I have attached the Final assignment since this assignment is building off of it this will be a big deal. I will let you pick the four domains that you are comfortable with and if you want I will help

Heterogeneity in Externalizing Problems at Age 3: Association With Age 15 Biological and Environmental Outcomes Kostas A. FantiUniversity of Cyprus Eva Kimonis The University of New South Wales Investigating heterogeneity in antisocial behavior early in life is essential for understanding the etiology, development, prognosis, and treatment of these problems. Data from the longitudinal National Institute of Child Health and Development (NICHD) study of Early Child Care were used to identify homoge- neous groups of young antisocial children differentiated on externalizing problems, internalizing prob- lems, and callous-unemotional (CU) traits using latent profile analysis (LPA). We examined how identified subgroups were differentiated on adolescent social, biological, cognitive, and environmental outcomes, controlling for dispositional and contextual antecedents during the first 2 years of life. The sample consisted of 1,167 children (52% male) followed from toddlerhood to adolescence. LPA identified a large “low problems” group (n 795; 49.9% male) as well as 3 antisocial groups at age 3:

the first scored high on internalizing and externalizing problems but low on CU traits (Ext/Int,n 125), the second scored high on CU traits and externalizing problems but low on internalizing problems (primary CU variant,n 135), and the third scored high on CU traits, internalizing, and externalizing problems (secondary CU variant,n 112), and these differences persisted into adolescence. Primary and secondary CU variants were further differentiated from one another on adolescent measures of aggression (reactive and relational), biological indices (cortisol, heart rate), cognitive abilities, and parental psy- chopathology, after controlling for early life risk factors (i.e., maternal sensitivity, difficult temperament, and maternal depression). We discuss implications of our findings for research, theory, and practice on early childhood externalizing problems.

Keywords:internalizing and externalizing problems, callous-unemotional traits, primary and secondary variants, cortisol and heart rate, parenting Antisocial behaviors developing early in life are a risk factor for stable and persistent problems that continue into later adulthood, placing the child on a developmental pathway of low academic achievement, poor parent and peer relations, and other maladaptive outcomes (Coie & Dodge, 1998;Fanti & Henrich, 2010;Keiley, Lofthouse, Bates, Dodge, & Pettit, 2003). Early in life these behaviors take the form of aggression, defiance, and destructive behavior, defined as externalizing problems, and have been found to precede child, adolescent, and adult antisocial problems (e.g., Achenbach, 1991;Fanti & Henrich, 2010;Leadbeater, Thompson, & Gruppuso, 2012;Moffitt, 1993). Importantly, these externaliz- ing problems are marked by heterogeneity, with existing subtyping approaches proposing distinct subgroups according to co- occurring psychopathology or deficits in prosocial development (i.e., CU traits).

Two approaches to understanding heterogeneity in antisocial behavior have received considerable attention in the literature.

The first differentiates antisocial children into those with andwithout internalizing problems, including anxiety and depres- sion (e.g.,Fanti & Henrich, 2010;Keiley et al., 2003). This research suggests that distinct subgroups of children exist as early as toddlerhood that are either characterized by high ex- ternalizing and normative levels of internalizing problems or by co-occurring externalizing and internalizing symptoms. Cross- sectional and longitudinal studies suggest that these two sub- groups can be differentiated on dispositional, environmental, and biological factors (Beauchaine, Gartner, & Hagen, 2000; Fanti & Henrich, 2010;Leadbeater et al., 2012;Schoorl, Van Rijn, De Wied, Van Goozen, & Swaab, 2016).

The second approach to subtyping antisocial children differen- tiates them into those with and without CU traits (i.e., low guilt, low empathy;Fanti, Demetriou, & Kimonis, 2013;Frick, Ray, Thornton, & Kahn, 2014), with emerging evidence supporting CU heterogeneity as early as preschool and toddlerhood (Kimonis, Fanti, Anastassiou-Hadjicharalambous, et al., 2016;Willoughby, Waschbusch, Moore, & Propper, 2011). Children with stable ex- ternalizing problems without CU traits tend to score high on anxiety and to be more fearful, whereas those with stable high externalizing problems with co-occurring CU traits tend to score at average levels of anxiety and low on behavioral and physiological measures of fear (Fanti, Panayiotou, Lazarou, Michael, & Geor- giou, 2016;Frick et al., 2014). Thus, the group of children with co-occurring externalizing and internalizing problems might re- semble antisocial youth low on CU traits.

This article was published Online First April 13, 2017.

Kostas A. Fanti, Department of Psychology, University of Cyprus; Eva Kimonis, School of Psychology, The University of New South Wales.

Correspondence concerning this article should be addressed to Kostas A.

Fanti, Department of Psychology, University of Cyprus, P.O. Box 20537, CY 1678, Nicosia, Cyprus. E-mail:[email protected] This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Developmental Psychology© 2017 American Psychological Association 2017, Vol. 53, No. 7, 1230 –12410012-1649/17/$12.00http://dx.doi.org/10.1037/dev0000317 1230 Adding complexity, accumulating evidence supports further het- erogeneity within antisocial children high on CU traits. This liter- ature consistently identifies two groups with equivalent high levels of CU traits, one with co-occurring internalizing problems (anxi- ety, depression) known as secondary CU variant, and a second with low to average anxiety levels known as primary CU variant (Fanti, Demetriou, et al., 2013;Kimonis, Skeem, Cauffman, & Dmitrieva, 2011). Exposure to adverse or abusive home environ- ments is thought to be critical to the development of secondary CU traits, whereas biological risk factors play a more prominent role in developmental accounts of the primary variant (Karpman, 1941; Kimonis et al., 2011). Secondary CU variants show biological correlates associated with trait anxiety (Kimonis, Fanti, Goulter, & Hall, 2016;Kimonis, Frick, Cauffman, Goldweber, & Skeem, 2012), which are in the opposing direction to fearless, emotionally hyporeactive primary variants, thus more closely resembling chil- dren with co-occurring externalizing and internalizing problems (seeFanti, 2016for a review).

The present study aims to integrate for the first time these separate literatures focused on subtyping antisocial children, to gain a more complete understanding of developmental outcomes of children with externalizing problems disaggregated on the basis of internalizing problemsandCU traits. We tested whether three groups of antisocial youth could be identified as early as age 3: a group scoring high on both externalizing and internalizing prob- lems but low on CU traits (Ext/Int); a group scoring high on CU traits and externalizing problems, but not internalizing problems (primary CU); and a group scoring high on CU traits, internalizing, and externalizing problems (secondary CU). We sought to com- pare identified subgroups on adolescent outcomes measured at age 15. Elucidating individual differences in the developmental out- comes of children with more homogeneous presentations of early antisocial behavior (i.e., subtypes) is of great importance toward improving models of antisocial behavior and providing avenues for prevention and intervention efforts.

The majority of studies subtyping antisocial youth on the basis of CU traits and internalizing problems focus on adolescents and rely on cross-sectional designs using retrospective reports of dys- functional parenting practices, highlighting the need for longitu- dinal research. As such, it is unclear whether subtyping models detailed above are useful for understanding antisocial behavior within young children. The current study aims to test whether antisocial subtypes can be identified in preschoolers, and whether they predict similar adolescent outcomes to those documented in the literature. Together, this literature finds that adolescent sec- ondary CU variants (high internalizing problems) show greater histories of social and environmental adversity and poorer self- regulation relative to primary variants (low internalizing), despite showing equivalent levels of CU traits (Kahn et al., 2013;Kimonis et al., 2011). Primary CU variants show low levels of emotional arousal fitting with “emotionally stable” descriptions (Hicks, Markon, Patrick, Krueger, & Newman, 2004;Kimonis et al., 2012, 2016), and are more likely to engage in relational forms of ag- gression necessitating better developed social skills to facilitate manipulation of others (Rosan, Frick, Gottlieb, & Fasicaru, 2015).

By contrast, secondary CU variants more closely resemble antiso- cial youth with comorbid internalizing problems with their pro- pensity toward emotional dysregulation and reactive aggression(Bubier & Drabick, 2009;Eisenberg et al., 2009;Fanti & Henrich, 2010).

The many differences between these antisocial subgroups sug- gests they are likely to also differ in other ways, such as in cognitive and academic functioning; factors that have been studied extensively in antisocial youth more generally. Low cognitive and academic functioning are common among children with external- izing problems, with or without co-occurring internalizing prob- lems, suggesting they may represent a general risk factor for antisocial subgroups (Fanti & Henrich, 2010;Masten et al., 2005; Moffitt, 1993). However, the poor self-regulation of Ext/Int and secondary CU variants suggests they could show poorer cognitive abilities (IQ/achievement) relative to the more emotionally and behaviorally regulated primary CU group. The primary group may drive hypotheses that individuals with CU traits show average to above average cognitive abilities (Cleckley, 1976). Agreeing with this suggestion, individuals with co-occurring internalizing and externalizing problems tend to exceed those exhibiting pure prob- lems in terms of greater functional interference in daily life and more impairment across domains such as educational and social functioning (Fanti & Henrich, 2010;Ingoldsby et al., 2006;New- man, Moffitt, Caspi, & Silva, 1998).

One reason for poor academic achievement among children with externalizing problems is the considerable disruption they cause within school contexts. Similarly, their home environments are considerably disruptive and marked by conflicted parent-child relationships (Ingoldsby et al., 2006). Parent-child conflict has been identified as a common risk factor for multiple forms of childhood psychopathology (Burt, Krueger, McGue, & Iacono, 2003), and thus may be relevant to all externalizing subtypes.

Importantly, the child’s antisocial behaviors might elicit parent- child conflict, which, in turn, exacerbates the child’s antisocial behaviors, particularly when the conflict occurs with a depressed mother (Lytton, 1990;Marmorstein & Iacono, 2004). In fact, parental psychopathology is a risk factor for both internalizing and externalizing problems (Fanti & Henrich, 2010;Fanti, Panayiotou, & Fanti, 2013;McLeod & Shanahan, 1996). For example, children of depressed mothers are almost five times more likely to develop depression by adolescence than those without depressed mothers (Murray et al., 2011), with both genetic and environmental mech- anisms contributing to this intergenerational transmission. Impor- tantly, externalizing, internalizing, and CU characteristics have been associated with increases in maternal depression and distress (Fanti & Munoz Centifanti, 2014; Fanti, Panayiotou, et al., 2013), although no prior work investigated such child effects by compar- ing Ext/Int, primary, and secondary CU subgroups.

To date, this line of research has been limited to environmental measures and cognitive tasks and is yet to examine whether differences between CU variants extend to neurobiological mea- sures. For instance, low heart rate and low cortisol levels are two biomarkers linked to antisocial behavior, and each has distin- guished externalizing individuals with and without CU traits, or with and without internalizing problems (Alink et al., 2008;Cima, Smeets, & Jelicic, 2008;Fanti, 2016;Frick et al., 2014;Loney, Butler, Lima, Counts, & Eckel, 2006;Ortiz & Raine, 2004;Raine, Venables, & Mednick, 1997;Stadler et al., 2011). However, across these literatures results are inconsistent and suggest the relation- ship between these biological indices and antisocial behavior may be more complex than previously thought (e.g.,Alink et al., 2008; This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 1231 HETEROGENEITY IN ANTISOCIAL BEHAVIOR Fanti, 2016;Hawes, Brennan, & Dadds, 2009). Thus, investigating heart rate and cortisol in antisocial subgroups has potential to shed light on the role of neurobiological factors in the development of antisocial behavior. Low heart rate might characterize emotionally hyporeactive antisocial individuals with primary CU traits, whereas hyperreactive children with co-occurring externalizing and internalizing problems might show high average heart rate (see Fanti, 2016for a review). Secondary CU variants might also be characterized by high heart rate due to their emotional dysregula- tion (Beauchaine et al., 2000), such that high heart rate may be a biological marker for antisocial behavior co-occurring with inter- nalizing problems irrespective of levels of CU traits (Fanti, 2016).

Findings regarding the association between cortisol and exter- nalizing problems are especially equivocal with some studies suggesting that externalizing problems are associated with low cortisol levels and others suggesting no association or a positive association (seeAlink et al., 2008for a meta-analysis). In fact, this meta-analysis identified only a small association between cortisol levels and externalizing problems (d .10), which was positive among preschoolers, negative among school-age children, and dropped to nonsignificance in adolescence. The authors concluded that cumulative exposure to high levels of environmental stress may explain the lower levels of basal cortisol identified in antiso- cial children by school age, and that heterogeneity in externalizing problems might account for the small effect size. The nonsignifi- cant association identified during adolescence calls for more re- search on this developmental period and the need for longitudinal research spanning early childhood to adolescence. Covariates: Early Risk Factors Investigation of adolescent outcomes associated with antisocial subgroups must consider the extensive literature on early risk factors, particularly those associated with CU traits. Low maternal sensitivity has emerged as an important early predictor of child externalizing problems at high levels of CU traits, likely reflecting its key role in the development of moral emotions and empathic responding (Frick & Morris, 2004;Kochanska, Kim, Boldt, & Yoon, 2013;Waller et al., 2014). In their longitudinal study, Barker and colleagues also found that maternal postnatal depres- sion (PND) and child temperament predicted the co-occurrence of externalizing problems and CU traits (Barker, Oliver, Viding, Salekin, & Maughan, 2011). Thus, we controlled for these age 3 antecedents to ensure that any observed associations between latent classes and adolescent outcomes were not accounted for by risk factors occurring during the first three years of life. The Present Study A large number of factors are associated with childhood anti- social behavior, but the many inconsistencies in the literature may be clarified by a more comprehensive consideration of heteroge- neity within this population. That is, studies that fail to consider comorbidity between externalizing problems, internalizing prob- lems, and CU traits together risk missing important distinctions between these antisocial subgroups that may be critical to inform- ing the field’s understanding of the development of antisocial behavior. The overarching aim of this study is to identify sub- groups of young children with externalizing problems differenti-ated on these variables. We hypothesize that we will identify three groups of children showing early childhood antisocial behavior, based on three separate but complementary lines of subtyping research aimed at better understanding heterogeneity in childhood antisocial behavior: externalizing/internalizing (Ext/Int), primary CU (Ext/CU) and secondary CU (Ext/Int/CU) variants.

We hypothesize that subtypes identified at preschool age will predict adolescent outcomes aligning with the existing literature base. That is, we hypothesize that primary and secondary CU variants will show the greatest externalizing problems and CU traits at age 15, with levels indistinguishable from one another, supporting their developmental stability. We predict that children showing co-occurring internalizing and externalizing problems (Ext/Int and secondary CU) will continue to show the greatest levels of adolescent internalizing problems. With regard to social, biological, cognitive, and parental measures, we hypothesize that primary CU variants will show higher self-regulation, social skills, relational aggression, and cognitive/achievement functioning, and lower heart rate relative to secondary CU variants. Due to the inconsistencies reported in the literature, we did not have any specific predictions for cortisol level differences. Further, we hy- pothesize that secondary CU variants will show greater reactive overt aggression compared with all other groups, and that comor- bid groups (Ext/Int, secondary CU) would both show poorer self- regulation, social skills, cognitive/academic functioning, and greater maternal depression relative to primary CU and low-risk groups. We predict parent-child conflict will be higher in all externalizing groups relative to low risk youth. To investigate the unique influence of these factors, we examine these associations covarying early risk factors associated with externalizing problems at high levels of CU traits (difficult temperament, maternal PND, and low sensitivity). The study’s hypotheses are presented inTable 1for clarity. Method Participants The present study used data from the NICHD Study of Early Child Care. This study was conducted by the NICHD Early Child Care Research Network supported by NICHD through a cooper- ative agreement that called for scientific collaboration between the grantees and NICHD staff. Participants were recruited from dif- ferent hospitals across 10 locations in the United States based on conditional random sampling that was used to assure that the sample was representative of single mothers, poverty status, ethnic minority, and low maternal education. Sampling information is described in detail elsewhere (https://www.nichd.nih.gov/research/ supported/seccyd/Pages/overview.aspx). The analyses for the cur- rent study were based on 1,167 children (51.7% male) whose mothers completed the Child Behavior Checklist (CBCL) at age 3.

At age 15, the sample size ranged from 834 to 975 adolescents based on the behavioral, questionnaire, and biological measures used in analyses. To retain all participants, outcome analyses were conducted using the multiple imputation feature in SPSS 22. The sample used for the current study was diverse in terms of ethnicity (77.5% were White, 6% were of Hispanic descent, 11.7% were African American, 1.3% were Asian, and 3.5% represented other minority groups) and family income (70% scored above the pov- This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 1232 FANTI AND KIMONIS erty threshold; 20.4% scored below the poverty threshold at some point during data collection; 9.6% scored below the poverty threshold continuously during data collection). Based on inter- views with the mothers during the first year of data collection, 10% of mothers had not completed high school and 21% were single.

Grouping Variables at Age 3 Mothers completed the preschool version of the Achenbach System of Empirically Based Assessment (ASEBA;Achenbach & Rescorla, 2000) at age 3. The current study used items related to externalizing problems (26 items), anxiety/depression (11 items), and CU traits (5 items; seeWilloughby, Mills-Koonce, Gottfred- son, & Wagner, 2014;Willoughby et al., 2011). Item 58 (punish- ment doesn’t change his or her behavior) that was included in the CU index was removed from the externalizing composite. Reli- ability/validity information of the Preschool ASEBA subscales and the CU measure based on the NICHD data are reported in multiple studies (e.g.,Fanti & Henrich, 2010;Fanti, Panayiotou, et al., 2013;Willoughby et al., 2011,2014).

Possible Early Life Covariates Maternal sensitivity.Maternal sensitivity was calculated from ratings of mothers’ behavior toward the child during a videotaped interaction using semistructured free-play conditions at 6, 15, and 24 months (NICHD Early Child Care Research Net- work, 1998). A summed composite score of maternal sensitivity at each age was created from coding of the videotapes for emotional support, positive affect, lack of hostility, and respect of the child’s autonomy or efforts using a 4-point scale. Scores were averaged from 6 to 24 months with good internal consistency ( .75).

Interrater reliabilities of the composites at each age based on intraclass correlations across raters ranged from .83 to .87.Child temperament.Temperament was assessed with the Infant Temperament Questionnaire (Medoff-Cooper, Carey, & McDevitt, 1993). When the child was 1 and 6 months old, mothers were asked to respond to 38 items which were developmentally appropriate for young infants (rated on a 1- to 6-point scale from “almost never”to“almost always”). The items provide scores on five subscales: Activity, Adaptability, Approach, Mood, and In- tensity, and all items can be combined into a single difficult temperament scale. Because the stability from 1 to 6 months was high,r .77,p .001, we created a difficult temperament scale based on the average score of all the items from the first and sixth month collection ( .70).

Postnatal depression.Mothers completed the Center for Ep- idemiological Studies Depression Scale (CES-D;Radloff, 1977) when the study children were 1 month old. Mothers rated the frequency of their own depressive symptoms during the past week ( .91). Age 15 Outcome Measures Externalizing and internalizing problems.Mothers rated adolescents’ anxiety/depression and externalizing problems at age 15 using the school age version of the ASEBA (Achenbach, 1991).

Items were summed to compute externalizing and internalizing problem scores.

CU traits.The Youth Psychopathic Traits Inventory (YPI; Andershed, Kerr, Stattin, & Levander, 2002) is a self-report ques- tionnaire that assesses three dimensions of psychopathy. In the current study, participating children completed the 15-item Callous-Unemotional (CU, i.e., Callousness, Unemotionality, and Remorselessness) subscale ( .82). Each item is scored on a 4-point Likert-type scale ranging from “Does not apply at all”to “Applies very well.” Table 1 Hypotheses in Relation to Current Study Findings Adolescent outcomes Ext/Int Secondary (Ext/Int/CU) Primary (Ext/CU) Current study findings Age 15 outcomes Externalizing problems Moderate High High All groups Low, secondary Ext/Int Internalizing problems High High Low Secondary, Ext/Int Primary, low CU traits Low High High Secondary, primary Ext/Int, low Social outcomes Relational aggression Low Low High Primary Secondary, low Reactive aggression Low High Low Secondary Primary, low Social skills Low Low Moderate Low Primary, secondary Biological outcomes Heart rate No hyp. High Low Secondary Primary Cortisol No hyp. No hyp. No hyp. Primary, low Secondary Cognitive outcomes Self-regulation Low Low High Primary, low Secondary WJ-R cognitive score Low Low High Primary, low Secondary WJ-R achievement Low Low High Primary, low Secondary Parenting outcomes Conflict High High High All groups Low Maternal depression High High Low Secondary Primary, low Note. The three columns associated with each identified group report how we expected each hypothetical group to score on adolescent outcome measures:

Low, moderate, or high. We also specify instances where we did not have a specific hypothesis (No hyp. no hypothesis). The last column of the table indicates how the groups were found to differ on each measure, enabling the reader to determine whether the hypotheses were supported. Ext/Int externalizing/internalizing; Low low risk group; CU callous-unemotional; WJ-R Woodcock-Johnson Psychoeducational Battery–Revised. This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 1233 HETEROGENEITY IN ANTISOCIAL BEHAVIOR Social Outcomes Aggression.The six items used to create the Relational Ag- gression scale, based on an adaptation of theCrick and Grotpeter (1995)relational aggression instrument, had modest internal reli- ability ( .68). Reactive overt aggression was also based on six items adapted from the rating instrument developed byDodge and Coie (1987), with a Cronbach’s alpha of 0.87. Adolescents rated how true each of the items is of them on a 4-point scale, ranging from “not at all true”to“completely true.” Social skills.Mothers completed the 40-item Social Skills Rating System (Gresham & Elliott, 1990), which provides a broad assessment of child social skills. The response levels range from 0 (Never)to2(Very Often). The total scale, which is based on four subscales (Cooperation, Assertiveness, Responsibility, and Self- Control), had high internal reliability ( .93) Biological Outcomes Heart rate.Heart rate readings were taken using a blood pressure cuff and stethoscope (i.e., no automated instruments).

Heart rate readings were taken a second time if the child appeared anxious or if the initial readings were high, and this second measurement was used in analyses. For uniformity across data collection sites, the heart rate was taken with the child seated using the nondominant arm. The majority of exams were conducted by nurse practitioners.

Cortisol.Adolescents and their parents were given detailed instructions on the saliva collection procedure and were trained in the use of salivettes for collections in the morning hours during three consecutive school days. Immediately after collection they completed a daily diary reporting on the time and date of sample collection, quality of sleep, medication used, average minutes elapsed from awakening to cortisol acquisition, and average time of awakening. They also reported their general sleep problems (7 items indicating greater difficulty falling and staying asleep) and smoking habits. These variables were included as covariates in analyses. Details on the assay procedure are provided inRoisman et al. (2009). Samples were assayed in duplicate and were aver- aged for analyses. Cortisol values (fxg/dl) were averaged over the 3 days of data collection since the correlations among the three samples were large by Cohen’s criteria (rs ranged from .38 to .52 across the 3 days of sampling; allps .001). Outlier values falling 3SDabove the mean were assigned the next highest value that was mean 3SD. Raw untransformed data were used in analyses since averaged cortisol values were only moderately skewed (skew 1.08).

Cognitive Outcomes Self-regulation.Self-regulation was assessed by averaging the standardized scores on a measure of self-reported impulse control and two behavioral self-control tasks, the Tower of London and the Stroop task. Self-reported impulse control was measured with a seven-item questionnaire assessing reactions to external constraints, taken from the Weinberger Adjustment Inventory (Weinberger & Schwartz, 1990). The measure asked adolescents to rate (1 falseto 5 true) how closely their behavior matched a series of statements (e.g., “I do things without giving themenough thought,” reverse scored), resulting in an impulse control composite score ( .82). Abehavioral measure of impulse controlwas generated with a computerized version of the Tower of London task (Steinberg et al., 2008), which was used to measure the adolescent’s ability to inhibit acting before a plan is fully formed. For analyses, the variable of interest was the average latency (in milliseconds) to first move for more complex five-, six-, and seven-move problems, indicating greater impulse control.

Response inhibitionwas assessed with the computerized version of the classic Stroop color-word task. Adolescents completed two 48-trial experimental blocks, the first of which included an equal mix of neutral and incongruent (e.g., the word “blue” printed in yellow) trials, and the second of which (‘unequal block’) included a greater number of neutral than incongruent trials. We calculated an interference score as the difference between the average re- sponse time on incongruent versus neutral trials within unequal blocks, and reverse-scored this value so that higher scores repre- sented stronger response inhibition.

Cognitive/academic achievement.The Woodcock-Johnson Psychoeducational Battery–Revised (WJ–R;Woodcock, Johnson, & Mather, 1990) consists of two major parts: the Tests of Cogni- tive Ability and the Tests of Achievement. At age 15, adolescents’ cognitive ability was assessed with two subscales, Picture Vocab- ulary and Verbal Analogies. Achievement was assessed using the Passage Comprehension and Applied Problems subscales. We used standard scores in analyses, which are based on a mean of 100 and a standard deviation of 15. Parenting Related Measures Conflict.The Child-Parent Relationship Scale (Pianta, 1992) is a 15-item questionnaire designed to assess the parent’s feelings and beliefs about his or her relationship with the adolescent on a 5-point, Likert-type scale ranging from 1 (Definitely does not apply)to5(Definitely applies). For the purposes of the current study, mothers reported on their Conflict (7 items, .87; i.e., always struggling with each other) with their adolescents.

Maternal psychopathology.Maternal depression was as- sessed at age 15 with the same CES-D measure used in infancy ( .92).

Plan of Analyses Identifying variants.LPA in Mplus 7 (Muthén & Muthén, 2010) was used to identify distinct subgroups of individuals based on their scores on CU traits, externalizing problems, and anxiety/ depression assessed at age 3. LPA identifies different latent classes by decomposing the covariance matrix to highlight relationships among individuals, and clusters individuals that are similar on the constellation of indicators into latent classes (Bauer & Curran, 2004). Models that specify different numbers of classes are tested.

The Bayesian information criterion (BIC) and Lo-Mendel-Rubin (LMR) statistics are used as statistical criteria to compare models to identify the optimal number of groups to retain (Nylund, Asp- arouhov, & Muthén, 2007). The model with the lowest BIC value is preferred. A nonsignificant chi-square value (p .05) for the LMR statistic suggests that a model with one fewer class is preferred. Further, average posterior probabilities and entropy val- ues equal to or greater than .80 indicate clear classification and This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 1234 FANTI AND KIMONIS greater power to predict class membership (Clark & Muthén, 2009).

Identifying potential covariates.We used multinomial logis- tic regression analyses (MLR) in IBM SPSS 22 to identify poten- tial early life covariates (i.e., difficult temperament, maternal de- pression, and maternal sensitivity) to include in analysis of covariance (ANCOVA).

Outcome analyses.We first used ANCOVA (controlling for significant covariates) to compare the identified groups on age 3 and age 15 CU traits, externalizing, and internalizing problems.

Second, we used ANCOVA to compare the groups identified at age 3 on adolescent outcomes collected at age 15. The dependent variables were grouped into four categories: (a) social, (b) biolog- ical, (c) cognitive, and (d) parent-related outcomes. Because com- paring groups identified with LPA might underestimate standard errors of the parameters,Clark and Muthén (2009)recommend using a more stringent alpha level for comparisons. Following this recommendation and because of the multiple comparisons per- formed in the current study, the Bonferroni correction was used in all analyses. Results LPA Findings To identify the optimal number of groups to retain, we com- pared two LPA models, with the first model excluding covariates and the second model including the three hypothesized covariates (difficult temperament, maternal depression, and maternal sensi- tivity):

1. For the model with no covariates, the BIC statistic in- creased from Class 5 (BIC 17,040.83) to Class 6 (BIC 17,073.25) and decreased from Class 4 (BIC 17,145.69) to Class 5. In addition, the LMR statistic fell out of significance for the six-class model (p .59).

Thus, the five-class model better represented the data based on the BIC and LMR statistics. The mean posterior probability scores ranged from .89 to .94 and the entropy value was .82.

2. For the model including the three covariates, the BIC statistic increased from Class 4 (BIC 16,870.55) to Class 5 (BIC 16,933.63) and decreased from Class 3 (BIC 16,908.41) to Class 4. In addition, the LMR statistic fell out of significance for the five-class model (p .46). As a result, the four-class model better repre- sented the data. The mean posterior probability scores ranged from .90 to .93 and the entropy value was .81, suggesting that the identified classes were well separated Because both the five-class (without covariates) and four-class (including covariates) models identified similar high-risk groups, we decided to use the more parsimonious four-class model includ- ing covariates.Figure 1shows standardizedz-scores by group on each grouping variable. The majority of the sample scored below average (n 795; 49.9 male) on all of the measures under investigation. This group is described hereafter as the “low prob- lem” group. Children in the Ext/Int group (n 125; 52.8% male)scored at approximately 1SDabove the mean on anxiety/depres- sion and externalizing problems, but below average on CU traits.

Children in the secondary CU group (n 112; 55.4% male) scored 1SDabove the mean on all LPA measures. Children in the primary CU group (n 135; 57.8% male) scored below average on anxiety/depression and approximately 1SDabove the mean on externalizing problems and CU traits. According to 2analyses, the identified groups were not differentiated in terms of ethnicity, 2(20,N 1,167) 13.88,p .54, or sex, 2(3,N 1,167) 6.30,p .09, and as a result these demographics were not included in further analyses. Testing for Potential Early Life Covariates The analysis comparing the identified groups using MLR was significant, 2(9,N 1,167) 114.82,p .001. Maternal sensitivity, 2(3,N 1,167) 26.59,p .001, difficult temper- ament, 2(3,N 1,167) 32.75,p .001, and maternal PND, 2(3,N 1,167) 21.54,p .001, significantly predicted the identified subgroups.Table 2incorporates odds ratios to compare the groups, which reflect the odds likelihood of being in one group over the other, based on the level of the independent variable. As shown inTable 2, children who experienced less maternal sensi- tivity were more likely to be in the secondary CU group compared with the low problem and Ext/Int groups. The primary CU group also scored higher than the low problem group. Children charac- terized by difficult temperament were more likely to be classified in the secondary, primary, and Ext/Int groups compared with the low problem group. Finally, children whose mothers scored high on PND were more likely to be classified in the secondary and Ext/Int groups than the low problem group. Thus, maternal sensi- tivity, difficult temperament, and maternal PND were included as covariates in all follow-up ANCOVAs.

Comparing the Identified Groups on Age 3 Grouping Variables The ANCOVA results, shown inTable 3, suggested that the low problem group had the lowest levels of externalizing problems and Figure 1.Groups identified using latent profile analysis at age 3 (N 1,167; 95% confidence intervals are included in the figure) after controlling for covariates. Anx/Dep anxiety/depression. This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 1235 HETEROGENEITY IN ANTISOCIAL BEHAVIOR CU traits, and scored similarly low on anxiety/depression as the primary CU group. The Ext/Int and secondary CU groups had the highest scores on anxiety/depression, reaching borderline clinical scores (Achenbach, 1991). The secondary CU group had the highest scores on externalizing problems (borderline clinical), followed by the primary CU and Ext/Int groups. The primary CU group also scored higher than the Ext/Int group. Regarding CU traits, the secondary group scored higher than the primary group, and both groups scored higher compared with the remaining groups.

Analyses of Age 15 Outcomes Analyses involving age 15 outcomes are shown inTables 3and 4. The reported results are based on analyses of the data imputed using the multiple imputation feature in SPSS 22. For all analyses Bonferroni post hoc tests are reported, which set the statistical significance level at .008 (i.e., for 6 comparisons). All findings are also reported inTable 1in relation to the study’s hypotheses.

Evidence for Stability of LPA-Related Variables From Age 3 to Age 15 As shown inTable 3, children in the low problem group scored lower on externalizing problems at age 15 compared with all other groups. The secondary group scored similarly on externalizing problems as the primary CU group and higher than the Ext/Intgroup, approaching borderline clinical levels. Children in the Ext/ Int and secondary CU groups continued scoring high on anxiety/ depression compared with primary CU and low problem groups.

Regarding CU traits, primary and secondary CU groups scored higher than all other groups, but did not differ from one another.

The findings support the stability of traits relevant to antisocial subtypes from toddlerhood to adolescence (12 years later).

Social outcomes.As shown inTable 4, primary CU variants were more likely to engage in relational aggression than the low problem and secondary CU groups. In contrast, secondary variants were more likely to engage in reactive aggression than the low problem and primary CU groups.Further, primary and secondary CU variants scored similarly on the measure of social skills and lower than the low problem group. The Ext/Int group did not differ from any other group on measures of aggression and social skills.

Biological measures.As reported inTable 4, the secondary CU group had higher heart rate compared to the primary group, with the remaining groups falling between these two groups and undifferentiated from one another. In terms of cortisol, the sec- ondary CU group scored significantly differently than the primary and low-risk groups, with the secondary group showing lower cortisol concentrations compared with both groups. In addition, group differences remained significant,F(3, 816) 2.94,p .05, 2 .01, after controlling for sleep problems, time awakening, medication used, smoking, and average minutes elapsed from awakening to cortisol acquisition. Table 2 Multinomial Logistic Regression Analyses Comparing the Identified Groups on Early Life Covariates Group comparisons based on odds ratios Covariates Primary vs. low Secondary vs. low Ext/Int vs. low Primary vs. Ext/Int Secondary vs. Ext/Int Primary vs. secondary Mat. sensitivity .68 (.55–.85).55 (.42–.70).84 (.64–1.09) .82 (.60–1.11) .65 (.47–.90)1.25 (.93–1.67) Difficult temp. 1.84 (1.22–2.76)2.44 (1.49–3.99)3.02 (1.89–4.83).61 (.35–1.06) .81 (.44–1.50) .75 (.42–1.33) Mat. depression 1.02 (.99–1.05) 1.04 (1.02–1.06)1.04 (1.02–1.06).99 (.97–1.02) 1.00 (.97–1.03) .98 (.96–1.02) Note.N 1,167. Antecedent variables are: maternal sensitivity (Mat. sensitivity), difficult temperament (Difficult temp.), and maternal depression (Mat.

depression); Ext/Int externalizing/internalizing. Comparisons are based on odds ratios; 95% confidence intervals are in parentheses. Only comparisons at p .008 are reported as significant. Table 3 Comparisons Between the Identified Groups on Age 3 (N 1,167) LPA Variables and Age 15 Outcomes Indicating Stability Over Time, After Controlling for Significant Early Life Covariates Age 3 and age 15 outcomes Low Ext/Int Secondary PrimaryFvaluedf 2 LPA measures—age 3 Externalizing problems 44.43 a(.22)58.07 b(.55)63.80 d(.60)59.75 c(.48)482.11 3 .56 Anxiety/depression 46.38 a(.26)61.99 b(.65)63.99 b(.70)48.03 a(.56)255.11 3 .40 CU traits 1.22 a(.04)1.50 a(.10)3.95 c(.11)3.29 b(.08)288.35 3 .43 Outcomes—age 15 Externalizing problems 48.06 a(.37)56.11 b(.90)60.59 c(.95)57.99 bc(.77)45.28 3 .11 Anxiety/depression 48.77 a(.37)57.93 b(.21)58.64 b(.98)50.61 a(.18)17.08 3 .04 CU traits 28.38 a(.19)27.79 a(.51)31.23 b(.54)30.64 b(.43)8.38 3 .03 Note. Estimated marginal means (SE); different superscripts ( a,b,c,d ) denote significant differences between groups in post hoc pairwise comparisons using the Bonferroni procedure. Only pairwise comparisons atp .008 are reported as significant. AllFstatistics are significant at thep .01 level.Tscores are reported for externalizing problems and anxiety/depression. LPA latent profile analysis; Ext/Int externalizing/internalizing; CU callous- unemotional. This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 1236 FANTI AND KIMONIS Cognitive measures.The secondary CU group was charac- terized by lower self-regulation compared with the primary CU and low problem groups (seeTable 4). Regarding WJ-R cognitive and achievement scores, the secondary CU group scored signifi- cantly lower than the low problem and primary CU groups, but not the Ext/Int group.

Parent-related measures.All externalizing groups scored higher on the measure of conflict with parents relative to the control group (seeTable 4). Further, mothers of children in the secondary CU group were more likely to be depressed compared to all other groups, except the Ext/Int group. Discussion The present longitudinal study set out to integrate parallel lines of research supporting heterogeneity in childhood externalizing problems by examining adolescent outcomes differentiating pre- schoolers subgrouped on the basis of externalizing and internaliz- ing symptoms and CU traits. We were particularly interested in whether primary and comorbid secondary variants of CU traits, previously identified only in adolescent samples, existed in young children. We were further interested in whether the secondary CU group could be differentiated from children with comorbid exter- nalizing and internalizing problems in the absence of CU traits, and whether these distinctions identified in early childhood dem- onstrated developmental stability into adolescence. Findings indi- cated that preschool children with elevated externalizing problems could be disaggregated into three subgroups, two of which showed nonnormative levels of CU traits ( 1SDabove the sample mean) and one that was low on CU traits. Comorbid groups (Ext/Int and secondary CU variants [Ext/Int/CU]) were undifferentiated across the full range of adolescent outcomes investigated, including anx- iety/depression levels, social, cognitive, biological, and parent- related outcomes. In contrast, there were a number of differences between high CU variants with primary variants showing greater self-regulation, relational aggression, intellectual and academic functioning, lower heart rate,highermorning cortisol levels, and less maternal depression than secondary variants. These differ-ences persisted after controlling for early life risk factors, includ- ing difficult temperament, maternal depression, and maternal sen- sitivity. We discuss these findings in turn below.

Our findings suggest that two prior lines of study, examining co-occurring externalizing and internalizing problems on the one hand, and externalizing problems with or without CU traits on the other, may be missing an important piece of the puzzle by failing to consider all three factors together: internalizing and externaliz- ing problems and CU traits. To date, high-anxious secondary CU variants have only been studied among adolescent populations and are similarly identified as lower in self-regulation and higher in reactive aggression and depression (e.g.,Fanti, Demetriou, et al., 2013;Kimonis et al., 2011). Adding to this work, current findings provide novel evidence that secondary variants also show lower IQ and achievement scores compared with primary CU variants. Also consistent with adolescent subtyping studies is our finding that primary and secondary CU variants showed comparable levels of CU traits and externalizing problems in adolescence (Fanti, Dem- etriou, et al., 2013), despite secondary variants presenting with the highest levels as preschoolers. Further, both groups showed low social skills during adolescence, such as assertiveness and respon- sibility, pointing to the low prosocial behavior of both groups.

These findings are important because CU traits have increas- ingly been used to designate an important subgroup of antisocial youth at heightened risk for early onset and persistent conduct problems, and are incorporated into theDiagnostic and Statistical Manual of Mental Disorders(5th ed.;American Psychiatric As- sociation, 2013) diagnosis of conduct disorder as a limited proso- cial emotions (LPE) specifier (Frick et al., 2014); however, youth scoring high on CU traits or diagnosed with conduct disorder with LPE could fall into either primary or secondary groups, which is likely to be relevant to their course, prognosis, and individual treatment needs, although these are questions in need of further study. The uptake and influence of the primary/secondary CU subtyping literature is likely to be a slow process just as it has taken decades to influence theory, research, and practice to recog- nize the importance of CU traits to antisocial behavior more Table 4 Comparisons Between Identified Groups on Adolescent (Age 15) Outcomes, After Controlling for Significant Early Life Covariates Age 15 outcomes Low Ext/Int Secondary PrimaryFdf 2 Social outcomes Relational aggression 1.32 a(.01)1.38 ab(.03)1.30 a(.03)1.44 b(.02)4.12 3 .01 Reactive aggression 1.63 a(.02)1.70 ab(.06)1.83 b(.05)1.60 a(.05)4.48 3 .02 Social skills 58.95 b(.54)57.42 ab(.95)55.16 a(.99)56.58 a(.89)8.34 3 .03 Biological outcomes Heart rate 72.53 ab(.19)72.51 ab(.46)73.38 b(.50)70.63 a(.40)2.63 3 .01 Cortisol .35 b(.01).33 ab(.03).29 a(.03).37 b(.02)3.88 3 .02 Cognitive outcomes Self-regulation .01 b(.02) .09 ab(.08) .21 a(.08).04 b(.04)2.87 3 .01 WJ-R cognitive score 107.15 b(.36)105.52 ab(1.71)102.22 a(1.73)107.34 b(.84)2.86 3 .01 WJ-R achievement 105.52 b(.36)103.56 ab(1.71)99.78 a(1.72)106.80 b(.83)4.99 3 .02 Parenting outcomes Conflict 16.83 a(.20)19.03 b(.50)19.65 b(.53)18.70 b(.43)13.48 3 .04 Maternal depression 10.12 a(.30)11.21 ab(.73)12.63 b(.79)10.13 a(.63)3.24 3 .01 Note. Estimated marginal means (SE); different superscripts ( a,b) denote significant differences between groups in post hoc pairwise comparisons using the Bonferroni procedure. Only comparisons atp .008 are reported as significant. AllFstatistics are significant at thep .05 level. Ext/Int externalizing/internalizing; WJ-R Woodcock-Johnson Psychoeducational Battery–Revised. This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 1237 HETEROGENEITY IN ANTISOCIAL BEHAVIOR generally. Relative to this broader CU traits literature, the study of primary and secondary variants is in its infancy with the first published empirical adult study in 2004 (Hicks et al., 2004) and youth study in 2009 (Vaughn, Edens, Howard, & Smith, 2009).

The burgeoning literature on CU variants, together with current study findings, support that this broader perspective can improve our understanding of antisocial behavior and has potential to clarify inconsistencies across the literature.

Comorbid internalizing problems are a key distinguishing factor between secondary and primary CU variants irrespective of develop- mental period. Secondary CU variants had significantly greater symp- toms of anxiety and depression than primary variants at age 3 and at age 15. Notably, children with externalizing problems without CU and those classified as secondary CU both showed high and nonsig- nificantly different levels of anxiety/depression in early childhood that persisted into adolescence, and both may be considered comorbid externalizing groups. Similarly, both groups also showed high and undifferentiated levels of maternal depression during the postnatal period and in adolescence relative to lower risk and primary CU groups. This suggests that heritable and learning factors that are central to theories of anxiety and depression may represent shared vulnerabilities to both comorbid presentations.

It is thought that both too high and too low levels of arousal in the context of emotionally charged early socializing experiences (i.e., parental punishment) negatively impact the development of moral emotions and contribute to antisocial behavior (Frick & Morris, 2004). For young children with secondary CU traits that show similar patterns of heightened emotionality to those with internalizing prob- lems, as reflected by high heart rate in the current study, it is likely that too high levels of emotional arousal contribute to this failure in moral socialization, although this is a question in need of further study.

Conversely, low heart rate is a long-established correlate of external- izing problems; however, our results suggest that inconsistent findings in the literature may be due to important differences between exter- nalizing subtypes. Consistent with a large body of research document- ing low levels of emotional arousal among antisocial youth with CU traits (Blair, 1999;Marsh, Gerber, & Peterson, 2008), primary CU variants showed significantly lower heart rate relative to secondary variants. Paired with their greater self-regulation compared with sec- ondary CU variants, this presentation fits adult conceptualizations of primary psychopathy as an emotionally stable and planful group of individuals. These suggestions are strengthened by current findings that childhood primary CU variants were more likely to engage in relational than reactive aggression, which requires a degree of plan- ning. By contrast, the average heart rate of secondary CU variants was highest within our sample, although undifferentiated from low prob- lem and Ext/Int groups after applying Bonferroni corrections. Simi- larly,Kimonis et al. (2012)found that primary and secondary CU variants identified in an incarcerated adolescent sample showed con- trasting patterns of emotional reactivity on a dot probe task. Whereas primary variants were underaroused to negative cues on the task, secondary variants were hyperaroused.

Cortisol emerged as a second potential biomarker differentiating externalizing subgroups. Secondary CU variants presented with the lowest morning cortisol concentrations and primary variants with the highest, and were the only groups significantly differentiated from one another. Although a priori predictions were not made, the some- what unexpected direction of these findings necessitates caution in their interpretation and a need for replication in future studies. Incon-sistencies and contradictions are embedded throughout the cortisol- antisocial behavior literature (Hawes, Brennan, & Dadds, 2009), with some studies reporting a negative relationship between diurnal corti- sol and psychopathic traits, and others finding no relationship (Cima et al., 2008;Holi, Auvinen-Lintunen, Lindberg, Tani, & Virkkunen, 2006). Mixed results may, in part, arise from differences in method- ology, such as collecting single-snapshot basal or afternoon measures, examining stress-reactive measures, or diurnal measures as we did in the present study. Another line of research relevant to our findings reports that low cortisol levels are associated with negative familial experiences, such as maternal insensitivity and history of maltreat- ment (Roisman et al., 2009;Tarullo & Gunnar, 2006). The attenuation hypothesis was posited to explain these findings, proposing that social stressors and chronic activation of the hypothalamic-pituitary-adrenal (HPA) axis in early development might result in attenuated cortisol levels (Susman, 2006), similar to the pattern observed among our secondary variants. The longitudinal nature of our study might further explain this association, sinceRuttle et al. (2011)found that while internalizing problems were concurrently associated with higher morning cortisol levels, internalizing problems assessed in childhood were associated with low cortisol levels during adolescence. Further, children with the most severe behavioral problems during childhood, which was the case for the secondary CU group, had the lowest cortisol levels during adolescence, pointing to greater dysregulation of the HPA axis. Thus, by taking heterogeneity into account and inves- tigating longitudinal associations, current findings provide support for disentangling the association between cortisol and antisocial behavior. Strengths, Limitations, and Conclusions There were a number of important strengths to this study, including its longitudinal design with a relatively large sample, the inclusion of a variety of measures assessed via multiple methods, and at a variety of relevant time points. Our findings must also be considered within the context of several study limitations. First, CU traits at age 3 were based on the CBCL and not an instrument designed to comprehen- sively assess CU traits, although prior work suggests the ASEBA provides a valid measure (Willoughby et al., 2014;Willoughby et al., 2011). Also, the YPI is a self-report instrument and while the validity of self-report measures of personality tends to increase with age as that of parent- and teacher-report declines from childhood to adoles- cence (Frick, Barry, & Kamphaus, 2010), self-report is susceptible to deception that is a cardinal feature of psychopathy. An additional limitation might be that grouping variables relied exclusively on maternal reports andassociations with mother-reported outcomes could be inflated due to shared method variance; however, mothers might be more observant of their children’s emotional characteristics compared with other informants (Keiley et al., 2003). Related to this limitation, measures of father psychopa- thology were not included in the current study. Finally, the current study was not genetically informed and thus many of these findings could reflect heritable, rather than contextual, processes.

Our findings have important implications for assessment, preven- tion, and treatment of preschoolers with externalizing problems. They are consistent with the broader research on childhood antisocial be- havior suggesting that the causes for these problems can vary greatly and that effective intervention requires both a comprehensive assess- ment of these various causal pathways and a matching of treatment to This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 1238 FANTI AND KIMONIS the unique needs of children across these pathways (Frick et al., 2014). In the current study, children identified into a high-risk group were at risk for problematic adolescent outcomes: Ext/Int and sec- ondary CU groups were at heightened risk for internalizing problems; secondary CU variants were particularly at risk for severe external- izing problems and reactive aggression; and primary CU variants were at greatest risk for relational aggression, relative to other exter- nalizing and low problem groups. These findings suggest that distin- guishing between externalizing subgroups on the basis of internaliz- ing problems and CU traits is likely to have clinical utility for understanding their prognosis.

Parent management training is the gold-standard intervention for young children with externalizing problems (Comer, Chow, Chan, Cooper-Vince, & Wilson, 2013); however, estimates suggest that approximately one third of treated children do not respond to tradi- tional intervention (Masi et al., 2011;Webster-Stratton & Hammond, 1997). Whereas children with anxiety problems tend to yield the greatest benefit from these programs, they are significantly less cost- effective for children with CU traits (Hawes, Price, & Dadds, 2014; cf.,Hyde et al., 2013;Waller et al., 2014). The reason for this relative treatment resistance appears to be that the externalizing problems of children with CU traits arise from some distinct factors from those typically targeted in traditional interventions. Recent efforts to im- prove treatment outcomes for children with CU traits by targeting their unique emotional deficits are promising (Dadds, Cauchi, Wimalaweera, Hawes, & Brennan, 2012;Datyner, Kimonis, Hunt, & Armstrong, 2016). Our findings suggest these efforts require expan- sion to consider the unique treatment needs of those children with and without comorbid internalizing problems. Integrating promising ad- aptations of traditional interventions to address early childhood inter- nalizing problems (Comer et al., 2013;Luby, Lenze, & Tillman, 2012) with those designed to address CU traits may help to improve outcomes for secondary CU variants that have not yet been the focus of intervention studies. Taking a more comprehensive approach to understanding childhood antisocial behavior that considers heteroge- neity in terms of internalizing and externalizing problems, and CU traits, has great potential to advance the field and improve lives. References Achenbach, T. M. (1991).Manual for the Child Behavior Checklist/4 –18 and 1991 profile. Burlington, VT: Department of Psychiatry, University of Vermont Burlington.

Achenbach, T. M., & Rescorla, L. A. (2000).CBCL/1 1/2–5: Empirically based scales for boys and girls; DSM-oriented scales for boys and girls.

Burlington, VT: University of Vermont, ASEBA.

Alink, L. R. A., van Ijzendoorn, M. H., Bakermans-Kranenburg, M. J., Mesman, J., Juffer, F., & Koot, H. M. (2008). Cortisol and externalizing behavior in children and adolescents: Mixed meta-analytic evidence for the inverse relation of basal cortisol and cortisol reactivity with exter- nalizing behavior.Developmental Psychobiology, 50,427– 450.http:// dx.doi.org/10.1002/dev.20300 American Psychiatric Association. (2013).Diagnostic and statistical man- ual of mental disorders(5th ed.). Washington, DC: Author.

Andershed, H., Kerr, M., Stattin, H., & Levander, S. (2002). Psychopathic traits in non-referred youths: A new assessment tool. In E. Blaauw & L.

Sheridan (Eds.),Psychopaths: Current international perspectives(pp.

131–158). The Hague, the Netherlands: Elsevier.

Barker, E. D., Oliver, B. R., Viding, E., Salekin, R. T., & Maughan, B.

(2011). The impact of prenatal maternal risk, fearless temperament and early parenting on adolescent callous-unemotional traits: A 14-yearlongitudinal investigation.Journal of Child Psychology and Psychiatry, 52,878 – 888.http://dx.doi.org/10.1111/j.1469-7610.2011.02397.x Bauer, D. J., & Curran, P. J. (2004). The integration of continuous and discrete latent variable models: Potential problems and promising op- portunities.Psychological Methods, 9,3–29.http://dx.doi.org/10.1037/ 1082-989X.9.1.3 Beauchaine, T. P., Gartner, J., & Hagen, B. (2000). Comorbid depression and heart rate variability as predictors of aggressive and hyperactive symptom responsiveness during inpatient treatment of conduct- disordered, ADHD boys.Aggressive Behavior, 26,425– 441.http://dx .doi.org/10.1002/1098-2337(200011)26:6 425::AID-AB2 3.0.CO;2-I Blair, R. J. R. (1999). Responsiveness to distress cues in the child with psychopathic tendencies.Personality and Individual Differences, 27, 135–145.http://dx.doi.org/10.1016/S0191-8869(98)00231-1 Bubier, J. L., & Drabick, D. A. (2009). Co-occurring anxiety and disruptive behavior disorders: The roles of anxious symptoms, reactive aggression, and shared risk processes.Clinical Psychology Review, 29,658 – 669.

http://dx.doi.org/10.1016/j.cpr.2009.08.005 Burt, S. A., Krueger, R. F., McGue, M., & Iacono, W. (2003). Parent-child conflict and the comorbidity among childhood externalizing disorders.

Archives of General Psychiatry, 60,505–513.http://dx.doi.org/10.1001/ archpsyc.60.5.505 Cima, M., Smeets, T., & Jelicic, M. (2008). Self-reported trauma, cortisol levels, and aggression in psychopathic and non-psychopathic prison inmates.Biological Psychology, 78,75– 86.http://dx.doi.org/10.1016/j .biopsycho.2007.12.011 Clark, S. L., & Muthén, B. (2009). Relating latent class analysis results to variables not included in the analysis. Retrieved fromhttps://www .statmodel.com/download/relatinglca.pdf Cleckley, H. (1976).The mask of sanity. St. Louis: Mosby.

Coie, J. D., & Dodge, K. A. (1998). Aggression and antisocial behavior. In W. Damon & N. Eisenberg (Eds.),Handbook of child psychology: Vol.

3.Social, emotional, and personality development(pp. 779 – 862). New York, NY: Wiley.

Comer, J. S., Chow, C., Chan, P. T., Cooper-Vince, C., & Wilson, L. A. S.

(2013). Psychosocial treatment efficacy for disruptive behavior prob- lems in very young children: A meta-analytic examination.Journal of the American Academy of Child & Adolescent Psychiatry, 52,26 –36.

http://dx.doi.org/10.1016/j.jaac.2012.10.001 Crick, N. R., & Grotpeter, J. K. (1995). Relational aggression, gender, and social-psychological adjustment.Child Development, 66,710 –722.

http://dx.doi.org/10.2307/1131945 Dadds, M. R., Cauchi, A. J., Wimalaweera, S., Hawes, D. J., & Brennan, J.

(2012). Outcomes, moderators, and mediators of empathic-emotion recog- nition training for complex conduct problems in childhood.Psychiatry Research, 199,201–207.http://dx.doi.org/10.1016/j.psychres.2012.04.033 Datyner, A., Kimonis, E. R., Hunt, E., & Armstrong, K. (2016). Using a novel emotional skills module to enhance empathic responding for a child with conduct disorder with limited prosocial emotions.Clinical Case Studies, 15,35–52.http://dx.doi.org/10.1177/1534650115588978 Dodge, K. A., & Coie, J. D. (1987). Social-information-processing factors in reactive and proactive aggression in children’s peer groups.Journal of Personality and Social Psychology, 53,1146 –1158.http://dx.doi.org/10 .1037/0022-3514.53.6.1146 Eisenberg, N., Valiente, C., Spinrad, T. L., Cumberland, A., Liew, J., Reiser, M.,...Losoya, S. H. (2009). Longitudinal relations of children’s effortful control, impulsivity, and negative emotionality to their exter- nalizing, internalizing, and co-occurring behavior problems.Develop- mental Psychology, 45,988 –1008.http://dx.doi.org/10.1037/a0016213 Fanti, K. A. (2016). Understanding heterogeneity in conduct disorder: A review of psychophysiological studies.Neuroscience and Biobehavioral Reviews.http://dx.doi.org/10.1016/j.neubiorev.2016.09.022 Fanti, K. A., & Munoz Centifanti, L. C. (2014). Childhood callous- unemotional traits moderate the relation between parenting distress and This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 1239 HETEROGENEITY IN ANTISOCIAL BEHAVIOR conduct problems over time.Child Psychiatry and Human Development, 45,173–184.http://dx.doi.org/10.1007/s10578-013-0389-3 Fanti, K. A., Demetriou, C. A., & Kimonis, E. R. (2013). Variants of callous-unemotional conduct problems in a community sample of ado- lescents.Journal of Youth and Adolescence, 42,964 –979.http://dx.doi .org/10.1007/s10964-013-9958-9 Fanti, K. A., & Henrich, C. C. (2010). Trajectories of pure and co- occurring internalizing and externalizing problems from age 2 to age 12:

Findings from the National Institute of Child Health and Human Devel- opment Study of Early Child Care.Developmental Psychology, 46, 1159 –1175.http://dx.doi.org/10.1037/a0020659 Fanti, K. A., Panayiotou, G., & Fanti, S. (2013). Associating parental to child psychological symptoms: Investigating a transactional model of development.Journal of Emotional and Behavioral Disorders, 21,193– 210.http://dx.doi.org/10.1177/1063426611432171 Fanti, K. A., Panayiotou, G., Lazarou, C., Michael, R., & Georgiou, G.

(2016). The better of two evils? Evidence that children exhibiting continuous conduct problems high or low on callous-unemotional traits score on opposite directions on physiological and behavioral measures of fear.Development and Psychopathology, 28,185–198.http://dx.doi .org/10.1017/S0954579415000371 Frick, P. J., Barry, C. T., & Kamphaus, R. W. (2010).Clinical Assessment of Child and Adolescent Personality and Behavior(Vol. 3). New York, NY: Springer.

Frick, P. J., & Morris, A. S. (2004). Temperament and developmental pathways to conduct problems.Journal of Clinical Child and Adolescent Psychology, 33,54 – 68.http://dx.doi.org/10.1207/S15374424JCCP 3301_6 Frick, P. J., Ray, J. V., Thornton, L. C., & Kahn, R. E. (2014). Can callous-unemotional traits enhance the understanding, diagnosis, and treatment of serious conduct problems in children and adolescents? A comprehensive review.Psychological Bulletin, 140,1–57.http://dx.doi .org/10.1037/a0033076 Gresham, F. M., & Elliott, S. N. (1990).Social Skills Rating System (SSRS). Circle Pines, MN: American Guidance Service.

Hawes, D. J., Brennan, J., & Dadds, M. R. (2009). Cortisol, callous- unemotional traits, and pathways to antisocial behavior.Current Opin- ion in Psychiatry, 22,357–362.http://dx.doi.org/10.1097/YCO .0b013e32832bfa6d Hawes, D. J., Price, M. J., & Dadds, M. R. (2014). Callous-unemotional traits and the treatment of conduct problems in childhood and adoles- cence: A comprehensive review.Clinical Child and Family Psychology Review, 17,248 –267.http://dx.doi.org/10.1007/s10567-014-0167-1 Hicks, B. M., Markon, K. E., Patrick, C. J., Krueger, R. F., & Newman, J. P. (2004). Identifying psychopathy subtypes on the basis of person- ality structure.Psychological Assessment, 16,276 –288.http://dx.doi .org/10.1037/1040-3590.16.3.276 Holi, M., Auvinen-Lintunen, L., Lindberg, N., Tani, P., & Virkkunen, M.

(2006). Inverse correlation between severity of psychopathic traits and serum cortisol levels in young adult violent male offenders.Psychopa- thology, 39,102–104.http://dx.doi.org/10.1159/000091021 Hyde, L. W., Shaw, D. S., Gardner, F., Cheong, J., Dishion, T. J., & Wilson, M. (2013). Dimensions of callousness in early childhood: Links to problem behavior and family intervention effectiveness.Development and Psychopathology, 25,347–363.http://dx.doi.org/10.1017/ S0954579412001101 Ingoldsby, E. M., Shaw, D. S., Winslow, E., Schonberg, M., Gilliom, M., & Criss, M. M. (2006). Neighborhood disadvantage, parent-child con- flict, neighborhood peer relationships, and early antisocial behavior problem trajectories.Journal of Abnormal Child Psychology, 34,293– 319.http://dx.doi.org/10.1007/s10802-006-9026-y Kahn, R. E., Frick, P. J., Youngstrom, E. A., Kogos Youngstrom, J., Feeny, N. C., & Findling, R. L. (2013). Distinguishing primary and secondary variants of callous-unemotional traits among adolescents in a clinic-referred sample.Psychological Assessment, 25,966 –978.http://dx.doi .org/10.1037/a0032880 Karpman, B. (1941). On the need of separating psychopathy into two distinct clinical types: The symptomatic and the idiopathic.Journal of Criminal Psychopathology, 3,112–137.

Keiley, M. K., Lofthouse, N., Bates, J. E., Dodge, K. A., & Pettit, G. S. (2003). Differential risks of covarying and pure components in mother and teacher reports of externalizing and internalizing behavior across ages 5 to 14.Journal of Abnormal Child Psychology, 31,267–283.

http://dx.doi.org/10.1023/A:1023277413027 Kimonis, E. R., Fanti, K. A., Anastassiou-Hadjicharalambous, X., Mertan, B., Goulter, N., & Katsimicha, E. (2016). Can callous-unemotional traits be reliably measured in preschoolers?Journal of Abnormal Child Psy- chology, 44,625– 638.

Kimonis, E. R., Fanti, K. A., Goulter, N., & Hall, J. (2016). Affective startle potentiation differentiates primary and secondary variants of juvenile psychopathy.Development and Psychopathology. Advance on- line publication.http://dx.doi.org/10.1017/S0954579416001206 Kimonis, E. R., Frick, P. J., Cauffman, E., Goldweber, A., & Skeem, J.

(2012). Primary and secondary variants of juvenile psychopathy differ in emotional processing.Development and Psychopathology, 24,1091– 1103.http://dx.doi.org/10.1017/S0954579412000557 Kimonis, E. R., Skeem, J. L., Cauffman, E., & Dmitrieva, J. (2011). Are secondary variants of juvenile psychopathy more reactively violent and less psychosocially mature than primary variants?Law and Human Behavior, 35,381–391.http://dx.doi.org/10.1007/s10979-010-9243-3 Kochanska, G., Kim, S., Boldt, L. J., & Yoon, J. E. (2013). Children’s callous-unemotional traits moderate links between their positive rela- tionships with parents at preschool age and externalizing behavior prob- lems at early school age.Journal of Child Psychology and Psychiatry, 54,1251–1260.http://dx.doi.org/10.1111/jcpp.12084 Leadbeater, B., Thompson, K., & Gruppuso, V. (2012). Co-occurring trajectories of symptoms of anxiety, depression, and oppositional defi- ance from adolescence to young adulthood.Journal of Clinical Child and Adolescent Psychology, 41,719 –730.http://dx.doi.org/10.1080/ 15374416.2012.694608 Loney, B. R., Butler, M. A., Lima, E. N., Counts, C. A., & Eckel, L. A.

(2006). The relation between salivary cortisol, callous-unemotional traits, and conduct problems in an adolescent non-referred sample.

Journal of Child Psychology and Psychiatry, 47,30 –36.http://dx.doi .org/10.1111/j.1469-7610.2005.01444.x Luby, J., Lenze, S., & Tillman, R. (2012). A novel early intervention for preschool depression: Findings from a pilot randomized controlled trial.

Journal of Child Psychology and Psychiatry, 53,313–322.http://dx.doi .org/10.1111/j.1469-7610.2011.02483.x Lytton, H. (1990). Child and parent effects in boys’ conduct disorder: A reinterpretation.Developmental Psychology, 26,683– 697.http://dx.doi .org/10.1037/0012-1649.26.5.683 Marmorstein, N. R., & Iacono, W. G. (2004). Major depression and conduct disorder in youth: Associations with parental psychopathology and parent-child conflict.Journal of Child Psychology and Psychiatry, 45,377–386.http://dx.doi.org/10.1111/j.1469-7610.2004.00228.x Marsh, R., Gerber, A. J., & Peterson, B. S. (2008). Neuroimaging studies of normal brain development and their relevance for understanding childhood neuropsychiatric disorders.Journal of the American Academy of Child & Adolescent Psychiatry, 47,1233–1251.http://dx.doi.org/10 .1097/CHI.0b013e318185e703 Masi, G., Manfredi, A., Milone, A., Muratori, P., Polidori, L., Ruglioni, L., & Muratori, F. (2011). Predictors of nonresponse to psychosocial treat- ment in children and adolescents with disruptive behavior disorders.

Journal of Child and Adolescent Psychopharmacology, 21,51–55.

http://dx.doi.org/10.1089/cap.2010.0039 Masten, A. S., Roisman, G. I., Long, J. D., Burt, K. B., Obradovi´ c, J., Riley, J. R.,...Tellegen, A. (2005). Developmental cascades: Linking This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 1240 FANTI AND KIMONIS academic achievement and externalizing and internalizing symptoms over 20 years.Developmental Psychology, 41,733–746.http://dx.doi .org/10.1037/0012-1649.41.5.733 McLeod, J. D., & Shanahan, M. J. (1996). Trajectories of poverty and children’s mental health.Journal of Health and Social Behavior, 37, 207–220.http://dx.doi.org/10.2307/2137292 Medoff-Cooper, B., Carey, W. B., & McDevitt, S. C. (1993). The Early Infancy Temperament Questionnaire.Journal of Developmental and Behavioral Pediatrics, 14,230 –235.http://dx.doi.org/10.1097/ 00004703-199308010-00004 Moffitt, T. E. (1993). Adolescence-limited and life-course-persistent anti- social behavior: A developmental taxonomy.Psychological Review, 100,674 –701.http://dx.doi.org/10.1037/0033-295X.100.4.674 Murray, L., Arteche, A., Fearon, P., Halligan, S., Goodyer, I., & Cooper, P. (2011). Maternal postnatal depression and the development of de- pression in offspring up to 16 years of age.Journal of the American Academy of Child & Adolescent Psychiatry, 50,460 – 470.http://dx.doi .org/10.1016/j.jaac.2011.02.001 Muthén, L. K., & Muthén, B. O. (2010).Mplus user’s guide(6th ed.). Los Angeles, CA: Author.

Newman, D. L., Moffitt, T. E., Caspi, A., & Silva, P. A. (1998). Comorbid mental disorders: Implications for treatment and sample selection.Jour- nal of Abnormal Psychology, 107,305–311.http://dx.doi.org/10.1037/ 0021-843X.107.2.305 Nylund, K. L., Asparouhov, T., & Muthén, B. (2007). Deciding on the number of classes in latent class analysis and growth mixture modeling.

A Monte Carlo simulation study.Structural Equation Modeling, 14, 535–569.http://dx.doi.org/10.1080/10705510701575396 Ortiz, J., & Raine, A. (2004). Heart rate level and antisocial behavior in children and adolescents: A meta-analysis.Journal of the American Academy of Child & Adolescent Psychiatry, 43,154 –162.http://dx.doi .org/10.1097/00004583-200402000-00010 Pianta, R. C. (1992).Child-Parent Relationship Scale(Unpublished mea- sure). University of Virginia, Charlottesville, VA.

Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general population.Applied Psychological Measurement, 1,385– 401.http://dx.doi.org/10.1177/014662167700100306 Raine, A., Venables, P. H., & Mednick, S. A. (1997). Low resting heart rate at age 3 years predisposes to aggression at age 11 years: Evidence from the Mauritius Child Health Project.Journal of the American Academy of Child & Adolescent Psychiatry, 36,1457–1464.http://dx.doi.org/10 .1097/00004583-199710000-00029 Roisman, G. I., Susman, E., Barnett-Walker, K., Booth-LaForce, C., Owen, M. T., Belsky, J.,...theNICHD Early Child Care Research Network.

(2009). Early family and child-care antecedents of awakening cortisol levels in adolescence.Child Development, 80,907–920.http://dx.doi .org/10.1111/j.1467-8624.2009.01305.x Rosan, A., Frick, P. J., Gottlieb, K. A., & Fasicaru, L. (2015). Callous- unemotional traits and anxiety in a sample of detained adolescents in Romania.Journal of Evidence-Based Psychotherapies, 15,79 –95.

Ruttle, P. L., Shirtcliff, E. A., Serbin, L. A., Ben-Dat Fisher, D., Stack, D. M., & Schwartzman, A. E. (2011). Disentangling psychobiological mechanisms underlying internalizing and externalizing behaviors in youth: Longitudinal and concurrent associations with cortisol.Hormones and Behavior, 59, 123–132.http://dx.doi.org/10.1016/j.yhbeh.2010.10.015 Schoorl, J., Van Rijn, S., De Wied, M., Van Goozen, S. H. M., & Swaab, H. (2016). Variability in emotional/behavioral problems in boys withoppositional defiant disorder or conduct disorder: The role of arousal.

European Child & Adolescent Psychiatry, 25,821– 830.http://dx.doi .org/10.1007/s00787-015-0790-5 Stadler, C., Kroeger, A., Weyers, P., Grasmann, D., Horschinek, M., Freitag, C., & Clement, H.-W. (2011). Cortisol reactivity in boys with attention-deficit/hyperactivity disorder and disruptive behavior prob- lems: The impact of callous unemotional traits.Psychiatry Research, 187,204 –209. http://dx.doi.org/10.1016/j.psychres.2010.05.004 Steinberg, L., Albert, D., Cauffman, E., Banich, M., Graham, S., & Woolard, J. (2008). Age differences in sensation seeking and impulsivity as indexed by behavior and self-report: Evidence for a dual systems model.Developmental Psychology, 44,1764 –1778.http://dx.doi.org/10 .1037/a0012955 Susman, E. J. (2006). Psychobiology of persistent antisocial behavior:

Stress, early vulnerabilities and the attenuation hypothesis.Neuroscience and Biobehavioral Reviews, 30,376 –389.http://dx.doi.org/10.1016/j .neubiorev.2005.08.002 Tarullo, A. R., & Gunnar, M. R. (2006). Child maltreatment and the developing HPA axis.Hormones and Behavior, 50,632– 639.http://dx .doi.org/10.1016/j.yhbeh.2006.06.010 The NICHD Early Child Care Research Network. (1998). Early child care and self-control, compliance, and problem behavior at twenty-four and thirty-six months.Child Development, 69,1145–1170.

Vaughn, M. G., Edens, J. F., Howard, M. O., & Smith, S. T. (2009). An investigation of primary and secondary psychopathy in a statewide sample of incarcerated youth.Youth Violence and Juvenile Justice, 7, 172–188.http://dx.doi.org/10.1177/1541204009333792 Waller, R., Gardner, F., Viding, E., Shaw, D. S., Dishion, T. J., Wilson, M. N., & Hyde, L. W. (2014). Bidirectional associations between pa- rental warmth, callous unemotional behavior, and behavior problems in high-risk preschoolers.Journal of Abnormal Child Psychology, 42, 1275–1285.http://dx.doi.org/10.1007/s10802-014-9871-z Webster-Stratton, C., & Hammond, M. (1997). Treating children with early-onset conduct problems: A comparison of child and parent training interventions.Journal of Consulting and Clinical Psychology, 65,93– 109.http://dx.doi.org/10.1037/0022-006X.65.1.93 Weinberger, D. A., & Schwartz, G. E. (1990). Distress and restraint as superordinate dimensions of self-reported adjustment: A typological perspective.Journal of Personality, 58,381– 417.http://dx.doi.org/10 .1111/j.1467-6494.1990.tb00235.x Willoughby, M. T., Mills-Koonce, W. R., Gottfredson, N. C., & Wagner, N. (2014). Measuring callous unemotional behaviors in early childhood:

Factor structure and the prediction of stable aggression in middle child- hood.Journal of Psychopathology and Behavioral Assessment, 36,30 – 42.http://dx.doi.org/10.1007/s10862-013-9379-9 Willoughby, M. T., Waschbusch, D. A., Moore, G. A., & Propper, C. B.

(2011). Using the ASEBA to screen for callous unemotional traits in early childhood: Factor structure, temporal stability, and utility.Journal of Psychopathology and Behavioral Assessment, 33,19 –30.http://dx .doi.org/10.1007/s10862-010-9195-4 Woodcock, R. W., Johnson, M. B., & Mather, N. (1990).Woodcock- Johnson Psycho-Educational Battery—Revised. Allen, TX: DLM Teaching Resources. Received April 15, 2016 Revision received December 9, 2016 Accepted January 23, 2017 This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 1241 HETEROGENEITY IN ANTISOCIAL BEHAVIOR