Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Construct validity of psychopathic personality traits in a cohort of young twins
(USC Thesis Other)
Construct validity of psychopathic personality traits in a cohort of young twins
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
CONSTRUCT VALIDITY OF PSYCHOPATHIC PERSONALITY TRAITS
IN A COHORT OF YOUNG TWINS
by
Joshua Isen
_______________________________________________________
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)
May 2010
Copyright 2010 Joshua Isen
ii
Acknowledgements
This study would not have been possible without the material and moral support
of many individuals. I would like to express a special thanks to my advisor Dr. Laura
Baker for her encouragement, mentorship, and friendship over the years. I am also
grateful to Dr. Michael Dawson for his invaluable instruction and guidance. I would like
to acknowledge Dr. Carol Prescott for her comments and guidance, as well as thank Drs.
Biing-Jiun Shen and Penelope Trickett.
I owe the deepest gratitude to my family for always maintaining their faith in me,
and continuously encouraging me to reach my potential. In particular, I would like to
thank my parents and brother for all their love and support.
I would like to thank the many staff members involved with the University of
Southern California Twin Project; I am indebted to them for their services in data
collection and recruitment of twins. Further thanks go to Hyeran Shin and Kelly Norris
for their valuable assistance in processing psychophysiological data. I would also like to
acknowledge Dr. Serena Bezdjian for her excellent work with the Childhood
Psychopathy Scale, which inspired me to undertake this study. Last but not least, I
appreciate the enormous contributions of the twins and their families who have
participated in this study.
iii
Table of Contents
Acknowledgements ii
List of Tables v
List of Figures vii
Abstract viii
Chapter One: Background and Significance 1
Conceptualization of Psychopathy 2
Psychopathy and Antisocial Behavior 4
Psychopathy and Anxiety 5
Psychopathy and Behavioral Inhibition System 6
Psychopathy and Skin Conductance Activity 7
Psychopathy and Intelligence 8
Aims and Specific Hypotheses 9
Chapter Two: Method 11
Participants 11
Procedure 12
Measures 13
Psychopathy 13
Antisocial Behavior 15
Anxiety/Depression 16
ADHD 17
Intelligence 17
Skin Conductance Activity 18
Psychophysiological Tasks 20
Shameful Questions Task 20
Countdown Task 22
Statistical Analyses 23
Measurement Model of Psychopathy 25
Etiological Overlap with Criterion Variables 30
Chapter Three: Results 33
Descriptive Statistics 33
Measurement Model of Psychopathy 35
Construct Validity 39
Associations with Antisocial Behavior 39
Associations with ADHD 42
Associations with Anxiety/Depression 43
Associations with Intelligence 44
Associations with Anticipatory SCRs 46
iv
Associations with SCRs to White Noise 47
Associations with SCR Magnitudes 48
Chapter Four: Discussion 50
General-Specific Factor Model 50
Construct Validity 51
Antisocial Behavior 51
ADHD 53
Anxiety/depression 54
Intelligence 55
Skin Conductance Activity 57
Limitations 59
Strengths 63
Concluding Remarks 63
References 65
Appendix: Supplementary Analyses, Figures, and Tables 70
v
List of Tables
Table 1 Formulae for calculating and standardizing the genetic and
environmental parameters of the general-specific factor model, as
depicted in Figure 4
32
Table 2 Means and standard deviations of the CPS subscales
33
Table 3 Means and standard deviations of the criterion measures
34
Table 4 Model-fitting results
36
Table 5 Standardized factor loadings from the general-specific factor model
in males (and females in parentheses)
37
Table 6 Standardized estimates of the genetic/environmental variance
components (95% confidence intervals in parentheses)
38
Table 7 Standardized estimates of the genetic/environmental variance
components for each criterion variable
39
Table 8 Genetic and environmental contributions to the association between
psychopathy and antisocial behavior (ASB)
41
Table 9 Genetic and environmental contributions to the association between
psychopathy and ADHD
43
Table 10 Genetic and environmental contributions to the association between
psychopathy and anxiety/depression
44
Table 11 Genetic and environmental contributions to the association between
psychopathy and IQ
45
Table 12 Phenotypic correlations between psychopathy and
Verbal/Performance IQ
46
Table 13 Genetic and environmental contributions to the association between
psychopathy and the frequency of anticipatory skin conductance
responses
47
Table 14 Genetic and environmental contributions to the association between
psychopathy and electrodermal responsiveness to signaled white-
noise bursts
48
vi
Table 15 Genetic and environmental contributions to the association between
psychopathy and electrodermal responsiveness to unsignaled white-
noise bursts
48
Table 16 Genetic and environmental contributions to the association between
psychopathy and mean SCR magnitude to shameful questions
49
Table A1 Childhood Psychopathy Scale (CPS)
71
Table A2 Conduct Disorder Symptoms (DISC-IV)
74
Table A3 Reactive Aggression (RPQ)
75
Table A4 Proactive Aggression (RPQ)
76
Table A5 Attention Problems (CBCL)
77
Table A6 Inattention Symptoms (DISC-IV)
78
Table A7 Hyperactivity/Impulsivity Symptoms (DISC-IV)
79
Table A8 Anxious/Depressed (CBCL)
80
Table A9 Generalized Anxiety Disorder Symptoms (DISC-IV)
81
Table A10 Major Depression Symptoms (DISC-IV)
82
Table A11
Intraclass twin correlations for antisocial behavior
83
Table A12
Intraclass twin correlations for ADHD
84
Table A13
Intraclass twin correlations for anxiety/depression
85
Table A14
Intraclass twin correlations for WASI subtests 86
Table A15
Intraclass twin correlations for frequency of SCRs during
Countdown task
87
Table A16 Intraclass twin correlations for SCR magnitude during Questions
task
88
vii
List of Figures
Figure 1 Illustration of an ACE model
25
Figure 2 A general-specific factor model of psychopathy
27
Figure 3 A second-order factor model of psychopathy
28
Figure 4 Path model depicting the structural associations between the CPS
factors and a latent criterion variable
31
Figure A1 Path diagram of a second-order factor model 70
viii
Abstract
A nomological network of constructs was developed in order to evaluate the
construct validity of psychopathy in preadolescent children. Theoretical
conceptualizations of psychopathy in adults emphasize its multi-faceted nature; the
interpersonal features of psychopathy are thought to be indicative of positive adjustment,
whereas the impulsive/behavioral traits are thought to represent risk for maladaptive
outcomes. Hence, if psychopathy is a valid construct in children, then different traits
should show divergent associations with respect to external criteria.
Subjects were 605 pairs of same-sex and opposite-sex twins recruited from
diverse communities throughout Los Angeles. Data were obtained from twins and their
caregivers during the 1
st
wave of an ongoing twin study, when the twins were 9-10 years
old. Psychopathy was assessed using a caregiver version of the Childhood Psychopathy
Scale (Lynam, 1997). Previous factor-analytic work has shown that the subscales of the
CPS can be organized into 2 factors: Charming/Manipulative and Callous/Disinhibited.
A general-specific factor model was fit to the data to allow for easier interpretation of the
associations between different facets of psychopathy and external criteria.
Construct validity of psychopathy was evaluated using a range of behavioral,
cognitive, and psychophysiological measures. The genetic and environmental
contributions to the associations with external criteria were examined using standard twin
models. The genetic variance that was common to all of the psychopathic traits (i.e., the
general factor) was highly correlated with measures of antisocial behavior and
inattention/hyperactivity. This general factor was also positively associated with
anxiety/depression, particularly in females. General psychopathy was inversely related to
ix
IQ and to electrodermal responses, but these associations were modest and mainly
confined to males. The interpersonal (Charming/Manipulative) factor of psychopathy, by
contrast, was not associated with external criteria in a theoretically expected manner.
1
Chapter One: Background and Significance
The study of personality has long been a central issue in the literature on
antisocial behavior. Over the last several decades, psychologists have reached a
consensus that individual differences in aggression and delinquency are shaped by
enduring traits and predispositions. There is overwhelming evidence that antisocial
behavior is heritable in children as well as adults (Waldman & Rhee, 2006). It has been
linked to a myriad of other heritable traits such as neuroticism, low IQ, and autonomic
underreactivity (Raine, 1993). Psychopathy is at the forefront of our understanding of the
intersection between personality and antisocial behavior.
The social costs exacted by antisocial behavior are immense. As a result, there is
much interest in identifying the traits that predict aggressive and delinquent outcomes.
Psychopathy contains many traits that predispose individuals to antisocial behavior (e.g.,
impulsiveness, poor anger control, deceitfulness, and lack of empathy). Many of these
traits are diagnostic criteria of Antisocial Personality Disorder (ASPD). There is much
confusion regarding the boundaries between psychopathy and ASPD. In part, this is
because psychopathy embraces the symptoms of ASPD as well as containing some
extraneous personality features.
Although a full-blown manifestation of psychopathy is not applicable to children,
psychopathic-like dimensions can be measured at an early age (Farrington, 2005; Lynam,
1997). Moreover, psychopathy displays predictive validity not only among incarcerated
males, but also in mixed-gender community samples, where psychopathic characteristics
are more benign (Bare, Hopko, & Armento, 2004; Blonigen et al., 2005).
2
Conceptualization of Psychopathy
There is little consensus over the nature and definition of psychopathy. One point
of agreement is that psychopathy is multi-faceted, comprising at least two factors. Ever
since Cleckley’s classic description of the clinical psychopath, various instruments have
been devised to measure psychopathy in adults. The Psychopathy Checklist-Revised
(PCL-R; Hare, 1991) has been the most widely-used instrument, especially for
incarcerated samples. Preliminary work on the PCL-R indicated that there were two
correlated factors: an interpersonal/affective factor (“selfish, callous, and remorseless use
of others”) and an impulsive/antisocial factor (“chronically unstable, antisocial, and
socially deviant lifestyle”). More recently, the PCL-R has been divided into four facets:
interpersonal, affective, lifestyle, and antisocial (Hare, 2003).
There is much controversy over whether antisocial behavior should be included
as a component of psychopathy. Cooke, Michie, & Hart (2006) argue that antisocial
behavior is a consequence, rather than a symptom, of psychopathy. By tainting a
personality syndrome with measures of antisocial behavior, it becomes more difficult to
elucidate the personality processes that are responsible for external phenomena. In other
words, there is little to be gained by incorporating antisocial behaviors as diagnostic
criteria. It muddles rather than clarifies our theoretical understanding of psychopathy.
Psychopathy is based on several seeming contradictions. On one hand, the
prototypical psychopath is described as showing good intelligence, an absence of
nervousness, and social potency (Cleckley, 1941; Lynam & Derefinko, 2006). In other
respects, the same individual shows profound ineptness: unreliability, impulsiveness, lack
of long-term goals, and proneness to fits of anger. The complex and multi-faceted
3
structure of psychopathy stems from the co-occurrence of socially advantageous features,
on one hand, and harmful, self-destructive traits on the other hand. It is this intriguing
paradox that led Hervey Cleckley (1941) to describe psychopathy as a ‘mask of sanity’.
Most conceptualizations of psychopathy have implicitly emphasized the
distinction between two broad factors: primary and secondary psychopathy (Skeem,
Poythress, Edens, Lilienfeld, & Cale, 2003; Lykken, 1995). In keeping with this
theoretical conceptualization, many instruments operationalize psychopathy in such a
manner that two factors emerge: an ‘interpersonal’ factor and an ‘impulsiveness’ factor.
The external correlates of the two factors are often distinct, especially when controlling
for their shared variance. The first, or primary, factor generally reflects positive
adjustment (e.g., extraversion and stress immunity), whereas the second factor is
typically associated with maladaptive traits (negative emotionality and poor self-control).
Given that the two factors tend to have distinct correlates with criterion measures,
use of the global (full-scale) psychopathy score may minimize or obscure the underlying
personality processes. The considerable covariance between the two factors can also lead
to suppressor effects. It is not unusual for the two PCL-R factors to exhibit opposing
relations with an external criterion when entered simultaneously in a prediction model
(Hicks & Patrick, 2006). This clearly threatens the integrity of the psychopathy construct,
which is typically modeled as a single higher-order dimension that is responsible for two
lower-order domains.
Patrick et al., (2007) investigated alternative ways of modeling the structure of
psychopathy in adult offenders. They used a hierarchical approach to account for the
general variance permeating all of the PCL-R items, and then modeled specific factors
4
(interpersonal, affective, and impulsivity) to account for the unique variance shared by
certain subsets of items. By adopting this approach, it became much easier to isolate the
specific aspects of psychopathy that were related to external variables. The general
psychopathy factor was found to be positively correlated with negative emotionality,
whereas the specific interpersonal factor was negatively correlated with negative
emotionality (Patrick et al., 2007). Moreover, they showed that the common variance
underlying all items of the PCL-R was essentially redundant with externalizing problems
(i.e., symptoms of antisocial personality disorder and child conduct disorder).
Psychopathy and Antisocial Behavior
There is much dispute over whether antisocial behavior is an actual component
per se or rather a characteristic adaptation of psychopathy. Investigators are in agreement
that psychopathic traits are intrinsically related to antisocial behavior. However, it is
uncertain whether antisocial behavior is a defining feature of psychopathy or simply a
consequence of the underlying personality disorder. In juveniles, there is evidence that
the personality traits of psychopathy predict delinquency (Lynam et al., 2009; Lynam,
1997). That is, predictive validity has been obtained for instruments such as the
Childhood Psychopathy Scale (CPS), which emphasize the personality features of
psychopathy rather than blatantly antisocial behaviors.
Bezdjian (2008) submitted the subscales of the Childhood Psychopathy Scale to
an exploratory factor analysis, and identified two correlated factors. The two factors –
Charming/Manipulative and Callous/Disinhibited – were both moderately correlated with
externalizing problems. For example, the Charming/Manipulative and
Callous/Disinhibited factors obtained correlations of 0.50 and 0.51, respectively, with
5
reactive aggression. Given that both factors were equally associated with antisocial
behavior, the discriminant validity of a two-factor conceptualization of psychopathy was
not demonstrated. Bezdjian’s (2008) results suggest that a hierarchical approach to
modeling the structure of psychopathy is warranted. That is, two highly related domains
appear to comprise a general construct.
The overlap between psychopathic traits and antisocial behavior appears to stem
from common genetic sources. Larsson et al. (2007) investigated the etiological
influences on psychopathic traits and antisocial behavior (ASB) in an adolescent twin
sample. A single genetic factor was found to explain the covariance among psychopathic
traits and ASB, whereas shared environmental influences were specific to ASB. Thus,
the two are distinguishable on account of the fact that shared environmental factors
influence ASB, but not psychopathy. Psychopathic traits may be more ‘endophenotypic’
in nature, or closer to the heritable processes that contribute to antisocial behavior
(Waldman & Rhee, 2006). In turn, family-wide environmental variables such as
socioeconomic status may shape ASB in additional ways. These subtle differences in
etiology may help to disentangle the constructs of psychopathy and ASB.
Psychopathy and Anxiety
Abnormally low anxiety/nervousness is considered a core feature of psychopathy
(Lykken, 1995). Cleckley (1988) describes the prototypical psychopath as showing a
“relative immunity from such anxiety and worry as might be judged normal or
appropriate in disturbing situations" (pp. 339-340). Low anxiety presumably facilitates
the commission of antisocial behavior. The fact that psychopaths lie and steal without
compunction would suggest that they do not experience anticipatory fear and anxiety. At
6
the opposite pole, individuals with a “nervous” or harm-avoidant temperament would
likely be too inhibited to engage in antisocial behavior. Higher social emotions such as
shame and embarrassment are also thought to be poorly developed in psychopaths. A
healthy dose of these “internalizing” states – shame, guilt, fear, anxiety/worry, and
embarrassment - should tend to inhibit antisocial behavior.
This leads to an apparent contradiction. Contrary to theory, individuals who
chronically engage in antisocial behavior tend to be more anxious/depressed than typical
individuals (Sareen et al., 2004). The covariance between externalizing and internalizing
problems in children is unusually high (Gjone & Stevenson, 1997). Conduct-disordered
children are at elevated risk for anxiety/depression, and there is high comorbidity
between antisocial personality disorder and anxiety disorders (Widiger, 2006). How can
this be reconciled with the hypothetically low anxiety of psychopaths?
This paradox may stem from the fact that the two psychopathy factors exhibit
divergent associations with anxiety. For example, emotional distress is strongly and
positively related to the impulsive/antisocial factor of the PCL-R, but is negatively
related to the interpersonal factor (Hicks & Patrick, 2006). As a result, higher levels of
anxiety/distress are observed in the bulk of antisocial individuals, who happen to not
possess the compensatory interpersonal features of psychopathy.
Psychopathy and Behavioral Inhibition System
Some aspects of anxiety can be explained by the Behavioral Inhibition System.
Fowles (2000) argues that an underactive Behavioral Inhibition System is characteristic
of psychopathy, particularly with respect to the interpersonal facet. This would account
for the glibness and social boldness of the prototypical psychopath. According to Gray’s
7
(1987) psychobiological model of personality, there are two orthogonal systems that
determine behavior: the Behavioral Inhibition System (BIS) and the Behavioral Approach
System. The BIS is activated in response to signals of punishment and non-reward,
leading to the cessation of behavior. The Behavioral Approach System (BAS), on the
other hand, is sensitive to novel stimuli as well as cues of reward, and leads to the
activation of behavior. There are large individual differences in the strength of these two
systems. Impulsiveness is an extreme manifestation of the BAS, whereas
fearfulness/shyness is an extreme manifestation of the BIS.
Gray’s personality system has important implications in understanding the causes
of antisocial behavior. Some individuals may engage in antisocial behavior due to the
inability to delay gratification (overactive BAS), whereas others may have a normal BAS,
but be relatively immune to shame and embarrassment (underactive BIS). This has
spurred some commentators (Fowles & Dindo, 2006) to suggest that psychopathy is a
dual-process syndrome involving the co-occurrence of fearlessness (low BIS) and
impulsiveness (high BAS).
Psychopathy and Skin Conductance Activity
The BIS is thought to be indexed by activity of the sympathetic nervous system
(Fowles, 2000). Since electrodermal activity (EDA) is one of the purest measures of the
sympathetic nervous system, most of the psychophysiological research on psychopathy
has focused on EDA measures. One of the more replicable findings is that psychopathic-
like individuals show reduced electrodermal reactivity to aversive stimuli (Lorber, 2004;
Benning et al., 2005; Blair, 1999). Psychopathy has been associated with lower
responsiveness to stimuli that signal the onset of aversive events (Fung et al., 2005; Hare,
8
Frazelle, & Cox, 1978). When imagining fearful or distressing situations, psychopaths
show less increase in skin conductance activity as compared to non-psychopaths (Patrick,
Cuthbert, & Lang, 1994).
Psychopathy and Intelligence
There appears to be an inverse association between externalizing problems and
intelligence. Twin studies indicate that the overlap between low IQ and antisocial
behavior in children is completely due to common genetic sources (Koenen et al., 2006).
Antisocial outcomes in adolescents are generally associated with lower IQ, especially the
verbal component of IQ. As a result, one might assume that psychopathy is inversely
related to intelligence. However, the literature shows that the various psychopathy facets
have divergent associations with IQ. The interpersonal factor of the PCL-R is positively
related to IQ, whereas the affective factor is negatively related to IQ (Vitacco et al., 2008;
Vitacco et al., 2005). A similar pattern is obtained in community samples of children
(Fontaine et al., 2008). Interpersonal glibness is positively related to IQ, whereas
callous/unemotional traits and impulsivity are negatively related to IQ. This is a rather
remarkable divergence, given that interpersonal and affective traits are often considered
part of the same primary dimension. It lends credence to Cleckley’s assertion that the
prototypical psychopath is not intellectually deficient, particularly if he is superficially
charming.
In summary, there is evidence that low IQ and high negative emotionality are
ubiquitous in individuals with externalizing psychopathology. However, according to
Cleckley’s (1976) influential treatise, the classic psychopath is distinguished by his ‘good
intelligence’ and absence of ‘nervousness’. These contradictions can potentially be
9
explained by invoking a hierarchical (i.e., general-specific) conceptualization of
psychopathy, in which there is a general factor and two specific factors accounting for the
covariance among the subscales. This would explain the overall coherence of the
psychopathy construct while also accounting for the traditional two-factor structure. A
general factor should permeate all of the psychopathy subscales, representing a broad
liability for antisocial behavior (Patrick et al., 2007). This general factor should be
associated with the typical correlates of externalizing problems in children (e.g., lower IQ
and greater anxiety/depression). However, the specific interpersonal facet (as represented
by charm/manipulation) should be related to good cognitive and psychosocial outcomes.
Aims and Specific Hypotheses
The goal of this dissertation is to develop a theoretical conceptualization of
psychopathy in children. To this end, I have developed a nomological network of
constructs related to psychopathy. The research plan for this dissertation can be divided
into two parts: 1) establishing an optimal measurement model of psychopathy, and 2)
examining its validity through multivariate genetic analyses of the relations between
psychopathy and various external criteria. For the first aim, I expect that a general-
specific factor model will provide the best fit to the data. The two CPS factors that
Bezdjian (2008) identified - Charming/Manipulative and Callous/Disinhibited – should
appropriately be modeled as specific factors (orthogonal to a general factor) rather than
as two correlated factors.
Psychopathy, as a whole, appears to encapsulate the personality traits that
predispose individuals to antisocial behavior. It should represent the genetic liability for
antisocial outcomes. Therefore, I hypothesize that the genetic influences operating on the
10
general factor (i.e., the common variance underlying all 12 psychopathy subscales) will
completely overlap with the genetic risk for antisocial behavior. Genetically speaking,
psychopathy and antisocial behavior should not be separable constructs.
Furthermore, the general psychopathy factor should account for the traditional
correlates of antisocial behavior in children, whereas the interpersonal factor should be
related to more positive outcomes. More specifically, IQ should be negatively correlated
with the general factor, but positively correlated with charm/manipulation. A similar
divergence should occur for anxiety/depression, with the general factor predicting greater
internalizing symptoms, and the interpersonal factor predicting fewer symptoms.
Finally, the specific factors should show distinct external correlates, especially
with respect to skin conductance responsiveness and ADHD symptoms.
The interpersonal domain, representing an underactive Behavioral Inhibition System,
should be negatively correlated with electrodermal indices of anticipatory fear and
embarrassment. This prediction stems from my observation that Charming/Manipulative
is negatively correlated with skin conductance orienting responses (Isen et al., 2010).
The Callous/Disinhibited factor, on the other hand, should be strongly associated with
ADHD symptoms. It should be noted that these predictions are largely speculative, given
that these facets have yet to be examined using a general-specific factor approach. That
is, they may function quite differently from how they appear to behave (when modeled as
correlated factors).
11
Chapter Two: Method
Participants
The sample was drawn from participants in the USC Twin Study of Risk Factors
for Antisocial Behavior, an ongoing longitudinal study of the interplay of genetic,
environmental, social, and biological factors on the development of antisocial behavior
(Baker et al., 2006). The twins were recruited from Greater Los Angeles, and the sample
is representative of the ethnic and socio-economic makeup of this region (Baker et al.,
2002). The ethnic distribution of the sample is as follows: 37.5% Hispanic, 26.6%
Caucasian, 14.3% Black, 4.5% Asian, 0.3% Other, and the remaining 16.7% of mixed
heritage.
The present research specifically deals with data from the first wave of
assessment in 2000-2004, when the children were 9-10 years old. The mean age at the
time of assessment was 9.56 years (standard deviation = 0.58). There were a total of 605
participating families (N = 1219 twins/triplets). The maximum sample size for analyses
was 1210, rather than 1219, because nine of the 605 families consisted of triplets. (Since
the inclusion of triplets would have complicated the statistical design, one triplet from
each of the families was randomly omitted).
Assessment of psychopathy and behavior problems was provided by caregivers,
the vast majority of whom (91.4%) were biological mothers (N = 553). Caregivers were
administered the questionnaires/interviews in either English (N = 492) or Spanish (N =
113), depending on their language proficiency and preference. The twins were required
to be proficient in the English language.
12
Zygosity was determined using DNA microsatellite analysis for 87% of the twin
pairs. For the remaining pairs, zygosity was established by questionnaire items about the
twins’ physical similarity and the frequency with which people confuse them. Because
this study is longitudinal, it was common to obtain multiple DNA samples over time,
which were then cross-checked with the questionnaire ratings. For approximately 10
pairs, there were major discrepancies between the DNA results and the ratings of physical
similarity. Ordinarily, the DNA results would serve as the final word, but sometimes the
DNA results were inconclusive. Discrepancies were resolved by evaluating the twins’
photographs; six judges independently rated whether a given pair was monozygotic (MZ)
or dizygotic (DZ). A consensus was reached in each case. The breakdown of the sample
in terms of gender and zygosity are as follows: MZ male (N = 277), MZ female (N =
278), DZ male (N = 170), DZ female (N = 195), and DZ opposite-sex (N = 299).
Procedure
Families took part in 6-8 hours of assessment during a laboratory visit, entailing
interview and neurocognitive measures for both the twins and their caretakers. (Only the
caregiver ratings of twins’ traits/behaviors, rather than self-reports, were used for the
present study). During the neurocognitive assessment, a 30-minute intelligence test was
administered to the twins according to standard protocol. The twins also underwent
psychophysiological testing, in which electrodermal, electrocortical, and cardiovascular
measures were collected. The psychophysiological protocol lasted approximately three
hours and included several tasks.
13
Measures
Psychopathy
Psychopathy was measured using a modified version of the Childhood
Psychopathy Scale (CPS; Lynam, 1997). The CPS was designed to operationalize in
childhood the personality traits of psychopathy, as measured in adults via the
Psychopathy Checklist-Revised (PCL-R; Hare, 1991). Several domains from the adult
instrument (e.g., promiscuous sexual behavior) were omitted in order to produce a
developmentally appropriate measure of psychopathy (Lynam, 1997). Items reflecting
blatant antisocial behavior (e.g., juvenile delinquency) were also excluded in order to
produce a more personality-based conceptualization (Lynam et al., 2005). As a result,
the CPS only assesses 14 of the 20 PCL-R criteria: Glibness, Grandiosity, Boredom
Susceptibility, Untruthfulness, Manipulation, Lack of Guilt, Poverty of Affect,
Callousness, Impulsiveness, Parasitic Lifestyle, Behavioral Dyscontrol, Lack of Planning,
Unreliability, and Failure to Accept Responsibility,
The original 41-item CPS has undergone several revisions and expansions. The
present version is a questionnaire of 58 items (“no” = 1, “yes” = 2) that shows good
criterion-related validity among youth offenders (Falkenbach, Poythress & Heide, 2003).
The items that comprise each subscale are listed in Table A1 in the appendix. Given that
Bezdjian (2008) reported the psychometric properties of these subscales in great detail, a
brief summary will suffice. The internal consistency (Cronbach’s α) of the subscales
ranged from a low of 0.33 (Failure to Accept Responsibility) to a high of 0.75
(Behavioral Dyscontrol). Furthermore, test-retest reliabilities were calculated for 60
14
individuals who were rated twice within a period of approximately six months. The test-
retest correlations ranged from 0.48 (Lack of Guilt) to 0.81 (Boredom Susceptibility).
Previous investigators (e.g., Falkenbach et al., 2003; Lynam et al., 2005) typically
organized the 14 subscales into two broad factors that reflect the traditional factor
structure in adults (Hare, 1991). This was based on rational considerations, rather than
arrived at empirically. However, upon submitting these subscales to a factor analysis
with oblique rotation, Bezdjian (2008) found that the two-factor solution was not
structured according to Hare’s original conceptualization. The first extracted factor
contained a composite of affective and impulsive behavioral traits, which she labeled
‘Callous/Disinhibited’. It is composed of seven subscales: Callousness, Poverty of Affect,
Unreliability, Lack of Planning, Impulsiveness, Behavioral Dyscontrol, and Boredom
Susceptibility. The next factor appeared to tap into a ‘Charming/Manipulative’
orientation: Glibness, Untruthfulness, Manipulation, Failure to Accept Responsibility,
and Parasitic Lifestyle. Confirmatory factor analyses (see Bezdjian, 2008) demonstrated
that this structure was superior to the original PCL-R framework and to Cooke and
Michie’s (2001) three-factor structure.
Bezdjian (2008) noted that two subscales – Grandiosity and Lack of Guilt – failed
to load on either factor. Grandiosity, in fact, was negatively correlated with the other
subscales. It was also negatively skewed, due to the fact that most individuals obtained
high scores on this measure. As one can see in Table A1, the face validity of Grandiosity
is questionable, as the items appear to reflect self-esteem rather than arrogance per se.
Consequently, these subscales were excluded from the two CPS factors that Bezdjian
(2008) used.
15
Antisocial Behavior
Antisocial behavior was indexed by the following three measures: reactive
aggression, proactive aggression, and conduct problems. Reactive and proactive
aggression was operationalized using the Reactive-Proactive Aggression Questionnaire
(RPQ; Raine et al., 2006). The RPQ consists of 23 items about various aggressive
behaviors, which are subdivided into reactive aggression (11 items) and proactive
aggression (12 items). Aggressive acts are considered reactive when they are generated
by provocation or anger/frustration. For example, a child who yells when upset or hits
back when teased/attacked would be engaging in reactive aggression. Proactive
aggression, by contrast, is not driven by intense emotions, but is initiated for instrumental
purposes. Threatening or bullying others would be an example of proactive aggression.
Caregivers responded to the items on a 3-point scale: Never (0), Sometimes (1), or Often
(2).
The two-factor structure of the RPQ was validated in the present sample by Baker
et al. (2008). Cronbach’s α for reactive and proactive aggression were 0.83 and 0.77,
respectively. Also, the test-retest reliability was obtained for 60 individuals over a period
of six months. Test-retest correlations were 0.81 and 0.79 for reactive and proactive
aggression, respectively.
Finally, the lifetime number of conduct disorder symptoms was assessed using the
Diagnostic Interview Schedule for Children (DISC-IV; Schaffer et al., 2000). Caregivers
reported on a wide range of delinquent behaviors, such as stealing/shoplifting, lying to
avoid obligations, skipping school, bullying/fighting, and fire setting. These behaviors
could occur at any point in the twin’s life. There are typically a total of 26 possible
16
symptoms on the DISC-IV. However, an item regarding sexual behavior was omitted
due to its inappropriateness in preadolescents, thereby yielding an effective total of 25
symptoms. Symptom counts were set to missing if subjects were missing data for one or
more symptoms. Only a single subject was disqualified on account of missing data.
More than half (54.5%) of the boys had at least one conduct disorder symptom,
and 33.7% of boys had at least 2 symptoms in their lifetime. By contrast, only 39.2% of
girls had one or more conduct disorder symptoms. Children qualified for a diagnosis of
conduct disorder if they manifested three or more criteria within a 12-month period. A
total of 30 children (2.5% of the sample) met criteria for conduct disorder at some point
in their lifetime: 18 males (3.1%) and 12 females (2.0%) .
Anxiety/Depression
Internalizing problems was assessed using the Anxious/Depressed scale of the
Child Behavior Checklist (CBCL; Achenbach, 1991) as well as DISC-IV symptom
counts of Generalized Anxiety Disorder (GAD) and Major Depression. The
Anxious/Depressed scale of the CBCL contains 14 items that pertain to personal distress.
Caregivers responded to each item on a three-point scale: 0 = Not True, 1 = Somewhat or
Sometimes True, and 2 = Very True or Often True. Raw scores were obtained by
summing the responses. Using the semi-structured interview format of the DISC-IV,
caregivers also reported on various anxiety (GAD) symptoms in their children, including
excessive worrying, irritability, and somatic complaints. Symptoms of Major
Depression included excessive self-blame, diminished pleasure, feelings of worthlessness,
fatigability, and depressed mood.
17
ADHD
Symptoms of inattention and hyperactivity were assessed using the ADHD
module of the DISC-IV. There were a total of 22 ADHD symptoms, divided equally
between Inattention (11 items) and Hyperactivity-Impulsivity (11 items). Subjects who
were missing data for more than two symptoms were omitted from analyses. This cut-off
was decided by the fact that 240 subjects (20% of the sample) had missing data for one or
two symptom(s). Ultimately, only a single subject was disqualified, as he was missing
data for 11 symptoms. In order to create a latent ADHD factor, the inattention and
hyperactivity symptom counts were supplemented by scores on the Attention Problems
subscale of the CBCL. Caregivers responded to 11 items using the following scale: 0 =
Not True, 1 = Somewhat or Sometimes True, and 2 = Very True or Often True. Raw
scores were calculated by summing the responses to all items. As can be seen in the
appendix, the choice to combine this measure with the two symptom counts was justified
because Attention Problems contains items that overlap with hyperactivity-impulsivity as
well as inattention. In other words, it is related to both components of ADHD, not solely
inattention (despite the ‘Attention’ Problems label). For example, it contains items such
as “impulsive or acts without thinking” and “can't sit still, restless, or hyperactive”.
Intelligence
Cognitive ability was assessed using the Wechsler Abbreviated Scale of
Intelligence (WASI; Wechsler, 1999). The WASI consists of four subtests: Vocabulary,
Similarities, Block Design, and Matrix Reasoning. Standardization data for the WASI
was obtained from a representative sample of Americans between the ages of 6 and 89
(Wechsler, 1999). Raw scores on each subtest were standardized with respect to the
18
subjects’ age, and ultimately converted into T scores (with a population mean of 50 and a
standard deviation of 10). Although the four subtests can yield a summary (full-scale) IQ
score, it was instead decided to create a latent ‘general ability’ factor using the four
indicators. The WASI also yields scores in the verbal and nonverbal domains -Verbal IQ
and Performance IQ, respectively. Verbal IQ is measured by the Vocabulary and
Similarities subtests, and Performance IQ is measured by Block Design and Matrix
Reasoning.
Skin Conductance Activity
Psychophysiological recording occurred in a laboratory where subjects were
seated in a room adjacent to the experimenter. The average temperature was 74 °F. Skin
conductance activity was recorded from bipolar leads on the distal phalanges of the index
and middle fingers using silver-silver chloride electrodes, which were placed on the non-
dominant hand. The conducting medium was K-Y lubricating jelly, surrounded by an
adhesive electrode collar (measuring 7 millimeters in diameter) that maintained full
contact with the skin.
Skin conductance activity was recorded through a bioamplifier manufactured by
the James Long Company (Caroga Lake, NY), with a low-pass filter set to 10 Hz. The
signal was digitized at a sampling rate of 512 Hz. The relevant variable is skin
conductance response (SCR), measured during two tasks: a ‘countdown’ task and an
embarrassing questions task. SCRs were scored using the digitized output from the SCR
channel of the James Long Company hardware. Two methods were used to analyze
SCRs: 1) counting up the number of nonspecific responses, and 2) measuring the
size/magnitude of a specific response (see Dawson, Schell, & Filion, 2007). The former
19
method was used to count the number of skin conductance responses occurring over an
extended period (12 seconds), whereas the latter method was used to calculate the SCR
magnitude to a specific stimulus.
A nonspecific skin conductance response (NS-SCR) was counted whenever a rise
in skin conductance level exceeded .05 microsiemens (μS) and appeared to have a slope
in excess of .05 μS/second. Several undergraduate assistants visually scanned the
waveform output of the SCR channel to determine if this threshold was reached.
Furthermore, video coding procedures were used to ensure that movement artifacts were
not responsible for the appearance of responses. That is, software from the James Long
Company was used to record the exact times of finger movements during the recording
session. If the time of an NS-SCR coincided with movement, then the response was not
considered valid.
The amplitude of a specific response was not scored manually, but was instead
scored through the James Long Company software (SCOR component of the Orientation
Response Analysis System). A rise in skin conductance was judged a response if it
occurred within a window of 1.5 – 4.5 seconds after stimulus onset. Additionally, the
slope in skin conductance level was required to exceed the baseline slope (at stimulus
onset) by a minimum of 0.05 μS/second. SCR amplitude was defined as the peak rise in
skin conductance occurring within seven seconds of response initiation. If no response
was detected, then the magnitude for that particular trial assumed a value of zero rather
than being omitted.
20
Psychophysiological Tasks
Shameful Questions Task
The shameful questions task consisted of 30 questions that were delivered to the
subject through earphones from a recorded female voice. Half of the questions were
neutral (e.g., “Have you ever eaten spaghetti?”), while the other 15 questions were
designed to cause discomfort (e.g., “Have you ever hit your twin?”). Subjects were
instructed to answer each question by selecting the appropriate button on a gadget (“yes”
or “no”). In order to insure that the experimental manipulation was effective, the three
most highly ‘loaded’ questions were selected for analyses: 1) “Have you ever smoked a
cigarette?”, 2) “Have you ever hit your twin?”, and 3) “Have you ever smoked
marijuana?”. The neutral question that immediately followed or preceded each of these
questions was used as a comparison: 1) “Have you ever been on a bus?”, 2) “Have you
ever been on an airplane?”, and 3) “Have you ever watched TV?”. The questions had a
duration of 1-2 seconds. Stimulus onset was defined as the beginning of the question.
The SCR magnitude to each stimulus was calculated using the James Long Company
software.
Many subjects had invalid data for this task due to various technical problems.
The most common problem was a too low volume setting (221 subjects were eliminated
on account of this). Other subjects were invalidated due to excessive movement or the
mishandling of earphones, resulting in 773 valid subjects. The data itself was frequently
invalid when subjects registered negative SCR amplitudes to certain questions. These
negative values were an artifact of the peculiar algorithm that the James Long Company
software employs, in which the skin conductance level (SCL) recorded during the
21
baseline period may in fact be higher than the SCL measured at peak amplitude. (This
can happen if a late-occurring motor response to the previous question interferes with the
baseline period of the present question). All negative values were set to missing. Overall,
38% and 44% of subjects had at least one invalid SCR magnitude for the loaded and
neutral conditions, respectively. As a result, the responses within each condition were
averaged to form a composite SCR measure rather than serving as indicators of a latent
factor. Over 98% of the 773 technically valid subjects had at least one non-missing value.
In other words, less than 2% of the sample had negative (invalid) SCR magnitudes across
all three loaded questions or across all three neutral questions.
SCR magnitudes were square root transformed in order to improve the symmetry
of their distribution. As expected, the mean SCR magnitude to the loaded questions was
larger than that for the neutral questions; t(753) = 7.32; p < .01. This suggests that the
loaded questions were eliciting deeper emotions (e.g., shame, guilt, embarrassment) or at
least triggering greater cognitive arousal. Test-retest reliability was examined in 41
subjects who had valid data across two testing occasions. The test-retest correlation was
calculated using Pearson’s r as well as Spearman’s rho. The latter is a non-parametric
statistic that is insensitive to outliers as well as other data points that can exert inordinate
influence on parametric statistics. Spearman’s rho is particularly useful in small samples,
where chance capitalization or a single outlier can lead to very inaccurate results. The
test-retest correlations were not impressive, but were significantly different from zero for
the loaded questions: Pearson’s r = .33 (p = .04) and Spearman’s rho = .39 (p = .01). For
the neutral questions, the correlations were r = .50 and rho = .55.
22
Countdown Task
In the countdown task, each subject was instructed to sit still and look at a
computer monitor. The subject was told beforehand that numbers would go down from
12 to 0, after which a loud noise would occur. During the countdown trial, each number
(from 12 to 1) was presented for a duration of 1 second. When the countdown reached
zero, a burst of white noise was presented through the earphones at 105 decibels. There
were three signaled trials, in which the 12-second countdowns preceded the delivery of
the white noise. Additionally, there were two unsignaled trials, where the noise blast was
presented without any countdown warning. The signaled and unsignaled trials were
presented in an alternating order. The interstimulus interval was approximately 45
seconds (measured from blast to blast).
Skin conductance data was valid for 747 subjects. The original pool was much
larger, but many subjects were rendered unusable due to the same technical problems that
plagued the questions task. In particular, the noise blast was inadvertently presented at a
sound-pressure level below 105 decibels for many subjects. Additionally, video coding
procedures identified dozens of subjects who were constantly moving their hand or body
during the task. Since movement-related responses cannot be distinguished from
legitimate non-specific SCRs (Dawson et al., 2007), it was decided to exclude these
subjects from analyses.
The number of SCRs during the countdown is thought to serve as an index of
anticipatory fear or anxiety (Hare, Frazelle, & Cox, 1978; Fung et al., 2005). We
obtained ratings from the subjects regarding how anxious they were as the numbers were
counting down. On a 5-point scale (ranging from 1=”not at all” to 5=”extremely”), the
23
mean rating provided by 745 subjects was 2.43. Thus, their subjective responses
indicated that the countdown was not as anxiety-provoking as we would have desired.
Subjects also rated whether the noise was loud, using the same 5-point scale. The mean
response was 3.78, nearly corresponding to “quite a bit” loud.
We counted the number of SCRs during four phases of the task: 1) the twelve
‘countdown’ seconds prior to the noise blast, 2) the twelve seconds following the
signaled blast, 3) the twelve seconds prior to the unsignaled blast, and 4) the twelve
seconds following the unsignaled blast. These four conditions can be referred to as
signaled anticipatory, signaled responsivity, unsignaled anticipatory, and unsignaled
responsivity, respectively. Since there were an unequal number of signaled and
unsignaled trials (3 vs. 2), we averaged the number of SCRs over trials rather than
summing them.
We expected that the signaled anticipatory (i.e., countdown) condition would
elicit more responses than the unsignaled anticipatory condition. The latter is simply a
baseline period during which any SCRs would be purely spontaneous. A t-test on
dependent samples confirmed that subjects were more electrodermally reactive during the
signaled anticipatory trials; t(746) = 8.03, p < .01. Responses during the unsignaled
anticipatory condition were not statistically reliable. The test-retest correlation for 36
subjects was not significant; Pearson’s r = .15 and Spearman’s rho = .06. However, test-
retest correlations were significant for the other three conditions.
Statistical Analyses
All analyses were conducted in Mx (Neale et al., 2006), a structural equation
modeling program. Models were fit to raw data using full information maximum-
24
likelihood techniques. This has the advantage of avoiding listwise deletion of subjects
who are missing data on certain measures. It also allows for subjects to be included in
analyses even if their co-twin is completely missing data. Sex-limitation models were
employed for all analyses, in which the variances/covariances among measures were free
to differ between males and females (Neale & Cardon, 1992).
Standard twin methodology was used to estimate the genetic and environmental
contributions to various phenotypes. Twin studies allow for the disentanglement of
genetic and environmental influences in a population by exploiting the fact that twin pairs
fall into two types: monozygotic (MZ) and dizygotic (DZ). MZ twins are genetically
identical, whereas DZ (i.e., ‘fraternal’) twins share, on average, half of their alleles. Thus,
any excess covariance observed in MZ twins can be attributed to genetic effects. All
analyses were based on an extension of the standard ACE model (Neale & Cardon, 1992).
That is, the total variance of a phenotype as well as the covariance among phenotypes
was partitioned into additive genetic (A), shared environmental (C), and nonshared
environmental (E) components.
The partitioning of trait variance into genetic and environmental components is
based on the disparity in phenotypic similarity between MZ and DZ twin pairs.
Covariance between siblings can be due to additive genetic (A) influences and/or shared
environmental (C) influences. As illustrated in Figure 1, the additive genetic correlation
is specified as 1.0 and 0.5 in MZ and DZ twins, respectively. The latent genetic (A)
factor entails all influences that lead to greater resemblance in MZ twins relative to DZ
twins. The latent shared environmental (C) factor entails all influences that are
obligatorily shared by family members, contributing equally to MZ and DZ similarity.
25
Nonshared environmental (E) influences refer to individual-specific experiences as well
as measurement error, and are thus uncorrelated between twins.
Figure 1. Illustration of an ACE model, in which a phenotype is influenced by three
latent factors – additive genetic (A), shared environmental (C), and nonshared
environmental (E) effects. The genetic correlation is 1 among MZ twins and is 0.5
among DZ twins. Although the path diagram is only shown in its simplest univariate
form, it can readily be generalized to more complex multivariate models.
Measurement Model of Psychopathy
The first aim is to compare a general-specific factor model of psychopathy to a
second-order factor model. This comparison was motivated by the work of Patrick et al.
(2007), who found that a general-specific (bifactor) model provided a better fit to the
PCL-R facets than did a higher-order factor model. General-specific factor models have
several advantages over second-order factor models, especially when a given construct is
composed of several highly related domains (Chen, West, & Sousa, 2006). The general
factor accounts for the considerable overlap among the domains, while the specific
Twin 1
A C E
Twin 2
A C E
1 0.5 (DZ); 1 (MZ)
26
factors account for the unique influence of each domain over and above the general factor.
This differs from second-order models, in which a higher-order factor accounts for the
association among the lower-order domains. Since the lower-order factors are correlated
with one another, the unique contribution of each factor in predicting external variables is
less interpretable.
Figures 2 and 3 illustrate the two models that were fit to psychopathic traits. A
general-specific factor model (Figure 2) was compared to a second-order factor model
(Figure 3). (Figure 1A in the appendix shows an equivalent version of the second-order
factor model, in order to emphasize its higher-order nature). Both models were based on
results from Bezdjian’s (2008) factor analysis of the CPS subscales. Her best-fitting
model was composed of two factors: Charming/Manipulative and Callous/Disinhibited.
27
Figure 2. A general-specific factor model of psychopathy (shown in one twin only).
The curved double-headed arrows represent the error variance for each of the measured
variables. Genetic and environmental influences on the latent psychopathy factors are
depicted in circles. In order to identify the model, the factor variances were set to 1 (i.e.,
the sum of A, C, and E was constrained to equal 1).
General
Psychopathy
Charming/
Manipulative
Callous/
Disinhibited
Glibness
Untruthfulness
Fail to accept
responsibility
Unreliability
Impulsiveness
Callousness
Manipulation
Parasitic
Lifestyle
Behavioral
Dyscontrol
Lack of
Planning
Poverty of
Affect
Boredom
Susceptible
A C E
A C E
A C E
28
Figure 3. A second-order factor model of psychopathy (shown in one twin only). The
curved double-headed arrows represent the error variance for each of the measured
variables. The genetic and environmental contributions to the variance/covariance of the
two factors are signified by arrows pointing from the circles to the ovals. In order to
identify the model, the factor variances were set to 1.
The relative fit of the models was assessed by a likelihood-ratio χ
2
–test, which
depends on the difference between the log likelihood (-2LL) of the two models. Given
that the second-order factor model is nested within the general-specific factor model
Charming/
Manipulative
Callous/
Disinhibited
Glibness
Untruthfulness
Fail to accept
responsibility
Unreliability
Impulsiveness
Callousness
Manipulation
Parasitic
Lifestyle
Behavioral
Dyscontrol
Lack of
Planning
Poverty of
Affect
Boredom
Susceptible
A
C
E
A
C
E
29
(Chen et al., 2006), a χ
2
-difference test was used to determine whether the constraints of
the second-order model led to a significant reduction in fit. In other words, the second-
order factor model was considered acceptable only if the change in χ
2
was trivial. Chen
et al. (2006) have shown that the power to discriminate between the two models is greater
than 0.99 when the sample size is 500. As the present sample contained more than 500
twins of both sexes, the power to differentiate between these models was excellent.
After selecting the appropriate model, it was necessary to determine whether the
Childhood Psychopathy Scale was measurement invariant with respect to sex. The first
step is to test for pattern invariance (Meredith, 1993). That is, are the factor loadings
similar for males and females? If the factor loadings cannot be constrained to be equal
between the sexes, then all subsequent analyses must be conducted separately in boys and
girls. This is because the construct of psychopathy, as interpreted by caregivers, may be
different for boys and girls.
Finally, a simplification of the etiology of psychopathy was undertaken. Based on
previous work by Bezdjian (2008), it was expected that shared environmental influences
on psychopathy would be negligible. Reduced models were tested, in which the C
parameters were dropped from the full ACE model. The optimal submodel was selected
on the basis of Akaike’s Information Criterion (AIC; Akaike, 1987), which is calculated
as -2LL (log likelihood) minus twice its degrees of freedom. Models with a more
negative AIC are superior, as they represent a better balance of parsimony and goodness
of fit.
30
Etiological Overlap with Criterion Variables
The phenotypic association between each psychopathy factor and a given criterion
variable was estimated from the genetic and environmental covariances. Figure 4
depicts a structural model in which genetic/environmental influences on all three
psychopathy factors covary with the latent criterion variable. (This model is an extension
of the general-specific framework, as it was assumed to be superior to the higher-order
factor structure). The a
44
, c
44
, and e
44
paths represent the genetic/environmental variances
that are specific to the criterion variable (i.e., not explained by psychopathy). Table 1
provides details on how the genetic and environmental parameters were standardized.
Heritability refers to the proportion of phenotypic (observed) variance that is attributable
to additive genetic factors. Standardized estimates of the variance components were
obtained for each criterion variable via this structural model, and are reported in Table 7.
The genetic correlation (r
g
) between a psychopathy factor and a criterion variable
is obtained by standardizing the genetic covariance (a
41
, a
42
, or a
43
) with respect to the
total genetic variance of the criterion variable. It represents the extent to which genetic
influences are common to both variables. However, estimates of r
g
are unstable if the
heritability of one or both variable(s) is weak. A more meaningful statistic is the genetic
contribution to the phenotypic correlation, which is a function of r
g
as well as the
heritabilities of both variables. It represents the proportion of the phenotypic correlation
that is attributable to genetic effects. Similarly, the nonshared environmental
contribution to the phenotypic association is a function of the environmental correlation
(r
e
) as well as the ‘environmentality’ of both variables.
31
Figure 4. Path model depicting the structural associations between the CPS factors and a
latent criterion variable (with three indicators). For simplicity as well as empirical
considerations, there are no shared environmental (C) influences on the psychopathy
factors. Genetic (A) and nonshared environmental (E) influences are free to covary
between each psychopathy factor and the criterion variable.
Psychopathy
Charming/
Manipulative
Callous/
Disinhibited
Glibness
Untruthfulness
Fail to accept
responsibility
Unreliability
Impulsiveness
Callousness
Manipulation
Parasitic
Lifestyle
Behavioral
Dyscontrol
Lack of
Planning
Poverty of
Affect
Boredom
Susceptible
Measure Y
Measure Z
External
Criterion
A1
A4
C4
E4
E1
Measure X
e
41
e
44
a
41
a
44
c
44
e
11
a
11
A2
E2
E3
A3
a
42
e
42
a
43
e
43
a
33
e
33
a
22
e
22
32
Table 1. Formulae for calculating and standardizing the genetic and environmental
parameters of the general-specific factor model, as depicted in Figure 4
Notes: H
gen
= heritability of the general psychopathy factor; H
cm
= heritability of the charming/manipulative factor; H
cd
=
heritability of the callous/disinhibited factor; H
crit
= heritability of the criterion variable
General Factor Charming/
Manipulative
Callous/
Disinhibited
Criterion
Phenotypic
Variance
a
11
2
+ e
11
2
a
22
2
+ e
22
2
a
33
2
+ e
33
2
a
44
2
+ c
44
2
+ e
44
2
+ a
41
2
+
e
41
2
+ a
42
2
+ e
42
2
+ a
43
2
+ e
43
2
Heritability (H) a
11
2
/( a
11
2
+ e
11
2
) a
22
2
/( a
22
2
+ e
22
2
) a
33
2
/( a
33
2
+ e
33
2
) (a
41
2
+ a
42
2
+ a
43
2
+ a
44
2
)/
Criterion Variance
Environmentality (E) e
11
2
/( a
11
2
+ e
11
2
) e
22
2
/( a
22
2
+ e
22
2
) e
33
2
/( a
33
2
+ e
33
2
) (e
41
2
+ e
42
2
+ e
43
2
+ e
44
2
)/
Criterion Variance
Genetic correlation (r
g
) a
41
2
* a
11
2
/
[a
11
2
(a
41
2
+ a
44
2
)]
1/2
a
42
2
* a
22
2
/
[a
22
2
(a
42
2
+ a
44
2
)]
1/2
a
43
2
* a
33
2
/
[a
33
2
(a
43
2
+ a
44
2
)]
1/2
Environmental
Correlation (r
e
)
e
41
2
* e
11
2
/
[e
11
2
(e
41
2
+ e
44
2
)]
1/2
e
42
2
* e
22
2
/
[e
22
2
(e
42
2
+ e
44
2
)]
1/2
e
43
2
* e
33
2
/
[e
33
2
(e
43
2
+ e
44
2
)]
1/2
Genetic contribution √H
gen
* r
g
*
√H
crit
√H
cm
* r
g
*
√H
crit
√H
cd
* r
g
*
√H
crit
Environmental
Contribution
√E
gen
* r
e
*
√E
crit
√E
cm
* r
e
*
√E
crit
√E
cd
* r
e
*
√E
crit
33
Chapter Three: Results
Descriptive Statistics
The means and standard deviations of each CPS subscale are listed in Table 2.
Data was available for the entire sample of 1210 subjects: 591 males and 619 females.
Because observations are correlated within same-sex twin pairs, it is not appropriate to
conduct a t-test on independent samples. Instead, mean sex differences were evaluated in
Mx using maximum likelihood testing (Neale et al., 2006). Univariate sex-limitation
ACE models were fitted to each subscale, and a chi-square difference test determined
whether the means could be constrained to be equal across sex. Males received
significantly higher scores on all but three of the subscales (Untruthfulness, Manipulation,
and Poverty of Affect).
Table 2. Means and standard deviations of the CPS subscales
Subscale Male Mean Female Mean
χ²
Glibness 1.41 ± 0.29 1.37 ± 0.26 6.11*
Untruthfulness 1.18 ± 0.22 1.16 ± 0.20 3.07
Failure to accept
Responsibility
1.46 ± 0.31 1.41 ± 0.32 7.08*
Manipulation 1.28 ± 0.30 1.25 ± 0.30 1.38
Parasitic Lifestyle 1.18 ± 0.20 1.15 ± 0.20 5.64*
Unreliability 1.13 ± 0.23 1.08 ± 0.18 15.22**
Behavioral
Dyscontrol
1.40 ± 0.40 1.34 ± 0.38 6.40*
Impulsiveness 1.34 ± 0.33 1.24 ± 0.29 25.29**
Lack of Planning 1.31 ± 0.34 1.21 ± 0.29 28.07**
Callousness 1.12 ± 0.18 1.08 ± 0.15 14.75**
Poverty of Affect 1.18 ± 0.16 1.18 ± 0.15 0.69
Boredom
Susceptibility
1.29 ± 0.24 1.21 ± 0.22 31.31**
* Mean sex difference is significant at p < .05
**Mean sex difference is significant at p < .01
34
Table 3. Means and standard deviations of the criterion measures
Measure Males Females
χ²
N Mean N Mean
Conduct Disorder
symptoms
587 1.45 ± 2.02 611 .90 ± 1.54 23.86**
Reactive aggression 591 1.71 ± 3.32 619 1.64 ± 3.29 10.41**
Proactive aggression 591 1.11 ± 1.72 619 1.08 ± 1.38 5.88*
CBCL Attention Problems 586 3.25 ± 3.25 618 2.23 ± 2.87 31.32**
Inattention symptoms 587 3.60 ± 3.20 607 2.37 ± 2.74 52.49**
Hyperactivity symptoms 587 2.81 ± 2.84 607 1.99 ± 2.51 25.78**
CBCL Anxious/Depressed 586 2.77 ± 3.07 618 3.04 ± 3.25 2.57
Generalized Anxiety
Disorder symptoms
588 2.81 ± 2.49 608 2.85 ± 2.44 0.14
Major Depression
symptoms
588 3.27 ± 3.07 607 3.21 ± 3.07 0.23
Vocabulary T-score 590 49.56 ± 9.54 617 49.80 ± 10.08 .09
Similarities T-score 590 51.07 ± 9.77 618 52.62 ± 9.67 8.14**
Block Design T-score 591 48.86 ± 10.02 617 47.08 ± 9.66 6.87*
Matrix Reasoning T-score 591 52.29 ± 10.13 617 52.39 ± 10.15 .052
SCRs during signaled
anticipation (countdown)
340 .90 ± .82 407 .64 ± .75 17.45**
SCRs during unsignaled
anticipation
340 .55 ± .68 407 .52 ± .63 0.11
SCRs after signaled blast 340 .99 ± .66 407 .87 ± .60 4.30*
SCRs after unsignaled blast 340 .93 ± .65 407 .86 ± .64 1.74
SCR magnitude to
shameful questions
358 .88 ± .61 402 .77 ± .58 3.78
SCR magnitude to neutral
questions
360 79 ± .62 399 .70 ± .59 2.06
** Mean sex difference is significant at p < .01
* Mean sex difference is significant at .01 < p < .05
Notes: CBCL = Child Behavior Checklist; SCRs = mean number of skin conductance
responses; SCR magnitude is measured in microsiemens (μS)
35
The number of males and females with valid data for each criterion measure is
presented in Table 3. The means and standard deviations are also listed for each sex. As
expected, male received significantly higher endorsements of aggression, conduct
disorder symptoms, and ADHD symptoms. There were no significant sex differences in
anxiety or depression. Intellectual performance was similar between the sexes, although
females scored higher on a verbal subtest (Similarities), and males scored higher on a
nonverbal subtest (Block Design). For the skin conductance measures, the general
pattern was that males were more responsive than females. Males produced significantly
more SCRs during the countdown trials, and were more responsive to the unsignaled
noise blasts. Additionally, they tended to be more electrodermally reactive to the
shameful and neutral questions, although neither of these differences reached statistical
significance.
Measurement Model of Psychopathy
In order to determine the optimal measurement model, the fit of the general-
specific factor model was compared to that of the second-order factor model. (Since the
latter is nested within the former, the two can be directly compared using a chi-square
difference test). Fit statistics are reported in Table 4. The second-order factor model led
to a significant loss of fit; Δχ²(22) = 359.74, p < .01. Furthermore, it yielded a high
correlation between the two factors: r = 0.72 in males and r = 0.68 in females. These
results indicate that a single factor directly contributes to the covariance among the
subscales. Thus, the constraints of the second-order factor model are too strict.
36
Table 4. Model-fitting results
Model
Comparison
CPS Model -2LL df AIC
Δχ²
Δdf pΔ
1 General-specific
factor model
63687.86 14412 34863.86
2 vs. 1 Higher-order factor
model
64047.60 14434 35179.60 359.74 22 < .01
3 vs. 1 Pattern Invariance
♂=♀
63777.85 14436 34905.85 89.99 24 < .01
4a vs. 1 No C in males 63687.86 14415 34857.86 0 3
4b vs. 1 No C in females 63691.32 14415 34681.32 3.52 3 0.32
4c vs. 1 No C in both sexes 63691.38 14418 34855.38 3.52 6 0.74
Notes: -2LL = -2 log likelihood; df = degrees of freedom; AIC = Akaike’s Information
Criterion; C = shared environmental influences; best-fitting model is highlighted in bold
As can be seen in Model 3 of Table 4, measurement invariance was not tenable
with respect to sex. Constraining the factor loadings to be equal across sex led to a
significant loss of fit; Δχ²(24) = 89.99, p < .01. The standardized factor loadings of the
general-specific model are presented in Table 5. Most subscales had substantial loadings
on the general factor. Notable exceptions include Glibness and Manipulation, which
loaded more strongly on the Charming/Manipulative factor than on the general factor for
both males and females. In addition, Lack of Planning had a much greater loading on the
Callous/Disinhibited factor (0.79 - 0.98) than on the general factor (0.15 - 0.23).
Callousness loaded very poorly on the supposed “Callous/Disinhibited” factor, prompting
me to relabel it “Nonplanning”.
37
The fact that the second specific factor was essentially captured by a single
subscale (Lack of Planning) raises the possibility that this dimension is not legitimately
distinct from the general factor. This concern was particularly relevant in the case of
females, where Lack of Planning had a loading of 0.98, and the other factor loadings
never exceeded 0.21 (including a negative loading for Behavioral Dyscontrol). However,
the elimination of this Nonplanning factor in females led to a significant reduction in fit;
Δχ²(8) = 129.41; p < .01 . Apparently, there is a non-trivial amount of common residual
variance among the 7 subscales that comprise this factor. Even an attempt to drop the
four indicators with the lowest loadings – Behavioral Dyscontrol, Callousness, Poverty of
Affect, and Boredom Susceptibility – was not successful; Δχ²(4) = 13.66; p = .01.
Table 5. Standardized factor loadings from the general-specific factor model in males
(and females in parentheses)
CPS Subscale General Factor Charming/
Manipulative
Nonplanning
Glibness .29 (.26) .37 (.42)
Untruthfulness .64 (.48) .14 (.39)
Failure to accept
responsibility
.54 (.42) .16 (.46)
Manipulation .26 (.39) .96 (.56)
Parasitic Lifestyle .66 (.55) .18 (.33)
Unreliability .56 (.51) .31 (.21)
Behavioral Dyscontrol .49 (.62) .12 (-.05)
Impulsiveness .60 (.60) .32 (.14)
Lack of Planning .23 (.15) .79 (.98)
Callousness .57 (.54) .16 (.02)
Poverty of Affect .48 (.58) .23 (.08)
Boredom Susceptibility .57 (.57) .16 (.09)
38
Shared environmental influences on the three factors were non-existent in
males, which is not surprising given that there is a lack of twin resemblance for
psychopathic traits among DZ males (Bezdjian, 2008). The role of shared environmental
factors was small and non-significant in females. Ultimately, the model could be
simplified by dropping C from both males and females; Δχ²(6) = 3.52, p = 0.74 (see
Model 4c in Table 4). This submodel had the lowest AIC value, and was used as the
basis for all subsequent analyses. Standardized estimates of the genetic and
environmental parameters from this best-fitting AE model are presented in Table 6. The
heritability of the general factor was similar in males and females: 73% and 72%,
respectively. Heritability estimates for the two specific factors ranged from 56% to 97%.
Although nonshared environmental influences were non-significant in two instances, the
E parameter was always retained; it is implausible for a psychological construct to be
100% heritable, even when taking into account the elimination of measurement error.
Table 6. Standardized estimates of the genetic/environmental variance components
(95% confidence intervals in parentheses)
CPS Factor Males Females
A E A E
General Factor .73
(.57, .84)
.27
(.16, .43)
.72
(.58, .83)
.28
(.17, .42)
Charming/
Manipulative
.62
(.49, .82)
.38
(.18, .51)
.87
(.64, 1.00)
.13
(0, .36)
Nonplanning .97
(.65, 1.00)
.03
(0, .35)
.56
(.46, .82)
.44
(.18, .54)
Notes: A = additive genetic effects (heritability); E = nonshared environmental effects
39
Table 7. Standardized estimates of the genetic/environmental variance components for
each criterion variable
Criterion Males Females
A C E A C E
Antisocial Behavior
.58* .15* .27* .61* .17* .22*
ADHD
.78* .00 .22* .66* .00 .34*
Anxious/Depressed
.68* .05 .27* .65* .02 .33*
IQ
.53* .41* .06 .36* .63* .01
Signaled Anticipatory
SCRs
.05 .35* .59* .08 .33* .59*
Unsignaled
Anticipatory SCRs
.13 .12 .75* .14 .19* .67*
Signaled Responsivity
SCRs
.48* .04 .48* .40* .08 .52*
Unsignaled
Responsivity SCRs
.08 .13 .78* .21* .10 .69*
SCR magnitude to
shameful questions
.01 .36* .63* .04 .19* .77*
SCR magnitude to
neutral questions
.27* .13 .60* .35* .06 .59*
*p < .05
Notes: A = Additive genetic effects; C = shared environmental effects; E = nonshared
environmental effects
Construct Validity
Associations with Antisocial Behavior
Measures of antisocial behavior (ASB) were next incorporated into the extended
structural model (as shown in Figure 4). The psychopathy factors were permitted to
simultaneously predict antisocial behavior. ASB was a latent variable composed of
reactive aggression, proactive aggression, and Conduct Disorder symptom counts. This
was considered a coherent construct, as the standardized factor loadings exceeded 0.65.
40
Specifically, the factor loadings for reactive aggression, proactive aggression, and
conduct disorder symptoms were 0.72, 0.74, and 0.67 in males; the respective loadings
were 0.79, 0.80, and 0.65 in females.
The genetic and environmental variance components of the latent ASB factor are
reported in Table 7. Genetic influences explained 58% to 61% of the total variance.
Although shared environmental influences were modest (15 – 17%), they could not be
dropped without a significant reduction in fit; Δχ² (2) = 30.3, p < .01. Shared
environmental influences were specific to the ASB factor, given that no evidence of C
was found for psychopathy.
The phenotypic correlation between the general factor and ASB was very high
(r > .80) in both sexes (see Table 8). More than two-thirds of the phenotypic correlation
could be attributed to genetic effects. The correlation between Charming/Manipulative
(C/M) and ASB was less strong, but still significant. In girls, the genetic contribution
(0.31) to the phenotypic correlation was complete. In both sexes, there was no
association between Nonplanning and ASB. The general factor and C/M combined to
explain 100% of the ASB nonshared environmental variance. For the general factor
alone, the environmental correlation (r
e
) approached unity. Moreover, the general and
C/M factors combined to explain nearly 100% of the ASB genetic variance in both males
and females. The genetic variance specific to the latent ASB factor was small and non-
significant across both sexes; Δχ²(2) = 1.40, p = .50. In other words, the genetic
influences on ASB were common to psychopathy.
41
Table 8. Genetic and environmental contributions to the association between
psychopathy and antisocial behavior (ASB)
CPS Factor Males Females
r Gen
contrib
Env
contrib
r
g
r
e
r Gen
contrib
Env
contrib
r
g
r
e
General .82* .55 .26 .86* .94* .80* .55 .26 .84* .98*
C/M .23* .14 .08 .24* .27 .31* .31 .00 .44* .00
Nonplanning .02 .06 -.04 .08 -.21 -.04 .02 -.06 .03 -.20
*p < .05
Notes: C/M = Charming/Manipulative; r = phenotypic correlation; r
g
= genetic correlation;
r
e
= environmental correlation
An important qualification is that much of the genetic overlap may have stemmed
from the overlapping content between the CPS and ASB measures. Several items from
the Behavioral Dyscontrol, Untruthfulness, and Callousness subscales were ‘tainted’ with
antisocial behavior. As a result, the associations between psychopathy and ASB were re-
analyzed after removing these 3 subscales from the entire model. The genetic correlation
between the general factor and ASB declined from 0.86 to 0.71 in males, and from 0.84
to 0.71 in females. Even so, the hypothesis that psychopathy explains all of the ASB
genetic variance was still not rejected; Δχ²(2) = 5.15, p = .08.
In order to illustrate the clinical significance of psychopathic-like traits, an
extreme-groups analysis was conducted on the general CPS factor in relation to conduct
disorder. The twelve CPS subscales were subjected to principal axis factoring in order to
extract factor scores for each subject. Subjects were ranked into upper and lower
quartiles on the basis of their regression scores on the first (general) factor. Not
surprisingly, none of the 301 subjects in the lowest quartile had ever met criteria for
conduct disorder (CD), whereas 27 out of 298 subjects in the upper quartile had a lifetime
42
diagnosis of CD. In other words, of the 30 subjects with a CD diagnosis, all but 3 had
psychopathy scores in the upper quartile. Ranking subjects’ scores into more extreme
groups only served to increase sensitivity. Of the 57 subjects who scored in the top 5%,
16 had a CD diagnosis. Put another way, more than half of the children with CD
diagnoses were represented amongst the top 5% of scorers.
Associations with ADHD
Symptom counts of Inattention and Hyperactivity-Impulsivity from the DISC-IV
were supplemented with scores on the CBCL Attention Problems scale. This allowed the
formation of an ADHD latent factor. The three scores were highly intercorrelated.
Standardized factor loadings for Inattention, Hyperactivity-Impulsivity, and Attention
Problems were 0.84, 0.70, and 0.76 in males. The factor loadings in females were 0.76,
0.71, and 0.79, respectively. The heritability of the latent ADHD factor was 78% in
males and 66% in females. The remainder of the variance could be attributed to
nonshared environmental influences.
The correlation between the General factor and the ADHD latent factor was high
in both sexes (see Table 9). The overlap was mainly due to genetic contributions. The
phenotypic correlation between C/M and ADHD was negligible in both sexes, although
the genetic and environmental factors appeared to be operating in opposing directions in
females. The association between Nonplanning and ADHD was much higher in males (r
= .57) than in females (r = 0.16). The correlation was due to genetic as well as
environmental contributions in males, whereas the correlation was mainly explained by
an environmental contribution in females. The genetic influences on psychopathy were
43
largely in common with ADHD. As a result, the genetic variance specific to the latent
ADHD factor was small and non-significant across both sexes; Δχ²(2) = 5.76, p = .06.
Table 9. Genetic and environmental contributions to the association between
psychopathy and ADHD
CPS Factor Males Females
r Gen
contrib
Env
contrib
r
g
r
e
r Gen
contrib
Env
contrib
r
g
r
e
General .62* .51 .11 .65* .50* .77* .57 .20 .83* .65*
C/M .06 .03 .02 .04 .10 .03 .19 -.16 .25 -.69
Nonplanning .57* .35 .22 .47* .86* .16* .05 .11 .07 .32*
*p < .05
Notes: C/M = Charming/Manipulative; r = phenotypic correlation; r
g
= genetic correlation;
r
e
= environmental correlation
The high genetic overlap may be due to the overlapping content between the CPS
items and ADHD symptoms. Two of the CPS subscales – Impulsiveness and Boredom
Susceptibility – are intimately related to symptoms of Inattention and Hyperactivity.
Additionally, Unreliability contains an item tapping into the inability to complete tasks,
which is a symptom of Inattention. Analyses were rerun after omitting these three
subscales from the model. The genetic correlations with respect to the general factor
slightly decreased to 0.62 in males, and dropped from 0.83 to 0.72 in females. The
decline was most dramatic for the Nonplanning factor in males, with r
g
dropping from
0.47 to 0.21. Consequently, the hypothesis that psychopathy entirely accounts for the
genetic influences on ADHD was rejected; Δχ² (2) = 29.53, p < .01.
Associations with Anxiety/Depression
A latent internalizing factor was created using scores on the Anxious/Depressed
CBCL scale as well as symptom counts of Generalized Anxiety Disorder and Major
44
Depression. The corresponding factor loadings were .67, .66, and .84 in males
and .73, .72, and .73 in females. Approximately two-thirds of the phenotypic variance
was due to genetic influences; the remainder was due to nonshared environmental factors.
Table 10. Genetic and environmental contributions to the association between
psychopathy and anxiety/depression
CPS Factor Males Females
r Gen
Contrib
Env
contrib
r
g
r
e
r Gen
contrib
Env
contrib
r
g
r
e
General .41* .33 .08 .47* .28 .62* .51 .11 .75* .36*
C/M .10 .11 -.01 .17 -.02 -.10 .12 -.22 .16 -.92*
Nonplanning .12 .13 -.01 .16 -.90 -.07 -.12 .06 -.20 .16
*p < .05
Notes: C/M = Charming/Manipulative; r = phenotypic correlation; r
g
= genetic correlation;
r
e
= environmental correlation
The phenotypic correlation between the general factor and anxiety/depression was
moderate in males (r = .41) and somewhat high in females (r = .62). As can be seen in
Table 10, genetic effects contributed to most of the correlation. Phenotypic correlations
with respect to the specific factors were all non-significant. Unlike the case for ASB and
ADHD, a significant amount of genetic variance was specific to anxiety/depression;
Δχ²(2) = 6.69, p = .04. However, this conclusion may only apply to males. The genetic
correlation in males was not high (r
g
= .47), whereas it appeared to approach unity in
females (r
g
= .75). When tested in females alone, the specific genetic influences were not
significant; Δχ² = 3.04, p = .08.
Associations with Intelligence
General intellectual ability was indexed by the four WASI subtests: Vocabulary,
Similarities, Block Design, and Matrix Reasoning. The verbal subtests loaded on the
45
general intelligence factor to a greater extent than did the nonverbal subtests. Factor
loadings for Vocabulary and Similarities ranged from 0.76 to 0.83, whereas loadings for
Block Design and Matrix Reasoning ranged from .46 to .59. The latent factor was very
successful at eliminating measurement error; nonshared environmental influences on IQ
were negligible. Genetic influences were relatively more important than shared
environmental influences in males (53% vs. 41%), while the reverse was true for females
(36% vs. 63%).
As expected, there was a negative correlation between the general psychopathy
factor and IQ. The phenotypic correlation was significant in males (r = -.19), but did not
reach statistical significance in females (r = -.10). The correlation between C/M and IQ
was negligible in both sexes. However, there was an inverse association between
Nonplanning and IQ, which was again stronger in males (see Table 11).
Table 11. Genetic and environmental contributions to the association between
psychopathy and IQ
CPS Factor Males Females
r Gen
contrib
Env
contrib
r
g
r
e
r Gen
contrib
Env
contrib
r
g
r
e
General -.19* -.19 -.01 -.30* -.04 -.10 -.14 .04 -.27* .69
C/M .03 .02 .02 .03 .11 .07 .07 .01 .17 .10
Nonplanning -.25* -.25 .00 -.34* -.92 -.15* -.12 -.03 -.24* -.47
*p < .05
Notes: C/M = Charming/Manipulative; r = phenotypic correlation; r
g
= genetic correlation;
r
e
= environmental correlation
For exploratory purposes, analyses were run separately with respect to Verbal IQ
and Performance IQ. Among adolescents, antisocial behavior is typically more related to
Verbal IQ than Performance IQ (Lynam, Moffitt, & Stouthamer-Loeber, 1993), although
46
this pattern does not necessarily hold for preadolescent children (see Isen, 2010). The
phenotypic correlations between each of the psychopathy factors and VIQ/PIQ are
reported in Table 12. No evidence of a differential association emerged in either sex.
The correlations were somewhat stronger in males than females, particularly with respect
to the Nonplanning factor.
Table 12. Phenotypic correlations between psychopathy and Verbal/Performance IQ
CPS factor Males Females
Verbal IQ Performance IQ Verbal IQ Performance IQ
General -.16 -.19 -.12 -.17
Manipulative .03 .01 .06 .12
Nonplanning -.21 -.14 -.07 .01
Associations with Anticipatory SCRs
Electrodermal responsiveness was measured as the average number of SCRs
produced during various trials of the countdown task. There were two anticipatory
conditions: signaled and unsignaled. As can be seen in Table 7, SCRs during the
signaled anticipatory condition were not heritable. Rather, about a third of the variance
(33%-35%) was due to shared environmental factors, and the remainder was due to
nonshared environmental effects/error. It is therefore not surprising that the psychopathy
factors failed to associate with anticipatory SCRs in all instances (see Table 13).
47
Table 13. Genetic and environmental contributions to the association between
psychopathy and the frequency of anticipatory skin conductance responses
CPS Factor Males Females
r Gen
contrib
Env
contrib
r
g
r
e
r Gen
contrib
Env
contrib
r
g
r
e
General -.10 -.03 -.07 -.17 -.16 .05 .12 -.07 .50 -.17
C/M -.05 -.13 .07 -.69 .15 .01 -.04 .05 -.17 .20
Nonplanning -.02 .04 -.06 .18 -.43 -.05 .03 -.09 .16 -.17
Notes: C/M = Charming/Manipulative; r = phenotypic correlation; r
g
= genetic correlation;
r
e
= environmental correlation
The covariation between psychopathy and anticipatory SCRs during the
unsignaled trials was also examined. These SCRs can essentially be treated as
spontaneous fluctuations or nonspecific responses (NS-SCRs). Consistent with the
unreliability of this measure, there was little twin resemblance for NS-SCRs. Nonshared
environmental factors accounted for 75% and 67% of the phenotypic variance in males
and females, respectively. It appears, then, that NS-SCRs during the unsignaled trials do
not have any meaningful, trait-like properties. Suffice it to say, there were no significant
correlations between psychopathy and NS-SCRs (coefficients ranged from -.02 to .06;
data not shown).
Associations with SCRs to White Noise
The two responsivity conditions both followed presentations of white noise, but
showed a different genetic/environmental structure. For the signaled responsivity
condition, SCRs were moderately heritable in both sexes. However, they were not
related to psychopathy (see Table 14). For the unsignaled condition, genetic influences
on SCRs were weak and did not reach statistical significance in males. Nevertheless, the
general psychopathy factor was negatively correlated with SCRs in males (r = -0.18; see
48
Table 15), but bore no association in females. None of the correlations with respect to
the specific factors were significant.
Table 14. Genetic and environmental contributions to the association between
psychopathy and electrodermal responsiveness to signaled white-noise bursts
CPS Factor Males Females
r Gen
contrib
Env
contrib
r
g
r
e
r Gen
contrib
Env
contrib
r
g
r
e
General -.04 -.01 -.04 -.01 -.10 -.02 -.07 .05 -.07 .13
C/M -.10 -.05 -.05 -.08 -.13 -.01 .03 -.04 .03 -.16
Nonplanning .10 .07 .03 .10 .92 -.10 -.11 .02 -.11 .03
Notes: C/M = Charming/Manipulative; r = phenotypic correlation; r
g
= genetic correlation;
r
e
= environmental correlation
Table 15. Genetic and environmental contributions to the association between
psychopathy and electrodermal responsiveness to unsignaled white-noise bursts
CPS Factor Males Females
r Gen
contrib
Env
contrib
r
g
r
e
r Gen
contrib
Env
contrib
r
g
r
e
General -.18* -.02 -.16 -.08 -.35* -.04 -.04 .00 -.10 .00
C/M .02 -.03 .05 -.15 .09 .02 -.03 .05 -.07 .18
Nonplanning .10 .08 .02 .27 .50 -.11 -.07 -.03 -.22 -.06
*p < .05
Notes: C/M = Charming/Manipulative; r = phenotypic correlation; r
g
= genetic correlation;
r
e
= environmental correlation
Associations with SCR Magnitudes
The genetic/environmental structure of the SCR magnitudes differed for the
shameful and neutral conditions. The SCR magnitudes elicited by the shameful questions
were not heritable, although there was evidence of shared environmental influences in
males (C = 36%) and females (C = 19%). Genetic influences on SCRs to the neutral
questions were significant in both sexes. There was no association between any of the
49
psychopathy factors and SCR magnitudes in females. In males, the C/M factor was
negatively (and very modestly) correlated with SCR magnitude to the shameful questions,
but otherwise the psychopathy factors were not associated with either SCR measure.
(Results are only shown for the SCRs to shameful questions; see Table 16).
Table 16. Genetic and environmental contributions to the association between
psychopathy and mean SCR magnitude to shameful questions
CPS Factor Males Females
r Gen
contrib
Env
contrib
r
g
r
e
r Gen
contrib
Env
contrib
r
g
r
e
General -.07 .00 -.07 -.05 -.16 .06 .02 .04 .11 .08
C/M -.14* -.05 -.09 -.63 -.19 .04 -.01 .05 -.06 .15
Nonplanning -.09 -.08 -.01 -.76 -.09 -.02 -.02 .00 -.14 .00
*p < .05
Notes: C/M = Charming/Manipulative; r = phenotypic correlation; r
g
= genetic correlation;
50
Chapter Four: Discussion
General-Specific Factor Model
The goal of this dissertation was to develop an optimal measurement model of
psychopathy in preadolescent children, and to evaluate it vis-à-vis a nomological network
of constructs related to psychopathy. Most of the traits measured by the Childhood
Psychopathy Scale (CPS) were highly interrelated, and thus influenced by a general
factor. Based on previous work by Bezdjian (2008), two specific factors were modeled:
Charming/Manipulative and Callous/Disinhibited. Relative to a higher-order model with
two correlated factors, a general-specific factor model yielded a better fit to the CPS data.
Glibness and Manipulation had higher loadings on the specific factor than on the
general factor, indicating that ‘Charming/Manipulative’ is an appropriate label for the
interpersonal facet of psychopathy. However, the “Callous/Disinhibited” factor appeared
to be a misnomer. Given that Callousness and various disinhibitory traits (e.g., Boredom
Susceptibility and Impulsiveness) loaded rather poorly on this factor, it was renamed to
Nonplanning.
A genetically informative twin design was used to shed light on the nature of the
psychopathy construct. Consistent with previous research conducted on the present twin
sample (Bezdjian, 2008; Baker et al., 2007), caregiver-reported psychopathic traits were
highly heritable in males as well as females. Shared environmental influences on
psychopathy were non-existent or negligible in both sexes. Despite the similar
heritability estimates across the sexes, the construct was not measurement invariant with
respect to sex. That is, pattern invariance was not tenable.
51
Construct Validity
The construct validity of psychopathy was assessed using several criterion
variables, including antisocial behavior, ADHD symptoms, anxiety/depression, skin
conductance responsivity, and IQ. Based on the structural models of Patrick et al. (2007),
it was hypothesized that the general psychopathy factor would be indicative of
externalizing psychopathology and other negative outcomes, whereas the specific
interpersonal factor would be associated with positive adjustment. Support was found for
the former, but not the latter.
Antisocial Behavior
Most investigators have demonstrated that common genetic influences account for
the overlap between psychopathy and antisocial behavior (Larsson et at., 2007; Waldman
& Rhee, 2006). The present results support the hypothesis that psychopathy and
antisocial behavior (ASB) covary due to common genetic influences. Genetically
speaking, the two constructs were indistinguishable. There was also complete overlap
between the nonshared environmental influences on the two constructs. Shared
environmental factors were present for ASB, but they were relatively minor – explaining
less than 17% of the total variance. Consequently, the phenotypic correlation between
the general psychopathy factor and ASB was very high (r > .80) in both males and
females.
One interpretation of the high genetic correlation between the general
psychopathy factor and ASB is that psychopathic personality traits predispose children
towards the development of aggression and conduct problems. For example, an
52
impulsive child who shows little interest in others’ feelings may be predisposed towards
acting in a hurtful or aggressive manner. The same child may constantly cheat and lie
because he/she is very verbally persuasive and charming. In other words, the personality
traits direct the child towards antisocial behavior, without actually encompassing the
behaviors in question.
Alternatively, the high correlation may be a spurious product of the overlapping
content between instruments. Although the CPS was revised in order to de-emphasize
blatant antisocial behaviors (Lynam et al., 2005), some items do in fact measure aspects
of reactive and proactive aggression. For example, the Behavioral Dyscontrol subscale
contains items pertaining to temper problems and low frustration tolerance. Callousness
contains a pair of items about acting mean and teasing/picking on others. Untruthfulness
obviously taps into the habit of lying, which is a symptom of Conduct Disorder. Due to
this contamination, the association between psychopathy and ASB was re-examined after
removing these subscales from the model. Remarkably, the purer psychopathy
dimensions still accounted for most of the genetic influences on ASB.
These results suggest that there is little usefulness in treating antisocial behavior
and psychopathy as separate entities. In essence, this was the perspective taken by Baker
et al. (2007); a single ASB dimension was created by combining measures of
psychopathy, conduct problems, and aggression. It appears that psychopathy broadly
captures an externalizing dimension, even when the former is denuded of blatantly
antisocial behaviors. This is consistent with the results of Patrick et al. (2007), who
53
demonstrated that the common variance of the PCL-R items is synonymous with the
externalizing spectrum of psychopathology.
In the present study, the phenotypic association between psychopathy and ASB
was not unity, owing to the fact that shared environmental influences were specific to
ASB (explaining 15% and 17% of the variance in males and females, respectively).
Shared environmental influences on ASB are commonly observed in children and young
adolescents, but appear to fade into obscurity by early adulthood (Burt et al., 2007). The
presence of shared environmental factors may stem from the close proximity of twins via
a similar rearing environment. Co-twins may cooperate with one another when engaging
in antisocial behavior. If not deliberately in ‘cahoots’, they are likely to be in close
proximity with one another, which may lead caregivers to ascribe shared responsibility
for the behavior (even if only one twin was truly at fault). Thus, observations of ASB in
one twin may be shaped by the extent to which his/her cotwin has externalizing problems.
Twin resemblance was substantial for each of the ASB measures in MZ as well as DZ
pairs, as demonstrated by the intraclass twin correlations (see appendix).
ADHD
The correlation between the general psychopathy factor and ADHD symptoms
was remarkably high in females. In males, both the general and Nonplanning factors
were strongly associated with ADHD symptoms. In combination, they explained over
70% of the ADHD variance. For both sexes, all of the genetic influences on the ADHD
latent factor could be explained by psychopathy. This suggests that psychopathy captures
the genetic predisposition for ADHD as well as ASB. Psychopathic personality traits
54
may be viewed as the common ingredient that predisposes individuals to a variety of
maladaptive behavioral outcomes.
The fact that the correlation between Nonplanning and ADHD was rather high in
males deserves further comment. It suggests that there is a psychopathic component,
above and beyond a general externalizing orientation, that is related to ADHD in males.
This is based on the observation that the Nonplanning factor was uncorrelated with ASB,
yet was moderately correlated with ADHD symptoms. One can argue that this
association was due to overlapping items. Several CPS subscales, such as Impulsiveness
and Boredom Susceptibility, were indeed contaminated with attention problems. When
eliminating these subscales from the structural model, the association was attenuated but
did not disappear. At any rate, the inclusion of ADHD-like items in a psychopathy
instrument is not accidental. Classical descriptions of psychopaths emphasize their
proneness to boredom and need for stimulation (Cleckley, 1976).
Anxiety/Depression
The general CPS factor was moderately correlated with anxiety/depression. The
genetic correlations were even higher, ranging from 0.47 in males to 0.75 in females.
Thus, children with high levels of psychopathic traits experience more personal distress.
This result can be explained by the fact that externalizing and internalizing problems are
highly comorbid in children. Gjone & Stevenson (1997) reported that the phenotypic
correlation between externalizing and internalizing problems on the Child Behavior
Checklist ranged from 0.51 to 0.58. However, they found that the association was mainly
55
mediated by environmental effects, whereas the overlap between psychopathy and
anxiety/depression was mainly due to common genetic influences in the present study.
It was hypothesized that Charming/Manipulative would show an inverse
association with internalizing symptoms. However, neither of the specific psychopathy
factors were related to anxiety/depression. This represents a blow to the construct
validity of childhood psychopathy, given that fearlessness and low anxiety are thought to
define the interpersonal facet (Fowles & Dindo, 2006; Benning et al., 2003). In
incarcerated adults, the interpersonal facet of the PCL-R is negatively associated with
distress/neuroticism (Hicks & Patrick, 2006; Patrick et al., 2007), despite the fact that the
PCL-R does not directly assess internalizing problems. Furthermore, in a community
sample of adolescent twins, Blonigen et al. (2005) found that the genetic risk for
internalizing psychopathology was negatively correlated with the Fearless Dominance
factor of the Psychopathic Personality Inventory (PPI; Lilienfeld & Andrews, 1996).
However, as its name would indicate, this facet contained items assessing stress
immunity, fearlessness, and social potency. Unlike the PPI, items of the CPS do not
directly measure low anxiety.
Intelligence
It is well documented that children with conduct problems obtain lower scores on
intelligence tests. Few studies have examined the genetic and environmental
contributions to the association between externalizing problems and low IQ. In a large
sample of young twins, Koenen et al. (2006) determined that the phenotypic correlation
between antisocial behavior and IQ is approximately -.25 in boys and -.10 in girls.
56
Moreover, the association in boys was completely mediated by genetic effects. A similar
pattern of results was found in the present study. The general psychopathy factor was
negatively correlated with IQ in males (r = -.19), and to a lesser extent in females (r = -
.10). Genetic effects were necessarily responsible for these associations, given that
nonshared environmental influences on the latent IQ factor were negligible. Familial
aggregation of IQ was so high (and measurement error so low) that the point-estimate of
E was 6% and 1% in males and females, respectively.
On top of the general factor, there was an inverse relationship between IQ and
Nonplanning that reached statistical significance in both sexes (r = -.25 in males; r = -.15
in females). This indicates that psychopathy consists of a specific component that is
independent of externalizing problems, yet related to lower cognitive ability. It is
interesting to speculate that this represents a purer ADHD-like path, in which low IQ may
be due to attention problems. Children who lack goals, and do not plan ahead, may
perform poorly on IQ tests for a variety of reasons. Caregivers may attribute their child’s
academic failure to laziness or a lack of striving, even though it might be due to lower
cognitive ability. When the child is brought into the laboratory, he/she will naturally
perform poorly on an intelligence test. Intellectual deficits may also hinder a child’s
ability to accomplish goals, and therefore he/she ultimately avoids striving for goals
altogether. Alternatively, the child may lack the motivation to perform to his/her
potential, and therefore does not exert much effort during testing. Unfortunately, the
causal mechanisms between low IQ and maladaptive personality traits cannot yet be
57
determined. We can only conclude that the association is due to common genetic
influences.
Contrary to theoretical expectations, the association between IQ and the
Charming/Manipulative factor was not significant. This appears to challenge the validity
of the CPS, given that previous studies have consistently shown that interpersonal
glibness is positively correlated with IQ. This finding has been replicated in clinical
samples of adolescents and adults (Salekin et al., 2004; Vitacco et al., 2008) as well as a
community sample of preadolescent twins (Fontaine et al., 2008). Verbal abilities, in
particular, appear to be related to the interpersonal dimension (Salekin et al., 2004).
A more in-depth association between various aspects of intelligence and
psychopathy was explored. There is a vast literature showing that delinquents perform
poorer in the verbal domain relative to the nonverbal domain of intellectual functioning.
In particular, adolescent delinquents generally obtain a lower Verbal IQ (VIQ) than
Performance IQ (PIQ) on the Wechsler scales (Isen, 2010). In the present sample, VIQ
and PIQ scores were obtained from the Wechsler Abbreviated Scale of Intelligence. The
general psychopathy and Nonplanning factors were not preferentially related to low VIQ,
but were modestly related to PIQ as well as VIQ.
Skin Conductance Activity
Adult psychopaths generally show reduced electrodermal reactivity to aversive
stimuli (Lorber, 2004; Fowles, 2000). Moreover, reduced responsiveness to aversive
stimuli is found in young males with externalizing problems and/or psychopathic
tendencies (Blair, 1999; Herpertz et al., 2003, Fung et al., 2005). The study by Fung et al.
58
(2005) is particularly relevant, as they administered a countdown paradigm that was
nearly identical to the one employed here. They found that adolescent males with high
scores on the CPS were less likely to respond electrodermally during the signaled
anticipatory phase, but did not differ from controls during unsignaled trials. Furthermore,
these boys were less likely to respond to white noise-bursts, regardless of whether it was
signaled or unsignaled.
In the present sample, electrodermal findings for the countdown task were
equivocal. Contrary to Fung et al. (2005), psychopathy was not related to the number of
SCRs during the signaled anticipatory phase. This suggests that low anticipatory anxiety
is not related to psychopathy in preadolescents. Alternatively, subjects may not have
considered the countdown trials to be very stressful. There was also little evidence that
psychopathy is negatively associated with lower responsiveness to aversive stimuli. Only
one significant association emerged, and this was confined to males in the unsignaled
responsivity condition. More specifically, there was a correlation of -.18 between the
general psychopathy factor and the number of SCRs elicited by the unsignaled white-
noise bursts. This could suggest that psychopathy is related to reduced startle responses
in particular, rather than a general hyporesponsiveness to aversive stimuli.
There was minor support for the hypothesis that psychopathy is related to lower
SCR magnitudes to negative conceptual stimuli. It was hypothesized that psychopathy
would be inversely related to SCR magnitudes to shameful questions, but not to neutral
questions. Indeed, there was a weak negative association in males between the
Charming/Manipulative factor and the mean SCR magnitude to shameful questions.
59
However, none of the other correlations were significant, and no associations emerged in
females. This is consistent with the idea that the association between psychopathy and
electrodermal reactivity is limited to the interpersonal facet in males (Isen et al., 2010).
The genetic/environmental structure of these physiological variables was
somewhat perplexing. Ironically, the two SCR measures that showed relations with
psychopathy were not heritable. The low heritabilities imply that reduced startle
responses and hyporeactivity to distressing stimuli cannot be viewed as endophenotypic
characteristics of psychopathy in preadolescent males. There was undoubtedly much
measurement error (as reflected by the high E estimates), but that cannot explain the
presence of a significant environmental correlation between general psychopathy and
SCRs to unsignaled noise blasts. Instead of engaging in wild speculation about the cause
of this association, it may be more appropriate to acknowledge that Type 1 error may
have been responsible for this result. This possibility is very real, given that 24 potential
associations were examined with respect to the countdown task.
Limitations
The partitioning of phenotypic variance into genetic and environmental
components is imprecise when using standard twin methodology. Since the etiological
influences on a trait are not actually observed, they must be interpreted with caution.
Additive genetic effects (A) and shared environmental effects (C) may not precisely
capture the broad influences that they purport to measure. For example, the effect of C
might be masked by genetic non-additivity (i.e., genetic dominance). In the absence of
60
alternative designs (e.g., measured genotypes or twins separated at birth), it is impossible
to isolate the etiological mechanisms.
Several variables appeared to show evidence of non-additive genetic influences.
In particular, point-estimates of C for the psychopathy and ADHD factors were exactly
zero, which is a tell-tale sign of genetic non-additivity. Nevertheless, practical
considerations rendered it unfeasible to decompose the genetic variance into additive and
non-additive components. There was little benefit in modeling non-additive genetic
effects in the structural analyses, as many of the criterion variables were best described
by an ACE model.
Standard twin models assume the absence of assortative (non-random) mating
with respect to a phenotype of interest – an assumption that was certainly violated. When
parents are correlated on some heritable trait (e.g., intelligence), their fraternal offspring
will likely share more than 50% of the alleles for the trait in question. Unless specifically
modeled, the effects of assortative mating will be subsumed under shared environmental
influences. A closely related phenomenon is ethnic stratification, given that ethnic
diversity is sustained by non-random mating (i.e., couples tend to have a similar heritage).
If there are mean ethnic differences on a variable of interest, then the standard ACE
model will apportion the between-group variance as shared environmental influences.
This issue is particularly relevant for IQ, insomuch that 17% of the variance could be
explained by ethnic/racial status. To the extent that ethnic/racial differences are purely
environmental in origin (e.g., due to acculturation), the interpretation of shared
environmental effects is valid. However, it cannot be discounted that there are
61
ethnic/racial differences in the frequency of certain alleles that might affect cognitive
ability and antisocial behavior. In this case, genetic variance may wrongly be
apportioned as shared environmental variance. The implication is that estimates of C
may be artificially inflated for IQ and antisocial behavior.
Another shortcoming is the lack of multivariate normality. The maximum
likelihood method assumes that all of the variables (and their linear combinations) are
normally distributed. However, this assumption was violated with respect to many of the
variables under study. In particular, symptoms of conduct disorder and proactive
aggression were rather infrequent in this community sample. The resultantly high
kurtosis could not be remedied by any of the commonly used transformations.
Additionally, several subscales of the CPS, especially Callousness and Unreliability,
were positively skewed and quite kurtotic. This lack of normality, however, did not
appear to prejudice the structure of the CPS. When measurement models were fit to the
CPS data after the scores were first transformed using Blom’s (1958) rank normalization
procedure, the same findings were obtained. That is, the general-specific factor structure
prevailed over the second-order factor structure, and the hypothesis of factorial invarance
with respect to sex could still be rejected. Standardized factor loadings did not
appreciably differ when transformed scores were employed instead of raw scores.
A further problem is that the construct of psychopathy, as measured by the
Childhood Psychopathy Scale, appears to be different for males and females. The
specific interpersonal factor (Charming/Manipulative) was more robust in females,
whereas the second factor was more robust in males, insomuch that it was not as
62
overpowered by a single subscale (Lack of Planning). Given that factorial invariance was
not tenable with respect to sex, the criterion validity of psychopathy must necessarily be
evaluated separately in males and females. Hence, interpretation of psychopathy is
complicated by the fact that the underlying processes may differ between males and
females.
A final limitation is the use of a single instrument to measure psychopathy. The
Childhood Psychopathy Scale is only one of several instruments that are available.
Relative to other instruments, the CPS is poor at assessing guiltlessness and grandiosity.
For example, none of the Grandiosity items relate to the core issue of whether the child
thinks he/she is better than other people. Rather, the items appear to measure whether the
child has a healthy self-esteem.
This has deeper implications for the construct validity of psychopathy. Our
understanding of a construct is limited by the instrument(s) used to operationalize it. For
example, an instrument that heavily emphasizes the remorselessness, haughtiness, and
social dominance of the prototypical psychopath will reveal, not surprisingly, that
psychopathy is inversely related to anxiety. The CPS does a poor job of measuring these
core psychopathic traits, whereas other instruments (e.g., the Psychopathic Personality
Inventory) operationalize them to a far greater extent. It should also be noted that the
CPS subscales were imported from an adult instrument, rather than developed according
to theory (Lynam, 1997). Thus, the poor validity of certain psychopathy facets in the
present sample may stem from weaknesses in the Childhood Psychopathy Scale, rather
63
than necessarily implying that the construct of childhood psychopathy is inherently
flawed.
Strengths
A major strength of this study is the use of an innovative model to measure
psychopathy and its relations with external criteria. The general-specific (bifactor)
approach has clear advantages over competing approaches in terms of statistical elegance
and theoretical interpretability. Since the factors are orthogonal to one another, the
model permits a clearer interpretation of how each psychopathy facet is specifically
related to external criteria. Another benefit is that the criterion variables were typically
measured with little error. The USC Twin Study of Risk Factors for Antisocial Behavior
uses a very comprehensive assessment procedure. Multiple instruments were typically
used to assess a given construct. As a result, measurements of antisocial behavior,
anxiety/depression, and inattention/hyperactivity were largely free of error.
Concluding Remarks
The general conclusion is that psychopathy, as traditionally conceptualized by
Cleckley (1941), is not a valid construct in the present sample of preadolescent twins.
This may be due to its poor operationalization by the Childhood Psychopathy Scale
(CPS). Psychopathy, as measured by the CPS, appears to be a cluster of maladaptive
traits that describes a general externalizing orientation, with the concomitant internalizing
problems and low IQ often observed in children with externalizing problems. The
specific interpersonal factor did not function in a theoretically meaningful manner. That
is, the ‘good intelligence’ and absence of nervousness that supposedly characterize
64
psychopathy (Cleckley, 1976) was not demonstrated. In particular, there was no
evidence for the negative relations between psychopathy and anxiety/distress often
observed in adults. On the contrary, the correlation between general psychopathy and
anxiety/depression was unusually high in females, to such an extent that discriminant
validity could hardly be demonstrated. Moreover, discriminant validity with respect to
inattention/hyperactivity was questionable. This suggests that childhood psychopathy, as
measured by the CPS in this young sample, may best be conceptualized as a single
dimension that represents a broad risk for psychopathology.
65
References
Achenbach, T.M. (1991). Manual for the Child Behavior Checklist. Burlington, VT:
University of Vermont, Department of Psychiatry.
Akaike, H. (1987). Factor analysis and AIC. Psychometrika, 52(3), 317-332.
Baker, L. A., Barton, M., & Raine, A. (2002). The Southern California Twin Register at
the University of Southern California. Twin Research, 5(5), 456-459.
Baker, L. A., Barton, M., Lozano, D. I., Raine, A., & Fowler, J. H. (2006). The Southern
California Twin Register at the University of Southern California: II. Twin
Research and Human Genetics, 9(6), 933-940.
Baker, L. A., Jacobson, K. C., Raine, A., Lozano, D. I., & Bezdjian, S. (2007). Genetic
and environmental bases of childhood antisocial behavior: A multi-informant twin
study. Journal of Abnormal Psychology, 116(2), 219-235.
Baker, L. A., Raine, A., Liu, J., & Jacobson, K. C. (2008). Differential genetic and
environmental influences on reactive and proactive aggression in children.
Journal of Abnormal Child Psychology, 36(8), 1265-1278.
Bare, R. L., Hopko, D. R., & Armento, M. E. A. (2004). The Relation of Psychopathic
Characteristics and Anxiety in Noncriminals: Physiological and Cognitive
Responses to Guided Imagery. Journal of Psychopathology and Behavioral
Assessment, 26(4), 225-232.
Benning, S. D., Patrick, C. J., Hicks, B. M., Blonigen, D. M., & Krueger, R. F. (2003).
Factor Structure of the Psychopathic Personality Inventory: Validity and
Implications for Clinical Assessment. Psychological Assessment, 15(3), 340-350.
Benning, S. D., Patrick, C. J., & Iacono, W. G. (2005). Psychopathy, startle blink
modulation, and electrodermal reactivity in twin men. Psychophysiology, 42(6),
753-762.
Bezdjian, S. (2008). Psychometric and biometric evaluation of psychopathic personality
traits in 9 and 10 year old twins (Doctoral Dissertation). Available from
Dissertations and Theses database. (UMI No. 3325200)
Blair, R. J. R. (1999). Responsiveness to distress cues in the child with psychopathic
tendencies. Personality and Individual Differences, 27(1), 135-145.
Blom, G. (1958). Statistical estimates and transformed beta-variables. New York: John
Wiley and Sons
66
Blonigen, D. M., Hicks, B. M., Krueger, R. F., Patrick, C. J., & Iacono, W. G. (2005).
Psychopathic personality traits: Heritability and genetic overlap with internalizing
and externalizing psychopathology. Psychological Medicine, 35(5), 637-648.
Burt, S. A., McGue, M., Carter, L. A., & Iacono, W. G. (2007). The different origins of
stability and change in antisocial personality disorder symptoms. Psychological
Medicine, 37(1), 27-38.
Chen, F. F., West, S. G., & Sousa, K. H. (2006). A Comparison of Bifactor and Second-
Order Models of Quality of Life. Multivariate Behavioral Research, 41(2), 189-
225.
Cleckley, H. (1941/1976). The mask of sanity. St. Louis, MO: Mosby
Cooke, D. J., Michie, C., & Hart, S. D. (2006). Facets of Clinical Psychopathy: Toward
Clearer Measurement. In C. J. Patrick (Ed.), Handbook of Psychopathy (pp. 91-
106). New York, NY: Guilford Press.
Dawson, M. E., Schell, A. M., & Filion, D. L. (2007). The electrodermal system. In J. T.
Cacioppo, L. G. Tassinary & G. G. Berntson (Eds.), Handbook of
psychophysiology (3rd ed., pp. 159-181). New York: Cambridge University Press.
Falkenbach, D. M., Poythress, N. G., & Heide, K. M. (2003). Psychopathic Features in a
Juvenile Diversion Population: Reliability and Predictive Validity of Two Self-
Report Measures. Behavioral Sciences & the Law, 21(6), 787-805.
Farrington, D. P. (2005). The Importance of Child and Adolescent Psychopathy. Journal
of Abnormal Child Psychology, 33(4), 489-497.
Fontaine, N., Barker, E. D., Salekin, R. T., & Viding, E. (2008). Dimensions of
psychopathy and their relationships to cognitive functioning in children. Journal
of Clinical Child and Adolescent Psychology, 37(3), 690-696.
Fowles, D. C. (2000). Electrodermal hyporeactivity and antisocial behavior: Does anxiety
mediate the relationship? Journal of Affective Disorders, 61(3), 177-189.
Fowles, D. C., & Dindo, L. (2006). A Dual-Deficit Model of Psychopathy. In C. J.
Patrick (Ed.), Handbook of Psychopathy (pp. 14-34). New York, NY: Guilford
Press.
Fung, M. T., Raine, A., Loeber, R., Lynam, D. R., Steinhauer, S. R., Venables, P. H., &
Stouthamer-Loeber, M. (2005). Reduced Electrodermal Activity in Psychopathy-
Prone Adolescents. Journal of Abnormal Psychology, 114(2), 187-196.
67
Gjone, H., & Stevenson, J. (1997). The association between internalizing and
externalizing behavior in childhood and early adolescence: Genetic or
environmental common influences? Journal of Abnormal Child Psychology 25(4),
277-286.
Gray, J. A. (1987). The psychology of fear and stress (2nd ed.). New York, NY:
Cambridge University Press.
Hare, R. (1991). The Hare Psychopathy Checklist - Revised. Toronto, Ontario, Canada:
Multi-Health Systems
Hare, R. (2003). The Hare Psychopathy Checklist – Revised, 2
nd
edition. Toronto, Ontario,
Canada: Multi-Health Systems
Hare, R. D., Frazelle, J., & Cox, D. N. (1978). Psychopathy and physiological responses
to threat of an aversive stimulus. Psychophysiology, 15(2), 165-172.
Herpertz, S. C., Mueller, B., Wenning, B., Qunaibi, M., Lichterfeld, C., & Herpertz-
Dahlmann, B. (2003). Autonomic responses in boys with externalizing disorders.
Journal of Neural Transmission, 110(10), 1181-1195.
Hicks, B. M., & Patrick, C. J. (2006). Psychopathy and negative emotionality: Analyses
of suppressor effects reveal distinct relations with emotional distress, fearfulness,
and anger-hostility. Journal of Abnormal Psychology, 115(2), 276-287.
Isen, J., Raine, A., Baker, L., Dawson, M., Bezdjian, S., & Lozano, D. I. (2010). Sex-
specific association between psychopathic traits and electrodermal reactivity in
children. Journal of Abnormal Psychology, 119(1), 216-225.
Isen, J. (2010). Meta-analytic assessment of Wechsler’s P>V sign in antisocial
populations. Clinical Psychology Review, 30(4), 423-435.
Koenen, K. C., Caspi, A., Moffitt, T. E., Rijsdijk, F., & Taylor, A. (2006). Genetic
influences on the overlap between low IQ and antisocial behavior in young
children. Journal of Abnormal Psychology, 115(4), 787-797.
Larsson, H., Tuvblad, C., Rijsdijk, F. V., Andershed, H., Grann, M., & Lichtenstein, P.
(2007). A common genetic factor explains the association between psychopathic
personality and antisocial behavior. Psychological Medicine, 37(1), 15-26.
Lilienfeld, S. O., & Andrews, B. P. (1996). Development and preliminary validation of a
self-report measure of psychopathic personality traits in noncriminal populations.
Journal of Personality Assessment, 66(3), 488-524.
68
Lorber, M. F. (2004). Psychophysiology of Aggression, Psychopathy, and Conduct
Problems: A Meta-Analysis. Psychological Bulletin, 130(4), 531-552.
Lykken, D. T. (1995). The antisocial personalities. Hillsdale, NJ: Lawrence Erlbaum
Associates, Inc.
Lynam, D., Moffitt, T. E., & Stouthamer-Loeber, M. (1993). Explaining the relation
between IQ and delinquency: Class, race, test motivation, school failure, or self-
control? Journal of Abnormal Psychology, 102(2), 187-196.
Lynam, D. R. (1997). Pursuing the psychopath: Capturing the fledgling psychopath in a
nomological net. Journal of Abnormal Psychology, 106(3), 425-438.
Lynam, D. R., Caspi, A., Moffitt, T. E., Raine, A., Loeber, R., & Stouthamer-Loeber, M.
(2005). Adolescent Psychopathy and the Big Five: Results from Two Samples.
Journal of Abnormal Child Psychology, 33(4), 431-443.
Lynam, D. R., & Derefinko, K. J. (2006). Psychopathy and personality. In C. J. Patrick
(Ed.), Handbook of psychopathy (pp. 133-155). New York: Guilford
Lynam, D. R., Miller, D. J., Vachon, D., Loeber, R., & Stouthamer-Loeber, M. (2009).
Psychopathy in adolescence predicts official reports of offending in adulthood.
Youth Violence and Juvenile Justice, 7(3), 189-207.
Meredith, W. (1993). Measurement invariance, factor analysis and factorial invariance.
Psychometrika, 58(4), 525-543.
Neale M.C., & Cardon L.R. (1992). Methodology for genetic studies of twins and families.
Dordrecht, Netherlands: Kluwer Academic Publisher.
Neale M.C., Boker S.M., Xie G., & Maes, H.H. (1994). Mx: Statistical Modeling, 7th
edition. Richmond, Virginia: Department of Psychiatry.
Patrick, C. J., Cuthbert, B. N., & Lang, P. J. (1994). Emotion in the criminal psychopath:
Fear image processing. Journal of Abnormal Psychology, 103(3), 523-534.
Patrick, C. J., Hicks, B. M., Nichol, P. E., & Krueger, R. F. (2007). A bifactor approach
to modeling the structure of the Psychopathy Checklist-Revised. Journal of
Personality Disorders, 21(2), 118-141.
Raine, A. (1993). The psychopathology of crime: Criminal behavior as a clinical
disorder. San Diego, CA: Academic Press
69
Raine, A., Dodge, K., Loeber, R., Gatzke-Kopp, L., Lynam, D., Reynolds, C.,
Stouthamer-Loeber, M., & Liu, J. (2006). The reactive-proactive aggression
questionnaire: Differential correlates of reactive and proactive aggression in
adolescent boys. Aggressive Behavior, 32(2), 159-171.
Salekin, R. T., Neumann, C. S., Leistico, A.-M. R., & Zalot, A. A. (2004). Psychopathy
in Youth and Intelligence: An Investigation of Cleckley's Hypothesis. Journal of
Clinical Child and Adolescent Psychology, 33(4), 731-742.
Sareen, J., Stein, M. B., Cox, B. J., & Hassard, S. T. (2004). Understanding comorbidity
of anxiety disorders with antisocial behavior: Findings from two large community
surveys. Journal of Nervous and Mental Disease, 192(3), 178-186.
Schaffer, D., Fisher, P., Lucas, C.P., Dulcan, M.K., & Schwab-Stone, M.E. (2000).
NIMH Diagnostic Interview Schedule for Children Version IV (NIMH DISC-IV):
Description, differences from previous version, and reliability of some common
diagnoses. Journal of the American Academy of Child & Adolescent Psychiatry
39: 28-38.
Skeem, J. L., Poythress, N., Edens, J. F., Lilienfeld, S. O., & Cale, E. M. (2003).
Psychopathic personality or personalities? Exploring potential variants of
psychopathy and their implications for risk assessment. Aggression and Violent
Behavior, 8(5), 513-546.
Vitacco, M. J., Neumann, C. S., & Jackson, R. L. (2005). Testing a Four-Factor Model of
Psychopathy and Its Association With Ethnicity, Gender, Intelligence, and
Violence. Journal of Consulting and Clinical Psychology, 73(3), 466-476.
Vitacco, M. J., Neumann, C. S., & Wodushek, T. (2008). Differential relationships
between the dimensions of psychopathy and intelligence: Replication with adult
jail inmates. Criminal Justice and Behavior, 35(1), 48-55.
Waldman, I. D., & Rhee, S. H. (2006). Genetic and Environmental Influences on
Psychopathy and Antisocial Behavior. In C. J. Patrick (Ed.), Handbook of
Psychopathy (pp. 205-228). New York, NY: Guilford Press.
Wechsler, D. (1999). Wechsler Abbreviated Scale of Intelligence (WASI). San Antonio,
TX: Harcourt Assessment.
Widiger, T. A. (2006). Psychopathy and DSM-IV Psychopathology. In C. J. Patrick (Ed.),
Handbook of Psychopathy (pp. 156-171). New York, NY: Guilford Press.
70
Appendix:
Supplementary Analyses, Figures, and Tables
Figure A1. Path diagram of a second-order factor model. This is equivalent to the
model shown in Figure 3. However, its higher-order nature is more apparent in this
diagram. Genetic and environmental influences operate on the lower-order factors
directly via the higher-order factor. The residual variances of the lower-order factors can
be partitioned into genetic (A
R
), shared environmental (C
R
), and nonshared
environmental (E
R
) components.
Charming/
Manipulative
Callous/
Disinhibited
Glibness
Untruthfulness
Fail to accept
responsibility
Unreliability
Impulsiveness
Callousness
Manipulation
Parasitic
Lifestyle
Behavioral
Dyscontrol
Lack of
Planning
Poverty of
Affect
Boredom
Susceptible
Psychopathy
A
C
E
A
R
C
R
E
R
A
R
C
R
E
R
1
1
71
Table A1. Childhood Psychopathy Scale (CPS)
Glibness
1. Does he/she try to be the center of attention?
2. Is he/she talkative?
3. Is he/she shy? (r)
4. Does he/she tell stories to make him/herself look good?
5. Does he/she show off to get people to pay attention to him/her?
Untruthfulness
1. Is he/she open and honest? (r)
2. Can he/she be trusted? (r)
3. Is he/she a good liar?
4. Will he/she usually tell a lie if he/she thinks he/she can get away with it?
5. Do people usually believe him/her when he/she tells a lie?
Manipulation
1. Does he/she try to act charming or likable in order to get his/her way?
2. Does he/she try to take advantage of other people?
3. Does he/she try to get others to do what he/she wants by getting on their good side?
Parasitic Lifestyle
1. Does he/she try to see how much he/she can get away with?
2. Does he/she take a lot and not give much in return?
3. Does he/she give or share things? (r)
4. Does he/she usually return what he/she borrows? (r)
Failure to Accept Responsibility
1. Does he/she try to blame other people for things that he/she has done?
2. Does he/she think he/she gets blamed for things he/she did not do?
3. When he/she gets in trouble, can he/she talk his/her way out of it?
Lack of Guilt
1. Does he/she usually feel bad or guilty after doing something wrong? (r)
2. Does it bother him/her when he/she does something wrong? (r)
3. Does he/she wish he/she could take back many things that he/she has done? (r)
Grandiosity
1. Does he/she need to have people tell him/her that he/she is doing well or ok? (r)
2. Does he/she have a low opinion or think badly of him/herself? (r)
3. Is he/she very sure of him/herself? (r)
72
Boredom Susceptibility
1. Is he/she easily bored?
2. Does he/she concentrate well on things? (r)
3. Does he/she stay away form scary things and places? (r)
4. Does he/she do dangerous things for the fun of it?
5. Does he/she need to have things be exciting?
Poverty of Affect
1. Is he/she a warm and kind person? (r)
2. Does he/she make close friends with other people? (r)
3. Are his/her moods unpredictable?
4. Is he/she open with his/her feelings? (r)
5. Do his/her feelings come and go quickly?
6. Is he/she protective of people who are close to him/her? (r)
7. Do his/her feelings sometimes seem fake?
8. Does he/she feel things very strongly? (r)
9. Do his/her feelings change often and quickly?
10. Are his/her feelings intense? (r)
Callousness
1. Is he/she kind and thoughtful of other people? (r)
2. Is he/she mean to other people?
3. Does he/she tease and pick on other people?
4. Is he/she able to tell how other people feel? (r)
5. Does he/she try not to hurt other people's feelings? (r)
Behavioral Dyscontrol
1. Is he/she easily frustrated?
2. Does he/she get irritated or mad over little things?
3. Does he/she lose his/her temper easily?
Lack of Planning
1. Does he/she plan things ahead? (r)
2. Does he/she set goals for him/herself and try to reach them? (r)
3. Does he/she think about what he/she wants to do with the rest of his/her life? (r)
Impulsivity
1. Does he/she think before doing or saying something? (r)
2. Does he/she think about his/her actions and behavior? (r)
3. Does he/she have a hard time waiting for things he/she wants?
73
Unreliability
1. When he/she starts working on something, does he/she stick with it? (r)
2. Can people count on him/her? (r)
3. Does he/she often break his/her promises?
Note: Reverse-scored items are denoted by r
74
Table A2. Conduct Disorder Symptoms (DISC-IV)
1. Ever secretly stolen money or other things from people he/she lives with?
2. Ever shoplifted?
3. Ever stolen from anyone else when they weren’t around or weren’t looking?
4. Ever faked a name on a check or used someone’s credit card without permission?
5. Ever snatched someone’s purse or jewelry?
6. Ever held someone up or attacked someone to steal from them?
7. Ever threatened someone in order to steal from them?
8. Ever gotten into trouble for staying out at night more than two hours past the time
he/she was supposed to be home?
9. Ever run away from home overnight?
10. Ever lied to get money or something else he/she wanted?
11. Ever lied to not pay back money or to get out of something important he/she was
supposed to do?
12. Ever skipped school?
13. Ever broken into a house, building, or car?
14. Ever broken or damaged someone else’s things on purpose?
15. Ever broken something or messed up some place on purpose?
16. Ever started a fire without permission?
17. Ever been physically cruel to an animal?
18. Ever bullied someone by hitting or threatening or scaring someone who is younger or
smaller than him/her or somebody who won't fight back?
19. Ever threatened someone or frightened someone on purpose?
20. Ever been in a physical fight in which someone was hurt or could have been hurt?
21. Ever tried to hurt someone badly or been physically cruel to someone?
22. Ever hurt or threatened someone with a weapon (like a bat, brick, broken bottle, knife,
or gun)?
23. Ever been expelled from school for misbehavior?
24. Ever been suspended from school for misbehavior?
25. Ever been in trouble with the police?
75
Table A3. Reactive Aggression (RPQ)
1. He/she yells at others when they annoy him/her.
2. He/she gets angry when others annoy him/her.
3. He/she gets angry when frustrated.
4. He/she has temper tantrums.
5. He/she damages things when he/she is mad.
6. He/she gets angry or mad when he/she doesn't get his/her way.
7. He/she gets angry or mad when he/she loses a game.
8. He/she gets angry when others threaten him/her.
9. He/she feels better after hitting or yelling at someone.
10. He/she hits others to defend him/herself.
11. He/she gets mad or hits others when they tease him/her.
76
Table A4. Proactive Aggression (RPQ)
1. He/she fights others to show who is on top.
2. He/she takes things from other kids.
3. He/she damages or breaks things for fun.
4. He/she gets into fights to be cool.
5. He/she hurts others to win a game.
6. He/she uses force to get others to do what he/she wants.
7. He/she uses force to get money or things from others.
8. He/she threatens and bullies other kids.
9. He/she makes prank phone calls just for fun.
10. He/she gets others to gang up on other kids.
11. He/she carries a weapon to use in a fight.
12. He/she yells at others so they will do things for him/her.
77
Table A5. Attention Problems (CBCL)
1. Acts too young for his/her age
2. Can't concentrate/ pay attention for long
3. Can't sit still, restless, or hyperactive
4. Confused or seems to be in a fog
5. Daydreams or gets lost in his/her thoughts
6. Impulsive or acts without thinking
7. Nervous, highstrung, or tense
8. Poor school work
9. Poorly coordinated or clumsy
10. Stares blankly
78
Table A6. Inattention Symptoms (DISC-IV)
In the last year…
1. Did he often have trouble keeping his mind on what he was doing for more than a
short time?
2. Did he often try not to do things where he would have needed to pay attention for a
long time?
3. Did he often dislike doing things where he had to pay attention for a long time?
4. Did he often find it hard to keep his mind on what he was doing when other things
were going on?
5. Was he disorganized?
6. Did he often have trouble finishing things he was supposed to do?
7. Did he often lose things he needed?
8. Did he often forget what he was supposed to be doing or what he had planned to do?
9. Has he often made a lot of mistakes because it's hard for him to do things carefully?
10. Did he often not listen when people were speaking to him?
11. Did he often not finish things because he started to do something else?
79
Table A7. Hyperactivity/Impulsivity Symptoms (DISC-IV)
In the last year…
1. Was he often "on the go" or did he move around as if he was "driven by a motor"?
2. Was often fidgety or restless?
3. Has he often left his seat when he wasn't supposed to?
4. Did he often climb on things or run around when he wasn't supposed to?
5. When he had to sit still, for say more than ten minutes, did he nearly always seem
restless, as if he wanted to kick his feet or get up and move around?
6. Did he often talk a lot more than other people his age?
7. Did he often make much more noise than other people his age when he was having
fun?
8. Did he often interrupt other people when they were talking or when they were busy?
9. Did he often butt in on what other people were doing?
10. Did he often blurt out answers before someone could finish asking the question?
11. Has he often had trouble waiting for his turn, like when he was standing in line... or
playing a game?
80
Table A8. Anxious/Depressed (CBCL)
1. Complains of loneliness
2. Cries a lot
3. Fears he/she might think or do something bad
4. Feels he/she has to be perfect
5. Feels or complains that no one loves him/her
6. Feels others are out to get him/her
7. Feels worthless or inferior
8. Nervous, highstrung, or tense
9. Too fearful or anxious
10. Feels too guilty
11. Self-conscious or easily embarrassed
12. Acts Suspicious
13. Unhappy, sad or depressed
14. Worries
81
Table A9. Generalized Anxiety Disorder Symptoms (DISC-IV)
In the last year…
1. Did he often seem very worried about going to play a sport or game or do some other
activity?
2. Did he often seem to worry a lot when he made small mistakes doing projects or
activities?
3. Did he often seem worried about being on time?
4. Has he often seemed very worried that he might have some sickness or illness?
5. Is he the kind of person who is often very tense, or who seems to find it very hard to
relax?
6. Has he often seemed worried that he has made a mistake of has done something the
wrong way?
7. Has he often seemed worried that he made a fool of himself in front of other people?
8. Has he often worried about whether other people liked him?
9. Has he often said he had headaches?
10. Has he often said he had a stomachache?
11. Has he often said he had other aches and pains?
82
Table A10. Major Depression Symptoms (DISC-IV)
In the last year…
1. Was there a time when he often seemed sad or depressed?
2. Was there a time when it seemed like nothing was fun for him and he just wasn't
interested in anything?
3. Was there a time when he often was grouchy or irritable and often in a bad mood,
when even little things would make him mad?
4. Was there a time when he lost weight?
5. Was there a time when he seemed to lose his appetite or ate a lot less than usual?
6. Was there a time when he gained a lot of weight?
7. Was there a time when he seemed to feel much hungrier than usual or when he ate a
lot more than usual?
8. Was there a time when he had trouble sleeping - that is, trouble falling asleep, staying
asleep, or waking up too early?
9. Was there a time when he slept more during the day than he usually does?
10. Was there a time when he seemed to do things like walking or talking much more
slowly than usual?
11. Was there a time when he often seemed restless... like he just had to keep walking
around?
12. Was there a time when he seemed to have less energy than he usually does?
13. Was there a time when doing even little things seemed to make him feel really tired?
14. Was there a time when he said his arms and legs felt heavy, like he was weighted
down by them?
15. Was there a time when he often blamed himself for bad things that happened?
16. Was there a time when he said he couldn't do anything well or that he wasn't as good
looking or as smart as other people?
17. Was there a time when it seemed like he couldn't think as clearly or as fast as usual?
18. Was there a time when he often seemed to have trouble keeping his mind on things?
19. Was there a time when it often seemed hard for him to make up his mind or to make
decisions?
20. Was there a time when he said he often thought about death or about people who had
died or about being dead himself?
21. Did he ever talk seriously about killing himself?
83
Table A11. Intraclass twin correlations for antisocial behavior
Zygosity
Reactive
Aggression
Proactive
Aggression
Conduct Disorder
symptom count
MZ Male
.48**
(N = 138)
.42**
(N = 138)
.63**
(N = 135)
MZ Female
.52**
(N = 143)
.44**
(N = 143)
.68**
(N = 140)
DZ Male
.32**
(N = 84)
.23*
(N = 84)
.43**
(N = 83)
DZ Female
.46**
(N = 93)
.40**
(N = 93)
.64**
(N = 90)
DZ Male-Female
.46**
(N = 147)
.33**
(N = 147)
.28**
(N = 146)
*p < .05; **p < .01
84
Table A12. Intraclass twin correlations for ADHD
Zygosity
Inattention
Symptoms
Hyperactivity-
Impulsivity
Attention Problems
MZ Male
.45**
(N = 135)
.56**
(N = 135)
.44**
(N = 136)
MZ Female
.28**
(N = 139)
.53**
(N = 139)
.46**
(N = 143)
DZ Male
.08
(N = 83)
.23*
(N = 83)
.41**
(N = 83)
DZ Female
-.06
(N = 89)
.18
(N = 89)
.21*
(N = 93)
DZ Male-Female
.23**
(N = 146)
.25**
(N = 146)
.16*
(N = 146)
*p < .05; **p < .01
85
Table A13. Intraclass twin correlations for anxiety/depression
Zygosity
Major Depression
Symptoms
Generalized
Anxiety Symptoms
Anxious/
Depressed
MZ Male
.53**
(N = 136)
.35**
(N = 136)
.51**
(N = 136)
MZ Female
.43**
(N = 139)
.32**
(N = 139)
.68**
(N = 143)
DZ Male
.42**
(N = 83)
.24*
(N = 83)
.31**
(N = 83)
DZ Female
.26*
(N = 90)
-.05
(N = 90)
.37**
(N = 93)
DZ Male-Female
.33**
(N = 145)
.37**
(N = 146)
.42**
(N = 146)
*p < .05; **p < .01
86
Table A14. Intraclass twin correlations for WASI subtests
Zygosity
Vocabulary Similarities Block Design Matrix
Reasoning
MZ Male
.72**
(N = 138)
.54**
(N = 138)
.55**
(N = 138)
.51**
(N = 138)
MZ Female
.72**
(N = 142)
.50**
(N = 143)
.65**
(N = 143)
.66**
(N = 143)
DZ Male
.49**
(N = 84)
.28*
(N = 84)
.32**
(N = 84)
.30**
(N = 84)
DZ Female
.67**
(N = 92)
.54**
(N = 92)
.38**
(N = 91)
.32**
(N = 91)
DZ Male-Female
.58**
(N = 146)
.49**
(N = 146)
.42**
(N = 147)
.41**
(N = 147)
*p < .05; **p < .01
87
Table A15. Intraclass twin correlations for frequency of SCRs during Countdown task
Zygosity
Anticipatory
SCRs
Nonspecific
SCRs
Signaled SCRs Unsignaled
SCRs
MZ Male
(N = 62)
.43** .26* .59** .21
MZ Female
(N = 81)
.43** .30** .44** .28*
DZ Male
(N = 39)
.39* .12 .21 .13
DZ Female
(N = 58)
.38** .29* .32* .14
DZ Male-Female
(N = 64)
.30* .28* .22 .23
*p < .05; **p < .01
88
Table A16. Intraclass twin correlations for SCR magnitude during Questions task
Zygosity
Shameful
Questions
Neutral
Questions
MZ Male
.34**
(N = 63)
.39**
(N = 64)
MZ Female
.25*
(N = 77)
.38**
(N = 75)
DZ Male
.39**
(N = 44)
.22
(N = 45)
DZ Female
.20
(N = 60)
.27*
(N = 60)
DZ Male-Female
.32**
(N = 73)
.06
(N = 71)
*p < .05; **p < .01
Abstract (if available)
Abstract
A nomological network of constructs was developed in order to evaluate the construct validity of psychopathy in preadolescent children. Theoretical conceptualizations of psychopathy in adults emphasize its multi-faceted nature
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Child psychopathy and orienting: novel support for electrodermal hyporeactivity
PDF
Psychometric and biometric evaluation of psychopathic personality traits in 9 and 10 year old twins
PDF
Multivariate biometric modeling among multiple traits across different raters in child externalizing behavior studies
PDF
Multiple informant-rated psychopathy in young adolescents: a multitrait-multimethod investigation of the Antisocial Process Screening Device
PDF
Biometric models of psychopathic traits in adolescence: a comparison of item-level and sum-score approaches
PDF
Patterns of EEG spectral power in 9-10 year old twins and their relationships with aggressive and nonaggressive antisocial behavior in childhood and adolescence
PDF
Different genetic and environmental structures for the overlap of three antisocial behavior factors with alcohol initiation
PDF
Pushmi-pullyu and little c: a search for the structure of personal creativity in a general population
PDF
Detecting joint interactions between sets of variables in the context of studies with a dichotomous phenotype, with applications to asthma susceptibility involving epigenetics and epistasis
PDF
Toward counteralgorithms: the contestation of interpretability in machine learning
Asset Metadata
Creator
Isen, Joshua
(author)
Core Title
Construct validity of psychopathic personality traits in a cohort of young twins
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Psychology
Publication Date
05/08/2010
Defense Date
03/22/2010
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
antisocial behavior,OAI-PMH Harvest,preadolescence,psychopathology,psychopathy,Twins
Place Name
California
(states),
Los Angeles
(city or populated place)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Baker, Laura A. (
committee chair
), Dawson, Michael E. (
committee member
), Prescott, Carol A. (
committee member
), Shen, Biing-Jiun (
committee member
), Trickett, Penelope K. (
committee member
)
Creator Email
isen@usc.edu,joshisen@hotmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m3056
Unique identifier
UC1274352
Identifier
etd-Isen-3641 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-338471 (legacy record id),usctheses-m3056 (legacy record id)
Legacy Identifier
etd-Isen-3641.pdf
Dmrecord
338471
Document Type
Dissertation
Rights
Isen, Joshua
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
antisocial behavior
preadolescence
psychopathology
psychopathy