Intelligence xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Intelligence
journal homepage: www.elsevier.com/locate/intell

The role of verbal intelligence in becoming a successful criminal: Results

from a longitudinal sample☆
Cashen M. Boccioa,⁎, Kevin M. Beavera,b, Joseph A. Schwartzc
a
College of Criminology and Criminal Justice, Florida State University, Tallahassee, FL 32306-1127, USA
b
Center for Social and Humanities Research, King Abdulaziz University, Jeddah, Saudi Arabia
c
School of Criminology and Criminal Justice, University of Nebraska at Omaha, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Intelligence has been linked with success across a wide array of life domains. To date, however, relatively little
Arrest research has examined whether intelligence may predict criminal success—that is, engaging in criminal beha-
Criminal behavior viors, but escaping detection and arrest. The current study addresses this gap in the literature by examining the
Intelligence associations among verbal intelligence, criminal involvement, and criminal justice processing (i.e., arrest) using
data from the National Longitudinal Study of Adolescent to Adult Health (Add Health). Our findings reveal that
verbal intelligence is associated with criminal justice processing, wherein individuals with higher verbal in-
telligence scores are more likely to avoid arrest for criminal behavior when compared with individuals with
comparatively lower verbal intelligence scores. We discuss the implications of these findings for future research.

1. Introduction intelligence appears to be among the strongest predictors of criminal

Intelligence has emerged as one of the most consistent, robust, and
strongest predictors of virtually every type of antisocial outcome.
Research has revealed, for instance, an inverse relationship between IQ
and self-reported criminal and delinquent behavior, such as drug use
and physical violence (Fergusson, Horwood, & Ridder, 2005; Herrnstein
& Murray, 1994; Latvala et al., 2009; Lynam, Moffitt, & Stouthamer-
Loeber, 1993; McNulty, Bellair, & Watts, 2013). Not only is IQ related
to self-reported behavior, it is also associated with official records of
criminal justice contact, including being incarcerated, being sentenced
to probation, and recidivating (Beaver et al., 2013; Ganzer & Sarason,
1973; Herrnstein & Murray, 1994; Moffitt, Gabrielli, Mednick, &
Schulsinger, 1981; Richter, Scheurer, Barnett, & Kröber, 1996;
Schwartz et al., 2015).1 IQ has also been found to be related to mea-
sures of criminal conduct across different levels of aggregation, in-
cluding neighborhood- and prison-level units of analysis (Bartels, Ryan,
Urban, & Glass, 2010; Beaver & Wright, 2011; Diamond, Morris, &
Barnes, 2012; McDaniel, 2006). Based on all of the available literature,
behavior.
Given the amount of research linking IQ to antisocial phenotypes, it
is interesting that there has not been more research devoted to ex-
amining whether IQ contributes to being a successful criminal—that is,
someone who engages in a crime, but successfully escapes detection
and apprehension. Part of the reason for the lack of IQ research on this
topic has to do with the fact that there has not been much research on
successful criminals (McCall, 1978; Watters & Biernacki, 1989; Wright,
Decker, Redfern, & Smith, 1993). This is somewhat surprising given
that elusive criminals may be among the most dangerous by posing a
serious threat to society (Lussier, Bouchard, & Beauregard, 2011).
Given the lack of research on successful criminals, there is not an ex-
isting knowledge base to draw from in regard to the role that IQ might
play in creating successful criminals. Even so, based on findings from
multiple fields of study, there are at least three reasons to suspect that
IQ would predict being a successful criminal.
First, and most straightforward, IQ has been shown to predict suc-
cess in almost every domain of life. Whether it is a successful marriage,

☆
This research uses data from Add Health, a program project designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris, and funded by a grant P01-HD31921 from
the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cooperative funding from 17 other agencies. Special acknowledgment is due to Ronald R.
Rindfuss and Barbara Entwisle for assistance in the original design. Persons interested in obtaining data files from Add Health should contact Add Health, Carolina Population Center, 123
W. Franklin Street, Chapel Hill, NC 27516-2524 (addhealth@unc.edu). No direct support was received from grant P01-HD31921 for this analysis.
⁎
Corresponding author.
E-mail address: cb13b@my.fsu.edu (C.M. Boccio).
1
Much of the existing research on the connection between IQ and criminal involvement has been focused on verbal IQ (Beaver et al., 2013; Ganzer & Sarason, 1973; Lynam et al., 1993;
McNulty et al., 2013; Schwartz et al., 2015). However, previous research has also linked performance IQ, full-scale IQ, and other measures of cognitive ability (e.g., spatial reasoning and
numerical reasoning) with criminal involvement and involvement with the criminal justice system (Fergusson et al., 2005; Lynam et al., 1993; Moffitt et al., 1981; Schwartz et al., 2015).

http://dx.doi.org/10.1016/j.intell.2017.10.003
Received 12 April 2017; Received in revised form 29 August 2017; Accepted 17 October 2017
0160-2896/ © 2017 Elsevier Inc. All rights reserved.

Please cite this article as: Boccio, C., Intelligence (2017), https://doi.org/10.1016/j.intell.2017.10.003
C.M. Boccio et al.
earning a hefty salary, or securing a high-quality employment, IQ tends
to be a key predictor, with higher IQ persons being, on average, more
successful than lower IQ persons (Deary et al., 2005; Herrnstein &
Murray, 1994; Schmidt & Hunter, 2004; Strenze, 2007; Zagorsky,
2007). While it may be somewhat odd to think of a criminal as being
successful, there are certainly elements that would align with success in
other areas of life. Success, for criminals, would entail engaging in a
crime without being arrested. That is why offenders go to great lengths
to conceal their identity, such as by wearing masks or trying to elim-
inate physical evidence that would tie them to a crime. Seen in this
way, one measure of criminal success is whether the criminal was able
to complete the crime and avoid detection and apprehension by the
criminal justice system.
Most criminal offenses are committed relatively impulsively,
without much forethought and planning (Gottfredson & Hirschi, 1990).
Perhaps this is part of the reason why rates of being a successful
criminal are not higher than what they are. With that said, there are
offenders who spend months planning their illegal offenses and ap-
proach crime with a significant amount of constraint (Keith, 1989;
Raine et al., 1998; Singh & Siddiqui, 2011; Sommers & Baskin, 1993).
White-collar offenders, for instance, likely spend more time co-
ordinating their criminal acts than do impulsive violent offenders. Si-
milarly, con artists and professional thieves also appear to spend more
time thinking through the crimes to be committed that would be the
most lucrative and yet would result in the lowest probability of arrest
(Cherbbonneau & Copes, 2006; Jacobs & Cherbonneau, 2014; Ouellet &
Bouchard, 2016; Vieraitis, Copes, Powell, & Pike, 2015). The ability to
plan a crime successfully, to estimate the odds of being apprehended
accurately, and to execute the crime effectively would all hinge, in large
part, on having a relatively high IQ. If these types of planned crimes
result in a lower chance of being arrested, then it stands to reason that
higher IQ offenders would be more successful at being a criminal than
lower IQ offenders.
Second, findings from official crime data indicate that, while a
significant amount of crimes are successfully solved, a vast amount of
crime goes unsolved (Paré, Felson, & Ouimet, 2007; Roberts, 2008;
United States Department of Justice, Federal Bureau of Investigation,
2013). The fact that so many offenders are able to escape detection and
arrest strongly suggests that something has to account for such varia-
tion. There are numerous possibilities that could be at play. Perhaps this
variation is inherently built into the criminal justice system and it
simply is not all that effective in identifying offenders (Keel, Jarvis, &
Muirhead, 2009; Paré et al., 2007). Almost certainly, some of the var-
iation is random, wherein criminals are more or less likely to be ap-
prehended based on chance factors (e.g., a police officer driving by at
precisely the same time a crime is occurring). But most applicable to the
current study is that there are likely individual-level factors that explain
at least some of the variation in the chances of being arrested for a
crime. While there are certainly a wide range of potential individual
differences that might explain this variation, perhaps the most likely
one is IQ.
The argument that IQ might be linked to the probability of being
arrested has been advanced previously and is widely referred to as the
differential detection explanation (Feldman, 1977; Hirschi &
Hindelang, 1977; Moffitt & Silva, 1988; Stark, 1975). The differential
detection explanation argument posits that the reason that persons with
lower IQs appear to be more criminal than persons with comparatively
higher IQs is because they are more likely to be arrested for their
crimes. Why?—according to this explanation, criminal offenders with
lower IQs are less able to escape detection and thus are more likely to
be processed through the criminal justice system than offenders with
higher IQs (Hirschi & Hindelang, 1977; Moffitt & Silva, 1988). The
evidence to date on this possibility has been somewhat mixed. Most
studies have revealed results that are inconsistent with differential
detection explanation (Hirschi & Hindelang, 1977; Moffitt & Silva,
1988); however, some studies have provided some support in favor of it
2
Intelligence xxx (xxxx) xxx–xxx
(Yun & Lee, 2013). If there is at least some differential detection that
occurs as a result of IQ, then higher IQ offenders would ultimately be
more successful at being a criminal when compared to their lower IQ
counterparts.
Third, it is possible that IQ is related to being a successful criminal
because the types of crimes committed may vary as a function of IQ. If
criminals with lower IQs commit crimes with higher arrest clearance
rates than offenders with higher IQs, then they will be more likely to be
arrested. According to the most recent data, there is tremendous var-
iation in arrest clearance rates by crime type (Paré et al., 2007; United
States Department of Justice, Federal Bureau of Investigation, 2013).
Overall, 46.8% of all violent crimes and 19.0% of all property crimes
were cleared by arrest in 2012 (United States Department of Justice,
Federal Bureau of Investigation, 2013). There is even significant var-
iation within these two general crime types (i.e., property and violent).
For instance, the arrest clearance rates for the violent crimes of murder
and rape are 62.5% and 40.1% respectively, whereas the arrest clear-
ance rates for the property crimes of larceny-theft and burglary are
22.0% and 12.7%, respectively (United States Department of Justice,
Federal Bureau of Investigation, 2013). What that necessarily means is
that if IQ is related to the type of crime being committed, then arrest
rates and, as a result, criminal success would likely vary across the IQ
spectrum.
To date, however, there has been very little quantitative evidence
bearing directly on the types of crimes committed by high-IQ vs.
comparatively low-IQ offenders. The research that does exist is far from
unequivocal and there does not seem to be much consistency in the
pattern of results published on IQ and crime type. In one study, the
findings revealed that high IQ offenders actually had a higher pre-
valence rate for most types of offenses when compared with a control
group (Oleson, 2016). This same study also showed that there were not
any significant differences between high IQ offenders and the control
group on the seriousness of the crime that was committed. Of course,
there are some types of crimes, such as white-collar crimes, where high
IQ offenders are likely to commit the vast majority of crimes. When
taken together, though, the results on the link between IQ and crime
type do not provide any clear-cut evidence of a consistent association.
But given the paucity of research on this topic, the possibility that IQ is
related to crime type remains an open-empirical question and thus
could potentially provide a reason why IQ is related to being a suc-
cessful criminal.
The current study is designed to examine directly the potential as-
sociation between IQ and rates of success for criminal offenders. To our
knowledge, this is the first study to examine this issue in this way.
Importantly, the goal is not to test the alternative explanations for the
potential association between IQ and being a successful criminal; ra-
ther, the study is focused on establishing whether there is or is not an
association. To address this possibility, we analyze data drawn from a
longitudinal and nationally representative sample of American youth.
2. Methods
2.1. Data
This study uses data drawn from the National Longitudinal Study of
Adolescent to Adult Health (Add Health). Add Health is a longitudinal
nationally representative sample of more than 90,000 American ado-
lescents that was originally collected during the 1994–1995 school
year. Three subsequent waves of data have been collected from a sub-
sample of the original respondents over the last two decades (Udry,
2003). The first wave of the study, completed when respondents were
between the ages of 12 and 21, asked questions covering a broad array
of topics including family structure, daily activities, and delinquent
involvement. Importantly, during the first wave, respondents were
administered the Picture Vocabulary Test (PVT) test which assess
verbal intelligence. The second wave, administered one year after the
C.M. Boccio et al. Intelligence xxx (xxxx) xxx–xxx

first wave to approximately 14,700 of the original respondents, asked
similar questions to the first wave, including items taping criminal in-
volvement. Five years later, the third wave of the study included ap-
proximately 15,000 of the original respondents who were between the
ages of 18 and 26, and asked questions related to labor market parti-
cipation, marital status, and criminal involvement. The fourth wave of
the study included more than 15,000 of the original respondents and
was completed in 2008, when the majority of the respondents were in
their mid-twenties to early-thirties, and asked questions pertaining to
household income, marital status, and criminal involvement (Harris
et al., 2003).
3. Measures
3.1. Outcome measures
3.1.1. Number of arrests
Number of arrests was measured using information from four items
asked during the fourth wave of the study. First, during the fourth wave
respondents were asked to indicate if they had ever been arrested
(0 = no, 1 = yes). Respondents who indicated that they had not been
arrested were instructed to skip the remaining questions pertaining to
the number of arrests. Second, respondents who indicated that they had
been arrested were asked to indicate how many times they have been
arrested (1 = once, 2 = more than once). Then, respondents who in-
dicated that they had been arrested more than once were asked how
many times they were arrested before and after their 18th birthday.
Responses to these items ranged from 1 to 32 arrests before the re-
spondents' 18th birthdays and from 1 to 56 arrests since the re-
spondents' 18th birthdays. Respondents who reported they had never
been arrested and skipped the three items pertaining to number of ar-
rests were coded as zero. Responses to the two items concerning
number of arrests before and after the respondents' 18th birthdays were
then summed along with respondents who indicated that they were
only arrested once, to reflect the total number of arrests for each re-
spondent. Values for this item ranged between 0 and 62 (Table 1).
3.1.2. Number of crimes (wave 1–wave 4)
The number of crimes committed by each respondent was calculated
by summing responses to items tapping criminal behavior at all four
waves. Number of crimes was measured at wave 1 using responses to 13
items designed to tap engagement in delinquent behavior that may
result in an arrest. For instance, respondents were asked to indicate how
often in the last 12 months they deliberately damaged property, stole
from a store, or were involved in a serious physical fight. Responses to
these items spanned from never, one or two times, three or four times,
to five or more times. Responses to these items were summed together
to create a rough count of the number of criminal acts committed at
wave 1. This list of criminal acts is similar to delinquency scales used
previously with Add Health data (Guo, Roettger, & Shih, 2007; Wright,
Beaver, DeLisi, & Vaughn, 2008). These same items were asked again at
wave 2 and used to create a similar scale of the number of criminal acts
committed at wave 2.
The number of crimes committed at wave 3 was measured using
responses to 12 items that were designed to tap criminal involvement.
Respondents, for instance, were asked to indicate how often in the last
12 months they committed burglary, were involved in a serious phy-
sical fight, used or threatened to use a weapon to get something, or
bought or sold stolen property. Responses to these items spanned from
never, one or two times, three or four times, to five or more times.
Reponses to these items were summed to create a rough count of the
number of criminal acts committed during the 12 months prior to wave
3. The number of crimes committed at wave 4 was measured using
responses to 14 items designed to tap criminal involvement over the last
year. Respondents, for example, were asked to indicate how often they
shot someone, used or threatened to use a weapon to get something,
sold drugs, or used someone else's credit card without their permission
or knowledge. Responses to these items were summed to create a rough
count of the number of crimes committed at wave 4. The items used to
measure the number of crimes committed at waves 3 and 4 have been
used previously to measure criminal involvement with Add Health data
(Guo et al., 2007; Wright et al., 2008). Finally, we summed the number
of criminal acts committed at all four waves to create a rough count of
the total number of crimes committed by respondents.

3.1.3. Rate of criminal failure (“fail rate”)

Table 1 After calculating the number of crimes committed by respondents
Descriptive statistics for scales and variables included in the analyses.
and the number of arrests reported by the respondents, we were able to
Mean Standard Range N create a variable measuring the rate at which respondents failed to “get
(percentage) deviation away with crimes” (fail rate). The fail rate variable was calculated by
dividing the respondent's number of arrests by the total number of
PVT scores 100.061 14.845 13–146 18795
crimes they committed across all four waves. This variable allows us to
Arrest
Yes (27.4) 4099 assess the number of times respondents were arrested compared to the
No (72.6) 10841 number of crimes they committed.
Number of arrests 0.755 2.410 0–62 14940
Number of crimes 6.790 8.291 0–66 9633
Number of 4.767 6.337 0–57 9633
nonviolent
3.2. Predictor variable
crimes
Number of violent 2.022 3.021 0–28 9633 3.2.1. Verbal intelligence (IQ)
crimes Verbal intelligence was measured using respondent's scores on the
Fail Rate (#arrests/
Picture Vocabulary Test (PVT). The PVT is an abridged version of the
#crimes)
(Arrested only) 0.411 0.684 0.017–10.5 2380 Peabody Picture Vocabulary Test which tests verbal aptitude and re-
Age 16.156 1.735 12–21 19758 ceptive vocabulary. Previous research has demonstrated that PPVT
Sex scores are correlated with scores from standard intelligence tests in-
Male (48.6) 9605 cluding the Wechsler Adult Intelligence Scale (WAIS; Carvajal, Shaffer,
Female (51.4) 10157
Race
& Weaver, 1989; Bell, Lassiter, Matthews, & Hutchinson, 2001; Hodapp
White (62.8) 12408 & Gerken, 1999; Dunn & Dunn, 1997). In addition, PVT scores have
Non-White (37.2) 7344 been used to measure verbal intelligence previously in the Add Health
Low self-control 0.001 0.655 −1.406–2.867 19762 data (Rowe, Jacobson, & Van den Oord, 1999). The Add Health con-
Maternal attachment −0.001 0.874 −6.137–0.450 18601
tains PVT score in raw, percentile and standardized forms. For this
Maternal 0.000 0.445 −0.855–1.487 18598
involvement study, we use the standardized form of the PVT scores measured at
wave 1.

3
C.M. Boccio et al.
3.3. Control variables
All of the analyses in the study were estimated using six control
variables measured at wave 1. Age was measured continuously in years.
Sex (0 = female, 1 = male) and race (0 = white, 1 = nonwhite) were
both measured dichotomously.
3.3.1. Low self-control
Low self-control was measured at wave 1 using responses to five
items designed to tap respondent's ability to focus and concentrate. For
instance, respondents were asked to indicate if they had difficulty
concentrating in school or had difficulty completing homework.
Responses to these five items were summed together to create a low
self-control index (α = 0.66). This low self-control index is similar to
those that have been used in previous studies analyzing the Add Health
data (Beaver, 2008; Perrone, Sullivan, Pratt, & Margaryan, 2004).
3.3.2. Maternal attachment
Maternal attachment was measured at wave 1 using responses to
two items tapping respondent's relationships with their mothers.
Specifically, respondents were asked to indicate how close they feel to
their mother and how much they think their mother cares about them.
Responses to these two items were summed together to create a ma-
ternal attachment index (α = 0.69). This maternal attachment index is
similar to those used in previous studies analyzing the Add Health data
(Beaver, 2008).
3.3.3. Maternal involvement
Maternal involvement was measured using responses to 10 items
asked during the first wave of the survey. Respondents, for instance,
were asked to indicate if they went to a movie, went shopping, or
worked on a project with their mother over the last four weeks.
Responses to these items were summed together to create an index of
maternal involvement (α = 0.55). This maternal involvement index is
similar to those used in previous studies analyzing the Add Health data
(Crosnoe & Elder, 2002).
3.4. Analytic strategy
The analysis for this study proceeded in a number of interlinked
steps. First, we examined the relationship between verbal IQ and
number of arrests using negative binomial regression in the full sample.
Second, we used logistic regression to test whether verbal IQ predicts
the odds of arrest for individuals in the top quartile (11 to 66 crimes)
and the bottom quartile (1 to 2 crimes) of the distribution of the
number of crimes committed. Third, we assessed the relationship be-
tween verbal IQ and the number of crimes committed using negative
binomial regression in the full sample. Then, given well established
disparities in criminal involvement and involvement with the criminal
justice system between males and females (Steffensmeier & Allan,
1996), we tested whether the relationships among verbal IQ, number of
arrests, and number of crimes committed differed across gender.
Finally, we tested whether verbal IQ is associated with “getting
away with crime.” In order to do so, we estimated two separate analyses
employing the fail rate measure. In the first analysis, we tested if verbal
IQ predicted respondent's fail rate among respondents who had been
arrested using ordinary least squares (OLS) regression in the full sample
and in males and females separately. In the second analysis we added a
small constant to both sides of the ratio used to construct the fail rate
measure and then tested whether verbal IQ predicts fail rate in the full
sample (also including respondents who had not been arrested) using
OLS regression. We then tested this same relationship in males and
females separately.
4
Intelligence xxx (xxxx) xxx–xxx
4. Results
The analysis for this study began by examining the relationship
between verbal IQ and number of arrests. As can be seen in Table 2,
verbal IQ has a significant negative association with number of arrests
for the full sample and for males and females. These findings indicate
that individuals with higher verbal IQs tend to report fewer arrests than
individuals with lower verbal IQs.
Next, we examined the relationship between verbal IQ and the
likelihood of arrest for respondents in the top quartile and the bottom
quartile of the distribution of the number of crimes committed.
Examination of Table 3 reveals that verbal IQ has a significant negative
association with the likelihood of arrest for respondents in the bottom
quartile of crimes committed (committed one to two crimes). This
finding indicates that individuals with higher verbal IQs who com-
mitted only one or two crimes were significantly less likely to be ar-
rested than individuals with lower verbal IQs who committed the same
number of crimes. Further examination of Table 3 reveals that verbal IQ
is not significantly associated with the likelihood of arrest for re-
spondents in the top quartile (11–66 crimes) of the distribution of the
number of crimes committed. This finding means that verbal IQ did not
influence the likelihood of arrest among respondents who committed a
large number of criminal acts.
In order to more fully explore the relationship between number of
criminal acts, verbal IQ, and probability of arrest, we plotted the pre-
dicted probabilities of arrest for offenders given their verbal IQs (high,
average, low) and the number of crimes committed. The verbal IQ ca-
tegories were separated into three groups with one group containing
individuals with verbal IQs 1 standard deviation above the mean (high),
on group containing individuals with verbal IQs one standard deviation
below the mean (low), and a group containing individuals with verbal
IQs near the mean (average). As can be seen in Fig. 1, individuals with
high verbal IQs were less likely to be arrested across the range of
number of criminal acts committed, however, as the number of crimes
committed increases the probability of arrest for individuals with high
verbal IQs becomes closer to the probability of arrest for individuals
with average and low verbal IQs. In addition, Fig. 1 also demonstrated
that the predicted probabilities of arrest for individuals with average
and low verbal IQs are roughly similar across the range of number of
criminal acts committed.
Then, we examined the relationship between verbal IQ and the
number of crimes committed by respondents. Table 4 reveals that
verbal IQ is not significantly associated with the number of crimes
committed in the full sample or in the male sample. In contrast, Table 4
also reveals that verbal IQ has a significant and negative association
with the number of crimes committed by females. The negative asso-
ciation between verbal IQ and number of crimes committed indicates
that females with higher verbal IQs tended to report fewer criminal acts
than females with lower verbal IQs. These findings indicate that verbal
IQ only appears to influence the number of criminal acts committed by
females and not males.
Given that verbal IQ appears to influence the number of arrests
reported by respondents, we then examined whether verbal IQ is as-
sociated with “getting away with crimes” among respondents who had
been arrested. In order to do so, we examined the association between
verbal IQ and criminal “fail rate,” or the rate at which respondents were
arrested per the number of crimes they committed. As can be seen in
Table 5, verbal IQ is significantly associated with fail rate in the full
sample and in the male sample. The coefficients for the associations
between verbal IQ and fail rate are negative indicating that respondents
with higher verbal IQs were more likely to get away with crime. On the
other hand, this finding means that respondents with lower verbal IQs
were more likely to apprehended and arrested for committing crimes. In
contrast, the coefficient for the relationship between verbal IQ and fail
rate is not significant for females, indicating that verbal IQ is not as-
sociated with getting away with crime for females.
C.M. Boccio et al. Intelligence xxx (xxxx) xxx–xxx

Table 2 Table 3
Negative binomial regression models for the association between intelligence and number Logistic regression models for the association between intelligence, bottom 25% of crimes
of arrests. committed, top 25% of crimes committed, and the likelihood of being arrested.

Full sample Males Females Odds of arrest

Coeff IRR Coeff IRR Coeff IRR Bottom 25% (1–2 crimes) Top 25% (11–66 crimes)

Predictor variables Coeff O.R. Coeff O.R.

PVT Score − 0.016 0.984⁎⁎ −0.015 0.985⁎⁎ − 0.017 0.983⁎⁎
(0.002) (0.002) (0.003) Predictor variables
Number of crimes 0.069 1.072⁎⁎ 0.063 1.065⁎⁎ 0.086 1.090⁎⁎ PVT score − 0.015 0.985⁎⁎ −0.005 0.995
(0.004) (0.004) (0.007) (0.005) (0.004)
Controls Number of crimes 0.410 1.506⁎⁎ 0.036 1.037⁎⁎
Age − 0.011 0.989 − 0.005 0.996 −0.033 0.967 (0.201) (0.006)
(0.015) (0.019) (0.026) Controls
Race 0.066 1.068 0.040 1.040 0.111 1.118 Age 0.042 1.043 0.031 1.032⁎
(0.059) (0.074) (0.099) (0.043) (0.033)
Sex 1.141 3.131⁎⁎ – – Race 0.220 1.247 0.049 1.050
(0.165) (0.176) (0.112)
Low self-control 0.243 1.275⁎⁎ 0.185 1.204⁎⁎ 0.292 1.338⁎⁎ Sex 1.030 2.802⁎⁎ 0.952 2.591⁎⁎
(0.051) (0.063) (0.087) (0.380) (0.275)
Maternal − 0.063 0.939⁎ −0.140 0.870⁎⁎ 0.015 1.015 Low self-control 0.310 1.364⁎⁎ −0.017 0.984
attachment (0.153) (0.073)
(0.028) (0.036) (0.047) Maternal attachment 0.114 1.121 −0.106 0.899⁎
Maternal − 0.084 0.919 −0.260 0.771⁎⁎ 0.172 1.187 (0.110) (0.049)
involvement Maternal involvement − 0.114 0.892 −0.035 0.966
(0.055) (0.058) (0.116) (0.141) (0.109)
N 8803 3880 4923 N 1947 1830

⁎ ⁎
p < 0.05. p < 0.05.
⁎⁎ ⁎⁎
p < 0.01. p < 0.01.

Finally, we examined the association between verbal IQ and fail rate
in the entire sample, including respondents who had not been arrested.
Table 6 reveals that, once again, verbal IQ is significantly and nega-
tively associated with fail rate in the full sample and in the male
sample. This finding indicates that individuals with higher verbal IQs
are more likely to avoid being arrested for committing crimes. Further
examination of Table 6, however, reveals that the association between
verbal IQ and fail rate is not significant for females.2
5. Discussion
There is likely no other variable that predicts success in such a wide
array of life outcomes as IQ. Whereas having a high IQ is associated
with more positive outcomes, having a comparatively lower IQ has the
opposite effect and is thus related with a host of negative outcomes
(Herrnstein & Murray, 1994; Strenze, 2007). Very little attention,
however, has been paid to the way in which verbal IQ might predict
success in areas of life that are largely considered antisocial or criminal.
The current study sought to partially address this gap in the literature
by examining whether verbal IQ was related to being a successful
criminal—that is, committing crimes and not being apprehended by the
criminal justice system. The results of the analyses revealed two key
findings.
First, and in line with previous research, verbal IQ was a statistically
significant predictor of criminal justice processing (i.e., being arrested),
wherein persons with comparatively lower verbal IQ scores were more
likely to be arrested (Beaver et al., 2013; Hirschi & Hindelang, 1977;
2
A reviewer suggested that it would interesting to test whether the relationships
among verbal IQ, number of crimes, and number of arrests varied by type of crime.
Therefore, we calculated separate counts for violent and nonviolent crimes, and then we
tested all of these same relationships with the two different types of crimes. In sum,
findings from the analyses split by violent and nonviolent crimes revealed that verbal IQ
is negatively associated with number of arrests and number of violent and nonviolent
crimes. In addition, verbal IQ appears to be negatively associated with fail rate for both
nonviolent and violent crimes in the full sample. However, these findings also reveal that
verbal IQ is not significantly associated with fail rate for violent crimes in the sample of
respondents who reported at least one arrest. The tables with these results are available
upon request.
Moffitt et al., 1981). Verbal IQ was also related to the total number of
self-reported crimes, but only for females, not males. This is an inter-
esting finding because previous research has indicated a significant
association between verbal IQ and criminal involvement for both males
and females (Beaver et al., 2013). What accounts for this divergent
finding remains unknown at this point. However, it is worth mentioning
that our study did not examine the type(s) of crime(s) being committed
by males and females nor did it examine at what point in the life course
the individual was engaging in crime. Perhaps as a result, the effect of
verbal IQ on criminal involvement was attenuated as the crime measure
was pooling together serious and non-serious crimes, many of which
were being committed during adolescence when crime and delinquency
are developmentally normal. Future research should examine the po-
tential reasons for this particular finding.
Second, and perhaps most interesting, verbal IQ was found to be
related to being a successful criminal, with higher verbal IQ criminals
being more likely to avoid apprehension when compared with lower
verbal IQ criminals. Additional analyses revealed that this verbal IQ-
criminal success nexus was confined only to males. Why this association
was only detected for males is not clear at this point. While admittedly
speculatively, we offer one possibility. Perhaps females engage more
frequently in crimes that have higher detection and arrest rates than do
males. There is some evidence to support this possibility. Females, for
instance, are disproportionately involved in theft and drug crimes,
crimes that have higher arrest rates than do other types of offenses
dominated by males, such as sexual assault and burglary (Steffensmeier
& Allan, 1996). If that is the case, then arrest rates for females may be
more function of the crime that they are committing than whether they
are adept criminals.
The verbal IQ-successful criminal association that was detected for
males is an intriguing finding and one that begs the question of the
underlying mechanism that accounts for this result. A closer look at the
analyses revealed that the effect of verbal IQ on being a successful
criminal was found only among individuals who were committing re-
latively few crimes (i.e., they scored in the bottom 25% of self-reported
crime), not among those who were committing a large number of
crimes (i.e., they scored in the top 25% of self-reported crime). This
finding might seem somewhat counterintuitive at first glance, but it

5
C.M. Boccio et al. Intelligence xxx (xxxx) xxx–xxx

Fig. 1. The relationship between verbal IQ, number of crimes,

and probability of arrest.

Table 4 criminal involvement was not entirely precise. Rather than having each
Negative binomial regression models for the association between intelligence and number subject report the total number of crimes that they committed, the re-
of crimes.
sponses were grouped into categories and capped at 5 crimes per year.
Full sample Males Females As a result, the most high-rate offenders would not necessarily be se-
parated with precision from medium-rate offenders. In addition, and
Coeff IRR Coeff IRR Coeff IRR directly related, the number of crimes measures employed in the cur-
rent study likely represent an underestimation of the overall number of
Predictor variables
PVT score − 0.001 0.999 0.002 1.002 − 0.004 0.996⁎⁎ criminal acts each respondent engaged in across all four waves of data
(0.001) (0.001) (0.001) collection, as the indices used only tap common forms of criminal be-
Controls havior as opposed to an exhaustive list. Collectively, these data lim-
Age − 0.076 0.927⁎⁎ −0.039 0.962⁎⁎ − 0.106 0.900⁎⁎ itations would be more likely to result in an attenuated effect of verbal
(0.007) (0.010) (0.010)
IQ on criminal success, increasing the probability of Type II error.
Race 0.120 1.127⁎⁎ 0.101 1.106⁎⁎ 0.141 1.152⁎⁎
(0.029) (0.040) (0.040) Future research employing more comprehensive criminal activity
Sex 0.557 1.745⁎⁎ – – measures would be beneficial in addressing this limitation. In addition,
(0.041) the Add Health only included one measure of verbal IQ: the PVT score.
Low self-control 0.627 1.871⁎⁎ 0.577 1.781⁎⁎ 0.669 1.952⁎⁎
While the Add Health PVT has been used widely in intelligence research
(0.035) (0.048) (0.051)
Maternal − 0.114 0.893⁎⁎ −0.082 0.921⁎⁎ − 0.126 0.882⁎⁎ (Rowe et al., 1999; Schwartz & Beaver, 2013), it would be interesting to
attachment determine whether the findings reported here would be robust to dif-
(0.013) (0.022) (0.016) ferent measures of IQ. Future research would benefit by addressing this
Maternal 0.027 1.028 0.042 1.043 0.023 1.023 possibility. Furthermore, previous research indicates that verbal IQ may
involvement
be related to attrition in longitudinal studies. For example, a study by
(0.027) (0.041) (0.037)
N 8803 3880 4923 Beaver (2013), indicates that participants who dropped out of the Add
Health study between waves one and four had verbal IQ scores 4.5
⁎⁎
p < 0.01. points lower on average than respondents who remained in the study.
These findings indicate that the analytic sample for this study may have
likely provides some great insight into why verbal IQ is related to being had verbal IQs that were slightly higher than the original sample on
a successful criminal. For low-rate offenders, verbal IQ matters more average. As a result, a different pattern of findings may have emerged
because the criminal justice system is not all that efficient at clearing using a different sample that had not experienced selective attrition.
crimes by arrests. In other words, most offenders can likely get away Rarely is there a study that shows that IQ does not predict human
with a few crimes without being apprehended. It is here—among low- outcomes. The current study does not deviate from this pattern of re-
rate offenders—where verbal IQ matters because those who are caught sults and once again underscores the role of verbal IQ in yet another
after only a crime or two are likely committing crimes without much outcome. Of course, the unique aspect of this outcome is that it is
thinking, planning, or coordinating. Among high-rate offenders, measuring a behavior—criminal involvement—that, for the most part,
though, the laws of probability catch up with all offenders—even if they would be viewed as being designed, in part, by low verbal IQ, not high
are intelligent and plan their crimes carefully. In this case, verbal IQ verbal IQ. When it comes to the effects of verbal IQ, however, it appears
matters less (or not at all) because the high rate of offending necessarily that how individuals define success—in this case, being a successful
means that the probability of arrest for each crime is compounded criminal—is more important than how society views and defines the
additively across their criminal career. The end result is a relatively behavior. Whatever the case, offenders who have relatively higher
high probability of arrest regardless of verbal IQ. verbal IQs are more likely to avoid apprehension and arrest for their
While this is one of the first studies to link verbal IQ to criminal crimes when compared to offenders who have relatively lower verbal
success, the results should be interpreted with caution in light of a IQs.
number of limitations. To begin with, the measurement of self-reported

6
C.M. Boccio et al. Intelligence xxx (xxxx) xxx–xxx

Table 5
OLS regression models for the association between intelligence and fail rate (arrested only).

Full sample Males Females

b Beta b Beta b Beta

Predictor variables
PVT score −0.003 −0.064⁎⁎ − 0.004 − 0.074⁎⁎ − 0.001 − 0.030
(0.001) (0.001) (0.002)
Controls
Age 0.015 0.037 0.016 0.038 0.005 0.013
(0.009) (0.012) (0.013)
Race 0.019 0.014 0.006 0.004 0.044 0.040
(0.031) (0.042) (0.043)
Sex 0.051 0.037 – –
(0.031)
Low self-control −0.081 −0.083⁎⁎ −0.069 − 0.064⁎ − 0.098 − 0.123⁎⁎
(0.022) (0.029) (0.030)
Maternal attachment − 0.002 −0.002 −0.022 −0.025 0.019 0.037
(0.017) (0.024) (0.021)
Maternal involvement − 0.044 −0.030 −0.071 −0.044 0.010 0.008
(0.033) (0.045) (0.048)
N 2137 1405 732

⁎
p < 0.05.
⁎⁎
p < 0.01.

Table 6
OLS regression models for the association between intelligence and fail rate (constant added).

Full sample Males Females

b Beta b Beta b Beta

Predictor variables
PVT score −0.002 −0.047⁎⁎ − 0.003 − 0.083⁎⁎ − 0.000 − 0.007
(0.000) (0.001) (0.000)
Controls
Age 0.020 0.064⁎⁎ 0.015 0.039⁎⁎ 0.023 0.088⁎⁎
(0.003) (0.006) (0.004)
Race −0.023 −0.022⁎ 0.000 0.000 − 0.038 − 0.044⁎⁎
(0.012) (0.021) (0.012)
Sex − 0.008 −0.008 – –
(0.011)
Low self-control −0.155 −0.199⁎⁎ − 0.132 − 0.144⁎⁎ − 0.174 − 0.269⁎⁎
(0.008) (0.015) (0.009)
Maternal attachment 0.009 0.015 −0.021 −0.025 0.021 0.045⁎⁎
(0.007) (0.014) (0.007)
Maternal involvement − 0.018 −0.016 −0.052 − 0.037⁎ 0.004 0.005
(0.012) (0.022) (0.013)
N 8803 3880 4923

⁎
p < 0.05.
⁎⁎
p < 0.01.

Appendix A. Correlation matrix of variables included in the analysis.

X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12

PVT score X1 1
Arrest (1/0) X2 − 0.05⁎ 1
Arrest number X3 − 0.05⁎ 0.51⁎ 1
Crime number X4 0.01 0.32⁎ 0.30⁎ 1
Nonviolent crime number X5 0.07⁎ 0.28⁎ 0.25⁎ 0.95⁎ 1
Violent crime number X6 − 0.11⁎ 0.30⁎ 0.30⁎ 0.75⁎ 0.51⁎ 1
Age X7 − 0.05⁎ −0.00 − 0.01 −0.05⁎ − 0.05⁎ −0.03⁎ 1
Sex X8 0.06⁎ 0.26⁎ 0.20⁎ 0.21⁎ 0.16⁎ 0.22⁎ 0.05⁎ 1
Race X9 − 0.28⁎ 0.03⁎ 0.03⁎ 0.02 − 0.02 0.08⁎ 0.03⁎ − 0.02⁎ 1
Low self-control X10 − 0.00 0.14⁎ 0.13⁎ 0.35⁎ 0.34⁎ 0.25⁎ 0.45⁎ 0.05⁎ − 0.03⁎ 1
Maternal attachment X11 − 0.01 −0.04⁎ − 0.05⁎ −0.12⁎ − 0.13⁎ −0.06⁎ − 0.08⁎ 0.07⁎ 0.01 − 0.20⁎ 1
Maternal involvement X12 0.08⁎ −0.06⁎ − 0.06⁎ −0.04⁎ − 0.04⁎ −0.05⁎ − 0.08⁎ − 0.16⁎ − 0.02 − 0.04⁎ 0.21⁎ 1
⁎

⁎
p < 0.05.

7
C.M. Boccio et al.
Appendix B. Supplementary data
Supplementary data to this article can be found online at https://doi.org/10.1016/j.intell.2017.10.003.
References
Bartels, J. M., Ryan, J. J., Urban, I. S., & Glass, I. A. (2010). Correlations between esti-
mates of state IQ and FBI crime statistics. Personality and Individual Differences, 48,
579–583.
Beaver, K. M. (2008). Nonshared environmental influences on adolescent delinquent in-
volvement and adult criminal behavior. Criminology, 46, 341–369.
Beaver, K. M. (2013). Intelligence and selective attrition in a nationally representative
and longitudinal sample of Americans. Personality and Individual Differences, 55,
157–161.
Beaver, K. M., Schwartz, J. A., Nedelec, J. L., Connolly, E. J., Boutwell, B. B., & Barnes, J.
C. (2013). Intelligence is associated with criminal justice processing: Arrest through
incarceration. Intelligence, 41, 277–288.
Beaver, K. M., & Wright, J. P. (2011). The association between county-level IQ and
county-level crime rates. Intelligence, 39, 22–26.
Bell, N. L., Lassiter, K. S., Matthews, T. D., & Hutchinson, M. B. (2001). Comparison of the
picture Peabody vocabulary test-third edition and Wechsler adult intelligence scale-
third edition with university students. Journal of Clinical Psychology, 57, 417–422.
Carvajal, H., Shaffer, C., & Weaver, K. A. (1989). Correlations of scores of maximum
security inmates on Wechsler adult intelligence scale-revised and Peabody picture
vocabulary test-revised. Psychological Reports, 65, 268–270.
Cherbbonneau, M., & Copes, H. (2006). ‘Drive it like you stole it’: Auto theft and the
illusion of normalcy. The British Journal of Criminology, 46, 193–211.
Crosnoe, R., & Elder, G. H. (2002). Family dynamics, supportive relationships, and edu-
cational resilience during adolescence. Journal of Family Issues, 25, 571–602.
Deary, I. J., Taylor, M. D., Hart, C. L., Wilson, V., Smith, G. D., Blane, D., & Starr, J. M.
(2005). Intergenerational social mobility and mid-life status attainment: Influences of
childhood intelligence, childhood social factors, and education. Intelligence, 33,
455–472.
Diamond, B., Morris, R. G., & Barnes, J. C. (2012). Individual and group IQ predict inmate
violence. Intelligence, 40, 115–122.
Dunn, L. M., & Dunn, L. M. (1997). Peabody picture vocabulary test-third edition. Circle
Pines, MN: American Guidance Service.
Feldman, M. P. (1977). Criminal behavior: A psychological analysis. New York, NY: Wiley.
Fergusson, D. M., Horwood, J., & Ridder, E. M. (2005). Show me the child at seven II:
Childhood intelligence and later outcomes in adolescence and young adulthood.
Journal of Child Psychology and Psychiatry, 46, 850–858.
Ganzer, V. J., & Sarason, I. G. (1973). Variables associated with recidivism among ju-
venile delinquents. Journal of Consulting and Clinical Psychology, 40, 1–5.
Gottfredson, M. R., & Hirschi, T. (1990). A general theory of crime. Stanford, CA: Stanford
University Press.
Guo, G., Roettger, M. E., & Shih, J. C. (2007). Contributions of the DAT1 and DRD2 genes
to serious and violent delinquency among adolescents and young adults. Human
Genetics, 121, 125–136.
Harris, K. M., Florey, F., Tabor, J., Bearman, P. S., Jones, J., & Udry, J. R. (2003). The
national longitudinal study of adolescent health: Research design. Retrieved from
http://www.cpc.unc.edu/projects/addhealth/design.
Herrnstein, R. J., & Murray, C. (1994). The bell curve: Intelligence and class structure in
American life. New York, NY: Free Press.
Hirschi, T., & Hindelang, M. J. (1977). Intelligence and delinquency: A revisionist review.
American Sociological Review, 42, 571–587.
Hodapp, A. F., & Gerken, K. C. (1999). Correlations between scores for Peabody picture
vocabulary test-III and the Wechsler intelligence scale for children-III. Psychological
Reports, 84, 1139–1142.
Jacobs, B. A., & Cherbonneau, M. (2014). Auto theft and restrictive deterrence. Justice
Quarterly, 31, 344–367.
Keel, T. G., Jarvis, J. P., & Muirhead, Y. E. (2009). An exploratory analysis of factors
affecting homicide investigations: Examining the dynamics of murder clearance rates.
Homicide Studies, 13, 50–68.
Keith, P. (1989). The business of crime: Research findings reveal economic motives. MTAS
Publications: Technical Bulletins.
Latvala, A., Castaneda, A. E., Perälä, J., Saarni, S. I., Aalto-Setälä, T., Lönnqvist, J., ...
Tuulio-Henriksson, A. (2009). Cognitive functioning in substance abuse and depen-
dence: A population-based study of young adults. Addiction, 104, 1558–1568.
Lussier, P., Bouchard, M., & Beauregard, E. (2011). Patterns of criminal achievement in
sexual offending: Unraveling the “successful” sex offender. Journal of Criminal Justice,
39, 433–444.
Lynam, D. R., Moffitt, T. E., & Stouthamer-Loeber, M. (1993). Explaining the relation
between IQ and delinquency: Class, race, test motivation, school failure, or self-
control? Journal of Abnormal Psychology, 102, 187–196.
McCall, G. (1978). Observing the law. New York, NY: Free Press.
Intelligence xxx (xxxx) xxx–xxx
McDaniel, M. A. (2006). Estimating state IQ: Measurement challenges and preliminary
correlates. Intelligence, 34, 607–619.
McNulty, T. L., Bellair, P. E., & Watts, S. J. (2013). Neighborhood disadvantage and
verbal ability as explanations of the Black-White difference in adolescent violence:
Toward an integrated model. Crime and Delinquency, 59, 140–160.
Moffitt, T. E., Gabrielli, W. F., Mednick, S. A., & Schulsinger, F. (1981). Socioeconomic
status, IQ, and delinquency. Journal of Abnormal Psychology, 90, 152–156.
Moffitt, T. E., & Silva, P. A. (1988). IQ and delinquency: A direct test of the differential
detection hypothesis. Journal of Abnormal Psychology, 97, 330–333.
Oleson, J. C. (2016). Criminal genus: A portrait of high-IQ offenders. Oakland, CA:
University of California Press.
Ouellet, F., & Bouchard, M. (2016). Only a matter of time? The role of criminal compe-
tence in avoiding arrest. Justice Quarterly, 34, 699–726. http://dx.doi.org/10.1080/
07418825.2016.1219761.
Paré, P.-P., Felson, R. B., & Ouimet, M. (2007). Community variation in crime clearance:
A multilevel analysis with comments on assessing police performance. Journal of
Quantitative Criminology, 23, 243–258.
Perrone, D., Sullivan, C. J., Pratt, T. C., & Margaryan, S. (2004). Parental efficacy, self-
control, and delinquency: A test of a general theory of crime on a nationally re-
presentative sample of youth. International Journal of Offender Therapy and
Comparative Criminology, 48, 298–312.
Raine, A., Meloy, J. R., Bihrle, S., Stoddard, J., LaCasse, L., & Buchsbaum, M. S. (1998).
Reduced prefrontal and increased subcortical brain functioning assessed using posi-
tron emission tomography in predatory and affective murders. Behavioral Sciences &
the Law, 16, 319–332.
Richter, P., Scheurer, H., Barnett, W., & Kröber, H. L. (1996). Forecasting recidivism in
delinquency by intelligence and related constructs. Medicine, Science, and the Law, 36,
337–342.
Roberts, A. (2008). The influences of incident and contextual characteristics on crime
clearance of nonlethal violence: A multilevel event history analysis. Journal of
Criminal Justice, 36, 61–71.
Rowe, D. C., Jacobson, K. C., & Van den Oord, E. J. (1999). Genetic and environmental
influences on vocabulary IQ: Parental education level as moderator. Child
Development, 70, 1151–1162.
Schmidt, F. L., & Hunter, J. (2004). General mental ability in the world of work:
Occupational attainment and job performance. Journal of Personality and Social
Psychology, 86, 162–173.
Schwartz, J. A., & Beaver, K. M. (2013). Examining the effects of dopamine genes on
verbal IQ within and between families. Journal of Neural Transmission, 120, 477–486.
Schwartz, J. A., Savolainen, J., Aaltonen, M., Merikukka, M., Paananen, R., & Gissler, M.
(2015). Intelligence and criminal behavior in a total birth cohort: An examination of
functional form, dimensions of intelligence, and the nature of offending. Intelligence,
51, 109–118.
Singh, A. K., & Siddiqui, A. T. (2011). New face of terror: Cyber threats, emails containing
viruses. Asian Journal of Technology & Management Research, 1 (ISSN: 2249-0892).
Sommers, I., & Baskin, D. R. (1993). The situational context of violent female offending.
Journal of Research in Crime and Delinquency, 30, 136–162.
Stark, R. (1975). Social problems. New York, NY: Random House.
Steffensmeier, D., & Allan, E. (1996). Gender and crime: Toward a gendered theory of
female offending. Annual Review of Sociology, 22, 459–487.
Strenze, T. (2007). Intelligence and socioeconomic success: A meta-analytic review of
longitudinal research. Intelligence, 35, 401–426.
Udry, J. R. (2003). The national longitudinal study of adolescent health (Add Health), wave
III, 2001–2002; wave IV, 2008–2009. Los Altos, CA: The Carolina Population Center.
United States Department of Justice, Federal Bureau of Investigation (Fall 2013). Crime in
the U.S. 2012. Retrieved 2/28/17, from https://ucr.fbi.gov/crime-in-the-u.s/2012/
crime-in-the-u.s.-2012/offenses-known-to-law-enforcement/clearancetopic.pdf.
Vieraitis, L. M., Copes, H., Powell, Z. A., & Pike, A. (2015). A little information goes a long
way: Expertise and identity theft. Aggression and Violent Behavior, 20, 10–18.
Watters, J., & Biernacki, P. (1989). Targeted sampling: Options for the study of hidden
populations. Social Problems, 36, 416–430.
Wright, J. P., Beaver, K. M., DeLisi, M., & Vaughn, M. (2008). Evidence of negligible
parenting influences on self-control, delinquent peers, and delinquency in a sample of
twins. Justice Quarterly, 25, 544–569.
Wright, R., Decker, S. H., Redfern, A. K., & Smith, D. L. (1993). A snowball's chance in
hell: Doing fieldwork with active residential burglars. Journal of Research in Crime and
Delinquency, 29, 148–161.
Yun, I., & Lee, J. (2013). IQ and delinquency: The differential detection hypothesis re-
visited. Youth Violence and Juvenile Justice, 11, 196–211.
Zagorsky, J. L. (2007). Do you have to be smart to be rich? The impact of IQ on wealth,
income and financial distress. Intelligence, 35, 489–501.