The Development of Callous-Unemotional Traits and Antisocial Behavior in Children: Are There Shared and/or Unique Predictors?

Abstract

Callous and unemotional (CU) traits have been linked to severe antisocial behavior in youth, but studies examining the etiology of CU traits are lacking. Based on prior research, it was hypothesized that childhood anxiety and parenting practices would interact to predict changes in CU traits over time. Hypotheses were tested using a sample of 120 moderate to highly aggressive fifth graders followed over a 1-year period. Although CU traits displayed moderate temporal stability and predicted increases in antisocial behavior, evidence suggested that these features were not immutable. Children exposed to lower levels of physical punishment showed decreases in CU traits over time, whereas higher levels of child-reported parental warmth and involvement predicted decreases in both CU traits and antisocial behavior over time. Lower levels of anxiety were uniquely related to increased CU traits for children who described their primary caregiver as exhibiting low warmth and involvement.

Data were collected with a grant from the Centers for Disease Control and Prevention (R49\CCR418569), with further support provided by the National Institute of Drug Abuse (DA08453; DA16135) and WT Grant Foundation (all awarded to John Lochman). Dustin Pardini received support from the National Science Foundation (SES-0215551) through the National Consortium on Violence Research. Any opinions, findings, conclusions, or recommendations expressed in this article are ours and do not necessarily reflect the views of the National Science Foundation.

Special thanks to Matthew McLean, Jamie Pardini, and Ernesta Pardini for their assistance on this project.

Notes

¹Confirmatory factor analysis was conducted to validate the notion that items measuring CU traits and antisocial behavior were best represented by two latent factors at Time 1 and Time 2, rather than a single construct. Models were specified using mean and variance-adjusted weighted least squares estimation with delta parameterization using Mplus 3.11 (L. Muthén & Muthén, 1998–2004). The relative fit of nested models was tested with an asymptotic chi-square difference test appropriate for use with a robust estimator (see B. Muthén, 2004). Absolute fit of the two-factor model was examined using the comparative fit index (CFI) and Tucker–Lewis index (TLI). Conventional guidelines of .90 or greater for “acceptable fit” and .95 or greater for “good fit” for the CFI and TLI were used (McDonald & Ho, 2002). At Time 1, the two-factor model fit significantly better than the one-factor model and provided an adequate overall fit to the data, Δχ²(1, N = 120) = 28.53, p < .0001, CFI = .94, TLI = .97. The two-factor model also provided a better fit to the data in comparison to the one-factor model at Time 2, Δχ²(1, N = 120) = 25.08, p < .0001, CFI = .93, TLI = .97.

Note: T1 = Time 1 assessment; T2 = Time 2 assessment; PT = combined parent and teacher report; P = parent report; T = teacher report; C = child report.

∗p < .05. ∗∗p < .01. ∗∗∗p < .001.

Note: All parameter estimates are standardized betas. CUT1 = callous-unemotional traits at Time 1; ABT1 = antisocial behavior at Time 1; T1 = Time 1; P = parent report; C = child report.

^† p < .08.

∗p < .05.

∗∗p < .01.

Note: All parameter estimates are standardized betas. Results are from separate linear regressions in which Time 2 (T2) levels of the dependent variable were regressed onto Time 1 (T1) levels of the dependent variable and a single predictor. PT = combined parent and teacher report P = parent report; C = child report.

∗p < .05.

∗∗p < .01.

²Based on cross-sectional research, some investigators have suggested that parenting practices may not be associated with the development of antisocial behavior for children exhibiting high CU traits (Oxford, Cavell, & Hughes, 2003; Wootton et al., 1997). Although not the focus of the investigation presented here, this notion was tested in post hoc analyses by creating interaction terms between Time 1 CU traits and each parenting scale and adding these to the regressions predicting Time 2 antisocial behavior. None of the interactions reached statistical significance (all ps < .05). Further details regarding these supplementary analyses are available upon request.

Note: Results are from separate linear regressions predicting Time 2 (T2) Callous-Unemotional Traits and T2 Antisocial Behavior. T1 = Time 1; PT = combined parent and teacher report; P = parent report; C = child report.

^† p = .06.

∗p < .05.

∗∗∗p < .001.