Personality Dimensions and Criminal Arrest
Jack Samuels, O. Joseph Bienvenu, Bernadette Cullen, Paul T. Costa Jr,
William W. Eaton, and Gerald Nestadt
Previous studies have implicated antisocial personality
disorder in criminal behavior, but little is known
about the association between "normal" personality
dimensions and arrest. We investigated the relationships
between these personality dimensions and prior
arrest in a sample of adults participating in a longitudinal
epidemiological study. Between 1993 and 1999,
psychiatrists re-examined subjects who were originally
interviewed in Baltimore in 1981 as part of the
Epidemiologic Catchment Area study; the psychiatrists
diagnosed axis I and axis II disorders according
to DSM-IV criteria. A total of 611 subjects also completed
the Revised NEO Personality Inventory (NEOPI-R),
which assesses five broad factors and 30 facets
of normal personality. History of criminal arrest in
Maryland in the period 1981 to 1993 was determined
from the state criminal justice database. Studenťs t
test and logistic regression were used to evaluate
relationships between NEO personality scores and
prior arrest. Controlling for demographic characteristics,
alcohol or drug use disorders, and DSM-IV personality
disorder scores, the odds of prior arrest increased
with scores on angry hostility, impulsiveness,
and excitement-seeking dimensions. Prior arrest was
inversely related to scores on trust, straightforwardness,
compliance, modesty, dutifulness, and deliberation
dimensions. The results suggest that specific
dimensions of normal personality are related to criminal
arrest in the community.
 2004 Elsevier Inc. All rights reserved.
CRIMINAL ACTIVITY is a leading cause of
morbidity and mortality in the United States
and adversely impacts the quality of life of communities
in this country.1,2 The etiology of these
behaviors appears complex, involving numerous
individual, interpersonal, neighborhood, and community
factors.3,4 Results from several recent studies
in patients,5-9 arrestees,10-13 and members of
birth cohorts14-17 suggest that individuals with psychiatric
disorders--especially schizophrenia, alcohol
and drug use disorders, and antisocial personality
disorder--are at increased risk of criminal
arrest.
Less is known about relationships between personality
dimensions and arrest. McMillen et al.18
found that individuals arrested multiple times for
driving under the influence of alcohol scored significantly
higher on scales of hostility, sensation
seeking, and psychopathic deviance than did firsttime
offenders. Ulrich et al.19 reported that criminal
offenders scored significantly different than
noncriminal controls on all ICD-10 personality disorder
dimensions. Johnson et al.20 found that, in a
community-based sample of adolescents, paranoid,
narcissistic, and passive-aggressive personality
traits were associated with the self-reported commission
of violent acts over the subsequent 10
years.
The focus of these studies has been on clinical
criteria for personality disorders. The overlap between
"abnormal" and "normal" conceptualizations
of personality is an area of active investigation,
and it has been proposed that normal
personality dimensions may provide additional and
sometimes richer clinical insight into an individuaľs
strengths and potential problems.21 In the current
study, we evaluated the relationship between
personality dimensions and documented prior arrest
in a sample of adults participating in a longitudinal
epidemiological study. The aim was to determine
if "normal" personality dimensions are
related to criminal arrest and if these relationships
are independent of other important demographic
and clinical features, including DSM-IV personality
disorder dimensions.
METHOD
Sample
Subjects participating in the Hopkins Epidemiology of Personality
Disorder Study were sampled from the Baltimore Epidemiological
Catchment Area follow-up survey, as described
previously.22 In brief, in 1981, a total of 3,481 adult household
residents of east Baltimore were sampled probabilistically and
From the Department of Psychiatry and Behavioral Sciences,
The Johns Hopkins University School of Medicine, Baltimore,
MD; Department of Mental Hygiene, Bloomberg School of
Public Health, Baltimore, MD; and the Gerontology Research
Center, National Institute on Aging, Baltimore, MD.
Supported by Grant No. MH50616 and MH47447 from the
National Institute on Mental Health, and a Johns Hopkins
University School of Medicine Clinical Scientist Award (to
J.B.).
Address correspondence to Jack Samuels, Ph.D., Department
of Psychiatry and Behavioral Sciences, Johns Hopkins Hospital,
Meyer 4-181, 600 N Wolfe St, Baltimore, MD 21287.
 2004 Elsevier Inc. All rights reserved.
0010-440X/04/4504-0012$30.00/0
doi:10.1016/j.comppsych.2004.03.013
Comprehensive Psychiatry, Vol. 45, No. 4 (July/August), 2004: pp 275-280 275
interviewed by lay interviewers using the Diagnostic Interview
Schedule (DIS)23; 810 of these individuals also were examined
by psychiatrists at that time. Between 1993 and 1996, a total of
1,920 (73%) of the surviving subjects were reinterviewed. From
these 1,920 subjects, we invited for psychiatric examination
subjects with a range of axis I psychopathology, including all
those who were identified by the DIS as having a lifetime
diagnosis of mania, depression, panic disorder, obsessive-compulsive
disorder, alcohol use disorders, or drug use disorders at
follow-up; had incident (i.e., between 1981 and follow-up)
DIS-ascertained social phobia, agoraphobia, or cognitive impairment;
or were examined by psychiatrists in 1981. In addition,
we randomly selected approximately 25% of the remaining
subjects for examination so as to have a sufficient number of
individuals not selected for psychopathology. After providing
informed consent, 816 subjects received the psychiatric examination
between 1993 and 1999.
The current study is restricted to the 611 subjects who completed
the normal personality inventory. The age of the subjects
ranged from 30 to 87 years. Of these, 230 (38%) were men and
381(62%) were women; 378 (62%) were Caucasian and 230
belonged to other racial/ethnic groups. Subjects completing the
inventory were slightly younger, on average, than the 205 who
did not (mean ages, 47.5 and 50.5 years, respectively) (t814
2.7, P .01), but the gender distributions of the two groups
were similar (62% and 69% female, respectively).
Diagnostic Procedures
The subjects were examined by psychiatrists using the World
Health Organization Schedules for Clinical Assessment in Neuropsychiatry
(SCAN) to evaluate axis I disorders.24 The psychiatrists
assessed axis II personality disorder criteria using the
personality disorder section of Standardized Psychiatric Examination
(SPE)25 supplemented by DSM-IV axis II criteria26 that
were not part of the original version; the reliability of the
original instrument has been reported previously.27 All DSM-IV
personality disorder criteria were included, and biographical
information, emphasizing interpersonal relationships, was collected
on each subject. The psychiatrist rated each personality
criterion on a 3-point scale ranging from 0 (absent) to 2 (trait
definitely present and has caused the subject distress and/or
disruption of social and occupational functioning); a score of 1
meant that the feature was present but did not cause the subject
substantial distress or dysfunction. A dimensional score was
calculated for each of the 10 DSM-IV personality disorders by
summing the scores for each constituent feature of the specific
disorder.27
Normal personality was assessed with the self-completed
paper-and-pencil form of the Revised NEO Personality Inventory
(NEO PI-R).28 This instrument assesses the five domains of
normal personality as construed by the Five-Factor Model:
neuroticism, extraversion, openness, agreeableness, and conscientiousness.
Each domain is represented by six specific facet
scales. Neuroticism facets are anxiety, angry hostility, depression,
self-consciousness, impulsiveness, and vulnerability. Extraversion
facets are warmth, gregariousness, assertiveness, activity,
excitement-seeking, and positive emotions. Openness
facets are fantasy, aesthetics, feelings, actions, ideas, and values.
Agreeableness facets are trust, straightforwardness, altruism,
compliance, modesty, and tender-mindedness. Conscientiousness
facets are competence, order, dutifulness,
achievement-striving, self-discipline, and deliberation. The t
scores for the five domains and 30 facets were calculated
according to the method of Costa and McCrae, which uses
different reference means and standard deviations for men and
women. These distributions have a mean of 50 and standard
deviation of 10. T-scores ranging from 45 to 55 are considered
"average." Scores less than 45 are considered "low," and those
greater than 55 are considered "high."28
Prior Arrest
Arrest in Maryland between 1981 and 1994 was assessed
using the State of Maryland Criminal Justice System database,
a computerized file of all criminal records of individuals arrested,
charged, and sentenced in the state of Maryland. The
information includes the date of arrest, reporting court, citation
number and description of the crime, verdict, and times of
confinement, suspension, and probation. For this report, we
categorized type of arrest charge into two groups: "violent"
(assault, battery, murder, rape, or weapons violation) and "nonviolent"
(theft, burglary, possession of illicit substances, and
other crimes against property but not against persons). We
treated arrest over the follow-up period (1981 to 1994) as a
dichotomy (ever/never). A subject may have been arrested multiple
times, for both violent and nonviolent crimes; if any of the
arrests were for violent crimes, then this person was considered
to have had a violent arrest. Most subjects with a violent arrest
were also charged for nonviolent behaviors. If a subject was
arrested only for nonviolent behaviors during the follow-up
period, then he/she was considered to have a nonviolent-only
arrest.
Statistical Analysis
The proportions of subjects with prior arrest were compared
across demographic and axis I categories with the chi-square
test. NEO personality scores in subjects with and without a prior
arrest were compared using Studenťs t test; given the 35 separate
NEO dimensions, we considered P values less than .002 as
statistically significant. Logistic regression analysis was used to
assess the relationships between DSM-IV personality disorder
dimensions and prior arrest.29 Logistic regression also was used
to evaluate the relationships between NEO personality scores
and prior arrest, controlling for demographic characteristics,
alcohol or drug use disorders, and DSM-IV personality disorder
dimensions that were associated with prior arrest.
RESULTS
Correlates of Prior Arrest
Of the 611 subjects, 79 (12.9%) had a history of
arrest since 1981: 46 (7.5%) for a nonviolent-only
crime, and 33 (5.4%) for a violent crime. The
number of separate arrests ranged from 1 to 28
(mean, 2.9). As shown in Table 1, the proportion of
subjects with a prior arrest was inversely related to
current age, from 21% in 30- to 39-year-olds to 2%
in those age 70 years and older. A greater proportion
of men than women (19% v 9%), and a greater
proportion of "others" than Caucasians (13% v
276 SAMUELS ET AL
5%), had been arrested. Subjects with a lifetimeever
diagnosis of schizophrenia or mania/hypomania
were not more likely to have been arrested;
however, prior arrest was significantly greater in
individuals who had ever had alcohol use disorders,
and nearly 50% of subjects who had ever had
other psychoactive drug use disorders had been
arrested.
Personality Disorder Scores and Prior Arrest
Prior arrest was associated with scores on several
personality disorder dimensions (Table 2). The
odds of prior arrest increased with the score on the
adult antisocial dimension (odds ratio [OR]
1.68), i.e., a 68% increase in odds of arrest per unit
increase in the antisocial score. In addition, the
odds of arrest increased with scores on paranoid,
borderline, and narcissistic dimensions. In contrast,
prior arrest was inversely related to obsessivecompulsive
personality score.
These odds did not change appreciably after
controlling for age, sex, or race (data not shown).
Controlling for alcohol/drug use disorders did not
substantially change the odds ratios for antisocial
score (OR 1.35; 95% confidence interval, 1.2 to
1.5, P .001) or obsessive compulsive score
(OR 0.84 [0.7 to 0.99], P .04), but it did
substantially reduce the magnitudes of the association
for paranoid (OR 1.08 [0.9 to 1.2], P
0.30), borderline (OR 1.09 [0.96 to 1.2], P
.18), and narcissistic (OR 1.18 [0.9 to 1.5], P
0.19) scores.
NEO Personality Dimensions and Prior Arrest
As shown in Table 3, compared to subjects without
prior arrest, arrested subjects had significantly
higher mean scores on neuroticism and two neuroticism
facets, angry hostility and impulsiveness
Arrested subjects also scored higher on excitement-seeking.
In addition, arrested subjects scored
lower on agreeableness and several of its facets:
trust, straightforwardness, compliance, and modesty.
Arrested subjects also scored low on dutifulness
and deliberation. In the nonarrested subjects,
the mean NEO scores all fell within the range (45
to 55) considered "average." In contrast, in the
arrested subjects, mean scores on several of the
NEO facets were either higher (angry hostility) or
lower (warmth, trust, compliance, and dutifulness)
than the average range.
The odds of prior arrest increased from 4% to
7% per unit increase on neuroticism, angry hostility,
impulsiveness, and excitement-seeking dimensions,
and decreased 4% to 5% per unit increase on
agreeableness, trust, straightforwardness, compliance,
modesty, dutifulness, and deliberation diTable
2. Relationship between DSM-IV Personality
Disorders Dimensions in 1993 to 1999 and Prior Arrest
(1981 to 1993)
Odds Ratio (95% CI) P Value
Paranoid 1.24 (1.1-1.4) .001
Schizoid 1.07 (0.9-1.3) .42
Schizotypal 1.09 (0.98-1.2) .11
Antisocial 1.68 (1.5-1.9) .001
Borderline 1.39 (1.2-1.6) .001
Histrionic 1.09 (0.99-1.2) .08
Narcissistic 1.57 (1.2-2.0) .001
Avoidant 1.04 (0.8-1.3) .73
Dependent 0.87 (0.6-1.3) .45
Obsessive-compulsive 0.84 (0.7-0.98) .03
Abbreviation: CI, confidence interval.
Table 1. Proportion of Subjects With Prior Arrest (1981 to
1993), by Demographic Features and Lifetime Ever
Psychiatric Disorders in 1993 to 1999
Prior Arrest,
N (%)
Test
Statistic P Value
Age at interview (yr)
30-39 (n 194) 41 (21.1)
40-49 (n 217) 26 (12.0)
50-59 (n 86) 9 (10.5)
60-69 (n 55) 2 (3.6)
70 (n 59) 1 (1.7) 2
(trend)
21.0 .001
Sex
Female (n 381) 35 (9.2)
Male (n 230) 44 (19.1) 2
1 12.6 .001
Race
White (n 388) 19 (5.2)
Other (n 210) 27 (12.9) 2
1 10.8 .001
Schizophrenia
No (n 602) 78 (13.0)
Yes (n 9) 1 (11.1) 2
1 0.03 .87
Major depression
No (n 468) 56 (12.0)
Yes (n 138) 22 (15.9) 2
1 1.5 .22
Mania or hypomania
No (n 582) 74 (12.7)
Yes (n 20) 4 (20.0) 2
1 0.91 .34
Drug use disorders
No (n 404) 17 (4.2)
Alcohol only (n 99) 14 (14.1)
Other drugs (n 102) 48 (47.1) 2
2 131.8 .001
PERSONALITY DIMENSIONS AND CRIMINAL ARREST 277
mensions (Table 4). The magnitudes of these relationships
did not appreciably change, when age,
sex, race/ethnicity, alcohol/drug use disorders, or
paranoid, antisocial, borderline, narcissistic, or obsessive-compulsive
personality disorder scores
were controlled in logistic regression models. Conversely,
the magnitude of the relationships between
the personality disorder scores and arrest did not
markedly change when included in these models.
For almost all NEO scores, there were only
small differences between subjects with a nonviolent-only
or violent prior arrest (data not shown).
However, violent arrestees had lower mean scores
than nonviolent-only arrestees on gregariousness
(45.8 v 52.0, P .02) and openness to feelings
(42.2 v 49.2, P .004).
DISCUSSION
Major Findings
We found that several "normal" personality dimensions
were associated with prior arrest over the
preceding 13 years. Compared to those who had
not been arrested, arrested subjects had significantly
higher mean scores on NEO scales of angry
hostility, impulsiveness, and excitement-seeking,
and the odds of prior arrest increased between 4%
and 7% per unit increase on these scales. This
suggests that the likelihood of arrest increases, in a
dimensional way, with readiness to experience anger
(angry hostility), inability to control urges (impulsiveness),
and craving excitement and stimulation
(excitement-seeking).28 In contrast, arrested
subjects had significantly lower mean scores on
NEO scales of trust, straightforwardness, compliance,
modesty, dutifulness, and deliberation, and
the odds of prior arrest decreased 4% to 5% per
unit increase on these scales. This suggests that the
likelihood of arrest decreases dimensionally with
the disposition to believe that others are honest and
well-intentioned (trust), unwillingness to manipulate
others (straightforwardness), tendency to inhibit
aggression and to forgive others (compliance),
degree of humility and self-effacement
(modesty), extent to which one is governed by
moral obligations (dutifulness), and tendency to
think carefully before acting (deliberation).27
The association between scores on these NEO
scales and prior arrest were independent of demoTable
4. Relationship Between NEO Scores in 1993 to 1999
and Prior Arrest (1981 to 1993)
Odds Ratio
(95% CI) P Value
Neuroticism 1.04 (1.01-1.06) .002
Angry hostility 1.06 (1.03-1.08) .001
Impulsiveness 1.04 (1.02-1.07) .002
Warmth 0.97 (0.95-0.99) .006
Excitement-seeking 1.07 (1.05-1.10) .001
Agreeableness 0.95 (0.93-0.98) .001
Trust 0.96 (0.95-0.99) .001
Straightforwardness 0.95 (0.93-0.98) .001
Compliance 0.95 (0.93-0.97) .001
Modesty 0.96 (0.94-0.99) .002
Dutifulness 0.96 (0.93-0.98) .001
Deliberation 0.96 (0.93-0.98) .001
Table 3. NEO Personality Scores in 1993 to 1999, by Prior
Arrest (1981 to 1993)
Prior Arrest
Test
Statistic
(t609)
P
Value
No
(n 532)
Yes
(n 79)
Neuroticism 49.9 53.6 3.09 .002
Anxiety 49.1 50.9 1.55 .12
Angry hostility 50.4 55.7 4.66 .001
Depression 50.9 53.5 2.06 .04
Self-consciousness 49.8 51.6 1.66 .10
Impulsiveness 48.6 52.0 3.15 .002
Vulnerability 50.7 51.9 0.91 .36
Extraversion 47.5 48.3 0.67 .50
Warmth 47.5 44.1 2.81 .005
Gregariousness 50.1 49.5 0.54 .59
Assertiveness 48.7 49.2 0.41 .68
Activity 47.5 48.5 0.93 .35
Excitement-seeking 48.0 52.9 4.83 .001
Positive emotions 47.7 47.6 0.11 .91
Openness 45.8 46.0 0.14 .89
Fantasy 48.3 47.3 0.93 .35
Aesthetics 48.7 49.8 0.93 .35
Feelings 47.7 46.2 1.24 .22
Actions 45.9 47.5 1.44 .15
Ideas 47.1 49.0 1.68 .09
Values 46.0 43.7 2.07 .04
Agreeableness 49.2 43.8 4.32 .001
Trust 46.6 42.2 3.42 .001
Straightforwardness 49.5 44.8 4.04 .001
Altruism 49.1 46.7 1.89 .06
Compliance 49.4 43.8 4.51 .001
Modesty 50.8 46.9 3.19 .001
Tender-mindedness 51.8 52.1 0.28 .78
Conscientiousness 47.0 44.3 2.31 .02
Competence 48.2 45.1 2.45 .02
Order 46.3 46.5 0.27 .79
Dutifulness 47.0 42.3 3.92 .001
Achievement-striving 46.9 46.4 0.42 .67
Self-discipline 46.3 46.0 0.23 .82
Deliberation 52.0 48.2 3.48 .001
278 SAMUELS ET AL
graphic characteristics and alcohol or drug use
disorders, which also were related to prior arrest.
In addition, the relationships between these NEO
dimensions and prior arrest were independent of
DSM-IV personality disorder dimensions (paranoid,
antisocial, borderline, narcissistic, and obsessive-compulsive)
that we found to be associated
with arrest. Conversely, these personality disorder
dimensions remained associated with arrest, even
after controlling for the NEO dimensions; however,
only antisocial and obsessive-compulsive
personality disorder dimensions were significantly
related to arrest, after controlling for alcohol/drug
use disorders. This suggests that the relationship
between these NEO scales and arrest is not mediated
by alcohol/drug use disorders or personality
disorder dimensions, including antisocial personality
disorder. Thus, "normal" personality traits, as
measured by the NEO, may provide insight into
identifying individuals vulnerable to criminal arrest,
apart from that provided by DSM-IV criteria
for antisocial and other personality disorders.
Strengths and Limitations
The study has advantages not available to the
few previous studies of the relationship between
personality dimensions and arrest. The subjects
were participants in an epidemiological follow-up
survey in the community and were not selected for
treatment or arrest. They underwent thorough evaluation
by psychiatrists of DSM-IV symptoms and
personality disorder criteria. In addition, the determination
of their arrest history was based on the
state criminal justice records, not on self-report.
However, several potential limitations of the
study must be addressed. First, because the state
criminal justice database did not include arrests in
states other than Maryland, we may have underestimated
the magnitude of the relationship between
specific personality dimensions and arrest. Second,
the assessment of personality dimensions was
based on self-report and may have been misrepresented
by some subjects, either deliberately or because
they lacked insight; it would have been useful
to include information from the observer-report
version of the NEO (Form R).28 Third, given that
personality was assessed currently and arrest retrospectively,
we cannot conclude definitively that
NEO personality features predict arrest; longitudinal
studies of the relationship are needed to more
rigorously support the findings. However, there is
evidence of the long-term stability of NEO personality
features in individuals who are at least 30
years old.30 Fourth, it is important to note that
arrest is not equivalent to criminal activity; a person
engaging in a criminal act may not be arrested,
whereas a person may be arrested for a crime that
he/she did not commit. Sociodemographic and personality
characteristics could influence the likelihood
of arrest, independently of criminal behav-
ior.2
Implications
Despite these potential limitations, our findings
support the notion that specific "normal" personality
dimensions are associated with arrest in the
community. These relationships are independent of
demographic characteristics, alcohol/drug use disorders,
and DSM-IV axis II personality disorder
dimensions, including antisocial personality
scores. There is often an interest in clinical and
other settings to identify individuals who, in certain
circumstances, may be at increased risk of
criminal activities and arrest. Our findings suggest
that scores on NEO facets may provide additional
useful information for this task, beyond that provided
by evaluation of DSM-IV personality disorder
criteria. Moreover, in settings where trained
clinicians are not available for a complete evaluation
of axis II criteria, or time is limited, an assessment
of antisocial personality criteria, as well as
completion of the NEO, may be useful in evaluating
the proneness of individuals to criminal arrest.
Additional research is needed to clarify the interaction
of personality features with environmental
characteristics that promote or inhibit criminal involvement
and arrest.31 Understanding these relationships
is important for developing strategies to
reduce the likelihood of criminal activity and arrest
in individuals with extreme scores on specific personality
dimensions.
REFERENCES
1. Mercy JA, Krug EG, Dahlberg LL, Zwi AB. Violence
and health: the United States in a global perspective. Am J Publ
Health 2003;92:256-261.
2. Freudenberg N. Jails, prisons, and the health of urban
populations: a review of the impact of the correctional system
on community health. J Urban Health 2001;78:214-235.
PERSONALITY DIMENSIONS AND CRIMINAL ARREST 279
3. Dahlberg LL, Potter LB. Youth violence: developmental
pathways and prevention challenges. Am J Prev Med 2001;
20(Suppl 1):3-14.
4. Galea S, Karpati A, Kennedy B. Social capital and violence
in the United States, 1974-1993. Soc Sci Med 2002;55:
1373-1383.
5. Holcomb WR, Ahr PR. Arrest rates among young adult
psychiatric patients treated in inpatient and outpatient settings.
Hosp Commun Psychiatry 1988;39:52-57.
6. Modestin J, Ammann R. Mental disorders and criminal
behaviour. Br J Psychiatry 1995;166:667-675.
7. Hodgins S. Epidemiological investigations of the associations
between major mental disorders and crime: methodological
limitations and validity of the conclusions. Soc Psychiatry
Psychiatr Epidemiol 1998;33(Suppl 1):S29-37.
8. Hernandez-Avila CA, Burleson JA, Poling J, Tennen H,
Rounsaville BJ, Kranzler HR. Personality and substance use
disorders as predictors of criminality. Compr Psychiatry 2000;
41:276-283.
9. Bovasso GB, Alterman AI, Cacciola JS, Rutherford MJ.
The prediction of violent and nonviolent criminal behavior in a
methadone maintenance population. J Personal Disord 2002;16:
360-373.
10. Teplin LA, Abram KM, McClelland GM. Prevalence of
psychiatric disorders among incarcerated women. I. Pretrial jail
detainees. Arch Gen Psychiatry 1996;53:505-512.
11. Wallace C, Mullen P, Burgess P, Palmer S, Ruschena D,
Browne C. Serious criminal offending and mental disorder.
Case linkage study. Br J Psychiatry 1998;172:477-484.
12. Eronen M, Hakola P, Tiihonen J. Mental disorders and
homicidal behavior in Finland. Arch Gen Psychiatry 1996;53:
497-501.
13. Hodgins S, Cote G. Major mental disorder and antisocial
personality disorder: a criminal combination. Bull Am Acad
Psychiatry Law 1993;21:155-160.
14. Hodgins S. Mental disorder, intellectual deficiency, and
crime: evidence from a birth cohort. Arch Gen Psychiatry
1992;49:476-483.
15. Hodgins S, Mednick SA, Brennan PA, Schulsinger F,
Engberg M. Mental disorder and crime: evidence from a Danish
birth cohort. Arch Gen Psychiatry 1996;53:489-496.
16. Tiihonen J, Isohanni M, Ra¨sa¨nen P, Koiramen M, Moring
J. Specific major mental disorders and criminality: a 26-year
prospective study of the 1966 northern Finland birth cohort.
Am J Psychiatry 1997;154:840-845.
17. Arseneault L, Moffitt TE, Caspi A, Taylor PJ. Mental
disorders and violence in a total birth cohort. Arch Gen Psychiatry
2000;57:979-986.
18. McMillen DL, Adams MS, Wells-Parker E, Pang MG,
Anderson BJ. Personality traits and behaviors of alcohol-impaired
drivers: a comparison of first and multiple offenders.
Addict Behav 1992;17:407-414.
19. Ullrich S, Borkenau P, Marneros A. Personality disorders
in offenders: categorical versus dimensional approaches. J Personal
Disord 2001;15:442-449.
20. Johnson JG, Cohen P, Smailes E, Kasen S, Oldham JM,
Skodol AE, et al. Adolescent personality disorders associated
with violence and criminal behavior during adolescence and
early adulthood. Am J Psychiatry 2000;157:1406-1412.
21. Costa PT Jr, Widiger TA (eds). Personality Disorders and
the Five-Factor Model of Personality. Ed. 2. Washington, DC:
American Psychological Association, 2002.
22. Samuels J, Eaton WW, Bienvenu III OJ, Brown CH,
Costa PT Jr, Nestadt G. Prevalence and correlates of personality
disorders in a community sample. Br J Psychiatry 2002;536-
542.
23. Robins LN, Helzer JE, Croughan J. National Institute of
Mental Health Diagnostic Interview Schedule, Version III.
Rockville, MD: NIMH, 1981.
24. Eaton WW, Anthony JC, Romanoski A, Tien A, Gallo J,
Cai G, et al. Onset and recovery from panic disorder in the
Baltimore ECA follow-up. Br J Psychiatry 1998;173:501-507.
25. Nestadt G, Romanoski AJ, Samuels JF, Folstein MF,
McHugh PR. The relationship between personality and axis I
disorders in the population: results from an epidemiological
survey. Am J Psychiatry 1992;149:1228-1233.
26. American Psychiatric Association. Diagnostic and Statistical
Manual of Mental Disorders, Fourth Edition (DSM-IV).
Washington, DC: APA, 1994.
27. Samuels JF, Nestadt G, Romanoski AJ, Folstein MF,
McHugh PR. DSM-III personality disorders in the community.
Am J Psychiatry 1994;151:1055-1062.
28. Costa PT Jr, McCrae RR. Revised NEO Personality
Inventory (NEO PI-R) and NEO Five-Factor Inventory (NEOFFI)
Professional Manual. Odessa, FL: Psychological Assessment
Resources, 1992.
29. Hosmer DW Jr, Lemeshow S. Applied Logistic Regression.
New York, NY: Wiley, 1989.
30. Costa PT Jr, McCrae RR. Still stable after all these years:
personality as a key to some issues in adulthood and old age. In:
Baltes PB, Brim OG (eds). Life Span Development and Behavior.
Vol. 3. New York, NY: Academic Press, 1980:141-157.
31. Krug E, Mercy J, Dahlberg L, Zwi AB. The world report
on violence and health. Lancet 2002;360:1083-1088.
280 SAMUELS ET AL