Criminal Justice and Behavior

[Link]

Executive Functioning and Risky Decision Making in Young Male Offenders

Eva M. Syngelaki, Simon C. Moore, Justin C. Savage, Graeme Fairchild and Stephanie H.M. Van
Goozen
Criminal Justice and Behavior 2009; 36; 1213
DOI: 10.1177/0093854809343095

The online version of this article can be found at:

[Link]

Published by:

[Link]

On behalf of:
International Association for Correctional and Forensic Psychology

Additional services and information for Criminal Justice and Behavior can be found at:

Email Alerts: [Link]

Subscriptions: [Link]

Reprints: [Link]

Permissions: [Link]

Citations [Link]

Downloaded from [Link] at Cardiff University on March 22, 2010

EXECUTIVE FUNCTIONING AND RISKY
DECISION MAKING IN YOUNG MALE
OFFENDERS
EVA M. SYNGELAKI
SIMON C. MOORE
JUSTIN C. SAVAGE
Cardiff University
GRAEME FAIRCHILD
Cambridge University
STEPHANIE H. M. VAN GOOZEN
Cardiff University

Executive functioning (EF) deficits have been implicated in antisocial behavior. EF constitutes a broad selection of functions,
and it is yet to be determined what specific aspects of EF are associated with antisocial behavior. This study examines IQ,
EF, and related decision-making processes in young males involved with the criminal justice system and their age-matched
male controls. The Risky Choice Task (RCT), the Wisconsin Card Sorting Test (WCST), and tests taken from the Cambridge
Neuropsychological Test Automated Battery (CANTAB) assessing spatial working memory, planning, and set-shifting ability
were administered. Young offenders showed lower estimated IQ scores and perseveration of responding, as indicated by
performance on the WCST. There were also specific EF impairments, such as problems in working memory and planning
(CANTAB). Finally, young offenders showed more risky decision-making than controls did, particularly after a small win.
The results on the RCT suggest altered reward mechanisms in young offenders, whereas the findings also support the notion
of EF difficulties related to dorsolateral and ventromedial prefrontal cortex functioning. The implications of these findings
for interventions with young offenders are discussed.

Keywords: executive functioning; decision making; young offenders; reward sensitivity; prefrontal cortex

V iolent or antisocial people often display disinhibited, impulsive, and risk-taking behav-
iors with little concern for the consequences of their actions. Moreover, they seem
unable to learn from their mistakes (this particularly applies to individuals with psycho-
pathic characteristics; see Huss & Langhinrichsen-Rohling, 2000). These patterns of
behavior are similar to those displayed by patients with ventromedial prefrontal cortex
deficits, which suggests that frontal lobe dysfunction may underlie some forms of antiso-
cial behavior (Seguin, 2004). This study examines neuropsychological processes associ-
ated with prefrontal lobe functioning in young offenders involved in antisocial behavior.
The role of neuropsychological function in the etiology of antisocial behavior has been
highlighted by numerous researchers (Ishikawa & Raine, 2003; Raine, 2002). According
to Moffitt’s (1993) developmental taxonomic theory of antisocial behavior, for example,

AUTHORS’ NOTE: The study was funded by project grants from the Economic and Social Research Council
(RES-164-25-0017) and the British Academy (LRG-39883) to Stephanie H. M. van Goozen and Simon C. Moore.
We thank our participants for taking part in the study and the Cardiff Youth Offending Service for helping with
recruitment. Correspondence may be addressed to Stephanie H. M. van Goozen, School of Psychology, Cardiff
University, CF10 3AT, Cardiff, United Kingdom; e-mail: vangoozens@[Link].

CRIMINAL JUSTICE AND BEHAVIOR, Vol. 36 No. 11, November 2009 1213-1227
DOI: 10.1177/0093854809343095
© 2009 International Association for Correctional and Forensic Psychology

1213
Downloaded from [Link] at Cardiff University on March 22, 2010
1214   Criminal Justice and Behavior

neuropsychological functioning deficits characterize life-course-persistent antisocial youths

(Moffitt & Caspi, 2001), and an extensive literature now implicates prefrontal cortex dys-
function, usually resulting in executive dysfunction (Pennington & Ozonoff, 1996), in the
expression of antisocial behavior (J. Blair, Mitchell, & Blair, 2005). A meta-analysis of
studies relating to developmental psychopathology and neuropsychological functioning
(Pennington & Ozonoff, 1996) found that executive dysfunction was associated with
Attention Deficit Hyperactivity Disorder (ADHD; American Psychiatric Association,
2000), with no evidence for executive deficits in Conduct Disorder (CD; American
Psychiatric Association, 2000) alone. Whereas Pennington and Ozonoff’s (1996) study is
informative with respect to the role of EF (executive functioning) in different developmen-
tal psychopathologies, the studies reviewed in relation to CD used some neuropsychological
tests that had not been as well-validated or had less clear links with EF compared to more
recent tasks (Morgan & Lilienfeld, 2000). Moreover, a number of studies in Pennington and
Ozonoff’s study used the Wisconsin Card Sorting Test (WCST; Heaton, 2005), one of the
most commonly used EF tests, with performance on this test relying heavily on the dorsola-
teral prefrontal cortex (Monchi, Petrides, Petre, Worsley, & Dagher, 2001), an area of the
brain not usually associated with antisocial behavior (J. Blair et al., 2005).
Contrary to Pennington and Ozonoff’s (1996) meta-analysis, support for EF impair-
ments in antisocial groups was found in a meta-analysis by Morgan and Lilienfeld (2000),
with the largest effect size found for the Porteus Mazes Q score assessing motoric control
and inhibition and with greater effect sizes in studies using criminal and delinquent groups
than clinically defined antisocial groups (i.e., Antisocial Personality Disorder, CD, psy-
chopathy). Even though effect sizes for the results of the meta-analysis were in the
medium to large range, potential confounds such as ADHD comorbidity were not taken
into account. Furthermore, although a larger range of EF tests was considered, no distinc-
tion was made between the different prefrontal brain regions (e.g., dorsolateral, orbitome-
dial) relevant to antisocial behavior (J. Blair et al., 2005). Apart from a few studies, the
majority of the literature has not investigated the role of specific frontal subregions in
aggressive and antisocial behavior, nor has it distinguished between different forms of
executive function. It is, therefore, important to further investigate the different executive
functioning processes associated with different frontal areas and their relationship to the
occurrence of antisocial behavior.
Although relatively few studies have investigated memory function in antisocial popula-
tions (Raine et al., 2005), some studies have reported memory impairments in these groups.
For example, boys with a history of physical aggression have been found to perform poorly
on tests of working memory compared to nonaggressive boys (Seguin, Boulerice, Harden,
Tremblay, & Pihl, 1999). This finding held after controlling for group differences in IQ and
ADHD symptomatology. Other studies found no evidence for working memory deficits in
antisocial children. For example, van Goozen et al. (2004) found that children with
Oppositional Defiant Disorder (ODD) and children with ODD comorbid with ADHD
(ODD/ADHD) had no problems with working memory, inhibition, planning, and impulsiv-
ity compared to controls and only performed worse on a response perseveration task assess-
ing reversal learning and punishment-and-reward sensitivity. To examine EF in more detail,
we employed tests intended to preferentially probe orbitofrontal and medial prefrontal
cortical processing. These tests were from the Cambridge Neuropsychological Test
Automated Battery (CANTAB; CeNeS Ltd., Cambridge, UK). The CANTAB is a battery

Downloaded from [Link] at Cardiff University on March 22, 2010

 Syngelaki et al. / EXECUTIVE FUNCTIONING AND RISKY DECISION MAKING   1215

of computerised tests administered with the aid of a touch-sensitive screen. It has been
extensively used to test various aspects of executive functioning in children (Lehto,
Juujärvi, Kooistra, & Pulkkinen, 2003) and adults (Robbins et al., 1998), and its usefulness
in different populations, ranging from normal healthy volunteers (Murphy, Smith, Cowen,
Robbins, & Sahakian, 2002) to those with ADHD (Mehta, Goodyer, & Sahakian, 2004) or
substance misuse problems (Townshend & Duka, 2001), makes this battery of tests suitable
for the purposes of this study.
In addition to the role played by the orbitofrontal cortex in higher order cognitive func-
tions such as EF, it also appears to be involved in emotional responding together with the
amygdala (Angrilli, Bianchin, Radaelli, Bertagnoni, & Pertile, 2008; Rudebeck, Bannerman,
& Rushworth, 2008). It has been suggested that a deficit in emotional processing can lead
to impaired decision making as reflected in problem behaviors such as risk taking and
impulsive behavior (Bechara, Damasio, & Damasio, 2000). A deficit in ventromedial
(orbitofrontal and medial frontal cortex) prefrontal cortex functioning has been associated
with decision-making impairments (Bechara, Damasio, Damasio, & Anderson, 1994).
Damasio’s somatic marker hypothesis proposes that somatic processes act as signals to the
emotional circuitry of the brain, particularly the ventromedial prefrontal cortex, thereby
facilitating decision making in cases of uncertainty and difficulty (Damasio, 1996). One
test that has been used to test the somatic marker hypothesis is the Iowa Gambling Task.
However, performance on this task might also relate to neuropsychological processes other
than somatic markers, such as working memory and sensitivity to reward and punishment
(Dunn, Dalgleish, & Lawrence, 2006).
Decision-making processes typically involve evaluating options in terms of prospective
outcomes (e.g., cost–benefit assessment of whether a positive outcome is more likely than a
negative one), which is part of the rational process in carrying out a particular action (Dunn
et al., 2006). The relationship between decision making and frontal lobe functioning (spe-
cifically the orbitofrontal cortex) has been demonstrated through studies of substance misuse
(London, Ernst, Grant, Bonson, & Weinstein, 2000). Using the Iowa Gambling Task, drug
users have been found to opt for short-term gains, which ultimately result in long-term
negative outcomes (Grant, Contoreggi, & London, 2000). Given their propensity to abuse
substances (Disney, Elkins, McGue, & Iacono, 1999; Lewis, Cloninger, & Pais, 1983)
decision-making impairments with altered reward and punishment processing may also be
present in antisocial individuals. To directly assess this, a modified version of the Risky
Choice Task (RCT; Rogers et al., 2003) was employed. The modified RCT was chosen in
preference to the Iowa Gambling Task because impaired performance on the Iowa Gambling
Task might at the same time reflect impairments in other processes, whereas the RCT is a
more direct measure of decision making and sensitivity to reward and punishment.
Participants’ IQ was taken into account when investigating executive functioning.
Antisocial behavior is related to both intellectual and neuropsychological functioning, with
antisocial groups’ scoring lower on intelligence tests than nonantisocial groups by about 8
points on average (Moffitt, 1990). Previous research has suggested that the frontal lobes are
mostly associated with fluid intelligence, which is usually assessed through performance
IQ tests, typically consisting of nonverbal tests of attention to detail, sequential reasoning,
manual design construction, visual puzzle solving, symbolic encoding and decoding, and
maze completion (Moffitt, 1990, p. 134). Duncan, Burgess, and Emslie (1995) found that
patients with frontal lobe lesions had significantly lower scores on a fluid intelligence test

Downloaded from [Link] at Cardiff University on March 22, 2010

1216   Criminal Justice and Behavior

compared to posterior lesion patients and healthy controls. However, other studies found
that delinquents show specific problems on verbal IQ tests (Lynam, Moffitt, & Stouthamer-
Loeber, 1993; Wolff, Waber, Bauermeister, Cohen, & Ferber, 1982). Because delinquent
behavior is linked to lower intelligence, less schooling, and social disadvantage (J. Blair
et al., 2005; Moffitt, 1990), we examined the role of IQ in executive functioning in young-
sters presenting with antisocial behavior.
Investigations into the neuropsychology of antisocial behavior, especially ones focussing
on EF, have so far yielded mixed results. The inconclusive evidence with respect to the role
of executive dysfunction in individuals presenting with antisocial behavior was the primary
motivation to conduct this study. Finding out more on the presence as well as the extent of
these impairments is important for the development of targeted interventions.
This study incorporates measures assessing IQ and global executive functioning as well
as more specific executive functions including working memory, planning, risk taking, and
sensitivity to reward and punishment. There were four hypotheses: First, young offenders
were expected to show lower estimated IQ scores compared to available norms. Second,
with respect to the WCST it was hypothesized that young offenders would not show global
executive problems (consistent with findings by Pennington & Ozonoff, 1996, and Fairchild
et al., 2009). Third, specific impairments in working memory, planning, and decision mak-
ing were expected in young offenders. Finally, young offenders were predicted to show not
only more risk taking behavior but also differential risk taking following receipt of reward
as opposed to punishment on laboratory-based gambling tests.

METHOD

PARTICIPANTS

One hundred and four 12- to 18-year-old young male offenders (mean age = 16.30 years,
SD = 1.51) were recruited from the Youth Offending Service (YOS) in Cardiff. Because
young offenders are an extremely difficult group to test, the number of participants per test
administered differed considerably. With respect to the measurement of intellectual func-
tioning, we used the Raven’s Standard Progressive Matrices (referred to throughout this
article as “the Raven”) for 33 young offenders (mean age = 16.88 years, SD = 1.21) and the
Wechsler Abbreviated Scale of Intelligence (WASI) for 71 young offenders (mean age =
16.04 years, SD = 1.57); we administered the CANTAB to 43 young offenders (mean
age = 15.96 years, SD = 1.57), the RCT to 103 young offenders (mean age = 16.33 years,
SD = 1.50), and the WCST to 103 young offenders (mean age = 16.29 years, SD = 1.51).
Participants (referred to throughout this article as “the YOS group”) had exhibited delin-
quent behavior at different levels of seriousness and were expected to differ in the extent
of their emotional and behavioral difficulties. The mean number of offences (as indexed
from their criminal records) of the YOS group was 9.4 (SD = 10.4), and the most frequent
types of offences were violence and theft.
The YOS group’s (n = 71) estimated IQ (i.e., WASI) scores were compared to population
norms. Young offenders completed three tests from the CANTAB and their data (n = 43)
were compared to existing age-matched norms for these tests, namely the Spatial Working
Memory test (SWM), the Stockings of Cambridge (SOC), and the Intra-Extra Dimensional

Downloaded from [Link] at Cardiff University on March 22, 2010

 Syngelaki et al. / EXECUTIVE FUNCTIONING AND RISKY DECISION MAKING   1217

Set Shift test (ID/ED). Because IQ can have a confounding influence on executive func-
tioning measures, only participants with an estimated average IQ were used in analyses of
the CANTAB tests (n = 43). Five young offenders were identified as having a below aver-
age intelligence based on their WASI (<79) or Raven (Grade 4 or 5) results. Of the remain-
ing 38 YOS participants, t tests indicated no significant difference between their IQ score
(M = 96.53, SD = 10.74) and published norms (M = 100.00, SD = 15.00; t = –1.30, p > .05).
These five participants were, however, included in the analyses of the WCST, and the RCT
as IQ was not correlated with any of these measures.
Data on the WCST (n = 103) were compared to available norm data taken from the
WCST–64 Card Version (WCST-64) Professional Manual (Kongs, Thomson, Iverson, &
Heaton, 2000).
YOS data on the RCT (n = 103) were compared to those of a normal control (NC) group
(Fairchild et al., 2009) consisting of 84 male adolescents, aged 14 to 18 years (mean age =
15.77, SD = 0.82). These controls had been recruited from secondary schools and colleges
in relatively deprived areas in Cambridge.

PROCEDURE

Young offenders were recruited in collaboration with their case workers at the Cardiff
YOS. Information letters and consent forms were given out to young people and their par-
ents or guardians if they were younger than 18 years old. Participants received a full expla-
nation of the procedure and the tasks they were required to undertake and were made aware
of their right to withdraw at any time.
Participants completed the following materials: the Wechsler Abbreviated Scale of
Intelligence (WASI; Wechsler, 1999) or the Raven (Raven, Raven, & Court, 2004), the
WCST (Heaton, 2005), and a modified version (Fairchild et al., 2009) of the RCT (Rogers
et al., 2003). Participants who agreed to take part in a further study also completed three
tasks from the CANTAB. These consisted of the SWM, ID/ED, and SOC. Each participant
received a £5 gift voucher for each hour of participation.

MEASURES

WASI. The WASI (Wechsler, 1999) provides a brief and reliable estimate of a person’s
intellectual functioning. The two-subscale version was used; participants completed the
vocabulary and block design components of the WASI, which tap into crystallized and fluid
abilities. A total t score from the individual t scores of the two subtests is calculated and
converted into an IQ estimate, which corresponds to respondent’s age.

The Raven. The Raven (Raven et al., 2004) was used initially to determine partici-
pants’ intellectual functioning. However, because the Raven does not give an estimated
IQ score nor accounts for verbal ability, it was subsequently replaced with the WASI.
The Raven is a culture-free test, eliminating the use of literacy skills to assess IQ. The
test comprises five 12-item sets, which give a total score between 0 and 60. Participants
have to choose one figure out of six or eight possible options to complete a pattern for
each item. Based on their total score, participants are assigned to a percentile of people
of the same age group. According to this percentile score they are then classified to a

Downloaded from [Link] at Cardiff University on March 22, 2010

1218   Criminal Justice and Behavior

grade. Grades range from 1 to 5; people who are classified in Grade 1 are categorised
as “intellectually superior” and those who are classified in Grade 5 are categorised as
“intellectually impaired.”

WCST. The WCST (Heaton, 2005) was used as a measure of set shifting ability. Participants
completed the computerised WCST-64, which is an abbreviated form of the standard
128-card version (Heaton, 1981). The WCST-64 maintains the task requirements of the
WCST. The following variables were used, consistent with previous research (Ernst et al.,
2003): total number of errors made (ranging from 0 to 64), perseverative errors (errors
made when continuing to sort cards according to the rule that was correct for the prior stage
of the task; ranging from 0 to 62), nonperseverative errors (ranging from 0 to 64), and the
number of categories completed (ranging from 0 to 6).

Decision-making task. A modified version of the RCT was used (Rogers et al., 2003).
This computer-based task assesses risk-taking behavior. The aim of this version of the RCT
is to win as many points as possible. Participants are told that they will see two wheels of
fortune on the computer screen, one on the left and one on the right, and they have to
choose the wheel that will give them the best chance of winning as many points as possible.
Each wheel consists of eight segments that have differing amounts they can win or lose
each time. The participants have to choose between control and experimental wheels. The
control wheels have a 50-50% chance of either winning or losing 10 points (thus an
expected value of 0; see below). The experimental wheel varies systematically in terms of
the probability of winning or losing points (75% or 25%), the magnitude of a gain (20 or
80 points), and the magnitude of a loss (20 or 80 points) to yield eight possible trial types.
These types vary in their relative expected value (delta expected value [∆EV], which is the
difference between the two wheels with which the participants are presented). For example,
in most trials the control wheel has an EV of 0 (.5 × 10 + .5 × –10), whereas an experimen-
tal wheel with 25% probability of losing 80 points and 75% probability of winning 20
points has an EV of –5 (.25 × –80 + .75 × 20). In this example, the ΔEV of choosing the
experimental gamble is –5.
Two additional trial types are included to measure framing effects. The negative framing
trial (0 + frame) involves a wheel with a guaranteed gain of 40 points and a wheel with a
50% probability of winning 80 points and a 50% probability of gaining 0 points. In com-
parison, the positive framing trial (0 – frame) includes a wheel with a guaranteed loss of
40 points and a wheel with a 50% probability of losing 80 points and a 50% probability of
losing 0 points. These two trial types are used to measure risk aversion and risk seeking.
The task consists of four blocks with 20 trials in each block, giving 80 trials in total. All 10
trial types are included twice in each block in a pseudo-randomised order, and the experi-
mental wheel can appear on either the left or right side of the display. The dependent meas-
ure is the number of times the experimental wheel is chosen over the control wheel,
expressed as a percentage. Further analysis examined participants’ performance data as
well as the choice of the experimental gamble by trial type. Finally, participants’ behavior
after various levels of wins or losses resulting from choosing the experimental, risky wheel
was examined. These were the percentage of times the experimental wheel was chosen
after a big loss, a small loss, a big win, and a small win. These four measures served as
indicators of sensitivity to reward and punishment. The following computerised tests from

Downloaded from [Link] at Cardiff University on March 22, 2010

 Syngelaki et al. / EXECUTIVE FUNCTIONING AND RISKY DECISION MAKING   1219

the CANTAB were used. All tasks were presented on a monitor with a touch-sensitive
screen, and responses were recorded via the touch-sensitive screen.

SWM. This is a test assessing the participant’s ability to retain spatial information in work-
ing memory while incorporating a strategic search element to index central executive func-
tion (Owen, Downes, Sahakian, Polkey, & Robbins, 1990). Participants are presented with a
number of coloured boxes and asked to find equal numbers of tokens hidden under the boxes.
The number of boxes gradually increases from four presentations of three boxes to four pres-
entations of each of four, six, and eight boxes. Participants are told that once a token is found,
the computer will never hide a token under the same box, allowing for different types of
errors recorded for this task. A “between error” occurs upon returning to a box that has
already been found to contain a token. An estimate of a successful strategy (Owen et al.,
1990, p. 1025) is calculated by adding the number of times participants start a new search
by opening a different box. A high score corresponds to a poor use of the strategy and a low
score reflects effective use. The variables between errors and strategy score from the SWM
task were compared to available norms.

ID/ED. The ID/ED is a test of attentional set shifting and reversal learning. Participants
are presented with two types of stimuli, colour-filled shapes and white lines, with the
combination of both types of stimuli comprising compound stimuli. Participants are asked
to choose the pattern they think is correct by learning a rule and following it. Participants’
ability to learn to attend to compound stimuli and to shift their attention from one type of
stimuli to another is examined. Participants progress by learning a set criterion (six con-
secutive correct responses) and the test comes to an end if a criterion is failed after 50
trials. The task consists of nine blocks and has two key stages, the intra-dimensional shift
(IED; block 6), and the extra-dimensional shift (EDS; block 8). The IED stage requires
participants to carry on attending to the previously relevant dimension, and the EDS stage
requires participants to switch their attention to the previously irrelevant dimension and
to learn which stimuli is correct in this dimension. These two key stages correspond to the
rules of the WCST. However, the ID/ED is intended to disaggregate confounded stages of
the WCST (Seguin, Sylver, & Lilienfeld, 2007) and dissociate response reversal from
attentional set-shifting (Mitchell, Colledge, Leonard, & Blair, 2002). Attentional set-
shifting corresponds to the EDS component of the task and is a component also measured
by the WCST. For our study, the variables used were the number of errors made in each
of the two key stages (IED and EDS), total number of errors made throughout the task,
and number of stages passed. Performance of the YOS group on the ID/ED was compared
to that of the norm.

SOC. The SOC is based on the Tower of London task (Shallice, 1982) and is used to
measure spatial planning ability. Two sets of coloured balls are presented, positioned in
vertically placed stockings hanging in three pockets. The positions of balls in both sets vary
for each trial, and the problems increase in difficulty with subjects’ having to complete two-,
three-, four-, and five-move problems. The aim of the task is to replicate the top arrangement
of balls in the bottom display of balls by moving one ball at a time. The number of trials in
which participants manage to complete the minimum problem solutions is measured. These
data were compared to available norms (CeNeS Ltd., Cambridge, UK).

Downloaded from [Link] at Cardiff University on March 22, 2010

1220   Criminal Justice and Behavior

DATA ANALYSES

Data on the Raven are only reported descriptively. Young offenders’ performances on the
WASI, the WCST, and CANTAB were compared to existing norms by using two-sample
t tests in Stata v. 10, entering the number of young offenders, their means and standard
deviations, together with these values for the norm groups. The young offenders’ data were
found to be normally distributed. With respect to the RCT task, control data were available
from Cambridge and mixed design ANOVAs were conducted with group (control vs.
offender) as the between-subjects factor with three various within-subject factors (tested
separately). These were as follows: round (1 to 4), trial type (1 to 10), and outcome (large
loss, small loss, large gain, small gain). Simple main effects were used for post hoc com-
parisons between the two groups. Where the assumption of sphericity was violated
Greenhouse-Geisser corrections were applied. Effect sizes are reported as Cohen’s d (small,
d = .2; medium, d = .5; large, d = .8; Cohen, 1988) and partial eta squared (ηp2; small ≥ .01,
medium ≥ .06, large ≥ .14; Cohen, 1988). These analyses were carried out using SPSS 14.0
(SPSS Inc., Chicago, IL).

RESULTS

YOS DATA COMPARISONS TO NORMS

The results of two-sample t tests comparing YOS means for estimated IQ, WCST, and
CANTAB variables to available norms are shown in Table 1. WCST data for one partici-
pant were missing and the IED data for one YOS participant were removed because he was
an outlier. Data on the Raven (n = 33) are only mentioned descriptively as these results
could not be included in any of the subsequent analyses because of their nominal nature. It
was found that 17 young offenders scored in the below average range, 15 in the average
range, and 1 in the above average range.
The YOS group differed significantly from norms in terms of estimated IQ and the
number of perseverative errors on the WCST. Importantly, the number of perseverative
errors was not significantly correlated with IQ (r = –.197, n = 70, p > .05, two-tailed), which
suggests that IQ did not influence these results. With respect to the CANTAB tests, the YOS
group performed significantly worse than the reported norms on nearly all of the variables.

BETWEEN-GROUP COMPARISONS: RCT DATA

Data were available for 103 YOS and 84 control participants. Data were analyzed by
running a mixed-model ANOVA with group as between-subjects factor and round (1 to 4)
as within-subjects factor. There was no main effect of round (p = .33); main effect of group,
F(1, 185) = 5.32, p < .05, ηp2 = 0.028, with controls scoring significantly more points (mean
number of points = 186.25, SE = 9.73) than YOS (mean number of points = 156.02, SE =
8.79); and no Round × Group interaction (p = .52).

CHOICE OF THE EXPERIMENTAL GAMBLE BY TRIAL TYPE

A mixed-model ANOVA with trial type (10 types differing in ΔEV; the ΔEV means that
the difference between the two wheels results in unfavourable or favourable contingencies

Downloaded from [Link] at Cardiff University on March 22, 2010

 Syngelaki et al. / EXECUTIVE FUNCTIONING AND RISKY DECISION MAKING   1221

TABLE 1: IQ, WCST, and CANTAB Scores for YOS and Norms

YOS Norms

Measure M SD M SD t p

Estimated IQ (n = 71) 92.52 12.1 100 15 –3.47 <0.001

WCST total errors (n = 103) 22.15 9.01 21.07 8.73 1.13 0.26
WCST perseverative errors (n = 103) 9.69 4.27 8.55 4.14 2.51 <0.05
WCST nonperseverative errors (n = 103) 12.46 6.97 12.52 6.80 –0.08 0.94
WCST categories completed (n = 103) 2.82 1.08 2.79 1.35 0.21 0.83
SWM strategy score (n = 38) 34.24 4.24 29.12 5.19 5.14 <0.001
SWM between errors (n = 38) 29.21 15.63 11.95 11.27 6.46 <0.001
SOC minimum move solutions (n = 38) 8.32 2.05 9.52 1.97 –2.92 <0.01
IED pre-ES errors (n = 37) 7.70 2.95 7.34 5.97 0.34 0.73
IED EDS errors (n = 37) 9.05 9.42 5.81 7.22 1.98 <0.05
IED total errors (n = 37) 20.22 10.64 14.61 8.72 2.87 <0.01
IED number of stages passed (n = 37) 8.54 0.84 8.79 0.74 –1.59 0.11

Note. IQ = Intelligence Quotient; WCST = Wisconsin Card Sorting Test; CANTAB = Cambridge Neuropsychological
Test Automated Battery; YOS = young offender group; SWM = Spatial Working Memory; SOC = Stockings of
Cambridge; IED = Intra-Extra-Dimensional Set Shift; pre-ES = pre-Extra-Dimensional Shift; EDS = Extra-
Dimensional Shift.

depending on the selection of the wheel) as within-subjects factor and group as between-
subjects factor was run. There was a main effect of trial type, F(4.55, 840.89) = 737.63, p <
.001, ηp2 = 0.80, a main effect of group, F(1, 185) = 35.22, p < .001, ηp2 = 0.16, and a Trial
Type × Group interaction, F(4.55, 840.89) = 5.85, p < .001, ηp2 = 0.031. The Trial × Group
interaction was further investigated by looking at how the two groups differed for each trial
type. The analyses showed that the YOS group gambled significantly more than controls
on all trial types, apart from those with a ΔEV of +10 and +55.
Figure 1 shows that decision making was influenced by the experimental contingencies,
with risk taking increasing as the probability of a win increased. In contrast, both groups
were less likely to gamble in trials with more unfavourable contingencies. YOS participants
generally took more risks, with the differences being most clear in the middle of the deci-
sion making curve.

GROUP DIFFERENCES IN DECISION MAKING FOLLOWING REWARD OR PUNISHMENT

One participant from the control group and one participant from the YOS group were
excluded from the analyses due to insufficient data (these participants did not experience
all types of trials outcome—one never experienced a large loss and one never experienced
a small loss). First, a one-way ANOVA was run to examine whether the YOS (n = 102) and
NC (n = 83) groups differed in terms of the overall proportion gambled across trials. The
YOS group gambled significantly more than the NC group did, F(1, 184) = 33.53, p < .001,
d = 0.88, with a mean overall proportion bet of 0.60 (SE = 0.01) for the YOS group and a
mean overall proportion bet of 0.52 (SE = 0.01) for the NC group. This finding means that
YOS participants chose the experimental gamble 60% of the time, whereas the controls
chose the experimental gamble 52% of the time.
Pearson correlations reveal that young offenders’ estimated IQ and controls’ estimated IQ
were not significantly related to risk-taking behavior (r = 0.025, n = 69, p = .84, two-tailed,
r = –.019, n = 84, p = .86, two-tailed, respectively); thus, IQ was not used as a covariate in

Downloaded from [Link] at Cardiff University on March 22, 2010

1222   Criminal Justice and Behavior

Experimental Gamble Selected

7

–55 –10 –40 5 0 –5 0 40 10 55

+frame – frame
ΔExpected Value of Difference Between Gambles

Controls (n = 84) YOS (n = 103)

Figure 1: Mean (±SEM) Number of Times (of 8 Possible Trials) the Experimental Gamble Was Chosen
Over the Control Gamble in the Risky Choice Task for Each Trial Type by Group
Note. Values on the x axis (trial types) are ordered according to the frequency that the control group chose the
riskier experimental gamble.

further analyses. Two young offenders did not have data on the RCT in these analyses,
hence the lower number of observations in the correlation.
Next, the RCT data were analysed to examine the effect of different outcomes on subse-
quent risk taking behavior. A mixed-model ANOVA was used with group (control vs.
young offenders) as the between-subjects factor (YOS n = 102; NC n = 83) and outcome
(big loss, big win, small loss, and small win) as within-subjects factor. There was no main
effect of outcome (p = .14) but there was a main effect of group, F(1, 183) = 19.25, p <
.001, ηp2 = 0.095, and a significant Group × Outcome interaction, F(3, 549) = 6.62, p <
.001, ηp2 = 0.035. Simple main effects revealed that the YOS group did not gamble more
than controls did after a large win (p = .83) nor after a large loss, F(1, 183) = 3.60, p = .059,
ηp2 = 0.019, but they did gamble significantly more than controls did after a small loss, F(1,
183) = 15.96, p < .001, ηp2 = 0.080, and after a small win, F(1, 183) = 39.77, p < .001, ηp2 =
0.179 (see Figure 2). The large effect size after a small win indicates that the significant
interaction was particularly driven by the young offenders’ stronger tendency to choose the
experimental gamble after the receipt of a small reward in the previous trial.

DISCUSSION

The purpose of this study was to explore neuropsychological functioning in an adoles-

cent group of young offenders because evidence of neuropsychological involvement in
antisocial behavior has been inconclusive, with mixed findings with respect to the nature

Downloaded from [Link] at Cardiff University on March 22, 2010

 Syngelaki et al. / EXECUTIVE FUNCTIONING AND RISKY DECISION MAKING   1223

0.70

0.65

0.60
Proportion of Time

0.55

0.50

0.45

0.40
BigLoss SmallLoss BigWin SmallWin
Outcome

Controls (n = 83) YOS (n = 102)

Figure 2: Proportion of Time Experimental Gamble Was Chosen Following a Specific Event (Big Loss,
Small Loss, Big Win, and Small Win) by Group

and extent of executive functioning deficits. The aim of this study was to find out whether
neuropsychological deficits (globally or more specifically) characterize youngsters who
engage in antisocial behavior. Consistent with our hypothesis and previous literature
(Moffitt, 1990; Lynam et al., 1993) the young offender group had a lower estimated IQ
score compared with existing norm data. After excluding participants with a below-average
estimated IQ score, a subsample of young offenders used for examining the CANTAB
results did not differ from their available norms in IQ.
The second hypothesis, involving the WCST and positing no global EF deficit in young
offenders, was partially supported. Young offenders differed on one variable of the WCST,
the number of perseverative errors, implying difficulty in reversal learning and changing
behavior that was once rewarded. This finding is in line with other research showing prob-
lems with perseveration under changing conditions of reward and punishment in antisocial
groups (J. Blair et al., 2005; van Goozen et al., 2004). The WCST has been found to acti-
vate different neural circuits during the different stages of performance (Monchi et al.,
2001), with the mid-dorsolateral prefrontal region’s being involved in working memory
and the mid-ventrolateral prefrontal region’s being active during attentional set shifting.
The present finding of impaired performance in shifting behavior, together with the finding
of impaired performance on the IE/ED test, is consistent with literature suggesting antiso-
cial groups do not show global prefrontal deficits (Fairchild et al., 2009; Pennington &
Ozonoff, 1996; Raine et al., 2005) but rather specific problems in reversal learning or
detecting altered contingencies (J. Blair et al., 2005; R. J. R. Blair, Colledge, & Mitchell,
2001; van Goozen et al., 2004).

Downloaded from [Link] at Cardiff University on March 22, 2010

1224   Criminal Justice and Behavior

As predicted, the YOS group exhibited a range of specific executive deficits as shown
by their performance on the CANTAB tests and the RCT. In particular, young offenders
committed more errors on the spatial working memory task and were less able to use a
systematic search strategy, which is associated with better performance on this task. Their
planning ability was also worse and they had trouble shifting their behavior and made more
errors in the IE/ED task, specifically the extra-dimensional shift of the test. Impaired per-
formance by the YOS group on these tests suggests that the dorsolateral prefrontal cortex
is compromised in this group. The ventrolateral prefrontal cortex might also be affected
because it is involved in set shifting (Monchi et al., 2001).
The YOS group generally demonstrated a greater propensity to gamble (choosing the
more risky option), but they especially did so after the receipt of a small win, as shown by
Figure 2 and the effect sizes of the differences, whereas the opposite was true for control
participants: Nonoffending male adolescents gambled least after the receipt of small
wins. IQ did not seem to influence these results as it was not correlated with performance
on the RCT. This result, of differential modulation of subsequent behavior following
receipt of small rewards, is consistent with findings in conduct disorder adolescents
(Fairchild et al., 2009). Furthermore, young offenders showed elevated risk-taking behav-
ior as the probability of win increased, as shown in Figure 1. A similar pattern of behavior
was observed in the control group. If the differences in gambling behavior between antiso-
cial and normal control groups are caused by the fact that antisocial individuals are
less satisfied with small rewards and risk-taking behavior is modulated by sensitivity to
reward rather than punishment, then this aspect of behavior requires manipulation when
considering interventions.
The findings on the CANTAB and the RCT are consistent with our hypotheses and pre-
vious research suggesting working memory impairments (Seguin et al., 1999), problems in
reversal learning (J. Blair et al., 2005), and the modulation of behavior following changes
in reward and punishment contingencies (Fairchild et al., 2009; van Goozen et al., 2004),
which are associated with dorsolateral prefrontal cortex functioning in the case of working
memory performance and with orbitofrontal and/or ventromedial cortex functioning in the
case of reversal learning and response perseveration (Clark et al., 2008; Clark, Cools, &
Robbins, 2004; Seguin, 2004). Future studies involving brain imaging techniques such as
fMRI are needed to confirm the involvement of these brain areas in relation to the execu-
tive deficits observed and would thus be an appropriate extension to this research.
This study has some strengths in comparison to previous studies. It provides information
on executive functioning deficits in young offenders recruited from the community, which
enables the presentation of some novel findings in a group of youngsters with varying
degrees and types of antisocial behavior. Also, previous studies on neuropsychological
functioning in antisocial groups often used a limited range of executive functioning tasks
(Giancola, & Zeichner, 1994; Rosse, Miller, & Deutsch, 1993) or used tasks not necessarily
tapping into the prefrontal regions that have been implicated in antisocial behavior (i.e.,
orbitomedial rather than dorsolateral prefrontal cortex). The present study addressed these
issues by employing a wider range of executive functioning tests. The findings suggest that
young offenders present with specific problems in working memory, planning and impaired
decision making, and risk-taking behavior.
The observed impairments are similar to the ones reported in CD and physically aggres-
sive youngsters (Fairchild et al., 2009; Seguin et al., 1999) and highlight the importance of

Downloaded from [Link] at Cardiff University on March 22, 2010

 Syngelaki et al. / EXECUTIVE FUNCTIONING AND RISKY DECISION MAKING   1225

the prefrontal cortex in these behavioral problems. Patients with ventromedial prefrontal
cortex lesions have been found to show increased betting on the Cambridge Gambling Task
regardless of the odds of winning (Clark et al., 2008), a finding similar to the behavior
shown by our young offenders on the RCT (except for risk taking after a large win). Our
results show that young offenders suppressed risk-taking behavior after a large win, simi-
larly to controls, but showed increased risk-taking behavior after receipt of a small win. This
finding replicates risk taking in early-onset conduct disordered youngsters (Fairchild et al.,
2009) but is a novel finding in young offenders. Outcomes like this, of differential sensitiv-
ity to reward and punishment, could have important implications for interventions in antiso-
cial groups, especially in light of the current reliance on the use of punishment in the
criminal justice system. Furthermore, if specific EF impairments are identified and found to
vary between antisocial youngsters, then appropriate interventions that take the individual’s
level of neuropsychological impairment into account should be taken into consideration.

REFERENCES

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). New York:
Author.
Angrilli, A., Bianchin, M., Radaelli, S., Bertagnoni, G., & Pertile, M. (2008). Reduced startle reflex and aversive noise per-
ception in patients with orbitofrontal cortex lesions. Neuropsychologia, 46, 1179-1184.
Bechara, A., Damasio, A. R., Damasio, H., & Anderson, S.W. (1994). Insensitivity to future consequences following damage
to human prefrontal cortex. Cognition, 50, 7-15.
Bechara, A., Damasio, H., & Damasio, A. R. (2000). Emotion, decision making and the orbitofrontal cortex. Cerebral Cortex,
10, 295-307.
Blair, J., Mitchell, D., & Blair, K. (2005). The psychopath: Emotion and the brain. Malden, MA: Blackwell.
Blair, R. J. R., Colledge, E., & Mitchell, D. G. V. (2001). Somatic markers and response reversal: Is there orbitofrontal cortex
dysfunction in boys with psychopathic tendencies? Journal of Abnormal Child Psychology, 29, 499-511.
Clark, L., Bechara, A., Damasio, H., Aitken, M. R. F., Sahakian, B. J., & Robbins, T. W. (2008). Differential effects of insu-
lar and ventromedial prefrontal cortex lesions on risky decision-making. Brain, 131, 1311-1322.
Clark, L., Cools, R., & Robbins, T. W. (2004). The neuropsychology of ventral prefrontal cortex: Decision-making and
reversal learning. Brain and Cognition, 55, 41-53.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Earlbaum
Associates.
Damasio, A. R. (1996). The somatic marker hypothesis and the possible functions of the prefrontal cortex. Philosophical
Transactions of the Royal Society of London B, 351, 1413-1420.
Disney, E. R., Elkins, I. J., McGue, M., & Iacono, W. G. (1999). Effects of ADHD, conduct disorder, and gender on substance
use and abuse in adolescence. American Journal of Psychiatry, 156, 1515-1521.
Duncan, J., Burgess, P., & Emslie, H. (1995). Fluid intelligence after frontal lobe lesions. Neuropsychologia, 33, 261-268.
Dunn, B. D., Dalgleish, T., & Lawrence, A. D. (2006). The somatic marker hypothesis: A critical evaluation. Neuroscience
and Biobehavioral Reviews, 30, 239-271.
Ernst, M., Grant, S. J., London, E. D., Contoreggi, C. S., Kimes, A. S., & Spurgeon, L. (2003). Decision making in adoles-
cents with behavior disorders and adults with substance abuse. American Journal of Psychiatry, 160, 33-40.
Fairchild, G., van Goozen, S., Stollery, S., Aitken, M., Savage, J., Moore, S., & Goodyer, I. (2009). Decision making and
executive function in male adolescents with early-onset or adolescence-onset Conduct Disorder and control participants.
Biological Psychiatry, 66, 162-168.
Giancola, P. R., & Zeichner, A. (1994). Neuropsychological performance on tests of frontal- lobe functioning and aggressive
behavior in men. Journal of Abnormal Psychology, 103, 832-835.
Grant, S., Contoreggi, C., & London, E.D. (2000). Drug abusers show impaired performance in a laboratory test of decision
making. Neuropsychologia, 38, 1180-1187.
Heaton, R. K. (2005). Wisconsin Card Sorting Test manual. Odessa, FL: Psychological Assesment Resources.
Huss, M. T., & Langhinrichsen-Rohling, J. (2000). Identification of the psychopathic batterer: The clinical, legal, and policy
implications. Aggression and Violent Behavior, 5, 403-422.
Ishikawa, S. S., & Raine, A. (2003). Prefrontal deficits and antisocial behavior: A causal model. In B. B. Lahey, T. E. Moffitt,
& A. Caspi (Eds), Causes of conduct disorder and juvenile delinquency (pp. 277-304). New York: Guilford.

Downloaded from [Link] at Cardiff University on March 22, 2010

1226   Criminal Justice and Behavior

Kongs, S. K., Thomson, L. L., Iverson, G. L., & Heaton, R. K. (2000). WCST-64: Wisconsin Card Sorting Test–64 card ver-
sion professional manual. Odessa, FL: Psychological Assessment Resources.
Lehto, J. E., Juujärvi, P., Kooistra, L., & Pulkkinen, L. (2003). Dimensions of executive functioning: Evidence from children.
British Journal of Developmental Psychology, 21, 59-80.
Lewis, C. E., Cloninger, C. R., & Pais, J. (1983). Alcoholism, antisocial personality and drug use in a criminal population.
Alcohol and Alcoholism, 18, 53-60.
London, E. D., Ernst, M., Grant, S., Bonson, K., & Weinstein, A. (2000). Orbitofrontal cortex and human drug abuse:
Functional imaging. Cerebral Cortex, 10, 334-342.
Lynam, D., Moffitt, T., & 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, 187-196.
Mehta, M. A., Goodyer, I. M., & Sahakian, B. J. (2004). Methylphenidate improves working memory and set-shifting in
ADHD: Relationships to baseline memory capacity. Journal of Child Psychology and Psychiatry, 45, 293-305.
Mitchell, D. G. V., Colledge, E., Leonard, A., & Blair, R. J. R. (2002). Risky decisions and response reversal: Is there evidence
of orbitofrontal cortex dysfunction in dysfunction in psychopathic individuals? Neuropsychologia, 40, 2013-2022.
Moffitt, T. E. (1990). The neuropsychology of juvenile delinquency: A critical review. Crime and Justice, 12, 99-169.
Moffitt, T. E. (1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy.
Psychological Review, 100, 674-701.
Moffitt, T. E., & Caspi, A. (2001). Childhood predictors differentiate life-course persistent and adolescence-limited antisocial
pathways among males and females. Development and Psychopathology, 13, 355-375.
Monchi, O., Petrides, M., Petre, V., Worsley, K., & Dagher, A. (2001). Wisconsin Card Sorting revisited: Distinct neural
circuits participating in different stages of the task identified by event-related functional magnetic resonance imaging.
Journal of Neuroscience, 21, 7733-7741.
Morgan, A. B., & Lilienfeld, S. O. (2000). A meta-analytic review of the relation between antisocial behavior and neuropsy-
chological measures of executive function. Clinical Psychology Review, 20, 113-136.
Murphy, F. C., Smith, K. A., Cowen, P. J., Robbins, T. W., & Sahakian, B. J. (2002). The effects of tryptophan depletion on
cognitive and affective processing in healthy volunteers. Psychopharmacology, 163, 42-53.
Owen, A. M., Downes, J. J., Sahakian, B. J., Polkey, C. E., & Robbins, T. W. (1990). Planning and spatial working memory
following frontal lobe lesions in man. Neuropsychologia, 28, 1021-1034.
Pennington, B. F., & Ozonoff, S. (1996). Executive functions and developmental psychopathology. Journal of Child
Psychology and Psychiatry, 37, 51-87.
Rahman, S., Sahakian, B. J., Cardinal, R. N., Rogers, R. D., & Robbins, T. W. (2001). Decision making and neuropsychiatry.
Trends in Cognitive Sciences, 5, 271-277.
Raine, A. (2002). Annotation: The role of prefrontal deficits, low autonomic arousal, and early health factors in the develop-
ment of antisocial and aggressive behavior. Journal of Child Psychology and Psychiatry, 43, 417-434.
Raine, A., Moffiitt, T. E., Caspi, A., Loeber, R., Stouthamer-Loeber, M., & Lynam, D. (2005). Neurocognitive impairments
in boys on the life-course persistent antisocial path. Journal of Abnormal Psychology, 114, 38-49.
Raven, J., Raven, J. C., & Court, J. H. (2004). The standard progressive matrices (SPM) (updated ed.). Oxford, UK: Harcourt
Assessment, OPP.
Robbins, T. W., James, M., Owen, A. M., Sahakian, B. J., Lawrence, A. D., McInnes, L., et al. (1998). A study of performance
on tests from the CANTAB battery sensitive to frontal lobe dysfunction in a large sample of normal volunteers:
Implications for theories of executive functioning and cognitive aging. Journal of the International Neuropsychological
Society, 4, 474-490.
Rogers, R. D., Tunbridge, E. M., Bhagwagar, Z., Drevets, W. C., Sahakian, B. J., & Carter, C. S. (2003). Trytophan depletion
alters the decision-making of healthy volunteers through altered processing of reward cues. Neuropsychopharmacology,
28, 153-162.
Rosse, R. B., Miller, M. W., & Deutsch, S. I. (1993). Violent antisocial behavior and Wisconsin card sorting test performance
in cocaine addicts. American Journal of Psychology, 150, 170.
Rudebeck, P. H., Bannerman, D. M., & Rushworth, M. F. S. (2008). The contribution of distinct subregions of the ventrome-
dial frontal cortex to emotion, social behavior, and decision making. Cognitive, Affective and Behavioral Neuroscience,
8, 485-497.
Seguin, J. R. (2004). Neurocognitive elements of antisocial behavior: Relevance of an orbitofrontal cortex account. Brain and
Cognition, 55, 185-197.
Seguin, J. R., Boulerice, B., Harden, P. W., Tremblay, R. E., & Pihl, R. O. (1999). Executive functions and physical aggres-
sion after controlling for attention deficit hyperactivity disorder, general memory, and IQ. Journal of Child Psychology
and Psychiatry, 40, 1197-1208.
Seguin, J. R., Sylver, P., & Lilienfeld, S. O. (2007). The neuropsychology of violence. In I. D. Waldman, D. J. Flannery, &
A. T. Vazsonyi (Ed.), The Cambridge handbook of violent behavior (pp. 187-214). New York: Cambridge University
Press.

Downloaded from [Link] at Cardiff University on March 22, 2010

 Syngelaki et al. / EXECUTIVE FUNCTIONING AND RISKY DECISION MAKING   1227

Shallice, T. (1982). Specific impairments of planning. Philosophical Transactions of the Royal Society of London B, 298,
199-209.
Townshend, J. M., & Duka, T. (2001). Attentional bias associated with alcohol cues: Differences between heavy and occa-
sional social drinkers. Psychopharmacology, 157, 67-74.
van Goozen, S. H. M., Cohen-Kettenis, P. T., Snoek, H., Matthys, W., Swaab-Barneveld, H., & van Engeland, H. (2004).
Executive functioning in children: A comparison of hospitalised ODD and ODD/ADHD children and normal controls.
Journal of Child Psychology and Psychiatry, 45, 284-292.
Wechsler, D. (1999). Wechsler Abbreviated Scale of Intelligence (WASI). New York: Psychological Corporation.
Wolff, P. H., Waber, D., Bauermeister, M., Cohen, C., & Ferber, R. (1982). The neuropsychological status of adolescent
delinquent boys. Journal of Child Psychology and Psychiatry, 23, 267-179.

Eva M. Syngelaki completed her PhD on the psychological causes of antisocial behaviour in young offenders at Cardiff
University’s School of Psychology. She is currently employed by the Cardiff Youth Offending Team.

Simon C. Moore is a senior lecturer in Cardiff University and researches the causes and correlates of violence, alcohol-related
harm, and human decision making, in particular how the propensity for risky decisions is mediated by environmental factors.

Justin C. Savage is a research associate at the Psychology Department of Cardiff University. He received a PhD in psychol-
ogy from Cardiff University and is currently investigating the impact of parental antisocial behaviour on infant well-being.

Graeme Fairchild is a research associate in the Developmental Psychiatry Section at Cambridge University. His research
interests are in the field of developmental psychopathology, with a particular focus on neuropsychological and neurobiologi-
cal influences on the development of externalizing disorders.

Stephanie H. M. van Goozen is a professor in the School of Psychology at Cardiff University (UK) and Leiden University
(NL). Her research is in the area of developmental psychopathology, particularly the neuropsychological and neurobiological
factors that influence young children’s and adolescent’s antisocial behavioural development.

Downloaded from [Link] at Cardiff University on March 22, 2010