J Abnorm Child Psychol (2011) 39:1013–1023 DOI 10.1007/s10802-011-9512-8
Effects of a Parenting Intervention on Features of Psychopathy in Children Renee McDonald & Mary Catherine Dodson & David Rosenfield & Ernest N. Jouriles
Published online: 10 May 2011 # Springer Science+Business Media, LLC 2011
Abstract This study examined whether Project Support, a parenting intervention shown to reduce child conduct problems, also exerts positive effects on features of psychopathy in children. Participants were 66 families (mothers and children) recruited from domestic violence shelters who participated in a randomized controlled trial evaluating Project Support. Each family included at least one child between the ages of 4 and 9 who was exhibiting clinical levels of conduct problems. Families were randomly assigned to the Project Support intervention condition or to an existing services comparison condition, and they were assessed on 6 occasions over 20 months, following their departure from the shelter. Children in families in the Project Support condition, compared with those in the comparison condition, exhibited greater reductions in features of psychopathy. Moreover, the changes in features of psychopathy remained after accounting for changes in conduct problems. Project Support’s effects on features of psychopathy were mediated by improvements in mothers’ harsh and inconsistent parenting. These findings on the effects of an intervention on features of psychopathy are the first from a randomized controlled trial. They inform the debate about whether features of psychopathy in children
This research was supported by grant R01-MH-53380, awarded by the National Institute of Mental Health, and 2005-JW-BX-K017, by the Office of Juvenile Justice and Delinquency Prevention, Office of Justice Programs, U.S. Department of Justice. Points of view or opinions in this document are those of the authors and do not represent the official position of the U.S. Department of Justice or other federal agencies. R. McDonald (*) : M. C. Dodson : D. Rosenfield : E. N. Jouriles Southern Methodist University, University Park, TX, USA e-mail: [email protected]
are responsive to intervention, and hold important implications for clinical practice. Keywords Child psychopathy . Intervention . Parenting . Conduct problems . Externalizing problems intimate partner violence
Psychopathy consists of a constellation of affective and interpersonal deficits coupled with behavioral impulsivity and aggressive and antisocial behavior. The affective and interpersonal deficits include lack of remorse for wrongdoing, diminished ability to empathize with others, poverty or shallowness of emotion, and callous treatment of others, especially for personal gain (Frick 2009; Frick et al. 2000). In the research on psychopathy, these key affective and interpersonal features are commonly referred to as callous/unemotional and narcissistic personality traits, and they are thought to lie at the heart of psychopathy. The phenomenon of psychopathy is increasingly conceptualized as a developmental phenomenon (e.g., Salekin and Frick 2005), with its affective and interpersonal features identified reliably in children as young as four years old (Dadds et al. 2005). There is understandable concern, and some controversy, over the use of the term “psychopathy” in reference to children (Edens et al. 2001; Frick 2002, 2009; Kotler and McMahon 2005). Nonetheless, recognizing the distinctiveness and importance of the features theorized to underlie the construct of psychopathy—particularly in the context of clinical levels of child conduct problems—may be critical for advancing interventions in this area (Salekin & Frick). By definition, psychopathy includes antisocial behaviors. Thus, as would be expected, such measures of psychopathy correlate positively with measures of child externalizing problems (see Burns 2000; Dadds et al. 2005,
for discussion).1 However, psychopathy and conduct problems are not isomorphic. The affective and interpersonal features that characterize psychopathy—some of which may overlap with but are not in themselves definitive of externalizing behaviors or disruptive behavior diagnoses— have been found to be important predictors of later antisocial and aggressive behavior. Indeed, these key features of psychopathy among children with conduct problems are associated with a particularly insidious and intractable course of antisocial behavior in adulthood, and elevations on the affective and interpersonal features are associated with increased likelihood of meeting diagnostic criteria for conduct disorder or oppositional defiant disorder (Frick and White 2008). There are other substantive and prognostic differences as well between children whose conduct problems are accompanied by these features of psychopathy and those whose are not (Blair et al. 2006; Kotler and McMahon 2005; Loeber et al. 2009). Such findings suggest that interventions that ameliorate child psychopathic features, in addition to child conduct problems, would offer a significant public health benefit. Research on the effects of interventions on features of psychopathy in pre-adolescent children is in its infancy. Nonetheless, there are strong and differing opinions about the degree to which psychopathic features can be changed (see Salekin 2002, 2010). Some have suggested that because psychopathic features are neither completely heritable nor completely deterministic, interventions targeting key aspects of the environment could influence the developmental course of psychopathic features (Pardini and Loeber 2007; Salekin et al. 2010). Consistent with this, some reviewers have concluded—primarily from results of research with adults and adolescents—that interventions can indeed exert positive effects on psychopathic features (Salekin 2002, 2010). However, there is disagreement about the quality of the evidence for this conclusion, and other reviewers have reached the opposite conclusion (e.g., Harris and Rice, 2006). Some researchers also doubt that treatment can have positive effects on psychopathic features because of the nature of psychopathy itself. The biological underpinnings and significant heritability of psychopathy, together with the lack of motivation to change and the affective deficits (lack of remorse, shallow emotion) that characterize psychopathy, are thought to pose inherent barriers to treatment engagement and therapeutic change (see Salekin 2010).
1 Features of psychopathy are also correlated with other psychological disorders in children, such as depression and anxiety (Kubak & Salekin, 2009). However, we limited the discussion of measurement overlap to conduct problem behavior because it is the focus of the study and the sample was recruited on the basis of elevated levels of conduct problems.
J Abnorm Child Psychol (2011) 39:1013–1023
Principles of developmental psychopathology suggest that if psychopathy is indeed a developmental phenomenon, it should be amenable to treatment if intervened upon early in its development. Consistent with the arguments for the mutability of psychopathic features, a developmental psychopathology approach suggests that transactional processes, in which salient environmental influences interact with a child’s psychological and biological predispositions (Todd et al. 1995), influence the development of psychopathic characteristics. This implies that psychopathic features gradually emerge and that their trajectory over time can be influenced by environmental factors. If so, intervening on the child’s environment early in the course of the development of psychopathy may alter its developmental trajectory. Although there is almost no scientific knowledge on the efficacy of interventions to reduce psychopathic features in young children, seminal researchers of psychopathy have pointed to parenting as a potentially important influence in the development of psychopathy (Hare 1970; W. McCord and J. McCord 1964), and consequently, as a potentially important target for intervention. Parenting exerts a key environmental influence on child behavior, and it is a robust predictor of child antisocial behavior (Patterson 1982). In particular, harsh and inconsistent parenting and parent-child aggression are associated with the development of child conduct problems (Gershoff 2002; Patterson 2002) and with the development of psychopathic features as well (Frick et al. 2003; Pardini et al. 2007). Moreover, certain aspects of parenting have been theorized to inhibit the development of psychopathic features in children. For example, firm and consistent discipline that focuses on obedience, without harshness, is believed to be important for staunching the further development of psychopathy, and empirical work has yielded results consistent with this supposition (e.g., Cornell and Frick 2007). Another view, also supported empirically, is that a strong, positive parent-child attachment, characterized by warm and responsive parenting and a focus on the positive aspects of the parent-child relationship, is important for fostering conscience development and empathy, and reducing the risk of further development of psychopathic features (Kochanska 1997; Pardini et al. 2007). Taken together, these two lines of thought suggest that interventions that help parents develop and consistently use firm responses to disobedience (Cornell and Frick 2007), but do so within the context of a warm and supportive parent-child relationship (Kochanska 1997), might reduce the likelihood that psychopathic features will develop. Interventions that target parental disciplinary strategies and the quality of the parent-child relationship are effective in reducing child conduct problems (Kaminski et al. 2008). However, researchers have only recently begun to examine
J Abnorm Child Psychol (2011) 39:1013–1023
whether or not parenting interventions are also effective for reducing psychopathic features in childhood. Two studies have considered child psychopathic features as an outcome variable in parenting interventions. In a sample of 4–8 year olds referred for conduct problems, Hawes and Dadds (2007) found that their parenting intervention—which focused on parental warmth and positive reinforcement, and on calm use of time-out for noncompliance—resulted in reductions in child callous/ unemotional traits at post-treatment that were maintained at the 6-month follow-up. As is commonly found, features of psychopathy and conduct problem scores were moderately correlated; however, changes in psychopathic features over time predicted antisocial behavior at 6-months posttreatment, but changes in antisocial behavior did not predict later psychopathic features. This study provides evidence consistent with the idea that child psychopathic features can improve over time, and suggests that treatment effects on features of psychopathy may not simply be an artifact of treatment effects on conduct problems. However, because this study did not include a control group, reductions in levels of psychopathic features cannot be definitively attributed to the intervention. Kolko and colleagues (Kolko et al. 2009) conducted a randomized controlled trial of an intervention for 6- to 11-year-old children with conduct problems, in which one group was served through community settings (e.g., home, school & neighborhood) and the other through a clinic. The intervention in both groups included parent training, but also included other treatment methods such as family therapy and cognitive-behavioral therapy for the children. Measures of the affective and interpersonal features of psychopathy (i.e., callous/unemotional traits and narcissism) from the Antisocial Process Screening Device (Frick and Hare 2001) were used to assess features of psychopathy. Results indicated that features of psychopathy were reduced in both groups of children, and the reductions were maintained over a 3-year follow-up period. Again, this study provides evidence that child psychopathic features can improve over time. However, the extent to which parenting changes may have accounted for the improvements in this study is indeterminate: The study did not include a no-treatment control group, so conclusions about the effects of the intervention on psychopathic features must be tempered accordingly. Clarifying whether parenting interventions can reduce child psychopathic features, and whether they do so as hypothesized, by altering harsh and inconsistent parenting practices and improving the quality of the parent-child relationship, has important implications for understanding the mutability of psychopathic features. In addition, because of the relation between child psychopathic features and conduct problems, it would be most valuable to determine if observed changes in psychopathic features
occur independently of changes in child conduct problems. In this study, we examine effects of Project Support—an intervention designed for children with conduct problems— on the affective and interpersonal features of psychopathy in children. Project Support is a parenting intervention that has been found to be effective in altering parenting and reducing conduct problems in samples of 4–9 year old children in families in which severe intimate partner violence has occurred (Jouriles et al. 2009, 2001; McDonald et al. 2006). Based on a transactional model of the development of psychopathic characteristics and the intervention findings reviewed above, we hypothesized that children in the Project Support condition would show improvements in the affective and interpersonal features of psychopathy, but children in a comparison condition would not. In addition, we expected that reductions in harsh and inconsistent parenting would mediate the effects of Project Support on features of psychopathy.
Method Sample and Procedures This study was a secondary data analysis of a randomized clinical trial evaluating effects of Project Support (Jouriles et al. 2009) on children’s conduct problems. Participants were 66 families (mothers and children between 4 and 9 years) in which the mothers had sought refuge at a domestic violence shelter. Mothers participated in an inshelter assessment to determine their family’s eligibility for the clinical trial. Eligible families were those in which the mother 1) reported at least one act of physical intimate partner violence (IPV) from a male partner during the previous 12 months; and 2) at least one child in the family between 4 and 9 years (who had accompanied the mother to the shelter) who met diagnostic criteria for oppositional defiant disorder or conduct disorder, as described in the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV; American Psychiatric Association 1994). The diagnostic eligibility criteria were assessed via a structured interview with mothers (during their shelter stay) focused on the symptoms of ODD and CD (see Jouriles et al. 2009; McDonald and Stephens 1998). After families left the shelter, they remained eligible for the study if the mother established a residence independent of her abusive partner. The baseline assessment was conducted in the families’ post-shelter residence. On average, these assessments occurred approximately 26 days following shelter departure. At the conclusion of the baseline assessment, families were randomly assigned to either the Project Support (n=32) condition or the services-as-usual comparison (n=34) condition. The average age of mothers in
the sample was 29.45 (SD = 5.21) years, with 11.63 (SD=1.13) years of education; the average age of the children was 6.16 (SD=1.66) years. There were 3.24 children per family, on average. Only one child per family participated however; for families with more than one eligible child in the age range, the youngest eligible child was chosen. Family income was $534 (SD=$537) per month on average. Mothers’ reports of ethnicity indicated that the sample consisted of approximately 40% Black, 40% White, and 20% Latino families. The groups did not differ on any of the measured demographic variables. Additional details about the sample and procedures are provided in Jouriles et al. (2009). Families in the Project Support condition received a family intervention that included two primary components: 1) parent-training, and 2) provision of instrumental and emotional support to mothers. The parenting skills were detailed in a manual that specified the particular skills to be taught, scenarios for practice role plays using the skills, and homework assignments. The parent-training component of the program included 12 child management skills (e.g., listening to your child, praising, reprimanding). The skills were presented in sequence, with the initial skills focusing on improving the quality of the mother-child relationship and increasing prosocial child behavior, and the latter skills focusing on reducing problematic behavior. The instrumental and emotional support component of the intervention was based on Sullivan and colleagues’ advocacy intervention for women departing from domestic violence shelters (Sullivan and Bybee 1999; Sullivan et al. 1992). Therapists regularly assessed and addressed safety concerns, provided emotional support to the mothers, assessed families’ current needs (e.g., food, transportation, etc.), offered referrals and help as indicated, and delivered donated goods such as furniture and small household items. Measures of treatment fidelity indicated that the parenting skills were delivered as intended (see Jouriles et al. 2009, for details). Families assigned to the Project Support condition could receive Project Support services for up to 8 months following shelter departure. Families received an average of 20 (SD=9, range=2 to 40) homebased treatment sessions during the 8-month intervention period (i.e., between the pre- and post-treatment assessments) following shelter departure. Project staff attempted to contact families in the comparison condition monthly, either in person or by telephone. These monthly contacts were structured so that the families could receive support services similar to those provided to Project Support families. In addition, no restrictions were placed on comparison families’ receipt of services from other sources; indeed, we encouraged them to make use of community resources. During the 8-month period following shelter departure, families assigned to the
J Abnorm Child Psychol (2011) 39:1013–1023
comparison condition averaged 3.7 (SD=2.66, range=0 to 9) contacts with project staff in which a safety issue was addressed, emotional support was provided, a referral was requested or offered, some form of instrumental support was provided, or the family received some combination of support services. Families in the comparison condition received no clinical services through our program or from project therapists that addressed parenting or child behavior. Among the 34 families in the comparison condition, 11 received some form of child mental health or parenting services outside of our project over the course of the 20month period following shelter departure. However, the services received were typically minimal. For example, six of these 11 families received three or fewer sessions of “counseling” for the target child. Assessments were conducted in families’ homes every 4 months over a 20-month period (baseline, 4, 8, 12, 16, and 20 months). A member of the research staff who was blind to the families’ experimental condition administered the measures. The assessments were quite lengthy: Each of the six assessments was administered over three separate days (2–3 hours per day) within a two-week period, and covered a broad array of constructs. Mothers in both conditions were paid for participating in the assessments. Also, to maximize sample retention, project staff attempted to contact all families monthly during the final 12 months of a family’s participation in the project (from 8 months to 20 months following shelter departure) to offer support. These contacts were structured similarly to those described above for families in the comparison condition. During this period, families in the Project Support condition averaged 5.4 (SD=4.84, range=0 to 16) contacts in which a safety issue was addressed, emotional support was provided, a referral was requested or offered, some form of instrumental support was provided, or the family received some combination of support services. Families in the comparison condition averaged 4.0 (SD=3.44, range=0 to 17) such contacts. Of the 66 participating families, 42 completed all six assessments, 9 completed five, 5 completed four, 4 completed three, and 6 completed two or fewer. Project Support families completed an average of 5 (SD=1.5) assessments as did comparison families (SD=1.3). Measures Features of psychopathy Mothers reported on their children’s psychopathic features on 16 items of the Psychopathy Screening Device (PSD; Frick et al. 1994) that were recommended for use on the basis of an early factoranalytic study by the authors of the scale. Mothers rated the extent to which each item characterized their child on a 3-point scale that ranged from 0-Not at all true to 2Definitely true. The PSD (and its later variant, the APSD;
J Abnorm Child Psychol (2011) 39:1013–1023
Frick and Hare 2001) is one of only two measures designed explicitly for assessing features of psychopathy in younger children (see Kotler and McMahon 2010), and it is the most widely researched measure in the small body of research on child psychopathic features (Forth and Book 2010). Although early work pointed to a 2-factor model (callous/unemotional traits and impulsivity/conduct problems; Frick et al. 1994) of psychopathic features in youth, more recent research also supports a 3-factor model consisting of callous/unemotional traits, narcissism, and impulsivity subscales (Forth and Book 2010; Kotler and McMahon 2005). Subscale and total scores from the full PSD have been found to correlate with DSM diagnoses of Conduct Disorder or Oppositional Defiant Disorder at levels suggesting that the measure is related to, but not fully overlapping with, these disorders (e.g., Frick et al. 1994). Using the 3-factor model, internal consistency for the subscales in this sample was callous/unemotional α=0.34, narcissism α=0.59, and impulsivity α=0.53. For the total PSD scale, α=0.68. Although the alphas for the subscales are low, they are consistent with other findings. For example, Dadds et al. 2005 reported alphas from 0.56 to 0.69 on the three subscales of the 20-item APSD, and lower internal consistency for the CU subscale has been noted by others (Kotler and McMahon 2010). Given the greater internal consistency of the total scale score, we used it in our initial analyses to test our hypotheses, and followed up with analyses evaluating treatment effects on the affective and interpersonal features (callous/unemotional and narcissism traits) of psychopathy. Harsh parenting Mothers’ psychological aggression and physical aggression toward the child were measured with the Psychological Aggression subscale and the Physical Violence subscale of the Revised Conflict Tactics Scale – Parent-Child (Straus et al. 1996). Mothers reported on these items, which include acts of psychological aggression (e.g., shouted, yelled or screamed; threatened to spank or hit but did not actually do it) and relatively minor (e.g., pushing, grabbing, and slapping) as well as more severe acts (e.g., kicking, biting, hitting with a fist) of physical aggression. Responses were made on a 7-point scale that ranged from 0Never to 6-More than 20 times. Coefficient alpha in this sample for the first assessment was 0.56 for psychological aggression and 0.54 for physical aggression. These reliabilities are consistent with those found by other researchers (0.55 and 0.60, respectively; Straus et al. 1996). The sum of the item scores was used in analyses. Inconsistent parenting Mothers completed the Consistency subscale of the Parenting Dimensions Inventory (PDI; Power 1993). This 4-item scale captures the degree to which mothers follow through on attempts to extinguish
unwanted behaviors. Sample items included “I always follow through on discipline for my child no matter how long it takes” and “My child can often talk me into letting him off easier than I had intended” (reverse scored). The response scale ranged from 1-Not at all descriptive of me to 6-Highly descriptive of me. Coefficient alpha for these items was 0.76 in this sample. Items were scored so that higher scores reflected more inconsistent parenting. Children’s conduct problems Mothers’ reports on the Externalizing Problems Scale of the Child Behavior Checklist (CBCL; Achenbach 1991) were used to assess children’s externalizing problems. The CBCL is widely used in research on child psychopathology and has welldocumented psychometric properties. Scores range from 0-Not True (as far as you know) to 2-Very True or Often True. T-scores for the Externalizing Problems subscale were used in analyses.
Results The mean PSD score at baseline was 13.45 (SD=3.97). As expected, this mean score is higher than that reported for community samples of children, as in the validation study of the 20-item PSD (Frick et al. 2000), M=9.8, SD=8.2 (we were unable to locate a study of a child-clinical sample that reported the mean total score on the 16-item PSD used in this study). The mean externalizing problems scale T-score was in the clinical range, M = 66.85 (SD = 7.95). The correlation of PSD scores with externalizing problems in our sample, r=0.39, p