The 4 Year Stability of Psychopathic Traits in Non ...labs.uno.edu/developmental-psychopathology/APSD-Self-Report/The_4... · The 4 Year Stability of Psychopathic Traits in Non-Referred

Behavioral Sciences and the Law

Behav. Sci. Law 21: 713–736 (2003)

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/bsl.568

The 4 Year Stability ofPsychopathic Traits inNon-Referred Youth

Paul J. Frick, Ph.D.,* Eva R. Kimonis, M.S.,Danielle M. Dandreaux, M.S., and Jamie M.Farell, M.S.

One significant limitation in research extending the

construct of psychopathy to youth has been the absence of

longitudinal studies testing the stability of psychopathic

traits prior to adulthood. To begin to address this limita-

tion, the current study estimated the stability of psycho-

pathic traits over a 4 year period in a sample of non-referred

children in the third, fourth, sixth, and seventh grades at

the first assessment. For parent ratings of psychopathic

traits, stability estimates using intra-class correlation coef-

ficients ranged from 0.80 to 0.88 across 2–4 years, with a

stability estimate of 0.93 across all four assessments. There

were also distinct trends in the patterns of stability found in

the sample. Specifically, children rated as being initially

high on these traits were more likely to be rated lower at

later assessments than was the case for children rated

initially low on these traits. Finally, the child’s level of

conduct problems, the socioeconomic status of the child’s

family, and the quality of parenting the child received were

the most consistent predictors of stability of psychopathic

traits. Copyright # 2003 John Wiley & Sons, Ltd.

Research extending the construct of psychopathy to youth has been quite promising

in a number of respects. For example, the presence of psychopathic features has been

proven to designate a subgroup of antisocial youth with more severe and more

aggressive patterns of antisocial behavior in forensic (Caputo, Frick, & Brodsky,

1999; Kruh, Frick, & Clements, in press), mental health (Christian, Frick, Hill, Tyler,

& Frazer, 1997), and community (Frick, Cornell, Barry, Bodin, & Dane, 2003;

Lynam, 1997) samples. Further, children with conduct problems who also show

psychopathic traits show a number of distinct characteristics compared to other

antisocial youth, mirroring research on antisocial adults with psychopathy

(Hare, 1998; Hart & Hare, 1997). For example, antisocial youth who also show

Copyright # 2003 John Wiley & Sons, Ltd.

*Correspondence to: Paul J. Frick, Department of Psychology, University of New Orleans, 2001 Geologyand Psychology Building, New Orleans, LA 70148, U.S.A. E-mail: [email protected] work was supported by grant MH55654 from the National Institute of Mental Health made to thefirst author.

Contract/grant sponsor: National Institute of Mental Health; contract/grant number: MH55654.

psychopathic traits demonstrate a stronger preference for novel, exciting, and dan-

gerous activities (Frick et al., 2003b; Frick, Lilienfeld, Ellis, Loney, & Silverthorn,

1999), are less sensitive to cues of punishment, especially when a reward-oriented

response set is primed (Barry et al., 2000; Fisher & Blair, 1998; Frick et al., 2003b)

and are less reactive to threatening and emotionally distressing stimuli (Blair, 1999;

Frick et al., 2003b; Loney, Frick, Clements, Ellis, & Kerlin, 2003).

Unfortunately, the research extending the construct of psychopathy to youth has

largely been cross-sectional in nature. As a result, there is limited evidence for the

predictive utility of psychopathic features prior to adulthood (Edens, Skeem, Cruise,

& Cauffman, 2001). There are several notable exceptions in which psychopathic

features have predicted subsequent delinquency, aggression, number of violent

offenses, and a shorter length of time to violent reoffending in institutionalized

antisocial youth (Brandt, Kennedy, Patrick, & Curtin, 1997; Forth, Hart, & Hare,

1990; Toupin, Mercier, Dery, Cote, & Hodgins, 1995). In one of the only studies to

test the predictive utility of psychopathic features in a non-referred sample of children,

Frick, Cornell, Barry, Bodin, and Dane (2003) reported that children with conduct

problems and psychopathic features showed more severe and more instrumental

aggression and had higher rates of self-reported delinquent acts one year later

compared with children with conduct problems but without psychopathic traits.

These results suggest that psychopathic features may designate an especially

severe, aggressive, and chronic type of disturbance in antisocial youth. However,

there are a number of limitations in this body of research. First, most studies have

relied on clinic-referred or forensic samples (except Frick et al., 2003a; Lynam,

1997) and focused on the adolescent age group (except Christian et al., 1997; Frick

et al., 2003a). Second, even the studies that have used prospective designs have had

limited follow-up periods, typically with follow-up periods of one (Frick et al.,

2003a) to two (Brandt et al., 1997; Forth et al., 1990) years. Third, while these few

prospective studies have tested the utility of psychopathic traits in youth for

predicting later antisocial and delinquent behavior, these studies have not tested

the stability of the actual traits associated with psychopathy.

While there is still much debate as to how best to conceptualize the core

dimensions of psychopathy (Hare, 2003), several recent factor analyses in adults

(Cooke & Michie, 2001) and children (Frick, Bodin, & Barry, 2000) have isolated

three core dimensions. One dimension focuses on a callous and unemotional

interpersonal style, the second dimension focuses on an inflated and egocentric

view of oneself, and the third focuses on an impulsive and antisocial lifestyle. A

recent critique of the literature extending the construct of psychopathy to youth

noted that some level of behavior in each of these dimensions is normative in youth,

particularly in adolescence (Seagrave & Grisso, 2002). As a result, the authors

suggested that, until research demonstrates that these traits designate a stable

dimension of personality in youth, comparable to the stability found in adult

samples, definitions of psychopathy run the risk of applying a potentially stigmatiz-

ing label to some youth who may only be showing a transient pattern of behavior

related to their developmental stage.

A number of important issues raised in this critique have been debated (see

Frick, 2002, Hart, Watt, & Vincent, 2002, & Lynam, 2002, for a more extended

discussion of these issues). One issue relates to the assumption that the stability of

psychopathic traits has been firmly established in adults. This assumption is largely

714 P. J. Frick et al.

Copyright # 2003 John Wiley & Sons, Ltd. Behav. Sci. Law 21: 713–736 (2003)

based on research showing that incarcerated adults with psychopathic traits often

do not respond to typical interventions (see, e.g., Ogloff, Wong, & Greenwood,

1990) and that these adults are at high risk for recidivism once released from prison

(see, e.g., Gendreau, Goggin, & Smith, 2002; Hemphill, Hare, & Wong, 1998;

Walters, 2003). However, recent reviews of the literature on the treatment of

psychopathy have called into question the extreme pessimism surrounding treat-

ment (Salekin, 2002; Skeem, Monahan, & Mulvey, 2003). More importantly, this

body of research does not provide a direct estimate of the stability of the core traits

associated with psychopathy in adults. One of the reasons for the paucity of

information on the stability of psychopathic traits is that the most widely used

method for assessing these traits relies heavily on historical information (Hare,

2003), making it difficult to assess changes in these traits over time. Therefore, with

three notable exceptions, the actual stability of these traits across adulthood has not

been tested.

First, Harpur and Hare (1994) compared the prevalence of diagnoses of

psychopathy and the mean level of psychopathic traits in six different age groups

of adult offenders, ranging in age from 16 to 70. These authors reported that overall

rates of psychopathy declined with age, especially after age 45. Also, this decline

tended to be stronger for the impulsive and antisocial lifestyle dimension of

psychopathy, whereas the average level of the interpersonal and affective traits of

psychopathy remained fairly stable across age groups. Importantly, this study was a

cross-sectional study showing cohort changes in the level of psychopathy and, thus,

did not test the stability of psychopathic traits directly. Second, Schroeder,

Schroeder, and Hare (1983) did directly assess stability in a sample of incarcerated

adults. These authors reported a stability coefficient of 0.89 for psychopathic traits

over a 10 months period. Although this does indicate substantial stability, the

follow-up period was rather limited. Third, in one of the best tests of the stability of

psychopathic traits in adults, Rutherford and colleagues (Rutherford, Cacciola,

Alterman, McKay, & Cook, 1999) reported 2 year stability estimates obtained in a

sample of 200 male and 25 female patients receiving methadone treatment for drug

abuse. The 2 year stability estimates reported in this study were 0.60 for men and

0.65 for women.

Taken together, there is not a great deal of evidence documenting the level of

stability of psychopathic traits in adult samples. However, even if one assumes that

these traits are relatively stable in adults, the contention that they should be

comparably stable in youth is still questionable. While there is some continuity in

personality traits across childhood (Buss, 1995), personality traits are generally

found to be less stable in children than in adults (see, e.g., McCrae et al., 2002;

Roberts & DelVecchio, 2000). This lower stability is not surprising, given the

developmental changes that are occurring in children that help to shape their

eventual personality. In fact, one of the primary motivations for extending the

construct of psychopathy to youth is to hopefully identify these traits when they are

more malleable (i.e. less stable) and amenable to treatment (Frick, 1998, 2001).

Thus, one would expect, and actually hope, that psychopathic traits are less stable in

youth and that factors could be identified that predict changes in these traits.

However, the level of stability could be so low that considering it a ‘‘personality’’

construct in children, which assumes at least a modest level of temporal stability, is

questionable (Buss, 1995).

Stability of childhood psychopathy 715


Based on these considerations, the primary purpose of the study was to determine

the stability of psychopathic traits across late childhood and into adolescence.

However, a second goal of the study was to begin to test variables that might

influence the stability of these traits. It was difficult to make hypotheses on potential

predictors of stability, given the absence of research in the area. However, several

variables were tested as potential predictors of stability based on the following

theoretical considerations.

First, we tested whether psychopathic traits are more stable in the context of

significant antisocial behavior. As mentioned previously, most of the available

research on psychopathy in both adults and youth has focused on forensic or

clinic-referred samples. As a result, much of the available research only pertains to

psychopathic individuals who show high levels of antisocial and criminal behavior.

However, most theories of psychopathy recognize that psychopathic traits can be

present in individuals who do not show significant antisocial behavior (Cleckley,

1976; Hare, 1998; Lilienfeld & Andrews, 1996). In the current study, we test

whether the stability of these traits is influenced by the level of one’s antisocial

behavior. Second, as noted above, youth with psychopathic traits show character-

istics associated with a temperament that has been labeled as low fearfulness or low

behavioral inhibition (Frick & Morris, in press), such as seeking out novel and

dangerous activities and showing a reward dominant cognitive style. These char-

acteristics have been critical to many theoretical models to explain the potential

causes of psychopathy (Hare, 1998; Lykken, 1995; Patrick, 1994). However, it is

also possible that such measures of behavioral inhibition predict the stability of these

traits, with the most temperamentally vulnerable individuals showing the most

stable manifestations of psychopathy. Third, while psychopathy and antisocial

behavior are not interchangeable constructs, they are clearly related. Therefore,

some of the factors within the child (e.g. low intelligence) and factors within his or

her psychosocial context (e.g. low socioeconomic status, dysfunctional parenting

practices) that have been found to predict a more stable pattern of antisocial

behavior in youth (see Frick & Loney, 1999, for a review) may also be related to

the stability of the psychopathic traits. The potential importance of parenting factors

is particularly notable, given evidence that parenting has a strong influence on the

development of guilt, empathy, and other aspects of conscience in children (Frick &

Morris, in press; Hoffman, 1994; Kochanska, 1995).

METHOD

Participants

To address the primary study questions, a community sample of youth was

recruited. However, one problem with using a community sample of youth is the

need to ensure that there are significant numbers of participants who score high on

measures of psychopathic traits. While there are no normative studies estimating the

prevalence of these traits, only about one-quarter of adult prisoners (Hart & Hare,

1997) and about one-third of clinic-referred children (Christian et al., 1997) show

high levels of these traits. Therefore, rates in a non-referred sample are likely to be

quite low. Further, in order to test the potential effects of antisocial behavior on the



stability of these traits, there had to be enough participants high on psychopathic

traits with and without significant antisocial behavior to make comparisons. As a

result, a two-step stratified random sampling procedure was employed to over-

sample youth high on psychopathic traits and to ensure sufficient numbers of these

youth with and without conduct problems.

In the first step of subject recruitment, 4,000 parents of children in the third,

fourth, sixth, and seventh grades of two school systems in a moderate sized city in

the southeastern United States received announcements about the study. The two

school systems were chosen because one served the immediate urban area and the

second served the surrounding region that was predominantly suburban and rural.

Those parents who agreed to have their child participate in the study completed

consent forms and a screening questionnaire (Gadow & Sprafkin, 1995) that

assessed for the presence of DSM-IV symptoms of Oppositional Defiant Disorder

(ODD) and Conduct Disorder (CD; American Psychiatric Association, 2000).

Parents also completed a questionnaire that assessed for each of three core

dimensions of psychopathy (Frick & Hare, 2001). Following receipt of the parents’

consent forms and rating scales, the child’s teacher completed analogous ques-

tionnaires. For each child who participated in this initial phase of screening, his or

her teacher received $10 for educational supplies for the classroom. This first phase

yielded a sample of 1136 children with a mean age of 10.65 (SD¼ 1.60) that was

53% female, 77% Caucasian, 19% African-American, and 21% receiving special

education services, all of which closely matched the overall demographics of the two

school systems. The range of Duncan’s Socioeconomic Index (SEI; Hauser &

Featherman, 1977) was 0–92.3, with a mean of 47.20 (SD¼ 23.80).

In the second phase of recruitment, the sample of 1136 children was divided into

four groups based on combined parent and teacher ratings of conduct problem

symptoms and the callous–unemotional (CU) dimension of psychopathy (Frick

et al., 2000). The CU dimension was chosen for sample recruitment because it is the

dimension of psychopathy that is most independent from antisocial behavior (Frick,

O’Brien, Wootton, & McBurnett, 1994; Frick et al., 2000). This allowed for the

recruitment of a group of children high on psychopathic traits but low on conduct

problems. Further, elevated rates of CU traits and conduct problems were defined

as being above the upper quartile for the community sample, whereas normative

levels were defined as being at or below the mean of the sample. This ensured

divergence between groups defined as being high and low on these dimensions.

The first group was below the mean on both dimensions (n¼ 225), a second

group was at or above the upper quartile on the conduct problem measure but below

the mean on the measure of CU traits (n¼ 66), one group was at or above the upper

quartile on the measure of CU traits but below the mean on the measure of conduct

problems (n¼ 77), and the last group was above the upper quartile on both

dimensions (n¼ 128). These four groups were then stratified on gender, ethnicity,

and socioeconomic status. Next, 25 children in each of the four groups were

recruited to participate in the longitudinal portion of the study, ensuring that about

half of each group came from the younger and older cohorts. These children were

selected through a stratified random sampling procedure to ensure that the four

groups matched the group from which they were sampled on the three stratification

variables. Errors in data collection resulted in the loss of two participants from the

group high on conduct problems but low on the measure of CU traits. From this



recruitment procedure, participants were 98 children that fell into the four groups

described above. Twenty-one per cent of the children were African-American (the

only ethnic group represented other than Caucasian); 47% were girls; the average

Duncan’s SEI was 46.67 (SD¼ 19.96). Children in the sample had an average

Kaufman Brief Intelligence (K-BIT; Kaufman & Kaufman, 1990) composite score

of 104.83 (SD¼ 12.88).

This sample of 98 children was invited to participate in yearly assessments. A

summary of the measures and sources of information is provided in Table 1. The

child’s score on the measure of psychopathy collected during the community

screening (time 1) was used as the initial level of these traits. At the first follow-

up (time 2), a comprehensive assessment was conducted and it included all of the

potential predictors of stability used in analyses. The last three waves of data

collection (times 3, 4, and 5) involved repeated ratings of psychopathic traits.

Measures: Repeated Assessment of Psychopathic Traits

Antisocial Process Screening Device (APSD; Frick & Hare, 2001)

The APSD, formerly known as the Psychopathy Screening Device, is a 20-item

behavior rating scale that was completed by each child’s parent and teacher at the first

assessment and then by each child and his or her parent at the last three follow-up

assessments (see Table 1). The APSD was designed to measure the same traits as

those assessed by the Psychopathy Checklist—Revised (PCL-R; Hare, 2003).

However, on the APSD, these traits are placed in a rating scale format and rated as

0 (Not at all true), 1 (Sometimes true), or 2 (Definitely true). Also, some items from

the PCL-R were deleted because they were inappropriate for children (e.g. parasitic

lifestyle) or the wording of items was changed to be more developmentally appro-

priate (Frick et al., 1994; Frick & Hare, 2001). Also, at the initial assessment, the

APSD item related to criminal behavior was not included at the request of the

participating school systems. Factor analyses conducted on the items collected in the

full community screening sample of 1136 children, and replicated in a sample of

clinic-referred children, revealed three dimensions on the APSD: a seven-item

Narcissism dimension (‘‘thinks he/she is more important than others’’), a five-item

Impulsivity dimension (e.g. ‘‘acts without thinking’’), and a six-item Callous–

Unemotional dimension (‘‘is concerned about the feelings of others’’—reverse

scored) (Frick et al., 2000). One item, ‘‘lies easily and skillfully’’ did not show unique

or consistent loadings and was not included in the calculation of subscale scores.

The choice of informants for the APSD at each stage of data collection was based

on developmental considerations and findings from past research. Parents and

teachers were included as informants at the initial assessment because the majority

of the sample was pre-adolescent at this time and parents and teachers seem to be

optimal informants for this age group in the assessment of many areas of psycho-

logical adjustment (Kamphaus & Frick, 2002). Further, there is evidence for the

construct validity of these informants on the APSD for pre-adolescent samples

(Frick & Hare, 2001). However, at later follow-up assessments, much of the sample

had transitioned into adolescence, when the validity of teacher report decreases and

the validity of self-report increases in the assessment of most areas of psychological



Table

1.

Su

mm

ary

of

mea

sure

sacr

oss

waves

of

data

collec

tion

Con

stru

ctM

easu

reM

ethod

/in

form

an

ts

Tim

e1:

com

munity

wid

esc

reen

ing

Psy

chophath

ictr

ait

sA

nti

soci

al

Pro

cess

Scr

een

ing

Dev

ice

(AP

SD

:F

rick

&H

are

,2001)

Pare

nt

an

dT

each

erT

ime

2:

com

pre

hen

sive

ass

essm

ent

ofpre

dic

tors

ofst

abi

lity

Dem

ogra

ph

icvari

ab

les

Age/

eth

nic

ity/g

end

erP

are

nt

Du

nca

n’s

Soci

oec

on

om

icIn

dex

(Hau

ser

&F

eath

erm

an

,1977)

Pare

nt

Kau

fman

Bri

efIn

tellig

ence

Tes

t(K

-BIT

;K

au

fman

&K

au

fman

,1990)

Sta

nd

ard

ized

test

Beh

avio

ral

inh

ibit

ion

Com

pu

teri

zed

Rew

ard

Dom

inan

ceT

ask

(O’B

rien

&F

rick

,1996)

Lab

ora

tory

task

Ch

ild

ren

’sS

ensa

tion

See

kin

gS

cale

(SS

SC

;R

uss

oet

al.,

1993)

You

tha)

Thri

llan

dA

dven

ture

See

kin

gS

ubsc

ale

Con

du

ctpro

ble

ms/

an

tiso

cial

beh

avio

rD

iagn

ost

icIn

terv

iew

Sch

edu

lefo

rC

hild

ren

—IV

Ed

itio

n(S

haff

er,

Fis

her

,L

uca

s,D

ulc

an

,&

Sch

wab

-Sto

ne,

2000)

Pare

nt

an

dyou

tha)

Opposi

tion

al

Defi

an

tan

dC

on

du

ctD

isord

erS

ym

pto

ms

Sel

f-R

eport

of

Del

inq

uen

cysc

ale

(SR

D;

Ellio

tt&

Aget

on

,1980)

You

thP

are

nti

ng

pra

ctic

esA

lab

am

aP

are

nti

ng

Qu

esti

on

nair

e(S

hel

ton

,F

rick

,&

Woott

on

,1996)

Pare

nt

an

dyou

tha)

Posi

tive

Pare

nti

ng¼

Pare

nta

lIn

volv

emen

tan

dP

osi

tive

Rei

nfo

rcem

ent

b)

Neg

ati

ve

Pare

nti

ng¼

Poor

Mon

itori

ng/S

uper

vis

ion

,In

con

sist

ent

Dis

cip

lin

e,H

ars

hD

isci

plin

eT

imes

3–5:

follo

w-u

pass

essm

ents

Psy

chopath

ictr

ait

sA

nti

soci

al

Pro

cess

Scr

een

ing

Dev

ice

(AP

SD

:F

rick

&H

are

,2001)

Pare

nt

an

dyou

th



adjustment (Kamphaus & Frick, 2002). Again, a number of studies have supported

the construct validity of self-report on the APSD for adolescents (Loney et al., 2003;

Pardini, Lochman, & Frick, 2003; Silverthorn, Frick, & Reynolds, 2001). The

parent report was also collected at the follow-up assessments to maintain one

informant that was consistent across assessments.

The distributions, internal consistency estimates, and cross-informant correlations

for the APSD total score and subscales at each wave of data collection are reported in

Table 2. It is of note that the internal consistency estimates were generally moderate

to high for the parent and teacher scales. However, the internal consistency estimates

for the self-report scales were lower, especially for the individual subscales. The cross-

informant correlations, using the single measure intraclass correlation coefficient

(absolute agreement definition) ranged from 0.27 (Narcissism) to 0.39 (Impulsivity)

between parents and teachers and from 0.29 (time 4 Narcissism) to 0.57 (time 4

Impulsivity) between parents and children (all p< 0.01).

Measures: Predictors of Stability (Antisocial Behavior)

The NIMH Diagnostic Interview Schedule for Children—Version 4

(DISC-IV; Shaffer et al., 2000)

The parent and child report on the DISC-IV was used to assess for all symptoms

of ODD and CD at the time 2 comprehensive assessment. The DISC-IV is a

highly structured interview designed to be administered by lay interviewers with

Table 2. Descriptive statistics for dimensions of psychopathy across waves of data collection

Parent report Teacher (time 1)/youthreport (time 3, 4, 5)

Mean Std. Alpha Mean Std. Alpha Cross-informantAPSD scale ICC

Time 1 (n¼ 98)Total 10.48 6.74 0.89 11.14 8.51 0.92 0.36***Callous–Unemotional 2.70 2.50 0.76 3.66 2.97 0.82 0.34***Narcissism 3.23 2.98 0.82 3.28 3.57 0.90 0.27**Impulsivity 4.03 2.06 0.72 3.66 2.83 0.83 0.39***Time 3 (n¼ 92)Total 10.37 6.32 0.88 10.71 5.14 0.81 0.54***Callous–Unemotional 2.41 2.05 0.74 2.72 1.66 0.50 0.50***Narcissism 3.22 2.87 0.82 3.15 2.04 0.63 0.31**Impulsivity 4.08 2.10 0.73 4.13 2.01 0.68 0.54***Time 4 (n¼ 86)Total 10.37 6.50 0.89 11.54 5.24 0.80 0.54***Callous–Unemotional 2.43 2.04 0.75 3.05 1.87 0.60 0.52***Narcissism 3.43 2.69 0.80 3.53 2.10 0.63 0.29**Impulsivity 4.10 2.15 0.75 4.13 2.01 0.68 0.57***Time 5 (n¼ 79)Total 10.22 5.68 0.86 11.51 4.78 0.77 0.47***Callous–Unemotional 2.69 1.98 0.72 3.01 1.81 0.52 0.34**Narcissism 3.09 2.51 0.78 3.52 1.94 0.55 0.32**Impulsivity 3.88 1.98 0.63 4.10 1.79 0.60 0.39***

APSD, Antisocial Process Screening Device (Frick & Hare, 2001); ICC, single measure intraclasscorrelation coefficient (absolute agreement definition). *p<0.05; **p<0.01; ***p<0.001.



appropriate training. It has proven to be highly reliable on both the symptom and

diagnostic level (Lahey et al., 1994). Interviewers were a licensed psychologist or

advanced graduate students in psychology who were trained in standardized

administration procedures for the DISC. Symptoms were considered present if

endorsed by either the parent or child. The correlations between parent and child

report of ODD/CD symptoms was 0.25 ( p< 0.05) and the coefficient alpha for the

combined report of OCC/CD symptoms was 0.87 in the current sample.

Self-Report of Delinquency Scale (SRD; Elliott & Ageton, 1980)

The SRD assesses the child’s self-report of 36 illegal juvenile acts. It was developed

from a list of all offenses reported in the Uniform Crime Report with a juvenile base

rate of greater than 1% (Elliott & Huizinga, 1984). Due to the relatively young age of

the sample at the time 2 comprehensive assessment, the three items related to sexual

behavior were omitted from this scale. Consistent with past uses of the scale (see,

e.g., Krueger et al., 1994), a composite measure was created summing the number

of delinquent acts committed (with a possible range of 0–33). This composite had a

coefficient alpha of 0.74.

Measures: Predictors of Stability (Behavioral Inhibition)

Reward Dominance Task (O’Brien & Frick, 1996)

Participants’ sensitivity to punishment cues once a reward-oriented response set is

established was assessed using the paradigm employed in previous research with

adults (e.g., Newman, Patterson, & Kosson, 1987) and children (Daugherty &

Quay, 1991). A more complete description of the specific computer task can be

found elsewhere (O’Brien, Frick, & Lyman, 1994; O’Brien & Frick, 1996). The task

consists of four games and participants were allowed to earn prizes based on their

scores across all four games. In each game, a stimulus (e.g. a fisherman) appeared on

the screen and the child chose whether or not to press a key to view the other side of

the stimulus (e.g. reel in the fishing pole) or to press a key to stop the game. For each

game, there was a successful outcome (e.g. a fish) or an unsuccessful outcome (e.g.

no fish) on the other side of the stimulus. Each child began with 50 points, and a

point was either added or taken away, depending on the outcome of each trial. The

proportion of successful outcomes across each successive ten trials decreased from

90 to 0% over 100 trials. The total number of trials played was recorded and was

used as a dependent variable in the analyses because this serves as an index of

continued playing despite an increased ratio of punished (loss of point) to rewarded

(gain of point) trials. Across the four games, a forced 5 second pause between each

trial and the presence of a visual tally of the number of points won were varied in a

counterbalanced fashion. As in some previous studies, the forced pause appeared to

interrupt the reward dominant response set (Newman et al., 1987; O’Brien et al.,

1994). Therefore, only the two conditions in which there was no forced pause prior

to being allowed to play the next trial were used in analyses. As in previous uses of

this computer task (O’Brien et al., 1994; O’Brien & Frick, 1996), no effect of order

or order by condition interactions were found on the task.



Thrill and Adventure Seeking Subscale of the Sensation Seeking Scale for Children(SSS-C; Russo et al., 1993)

The SSS-C was used to measure participants’ preferences for novel and dangerous

activities. The scale was modified for use in the study. As in the original version of

the scale, participants chose between a pair of statements to indicate which was

more true of him/her. One statement (e.g. ‘‘I enjoy the feeling of riding my bike fast

down a big hill’’) describes sensation-seeking behaviors, whereas the other state-

ment describes a preference against sensation seeking behaviors (e.g. ‘‘Riding my

bike fast down a hill is scary for me’’). However, to increase the variance in scores,

the modified version also asked each child to rate how well the chosen behavior

described him or her by selecting either ‘‘sort of true for me’’ or ‘‘really true for me.’’

This modification created a four point scale for each item. The modified version of

SSS-C led to an internally consistent TAS subscale (alpha¼ 0.84). Although the full

SSS-C was administered, only the TAS scale was used in analyses. Its content

assesses fearlessness, the aspect of sensation seeking most specifically linked to

behavioral inhibition (Kagan & Snidman, 1991). Also, it is the scale from the SSS-C

that has been most strongly related to psychopathic traits in children (Frick et al.,

1994; Frick et al., 1999b).

Measures: Predictors of Stability (Parenting)

Alabama Parenting Questionnaire (APQ; Shelton et al., 1996)

The APQ includes 42 items assessing five parenting constructs that have been

consistently related to antisocial and delinquent behavior in past research (Shelton

et al., 1996). These constructs are parental involvement (10 items); parental use of

positive reinforcement (6 items); poor parental monitoring and supervision

(10 items), parental use of inconsistent discipline (6 items), and parental discipline

(10 items; 3 pertaining to corporal and 7 to non-corporal methods of discipline).

There are four assessment formats for the APQ. There is a global report in which

parenting behaviors are rated by the parents and children as to their typical

frequency on a five-point scale from 1 (‘‘Never’’) to 5 (‘‘Always’’). In the telephone

format, parents and children are called four times at least three days apart and asked

how often in the past three days each parenting behavior had taken place. The score

for each item is the average frequency across the four interviews.

Past research has indicated that the parenting constructs can be combined into

two composites: a positive parenting composite involving the parental involvement

and use of positive reinforcement scales and a negative composite involving the poor

monitoring and supervision, inconsistent discipline, and corporal punishment scales

(Shelton et al., 1996; Frick, Christian, Wootton, 1999a). To form these composites,

subscales were converted to z-scores and summed within each of the assessment

formats. To reduce the number of variables further, the composites of the global

report and interview formats were also combined for each informant. For the parent

report, the correlation between the global and interview formats for the positive

parenting composite was 0.44 ( p< 0.001) and that for the negative parenting



composite was 0.37 ( p< 0.001). For the child report, the correlations were 0.44

( p< 0.001) and 0.49 ( p< 0.001), respectively. However, the decision was made a

priori not to collapse across informants. This decision was based on past research

suggesting that the child report of positive parenting and the parent report of

negative parenting may be most strongly associated with problem behavior in late

childhood and early adolescence (Frick et al., 1999a). In the current sample, the

correlations between the child and parent reported composites were 0.44

( p< 0.001) and 0.26 ( p< 0.05) for the positive and negative parenting composites,

respectively.

Procedures

Using the stratified sampling procedure described above, parents and children who

participated in the community-wide screening were contacted and invited to

participate in a longitudinal study of children’s personality that included the

measures used in the current study (see Frick et al., 2003b, for a more complete

description of the full assessment procedures). Those who were contacted but

refused to participate were replaced by someone in the same group with similar

demographic characteristics until 25 participants had been recruited for each group.

Two participants were lost due to errors in data collection, and both came from the

high conduct problem and low CU traits group.

For the time 2 comprehensive assessment, participants were tested in two

sessions with the procedures standardized for all participants. The first session

started with an informed consent procedure conducted with the parent and child

together. They were then separated and parents were administered a semi-

structured interview to obtain demographic information, followed by the DISC-

IV interview and the APQ ratings. In a separate room, the children were adminis-

tered the K-BIT, followed by the youth version of the DISC-IV. At a second testing

session, the child completed all of the self-report questionnaires and the Reward

Dominance Task. Within one month of this assessment, all four APQ telephone

interviews were completed. Parents received $65.00 for their participation in the

comprehensive assessment procedures and the youth received a $15.00 gift certi-

ficate to either a local music store or book store. The remaining three follow-up

assessments took place as close to the 1 year anniversary of the initial assessment as

possible. To reduce attrition, the APSD measures were mailed to participants with a

self-addressed stamped envelope as part of the follow-up assessments. Parents and

youth were provided with separate return envelopes to increase the confidentiality of

each person’s responses. Parents and youth were also compensated for participation

in the follow-up assessments.

RESULTS

Attrition

Ninety-two per cent of the 98 participants at the time 2 comprehensive assessment

completed at least two of the follow-up assessments and 73% completed all

three. The average length of time between the initial screening assessment and first



follow-up assessment was 25.43 months (SD¼ 4.96). Subsequent follow-up assess-

ments were scheduled as close to the 1 year anniversary of the first follow-up

assessment as possible, with the average length between the follow-up assessments

being 12.63 months (SD¼ 1.82) and 13.38 months (SD¼ 2.82), respectively. This

led to the average length of time between the screening and the third follow-up

assessment being 50.91 (SD¼ 4.44). Importantly, there were no differences in

attrition rates between the groups high and low on CU traits, with 92% of both

groups completing at least two follow-up assessments and with 74% and 71%,

respectively, completing all three follow-up assessments. Seventy-nine participants

completed the last follow-up assessment and this group was compared with the 19

participants who participated in the time 2 comprehensive assessment but did not

complete the last follow-up on each of the predictors of stability tested in analyses.

The only variable on which the groups differed was on Duncan’s SEI

(t(df¼ 96)¼ 2.06, p< 0.04; eta-squared¼ 0.04), with the group not completing

the last assessment being from a significantly higher socioeconomic status

(M¼ 61.01; SD¼ 19.48) than the group completing the last follow-up assessment

(M¼ 48.80; SD¼ 19.70). None of the effects for other predictors of stability

approached significance, with effect sizes (eta-squared) ranging from 0.00 to 0.02.

Stability of Psychopathic Traits

The stabilities of psychopathic traits across 2, 3, and 4 years are provided in Table 3.

The stability estimates across time for parent report were generally quite high, with

Table 3. Stability of psychopathic traits

APSD total CU Narcissism ImpulsivityICC ICC ICC ICC

Parent report of psychopathic traits at time 1Within-informant stability (parent report)

Time 3 0.88*** 0.76*** 0.88*** 0.86***Time 4 0.87*** 0.86*** 0.84*** 0.73***Time 5 0.80*** 0.71*** 0.77*** 0.72***Stability 0.93*** 0.90*** 0.92*** 0.87***

Cross-informant stability (youth report)Time 3 0.46** 0.44** 0.36* 0.52**Time 4 0.53** 0.62*** 0.31* 0.56***Time 5 0.51** 0.46** 0.38* 0.43*Stability 0.79*** 0.75*** 0.65*** 0.79***

Teacher report of psychopathic traits at time 1Cross-informant stability (parent report)

Time 3 0.61*** 0.50*** 0.50*** 0.52**Time 4 0.63*** 0.56*** 0.41** 0.44**Time 5 0.59*** 0.50*** 0.47** 0.36*Stability 0.86*** 0.80*** 0.82*** 0.75***

Cross-informant stability (youth report)Time 3 0.43** 0.39** 0.19 0.32*Time 4 0.50*** 0.50** 0.07 0.35*Time 5 0.49** 0.30 0.16 0.39*Stability 0.76*** 0.67*** 0.52** 0.71***

CU, Callous–Unemotional Traits; APSD, Antisocial Process Screening Device (Frick & Hare, 2001).ICC, average measure intraclass correlation coefficient (absolute agreement definition); *p<0.05;**p<0.01; ***p<0.001.



average measure intraclass correlation coefficient (absolute agreement definition)

estimates of 0.93 ( p< 0.001) across all time points for the total APSD scores.

Overall stability estimates ranged from 0.87 (Impulsivity) to 0.92 (Naricissism)

(both p< 0.001) for the individual scales. These coefficients suggest that a sub-

stantial amount of variance in parent ratings is consistent across time. The intraclass

correlation coefficients (ICC) for the four year stability of parenting ratings ranged

from 0.71 for the CU scale of the APSD to 0.80 for the total APSD score (all

p< 0.001).

The cross-informant stability estimates reported in Table 3 were lower but still

indicated a substantial amount of variance was consistent across time. For example,

the ICC stability estimates for parent report at time 1 predicting later youth self-

reports of psychopathy ranged from 0.65 for the Narcissism scale of the APSD to

0.79 for the total and the Impulsivity scale of the APSD (all p< 0.001). The four

year coefficients ranged from 0.38 ( p< 0.05) for Narcissism to 0.51 ( p< 0.001) for

the total score. Similarly, for teacher report at time 1 predicting later parent report,

the stability estimates ranged from 0.75 to 0.86 (all p< 0.01) across all assessments,

with the 4 year stability coefficients ranging from 0.36 ( p< 0.05) to 0.59

( p< 0.001). The cross-informant correlations for teacher report predicting later

youth self- report were somewhat more variable, due largely to the much lower

cross-informant stability for the Naricissism dimension, where there was an ICC of

0.52 ( p< 0.01) across all assessments, with a 4 year ICC of 0.16 (ns).

Importantly, these cross-informant stability coefficients confound variation in

stability with variation due to different raters. As noted in Table 2, the cross-

informant ICC for parent and teacher report on the APSD ranged from 0.27

( p< 0.001) for the Narcissism scale to 0.39 ( p< 0.001) for the total APSD score.

Across the three follow-up assessments, the average cross-informant ICC for parent

and youth report ranged from 0.31 for the Narcissism scale to 0.52 for the total

score. When the stability coefficients were corrected for attenuation based on these

estimates of cross-informant agreement (Ghiselli, Campbell, & Zedeck, 1981), all of

the overall estimates of stability met or exceeded 1.00 and only three of the

individual ICC estimates fell below 1.00. Such corrections could not be conducted

for the teacher reports predicting later self-report because there was no estimate of

concurrent correlations between teacher and self-reports. However, these analyses

suggest that much of the variation in the cross-informant correlations across time

could be attributed to the level of agreement between informants.

For these overall analyses of stability, the younger (grades 3 and 4 at the initial

assessment) and older (grades 5 and 6 at the initial assessment) cohorts were

collapsed to estimate stability. When separate stability estimates were conducted for

the two cohorts, very similar stability estimates were obtained for the parent ratings

across time. Specifically, the ICC across the four time points for the total APSD

score was 0.93 in the younger cohort and 0.91 in the older cohort (both p< 0.001)

and this was similar across the APSD subscales, with ICCs ranging from 0.91 to

0.92 in the younger cohort and 0.81 to 0.91 in the older cohort (all p< 0.001). For

teacher report predicting later parent report, the stability estimate was somewhat

stronger for the younger (0.89) than older (0.80) cohorts (both p< 0.001) and this

was largely due to the higher ICC estimates for the Impulsivity scale of the APSD in

the younger (0.81) compared with the older (0.64) cohort (both p< 0.001). For

both parent and teacher report predicting later youth self-report, the opposite



pattern was found, with stability estimates being somewhat lower in the younger

cohort than in the older cohort on the total APSD score (0.70 and 0.68 for the

younger cohort and 0.87 and 0.82 for the older cohort, respectively; all p< 0.001).

These lower estimates for cross-informant ratings in the younger cohort were fairly

consistent across the various APSD subscales.

The initial test of potential predictors of stability involved correlating each of the

hypothesized predictors with time 5 APSD scores, after controlling for time 1 APSD

scores. Thus, while not being direct predictors of stability, these analyses determine

whether variables were associated with change in APSD scores from the initial

assessment to the last follow-up. These partial correlations are provided in Table 4.

Only four variables predicted total APSD scores, after controlling for the initial level

of these traits. Socioeconomic status was negatively associated with later APSD

scores (�0.39, p< 0.01), whereas conduct problems (0.30, p< 0.05), total delin-

quency (0.23, p< 0.05), and the negative parenting composite (0.31, p< 0.05) were

all positively associated with later APSD scores. A similar pattern of results emerged

across APSD subscales. However, the partial correlations for total conduct pro-

blems and total delinquency did not reach significance in predicting CU traits (0.21

and 0.20) or impulsivity (0.14 and 0.17), although both were in the expected

direction. Also, socioeconomic status did not reach significance for predicting

impulsivity (�0.15), although it also showed the same trend as the other APSD

scales. Another consistent pattern of note is that the positive parenting composite by

child report was consistently negatively correlated with the APSD scores (ranging

from �0.11 to �0.22), although none of these correlations reached significance.

Table 4. Predictors of psychopathic traits controlling for initial level

Time 5 APSD scores

Time 2 predictors Total Callous–Unemotional Narcissism Impulsivity

DemographicAge �0.06 0.13 �0.14 �0.08Sex �0.15 �0.07 �0.12 �0.11Ethnicity �0.06 �0.06 �0.09 �0.06Socioeconomic status �0.39** �0.46*** �0.32* �0.15K-BIT Composite 0.09 �0.09 0.07 0.17

Antisocial/conductODD/CD symptoms 0.30* 0.21 0.32* 0.14Total delinquency 0.23* 0.20 0.06 0.17

Behavioral inhibitionReward Dominance 0.03 0.11 0.08 0.00Thrill-Seeking 0.05 �0.13 0.06 0.11

ParentingPositive Parenting (P) �0.07 �0.21 0.02 �0.03Positive Parenting (Y) �0.21 �0.21 �0.11 �0.22Negative Parenting (P) 0.31* 0.27* 0.29* 0.28*Negative Parenting (Y) 0.20 0.25* 0.18 0.08

Correlations reported in body of table are partial correlations controlling for time 1 scores from APSD.APSD, Antisocial Process Screening Device (Frick & Hare, 2001). Sex was coded Boy¼1 and Girl¼2;ethnicity was coded Caucasian¼1 and African American¼ 2; socioeconomic status, Duncan’sSocioeconomic Index (Hauser & Featherman, 1977); K-BIT, Composite Index from the KaufmanBrief Intelligence Test (Kaufman & Kaufman, 1990); ODD, Oppositional Defiant Disorder; CD,Conduct Disorder; (P) parent report; (Y) youth report; *p< 0.05; **p< 0.01; ***p< 0.001.



Patterns of Stability within the Sample

Previously, we provided our justification for oversampling groups of children with

CU traits and conduct problems. However, this methodology led to the recruitment

of extreme groups that distorted the distribution of CU traits in the sample. Given

the correlations among APSD dimensions (0.42–0.63, all p< 0.001, for parent

report at time 1 and 0.47–0.72, all p< 0.001, for teacher report at time 1), it is likely

that this methodology also distorted the distribution of all the APSD scales some-

what. Thus, a second method of assessing stability and determining predictors of

stability was used, which (i) utilized the extreme group sample recruitment design

and (ii) utilized normative cut-offs from the large community sample screened to

recruit the current study participants. Also, these analyses were limited to the 71

children who (i) were either above the upper quartile of the community sample

according to parental report of CU traits or at or below the mean according to

parental report of CU traits at the time 1 assessment and (ii) provided data at each of

the last three follow-up assessment points. This led to two groups of children, one

that was initially rated as being high on CU traits (n¼ 37) and one that was initially

rated as low on these traits (n¼ 34). Five patterns of stability were assessed for these

two groups. One pattern of stability (n¼ 24) included children who scored at or

below the full community sample mean on parental ratings of CU traits at all three of

the follow-up assessment points. Three additional groups designated those who

crossed the upper quartile of CU traits based on the full community sample mean at

either one (n¼ 6), two (n¼ 3), or all three (n¼ 11) follow-up assessments. A final

group consisted of those children who did not fall into any of these stability groups

(n¼ 27).

The distribution of these patterns of stability across the two groups differing on

their initial level of CU traits is provided in Table 5. As evidenced in this table, the

patterns of stability were not distributed equally across the two groups

(�2(df¼ 4)¼ 24.40, p< 0.001). Specifically, 83% (n¼ 20) of the group that showed

stable low scores across follow-up assessments were in the group who initially scored

low on CU traits. Further, all 11 of the children that exceeded the normative cut-off

at all three assessment points came from the group that was initially rated by parents

as high on CU traits. Also, it was less likely for children who initially scored low on

CU traits to obtain high scores at later assessment points than it was for children

Table 5. Continuity in children high and low on callous–unemotional traits at time 1

Time 1 status

Low callous–unemotional (n¼ 34) High callous–unemotional (n¼37)

n (column/row) n (column/row)

Stable low n¼20 (59%/83%) n¼4 (11%/17%)High at one follow-up n¼2 (6%/33%) n¼4 (11%/67%)High at two follow-ups n¼2 (6%/67%) n¼1 (3%/33%)High at three follow-ups n¼0 (0%/0%) n¼11 (30%/100%)Other n¼10 (29%/37%) n¼17 (46%/63%)

Relationship between time 1 status and stability groups was significant (�2 (df¼4)¼24.40, p< 0.001).



who were initially high on CU traits to score lower on these traits at later

assessments. Specifically, of those children who were initially rated low on CU

traits, only 4 (12%) were rated high on CU traits at any later point in time and only

19 (29%) even fell between the mean and upper quartile of the community sample at

any later assessment point. However, while it was rare for a child initially rated as

high on CU traits to score below the community sample mean at each follow-up

assessment (n¼ 4; 11%), a substantial number never crossed the normative cut-off

again at any of the follow-up assessments (n¼ 17; 46%).

Based on the patterns of stability reported in Table 5, there were eight children

who showed somewhat transitory elevations of CU traits, being rated as high on CU

traits by parents at the initial assessment but scoring below the community sample

mean at all assessment points or scoring high at only one follow-up point. In

contrast, there were 12 children who showed more stable patterns of CU traits,

initially being rated high and also being rated above the normative cut-off at two or

more follow-up assessments. These two groups of children were compared on all the

predictors of persistence to the 20 children who scored at or below the community

sample mean on CU traits at all assessment points. The only variables to differ-

entiate among these groups were socioeconomic status (F(2,37)¼ 6.88, p< 0.01;

eta-squared¼ 0.27), intelligence (F(2,37)¼ 5.87, p< 0.01; eta-squared¼ 0.24),

conduct problem symptoms (F(2,37)¼ 6.59, p< 0.01; eta-squared¼ 0.26), trials

played on the reward dominance task (F(2,37)¼ 3.81, p< 0.05; eta-

squared¼ 0.17), the positive parenting composite based on youth report

(F(2,35)¼ 3.43, p< 0.05; eta-squared¼ 0.16), and the negative parenting compo-

site based on youth report (F(2,35)¼ 5.91, p< 0.01; eta-squared¼ 0.25). Because

of the group differences in intelligence and socioeconomic status, the analyses for

the other variables that differentiated groups were repeated using intelligence and

socioeconomic status as covariates. The only two findings to remain significant were

for the level of conduct problems (F(4,35)¼ 4.61, p< 0.05; eta-squared¼ 0.21) and

the positive parenting composite based on youth report (F(4,33)¼ 3.57, p< 0.05;

eta-squared¼ 0.18).

The pattern of results for socioeconomic status, intelligence, and the two

variables that differentiated groups after controlling for these variables, are repre-

sented in Figure 1. As evident in this figure, the group that showed a decline in CU

traits (desisters) were very similar to the group that was low on CU traits (stable low)

across the entire study period. In fact, using Tukey’s procedures for pairwise

comparisons, there were no significant differences between these two groups on

any predictor of persistence. In contrast, the group that showed relatively stable

patterns of the CU traits (stable high) differed from the stable low group on all four

variables, and differed from the desister group on intelligence test scores and level of

conduct problems.

DISCUSSION

The primary purpose of the current study was to provide initial estimates of the

stability of psychopathic traits over an extended follow-up period (i.e. 4 years). For

parent report, which was the one method consistent across all follow-up assess-

ments, the composite index of psychopathic traits showed an average measure



intraclass correlation coefficient of 0.93 ( p< 0.001) across all time points, with

estimates ranging from 0.87 (Impulsivity) to 0.92 (Naricissism) (both p< 0.001) for

the individual dimensions of psychopathy. These coefficients suggest that a sub-

stantial amount of variance in parent ratings is consistent across time. Further, the

ICC estimates for the 4 year stability of parent ratings ranged from 0.71 for the CU

scale of the APSD to 0.80 for the Total APSD score (all p< 0.001). These

coefficients also indicate a substantial degree of stability in parent ratings, especially

given that these ICC estimates, in contrast to standard correlation coefficients, take

into account both changes in the relative ranking of a child’s scores across time and

also absolute changes in the child’s score over time.

To interpret these stability estimates, several comparisons can be made. First, as

reviewed in the introduction, there are few studies directly testing the stability of

psychopathic traits in adults. One such study reported 2 year stability estimates of

0.60 for men and 0.65 for women for total psychopathy scores (Rutherford et al.,

1999). While it is somewhat difficult to compare the ICC estimates obtained in the

current study with the correlation coefficients reported by Rutherford et al. (1999),

they do suggest that parent ratings of psychopathic traits in youth show comparable

or higher coefficients than those reported in adults. Second, the stability estimates

obtained in the current sample of youth can be compared with the stability of parent

ratings of other psychological traits in children. This is important for a number of

reasons. For example, it shows how comparable the stability of psychopathy is to

other forms of psychopathology in children. Further, it helps to determine whether

parents show variability in their ratings across time or whether their ratings are

inflated by stable perceptions that do not reflect actual changes in the child’s

adjustment. In two studies of the stability of parent ratings of children’s adjustment

in community samples over 3 years (ages 4–16; McConaughy, Stanger. &

Achenbach, 1992) or 4 years (ages 4–12; Verhulst, Koot, & Berden, 1990), the

mean stability estimates across different types of child adjustment were 0.46 (range

Figure 1. Variables that differentiated most and least stable groups. All variables are coverted to Z-scoresand the means for conduct problems and parenting are adjusted for the covariates intelligence and

socioeconomic status.



of 0.30–0.58) and 0.41 (range of 0.11–0.67), respectively. Again, these stability

estimates are much lower than those found for parent ratings of psychopathy in the

current study. Third, the estimates of stability obtained for psychopathic traits in the

current study can be compared with those found for other measures of personality in

both adults in children. A meta-analysis of 152 longitudinal studies providing 3,217

test–retest intervals reported average trait consistency of 0.31 for samples of children

and adolescents, 0.54 during the college years and young adulthood, 0.64 in middle

adulthood, and 0.74 between the ages of 50 and 70 (Roberts & DelVecchio, 2000).

Importantly, these meta-analytic estimates of the stability of personality traits were

adjusted to reflect a constant test–retest interval of 6.7 years, which is somewhat

longer than the 4 year interval used in the current study. However, the stability

estimates of psychopathic traits obtained in the current study appear to be somewhat

higher than is typically found for measures of personality in child and adolescent

samples and are comparable to those found in adult samples.

In the current study, the stability estimates were substantially lower when

different informants were used at initial and follow-up assessments. Specifically,

for parent-report predicting youth self-report of psychopathy, the overall stability

estimates ranged from 0.65 to 0.79 (all p< 0.01) for the various dimension of

psychopathy studied, with the 4 year ICC estimates ranging from 0.38 to 0.51.

However, these estimates confound variance related to temporal stability and

variance related to differences in raters. As noted in Table 2, the contemporaneous

cross-informant correlations on the dimensions of psychopathy were relatively

modest. Further, when the cross-informant stability coefficients were corrected

for attenuation based on the cross-informant reliability, it is clear that much of the

decrease in the magnitude of the stability estimates was due to differing reporters of

psychopathy.

The modest cross-informant correlations reported in this and other studies of

youth (see Frick & Hare, 2001, for a review) have sometimes been viewed as a

limitation in measures of psychopathic traits for youth (Seagrave & Grisso, 2002).

However, this level of cross-informant consistency is actually somewhat better than

is typically found when assessing childhood psychopathology, where correlations are

typically around 0.27 for parent and teacher ratings, around 0.25 for parent and

child self-report ratings, and around 0.20 for teacher and child self-report ratings

(Achenbach, McConaughy, & Howell, 1987). The cross-informant correlations

reported in the current study are, in fact, comparable to those found for differing

informants reporting on personality traits in adult samples. Specifically, the average

correlation between self- and spouse ratings of specific facets of personality ranged

from 0.24 to 0.66 with a mean of 0.39 in a sample of 47 married couples (McCrae,

Stone, Fagan, & Costa, 1998). Similarly, in a sample of college undergraduates

(n¼ 184), the average correlation between self-ratings of personality and ratings

made by parents, college acquaintances, and hometown acquaintances was reported

as being 0.37, 0.36, and 0.30, respectively (Funder, Kolar, & Blackman, 1995).

While these data support the stability of psychopathic traits in youth, they also

clearly indicate that there is some variability in the level of these traits across time.

Further, our data suggest that the pattern of change tends to be one in which some

youth high on psychopathic traits improve and show less severe levels of these traits

over time. It was less likely for youth who scored initially low on these traits in

childhood to develop significant levels of these traits later in childhood and



adolescence. The important question resulting from these findings is what factors

might predict these variations in stability? Our data suggest that factors related to

stability may be very different from the variables associated with initial levels of these

traits. Specifically, in the current sample (Frick et al., 2003b) and in other samples

(Barry et al., 2000; Blair, 1999; Fisher & Blair, 1998; Loney et al., 2003), measures

of behavioral inhibition (e.g. thrill seeking behavior, a reward dominant response

style) have been associated with psychopathic traits, especially the CU dimension.

However, in the current study, these measures showed little predictive utility after

controlling for initial levels of these traits.

Instead, the two most consistent predictors of stability across the different

methods for estimating stability were factors related to the child’s psychosocial

context, such as socioeconomic status and quality of parenting. These associations

may seem somewhat surprising, given past findings that measures of the quality of

parenting tend to be less related to the antisocial behavior of children with CU traits

than for children without these traits (Wootton, Frick, Shelton, & Silverthorn,

1999). However, this earlier study did not test the association between parenting

practices and CU traits, but focused on the association between parenting and

antisocial behavior in children with and without CU traits. Further, these results are

consistent with research demonstrating the importance of parenting in the devel-

opment of guilt and empathy and other aspects of conscience (Frick & Morris, in

press; Hoffman, 1994; Kochanska, 1995). Finally, it is quite possible that, although

a child’s temperament may play a larger role in placing a child at risk for problems in

conscience development (Frick & Morris, in press), the quality of later socialization

that is experienced by the child may play a more important role in determining

whether the child overcomes this risk later in development.

The only other variable to show some consistency in predicting the stability of

psychopathic traits in our non-referred sample was the child’s level of conduct

problems. That is, children with higher levels of conduct problem behavior tended

to have more stable levels of psychopathic traits, even after controlling for initial

levels of these traits. This finding suggest that studies examining psychopathic traits

in antisocial and criminal samples may overestimate the stability of these traits in

individuals who show less extreme antisocial behavior. Unfortunately, there is much

less research on individuals with psychopathic traits who do not also show antisocial

and criminal behavior (Hare, 1998; Lilienfeld & Andrews, 1996). However, our

findings suggest that this may be an important group to study, not only to determine

what prevents them from showing severe antisocial behavior despite their callous,

impulsive, and narcissistic traits, but also to determine what might make their

personality traits less stable over time. It is possible that chronically engaging in

antisocial, aggressive, and criminal behavior over time further desensitizes indivi-

duals to the consequences of their behavior on themselves and others, making their

callous personality traits more intractable.

All of these interpretations need to be made in the context of a number of

limitations. First, the current sample was very modest in size. This may have

prevented some significant predictors of stability from emerging due to a lack of

statistical power. Also, although there was little evidence for selective attrition, the

small initial sample magnifies the effects of even the minimal attrition that occurred

in our sample. The small sample also prevented us from using more sophisticated

statistical methods for estimating stability in our measure of psychopathic traits and



testing predictors of stability (e.g. latent growth curve methodology). Second, the

recruitment strategy was designed to ensure a significant number of children high on

psychopathic traits and a significant number of these children with and without

conduct problems. However, this strategy resulted in an atypical distribution of

scores in our sample, which could have influenced our correlations. Although we

tried to overcome this limitation by estimating stability and testing predictors of

stability using multiple methods, the atypical distribution created by this method of

recruitment could influence the generalizability of the results to other samples with

different distributions. For example, if the highest scorers on the measure of

psychopathy also show the highest level of stability, our sampling strategy that

oversampled such individuals could have inflated estimates of stability. On the other

hand, recruiting children with the highest scores increases the potential effects of

regression to the mean for later ratings and, thus, decreases the obtained stability

estimates. Third, the tests of predictors of stability were largely exploratory in

nature, given the absence of past research on such predictors. The nature of the

research question, as well as the modest sample size, led to our decision not to

control for inflated Type I rates in analyses. However, this methodology makes it

possible that some associations may not be replicated in other samples. Fourth, the

assessment of psychopathy relied on a rating scale method to assess psychopathy,

rather than on the strategy of combining interview results with historical records

used by the PCL-R (Hare, 2003). Because it does not rely on historical information,

this rating method allowed us to estimate change in scores across time. Further,

there is a body of research supporting the validity of this method of assessing

psychopathic traits in youth (Frick & Hare, 2001). However, it is unclear how well

these results will generalize to other methods of assessing psychopathy. Also, the

ratings at the follow-up assessment were conducted by mail and this may have

affected a child’s responses due to concerns about privacy.

A critically important limitation of the current findings is that, although our

results demonstrate substantial stability in psychopathic traits over a 4 year period in

older children and young adolescents, these data do not provide an estimate of how

many youth who score high on these traits in childhood will eventually meet

established criteria for a diagnosis of psychopathy as adults. Many of the concerns

about applying the construct of psychopathy to children are based on the possibility

that there will be an assumption that youth who score high on childhood measures of

psychopathy will be highly likely to be diagnosed as psychopathic as adults (Seagrave

& Grisso, 2002). Clearly, there are no data to determine the actual risk for adult

diagnoses in children who score high on psychopathic traits. More importantly,

there are important statistical, theoretical, and developmental issues to suggest that

one should expect very few children with psychopathic traits to be so diagnosed as

adults. First, as reviewed previously, much of the work on adult psychopathy has

focused on designating a small subgroup of incarcerated adults (Hart & Hare,

1997). As result, the existing diagnosis designates an outcome that is likely to be

found at a very low base rate in non-referred samples. Predicting very low base rate

phenomena is very difficult and almost always associated with high rates of false

positive predictions (Kraemer, 1992). Second, even in adults, there has been very

limited work on the expression of psychopathic traits in non-referred samples.

Further, there is still considerable debate as to whether these traits in samples of

individuals without severe antisocial behavior are measuring the same construct



(Hare, 1998; Lilienfeld & Andrews, 1996). As a result, even if these traits remain

stable into adulthood, this does not necessarily mean that this will strongly predict

the personality pattern associated with severe antisocial and violent behavior that

has been the focus of most of the research on psychopathy in adult samples. Third,

and perhaps most importantly, a hallmark of taking a developmental approach to

understanding psychopathological conditions is that, as one attempts to understand

the developmental processes that can lead to an outcome of interest, one should

assume that the pathways are changeable (Cicchetti & Rogosch, 1996). Further, the

earlier in development one attempts to study, the greater the degree of change that

should be expected. This is because of both the longer period of time during which

biological and contextual forces can operate to deflect children from a typical

pathway and the greater malleability of children’s personality earlier in development

(McCrae et al., 2002; Roberts & DelVecchio, 2000).

This last issue clearly illustrates both the great potential danger and the great

potential benefit of extending the construct of psychopathy to youth. If develop-

mental considerations are ignored, assumptions about the long term outcome of

these children can be made erroneously, with great potential consequences to the

child. If, however, the extension of the construct of psychopathy is made with

appropriate attention to developmental considerations, this research could provide

important information for designing more effective treatment for children with

severe conduct problems (Frick, 2001) and potentially prevent the development of a

personality disorder that can have a huge detrimental effect on society. Our findings

suggest that psychopathic traits are more stable than other problems in adjustment

experienced by youth and are comparable in stability to personality traits assessed in

adults. However, our results also indicate that some change in a child’s level of these

traits is possible over time and more research is needed to uncover what factors may

influence this change. Our exploratory results on predictors of stability suggest that

the factors influencing change may be quite different from those predicting the

initial level of psychopathic traits. Specifically, factors in the child’s psychosocial

ecology may play a more important role in the continuity of these traits than in the

initial risk for developing them. If such findings are replicated, they could have

important implications for designing more effective interventions for children with

psychopathic tendencies. Such interventions could focus not only on removing

factors that can place a child at risk for developing these traits, but also on factors

that may promote prosocial behavior in children who already manifest them.

REFERENCES

Achenbach, T. M., McConaughy, S. H., & Howell, C. T. (1987). Child/adolescent behavioral andemotional problems: Implications of cross-informant correlations for situational specificity. Psycholo-gical Bulletin, 101, 213–232.

American Psychiatric Association. (2000). The diagnostic and statistical manual of mental disorders (4thed.—text revision). Washington, DC: Author.

Barry, C. T., Frick, P. J., Grooms, T., McCoy, M. G., Ellis, M. L., & Loney, B. R. (2000). Theimportance of callous–unemotional traits for extending the concept of psychopathy to children. Journalof Abnormal Psychology, 109, 335–340.

Blair, R. J. R. (1999). Responsiveness to distress cues in the child with psychopathic tendencies.Personality and Individual Differences, 27, 135–145.

Brandt, J. R., Kennedy, W. A., Patrick, C. J., & Curtin, J. J. (1997). Assessment of psychopathy in apopulation of incarcerated adolescent offenders. Psychological Assessment, 9, 429–435.



Buss, A. H. (1995). Personality: Temperament, social behavior, and the self. Boston, MA: Allyn and Bacon.Caputo, A. A., Frick, P. J., & Brodsky, S. L. (1999). Family violence and juvenile sex offending: Potential

mediating roles of psychopathic traits and negative attitudes toward women. Criminal Justice andBehavior, 26, 338–356.

Christian, R. E., Frick, P. J., Hill, N. L., Tyler, L., & Frazer, D. R. (1997). Psychopathy and conductproblems in children: II. Implications for subtyping children with conduct problems. Journal of theAmerican Academy of Child and Adolescent Psychiatry, 36, 233–241.

Cicchetti, D., & Rogosch, F. A. (1996). Equifinality and mulifinality in developmental psychopathology.Development and Psychopathology, 8, 597–600.

Cleckley, H. (1976). The mask of sanity (5th ed.). St. Louis, MO: Mosby.Cooke, D. J., & Michie, C. (2001). Refining the construct of psychopathy: Towards a hierarchical model.

Psychological Assessment, 13, 171–188.Daugherty, T. K., & Quay, H. C. (1991). Response perseveration and delayed responding in childhood

behavior disorders. Journal of Child Psychology and Psychiatry, 32, 453–461.Edens, J., Skeem, J., Cruise, K., & Cauffman, E. (2001). The assessment of juvenile psychopathy and its

association with violence: A critical review. Behavioral Sciences and the Law, 19, 53–80.Elliott, D. S., & Ageton, S. S. (1980). Reconciling race and class differences in self-reported and official

estimates of delinquency. American Sociological Review, 45, 95–110.Elliott, D. S., & Huizinga, D. (1984). The relationship between delinquent behavior and ADM problems.

Boulder, CO: Behavioral Research Institute.Fisher, L., & Blair, R. J. R. (1998). Cognitive impairment and its relationship to psychopathic tendencies

in children with emotional and behavioral difficulties. Journal of Abnormal Child Psychology, 26, 511–519.

Forth, A. E., Hart, S. D., & Hare, R. D. (1990). Assessment of psychopathy in male young offenders.Psychological Assessment, 2, 342–344.

Frick, P. J. (1998). Conduct disorders and severe antisocial behavior. New York: Plenum.Frick, P. J. (2001). Effective interventions for children and adolescents with conduct disorder. The

Canadian Journal of Psychiatry, 46, 26–37.Frick, P. J. (2002). Juvenile psychopathy from a developmental perspective: Implications for construct

development and use in forensic assessments. Law and Human Behavior, 26, 247–253.Frick, P. J., Bodin, S. D., & Barry, C. T. (2000). Psychopathic traits and conduct problems in community

and clinic-referred samples of children: Further development of the Psychopathy Screening Device.Psychological Assessment, 12, 382–393.

Frick, P. J., Christian, R. C., & Wootton, J. M. (1999a). Age trends in the association between parentingpractices and conduct problems. Behavior Modification, 23, 106–128.

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

Frick, P. J., Cornell, A. H., Bodin, S. D., Dane, H. A., Barry, C. T., & Loney, B. R. (2003b). Callous–unemotional traits and developmental pathways to severe conduct problems. Developmental Psychology,39, 246–260.

Frick, P. J., & Hare, R. D. (2001). The antisocial process screening device (APSD). Toronto: Multi-HealthSystems.

Frick, P. J., Lilienfeld, S. O., Ellis, M. L., Loney, B. R., & Silverthorn, P. (1999b). The associationbetween anxiety and psychopathy dimensions in children. Journal of Abnormal Child Psychology, 27,381–390.

Frick, P. J., & Loney, B. R. (1999). Outcomes of children and adolescents with conduct disorder andoppositional defiant disorder. In H. C. Quay, & A. Hogan (Eds.), Handbook of disruptive behaviordisorders (pp. 507–524). New York: Plenum.

Frick, P. J., & Morris, A. S. (in press). Temperament and developmental pathways to severe conductproblems. Journal of Clinical Child and Adolescent Psychology.

Frick, P. J., O’Brien, B. S., Wootton, J. M., & McBurnett, K. (1994). Psychopathy and conduct problemsin children. Journal of Abnormal Psychology, 103, 700–707.

Funder, D. C., Kolar, D. C., & Blackman, M. C. (1995). Agreement among judges of personality:Interpersonal relations, similarity, and acquaintanceship. Journal of Personality and Social Psychology,69, 656–672.

Gadow, K. D., & Sprafkin, J. (1995). Manual for the Child Symptom Inventory (4th ed.). Stony Brook, NY:Checkmate Plus.

Gendreau, P., Goggin, C., & Smith, P. (2002). Is the PCL-R really the ‘‘unparalleled’ measure ofoffender risk? A lesson in knowledge accumulation. Criminal Justice and Behavior, 29, 397–426.

Ghiselli, E. E., Campbell, J. P., & Zedeck, S. (1981). Measurement theory for the behavioral sciences. SanFransisco: Freeman.

Hare, R. D. (1998). Psychopathy, affect, and behavior. In D. J. Cooke, A. E. Forth, & R. D. Hare (Eds.),Psychopathy: Theory, research, and implications for society (pp. 105–138). Dordrecht: Kluwer.



Hare, R. D. (2003). Hare Psychopathy Checklist—Revised (PCL-R): Second edition, technical manual.Toronto: Multi-Health Systems.

Harpur, T. J., & Hare, R. D. (1994). Assessment of psychopathy as a function of age. Journal of AbnormalPsychology, 103, 604–609.

Hart, S. D., & Hare, R. D. (1997). Psychopathy: Assessment and association with criminal conduct. InD. M. Stoff, J. Brieling, & J. Maser (Eds.), Handbook of antisocial behavior (pp. 22–35). New York:Wiley.

Hart, S. D., Watt, K. A., & Vincent, G. M. (2002). Commentary on Seagrave and Grisso: Impressions ofthe state of the art. Law and Human Behavior, 26, 241–246.

Hauser, R. M., & Featherman, D. L. (1977). The process of stratification. New York: Academic.Hemphill, J. F., Hare, R. D., & Wong, S. (1998). Psychopathy and recidivism: A review. Legal and

Criminological Psychology, 3, 139–170.Hoffman, M. L. (1994). Discipline and internalization. Developmental Psychology, 30, 26–28.Kagan, J., & Snidman, N. (1991). Temperamental factors in human development. American Psychologist,

46, 856–862.Kamphaus, R. W., & Frick, P. J. (2002). Clinical assessment of child and adolescent personality and behavior

(2nd ed.). Boston, MA: Allyn and Bacon.Kaufman, A. S., & Kaufman, N. L. (1990). Kaufman brief intelligence test manual. Circle Pines, MN:

American Guidance Service.Kochanska, G. (1995). Children’s temperament, mothers’ discipline, and security of attachment:

Multiple pathways to emerging internalization. Child Development, 66, 597–615.Kraemer, H. C. (1992). Evaluating medical tests. Newbury Park, CA: Sage.Krueger, R. F., Schmutte, P. S., Caspi, A., Moffitt, T. E., Campbell, K., & Silva, P. A. (1994).

Personality traits are linked to crime among men and women: Evidence from a birth cohort. Journal ofAbnormal Psychology, 103, 328–338.

Kruh, I. P., Frick, P. J., & Clements, C. B. (in press). Historical and personality correlates to the violencepatterns of juveniles tried as adults. Criminal Justice and Behavior.

Lahey, B. B., Applegate, B., Barkley, R. A., Garfinkel, B., McBurnett, K., Kerdyk, L., Greenhill, L.,Hynd, G. W., Frick, P. J., Newcorn, J., Biederman, J., Ollendick, T., Hart, E. L., Perez, D., Waldman,I., & Shaffer, D. (1994). DSM-IV field trials for oppositional defiant disorder and conduct disorder inchildren and adolescents. American Journal of Psychiatry, 151, 1163–1171.

Lilienfeld, S. O., & Andrews, B. P. (1996). Development and preliminary validation of a self-reportmeasure of psychopathic personality. Journal of Personality Assessment, 66, 488–524.

Loney, B. R., Frick, P. J., Clements, C. B., Ellis, M. L., & Kerlin, K. (2003). Callous–unemotional traits,impulsivity, and emotional processing in antisocial adolescents. Journal of Clinical Child and AdolescentPsychology, 32, 66–80.

Lykken, D. T. (1995). The antisocial personalities. Hillsdale, NJ: Erlbaum.Lynam, D. R. (1997). Pursuing the psychopath: Capturing the fledgling psychopath in a nomological net.

Journal of Abnormal Psychology, 106, 425–438.Lynam, D. R. (2002). Fledgling psychopathy: A view from personality theory. Law and Human Behavior,

26, 255–259.McConaughy, S., Stanger, C., & Achenbach, T. M. (1992). Three-year course of behavioral/emotional

problems in a national sample of 4- to 16-year olds: I. Agreement among informants. Journal of theAmerican Academy of Child and Adolescent Psychiatry, 31, 932–940.

McCrae, R. R., Costa, P. T., Terracciano, A., Parker, W. D., Mills, C. J., De Fruyt, F., & Mervielde, I.(2002). Personality trait development from age 12 to age 18: Longitudinal, cross-sectional, and cross-cultural analyses. Journal of Personality and Social Psychology, 83, 1456–1468.

McCrae, R. R., Stone, S. V., Fagan, P. J., Costa, P. T. (1998). Indentifying causes and disagreementbetween self-reports and spouse ratings of personality. Journal of Personality, 66, 285–312.

Newman, J. P., Patterson, C. M., & Kosson, D. S. (1987). Response perseveration in psychopaths.Journal of Abnormal Psychology, 96, 145–148.

O’Brien, B. S., & Frick, P. J. (1996). Reward dominance: Associations with anxiety, conduct problems,and psychopathy in children. Journal of Abnormal Child Psychology, 24, 223–240.

O’Brien, B. S., Frick, P. J., & Lyman, R. D. (1994). Reward dominance among children with disruptivebehavior disorders. Journal of Psychopathology and Behavioral Assessment, 16, 131–145.

Ogloff, J., Wong, S., & Greenwood, A. (1990). Treating criminal psychopaths in a therapeuticcommunity program. Behavioral Sciences and the Law, 8, 181–190.

Pardini, D. A., Lochman, J. E., & Frick, P. J. (2003). Callous/unemotional traits and social cognitiveprocesses in adjudicated youth. Journal of the American Academic of Child and Adolescent Psychiatry, 42,364–371.

Patrick, C. J. (1994). Emotion and psychopathy: Startling new insights. Psychophysiology, 31, 319–330.Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency of personality traits

from childhood to old age: A quantitative review of longitudinal studies. Psychological Bulletin, 126,3–25.



Russo, M. F., Stokes, G. S., Lahey, B. B., Christ, M. A. G., McBurnett, K., Loeber, R., Stouthamer-Loeber, M., & Green, S. M. (1993). A sensation-seeking scale for children: Further refinement andpsychometric development. Journal of Psychopathology and Behavioral Assessment, 15, 69–86.

Rutherford, M. J., Cacciola, J. S., Alterman, A. I., McKay, J. R., & Cook, T. G. (1999). Two-year test–retest reliability of the Psychopathy Checklist—Revised in methodone patients. Assessment, 6, 285–291.

Salekin, R. T. (2002). Psychopathy and therapeutic pessimism: Clinical lore or clinical reality? ClinicalPsychology Review, 22, 79–112.

Schroeder, M. L., Schroeder, K. G., & Hare, R. D. (1983). Generalizability of a checklist for theassessment of psychopathy. Journal of Consulting and Clinical Psychology, 51, 511–516.

Seagrave, D., & Grisso, T. (2002). Adolescent development and the measurement of juvenile ofpsychopathy. Law and Human Behavior, 26, 219–239.

Shaffer, D., Fisher, P., Lucas, C. P., Dulcan, M. K., & Schwab-Stone, M. E. (2000). NIMH diagnosticinterview schedule for children version IV (NIMH DISC-IV): Description, differences from previousversions, and reliability of some common diagnoses. Journal of the American Academy of Child andAdolescent Psychiatry, 39, 28–38.

Shelton, K. K., Frick, P. J., & Wootton, J. (1996). The assessment of parenting practices in families ofelementary school-aged children. Journal of Clinical Child Psychology, 25, 317–327.

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

Skeem, J. L., Monahan, J., & Mulvey, E. P. (2002). Psychopathy, treatment involvement, and subsequentviolence among civil psychiatric patients. Law and Human Behavior, 26, 577–603.

Toupin, J., Mercier, H., Dery, M., Cote, G., & Hodgins, S. (1995). Validity of the PCL-R foradolescents. Issues in Criminological and Legal Psychology, 24, 143–145.

Verhulst, F. C., Koot, H. M., & Berden, G. F. (1990). Four-year follow-up of an epidemiological sample.Journal of the American Academy of Child and Adolescent Psychiatry, 29, 440–448.

Walters, G. D. (2003). Predicting criminal justice outcomes with the Psychopathy Checklist and LifestyleCriminality Screening Form: A meta-analytic comparison. Behavioral Sciences and the Law, 21, 89–102.

Wootton, J. M., Frick, P. J., Shelton, K. K., & Silverthorn, P. (1997). Ineffective parenting and childhoodconduct problems: The moderating role of callous–unemotional traits. Journal of Consulting and ClinicalPsychology, 65, 301–308.



Documents

The 4 Year Stability of Psychopathic Traits in Non ...labs.uno.edu/developmental-psychopathology/APSD-Self-Report/The_4... · The 4 Year Stability of Psychopathic Traits in Non-Referred