Antisocial personality disorder

Psychopathy is a devastating problem for society, especially when it appears in individuals with repertoires of antisocial behavior (Hart & Hare, 1997). Because psychopathy is investigated almost exclusively in adulthood, little is known about the potential childhood precursors of this constellation of personality traits. In this study, we use the California Child Q-Set (CCQ; Block, 1961), a well-investigated ipsative measure of personality traits, to measure psychopathic traits in childhood, allowing examination of the concurrent and prospective correlates of CCQ psychopathy dimension scores in a clinical and comparison sample of boys. Studying childhood correlates of psychopathic traits is essential for mental health professionals to identify children at risk for serious antisocial behavior and perhaps use appropriate early intervention strategies.

In his 1941 book The Mask of Sanity, Cleckley was one of the first to describe psychopathy, noting that personality traits such as guiltlessness, incapacity for love, impulsivity, and emotional shallowness comprise a serious antisocial syndrome (Cleckley, 1941). Today psychopathy continues to be viewed largely as an adult personality construct, meaning that it is a constellation of stable, inflexible, and maladaptive personality characteristics that can exist in people 18 years of age or older. Psychopathy is characterized by personality traits such as glibness, superficial charm, egocentricity, manipulativeness, and callousness as well as shallow affect, impulsivity, and poor response to punishment (Hart & Hare, 1997). Adults displaying psychopathy comprise a small subgroup of offenders, those who exhibit persistent, frequent, varied, severe, and violent conduct problems that often contribute to a lifestyle of crime. Research indicates that 5-6% of adult offenders, typically those that meet criteria for psychopathy, are responsible for 50-60% of known crimes (Lynam, 1998). The inflated pattern of illicit behavior for this subgroup is attributed to a lack of empathy as well as an underdeveloped fear of punishment, which when present may help to inhibit serious criminal behavior (Hare, Forth, & Strachan, 1992). Ironically, psychopathy can be reliably diagnosed but ineffectively treated in adulthood, as evidenced by the high proportion of incarcerated individuals with psychopathy who are repeat offenders (Hare et al., 1992; Hart & Hare, 1997).

The fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV; American Psychiatric Association [APA], 1994) contains the category of antisocial personality disorder (APD), which is based largely on behavioral criteria intermixed with several affective and interpersonal personality characteristics in a single criterion set. Yet many researchers conceptualize adult antisocial behavior as comprising two separable dimensions: affective/interpersonal and behavioral. The prototype for such a dual-factor structure is Hare's well-validated measure of adult psychopathy, the Psychopathy Checklist--Revised (PCL-R), for which factor 1 reflects the affective and interpersonal features of psychopathy and factor 2 reflects antisocial behavior per se (Hare, 1991). This two-factor structure implies that antisocial adults can be high on factor 1 only (i.e., "white collar" criminals), high on factor 2 only (i.e., "garden variety" antisocial types like petty thieves), or high on both factors (i.e., psychopathic criminals). Because the DSM-IV criteria for APD resemble the contents of factor 2 of the PCL-R (with a mixture of some facets of factor 1), the APD diagnosis appears to be a broad diagnostic category that does not discriminate between potentially important subtypes of criminals.

Despite general consensus that psychopathic characteristics would begin to manifest themselves at an early age and blossom into full-blown psychopathy during adulthood, such continuity is largely assumed (Forth & Mailloux, 2000). In fact, psychopathy is correlated with an early onset of antisocial behavior, but this association is derived primarily from follow-back studies, not from more accurate prospective longitudinal studies (Forth & Mailloux, 2000). In fact, little empirical research has focused on the identification and development of psychopathy in childhood and adolescence because all Axis II designations are reserved for adults (APA, 1994). The interest in the developmental aspects of psychopathy has recently intensified (e.g., Frick & Ellis, 1999; Lynam, 1996), but the presence and development of psychopathic traits in children remain poorly understood.

Several review articles note a retrospective link between adult psychopathy and childhood conduct problems marked by such features as early onset, frequent and versatile offending, stable and impulsive violence, and impulsivity (Hart & Hare, 1997; Lilienfeld & Waldman, 1990; Lynam, 1996). However, research clearly shows multifinality with respect to severe antisocial behavior, meaning that not all severe childhood conduct problems develop into full-blown adult psychopathy (Hinshaw & Lee, 2003). Antisocial youths' developmental trajectories differ with regard to factors such as the age of onset of antisocial behavior, influences of parents, patterns of offending, levels of impulsivity, types of aggression, rates of desistance, and the influence of peers, among other variables. To help explain the heterogeneity of conduct problems in children and to predict adult psychopathy, it is essential to distinguish more homogenous subtypes of antisocial children.

One means of subtyping antisocial youths is based on patterns of comorbidity among disruptive disorders. Compared to children with pure attention-deficit/hyperactivity disorder (ADHD), pure conduct disorder (CD), and diagnosis-free controls, the subgroup with the comorbid presence of ADHD and CD (ADHD-CD) display earlyonset, frequent, severe, cross-situational, and versatile forms of externalizing behavior ranging from aggression to stealing to substance use (Hinshaw, 1994; Loeber, Brinthaupt, & Green 1990; Lynam, 1996). This broad pattern of antisocial behavior resembles the antisocial behavior of adults with psychopathy. Children with ADHD-CD also show response patterns on laboratory measures that suggest lower autonomic reactivity, problems inhibiting a dominant response, and deficits in executive functioning that resemble the neurocognitive and psychophysiological functioning of adult psychopathy (Lilienfeld & Waldman, 1990; Lynam, 1996). In a large community sample, Lynam (1998) demonstrated that boys with ADHD-CD had more psychopathic characteristics, aggressive behavior, disinhibition, neuropsychological impairment, and higher rates and versatility of delinquent acts than did boys with ADHD and CD alone. It appears that ADHD children with conduct problems are at risk for adult psychopathy.

Although ADHD-CD comorbidity may constitute a precursor of adult psychopathy, many children with ADHD-CD do not have psychopathic outcomes (Frick & Ellis, 1999; Hinshaw, 1994), so further subtyping is required to accurately predict adult psychopathy. Frick and colleagues contend that only a subset of children with ADHD-CD--those who also show a constellation of emotional and behavioral responding termed callous/unemotional--are likely to develop psychopathy. Using the Antisocial Process Screening Device (APSD; formerly known as the Psychopathy Screening Device; Frick & Hare, 2002), a psychopathic trait rating scale for children based on the PCL-R with a similar two-factor structure (Frick, Bodin, & Christopher, 2000), this investigative team has associated high levels of callous/unemotional traits with adult psychopathic characteristics such as thrill seeking, reward-dominated response style, low levels of anxiety, and no distress about behavior problems (Barry et al., 2000). Other correlates include childhood onset behavior problems and adjudication (Silverthorn, Frick, & Reynolds, 2001), high rates and a great variety of CD symptoms and police contacts (Christian, Frick, Hill, Tyler, & Frazer, 1997), deficits in processing emotional expressions (Blair, Colledge, Murray, & Mitchell, 2001), a preference for thrill and adventure seeking (Frick, O'Brian, Wootton, & McBurnett, 1994), associations with juvenile sex offending (Caputo, Frick, & Brodsky, 1999), and a greater sensitivity to rewards than punishments (O'Brian & Frick, 1996). Furthermore, children with callous/unemotional traits appear to show an imperviousness to typical parental discipline patterns, suggestive of lowered responsiveness to socialization in this subgroup (Wootton, Frick, Shelton, & Silverthorn, 1997). Overall, a subgroup of conduct problem children with high levels of callous/unemotional traits appear to have the worst forms of antisocial behavior as well as other deficits that resemble those of adult psychopathy.

Crucially, however, prospective investigations are rare in this area, and extant investigations have begun with adolescent rather than child samples with brief follow-up periods (e.g., Vitacco, Neuman, Robertson, & Durrant, 2002). In the sole prospective longitudinal study beginning in childhood, Frick, Cornell, Barry, Bodin, and Dane (2003) used the APSD to show that children with high levels of callous/unemotional traits had higher levels of conduct problems and delinquency severity at a 1-year follow-up than did a control group.

Peer relationship problems during childhood, particularly peer rejection, often have negative and lasting impact on social and emotional adjustment (Parker & Asher, 1987). Children with ADHD and conduct problems have especially high rates of peer rejection (Erhardt & Hinshaw, 1994). No investigations, however, have examined the peer relationships of children or adolescents with psychopathic traits. This study marks the first empirical work to our knowledge that examines peer sociometric correlates for children with psychopathic traits.

In sum, there is a need to further explore the role of psychopathic traits in childhood. We provide an alternative means of appraising such traits. Because specific childhood psychopathy measures did not exist at the time of our data collection, we developed a post hoc dimensional measure of psychopathic traits from the CCQ, a broadband measure of childhood personality that allows researchers to map personality traits onto antisocial behavior constellations (e.g., John, Caspi, Robins, Moffit, & Stouthamer-Loeber, 1994). Also, Reise and Wink (1995) used a Q-sort methodology to measure psychopathy in adults, showing good reliabilities and preliminary convergent and divergent validity with other personality constructs.

Our overall purpose is to ascertain the behavioral and social correlates of psychopathic traits in childhood, both concurrently and prospectively. Concurrent criterion measures include diagnosable childhood externalizing disorders and symptoms as well as objective measures of non-compliance, overt aggression, and covert antisocial behavior. Also, in an exploratory analysis, we examine the relationship between the CCQ psychopathy dimension and peer social status, predicting positive correlations between CCQ psychopathy dimension scores and sociometric indices of peer rejection and negative correlations with peer popularity. Finally, we examine the prospective correlates of the CCQ psychopathy dimension by predicting adolescent delinquency severity from childhood CCQ psychopathy dimension scores at a 5-7-year prospective follow-up assessment.

Although our sample of clinical and comparison children was originally recruited for a general study of the developmental psychopathology of ADHD and not psychopathic traits per se, we believe that the participants are potentially enriched in the display of such traits because ADHD and conduct problems are thought to be risk factors for adult psychopathy (Lynam, 1996). Thus, this sample provides an opportunity to study a construct with a low base-rate in the general population, and the results may help to illuminate the child and adolescent manifestations of psychopathic traits.

METHOD

Overview of Data Collection

The baseline data for this research project were collected during summer research camps for boys with ADHD and comparison boys held in 1991, 1994, and 1995. These 6-week enrichment programs, free to participants, were funded by a grant from the National Institute of Mental Health. Previously unfamiliar boys were mixed together by clinical and comparison status and divided into two groups by age (6-9.5 and 9.5-12 years old), with each age group participating together in classroom, playground, and small group activities. Large group activities were observed in vivo daily, yielding measures of non-compliance and overt aggression. Laboratory measures of covert antisocial behavior as well as sociometric interviews were obtained during the last week of each summer program. Following each program, CCQ personality measures were gathered for each child from staff members who worked extensively with each boy yet were unaware of diagnostic status. Between 5 and 7 years following each program, staff contacted families of participants regarding a prospective follow-up assessment in which data relevant to antisocial behavior and delinquency were collected. Although a number of studies have been published on this sample of boys (e.g., Hinshaw & Melnick, 1995; Hinshaw, Simmel, & Heller, 1995; Hinshaw, Zupan, Simmel, Nigg, & Melnick, 1997; Lee & Hinshaw, in press), no previous studies have examined psychopathic traits in these children. (Note that this investigation does not include a 1993 cohort for which CCQ ratings were not performed by staff because of funding limitations.)

Participants with ADHD underwent double blind, placebo-controlled trials of methylphenidate (Nigg, Hinshaw, & Halperin, 1996). Observational data for this study come from unmedicated or placebo periods only.

Participants

The sample for this study consists of 122 boys (66 ADHD, 56 comparison), aged 6-12 years. Boys with ADHD were recruited through physician and mental health referrals and via parent and self-help groups. Initial inclusion criteria included a community diagnosis of ADHD and treatment with stimulant medication for at least 4 months prior to the beginning of the summer camp. Exclusionary criteria for all participants included IQ scores of less than 70, sensory impairment, neurological disorder, psychosis, and medical conditions that would prevent active participation in the summer camp program. Comparison children were recruited through newspaper advertisements or announcements placed in local service agencies or schools. Initially eligible boys and their parents underwent the same empirically based assessment as the children with ADHD; parent or teacher scores that were consistently in the clinical range served as exclusionary criteria. Subclinical scores were not considered exclusionary in order to avoid problems associated with a "super-normal" comparison group. Because of the many referral sources, comparative analyses of boys recruited from different referral sources were underpowered to detect significant differences.

For the boys with ADHD, inclusion criteria consisted of surpassing established cutoff scores for hyperactivity or ADHD on at least three of the following four measures: (1) the Conners Abbreviated Symptom Questionnaire (CASQ; Goyette, Conners, & Ulrich, 1978); (2) the Attention Problem Scale of Achenbach's Child Behavior Checklist (CBCL; Achenbach, 1991); (3) structured interviews with parents using an abridged version of the Diagnostic Interview for Children and Adolescents (DICA; Welner, Reich, Herjanic, & Jung, 1987); and (4) symptoms of ADHD from the Disruptive Behavior Disorders Checklist (DBD; Pelham, Gnagy, Greenslade, & Milich, 1992). Parents also reported that symptoms or impairment emerged at or before 6 years of age. The DBD and the Swanson, Nolan, and Pelham Checklist (SNAP; Swanson, 1992) also yielded diagnoses of ODD. All diagnoses were based on the criteria from the revised third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R; APA, 1987).

Table I shows the baseline sample's continuous demographic variables. The total sample mean age was 9.3 years with no significant difference between ADHD and comparison groups. Boys with ADHD scored significantly lower than comparison boys on Verbal IQ and reading achievement as measured by the Wechsler Intelligence Scale for Children--Third Edition (WISC-III; Wechsler, 1991) and the Woodcock-Johnson Psychoeducational Battery--Revised (Woodcock & Johnson, 1990), respectively. Comparison children had lower family incomes than boys with ADHD, but the groups did not differ on the Hollingshead measure of socioeconomic status (Hollingshead, 1975). As expected from inclusion criteria, the groups differed with respect to precamp maternal ratings of externalizing behavior, with ADHD boys' mean behaviors in the clinical range, and comparison boys' behavioral scores nearly identical to national norms.

The total sample across summer camps was diverse with regard to ethnicity: 57.4% were Caucasian, 15.6% were African American, 9.0% were Latino, 14.8% were Asian, and 3.3% were Native American. There were diagnostic group differences with regard to ethnicity, such that more comparison boys were non-Caucasian than were boys with ADHD ([chi square] = 16.72, p < .001). Of the total sample, 21% lived with a single parent and 79% lived with two parents. The diagnostic groups did not differ significantly with regard to living in two-parent versus one-parent households or adoptive households.

Baseline Measures

ADHD and ODD

ADHD and ODD diagnostic protocols are discussed above.

CD-Like Symptoms

This sample was selected for ADHD status, not CD status, and we did not collect information regarding CD diagnoses given the young age of many of the participants. Consequently, we used mother-reported externalizing scores on the CBCL as a measure of CD-like behaviors. Because the CBCL is considered to be a reliable and valid measure of CD-like symptoms (Keenan & Wakschlag, 2000; Lowe, 1998), we consider CBCL externalizing scores to be an approximation of CD status. The CBCL is a standardized and widely used behavior rating scale that contains 113 items that are rated on a 0-2 metric (Achenbach, 1991). CBCL externalizing scores are derived from the sum of the Aggressive Behavior subscale, which assesses a wide variety of conduct problems ranging from opposition to noncompliance to overt aggression, and the Delinquency subscale, which measures a variety of covert and illegal acts. Mother-reported CBCL scores were obtained prior to the summer camps.

Psychopathic Traits

The boys' personalities were measured with the California Child Q-Set (CCQ; Block, 1961), an ipsative method for capturing personality characteristics for quantitative comparisons. Shortly after each camp, three counselors who were very familiar with a classroom of children, but who were blind to diagnostic status, dimensional measures of externalizing behavior, and objective observations of behaviors, performed a CCQ-sort for each child. The counselors forced 100 personality traits into nine equal piles representing a continuum of salient traits with the middle pile (pile 5) having 12 entries and all other piles having 11 entries each. Pile 9, the high anchor point, represented the most descriptive traits, while pile 1, the low anchor point, represented the least descriptive traits for each child. This Q-Sort process took each counselor approximately 15-30 min per child. Interrater reliability was established using an alpha coefficient with raters considered "items" (Nunnally, 1970). Across raters, alphas ranged from .50 to .96 (mean = .74). Raters' scores were then averaged to produce a single CCQ profile for each boy.

Subsequently, six national psychopathy experts were asked to perform CCQ-sorts for a prototype of child psychopathy, parallel to prior procedures regarding ADHD and ADD-only prototypes (Hinshaw, Ablon, Kremen, & Zupan, 1997) and similar to Reise and Oliver's (1994) procedure regarding an adult psychopathy prototype. Given the strong expert interrater reliability ([alpha] = .93), we averaged the experts' data to create a composite CCQ psychopathy prototype. Each CCQ profile from the summer camps was then correlated with the expert-derived composite CCQ psychopathy prototype to establish a CCQ psychopathy dimension score for each boy. Theoretically, scores could range from 1.0, the prototype of child psychopathy, to--1.0, the antithesis of child psychopathy.

To prevent the possibility that shared predictor and outcome items would spuriously inflate any associations between baseline CCQ psychopathy dimension scores and adolescent delinquency severity, we recalculated baseline CCQ psychopathy dimension scores after excluding CCQ behavioral items that were similar to items on the outcome measures. For this "decontamination" of the CCQ psychopathy dimension predictor, the authors and an expert in developmental psychopathology selected CCQ items meeting two criteria: (a) items that were behavioral in nature (rather then affective/interpersonal), and (b) items that reiterated CBCL, TRF, YSR, and SRD items that were used to generate the delinquency severity outcome measure. We chose to retain affective items because affective characteristics are thought to be core features of psychopathy that we hoped to capture with the CCQ psychopathy dimension. The experts had 100% agreement as to which items met the above criteria. Specifically, we dropped the following CCQ items: "is obedient and compliant," "teases other children (including siblings)," "is aggressive (physically and verbally)," "is stubborn," and "behaves in a dominating manner with others." We then calculated modified CCQ psychopathy dimension scores without these five items in the identical way that we calculated the original CCQ psychopathy dimension scores.

Noncompliance and Overt Antisocial Behavior Observation

A validated time sampling system, developed to observe peer group activities, yielded externalizing measures of noncompliance and overt aggression (Hinshaw & Renfro, 1999; Hinshaw, Zupan, et al., 1997). Teams of four undergraduate trained observers, blind to diagnostic and medication status, watched 1-hr activities from the periphery and rated children on behavior using rosters of the boys' names in randomized orders. Using a pacing tape that cued 3 s to find a child, 5 s to observe the child, and 3 s to record an observation, naturalistic noncompliance, verbal and physical aggression, and other behavioral frequencies were established for each child. Noncompliant behavior was defined as annoying, rule-violating, and intrusive behavior that fell short of teasing or actual physical contact. Verbal aggression was defined by taunts, swearing, or threats directed toward a peer or adult, and hits, kicks, or shoves defined physical aggression. Because the base rate was low for physical aggression, verbal and physical aggression frequencies were combined to establish an overt aggression score for each child. Scores were calculated by dividing the overall frequency of a behavior by the total number of observation intervals yielding proportions. Comparison boys averaged over 300 observations, but boys with ADHD averaged 100 observations for placebo periods. For all participants, the overall base rate of overt aggression was 2% and noncompliance was 12%.

For some of the ratings, an observer pair unknowingly had identical rosters that allowed interobserver agreement calculations. Behavior was counted as occurring only if the two observers agreed. Across all observer pairs, the occurrence-only agreement for overt aggression was 50% and for noncompliance 65% (see Hinshaw & Melnick, 1995). Although these figures are not overly strong, they are rigorous and conservative occurrence-only statistics that have proved valid in past research (Erhardt & Hinshaw, 1994; Hinshaw, Zupan, et al., 1997).

Covert Antisocial Behavior Laboratory Measure

As first described by Hinshaw, Heller, and McHale (1992), boys participated in a laboratory paradigm that elicited covert antisocial behavior during the final week of camp. Boys completed a boring worksheet in a private room while temptation was present in the form of dollar bills, coins, and desirable small toys. After completing the worksheet, stolen items as well as defaced or destroyed property were recorded as an index of covert antisocial behavior. Hinshaw et al. (1995) extensively validated this measure, demonstrating good test-retest reliability and strong external criterion validity. Covert antisocial behavior and observed overt aggression were modestly associated (r = .23, p < .05), indicating that these two measures are tapping distinct aspects of antisocial behavior.

Peer Sociometrics

During the final week of summer camp, each boy was asked in a private interview to name the three classmates with whom he would most like to be friends, and the three with whom he would least like to be friends. Picture boards were used to facilitate this process. To adjust for cross-class comparisons, positive and negative nomination proportion scores were calculated by dividing the number of nominations received by the number of classmates. All of the boys with ADHD were on active stimulant medication when they gave their nominations. However, the appraisals that each boy received reflect peer impressions from the whole summer, which included an equal number of medicated and nonmedicated days for boys with ADHD. See Hinshaw and Melnick (1995) for details.

Follow-Up Procedures

Five years following each summer program, participants were invited to participate in a follow-up assessment study. One hundred six participants out of 122 total (87%) were involved in some aspect of the follow-up assessments. At follow-up, the age range was 11.3-19.3 years with a mean age of 14.6 years. Attrition occurred despite extensive staff efforts to retain the sample such as sending birthday cards and making regular check-up telephone calls. Of this group of 106, 83 families provided at least one of the four primary antisocial behavior measures that were independently rated for delinquency severity. For the remaining 23 families, who were not able to complete the standard questionnaires and rating scales, in-depth telephone interviews were conducted with the parent(s) by the principal investigator. Because of difficulties in tracking some of the camp participants, some follow-up assessments were not completed until up to 1.5 years following initial contact.

Chi-square and t-test analyses show that the follow-up sample did not differ from the baseline sample with regard to age. Verbal IQ, reading level, or ethnicity. However, the mean income level of the follow-up sample dropped slightly from baseline ([chi square] = 10.43, p < .05). Also, more of the one-parent families chose not to complete follow-up assessments than the two-parent families ([chi square] = 6.11, p < .05).

Follow-Up Measure

Adolescent Delinquency

On the basis of data collected at the follow-up visit, a multiple-source, multiple-method procedure was used to assess delinquency severity (see Lee & Hinshaw, in press; Loeber, Stouthamer-Loeber, Van Kamen, & Farrington, 1991). Achenbach's Teacher Report Form (TRF), Youth Self-Report (YSR), and the CBCL contain narrowband Delinquent Behavior subscales (Achenbach, 1991). T scores were calculated to help gauge clinically significant delinquent behavior. To further assess self-reported delinquency (the preferred method of assessing covert antisocial behavior), the Self-Reported Delinquency (SRD) measure was used (Elliot, Huizinga, & Ageton, 1985). This is a 47-item questionnaire that makes distinctions between overt, covert, destructive, and nondestructive offenses and it assesses the context and frequency of offending. To consolidate these multiple-sources into a single severity dimension of delinquency, Loeber and colleagues' (1991) validated conglomerate delinquency assessment procedure (Loeber et al., 1991) was used. On the basis of CBCL, TRF, YSR, and SRD delinquency scores, trained research assistants assign each boy a severity score ranging from 0 to 3. A score of 0 denotes no delinquency; 1 refers to minor forms of delinquency such as theft valued below five dollars; 2 includes delinquent behaviors outside of the home such as joyriding and fighting; and 3 denotes the most severe forms of delinquency involving interpersonal harm such as physical assault or robbery. A random subsample of boys who completed the follow-up visit (approximately 20% of the total sample) was used to test interrater reliability. Kappas ranged from .65 to .85, indicating very good to excellent agreement.

RESULTS

Initially, we examined variable distributions. Two variables, overt and covert antisocial behavior, had positively skewed distributions. To help transform these variables into more normal distributions for subsequent hierarchal multiple-regression analyses, we used a square root transformation (Osborne, 2002). We also computed Fisher r-to-z transformations to standardize and spread-out CCQ psychopathy dimension scores. Furthermore, the overt antisocial behavior variable had one significant outlier (9 standard deviations above the mean), and we recomputed that score as .5 standard deviations greater than the second largest value, maintaining the rank order of the boys for that variable (see Hinshaw, Carte, Sami, Treuting, & Zupan, 2002).

In the following analyses, we demonstrate the independent effects of psychopathic traits on concurrent and prospective criterion variables while controlling for ADHD diagnoses, ODD symptoms, and CD-like symptoms with hierarchal linear multiple-regression analyses, a technique that tests the incremental associations between independent predictors and criterion variables (Pedhazur, 1982). In the multiple-regression equations for concurrent criteria, ADHD status was entered first, ODD status second, CBCL externalizing scores third, followed by CCQ psychopathy dimension scores. We controlled for ADHD, ODD, and CD-like symptoms because of their correlations with antisocial domains and with CCQ psychopathy dimension scores (Table II), providing a stringent test of psychopathic traits. Note that we did not control for age in the concurrent multiple-regression equations because it did not covary with the predictors or criterion measures.

In contrast, we controlled for baseline age as well as ADHD, ODD, and CD-like symptoms while predicting adolescent delinquency severity because age significantly correlated with delinquency (Table II), entering it as the first step in the regression equations. We performed a very stringent test of the unique contribution of psychopathic traits prediction to adolescent delinquency severity by additionally controlling for baseline levels of observed covert and overt antisocial behavior. Further, we conducted parallel multiple-regression analyses predicting to adolescent delinquency severity with the modified CCQ psychopathy dimension in order to control for possible shared item confounds. Although we used linear multiple-regression equations to document incremental percent of variance, we also conducted parallel logistic multiple-regression analyses to be certain that our findings were not artifacts of linear regression techniques, a method that technically is not used with ordinal outcome variables. Finally, we explored the CCQ psychopathy dimension's individual predictive classifications to the presence or absence of delinquency (see Hinshaw et al., 2002).

CCQ Psychopathy Dimension

Using within-subject, across-item Pearson correlation coefficients, we successfully placed each subject on a CCQ psychopathy dimension. CCQ psychopathy dimension scores were continuously distributed in this sample. CCQ psychopathy dimension scores had a mean of 0, median of -.10, a minimum score of -.70, and a maximum score of 0.81 (Table I). Descriptive statistics also show the salience, means, and standard deviations for the most and least descriptive CCQ items from the expert-derived child psychopathy prototype (Table III).

Externalizing Diagnoses and Symptoms

In terms of "known groups" validation, we performed independent sample t-tests and found that boys with ADHD had higher scores on the CCQ psychopathy dimension than did control boys (Table I). This result held with control of ODD status using an ANCOVA: [F.sub.1,121] = 19.66, p < .001. Additionally, boys with ODD had higher CCQ psychopathy dimension scores than did boys without ODD: t(120) = 4.0, p < .001; Cohen's d = .95. However, with control of ADHD status using an ANCOVA, the association of ODD status with CCQ psychopathy dimension scores was reduced to nonsignificance: [F.sub.1,121] = 3.5, p = .06. We also found that CCQ psychopathy dimension scores were positively and significantly correlated with CBCL externalizing scores (r = .56, p < .01; Table II).

Concurrent Antisocial Behavior

To demonstrate a convergence between children's psychopathic traits and objective indicators of antisocial behavior, we calculated zero-order Pearson correlations between CCQ psychopathy dimension scores and concurrent criterion scores (Table II). Results indicate significant associations between CCQ psychopathy dimension scores, on the one hand, and observed noncompliance (r = .64, p < .001), overt aggression (r = .70, p < .001), and covert antisocial behavior (r = .53, p < .001), on the other. Furthermore, results of three multiple-regression equations established significant associations between CCQ psychopathy dimension scores and these criterion variables with statistical control of externalizing diagnoses and symptoms (Table IV). The proportions of variance accounted for were large: 14.8, 24.6, and 17.8% of incremental variance, respectively. These results were upheld when we also controlled for Verbal IQ in the multiple-regression equations.

Peer Social Status

CCQ psychopathy dimension scores were significantly correlated with negative peer nominations (r = .63, p < .001), but not with positive peer nominations (r = -.18; Table II). As shown in Table V, CCQ psychopathy dimension scores incrementally predicted negative peer nominations when placed in a hierarchal multiple-regression equation with control of externalizing variables, significantly accounting for an additional 17.9% of the negative peer nomination variance.

Adolescent Delinquency Severity

At the 5-7-year follow-up, 39.1% of the boys received a score of 0 (no delinquency), 34.8% received a score of 1 (minor delinquency), 20.9% received a score of 2 (delinquency outside of the home), and 5.2% received a score of 3 (severe delinquency). Pearson correlations between CCQ psychopathy dimension scores and delinquency severity scores indicated a moderate degree of association (r = .62, p < .001; Table II). (Note that the Pearson correlations were virtually identical to Spearman [rho] correlations, which are typically used to correlate ordinal variables; we used Pearson correlations for consistency with other analyses.) When utilizing multiple-regression procedures parallel to those above, we found that the CCQ psychopathy dimension remained a significant predictor of adolescent delinquency severity even with control of childhood age, ADHD, ODD, and CD-like symptoms (Table VI). CCQ psychopathy dimension scores accounted for 12.0% of incremental adolescent delinquency severity. To provide even more stringent controls, we repeated the regression analyses while covarying objective measures of observed antisocial behavior at baseline as well. Table VI indicates that the CCQ psychopathy dimension continued to predict adolescent delinquency severity, accounting for an additional 6.6% of its variance. Note that these results were replicated when we also controlled for Verbal IQ in the multiple-regression equations.

Importantly, we repeated these analyses using modified CCQ psychopathy dimension scores that lacked behavioral items shared by predictor and outcome variables. The results of this extremely stringent test were virtually identical to the analyses with the complete CCQ psychopathy dimension. Modified CCQ psychopathy dimension scores significantly accounted for 11.4% of incremental adolescent delinquency severity variance with control of age, ADHD, ODD, and CD-like symptoms, and 6.1% of adolescent delinquency severity variance with additional control of observed antisocial behaviors.

To test for the presence or absence of delinquency (a dichotomous variable), we conducted parallel logistic multiple-regression analyses. To create a dichotomous outcome variable, we collapsed delinquency severity scores into two categories: scores of 0 and 1 comprised group A (no or minor delinquency inside the home) and scores of 2 and 3 constituted group B (moderate to severe delinquency outside of the home). Out of 106 boys at follow-up, 77 were placed into group A and 29 were placed into group B. Results showed that the CCQ psychopathy dimension had highly significant associations with delinquency membership with control of age, ADHD, ODD, and CD-like symptoms (Wald = 11.65, p = .001, odds ratio = 14.9, 95% confidence interval = 3.2-70.3) and with additional control of observed covert and overt antisocial behavior (Wald = 9.69, p = .002, odds ratio = 34.6, 95% confidence interval = 3.7-322.1).

Finally, to test the accuracy of the CCQ psychopathy dimension's individual classification predictions to the presence or absence of delinquency, we split CCQ psychopathy dimension scores at the median and cross-classified the high and low psychopathic traits groups with the dichotomous delinquency outcome score (Table VII). Results indicate that the CCQ psychopathy dimension has the following psychometric properties in this sample: sensitivity = .69, specificity = .74, positive predictive power = .50, negative predictive power = .86, hit rate = .73. Although the majority of moderate to severe delinquent boys were captured by the CCQ psychopathy dimension, a high number of false positive predictions were evident, totalling 50% of those high on childhood psychopathic traits.

DISCUSSION

To measure psychopathic traits in childhood, we obtained counselor CCQ-sorts of participants in a naturalistic summer camp and correlated these CCQ-sorts with an expert-derived prototype of child psychopathy to yield CCQ psychopathy dimension scores for all participants. These dimensional scores were significantly associated with externalizing diagnoses and symptoms as well as objectively observed indicators of antisocial behavior. Even with control of ADHD, ODD, and CD-like symptoms, CCQ psychopathy dimension scores accounted for substantial variance in observed noncompliance, observed overt and covert antisocial behavior, and peer rejection. In addition, CCQ psychopathy dimension scores predicted the severity of delinquency at a 5-7-year prospective follow-up during adolescence with stringent control of age, externalizing disorders and symptoms, and observed antisocial behavior at baseline. This longitudinal prediction was preserved with use of a modified CCQ psychopathy dimension that eliminated behavioral items from the CCQ that may have been contaminated with the criterion. Note, however, that individual predictions yielded a plethora of false positive designations.

Psychopathic Traits and Externalizing Disorders

Results indicate significant associations between CCQ psychopathy dimension scores and ADHD, ODD, CD-like symptoms. Given the concurrent associations, the cause-effect relationships are unknown. It is possible that associations exist because CCQ psychopathy dimension scores are simply a proxy for severe externalizing disorders. For example, one of the most characteristic items on the CCQ psychopathy prototype is "unable to delay gratification," which is similar to the impulsivity of ADHD. Alternatively, severe externalizing disorders may manifest psychopathic traits creating secondary psychopathy, a condition where psychopathy is not the primary disorder, but the disorder is manifested in psychopathic traits. However, when we controlled for externalizing features in our multiple-regression analyses, the CCQ psychopathy dimension continued to predict antisocial criteria, suggesting that the CCQ psychopathy dimension taps personality traits that are broader than severe externalizing behaviors. As suggested by Frick et al. (1999), the construct of childhood psychopathy may be best represented by both externalizing symptoms and early manifestations of psychopathic traits.

Psychopathic Traits and Antisocial Behavior

We showed strong associations between the CCQ psychopathy dimension, concurrent noncompliance, verbal and physical aggression (overt antisocial behavior), and stealing and property destruction (covert antisocial behavior) even with control of externalizing disorders and symptoms. Apparently, boys aged 6-12 with psychopathic traits share some of adult psychopaths' behavioral profiles. Future studies should examine additional criteria because psychopathic traits in childhood may manifest themselves in other domains besides antisocial behavior.

This study marks one of the first longitudinal investigations of children with psychopathic traits. We found that the CCQ psychopathy dimension predicts the severity of adolescent delinquency as far as 5-7-years in the future even with extensive control of externalizing diagnoses, symptoms, and behaviors. These results are compelling given our multiple-source, multiple-method algorithm to ascertain delinquency severity, the robust sample retention at follow-up (i.e., ample power), and our efforts to show an effect with both ordinal and dichotomous delinquency outcome variables that are clinically interpretable. On the basis of individual predictive classifications, however, it appears that false positive predictions predominate when CCQ psychopathy dimension scores in childhood are used to predict the presence or absence of moderate to severe delinquency.

To establish a purer measure of the affective and interpersonal features of childhood psychopathy without the confounds of behavioral features, we "decontaminated" the CCQ psychopathy dimension by eliminating a few behavioral items that were similar to or were direct precursors of the adolescent delinquency severity measure. Parallel analyses conducted with the modified CCQ psychopathy dimension continued to predict adolescent delinquency severity with almost identical effect sizes. Thus, affective and interpersonal personality traits appear to play a role in the prediction of delinquency.

Psychopathic Traits and Peer Relationships

This study marks the first exploratory examination of psychopathic traits in childhood and peer relationships. We found that the CCQ psychopathy dimension predicted peer rejection above and beyond known diagnostic and behavioral risk-factors. In short, boys with psychopathic traits were very likely to be rejected by peers in this sample. These findings are important because peer rejection is highly predictive of future psychological maladjustment and it is linked to antisocial developmental trajectories (Parker & Asher, 1987). Because the CCQ psychopathy dimension predicts peer rejection better than ADHD status alone, there may be more to peer rejection than the annoyance generated by the hyperactive and impulsive behavior patterns of ADHD. Surprisingly, our results showed a nonsignificant relationship between peer popularity and CCQ psychopathy dimension scores. It may be the case that the presence of psychopathic traits in childhood is essentially unrelated to peer popularity. Additional studies are needed to confirm the sociometric results from this exploratory study.

Developmental Issues

At baseline, age did not relate to CCQ psychopathy dimension scores, suggesting that psychopathic traits are represented relatively equally in boys aged 6-12. We therefore did not control for age in our concurrent predictions. As expected, age did relate to adolescent delinquency severity, indicating that older adolescents exhibit worse delinquency, so we controlled for age in our prospective predictions. Psychopathic traits help to explain antisocial behavior regardless of youths' developmental levels, a finding that counters the dissenting notion that psychopathic traits cannot be measured in childhood because of developmental confounds.

Assessment Implications

The expert-based CCQ psychopathy prototype offers an empirically derived qualitative and quantitative description of psychopathy in childhood. Qualitatively, the CCQ psychopathy prototype offers clinicians a list of personality characteristics to examine in children at-risk for conduct problems. Note that these traits are a mix of affective, interpersonal, and behavioral characteristics. Quantitatively, psychologists can profile individual children with the CCQ and place them on a psychopathy dimension, specifying the magnitude of psychopathic traits within an individual child.

The CCQ psychopathy dimension utilizes trained observers rather than self-reports. Observations of psychopathic traits are thought to be a more valid source of information than self-reports because individuals high on psychopathy often lie, manipulate, minimize, and distort facts (Lilienfeld, 1998). It is unknown how much time is needed to accurately assess a child's personality with the CCQ. Our informants for the boys' CCQ-sorts were summer camp staff who knew each subject for 6 weeks, which is not practical for most clinical or research applications. Also, despite increased objectivity and training, it is unclear whether counselor ratings are more accurate than the reports of other raters such as parents or teachers. Finally, the specificity of our informants' CCQ-sorts is unknown. Our informants may have rated according to halo effects, not differentiating psychopathic traits from severe ADHD. Alternatively, their familiarity with ADHD may have helped them differentiate fine-grained personality traits.

In contrast to other measures of psychopathic traits in childhood such as the Antisocial Process Screening Device (APSD; Frick & Hare, 2002), the CCQ psychopathy dimension conceptualizes psychopathy as a child personality construct as opposed to an extension of an adult phenomenon. Expert generated prototypes of child and adult psychopathy recognize differences between these constructs (Salekin, Rogers, & Machin, 2001). A direct comparison of the most salient CCQ psychopathy prototype items with the APSD items shows that the CCQ yields a slightly different picture of psychopathy. For example, the CCQ reveals a prototypic child with psychopathic traits to have inappropriate emotive behavior, such as overreactions and a propensity for irritation and anger, while the APSD items account for shallow or absent emotionality. Yet the CCQ psychopathy dimension also shares some callous/unemotional traits with the APSD, suggesting that psychopathic traits may be stable from childhood into adulthood.

The wide variety of CCQ traits assures that the CCQ psychopathy dimension does not suffer from construct underrepresentation, an important consideration when measuring the magnitude of psychopathy in children, a construct that has not been defined adequately at present. Moreover, accounting for a variety of traits is important because psychopathic traits may interact with nonpsychopathic traits to generate diverse outcomes. Indeed, unknown constellations of personality traits may act as protective or accentuating factors. In short, the CCQ psychopathy dimension uses all aspects of a child's personality to gauge psychopathy and it accounts for nonpsychopathic traits when used for predictions.

The predictive efficacy of the CCQ psychopathy dimension may be partially related to its dimensional approach to the presence of psychopathic traits in children. In general, many researchers prefer dimensional predictors because dimensions capture the subtle aspects of a construct whereas categorical variables are thought to only truly capture the extremes of a construct (Fergusson & Horwood, 1995). The CCQ psychopathy dimension had a continuous distribution in this sample, so scores falling in the middle of the sample may have helped to predict criterion measures.

On the other hand, using dimensional predictor variables is uncommon because the majority of intervention and assessment efforts rely heavily on categorical DSM-IV diagnoses. It would be useful to create a diagnostic child psychopathy category with the CCQ. However, to do so would require the CCQ psychopathy dimension to be standardized on a large, representative community sample to establish clinical cutoff scores. At this time there are no clinical cutoff scores for the CCQ psychopathy dimension, so the measure offers little information for clinicians to classify a particular boy.

Our data herein establish, in preliminary fashion, external validity for the CCQ psychopathy dimension (Anastasi, 1988). Because there were no other measures of psychopathic traits in childhood at the time of data collection, it was impossible to establish convergent validity with other measures of psychopathic traits. However, the literature suggests strongly that adults with psychopathy engage in severe antisocial behavior in their youth, so we expected that such antisocial behavior patterns are present in children with psychopathic traits. Our significant concurrent and longitudinal predictions confirmed our expectations. Given our valid, objective nature of assessing antisocial behavior (Hinshaw et al., 1995; Hinshaw & Renfro, 1999), the convergent validity is noteworthy.

Regarding the predictive accuracy of the CCQ psychopathy dimension to the presence or absence of adolescent delinquency, we found that over two thirds of the moderately to severely delinquent boys had scored above the median on the CCQ psychopathy dimension during childhood (sensitivity = .69). However, consonant with perspectives emphasizing the multifinality of childhood antisocial behavior (Hinshaw, 1994; Hinshaw & Lee, 2003), the CCQ psychopathy dimension yielded a number of false positive predictions (positive predictive power = .50). Given our uncertain predictions, it is of the utmost importance that others do not use the CCQ to label children as "psychopaths" (Edens, Skeem, Cruise, & Cauffman, 2001).

Future studies must address the factor structure of the CCQ psychopathy dimension. Even though the CCQ is intended to be a measure of personality, its items constitute a blend of behavioral, affective, and interpersonal items. Although we partially distilled personality traits by eliminating behavioral items from the CCQ that showed clear overlap with the delinquency severity criterion measure and by controlling for conduct problems, it is still unclear whether personality related items are making incremental predictions. A factor analysis is required to disentangle the structure of the CCQ. We could not perform such an analysis because of our small sample size in relation to the 100 items that make up the CCQ.

Limitations

First, this study has a small sample, yielding limited statistical power. Second, our use of a clinical sample of boys with ADHD rather than a community sample of boys with ADHD-CD or CD limits generalizability. Our ADHD sample may be less impaired because of their less severe externalizing behaviors compared to possible ADHD-CD or CD samples (Hinshaw & Lee, 2003). Third, our results only speak to the role of psychopathic traits in male antisocial behavior, not female manifestations. Indeed, whether the etiology of antisocial behavior across the sexes is similar or different is not currently known (Hinshaw & Lee, 2003). Fourth, the lack of convergence with other measures of psychopathic traits limits our conclusions about external validity. Fifth, the CCQ psychopathy dimension factor structure is unknown, so it is unclear if affective, interpersonal, or behavioral items are making incremental predictions.

Finally, we could not fully establish whether the CCQ psychopathy dimension converges on diagnostic risk factors for adult psychopathy (i.e., ADHD-CD) because we did not evaluate participants for CD. Our use of CBCL externalizing scores as a proxy for CD has some pitfalls. CD-like symptoms were measured by a single-source and a single-method in contrast to ADHD and ODD diagnoses which were measured by multiple-sources and multiple-methods, a superior methodology (Hinshaw & Zupan, 1997). Also, CBCL externalizing scores are a dimensional variable in contrast to the categorical variables of ADHD and ODD diagnoses. This is problematic because the theory that children with ADHD-CD are at-risk for psychopathy is based on categorical diagnostic status, making direct comparisons impossible. Thus, this study can only partially and preliminarily lend support to Lynam's hypothesis that ADHD-CD boys are "fledgling psychopaths" (Lynam, 1996, 1997, 1998).

We highlight that this study must be interpreted with caution. This study marks the first use of the CCQ psychopathy dimension, the first evaluation of the peer status of boys with psychopathic traits, and one of the first longitudinal studies of preadolescent boys with psychopathic traits. These analyses must be replicated with separate, larger, community samples for verification. Also, we did not account for moderating or mediating variables herein. Most importantly, we reiterate that the CCQ psychopathy dimension should not be used as a clinical instrument at this time given its preliminary external and predictive validity. A false label of "psychopath" is potentially devastating.

Conclusion

We contend that some children manifest psychopathic traits at an early age, and the magnitude of these traits may be an important predictor to later developmental problems. It remains to be seen whether childhood psychopathic traits and associated antisocial behavior represent an underlying psychopathic personality taxon embedded in a discrete etiology. The need for additional research related to both conceptual and clinical issues is strong, given the devastation associated with youth antisocial behavior.

ACKNOWLEDGMENTS

We thank Elizabeth Owens and Brian Zupan for statistical consultation, Steve Lee for implementing the delinquency severity outcome measure, and the psychopathy experts--Paul Frick, Sherryl Goodman, Robert Hare, Scott Lilienfeld, Donald Lynam, and Adrian Raine--for providing CCQ psychopathy prototypes. This work was supported by National Institute of Mental Health grant 45064, awarded to Stephen P. Hinshaw.

Received April 17, 2003; revision received March 24, 2004; accepted March 31, 2004

REFERENCES

Achenbach, T. M. (1991). Manual for the child behavior checklist14-18 and 1991 profile. Barlington. VT: University of Vermont, Department of Psychiatry.

American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (3rd ed., Rev.), Washington, DC: Author.

American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.), Washington, DC: Author.

Anastasi, A. (1988). Psychological testing (6th ed.) New York: Macmillan.

Barry, C. T., Frick, P. J., DeShazo, T. M., McCoy, M. G., Ellis, M., & Loney, B. R. (2000). The importance of callous-unemotional traits for extending the concept of psychopathy to children. Journal of Abnormal Psychology, 109, 335-340.

Blair, R. J. R., Colledge, E., Murray, L., & Mitchell, G. G. V. (2001). A selective impairment in the processing of sad and fearful expressions in children with psychopathic tendencies. Journal of Abnormal Child Psychology, 29, 491-498.

Block, J. (1961). The Q-sort method in personality assessment and psychiatric research. Palo Alto, CA: Consulting Psychologists Press.

Caputo, A. A., Frick, P. J., & Brodsky, S. L. (1999). Family violence and juvenile sex offending: Potential mediating roles of psychopathic traits and negative attitudes toward women. Criminal Justice and Behavior, 26, 338-356.

Christian, R., Frick, P. J., Hill, N., Tyler, L. A., & Frazer, D. (1997). Psychopathy and conduct problems in children: Subtyping children with conduct problems based on their interpersonal and affective style. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 233-241.

Cleckley, H. (1941). The mask of sanity. St. Louis, MO: C. V. Mosby.

Edens, J. F., Skeen, J. L., Cruise, K. R., & Cauffman, E. (2001). Assessment of juvenile psychopathy and its association with violence: A critical review. Behavioral Science and the Law, 19, 53-80.

Elliot, D. S., Huizinga. D., & Ageton, S. S. (1985). Explaining delinquency and drug use. Thousand Oaks, CA: Sage.

Erhardt, D., & Hinshaw, S. P. (1994). Initial sociometric impressions of attention deficit-hyperactivity disorder and comparison boys: Predictors from social behaviors and non-behavioral variables. Journal of Consulting and Clinical Psychology, 62, 833-842.

Fergusson, D. M., & Horwood, L. J. (1995). Predictive validity of categorically and dimensionally scored measures of disruptive childhood behaviors. Journal of the American Academy of Child and Adolescent Psychiatry, 34, 477-485.

Forth, A. E., & Mailloux, D. L. (2000). Psychopathy in youth: What do we know? In C. B. Gacono (Ed.), Clinical and forensic assessment of psychopathy: A practitioner's guide (pp. 25-54). Mahwah, NJ: Erlbaum.

Frick, P. J., Bodin, S. D., & Christopher, T. (2000). Psychopathic traits and conduct problems in community and clinic-referred samples of children: Further development of the Psychopathy Screening Device. Psychological Assessment, 12, 382-393.

Frick, P. J., Cornell, A. H., Barry, C. T., Bodin, S. D., & Dane, H. A. (2003). Callous-unemotional traits and conduct problems in the prediction of conduct problem severity, aggression, and self-report of delinquency. Journal of Abnormal Child Psychology, 31, 457-470.

Frick, P. J., & Ellis, M. (1999). Callous-unemotional traits and subtypes of conduct disorder. Clinical Child and Family Psychology Review, 2, 149-168.

Frick, P. J., & Hare, R. D. (2002). Antisocial Process Screening Device. Toronto, Ontario, Canada: Multi-Health Systems.

Frick, P. J., O'Brian, B. S., Wootton, J. M., & McBurnett, K. (1994). Psychopathy and conduct problems in children. Journal of Abnormal Psychology, 103, 700-707.

Goyette, C. H., Conners, C. K., & Ulrich, R. F. (1978). Normative data on the revised Conners parent and teacher rating scale. Journal of Abnormal Child Psychology, 6, 221-236.

Hare, R. D. (1991). The Hare psychopathy checklist--Revised. Toronto, Ontario, Canada: Multi-Health Systems.

Hare, R. D., Forth, A. E., & Strachan, K. E. (1992). Psychopathy and crime across the life span. In R. D. Peters, R. J. McMahon, & V. L. Quinsey (Eds.), Aggression and violence across the life span (pp. 285-300). Newbury Park, CA: Sage.

Hart, S. D., & Hare, R. D. (1997). Psychopathy: Assessment and association with criminal conduct. In D. M. Stoff, J. Breiling, & J. D. Maser (Eds.), Handbook of antisocial behavior (pp. 22-34). New York: Wiley.

Hinshaw, S. P. (1994). Conduct disorder in children: Conceptualization, diagnosis, comorbidity, and risk status for antisocial functioning in adulthood. In D. Fowls, P. Sutker, & S. Goodman (Eds.), Progress in experimental personality research and psychopathology research: Special progress on psychopathy and antisocial personality (Vol. 17, pp. 3-44). New York: Springer.

Hinshaw, S. P., Ablon, S., Kremen, A., & Zupan, B. (1997). Attention deficits with and without hyperactivity: An ipsative, egopsychological perspective. Poster session presented at The International Society for Research in Child and Adolescent Psychopathology, Paris, France.

Hinshaw, S. P., Carte, E. T., Sami, N., Treuting, J. J., & Zupan, B. A. (2002). Preadolescent girls with attention-deficit/hyperactivity disorder. II: Neuropsychological performance in relation to subtypes and individual classification. Journal of Consulting and Clinical Psychology, 70, 1099-1111.

Hinshaw, S. P., Heller, T., & McHale, J. P. (1992). Covert antisocial behavior in boys with attention deficit-hyperactivity disorder: External validation and effects of methylphenidate. Journal of Consulting and Clinical Psychology, 60, 274-281.

Hinshaw, S. P., & Lee, S. S. (2003). Oppositional defiant and conduct disorder. In E. J. Mash & R. J. Barkley (Eds.), Child psychopathology (2nd ed., pp. 144-198). New York: Guilford Press.

Hinshaw, S. P., & Melnick, S. M. (1995). Peer relationships in boys with attention-deficit hyperactivity disorder with and without comorbid aggression. Development and Psychopathology, 7, 627-647.

Hinshaw, S. P., & Renfro, R. (1999). Social behavior observation system. Unpublished manuscript, University of California, Berkeley.

Hinshaw, S. P., Simmel, C., & Heller, T. (1995). Multi-method assessment of covert antisocial behavior in children: Laboratory observations, adult ratings, and child self-reports. Psychological Assessment, 7, 209-219.

Hinshaw, S. P., & Zupan, B. A. (1997). Assessment of antisocial behavior in children and adolescents. In D. M. Stoff, J. Breiling, & J. D. Maser (Eds.), Handbook of antisocial behavior (pp. 36-50). New York: Wiley.

Hinshaw, S. P., Zupan, B. A., Simmel, C., Nigg, J. T., & Melnick, S. (1997). Peer status in boys with and without attention deficit-hyperactivity disorder: Predictors from overt and covert antisocial behavior, social isolation, and authoritative parenting beliefs. Child Development, 68, 880-896.

Hollingshead, A. B. (1975). Four factor index of social status. Unpublished manuscript. Department of Sociology, Yale University, New Haven, CT.

John, O. J., Caspi, A., Robins, R. W., Moffit, T. E., & Stouthamer-Loeber, M. (1994). The "little five": Exploring the nomological network of the five-factor model of personality in adolescent boys. Child Development, 65, 160-178.

Keenan, K., & Wakschlag, L. S. (2000). More than the terrible twos: The nature and severity of behavioral problems in clinic-referred preschool children. Journal Abnormal Child Psychology, 1, 33-36.

Lee, S. S., & Hinshaw, S. P. (in press). Severity of adolescent delinquency among boys with and without attention-deficit hyperactivity disorder: Predictions from early antisocial behavior and peer status. Journal of Clinical Child and Adolescent Psychology.

Lilienfeld, S. O. (1998). Methodological advances and developments in the assessment of psychopathy. Behavior Research and Therapy, 36, 99-125.

Lilienfeld, S. O., & Waldman, I. D. (1990). The relation between childhood attention deficit hyperactivity disorder and adult antisocial behavior reexamined: The problem of heterogeneity. Clinical Psychology Review, 10, 699-725.

Loeber, R., Brinthaupt, V. P., & Green, S. (1990). Attention deficits, impulsivity, and hyperactivity with and without conduct problems: Relationships to delinquency and unique contextual factors. In R. J. McMahon & R. D. Peters (Eds.), Behavior disorders of adolescence. New York, Plenum Press.

Loeber, R., Stouthamer-Lober, M., Van Kamen, W. B., & Farrington, D. P. (1991). Initiation, escalation, and desistance in juvenile offending and their correlates. Journal of Criminal law and Criminology, 82, 36-82.

Lowe, L. A. (1998). Using the child behavior checklist in assessing conduct disorder: Issues of reliability and validity. Research on Social Work Practice, 3, 268-301.

Lynam, D. R. (1996). Early identification of chronic offenders: Who is the fledgling psychopath? Psychological Bulletin, 120, 209-234.

Lynam, D. R. (1997), Pursuing the psychopath: Capturing the psychopath in a nomological net. Journal of Abnormal Psychology, 10, 425-438.

Lynam, D. R. (1998). Early identification of the fledgling psychopath: Locating the psychopathic child in the current nomenclature. Journal of Abnormal Psychology, 107, 566-575.

Nigg, J. T., Hinshaw, S. P., & Halperin, J. M. (1996). The continuous performance test in boys with ADHD: Methylphenidate dose response and relationships with observed behavior. Journal of Clinical Child Psychology, 25, 330-340.

Nunnally, J. (1970). Introduction to psychological measurement. New York: McGraw-Hill.

O'Brian, B. S., & Frick, P. J. (1996). Reward dominance: Association with anxiety, conduct problems, and psychopathy in children. Clinical Psychology Review, 13, 699-720.

Osborne, J. (2002). Notes on the use of data transformation. Practical Assessment; Research, and Evaluation, 8, 1-11.

Parker, J. G., & Asher, S. R. (1987). Peer relations and later personal adjustment: Are low-accepted children at-risk? Psychological Bulletin, 102, 357-389.

Pedhazur, E. J. (1982). Multiple-regression in behavioral research: Explanation and prediction (2nd ed.). New York: Holt, Rinehart, & Winston.

Pelham, W. E., Gnagy, E. M., Greenslade, L., & Milich, R. (1992). Teacher ratings from the DSM-III-R symptoms for the disruptive behavior disorders. Journal of the American Academy of Child and Adolescent Psychiatry, 31, 210-218.

Reise, S. P., & Oliver, C. J. (1994). Development of the California q-set indicator of primary psychopathy. Journal of Personality Assessment, 62, 130-144.

Reise, S. P., & Wink, P. (1995). Psychological implications of the psychopathy Q-sort. Journal of Personality Assessment, 65, 300-312.

Salekin, R. T., Rogers, R., & Machin, D. (2001). Psychopathy in youth: Pursuing diagnostic clarity. Journal of Youth and Adolescence, 30, 173-195.

Silverthorn, P., Frick, P. J., & Reynolds, R. (2001). Timing of onset and correlates of severe conduct problems in adjudicated boys and girls. Journal of Psychopathology and Behavioral Assessment, 23, 171-181.

Swanson, J. M. (1992). Assessment and treatment of ADD students. Irvine, CA: K. C. Press.

Vitacco, M. J., Neuman, C. S., Robertson, A. A., & Durrant, S. L. (2002). Contributions of impulsivity and callousness in the assessment of adjudicated male adolescents: A prospective study. Journal of Personality Assessment, 78, 87-103.

Wechsler, D. (1991). Wechsler intelligence scale for children (3rd ed.). New York: Psychological Corporation.

Welner, Z., Reich, W., Herjanic, B., & Jung, K. (1987). Reliability, validity, and parent-child agreement studies of the diagnostic interview for children and adolescents (DICA). Journal of the American Academy of Child and Adolescent Psychiatry, 26, 649-653.

Woodcock, R. W., & Johnson, M. B. (1990). Woodcock-Johnson psychoeducational battery--revised. Allen, TX: DLM Teaching Resources.

Wootton, J. M., Frick, P. J., Shelton, K. K., & Silverthorn, P. (1997). Ineffective parenting and childhood conduct problems: The moderating role of callous-unemotional traits. Journal of Consulting and Clinical Psychology, 65, 301-308.

Auran Piatigorsky (1) and Stephen P. Hinshaw (1,2)

(1) Department of Psychology, University of California, Berkeley, California.

(2) Address all correspondence to Stephen P. Hinshaw, Department of Psychology, 2205 Tolman Hall #1650, University of California, Berkeley, California 94720-1650; e-mail: hinshaw@socrates.berkeley.edu.

COPYRIGHT 2004 Plenum Publishing Corporation
COPYRIGHT 2004 Gale Group