Research Article LISA J.
COHEN, PhD
Attraction Versus Action in Pedophilic Desire: The Role
of Personality Traits and Childhood Experience
Objective: Comparison of pedophilic individuals
who do and do not refrain from sexually engag-
ing with children may offer critically important
information regarding the differential contrib-
utors to pedophilic attraction versus behavior.
This study compared 5 traits that are potentially
contributory to pedophilic attraction or behav-
ior in both minor-attracted persons (MAPs) who
refrain from sexually engaging with minors
(nonacting MAPs) and those who have acted on
Downloaded from http://journals.lww.com/practicalpsychiatry by BhDMf5ePHKbH4TTImqenVNlUEzghFiXbUN+Yoi5xc8u4us9Hpzz9SCP5Q8E4p/c1 on 11/06/2018
pedophilic attractions and subsequently entered
the criminal justice system (forensic MAPs).
Methods: Subjects included 195 nonacting MAPs,
50 forensic MAPs, and 60 healthy controls. Data
on nonacting MAPs were drawn from an online
survey, and data on the other 2 groups were
based on prior in-person evaluations. Measures
included the Millon Clinical Multiaxial Inven-
tory-II (MCMI-II), Barratt Impulsiveness Scale
(BIS-11), the MAP Questionnaire, and the Sexual
History Questionnaire (SHQ).
Results: Both MAP groups scored higher than
healthy controls on the domains of socially inhi-
bited personality traits, propensity toward cog-
nitive distortions, and subjects’ own childhood
sexual abuse (CSA). Forensic MAPs scored higher
than nonacting MAPs on the CSA domain, but
the 2 MAP groups differed little on the other 2
domains. Forensic MAPs also scored higher than
the other 2 groups on the antisocial domain,
whereas nonacting MAPs did not differ from
controls on this measure. Nonacting MAPs scored
higher than controls on impulsivity.
Conclusions: Antisocial personality traits may be
a primary driver of pedophilic behavior that is
unrelated to pedophilic attraction. Socially inhi-
bited personality traits and propensity toward
cognitive distortions are associated with pedo-
philic attraction, although the direction of cau-
sation is not clear. CSA seems to increase the risk
of both attraction and behavior.
(Journal of Psychiatric Practice 2018;24;374–387)
KEY WORDS: pedophilia, forensic, minor-attracted
persons, sexual abuse, sex offense
374 November 2018
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.
NWAYIEZE NDUKWE, MD, MPH
RACHEL SIEGFRIED, MA
IRINA KOPEYKINA
ZIMRI S. YASEEN, MD
IGOR GALYNKER, MD, PhD
A number of authors have previously divided risk fac-
tors for pedophilic behavior into motivational traits,
those that might contribute to the pedophilic attraction,
and disinhibiting traits, those that might facilitate
illegal behavior.1–5 We believe this distinction offers a
useful heuristic, as it can help identify those risk factors
that are most pertinent to public safety and therefore
of highest priority as therapeutic targets. With a few
exceptions, however (eg, Bailey et al1), the literature on
pedophilia has been restricted to subjects who have
engaged in illegal pedophilic behavior and subsequently
entered the criminal justice system. Such a selection
bias precludes effective differentiation of the correlates
of pedophilic attraction and behavior. Thus, a compar-
ison of individuals with pedophilia who have refrained
from sexually engaging with children and their forensic
counterparts may offer critically important information
regarding the differential contributors to pedophilic
attraction versus behavior.
Existing research on pedophilia, conducted largely
with forensic populations, has identified several traits
which we have posited may potentially contribute to
either motivation for or failure to inhibit pedophilic
behavior. These traits include a history of experiencing
childhood sexual abuse (CSA),6–8 social anxiety/low
self-esteem,6,9–11 high levels of antisocial traits,11,12 a
propensity toward cognitive distortions,13–15 and high
impulsivity.16,17 We have previously posited that the
first 2 traits contribute to motivation and the last 3 to
inhibitory failure.14 However, as data in past studies
have largely been drawn from forensic populations,
these findings may be confounded by traits related to
criminality, such as antisocial personality traits and
impulsivity, as well as the psychological sequelae of
COHEN, YASEEN, and GALYNKER: Icahn School of
Medicine at Mount Sinai, Mount Sinai Beth Israel,
New York, NY; NDUKWE, KOPEYKINA: Mount Sinai Beth
Israel, New York, NY; SIEGFRIED: Long Island University
Brooklyn, New York, NY
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.
Please send correspondence to: Lisa J. Cohen, PhD, 317 E.
17th St., 5F02, New York, NY 10003 (e-mail: lisaj.
cohen@mountsinai.org).
The authors declare no conflicts of interest.
DOI: 10.1097/PRA.0000000000000348
Journal of Psychiatric Practice Vol. 24, No. 6
 ATTRACTION VERSUS ACTION IN PEDOPHILIC DESIRE
incarceration or criminal prosecution.18,19 Therefore,
research into self-identified adults with pedophilic att-
raction who successfully refrain from sexually engaging
with children and who remain outside the criminal
justice system will remove the confound of criminality
and potentially shed greater light on the distinction
between pedophilic feelings and pedophilic behavior.
In this study, we use the term minor-attracted per-
sons (MAPs) to refer to individuals who experience
persistent attraction to prepubescent or pubescent
children. This term is preferred by individuals who
self-identify as attracted to underage minors because it
is perceived to carry less stigma than the term pedo-
phile, incorporates attraction to both prepubescent and
pubescent children, and does not imply behavior.20
This study compares MAPs who refrain from sexually
engaging with minors (nonacting MAPs) with 2 control
groups, MAPs who have acted on their pedophilic
attractions and were subsequently involved with the
criminal justice system (forensic MAPs) and healthy
controls, on the 5 traits discussed below.
POTENTIALLY MOTIVATING TRAITS
Socially Inhibited Personality Traits
MAPs have been hypothesized to seek sexual relations
with children in response to social anxiety, poor self-
confidence, or other personality traits that inhibit
appropriate sexual relations with adults. MAP subjects
have indeed been shown to have elevated rates of these
traits relative to healthy controls.6,14,21 However,
studies have also shown that these individuals report
feelings of shame, low self-esteem, and social avoidance
as a result of their pedophilic feelings and/or behavior,
making the direction of causality unclear.14
History of CSA
Research on adults with pedophilia has also shown
that this population has a significantly higher inci-
dence of sexual abuse in their own childhoods (CSA)
compared with the general population.6,22,23 While
there is an association between a history of CSA and
sex offending in general, the association between CSA
and sexual offenses against children specifically is
much stronger.8,10 Thus, CSA appears to play an
important role in the development of pedophilic
attraction and tendencies.10,22 It is not clear at this
Journal of Psychiatric Practice Vol. 24, No. 6
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.
point, however, whether CSA pertains to both attrac-
tion and behavior or to just one and not the other.
We have previously hypothesized that individuals’
own CSA can trigger pedophilic attraction by altering
sexual neurodevelopment, whereas acting on such
attraction depends upon additional disinhibiting
factors.9
POTENTIALLY DISINHIBITING TRAITS
Antisocial Personality Traits
Engaging in sexual activity with children violates
strongly held taboos as well as criminal law. Thus,
it stands to reason that participation in child sexual
abuse may be associated with elevated antisocial
traits, and many studies have supported this
finding.3,11,24–26 However, the degree of antisocial
traits appears to vary across subgroups.15,16 Hence,
it is possible that those individuals with pedophilic
attraction who succeed in refraining from sexual
activity with children may have significantly lower
antisocial traits, if any at all.
Propensity Toward Cognitive Distortions
Another finding with robust support in the literature is
the tendency in sex offenders against children toward
grossly distorted thinking.13,21,22 This literature, which
is largely drawn from forensic populations, has found
that MAPs commonly experience their sexual desires as
“ego-syntonic,” meaning they consider sexual attraction
to children to be acceptable and consistent with their
overall beliefs. In DSM-5, the Cluster A personality
disorders (schizoid, paranoid, and schizotypal person-
ality disorders) are characterized by a propensity
toward ideational distortions. Relatedly, several studies
have noted high levels of cluster A personality traits in
the forensic pedophilic population.15,21 It should be
noted, however, that these traits may be related to the
failure to inhibit pedophilic behavior or even to the
sequelae of incarceration,18,27 and thus the level of such
traits may be significantly lower in a nonforensic
population.
Impulsivity
Considerable data also point to high levels of
impulsivity or impulse control disorders in forensic
November 2018 375
 ATTRACTION VERSUS ACTION IN PEDOPHILIC DESIRE
samples of individuals with pedophilia,11,16 although
some studies have found that scores on impulsivity
measures in pedophilic groups were virtually indis-
tinguishable from those of healthy controls.3 Thus, the
data on impulsivity are not consistent and may reflect
comorbid psychopathology in those minor-attracted
adults with the most severe psychopathology and/or
those most likely to engage sexually with children and
subsequently get caught.
NONFORENSIC VERSUS FORENSIC MAPS
There is a limited body of literature comparing
these 5 traits across various samples of forensic and
nonforensic populations. Samples have included
acting and nonacting MAPs living in the
community,1,28 pornography offenders with and
without additional touching offenses,29 and acting
MAPs within or outside the criminal justice
system.30 Two studies have found that nonacting
MAPs reported a lower frequency of CSA than act-
ing MAPs living in the community,1,28 but another
study failed to find a difference between pornog-
raphy offenders with and without additional
touching offenses.31 Surprisingly, 1 study found
that noncontact MAP offenders (ie, those with
nontouching offenses) had higher antisocial scores
than did child molestation and child pornography
offenders.29 Although we could not find studies
specifically comparing acting and nonacting MAPs
on the propensity toward cognitive distortions, 1
study found that the 2 groups held comparably
positive attitudes toward adult-child sexual
activity.1 In addition, Webb et al32 did not find any
differences between Internet sex offenders and child
molesters on the Millon Clinical Multiaxial Inven-
tory-III (MCMI-III),33 which includes measures of
personality traits related to impulsivity, propensity
toward cognitive distortions, and socially inhibited
personality traits.
In sum, the extant data suggest a significant need
for additional research comparing forensic and
community-based, nonacting MAPs. In light of this,
we compared forensic MAPs, nonacting MAPs, and
healthy controls on the 5 related variables dis-
cussed above. Our overall hypothesis was that the 2
MAP groups would share potentially motivating
traits but would differ on potentially disinhibiting
traits. Our specific hypotheses were as follows:
376 November 2018
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.
H1: Both forensic MAPs and nonacting MAPS
would score higher than healthy controls on the
putatively motivating traits of socially inhibited
personality and history of CSA but would not differ
significantly from one another on these variables.
H2: Forensic MAPs would score higher than both
nonacting MAPs and healthy controls on the puta-
tively disinhibiting traits of impulsivity, propensity
for cognitive distortions, and antisocial personality,
whereas nonacting MAPs would not differ from
controls on these variables.
METHODS
This study utilized a web-based survey to access
MAPs living in the community who have success-
fully refrained from acting on their pedophilic
desires. The study is the result of a collaboration
between researchers at Mount Sinai Beth Israel
and an organization named B4U-ACT, which rep-
resents a collaborative effort between MAPs who
commit to living within the law and mental health
professionals to promote communication and
understanding between the 2 groups. This organ-
ization provides a unique opportunity to examine a
subgroup of adults with pedophilic attraction that
has not been adequately studied. Data from this
sample were compared with data previously col-
lected from a forensic sample of individuals with
pedophilia treated in a clinic specializing in the
treatment of sex offenders. Data from a control
group of healthy controls, collected as part of the
earlier studies of forensic MAPs, were also utilized
in the study. Analysis of the earlier data set has
resulted in several prior publications.2,3,9,21,34
Subjects
Nonacting MAPs
Nonacting MAPs were recruited by the science
director of B4U-ACT, who contacted potential sub-
jects through the organization’s mailing list and
other means of contact with qualified candidates.
Potential subjects were subsequently provided with
the Internet website address at http://www.
surveymonkey.com where they could complete the
anonymous Internet survey. The survey link was
made available online for a period of 7 months, from
November 2014 to May 2015. After survey data
Journal of Psychiatric Practice Vol. 24, No. 6
 ATTRACTION VERSUS ACTION IN PEDOPHILIC DESIRE
collection was completed, the subjects’ responses
were reviewed for inclusion in the category of non-
acting MAPs. This classification was based on the
following criteria:
 Acknowledgment of persistent sexual attraction
to prepubescent and/or pubescent children on
≥ 1 of 8 queries on the survey instrument.
 Answering “No” to repeated questions on multi-
ple measures which directly ask about whether
subjects have ever engaged in sexual activity
with underage children.
 Consistently answering in a negative manner to
questions about legal history with regard to
sexual offenses against minors.
 Answering in a consistently negative manner on
multiple questions that allowed for narrative
answers.
Only MAP subjects whose answers were fully con-
sistent with never having engaged in sexual activity
with an underage child when they were 18 or older or
at least 5 years older than the child were coded as
nonacting MAPs and included in the data analysis. Of
note, comorbid psychiatric disorders were not an
exclusion criterion in this subject group.
Forensic MAPs
As noted above, data from forensic MAPs were
drawn from previously collected data sets (collected
in 2 different studies conducted between 2000 and
2008). In total, 50 males with pedophilic behavior
were recruited from an outpatient clinic specializ-
ing in the treatment of sex offenders. All subjects
admitted to committing and were charged for or
convicted of a sexual offense against a child age 13
or younger when the subject was at least 18 years
old or at least 5 years older than the child. Data
were collected via interviews and self-report ques-
tionnaires administered in person.
Healthy Controls
Data on healthy controls (n = 60) were collected as
part of the study of forensic MAPs. Control subjects
were recruited via advertising in New York City.
Exclusion criteria for the controls included history
of pedophilia or of sexual activity as an adult with
anyone 15 years of age or younger, or with anyone
at least 5 years younger than the subject when the
subject was younger than 18. Exclusion criteria for
Journal of Psychiatric Practice Vol. 24, No. 6
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.
both the healthy control group and the forensic
MAP group included meeting DSM-IV criteria for a
substance-use disorder within 6 months before the
study as well as any significant Axis I psychiatric
disorder other than pedophilia, such as major
depressive disorder with psychosis, recurrent psy-
chotic disorders, bipolar disorder, or any disorder
requiring hospitalization or disability leave. Data
were collected via interviews and self-report ques-
tionnaires administered in person.
Assessment Measures
All measures listed below were administered in the
B4U-ACT survey study and at least 1 of the earlier
studies.
Background Information
Basic demographic information (age, sex, gender,
education) and legal and clinical history (past
experiences of psychotherapy and psychiatric med-
ication) were obtained for all subjects using demo-
graphic and background questionnaires developed
for each study.
Sexual History Questionnaire (SHQ)
The SHQ, a detailed 72-item self-report ques-
tionnaire assessing childhood sexual experiences,
sexual interest, and pedophilic behavior, was pre-
viously developed by the authors of this study.34,35
The SHQ was used to assess the subjects’ own his-
tory of CSA. To increase the reliability of subjects’
reported sexual abuse history, childhood sexual
history was assessed with multiple questions. Four
measures of CSA were calculated: whether the
subject reported experiencing adult advances as a
child, whether the subject was age 13 or younger at
first sexual contact, whether the subject was age 13
or younger with a partner 5 years older at first
sexual contact, and the age difference between the
subject and the first sexual partner. To test the
construct validity of the SHQ items, each item was
correlated with the Child Trauma Questionnaire
Sexual Abuse scale36 for all subjects in the B4U-
ACT study (N = 291 to 373). The Spearman ρ cor-
relations ranged from 0.379 to 0.697 (P < 0.001).
The SHQ was administered to all subjects included
in the analyses presented here; however, there were
some missing data as detailed below.
November 2018 377
 ATTRACTION VERSUS ACTION IN PEDOPHILIC DESIRE
MCMI-II
The MCMI-II is a 175-item questionnaire that meas-
ures DSM-III-R personality disorders as well as
several Axis I disorders and syndromes.37 Dimensional
scales from the MCMI-II and a later edition, the
MCMI-III, have been well correlated (r > 0.5) with
related measures from the Symptom Checklist
90-Revised (SCL-90-R)38 and the Minnesota Multi-
phasic Personality Inventory (MMPI).39 Scale scores
were calculated according to the scoring key in the
MCMI-II manual.37 No corrections were used in scor-
ing. The avoidant, dependent, and self-defeating per-
sonality disorder plus anxiety and dysthymic disorder
scales were used to assess socially inhibiting traits and
the schizotypal, schizoid, and paranoid personality
disorder plus thought and delusional disorder scales
were used to assess the propensity for cognitive
distortions. The antisocial and aggressive-sadistic
personality disorders scales were used to measure
antisocial personality traits. The MCMI-II Desirability
scale was used to assess the likelihood that subjects
would portray themselves in an unrealistically positive
manner, a significant concern when studying subjects
with pedophilic tendencies. The MCMI-II was admin-
istered to the B4U-ACT sample and in the first of the 2
earlier studies (N = 20 forensic MAPs and 24 healthy
controls).21
Barratt Impulsiveness Scale (BIS-11)
The Barratt Impulsiveness Scale (BIS-11)40 is a 34-item
questionnaire that was developed to measure impul-
sivity as a stable personality trait, distinct from anxiety
or aggressiveness. It is a Likert-type self-report ques-
tionnaire that assesses 3 components of impulsivity:
motor impulsiveness (acting without thinking); atten-
tional impulsiveness (not focusing on the task at hand);
and nonplanning impulsiveness (not planning or
thinking ahead carefully). The revised version, the BIS-
11, has good internal consistency and reliability.40 The
total score is the sum of all items. The BIS-11 was
administered to the B4U-ACT sample (N = 194 with 1
subject missing data) and in the second of the 2 earlier
studies (N = 26 forensic MAPs and 29 healthy
controls).2
MAP Questionnaire
The MAP Questionnaire was developed by the first
author (L.J.C.) in collaboration with B4U-ACT to
investigate subjects’ personal experience of being an
MAP, including their history of attraction to and
378 November 2018
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.
sexual behavior with minors, ease or difficulty of
inhibiting their attraction to children, attitudes toward
and experiences of mental health care, and experiences
of MAP-related stigma. This questionnaire was
administered only to the nonacting MAPs in the B4U-
ACT study.
Missing Data
Because the B4U-ACT study utilized an anonymous
online survey, in which research staff could not review
responses during data collection, missing data
presented more of a problem than with the earlier face-
to-face studies. To preserve sample size and avoid
needless loss of data, particularly in multivariable
analyses, imputed values for missing data from the
B4U-ACT subjects were calculated when at least 75%
of the data were present. For the BIS scale, each sub-
ject’s missing items were replaced with the available
mean score for that subject. An average of 1.6 items
was imputed across 29 subjects; 1 subject was excluded
for missing 33% of the items. For demographic data
and the 13 MCMI scales, group means were entered for
missing values. An average of 14 (9.3%) subjects’ scores
were imputed across the 13 MCMI scales. To preserve
sample size in multivariable analyses, missing values
for demographic covariates were imputed for all 3
subject groups, involving 9 subjects for age, 4 for sex,
and 3 for education. For the SHQ variables, missing
values had less impact on multivariable analyses and
therefore were not imputed.
Informed Consent
We obtained a waiver of consent and a waiver of
research authorization from the Mount Sinai Beth
Israel IRB for gathering the survey data. With regard
to the earlier studies, all healthy control and forensic
MAP subjects signed informed consent forms approved
by the Beth Israel Medical Center IRB and received
$75 reimbursement for their time. All data collection
occurred within IRB-approved research protocols.
Additional Confidentiality Measures for the
Survey Study
Because of the sensitivity of the information col-
lected in the B4U-ACT survey study, special pre-
cautions were taken to protect subjects’ confidentiality.
All information filled out by participants was fully
anonymous. Except for basic demographic information
Journal of Psychiatric Practice Vol. 24, No. 6
 ATTRACTION VERSUS ACTION IN PEDOPHILIC DESIRE
(sex, age in 5-year ranges, education level, ethnicity),
we did not collect any identifying information during
the course of this study. Although Survey Monkey
was able to identify the IP address of the survey
respondent, we followed Survey Monkey’s guidelines
to ensure that neither email addresses nor IP
addresses were linked to the surveys nor saved in
the survey data. Moreover, Survey Monkey was not
able to connect the IP addresses to the survey con-
tent. We also encrypted the survey results and
obtained a Confidentiality Certificate from the fed-
eral government, which protects the confidentiality
of the data even from subpoena. A federal Certificate
of Confidentiality was also obtained for the second of
the earlier studies.
Because the federal government recognizes the
necessity of anonymity and/or confidentiality in
research into potentially illegal behavior, reporting
requirements were waived for this study. Likewise,
as described above, the anonymity precluded iden-
tification of any individual subject.
Statistical Analyses
Descriptive analyses, using means and frequencies/
percentages, were used to characterize the demographic
and clinical features of this sample as a whole.
All demographic variables that significantly differed
across groups were entered as covariates in the pri-
mary analyses. Multivariate general linear models
with simple contrasts were used to compare MCMI-II
data across the 3 study groups. Three models were
tested, one each for the socially inhibited, antisocial,
and cognitively distorting domains. MCMI-II and BIS-
11 data were analyzed separately as they were
administered to different subjects in the forensic MAP
and healthy control groups. The BIS-11 total score was
compared across groups in a univariate general linear
model. The SHQ data included both categorical and
continuous variables. Each of 3 categorical SHQ vari-
ables was compared across groups in 3 multivariate
logistic regression models (forensic MAPs vs. nonacting
MAPs, forensic MAPs vs. healthy controls, and non-
acting MAPs vs. healthy controls). These predictors
were analyzed separately due to multicollinearity.
Continuous SHQ variables were assessed in a single
multivariate general linear model. Bonferroni correc-
tions were applied with alpha levels adjusted for the
number of comparisons in each model.
Journal of Psychiatric Practice Vol. 24, No. 6
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.
RESULTS
Subjects
The sample was comprised of 305 subjects, 195
nonacting MAPs from the B4U-ACT sample, 50
forensic MAPs, and 60 healthy controls. The non-
acting MAPs represented 25% out of a total of 780
B4U-ACT survey respondents and 61% of the 315
study completers as defined by Survey Monkey (ie,
those who answered the last question on the sur-
vey). Four (2.1%) of the nonacting MAPs in the
current sample were study noncompleters.
Assessment of Potential Demographic
Confounds
Demographic characteristics are presented in Table 1.
Years of education significantly differed across groups
(F2,304 = 16.9, P < 0.001) with each group significantly
differing from each other in Scheffe pairwise com-
parisons. The healthy controls group scored higher
(15.28 ± 1.8) than the nonacting MAPs (14.09 ± 2.5)
who scored higher than the forensic MAPs
(12.43 ± 3.4). Age and sex also significantly differed
across groups. By pairwise comparison, forensic MAP
subjects, at ∼40 years of age, were significantly older
than nonacting MAPs at 31 years of age. Approx-
imately 16% of the healthy controls were either
female or transgender (4 with missing data), com-
pared with 8% of the nonacting MAPs and 0% of the
forensic MAPs. Therefore, age, sex, and education
were entered as covariates in all later analyses.
Assessment of Self-Report Bias
To test whether nonacting subjects with pedophilia
were more likely to respond in socially desirable
ways than the other 2 subject groups, the MCMI-II
Desirability scale was compared across groups by
univariate general linear model. The univariate
comparison was statistically significant (F2,239 = 28.2,
P < 0.001). Surprisingly, by Scheffe pairwise compar-
isons, the nonacting MAPs had significantly lower
MCMI-II Desirability scores (9.71 ± 3.3) than either the
healthy controls (13.42 ± 3.8) or the forensic MAPs
(14.80 ± 5.1). The latter 2 groups did not differ from
each other. As this finding supports the validity of the
nonacting MAP subjects’ self-report of never having
acted on their pedophilic inclinations, the MCMI-II
November 2018 379
 ATTRACTION VERSUS ACTION IN PEDOPHILIC DESIRE

TABLE 1. Demographic Characteristics and Social Desirability Scores

FM (N = 50) NM (N = 195) HC (N = 60) Statistic Pairwise Comparisons

Male [N (%)]† 50 (100.0) 179 (91.8) 47 (83.9) χ2 = 9.0* FM > NM, HC

Mean ± SD
Age (y) 39.98 ± 11.8 31.12 ± 11.9 34.44 ± 10.9 F2,304 = 9.6** FM > NM
Years of education 12.43 ± 3.4 14.09 ± 2.5 15.28 ± 1.8 F2,304 = 16.9** HC > NM > FM
MCMI Desirability‡ 14.80 ± 5.1 9.71 ± 3.3 13.42 ± 3.8 F2,239 = 28.2** FM, HC > NM

†Imputed values for missing data are included in descriptive data for continuous variables but not for categorical variables.
‡For MCMI variables, N = 20 for FM and N = 24 for HC.
FM indicates forensic MAP; HC, healthy controls; MAP, minor-attracted person; MCMI, Millon Clinical Multiaxial Inventory;
NM, nonacting MAPs.
*Pr0.05.
**Pr0.001.

Desirability scale was not entered as a covariate in the
primary analyses. Nonetheless, the effect of social
desirability was assessed in secondary analyses.
Comparison of Potentially Motivating
and Disinhibiting Traits
Table 2 presents the comparison of the potentially
motivating factor of socially inhibited personality
traits as assessed by the MCMI-II scales of avoi-
dant, dependent, and self-defeating personality and
anxiety and dysthymic disorder scales. There was a
significant difference across groups by the omnibus
test as well as by each univariate F test except for
dependent personality disorder. By simple con-
trasts, nonacting MAPs scored higher than healthy
controls on all scales. Forensic MAPs scored sig-
nificantly higher than healthy controls on all scales
but dependent personality disorder. The 2 MAP
groups did not differ from each other.
Table 2 also presents group comparisons on the
potentially disinhibiting factors of propensity for
cognitive distortions, antisocial traits, and impul-
sivity. On the omnibus test, there was a significant
difference across groups for propensity for cognitive
distortions. The groups also differed on all uni-
variate F tests. However, the pattern of pairwise
differences varied across the individual scales. Both
MAP groups scored higher than controls on schizo-
typal and thought disorder scales; the forensic
MAPs scored higher than the healthy controls on
paranoid and delusional scales; and the nonacting
MAPs scored higher than the other 2 groups on the
schizoid personality scale.
There were significant differences between
groups on antisocial traits, both by omnibus test
and each univariate test. On both scales in this
domain, however, the forensic MAPs scored higher
than the nonacting MAPs and also scored higher
than the healthy controls on the MCMI-II antisocial
scale. Importantly, the nonacting MAPs did not
differ from the healthy controls on either antisocial
subscale, with virtually identical scores on the
MCMI-II antisocial scale.
There was also a significant difference across
groups on the BIS-11 impulsivity scale. Unexpect-
edly, pairwise comparisons showed that the non-
acting but not forensic MAPs scored significantly
higher than controls on impulsivity.
Because educational attainment may reflect the
effect of personality pathology as well as an influ-
ence on it, we repeated these analyses removing
education as a covariate. Our findings remained the
same except that forensic MAPs now scored sig-
nificantly higher than nonacting MAPs on paranoid
and delusional scales and MAP groups no longer
differed on avoidant scales.
Childhood Sexual History
The 3 groups differed significantly with regard to all
sexual history variables measured on the SHQ.
There was a significant omnibus difference as well
as significant univariate F tests with regard to age

380 November 2018 Journal of Psychiatric Practice Vol. 24, No. 6

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.

 ATTRACTION VERSUS ACTION IN PEDOPHILIC DESIRE

TABLE 2. Comparison of Potentially Motivating or Disinhibiting Traits Across Subject Groups

Simple
FM NM HC Statistic Contrasts

MCMI-II potentially motivating traits

Socially inhibited traits Wilk’s λ (10,458) = 0.825, P < 0.001
(mean ± SD)†
Avoidant PD 22.30 ± 16.6 29.78 ± 15.0 10.38 ± 6.9 F5,238 = 12.6** NM > FM > HC
Dependent PD 27.05 ± 6.7 29.66 ± 8.8 25.33 ± 5.9 F5,238 = 2.0# NM > HC
Self-defeating PD 21.25 ± 16.4 21.76 ± 12.9 8.92 ± 8.04 F5,238 = 9.7** FM, NM > HC
Anxiety 11.40 ± 12.0 12.20 ± 10.3 3.25 ± 3.9 F5,238 = 7.1** FM, NM > HC
Dysthymia 19.10 ± 19.3 24.69 ± 17.5 5.08 ± 5.1 F5,238 = 11.0** FM, NM > HC
MCMI-II potentially disinhibiting traits
Propensity to cognitive Wilk’s λ (10,458) = 0.789, P < 0.001
distortions (mean ± SD)†
Paranoid PD 30.90 ± 16.2 25.98 ± 11.3 21.50 ± 12.9 F5,238 = 3.6* FM > HC
Schizoid PD 20.55 ± 8.1 29.70 ± 10.4 16.42 ± 5.1 F5,238 = 10.2** NM > FM, HC
Schizotypal PD 20.90 ± 17.7 27.22 ± 14.3 8.79 ± 6.3 F5,238 = 12.3** FM, NM > HC
Thought disorder 14.25 ± 12.7 15.76 ± 10.1 5.29 ± 5.9 F5,238 = 10.6** FM, NM > HC
Delusional disorder 15.20 ± 9.6 12.83 ± 6.6 9.50 ± 5.8 F5,238 = 5.3** FM > HC
Antisocial traits Wilk’s λ (4464) = 0.924, P = 0.001
(mean ± SD)‡
Antisocial PD 34.75 ± 13.9 22.78 ± 11.4 22.91 ± 12.3 F5,238 = 6.9** FM > NM, HC
Aggressive-sadistic PD 33.00 ± 12.7 23.62 ± 10.3 26.04 ± 12.2 F5,238 = 4.8** FM > NM
BIS-11 impulsivity (mean ± SD)
BIS-11 total 58.38 ± 12.7 63.07 ± 11.5 57.34 ± 10.1 F5,249 = 8.05** NM > HC

Covaried for age, sex, and years of education.

†Bonferroni-corrected alpha = 0.01.
‡Bonferroni-corrected alpha = 0.025.
BIS indicates Barratt Impulsiveness Scale; FM, forensic MAP; HC, healthy controls; MAP, minor-attracted person; MCMI,
Millon Clinical Multiaxial Inventory; NM, nonacting MAP; PD, personality disorder.
#P ≤ 0.1.
*P ≤ 0.01.
**P ≤ 0.001.

at first sexual contact and age difference between
subject and first partner. Note that, on these vari-
ables, 45 nonacting MAPs who reported no prior
sexual contact at all were excluded from the anal-
ysis. Simple contrasts showed that both MAP
groups were younger at first sexual contact than
healthy controls, although forensic MAPs had a
larger age difference between the subject and first
partner than either of the other 2 groups (Table 3).
Data on categorical variables are presented in
Table 4. By multivariate logistic regression, for-
ensic MAPs compared with healthy controls were
more likely to report adult sexual advances during
childhood [adjusted odds ratio (AOR) = 15.11; 95%
confidence interval (CI), 3.0-75.7], first sexual
encounter at age 13 or younger (AOR = 6.47; 95%
CI, 2.2-18.8), and first sexual encounter at or
below age 13 and partner over 5 years older
(AOR = 13.31; 95% CI, 2.6-67.5). Nonacting MAPs
were also more likely than healthy controls to
report adult sexual advances (AOR = 5.80; 95%
CI, 1.6-20.5) and first sexual contact at or below
age 13 (AOR = 2.93; 95% CI, 1.3-6.6). In addition,
forensic MAPs were more likely than nonacting
MAPs to report adult sexual advances (AOR =
3.90; 95% CI, 1.9-8.2), to have first sexual contact
at age 13 or below (AOR = 2.43; 95% CI, 1.1-5.2)
and to have first sex at or below age 13 with a
partner at least 5 years older (AOR = 4.80; 95%
CI, 2.1-10.8).

Journal of Psychiatric Practice Vol. 24, No. 6 November 2018 381

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.

 ATTRACTION VERSUS ACTION IN PEDOPHILIC DESIRE

TABLE 3. Comparison of Age of First Sexual Encounter Across Subject Groups

FM NM HC Simple
(N = 47) (N = 129)* (N = 55) Statistic Contrasts

Potentially motivating traits

Subjects’ own childhood history of sexual abuse†
Age of self and first partner Wilk’s λ (4396) = 0.916, P = 0.002
at first contact (mean ± SD)
Age at first sexual 11.74 ± 5.3 13.98 ± 7.3 16.53 ± 2.5 F5,233 = 5.5 HC > FM,
encounter P < 0.001 HC > NM
Age difference subject and 11.41 ± 11.5 3.02 ± 8.4 2.33 ± 8.1 F5,224 = 7.2 FM > HC,
first sexual partner P < 0.001 FM > NM

Covaried for age, sex, and years of education.

*NM subject who reported never having sex (N = 45) were excluded from this analysis; additional variation in sample size due to
missing data.
†Bonferroni-corrected alpha = 0.025.
FM indicates forensic MAP; HC, healthy controls; MAP, minor-attracted person; NM, nonacting MAP.

Adjustment for Type 1 Error Adjustment for Social Desirability Bias

The Bonferroni adjusted alpha level is presented in a
footnote for each table. As is evident in the tables, the
vast majority of the P-values presented fall well
below the adjusted alpha levels, suggesting adequate
protection from type 1 error.
When the primary analyses were repeated with
MCMI-II Desirability score entered as a covariate,
results remained essentially unchanged with the
exception of the 2 MAP groups no longer differing
on aggressive-sadistic or avoidant and nonacting

TABLE 4. Logistic Regression Analyses: Early Sexual History Across Subject Groups

AOR: AOR: AOR:

FM NM HC FM vs. NM NM vs. HC FM vs. HC

Potentially motivating traits [n (%)]

Subjects’ own history of childhood sexual abuse†
First sexual encounter rage 13 and Z 5 y older
Yes 23 (52.3) 23 (13.2) 3 (5.5) 4.80*** 2.65 13.31***
No 21 (47.7) 151 (86.8) 52 (94.5) — — —
First sexual encounter age 13 or younger
Yes 32 (68.1) 64 (36.8) 9 (16.4) 2.43* 2.93** 6.47**
No 15 (31.9) 110 (63.2) 46 (83.6) — — —
Adult sexual advances < age 18
Yes 27 (56.3) 40 (20.6) 3 (5.5) 3.90*** 5.80** 15.11***
No 21 (43.8) 154 (79.4) 52 (94.5) — — —

Three groups of multivariate logistic regression models were tested: FM versus NM, NM versus HC, and FM versus HC. All
models covaried for age, sex, and years of education.
†Bonferroni-corrected alpha = 0.017.
AOR indicates adjusted odds ratio; FM, forensic MAP; HC, healthy controls; MAP, minor-attracted person; NM, nonacting MAP.
*P ≤ 0.05.
**P ≤ 0.01.
***P ≤ 0.001.

382 November 2018 Journal of Psychiatric Practice Vol. 24, No. 6

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.

 ATTRACTION VERSUS ACTION IN PEDOPHILIC DESIRE
MAPs now scoring higher than controls on paranoid
and delusional.
DISCUSSION
The purpose of this study was to compare a group of
community-based MAPs who have refrained from
sexually engaging with children with a group of
MAPs recruited via the criminal justice system as
well as a sample of healthy controls on 5 traits
putatively associated with either the motivation for
or failure to inhibit pedophilic behavior.2,3 We
hypothesized that both MAP groups would exhibit
higher rates of the potentially motivating features
of socially inhibited personality traits and CSA
histories compared with healthy controls, but that
the 2 MAP groups would not differ from one another
on these traits. We also hypothesized that forensic
MAPs would show higher rates of the potentially
disinhibiting traits of propensity toward cognitive
distortions, impulsivity, and antisocial features
compared with both the healthy controls and the
nonacting MAPs. The results offered partial sup-
port for these hypotheses.
In our sample, both MAP groups scored higher than
healthy controls on socially inhibited personality traits,
specifically avoidant and self-defeating personality
disorder traits as well as anxiety and dysthymia
symptoms. This profile is consistent with our research
hypothesis and earlier research.11,14,24,32,41,42 The 2
MAP groups only differed from one another on avoi-
dant personality disorder traits. Although we could not
find other studies that precisely replicated this com-
parison, this finding is consistent with that of a similar
study comparing Internet child pornography offenders
with and without additional touch offenses which
found no significant difference between the 2 groups
using the MCMI-III.32
As such, our results suggest that socially inhibited
traits may be a correlate of pedophilic attraction rather
than pedophilic behaviors per se. Importantly, it is not
clear from these data whether such traits are a cause
or an effect of the attraction to minors. It has pre-
viously been hypothesized that people who feel
uncomfortable relating to adults might find it easier to
relate sexually to children.41,42 However, MAPs also
report intense experiences of stigma,43,44 which is
strongly associated with increased social insecurity
and inhibition.45,46 Thus, socially inhibited personality
traits may predispose MAPs to develop sexual interest
Journal of Psychiatric Practice Vol. 24, No. 6
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.
in children rather than adults but may also develop in
response to MAPs’ pedophilic preferences regardless of
their actual behavior, as a function of the profound
stigma against pedophilia in the wider society.20,47,48
Importantly, our own data suggest that the age of
onset of pedophilic attraction is in early adolescence
(13 ± 3.8 y).49 Thus, it is possible that pedophilic ten-
dencies could precede personality consolidation.
In addition, both MAP groups scored higher than
healthy controls on a number of variables related to
CSA history. Such findings are consistent with a
large body of prior research, adding support to the
contributory role of childhood sexual experience in
the development of pedophilic attraction.6–8,22
Contrary to our hypotheses, forensic MAPs were
more likely to be 13 or younger at first sexual contact
than MAP nonactors, had a greater age difference
with first partner at first sexual contact, and were
more likely to report sexual advances from adults in
childhood. These findings are consistent with other
studies that have compared MAPs who have and have
not sexually engaged with children,1,30,49 although
they are inconsistent with other studies.31 This sug-
gests that CSA plays a contributory role in both
pedophilic behavior and attraction. It is possible that
individuals with more severe or frequent incidents of
premature sexual exposure in childhood may be at
greater risk for failure to inhibit pedophilic behavior
in adulthood.
Given the elevated level of antisocial traits in
forensic versus nonacting MAPs, it is also possible
that antisocial traits mediate the relationship
between premature sexual exposure and failure to
inhibit pedophilic behavior. As our group has
speculated in a prior report,9 sexual abuse survivors
who develop pedophilic attraction in the absence of
antisocial traits may be more motivated to refrain
from sexual engagement with children. Fur-
thermore, it is notable that a significant number of
subjects in both MAP groups denied any history of
CSA, suggesting that premature sexual exposure is
not the only pathway to pedophilic attraction.
With regard to the potentially disinhibiting traits,
we found both MAP groups scored higher than healthy
controls on propensity toward cognitive distortions.
More specifically, both MAP groups had elevated
schizotypal traits and thought disorder symptoms and
at least 1 MAP group scored higher than controls on
each of the 3 other traits in this domain. These findings
are consistent with past research,13–15 and support the
November 2018 383
 ATTRACTION VERSUS ACTION IN PEDOPHILIC DESIRE
notion that tendencies toward unconventional, inac-
curate, and illogical thinking are common among
individuals with pedophilia.13,21,22,50 The fact that
these tendencies are elevated even in MAPs who have
not engaged in illegal sexual behavior with children is
notable and may well relate to the above-mentioned
stigma against pedophilia, so that MAPs may need to
downplay to themselves the potential harm of acting
on their sexual attractions in an effort to protect their
own self-esteem. Alternatively, as schizotypy has been
shown to have high genetic loading,51 it is possible that
comorbid pedophilia and schizotypal traits may reflect
common genetic influences.
In terms of antisocial traits, no significant difference
was found between MAP nonactors and healthy con-
trols. Forensic MAPs, however, were found to score
significantly higher than both healthy controls and
nonactors on antisocial traits and higher than non-
acting MAPs on aggressive-sadistic traits (although
this latter result did not survive addition of MCMI-II
Desirability as a covariate). This finding regarding
forensic MAPs is consistent with the frequent finding
of elevated antisocial traits in forensic populations in
general12,52 and in sexual offenders against children in
particular.11 That MAP nonactors in this study did not
exhibit elevated antisocial features is an important
finding and suggests that such features may be critical
contributors to the failure to inhibit illegal pedophilic
behavior but may be unrelated to pedophilic attraction
in itself. Moreover, as forensic MAPs did not show
evidence of generalized impulsivity, it is likely to be the
attitudinal vs. the behavioral component of antisocial
personality disorder (c.f. Hare53)—that is, the defi-
ciency of empathy or remorse rather than generalized
behavioral disinhibition—that serves to disinhibit
pedophilic behavior among adults with persistent
sexual attraction to minors.
Unexpectedly, nonacting MAPs scored higher
than either forensic MAPs or controls on impulsiv-
ity although forensic MAPs did not differ from
controls. This finding is consistent with some past
studies3,54 but not others16,17 and suggests that
impulsivity may only be a factor in a subgroup of
MAPs and is not necessary for acting upon pedo-
philic desire. It is possible that impulsivity in non-
acting MAPs is related to their higher anxiety and
dysthymia scores (r = 0.35 and 0.34, respectively),
reflecting a restless discomfort.
In sum, our findings partially support our
hypotheses and suggest that socially inhibited
384 November 2018
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.
personality traits and a propensity toward cognitive
distortions are associated with pedophilic attrac-
tion, regardless of history of sexually engaging with
children. It is unclear to what extent these person-
ality traits serve as a cause or an effect of pedophilic
attraction, however, as the impact of the intense
stigma against pedophilia should not be dis-
regarded. CSA appears to contribute to both pedo-
philic attraction and to decreased behavioral
inhibition with regard to that attraction. Finally,
antisocial traits appear to be the central risk factor
for behavioral disinhibition among the traits stud-
ied in this analysis but appear unrelated to pedo-
philic attraction.
Clinical Implications
The findings presented here have significant clinical
implications. All of the traits studied, with the
exception of impulsivity, are associated with either
pedophilic attraction or behavior and thereby warrant
clinical attention. In particular, our findings highlight
the importance of assessing and addressing antisocial
tendencies in treatment and underscore the dis-
tinction between attraction and behavior. Although
antisocial personality disorder and traits have tradi-
tionally been seen as treatment resistant,55 successful
results have been reported across multiple modalities.
Contingency management, utilizing rewards for
negative urine samples, has demonstrated efficacy in
the treatment of cocaine dependence in men with
antisocial personality disorder.56 Although external
structure may be important for patients with anti-
social personality disorder who lack internal motiva-
tion for change, treatments reliant upon intrinsic
motivation have also been reported to be effective in
patients with antisocial personality disorder traits.
Mentalization-based treatment of outpatients with
comorbid borderline and antisocial personality dis-
orders yielded reductions in anger, hostility, paranoia,
self-directed aggression, and suicide attempts, as well
as a reduction in psychiatric symptoms and an
improvement in mood and interpersonal and social
functioning.57 Likewise, following an integrative
treatment combining psychodynamic, cognitive
behavioral, and relaxation techniques, prison inmates
showed reduced recidivism and improved occupa-
tional and social functioning over a 5-year follow-up
period.58
Journal of Psychiatric Practice Vol. 24, No. 6
 ATTRACTION VERSUS ACTION IN PEDOPHILIC DESIRE
In addition, the well-documented relationship
between CSA and both adult pedophilic attraction and
behavior supports the need for prevention and early
intervention efforts with victims of sexual abuse.
A robust literature has reported the efficacy of multiple
forms of treatment, including cognitive therapy,59 group
therapy,60 mentalization-based therapy,61 and intensive
multimodal approaches,62 for the sequelae of CSA in
both children and adults.
Finally, the relationship of perceived stigma to the
traits studied here is a largely unexplored and poten-
tially valuable area of investigation for both clinicians
and researchers. To the extent that maladaptive
responses to perceived stigma impede healthy rela-
tionships, damage self-esteem, and distort cognition,
treatment effectiveness may be compromised. This
points to the profound need for mental health treatment
for MAPs that can (a) distinguish between pedophilic
attraction and behavior, (b) help MAPs negotiate the
profound difficulties of living with pedophilic attraction,
while at the same time (c) firmly condemning any sex-
ual contact between children and adults, and (d) work-
ing in conjunction with the legal system if necessary.
This study should be considered in the context of its
limitations. First, the samples were studied separately,
with a time gap between data collection, which may
have introduced undetected bias in data collection
methods or a possible cohort effect. Second, MCMI-II
and BIS-11 data were collected only from a subset of the
forensic and healthy control groups. Third, the inclusion
criteria for the B4U-ACT but not the forensic MAP
samples included attraction to pubescent as well as
prepubescent children. However, in our survey data, all
of the nonacting B4U-ACT subjects endorsed sexual
attraction either to “boys,” “girls,” or children aged 13 or
younger. This, in conjunction with subjects’ participat-
ing in research specifically described as addressing
“adults who are sexually attracted to children,” supports
our assumption that all MAP subjects experience per-
sistent pedophilic attraction. In addition, the analyses
excluded consideration of other personality disorders
that might be relevant in this population, such as bor-
derline and narcissistic personality disorders. Finally,
data on nonacting MAPs were gathered by Internet
survey, which poses multiple risks of self-report bias.
However, we attempted to circumvent such risks with
strong efforts to protect confidentiality, including a
federal guarantee of confidentiality, investigation of
response consistency across multiple instruments, and
collaboration with B4U-ACT, which allowed us access to
Journal of Psychiatric Practice Vol. 24, No. 6
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.
a large population of self-identified MAPs. Taken in the
context of these limitations, however, our study adds
valuable information to the sparse but growing liter-
ature on the difference between MAPs who do and do
not successfully refrain from sexually engaging with
children. Such information can shed critically important
light on the different risk factors for pedophilic attrac-
tion versus behavior to aid the development of effective
intervention and prevention efforts.
REFERENCES
1. Bailey JM, Hsu KJ, Bernhard PA. An Internet study of
men sexually attracted to children: sexual attraction
patterns. J Abnorm Psychol. 2016;125:976–988.
2. Cohen LJ, Galynker I. Identifying psychological traits
potentially subserving aberrant motivation or inhibitory
failure in pedophilic behavior. Isr J Psychiatry Relat Sci.
2012;49:281–290.
3. Cohen LJ, Grebchenko YF, Steinfeld M, et al. Comparison
of personality traits in pedophiles, abstinent opiate addicts,
and healthy controls: considering pedophilia as an addictive
behavior. J Nerv Ment Dis. 2008;196:829–837.
4. Seto MC. Pedophilia and Sexual Offending Against
Children: Theory, Assessment, and Intervention. Wash-
ington, DC: American Psychiatric Association; 2007.
5. Seto MC. Internet Sex Offenders. Washington, DC:
American Psychological Association; 2013.
6. Bagley C, Wood M, Young L. Victim to abuser: mental
health and behavioral sequels of child sexual abuse in a
community survey of young adult males. Child Abuse
Negl. 1994;18:683–697.
7. Jespersen AF, Lalumière ML, Seto MC. Sexual abuse
history among adult sex offenders and non-sex offenders:
a meta-analysis. Child Abuse Negl. 2009;33:179–192.
8. Seghorn TK, Prentky RA, Boucher RJ. Childhood sexual
abuse in the lives of sexually aggressive offenders. J Am
Acad Child Adolesc Psychiatry. 1987;26:262–267.
9. Cohen LJ, Nikiforov K, Gans S, et al. Heterosexual male
perpetrators of childhood sexual abuse: a preliminary
neuropsychiatric model. Psychiatr Q. 2002;73:313–336.
10. Freund K, Kuban M. The basis of the abused abuser
theory of pedophilia: a further elaboration on an earlier
study. Arch Sex Behav. 1994;23:553–563.
11. Raymond NC, Coleman E, Ohlerking F, et al. Psychiatric
comorbidity in pedophilic sex offenders. Am J Psychiatry.
1999;156:786–788.
12. Ahlmeyer S, Kleinsasser D, Stoner J, et al. Psychopa-
thology of incarcerated sex offenders. J Pers Disord. 2003;
17:306–318.
13. Blumenthal S, Gudjonsson G, Burns J. Cognitive dis-
tortions and blame attribution in sex offenders against
adults and children. Child Abuse Negl. 1999;23:129–143.
14. Cohen LJ, Galynker I. Psychopathology and personality
traits of pedophiles: issues for diagnosis and treatment.
Psychiatr Times. 2009;26:25–30. Available at: www.
psychiatrictimes.com/addiction/psychopathology-and-perso-
nality-traits-pedophiles. Accessed September 1, 2018.
15. Henderson MC, Kalichman SC. Sexually deviant
behavior and schizotypy: a theoretical perspective with
supportive data. Psychiatr Q. 1990;61:273–284.
November 2018 385
 ATTRACTION VERSUS ACTION IN PEDOPHILIC DESIRE
16. Galli V, McElroy SL, Soutullo CA, et al. The psychiatric
diagnoses of twenty-two adolescents who have sexually
molested other children. Compr Psychiatry. 1999;40:85–88.
17. Schiffer B, Peschel T, Paul T, et al. Structural brain
abnormalities in the frontostriatal system and cerebel-
lum in pedophilia. J Psychiatr Res. 2007;41:753–762.
18. Tomar S. The psychological effects of incarceration on
inmates: can we promote positive emotion in inmates?
Delhi Psychiatry J. 2013;16:66–72.
19. Jeglic EL, Spada A, Mercado CC. An examination of
suicide attempts among incarcerated sex offenders. Sex
Abuse. 2013;25:21–40.
20. Kramer R. The DSM and the stigmatization of people who
are attracted to minors. In: Pedophilia, minor-attracted
persons, and the DSM: issues and controversies. Sympo-
sium conducted at the meeting of the B4U-ACT Inc.,
Baltimore, MD, August 17, 2011. Available at: www.
b4uact.org/wp-content/uploads/2014/08/2011-Symposium-
Proceedings-Part-1.pdf. Accessed September 1, 2018.
21. Cohen LJ, McGeoch PG, Watras-Gans S, et al. Person-
ality impairment in male pedophiles. J Clin Psychiatry.
2002;63:912–919.
22. Haywood TW, Kravitz HM, Wasyliw OE, et al. Cycle of
abuse and psychopathology in cleric and noncleric
molesters of children and adolescents. Child Abuse Negl.
1996;20:1233–1243.
23. Seto MC, Laumiere M. What is so special about male
adolescent sexual offending? A review and test of
explanations through meta-analysis. Psychol Bull. 2010;
136:526–575.
24. Hall RC, Hall RC. A profile of pedophilia: definition,
characteristics of offenders, recidivism, treatment out-
comes, and forensic issues. Mayo Clin Proc. 2007;82:
457–471.
25. Kalichman SC. Psychopathology and personality charac-
teristics of criminal sexual offenders as a function of
victim age. Arch Sex Behav. 1991;20:187–197.
26. Kruger TH, Schiffer B. Neurocognitive and personality
factors in homo-and heterosexual pedophiles and con-
trols. J Sex Med. 2011;8:1650–1659.
27. Grassian S. Psychiatric effects of solitary confinement.
Wash U J Law Policy. 2006;22:325–383.
28. Schaefer GA, Mundt IA, Feelgood S, et al. Potential and
Dunkelfeld offenders: two neglected target groups for
prevention of child sexual abuse. Int J Law Psychiatry.
2010;33:154–163.
29. Jung S, Ennis L, Stein S, et al. Child pornography
possessors: comparisons and contrasts with contact-and
non-contact sex offenders. J Sex Aggress. 2013;19:
295–310.
30. Neutze J, Grundmann D, Scherner G, et al. Undetected
and detected child sexual abuse and child pornography
offenders. Int J Law Psychiatry. 2012;35:168–175.
31. McCarthy JA. Internet sexual activity: a comparison
between contact and non-contact child pornography
offenders. J Sex Aggress. 2010;16:181–195.
32. Webb L, Craissati J, Keen S. Characteristics of internet
child pornography offenders: a comparison with child
molesters. Sex Abuse. 2007;19:449–465.
33. Millon T. MCMI-III: Millon Clinical Multiaxial Inven-
tory, 4th ed. Minneapolis, MN: Pearson; 2009.
34. Cohen LJ, Forman H, Steinfeld M, et al. Comparison of
childhood sexual histories in subjects with pedophilia or
opiate addiction and healthy controls: is childhood sexual
abuse a risk factor for addictions? J Psychiatr Pract.
2010;16:394–404.
386 November 2018
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.
35. Cohen LJ, McGeoch PG, Watras Gans S, et al. Childhood
sexual history of 20 male pedophiles vs. 24 male healthy
control subjects. J Nerv Ment Dis. 2002;190:757–766.
36. Bernstein DP, Fink L, Handelsman L, et al. Initial reliability
and validity of a new retrospective measure of child abuse
and neglect. Am J Psychiatry. 1994;151:1132–1136.
37. Millon T, Wetzler S. Millon Clinical Multiaxial Inventory:
I & II. J Couns Dev. 1992;70:421–426.
38. Derogatis LR. Symptom Checklist-90 Revised. Pearson:
Clinical Psychology. 1994.
39. Drayton M. The Minnesota Multiphasic Personality Inven-
tory-2 (MMPI-2). Occup Med (Lond). 2009;59:135–136.
40. Patton JH, Stanford MS. Factor structure of the Barratt
Impulsiveness Scale. J Clin Psychol. 1995;51:768–774.
41. Cohen LJ, Galynker II. Clinical features of pedophilia
and implications for treatment. J Psychiatr Pract. 2002;
8:276–289.
42. Egan V, Kavanagh B, Blair M. Sexual offenders against
children: the influence of personality and obsessionality
on cognitive distortions. Sex Abuse. 2005;17:223–240.
43. Cantor JM, McPhail IV. Non-offending pedophiles. Curr
Sex Health Rep. 2016;8:121–128.
44. Houtepen JA, Sijtsema JJ, Bogaerts S. Being sexually
attracted to minors: sexual development, coping with
forbidden feelings, and relieving sexual arousal in self-
identified pedophiles. J Sex Marital Ther. 2016;42:48–69.
45. Maruna S, Immarigeon R. After Crime and Punishment:
Pathways to Offender Reintegration. Collompton, UK: Willan;
2011.
46. Levy B, Tsoy E, Brodt M, et al. Stigma, social anxiety, and
illness severity in bipolar disorder: implications for treatment.
Ann Clin Psychiatry. 2015;27:55–64.
47. Jahnke S, Imhoff R, Hoyer J. Stigmatization of people
with pedophilia: two comparative surveys. Arch Sex
Behav. 2015;44:21–34.
48. Harper CA, Bartels RM, Hogue TE. Reducing stigma and
punitive attitudes toward pedophiles through narrative
humanization. Sex Abuse. 2018;30:533–555.
49. Cohen LJ, Ndukwe N, Yaseen Z, et al. Comparison of self-
identified minor attracted persons who have and have not
successfully refrained from sexual activity with children.
J Sex Marital Ther. 2018;44:217–230.
50. Geer JH, Estupinan LA, Manguno-Mire GM. Empathy,
social skills, and other relevant cognitive processes in rapists
and child molesters. Aggress Violent Behav. 2000;5:99–126.
51. Torgersen S, Edvardsen J, Øien PA, et al. Schizotypal
personality disorder inside and outside the schizophrenic
spectrum. Schizophr Res. 2002;54:33–38.
52. Dolan M, Rogers P, Soothill K, et al. Handbook of Forensic
Mental Health. Cullompton, UK: Willan Publishing; 2008.
53. Hare RD. The Hare Psychopathy Checklist-Revised: Manual.
North Tanawanda, NY: Multi-Health Systems Inc; 1991.
54. Cantor JM. Understanding MRI research on pedophilia. Fall
2008 issue of the Association for the Treatment of Sexual
Abusers (ATSA). 2008. Available at: http://individual.
utoronto.ca/james_cantor/blog2.html. Accessed September 1,
2018.
55. Wilson HA. Can antisocial personality disorder be treated? A
meta-analysis examining the effectiveness of treatment in
reducing recidivism for individuals diagnosed with ASPD.
Int J Forensic Ment Health. 2014;13:36–46.
56. Messina N, Farabee D, Rawson R. Treatment responsiv-
ity of cocaine-dependent patients with antisocial person-
ality disorder to cognitive–behavioral and contingency
management interventions. J Consult Clin Psychol. 2003;
71:320–329.
Journal of Psychiatric Practice Vol. 24, No. 6
 ATTRACTION VERSUS ACTION IN PEDOPHILIC DESIRE

57. Bateman A, O’Connell J, Lorenzini N, et al. A rando-
mised controlled trial of mentalization-based treatment
versus structured clinical management for patients with
comorbid borderline personality disorder and antisocial
personality disorder. BMC Psychiatry. 2016;16:304.
58. Krampen G. Psychotherapeutic processes and outcomes
in outpatient treatment of antisocial behavior: an
integrative psychotherapy approach. J Psychother Integr.
2009;19:213–230.
59. Poleshuck E, Talbot N, Su H, et al. Pain as a predictor of
depression treatment outcomes in women with childhood
sexual abuse. Compr Psychiatry. 2009;50:215–220.
60. Sikkema K, Hansen N, Kochman A, et al. Outcomes from
a group intervention for coping with HIV/AIDS and
childhood sexual abuse: reductions in traumatic stress.
AIDS Behav. 2007;11:49–60.
61. Fonagy P, Luyten P. A developmental, mentalization-based
approach to the understanding and treatment of borderline
personality disorder. Dev Psychopathol. 2009;21:1355–1381.
62. Linder JM, Silverstone PH. Initial long-term findings
from a multimodal treatment program for child sexual
abuse victims demonstrate reduction of PTSD frequency
and symptoms. J Child Adolesc Behav. 2016;4:297. Doi:
10.4172/2375-4494.1000297.

Journal of Psychiatric Practice Vol. 24, No. 6 November 2018 387

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.