parenting practices may be more influential to the antisocial behavior of children with significant CU traits than others, especially when considering specific symptom presentations. Results have clinical implications for developing individualized parent-directed interventions for children with CP and high levels of CU, such as programs designed to enhance parental supervision for children with high CU trait levels and a strong CD symptom profile, or to emphasize the use of effective discipline strategies for children with high CU levels and a strong ODD symptom profile. Future studies are charged with investigating the longitudinal effects of specific parenting practices on different aspects of child CP across CU trait levels, as well as the potential for improved treatment development and selection based on these clinical features. Acknowledgments This research is based on data collected by Daniel A. Waschbusch through the Behavior Education Support and Treatment (BEST) program. Funding for the BEST project was partially provided by grants to Daniel Waschbusch from the Nova Scotia Health Research Foundation (No. 304E) and the Social Science and Humanities Research Council of Canada (839-2000-1061).

12.

13.

14.

15.

16.

17.

18.

References 19. 1. Costello EJ, Mustillo S, Erkanli A, Keeler G, Angold A (2003) Prevalence and development of psychiatric disorders in childhood and adolesence. Arch Gen Psychiatry 60(9):837–844 2. Patterson GR, DeBaryshe BD, Ramsey E (1989) A developmental perspective on antisocial behavior. Am Psychol 44(2):329 3. Burke JD, Loeber R, Birmaher B (2002) Oppositional defiant disorder and conduct disorder: a review of the past 10 years, part II. J Am Acad Child Adolesc Psychiatry 41(11):1275–1293 4. Hinshaw SP, Owens EB, Wells KC, Kraemer HC, Abikoff HB, Arnold LE et al (2000) Family process and treatment outcome in the MTA: negative/ineffective parenting practices in relation to multimodal treatment. J Abnorm Child Psychol 28(6):555–568 5. Loeber R, Stouthamer-Loeber M (1986) Family factors as correlates and predictors of juvenile conduct problems and delinquency. In: Tonry M, Morris N (eds) Crime and justice. University of Chicago Press, Chicago, IL, pp 29–149 6. Loeber R, Burke JD, Pardini DA (2009) Development and etiology of disruptive and delinquent behavior. Annu Rev Clin Psychol 5:291–310 7. Frick PJ, Nigg JT (2012) Current issues in the diagnosis of attention deficit hyperactivity disorder, oppositional defiant disorder, and conduct disorder. Annu Rev Clin Psychol 8:77–107 8. Frick PJ, Ray JV, Thornton LC, Kahn RE (2014) Can callous– unemotional traits enhance the understanding, diagnosis, and treatment of serious conduct problems in children and adolescents? A comprehensive review. Psychol Bull 140(1):1–57 9. Blair RJR, Peschardt KS, Budhani S, Mitchell DGV, Pine DS (2006) The development of psychopathy. J Child Psychol Psychiatry 47(3–4):262–275 10. Anastassiou-Hadjicharalambous X, Warden D (2008) Physiologically-indexed and self-perceived affective empathy in conduct-disordered children high and low on callous–unemotional traits. Child Psychiatry Hum Dev 39(4):503–517 11. Marsh AA, Finger EC, Mitchell DGV, Reid ME, Sims C, Kosson DS et al (2008) Reduced amygdala response to fearful expression

123

20.

21.

22.

23.

24.

25.

26.

27.

in children and adolescents with callous–unemotional traits and disruptive behavior disorders. Am J Psychiatry 165(6):712–720 Hawes DJ, Price MJ, Dadds MR (2014) Callous–unemotional traits and the treatment of conduct problems in childhood and adolescence: a comprehensive review. Clin Child Fam Psychol Rev 17(3):248–267 Frick PJ, Cornell AH, Barry CT, Bodin SD, Dane HE (2003) Callous–unemotional traits and conduct problems in the prediction of conduct problem severity, aggression, and self-report of delinquency. J Abnorm Child Psychol 31(4):457–470 Frick PJ, Stickle TR, Dandreaux DM, Farrell JM, Kimonis ER (2005) Callous–unemotional traits in predicting the severity and stability of conduct problems and delinquency. J Abnorm Child Psychol 33(4):471–487 Waschbusch DA, Porter S, Carrey N, Kazmi SO, Roach KA, D’Amico DA (2004) Investigation of the heterogeneity of disruptive behaviour in elementary-age children. Can J Behav Sci 36(2):97–112 Waller R, Gardner F, Hyde LW (2013) What are the associations between parenting, callous–unemotional traits, and antisocial behavior in youth? A systematic review of evidence. Clin Psychol Rev 33(4):593–608 Hipwell AE, Pardini DA, Loeber R, Sembower M, Keenan K, Stouthamer-Loeber M (2007) Callous–unemotional behaviors in young girls: shared and unique effects relative to conduct problems. J Clin Child Adolesc Psychol 36(3):293–304 Edens JF, Skopp NA, Cahill MA (2008) Psychopathic features moderate the relationship between harsh and inconsistent parental discipline and adolescent antisocial behavior. J Clin Child Adolesc Psychol 37(2):472–476 Wootton JM, Frick PJ, Shelton KK, Silverthorn P (1997) Ineffective parenting and childhood conduct problems: the moderating role of callous–unemotional traits. J Consult Clin Psychol 65(2):301–308 Oxford M, Cavell TA, Hughes JN (2003) Callous/unemotional traits moderate the relation between ineffective parenting and child externalizing problems: a partial replication and extension. J Clin Child Adolesc Psychol 32(4):577–585 Falk AE, Lee SS (2012) Parenting behavior and conduct problems in children with and without attention-deficit/hyperactity disorder: moderation by callous–unemotional traits. J Psychopathol Behav Assess 34(2):172–181 Pasalich DS, Dadds MR, Hawes DJ, Brennan J (2011) Do callous– unemotional traits moderate the relative importance of parental coercion versus warmth in child conduct problems? An observational study. J Child Psychol Psychiatry 52(12):1308–1315 Kroneman LM, Hipwell AE, Loeber R, Koot HM, Pardini DA (2011) Contextual risk factors as predictors of disruptive behavior disorder trajectories in girls: the moderating effect of callous– unemotional features. J Child Psychol Psychiatry 52(2):167–175 Kochanska G, Kim S, Boldt LJ, Yoon JE (2013) Children’s callous–unemotional traits moderate links between their positive relationships with parents at preschool age and externalizing behavior problems at early school age. J Child Psychol Psychiatry 54(11):1251–1260 Waller R, Gardner F, Shaw DS, Dishion TJ, Wilson MN, Hyde LW (2014) Callous–unemotional behavior and early-childhood onset of behavior problems: the role of parental harshness and warmth. J Clin Child Adolesc Psychol. http://www.ncbi.nlm.nih. gov/pubmed/24661288 Mun˜oz LC, Pakalniskiene V, Frick PJ (2011) Parental monitoring and youth behavior problems: moderation by callous–unemotional traits over time. Eur Child Adolesc Psychiatry 20(5): 261–269 Hyde LW, Shaw DS, Gardner F, Cheong J, Dishion TJ, Wilson M (2013) Dimensions of callousness in early childhood: links to

Child Psychiatry Hum Dev

28.

29.

30.

31.

32.

33.

34. 35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

problem behavior and family intervention effectiveness. Dev Psychopathol 25(02):347–363 Kimonis ER, Frick PJ, Boris NW, Smyke AT, Cornell AH, Farrell JM et al (2006) Callous–unemotional features, behavioral inhibition, and parenting: independent predictors of aggression in a high-risk preschool sample. J Child Fam Stud 15(6):745–756 Frick PJ, Christian RE, Wooton JM (1999) Age trends in the association between parenting practices and conduct problems. Behav Modif 23(1):106–128 Dishion TJ, McMahon RJ (1998) Parental monitoring and the prevention of child and adolescent problem behavior: a conceptual and empirical formulation. Clin Child Fam Psychol Rev 1(1):61–75 Kimonis ER, Frick PJ, Barry CT (2004) Callous–unemotional traits and delinquent peer affiliation. J Consult Clin Psychol 72(6):956–966 Simons RL, Lin KH, Gordon LC, Brody GH, Murry V, Conger RD (2002) Community differences in the association between parenting practices and child conduct problems. J Marriage Fam 64(2):331–345 Howard AL, Kimonis ER, Mun˜oz LC, Frick PJ (2012) Violence exposure mediates the relation between callous–unemotional traits and offending patterns in adolescents. J Abnorm Child Psychol 40(8):1237–1247 Margolin G, Gordis EB (2000) The effects of family and community violence on children. Annu Rev Psychol 51:445–479 Loeber R, Keenan K, Lahey BB, Green SM, Thomas C (1993) Evidence for developmentally based diagnoses of oppositional defiant disorder and conduct disorder. J Abnorm Child Psychol 21(4):377–410 Rowe R, Costello EJ, Angold A, Copeland WE, Maughan B (2010) Developmental pathways in oppositional defiant disorder and conduct disorder. J Abnorm Psychol 119(4):726 Rowe R, Maughan B, Pickles A, Costello EJ, Angold A (2002) The relationship between DSM-IV oppositional defiant disorder and conduct disorder: findings from the Great Smoky Mountains Study. J Child Psychol Psychiatry 43(3):365–373 Pardini DA, Fite PJ (2010) Symptoms of conduct disorder, oppositional defiant disorder, attention-deficit/hyperactivity disorder, and callous–unemotional traits as unique predictors of psychosocial maladjustment in boys: advancing an evidence base for DSMV. J Am Acad Child Adolesc Psychiatry 49(11):1134–1144 Burke JD, Pardini DA, Loeber R (2008) Reciprocal relationships between parenting behavior and disruptive psychopathology from childhood through adolescence. J Abnorm Child Psychol 36(5):679–692 Achenbach TM, McConaughy SH, Howell CT (1987) Child/ adolescent behavioral and emotional problems: implications of cross-informant correlations for situational specificity. Psychol Bull 101(2):213–232 De Los Reyes A, Kazdin AE (2005) Informant discrepancies in the assesment of childhood psychopathology: a critical review, theoretical framework, and recommendations for future study. Psychol Bull 131(4):483–509 Hinshaw SP (1987) On the distinction between attentional deficits/hyperactivity and conduct problems/aggression in child psychopathology. Psychol Rev 101(3):443–463 Loeber R, Keenan K (1994) Interaction between conduct disorder and its comorbid conditions: effects of age and gender. Clin Psychol Rev 14(6):497–523 Waschbusch DA (2002) A meta-analytic examination of comorbid hyperactive–impulsive-attention problems and conduct problems. Psychol Bull 128(1):118–150 Podolski CL, Nigg JT (2001) Parent stress and coping in relation to child ADHD severity and associated child disruptive behavior problems. J Clin Child Psychol 30(4):503–513

46. Deault LC (2010) A systematic review of parenting in relation to the development of comorbidities and functional impairments in children with attention-deficit/hyperactivity disorder (ADHD). Child Psychiatry Hum Dev 41(2):168–192 47. Waschbusch DA, Pelham WE, Massetti GM, Northern Partners In Action for Children and Youth (2005) The Behavior Education Support and Treatment (BEST) school intervention program: pilot project data examining school-wide, targetedschool, and targeted-home approaches. J Atten Disord 9(1): 313–322 48. Hollingshead AB (1975) Four factor index of social status. Yale University, New Haven, CT 49. Waschbusch DA, Sparkes SJ, Northern Region Partners in Action for Children and Youth (2003) Rating scale assessment of attention-deficit/hyperactivity disorder (ADHD) and oppositional defiant disorder (ODD) symptoms: is there a normal distribution and does it matter? J Psychoeduc Assess 21(3):261–281 50. Waschbusch DA, Elgar FJ (2007) Development and validation of the conduct disorder rating scale. Assessment 14(1):65–74 51. Milich R, Loney J, Landau S (1982) Independent dimensions of hyperactivity and aggression: a validation with playroom observation data. J Abnorm Psychol 91(3):183–198 52. Pelham WE, Milich R, Murphy DA, Murphy HA (1989) Normative data on the IOWA Conners teacher rating scale. J Clin Child Psychol 18(3):259–262 53. Waschbusch DA, Willoughby MT (2008) Parent and teacher ratings on the IOWA Conners Rating Scale. J Psychopathol Behav Assess 30(3):180–192 54. Dodge KA, Coie JD (1987) Social information processing factors in reactive and proactive aggression in children’s peer groups. J Pers Soc Psychol 53(6):1146–1158 55. Crick NR, Grotpeter JK (1995) Relational aggression, gender, and social-psychological adjustment. Child Dev 66(3):710–722 56. Waschbusch DA, Pelham WE, Jennings JR, Greiner AR, Tarter RE, Moss HB (2002) Reactive aggression in boys with disruptive behavior disorders: behavior, physiology, and affect. J Abnorm Child Psychol 30(6):641–656 57. Frick PJ, Hare RD (2001) Antisocial processes screening device: technical manual. Multi-Health Systems, Toronto 58. Waschbusch DA, Graziano PA, Willoughby MT, Pelham WE (2014) Classroom rule violations in elementary school students with callous–unemotional traits. J Emot Behav Disord. doi:10. 1177/1063426614552903 59. Crum KI, Waschbusch DA, Willoughby MT (in press) Callous– unemotional traits, behavior disorders, and the student–teacher relationship in elementary school students. J Emot Behav Disord. doi:10.1177/1063426615569533 60. Waschbusch DA, Carrey NJ, Willoughby MT, King S, Andrade BF (2007) Effects of methylphenidate and behavior modification on the social and academic behavior of children with disruptive behavior disorders: the moderating role of callous/unemotional traits. J Clin Child Adolesc Psychol 36(4):629–644 61. Loney BR, Frick PJ, Ellis M, McCoy MG (1998) Intelligence, callous–unemotional traits, and antisocial behavior. J Psychopathol Behav Assess 20(3):231–247 62. Waschbusch DA, Willoughby MT (2008) Attention-deficit/hyperactivity disorder and callous–unemotional traits as moderators of conduct problems when examining impairment and aggression in elementary school children. Aggress Behav 34(2):139–153 63. Andrade BF, Sorge GB, Djordjevic D, Naber AR (2014) Callous– unemotional traits influence the severity of peer problems in children with impulsive/overactive and oppositional/defiant behaviors. J Child Fam Stud. doi:10.1007/s10826-014-0021-6 64. Frick PJ (1991) Alabama parenting questionnaire. University of Alabama, Tuscaloosa, AL

123

Child Psychiatry Hum Dev 65. Hawes DJ, Dadds MR (2006) Assessing parenting practices through parent-report and direct observation during parenttraining. J Child Fam Stud 15(5):555–568 66. Essau CA, Sasagawa S, Frick PJ (2006) Callous–unemotional traits in a community sample of adolescents. Assessment 13(4):454–469 67. Little RJA (1988) A test of missing completely at random for multivariate data with missing values. J Am Stat Assoc 83(404): 1198–1202 68. Raudenbush SW, Bryk AS (2002) Hierarchical linear models: applications and data analysis methods. Sage, Thousand Oaks, CA 69. Muthe´n LK, Muthe´n BO (1998–2012) Mplus user’s guide, 7th edn. Muthe´n & Muthe´n, Los Angeles, CA 70. Kreft IGG, De Leeuw J, Aiken LS (1995) The effect of different forms of centering in hierarchical linear models. Multivar Behav Res 30(1):1–21 71. Hu LT, Bentler PM (1999) Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct Equ Modeling 6(1):1–55 72. Ryu E, West SG (2009) Level-specific evaluation of model fit in multilevel structural equation modeling. Struct Equ Model 16(4):583–601 73. Preacher KJ, Curran PJ, Bauer DJ (2006) Computational tools for probing interactions in multiple linear regression, multilevel modeling, and latent curve analysis. J Educ Behav Stat 31(4):437–448 74. Patterson GR, Reid JB, Dishion TJ (1992) A social learning approach 4: antisocial boys. Castalia, Eugene, OR 75. Patterson GR, Reid JB, Eddy JM (2002) A brief history of the Oregon model. In: Reid JB, Patterson GR, Snyder J (eds) Antisocial behavior in children and adolescents: a developmental analysis and model for intervention. American Psychological Association, Washington, DC 76. De Los Reyes A, Thomas SA, Goodman KL, Kundey SMA (2013) Principles underlying the use of multiple informants’ reports. Annu Rev Clin Psychol 9:123–149

123

77. Pfiffner LJ, McBurnett K, Rathouz PJ, Judice S (2005) Family correlates of oppositional and conduct disorders in children with attention deficit/hyperactivity disorder. J Abnorm Child Psychol 33(5):551–563 78. Pasalich DS, Dadds MR, Hawes DJ, Brennan J (2012) Attachment and callous–unemotional traits in children with early-onset conduct problems. J Child Psychol Psychiatry 53(8):838–845 79. Hawes DJ, Dadds MR, Frost ADJ, Hasking PA (2011) Do childhood callous–unemotional traits drive change in parenting practices? J Clin Child Adolesc Psychol 40(4):507–518 80. Haas SM, Waschbusch DA (2012) Callous–unemotional traits and their relevance to ADHD. ADHD Rep 20(3):5–9 81. Pelham WE, Waschbusch DA (2004) Assessment and treatment of attention-deficit/hyperactivity disorder (ADHD) in schools. In: Brown RT (ed) Handbook of pediatric psychology in schools. Lawrence Erlbaum Associates, Manweh, NJ, pp 405–430 82. Webster-Stratton CH, Reid JM, Beauchaine TP (2011) Combining parent and child training for young children with ADHD. J Clin Child Adolesc Psychol 40(2):191–203 83. Greene RW, Beszterczey SK, Katzenstein T, Park K, Goring J (2002) Are students with ADHD more stressful to teach? Patterns of teacher stress in an elementary school sample. J Emot Behav Disord 10(2):79–89 84. Nova Scotia Department of Finance (2003) 2001 Census of Canada: Nova Scotia perspective. Government of Canada. http:// www.gov.ns.ca/finance/publish/CENSUS/Census%201.pdf 85. Salihovic K, Kerr M, Ozdemir M, Pakalniskiene V (2012) Direction of effects between adolescent psychopathic traits and parental behaviors. J Abnorm Child Psychol 40(6):957–969 86. Dadds MR, Cauchi AJ, Wimalaweera S, Hawes DJ, Brennan J (2012) Outcomes, moderators, and mediators of empathic-emotion recognition training for complex conduct problems in childhood. Psychiatry Res 199(3):201–207

Effects of Callous-Unemotional Traits on the Association Between Parenting and Child Conduct Problems.

The current study investigated whether and how callous-unemotional traits (CU) moderated the association between specific parenting practices and chil...
287KB Sizes 1 Downloads 6 Views