ArticlePDF Available

Hostility and Recidivism in Sexual Offenders

Springer Nature
Archives of Sexual Behavior
Authors:
  • University of Ottawa Psychology & Psychiatry (1975-2011) RETIRED

Abstract

In this study, we examined the association of hostility, as measured by the Buss-Durkee Hostility Inventory (BDHI), with offence characteristics and recidivism in 656 adult male sexual offenders. Hostility was significantly associated with having prior violent charges, the use of violence in the index sexual offence, sexual recidivism, and violent recidivism. After controlling for risk level, as measured by a modified version of the Rapid Risk Assessment for Sexual Offence Recidivism (RRASOR-mod), the significant association between hostility and sexual and violent recidivism remained. When examined by type of offender, hostility was significantly associated with recidivism in intrafamilial and extrafamilial child molesters, but not in rapists or mixed offenders. Given the predictive value of hostility independent of the RRASOR-mod, the present findings confirm and encourage treatment efforts directed toward the management of hostility and anger in sexual offenders.
Archives of Sexual Behavior, Vol. 34, No. 3, June 2005, pp. 277–283 ( C
2005)
DOI: 10.1007/s10508-005-3116-8
Hostility and Recidivism in Sexual Offenders
Philip Firestone, Ph.D.,1,3Kevin L. Nunes, B.A.,1Heather Moulden, B.A.,1
Ian Broom, M.A.,1and John M. Bradford, M.D.2
Received January 28, 2004; revision received April 6, 2004; accepted May 5, 2004
In this study, we examined the association of hostility, as measured by the Buss-Durkee Hostility
Inventory (BDHI), with offence characteristics and recidivism in 656 adult male sexual offenders.
Hostility was significantly associated with having prior violent charges, the use of violence in the index
sexual offence, sexual recidivism, and violent recidivism. After controlling for risk level, as measured
by a modified version of the Rapid Risk Assessment for Sexual Offence Recidivism (RRASOR-
mod), the significant association between hostility and sexual and violent recidivism remained. When
examined by type of offender, hostility was significantly associated with recidivism in intrafamilial
and extrafamilial child molesters, but not in rapists or mixed offenders. Given the predictive value
of hostility independent of the RRASOR-mod, the present findings confirm and encourage treatment
efforts directed toward the management of hostility and anger in sexual offenders.
KEY WORDS: sexual offenders; hostility; recidivism; Buss-Durkee Hostility Inventory; RRASOR.
INTRODUCTION
Hostility can be described as involving angry feelings
and connoting “a complex set of attitudes that motivate
aggressive behaviors directed toward destroying objects
or injuring other people” (Spielberger, Jacobs, Russell,
& Crane, 1983, p. 162). Etiological theories of sexual
offending consistently refer to the role of hostility in
sexual aggression. Marshall and Barbaree (1990) at-
tributed feelings of hostility observed in sexual offenders
to poor socialization, which prevents these individuals
from acquiring appropriate inhibitory controls over sex
and aggression. In their theory of sexual offending, Hall
and Hirschman (1991) postulated that negative affective
states, such as hostility, precede and facilitate sexual
aggression, whereas other factors, such as guilt and
empathy, fail to inhibit feelings of hostility. In examining
the offence process in child molesters, hostile feelings
1School of Psychology and Department of Psychiatry, University of
Ottawa, Ottawa, Ontario, Canada.
2Sexual Behaviours Clinic and Forensic Service, Royal OttawaHospital,
Ottawa, Ontario, Canada.
3Towhom correspondence should be addressed at School of Psychology,
University of Ottawa, 120 University Private, Ottawa, Ontario K1N
6N5, Canada; e-mail: fireston@uottawa.ca.
toward women and society have been cited as affective
pre-offence states (Proulx, Perrault, & Ouimet, 1999) and
anger and hostility have been associated with some sexual
offender typologies (Prentky & Knight, 1991).
Empirical work has supported these theories of
hostility in sexual offenders. Higher levels of both general
and victim specific hostility have been documented in the
sexual offender literature. Child molesters have reported
greater levels of trait anger and hostility compared to
nonsex offender comparison groups (Kalichman, 1991)
and hostility levels in rapists have also been reported
to be higher than in comparison groups (Hudson &
Ward, 1997; Marshall & Moulden, 2001; Rada, Laws,
Kellner, Stivastava, & Peake, 1983). Measures of hostility
have successfully differentiated between groups of sexual
offenders based on the level of violence involved in
their offence. In one study, researchers found significantly
higher hostility for sexual offenders who used excessive
physical force in their past offences compared to sexual
offenders who did not use excessive physical force (Rada
et al., 1983).
Although empirical support exists, research on the
relationship between hostility and sexual offending has
not been conclusive. Some researchers have not found
a significant association between hostility and the level
of violence used in offenders’ last offence, the number of
277
0004-0002/05/0600-0277/0 C
2005 Springer Science+Business Media, Inc.
278 Firestone, Nunes, Moulden, Broom, and Bradford
prior violent and nonviolent convictions (Holland, Levi, &
Beckett, 1983), or ratings of aggressiveness in psychiatric
patients (Edmunds, 1976). In other studies, no differences
have been found between child molesters or rapists on
measures of hostility when compared to nonsex offender
comparison groups (Overholser & Beck, 1986; Seidman,
Marshall, Hudson, & Robertson, 1994).
The lack of significant differences obtained by some
researchers may simply reflect Type-II errors. Alternately,
a theoretical explanation for non-significant differences is
that sexual offenders may differ from comparison groups
in terms of how they regulate or express their emotions
rather than their experience of those emotions (Lee,
Pattison, Jackson, & Ward, 2001). In an examination of di-
mensions of deficits in sexual offenders, Lee et al. (2001)
found a relationship between anger/hostility expression
and the type of paraphilia, such that pedophiles and exhibi-
tionists suppressed or turned anger inward, rapists turned
anger outward, and multiple paraphiliacs both suppressed
and acted out anger; however, although differences in
expression were detected, anger/hostility remained one
of only two common features of psychopathology for
each of the four targeted paraphilias (rapists, pedophiles,
exhibitionists, and multiple paraphiliacs). Based on some
evidence that hostility is higher in sexual offenders, and
associated with various types of sexual offending, one
would expect hostility to be related to future sexual
offending and would be relevant for the prediction of
recidivism across offender types.
Researchers have examined the degree to which hos-
tility is predictive of recidivism in sexual offenders. In a
recent meta-analysis, Hanson and Morton (2003) reported
a small but significant association between hostility and
recidivism in sexual offenders. Of the published studies
included in that meta-analysis, a significant proportion
found that hostility in sexual offenders predicted sexual
and/or violent recidivism. Quinsey, Khanna, and Malcolm
(1998) reported that hostility was significantly higher in
sexual and violent recidivists compared to non-recidivists.
In fact, hostility was the best predictor of recidivism
compared to the other self-report psychometric measures
examined (r=.18 for sexual recidivism; r=.22 for
violent [including sexual] recidivism). Other researchers
found that rumination of anger was related to sexual
recidivism in groups of mixed child molesters (Hudson,
Wales, Bakker, & Ward, 2002; Thornton, 2002).
The purpose of the present study was to examine the
association of hostility with offence characteristics and
recidivism in a sample of sexual offenders. The degree to
which hostility was related to prior sexual and violent
offences, the use of violence in sexual offences, and
sexual and violent recidivism was assessed. In addition,
we examined whether or not hostility would add to
prediction of recidivism beyond that of a well validated
risk assessment instrument (Doren, 2002; Hanson &
Morton, 2003), the Rapid Risk Assessment for Sexual
Offence Recidivism (RRASOR; Hanson, 1997).
METHOD
Participants
All 656 participants were assessed at the Royal
Ottawa Hospital, Sexual Behaviours Clinic, between 1982
and 1995. The participants were males, 18 years of age or
older at the time of their index offences, and had all been
convicted of a hands-on sexual offence against an adult or
a child (i.e., under the age of 16 at the time of the offence).
Offenders were classified based on their index and prior
sexual offences: 298 incest offenders (only related child
victims), 205 extrafamilial child molesters (only unrelated
child victims), 89 rapists (only adult victims), and 64
mixed offenders (i.e., falling into two or more of the
above offender categories). Descriptive statistics by type
of offender are presented in Table I. The majority of the
participants were assessed just prior to or just after their
court appearance or sentencing. Portions of this sample
have been examined in previous research (e.g., Firestone
et al., 1999, 2000; Nunes, Firestone, Bradford, Greenberg,
& Broom, 2002), but a longer follow-up time was used in
the current study.
Measures
Buss-Durkee Hostility Inventory (BDHI)
The Buss-Durkee Hostility Inventory (BDHI; Buss
& Durkee, 1957) is a 66-item true–false self-report
questionnaire that was developed with a non-forensic
sample and designed to measure 7 subtypes of hostility
(Assault, Indirect Hostility, Irritability, Negativism, Re-
sentment, Suspicion, and Verbal Hostility). Summing the
Hostility scale scores provides a Total Hostility score that
can range from 0 to 66, with higher scores reflecting
greater hostility. Test–retest reliability of .82 for the
BDHI total score over a 2-week period has been reported
(Biaggio, Supple, & Curtis, 1981). Biaggio (1980) found
BDHI scores were significantly correlated with anger
questionnaires, including the Anger Self-Report (Zelin,
Alder, & Myerson, 1972), Reaction Inventory (Evans
& Stangeland, 1971), and the Novaco Anger Inventory
(Novaco, 1975).
Hostility and Recidivism 279
Tab l e I . Descriptive Statistics (Mean [SD] or Percentage) by Offender Type
Variable ICMaECMbRapistcMixedd
BDHI total score 26.20 (12.09) 26.81 (13.26) 29.55 (12.87) 26.98 (13.62)
RRASOR-mod 0.41 (0.70) 2.10 (1.11) 1.44 (0.72) 2.20 (1.36)
Age at assessment 40.73 (11.05) 37.70 (12.46) 31.18 (10.40) 41.00 (12.89)
Education (years) 10.03 (2.67) 11.12 (3.66) 10.49 (3.22) 10.31 (3.42)
Ever married 86.3% 49.8% 36.0% 63.5%
Any prior sexual charges 18.1% 32.2% 20.2% 35.9%
Any prior violent charges 12.8% 20.1% 50.6% 25.0%
Violence in index offence 12.0% 11.6% 62.5% 18.6%
More than 1 victim 33.3% 51.6% 34.6% 98.4%
Sexual recidivism 10.4% 23.9% 25.8% 40.6%
Violent recidivism 20.1% 36.1% 40.4% 40.6%
Note. ICM =Intrafamilial child molesters. ECM =Extrafamilial child molesters.
anranged from 210 to 298.
bnranged from 189 to 205.
cnranged from 80 to 89.
dnranged from 59 to 64.
The Rapid Risk Assessment for Sexual
Offence Recidivism (RRASOR)
The RRASOR (Hanson, 1997) is an actuarial in-
strument designed to assess risk for sexual recidivism. It
consists of four items: (1) prior sexual offences, (2) age at
release, (3) victim gender, and (4) relationship to victim.
The coding rules were as follows: For the prior sexual
offences item, a score of 0 was assigned for 0 charges
or convictions, a score of 1 was given for 1 or 2 charges
or 1 conviction, a score of 2 was assigned for 3, 4, or
5 charges or 2 or 3 convictions, and a score of 3 was
given for 6 or more charges or 4 or more convictions.
The age item was scored 0 if the offender was 25-years-
old or older and 1 if the offender was under 25-years of
age when released to the community. The victim gender
item was scored 0 if there were only female victims and
1 if there were any male victims. The relationship item
was scored 0 if all victims were related to the offender
and 1 if any victims were unrelated. RRASOR scores can
range from 0 to 6, with higher scores reflecting higher
risk. These four items were selected from a larger pool
of variables through multivariate statistical procedures.
Relatively good associations have been demonstrated with
the RRASOR and sexual recidivism (r=.27 to .28) and
violent (including sexual) recidivism (r=.22) (Hanson,
1997; Hanson & Thornton, 2000). In their meta-analysis,
Hanson and Morton (2003) found average ds of .59 and
.34 for sexual and violent recidivism, respectively.
In the current study, the RRASOR was scored retro-
spectively and generally followed the coding guidelines
outlined by Hanson (1997). Due to the nature of our
database, however, some deviations from the guidelines
were necessary. Specifically, we did not make a distinction
between formal charges that did and did not result in
conviction. In addition, victim gender and relationship
to victim were coded from information pertaining only to
the index offence. Due to these deviations from the coding
guidelines in the present study, the instrument would be
most accurately described as a modified RRASOR; thus, it
was referred to as the RRASOR-mod. For some offenders
age at release was known. For others, it was estimated as
their age at two-thirds of their sentence.
Procedure
Offenders were assessed at a forensic psychiatric unit
regarding their index sexual offences and consented to the
use of their data for research. The research was approved
by the Royal Ottawa Hospital Ethics committee. Most
data were gathered at the time of assessment through
file reviews, interviews, and psychometric tests. Only a
portion of the data collected in these assessments was
examined here. Violence in the index offence was coded
as present if there had been threat of assault with a
weapon, minor injury with or without a weapon, severe
beating with or without a weapon, potential homicide,
homicide, or homicide with post-death mutilation. Self-
reported number of victims was coded as 1 vs. more than 1.
Offence information (i.e., prior and index offences)
was gathered from the Canadian Police Information
Center (CPIC) at the Ottawa Police Station, a national
database of criminal arrests and convictions including
INTERPOL reports from the Royal Canadian Mounted
Police. For an offender to be considered eligible to
280 Firestone, Nunes, Moulden, Broom, and Bradford
reoffend, he must have been free to commit a crime; that is,
he could not have been incarcerated or in secure custody
for reasons of mental illness. Recidivism information
was gathered from the CPIC records. Recidivism was
coded as the first offense committed when eligible. Sexual
recidivism was defined as a charge or conviction for a
sexual offence after the index offence. Violent recidivism
included a charge or conviction for a violent or sexual
offence after the index offence.
RESULTS
To limit the number of comparisons involving the
BDHI, a principal components extraction with varimax
rotation was performed on the 7 subscales; this procedure
has been utilized by other researchers to extract compo-
nents from the BDHI subscales (e.g., Holland et al., 1983).
Only one component was extracted (eigenvalue =4.12),
accounting for 58.9% of the variance. Accordingly, only
the BDHI total score, rather than the individual subscales,
was examined in subsequent analyses.
Descriptive statistics are presented in Table II, as are
correlations of variables of interest with the BDHI. The
BDHI was not significantly correlated with RRASOR-
mod scores. Although the BDHI was not significantly as-
sociated with any prior sexual charges, it was significantly
correlated with any prior violent charges and the use of
violence in the index offence. The BDHI was significantly
correlated with both sexual and violent recidivism.
The base rates of sexual recidivism and violent
recidivism were, respectively, 19.7 and 29.9%. Aver-
age follow-up time was 12.18 years (SD =3.33). The
RRASOR-mod was also significantly correlated with
sexual recidivism, r(637) =.22, p<.001, and violent
recidivism, r(637) =.20, p<.001. With regard to sexual
recidivism, mean BDHI total score was 30.55 (SD =
13.12) among recidivists and 26.03 (SD =12.50) among
Table II. Descriptive Statistics and Correlations with BDHI
Variable NMor % SD r
BDHI total score 656 26.92 12.74
RRASOR-mod 637 1.25 1.22 .06
Age at assessment 654 38.54 12.02 .24
Education (years) 602 10.48 3.24 .20
Ever married 572 63.1% .06
Any prior sexual charges 656 24.5% .05
Any prior violent charges 655 21.4% .18
Violence in index offence 612 19.1% .20
Sexual recidivism 656 19.7% .14
Violent recidivism 656 29.9% .21
p<.001.
TableIII. Sequential Logistic Regression Predicting Sexual Recidivism
from RRASOR-mod and BDHI
Scale BSEBWald Odds ratio 95% CI
Block 1
RRASOR-mod 0.42 .08 29.34∗∗ 1.52 1.31–1.77
Block 2
RRASOR-mod 0.42 .08 28.11∗∗ 1.51 1.30–1.76
BDHI 0.03 .01 9.531.03 1.01–1.04
Note. χ2(1,N =637) =29.44 for Block 1 (p<.001).
χ2(1,N =637) =9.60 for Block 2 (p<.01).
SE =Standard error.
CI =Confidence interval.
p<.01.
∗∗p<.001.
nonrecidivists; mean RRASOR-mod score was 1.80
(SD =1.35) among recidivists and 1.11 (SD =1.15)
among nonrecidivists. In terms of violent recidivism,
mean BDHI total scores were 30.99 (SD =12.61) among
recidivists and 25.19 (SD =12.41) among nonrecidi-
vists; mean RRASOR-mod scores were 1.67 (SD =
1.34) among recidivists and 1.11 (SD =1.16) among
nonrecidivists.
We were interested in whether hostility would
contribute independently to the prediction of sexual and
violent recidivism even after RRASOR-mod scores were
taken into account. A sequential logistic regression was
performed to determine whether the BDHI and RRASOR-
mod contributed uniquely to the prediction of sexual
recidivism. The RRASOR-mod was entered first followed
by the BDHI. As can be seen in Table III, the BDHI
contributed uniquely to the prediction of sexual recidi-
vism. A parallel sequential logistic regression for violent
recidivism was performed; the results are presented in
Table IV. As with sexual recidivism, the BDHI contributed
uniquely to the prediction of violent recidivism.
Tab l e I V. Sequential Logistic Regression Predicting Nonsexual Violent
(Including Sexual) Recidivism from RRASOR-mod and BDHI
Scale BSEBWald Odds ratio 95% CI
Block 1
RRASOR-mod 0.34 .07 23.011.40 1.22–1.61
Block 2
RRASOR-mod 0.33 .07 21.181.39 1.21–1.60
BDHI 0.04 .01 23.861.04 1.02–1.05
Note. χ2(1,N =637) =26.49 for Block 1 (p<.001).
χ2(1,N =637) =24.74 for Block 2 (p<.001).
SE =Standard Error.
CI =Confidence Interval.
p<.001.
Hostility and Recidivism 281
Tab l e V. BDHI Correlations with Sexual Recidivism by Type of Offender
Sexual recidivism
No Yes
Type of offender n M SD n M SD r rp
Intrafamilial child molesters 267 25.46a11.72 31 32.55a13.51 .18∗∗ .16
Extrafamilial child molesters 156 25.62ab 12.80 49 30.63a14.10 .16.18
Rapists 66 29.09b13.56 23 30.87a10.77 .06 .03
Mixed 38 26.47ab 14.26 26 27.73a12.86 .05 .04
Note. Means in each column not sharing same superscript letter are significantly different (p<.05). In partial
correlations, n=289 for intrafamilial child molesters, n=197 for extrafamilial child molesters, n=81 for
rapists, and n=58 for mixed offenders.
p<.05. ∗∗p<.01.
Although hostility was associated with recidivism
for the entire sample, it was important to assess whether
this association was present in the different types of
sexual offenders. Correlations between the BDHI and
sexual recidivism and violent recidivism are reported,
respectively, in Tables V and VI for intrafamilial child
molesters, extrafamilial child molesters, rapists, and
mixed offenders. Partial correlations between the BDHI
and recidivism, while controlling for the RRASOR, are
also provided in Tables V and VI. For both sexual
and violent recidivism, small significant correlations and
partial correlations were found with hostility in the
intrafamilial and extrafamilial child molesters, but not in
the rapists or mixed offenders. However, these correlations
were not significantly different from one another.
To compare types of offenders, t-tests were con-
ducted. As can be seen in Tables V and VI, the failure
to find significant correlations between the BDHI and
recidivism in the rapists and mixed offenders was not
indicative of significantly lower BDHI scores in these
groups relative to the intrafamilial or extrafamilial child
molesters.
DISCUSSION
We found that hostility, as measured by the BDHI,
was associated with a higher likelihood of sexual and
violent recidivism. This result is generally consistent with
the findings of other researchers (Hanson & Morton,
2003). More importantly, we found that hostility added to
the prediction of sexual and violent recidivism above and
beyond the RRASOR-mod. This suggests that hostility
may be an important construct to assess in addition
to the predictors addressed by the RRASOR-mod (i.e.,
prior sex offences, age, victim gender, and relationship
to victim). When the types of offenders were examined
separately, hostility was significantly associated with both
sexual and violent recidivism even after controlling for the
RRASOR-mod in the intrafamilial and extrafamilial child
molesters, but not in the rapists and mixed offenders.
In the present study, we attempted to contribute to
the hostility literature and to clarify some discrepancies.
Although not related to previous sexual offences, hostility
was related to previous violent offences. Given that
the BDHI is a measure of general hostility, it may
Tab l e V I. BDHI Correlations with Violent Recidivism by Type of Offender
Violent recidivism
No Yes
Type of offender n M SD n M SD r rp
Intrafamilial child molesters 238 24.66a11.48 60 32.32a12.59 .26.25
Extrafamilial child molesters 131 24.34a12.62 74 31.19a13.33 .25.26
Rapists 53 28.74b14.08 36 30.75a10.91 .08 .05
Mixed 38 26.47ab 14.26 26 27.73a12.86 .05 .04
Note. Means in each column not sharing same superscript letter are significantly different (p<.05). In partial
correlations, n=289 for intrafamilial child molesters, n=197 for extrafamilial child molesters, n=81 for
rapists, and n=58 for mixed offenders.
p<.001.
282 Firestone, Nunes, Moulden, Broom, and Bradford
more accurately capture general violent tendencies versus
those sexual in nature; however, consistent with previous
literature, hostility was predictive of recidivism in this
sample (Firestone et al., 1999, 2000; Thornton, 2002;
Ward et al., 2002) and complemented the RRASOR-
mod in the prediction of recidivism. Given the predictive
value of hostility beyond the RRASOR-mod, the present
findings confirm and encourage treatment efforts directed
toward the management of hostility and anger in sexual
offenders.
The present study had some limitations, such as
reliance on police reported sexual and violent offences,
which obviously fails to capture the behavioral and risk
dynamics of those sexual offenders who do reoffend but
are not caught. In addition, because we used a modified
version of the RRASOR, it cannot be assumed that
hostility would not be redundant to the actual RRASOR
in predicting recidivism. There is evidence that different
types of sexual offenders are characterized by different
risk factors and recidivism rates (Firestone et al., 1999);
thus, future research on the relationship between hostility
and recidivism, particularly in rapists, would be valuable.
A strength of the study is the population from which
the sample was drawn. This population is unique in that
it comprised men who have been convicted of a sexual
offence and have been assessed in a sexual behavior
clinic located in the forensic unit of a large psychiatric
hospital. Many of these offenders were not sentenced
to time in prison for their offences (Nunes, Firestone,
Jensen, Wexler, & Bradford, 2005). Therefore, this group
may be more representative of the wide array of men
processed through the courts than those used in previous
studies, often drawn from maximum security institutions.
The present results suggest that previous findings of a
relationship between hostility and recidivism in sexual
offenders generally hold true when tested on a lower risk
population, and that hostility may be an important factor
to consider in addition to constructs tapped by instruments
such as the RRASOR-mod.
REFERENCES
Biaggio, M. (1980). Assessment of anger arousal. Journal of Personality
Assessment,44, 289–298.
Biaggio, M. K., Supple, K., & Curtis, N. (1981). Reliability and validity
of four anger scales. Journal of Personality Assessment,45, 639–
648.
Buss, A. H., & Durkee, A. (1957). An inventory for assessing different
kinds of hostility. Journal of Consulting Psychology,21, 343–
349.
Doren, D. M. (2002). Evaluating sex offenders: A manual for civil
commitments and beyond. Thousand Oaks, CA: Sage.
Edmunds, G. (1976). The predictive validity of the Buss-Durkee
Inventory. Journal of Clinical Psychology,32, 818–820.
Evans, D. R., & Stangeland, M. (1971). Development of the Reaction
Inventory to measure anger. Psychological Reports,29, 412–
414.
Firestone, P., Bradford, J. M., McCoy, M., Greenberg, D. M., Curry, S.,
& Larose, M. R. (2000). Prediction of recidivism in extra-familial
child molesters based on court-related assessments. Sexual Abuse:
A Journal of Research and Treatment,12, 203–222.
Firestone, P., Bradford, J. M., McCoy, M., Greenberg, D. M., Larose,
M. R., & Curry, S. (1999). Prediction of recidivism in incest
offenders. Journal of Interpersonal Violence,14, 511–531.
Hall, G. C. N., & Hirschman, R. (1991). Toward a theory of sexual
aggression: A quadripartite model. Journal of Consulting and
Clinical Psychology,59, 643–669.
Hanson, R. K. (1997). The development of a brief actuarial risk scale
for sexual offence recidivism. Ottawa, Ontario: Department of the
Solicitor General of Canada. Retrieved from www.sgc.gc.ca.
Hanson, R. K., & Harris, A. J. R. (2000). Where should we intervene?
Dynamic predictors of sexual offence recidivism. Criminal Justice
and Behavior,27, 6–35.
Hanson, R. K., & Morton, K. E. (2003, June). Recidivism risk factors
for sexual offenders: An updated meta-analysis. Paper presented at
the Annual Convention of the Canadian Psychological Association,
Hamilton, Ontario.
Hanson, R. K., & Thornton, D. (2000). Improving risk assessments for
sex offenders: A comparison of three actuarial scales. Law and
Human Behavior,24, 119–136.
Holland, T. R., Levi, M., & Beckett, G. E. (1983). Ethnicity, criminality,
and the Buss-Durkee Hostility Inventory. Journal of Personality
Assessment,47, 375–379.
Hudson, M. S., Wales, D. S., Bakker, L., & Ward, T. (2002). Dynamic
risk factors: The Kia Marama evaluation. Sexual Abuse: A Journal
of Research and Treatment,14, 103–119.
Hudson, M. S., & Ward, T. (1997). Intimacy, loneliness, and attachment
styles in sexual offenders. Journal of Interpersonal Violence,12,
323–339.
Kalichman, S. C. (1991). Psychopathology and personality character-
istics of criminal sexual offenders as a function of victim age.
Archives of Sexual Behavior,20, 187–197.
Lee, J. K. P., Pattison, P., Jackson, H. J., & Ward, T. (2001). The general,
common, and specific factors of psychopathology for different
types of paraphilias. Criminal Justice and Behavior,28, 227–
266.
Marshall, W. L., & Barbaree, H. E. (1990). An integrated theory of the
etiology of sexual offending. In W. L. Marshall, D. R. Laws, & H. E.
Barbaree (Eds.), Handbook of sexual assault: Issues, theories, and
treatment of the offender (pp. 257–275). New York: Plenum.
Marshall, W. L., & Moulden, H. (2001). Hostility toward women and
victim empathy in rapists. Sexual Abuse: A Journal of Research
and Treatment,13, 249–255.
Novaco, R. W. (1975). Anger control: The development and evaluation
of an experimental treatment. Lexington, MA: Lexington.
Nunes, K. L., Firestone, P., Bradford, J. M., Greenberg, D. M., &
Broom, I. (2002). A comparison of the modified versions of the
Static-99 and the Sex Offender Risk Appraisal Guide (SORAG).
Sexual Abuse: A Journal of Research and Treatment,14, 249–
265.
Nunes, K. L., Firestone, P., Jensen, T., Wexler, A. F., & Bradford,
J. M. (2005). Incarceration and recidivism in a sample of sexual
offenders. Manuscript in preparation.
Overholser, J. C., & Beck, S. (1986). Multimethod assessment of
rapists, child molesters, and three control groups on behavioral
and psychological measures. Journal of Consulting and Clinical
Psychology,54, 682–687.
Prentky, R. A., & Knight, R. A. (1991). Identifying critical dimensions
for discriminating among rapists. Journal of Consulting and
Clinical Psychology,59, 643–661.
Proulx, J., Perreault, C., & Ouimet, M. (1999). Pathways in the offending
process of extrafamilial sexual child molesters. Sexual Abuse: A
Journal of Research and Treatment,11, 117–129.
Hostility and Recidivism 283
Quinsey, V. L., Khanna, A., & Malcolm, P. B. (1998). A retrospec-
tive evaluation of the Regional Treatment Centre sex offender
treatment program. Journal of Interpersonal Violence,13, 621–
644.
Rada, R. T., Laws, D. R., Kellner, R., Stivastava, L., & Peake, G. (1983).
Plasma androgens in violent and nonviolent sex offenders. Bulletin
of the American Academy of Psychiatry and the Law,11, 149–158.
Seidman, B. T., Marshall, W. L., Hudson, S. M., & Robertson, P. J.
(1994). An examination of intimacy and loneliness in sex offenders.
Journal of Interpersonal Violence,9, 518–534.
Spielberger, C. D., Jacobs, G., Russell, S., & Crane, R. S. (1983).
Assessment of anger: The State-Trait Anger Scale. In J. N. Butcher
& C. D. Spielberger (Eds.), Advances in personality assessment
(Vol. 2, pp. 161–189). Hillsdale, NJ: Erlbaum.
Thornton, D. (2002). Construction and testing a framework for dynamic
risk assessment. Sexual Abuse: A Journal of Research and Treat-
ment,14, 139–153.
Zelin, M. L., Alder, G., & Myerson, P. G. (1972). Anger self-report: An
objective questionnaire for the measurement of aggression. Journal
of Consulting and Clinical Psychology,39, 340.
... Providing some support for Ellis ' and Tafrate's (1997) claims, Tafrate et al. (2002) found (N = 91) that demandingness was present in 90% of anger episodes, global ratings of others was present 53% of the time, low frustration tolerance was 46%, distortion in reality was 32%, awfulizing was 31%, and global ratings of self was present 24% (see DiGiuseppe et al. (2023) for a critique of Ellis' model of these cognitions). Additionally, hostility is believed to be common in dysfunctional anger, and is a core mechanism in violent behavior (Dodge, 1991;Firestone et al., 2005;Hutchings et al., 2010). Thoughts of revenge are also associated with increased state anger intensity. ...
... They found that four studies reported small effect size, three found a medium effect size, and two found a large effect size, ranging from d = 0.10 to 1.14. In fact, despite there being a large effect size for the relationship between cognitive distortions and violence (Chereji et al., 2012), only a few cognitions have been identified leading to physical aggression (code of honor (McGill et al., 2021), retribution (DiBlasi & Kassinove, 2021DiBlasi et al., 2024), hostility (Firestone et al., 2005;Hutchings et al., 2010)), and they have not all been examined in the context of anger. There may be specific cognitions in anger episodes that lead to physical aggression compared to anger episodes that do not lead to physical aggression. ...
... As such, these cognitions were examined in this study. Although hostility is supported by previous research (Dodge, 1991;Firestone et al., 2005;Hutchings et al., 2010), it was not easily able to be coded by the research assistants, and thus it was not included. ...
Article
Full-text available
Background Anger is a primary emotion that can lead to physical aggression 10 -25% of the time. Given this, it is important to understand which cognitions are present when anger leads to physical aggression. Method Using Prolific, 302 participants were randomly assigned to write about a time they were angry that they did or did not physically aggress. Using a mixed method approach, participants answered close-ended questions regarding the presence of eight angry cognitions. Additionally, their open-ended answers were coded for the same angry cognitions. Results Participants who physically aggressed (PA) when angry were more likely to endorse thoughts of just desserts and have more coded cognitions of low frustration tolerance and retribution than participants who did not physically aggress when angry (NPA). Participants in the NPA condition were more likely to endorse and have coded thoughts of demandingness than participants in the PA condition. Moreover, participants in the PA condition were more likely to have more coded angry cognitions in general than participants in the NPA condition. Conclusion Total angry cognitions, and specific cognitions (just desserts, low frustration tolerance, and retribution) may be important treatment targets for anger and physical aggression. Implications and additional treatment targets are discussed.
... Providing some support for Ellis' and Tafrate's (1997) claims, Tafrate et al. (2002) found (N = 91) that demandingness was present in 90% of anger episodes, global ratings of others was present 53% of the time, low frustration tolerance was 46%, distortion in reality was 32%, awfulizing was 31%, and global ratings of self was present 24% (see DiGiuseppe et al., (2023) for a critique of Ellis' model of these cognitions). Additionally, hostility is believed to be common in dysfunctional anger, and is a core mechanism in violent behavior (Dodge, 1991;Firestone et al., 2005;Hutchings et al., 2010). Thoughts of revenge are also associated with increased state anger intensity (DiGiuseppe & Froh, 2002). ...
Preprint
Full-text available
Background Anger is a primary emotion that can lead to physical aggression 10% -25% of the time. Given this, it is important to understand which cognitions are present when anger leads to physical aggression. Method Using Prolific, 302 participants were randomly assigned to write about a time they were angry that they did or did not physically aggress. Using a mixed method approach, participants answered close-ended questions regarding the presence of eight angry cognitions. Additionally, their open-ended answers were coded for the same angry cognitions. Results Participants who physically aggressed (PA) when angry were more likely to endorse thoughts of just desserts and have more coded cognitions of low frustration tolerance and retribution than participants who did not physically aggress when angry (NPA). Participants in the NPA condition were more likely to endorse and have coded thoughts of demandingness than participants in the PA condition. Moreover, participants in the PA condition were more likely to have more coded angry cognitions in general than participants in the NPA condition. Conclusion Total angry cognitions, and specific cognitions (just desserts, low frustration tolerance, and retribution) may be important treatment targets for anger and physical aggression. Implications and additional treatment targets are discussed.
... Cognitive distortions -habitual ways of thinking that are often inaccurate and negatively biased -can play a major role in both sexual and non-sexual violent offenses (Smallbone & Milne, 2000;Firestone et al., 2005;Simons, 2015;Helmond et al., 2014). For example, child sex offenders tend to perceive children as sexual beings and their urges as uncontrollable (Ward & Keenan, 1999;Whitaker et al., 2008). ...
Article
Full-text available
Sex offenders and non-sexual violent offenders display distinct cognitive distortions that shape their offenses. Some evidence suggests that early maladaptive schemas which reflect individuals' fundamental cognitions and beliefs, contribute to these cognitive distortions. However, these have not been systematically appraised. We conducted a systematic review to synthesize the evidence regarding the schema profiles of sex offenders and non-sexual violent offenders, and to explore differences in schemas between these offender types. Eight studies met the inclusion criteria. Within the sex offender category, rapists and child sex offenders endorsed similar schemas. For non-sexual violent offenders, schemas related to self-control issues and self-entitlement were most prominent. Sex offenders demonstrated more schemas that endorsed negative perceptions of themselves and others than violent offenders. Findings also suggest that atypical sexual interests and mental disorders may be important. This evidence highlights the need for a more fine-tuned understanding of the cognitive profiles of sexual and non-sexual violent offenders and for developing schema-based offender management and rehabilitation services.
... It should be noted that our study primarily focused on patients at high risk of suicide or aggression. Such individuals, characterized by heightened hostility and aggression, may inherently exhibit increased sexual behaviors (42). Yet, there are constraints to our study. ...
Article
Full-text available
Background Sexual dysfunction is commonly observed in individuals with Major Depressive Disorder (MDD), along with various psychological symptoms such as anxiety, somatic complaints, interpersonal sensitivity, and obsessive-compulsive tendencies. However, there is a research gap in understanding the impact of these psychological symptoms on sexual functioning in MDD. Furthermore, there is limited data on the incidence of sexual dysfunction among drug-naive MDD patients in West China. This study aims to determine the prevalence of sexual dysfunction in this patient population and explore its association with other psychological indicators. Methods We conducted a retrospective analysis of patient data from October 2020 to September 2022 using propensity score matching. A focused group of 165 males and 490 females was selected from a total of 1941 MDD patients. This allowed for a comparative analysis of demographic data, as well as scores from the Self-Rating Depression Scale (SDS), Self-Rating Anxiety Scale (SAS), and Symptom Checklist-90 (SCL-90), the Arizona Sexual Experience Scale (ASEX). Results Our findings reveal that 46.2% of drug-naive MDD patients experienced sexual dysfunction. Notably, there was a higher prevalence of sexual dysfunction among female patients (50.3%) compared to males (37.5%). MDD patients without sexual dysfunction consistently exhibited higher SDS scores than those with sexual dysfunction (p < 0.01), There were no statistically significant differences between male and female MDD patients with or without concomitant sexual dysfunction in terms of Somatic complaints, Obsessive-compulsive, Interpersonal sensitivity, Anxiety, Phobic anxiety, Paranoid ideation, Psychoticism and Diet/sleep difficulties (p > 0.05). In addition, male MDD patients with sexual dysfunction showed a emerging trend towards elevated Hostility scores on the SCL-90 (p = 0.058), male MDD patients with sexual dysfunction showed an increasing trend in hostility scores on the SCL-90, whereas female MDD patients with sexual dysfunction did not show such a trend. Conclusion The study highlights a significant gender difference in the prevalence of sexual dysfunction among MDD patients, with females being more susceptible than males. There is a positive correlation between the severity of depression and sexual dysfunction in both genders. Interestingly, male MDD patients demonstrated a potential protective effect of hostility against sexual dysfunction, which was not observed in female patients.
... In line with prior studies on recidivism among sexual offenders (Firestone et al., 2005), hostility was also found to be a direct predictor of violent recidivism. High scorers on hostility suspect others of negative intentions and attribute hostile motives to systems, others, or institutions (Spreen et al., 2014). ...
Article
Full-text available
In this study, we investigated the potential moderating role of instability in living arrangements in the association between 14 clinical factors of the Historical Clinical Future Revised (HKT-R) and 5-year violent recidivism. The sample included 316 male forensic psychiatric patients who were unconditionally released from Dutch forensic psychiatric institutions between 2004 and 2008. Results showed that instability in living arrangements did not moderate the association between the clinical factors and violent recidivism, however, it directly and positively predicted (OR = 3.28) violent recidivism. Likewise, hostility (OR = 4.36) and addiction (OR = 1.65) were directly and positively related to violent recidivism, while psychotic symptoms (OR = 0.25) were inversely related.
... As expected, patients who were characterized by higher levels of impulsivity and hostility at the time of unconditional release were more likely to violently reoffend up to 5 years from release, which is in line with previous research (Douglas & Skeem, 2005;Spreen et al., 2014). However, in previous studies, the link between hostility and recidivism was mainly determined in a sample of sexual offenders (Firestone et al., 2005;Pettersen et al., 2015), while the current study demonstrates this relationship also in a sample of violent offenders. Moreover, the results showed that influence of risky network members is the strongest predictor Note. ...
Article
Full-text available
This study investigated which risk and protective factors, based on the 14 clinical indicators of the Historical-Clinical-Future Revised, significantly predicted violent recidivism in a sample of 315 male forensic psychiatric patients. Additionally, it was investigated whether these associations were moderated by intellectual ability. Regarding risk factors, a stronger influence of risky network members, and higher levels of hostility, impulsivity, and addiction significantly predicted violent recidivism. Likewise, regarding protective factors, poorer social and labor skills, and a lower degree of patient’s acceptance of crime responsibility were significant predictors of violent recidivism. Contrary to our expectations, better coping skills and more insight into risky behaviors that can lead to relapse also contributed significantly to an increased likelihood of violent recidivism. Intellectual ability had no significant moderating effect on the associations between the factors and violent recidivism. The results offer an insight into which factors need to be prioritized during treatment.
Article
Introduction: Aim of the research presented in this article is investigation of indirect aggression among inmates. Indirect aggression covers aggressive behaviors toward social position and reputation of the victim, such as ridiculing, disseminating of fake information and social exclusion. Relationship between use of indirect aggression and being victim of indirect aggression in penitentiary context was examined. Sex differences in indirect aggression also was considered. Moreover, direct aggression, anger and hostility were tested as predictors of indirect aggression.Method: Two versions of Indirect aggression scale (IAS-Aggressor, IAS-Target) as well as Buss-Perry Aggression Questionnaire to examined selected psychological variables were used. Cumulatively 78 penitentiary recidivists and first time sentenced were examined (40 males, 38 females).Results: Relationship between inmate’s indirect aggression use and its experience was revealed. The research show also that male offenders more often use indirect aggression than sentenced woman. Physical aggression and hostility predicted indirect aggression in investigated penitentiary sample.Key practitioner massages: Presented research suggest that indirect aggression among inmates can be a serious threat during incarceration. From that reason early diagnosis and prevention of covert aggressive behaviors among inmates is needed.
Article
Full-text available
The Buss-Durkee Hostility Inventory (BDHI) is an important assessment scale of hostility in forensic psychiatry. We analyzed the validity and reliability of a Papiamento translation of the BDHI in 134 pre-trial defendants in Curaçao using Exploratory Structural Equation Modeling (ESEM). The reliability of the Direct and Indirect Hostility BHDI-P subscales were good and the reliability of the Social Desirability poor. There was a negative correlation between Direct Hostility and Agreeableness and a positive correlation between Indirect Hostility and Anxiety. We conclude that the BDHI-P has an acceptable measurement quality when used in defendants.
Chapter
This chapter will discuss ethical issues connected with neuroscientific research on prisoners, focusing specifically on whether participating in such research should always be disregarded when making decisions about early release from prison. It was once routine in some jurisdictions for a prisoner’s participation in medical research to count as “good behaviour”, which could be given weight in decisions about early release. However, medical ethicists now widely regard this practice as ethically problematic, because prisoners might feel pressurised to participate in medical research programmes by the desire to escape a coercive and harsh prison environment, and thus their agreement to take part in the programme might not satisfy the freedom requirement for valid consent. Yet, there may be a tension between traditional medical ethics and theories of punishment (a tension which also arises in other contexts at the interface between medicine and punishment). From the perspective of retributive or communication/reform-based theories of punishment, it might seem unfair for parole boards to disregard willingness to participate in medical research, if the prisoner were motivated by altruistic considerations. This altruism might indicate that the prisoner “deserves” a shorter sentence or may be a sign that the prisoner has reformed. This chapter will consider whether/how this tension between traditional medical ethics and penal theories can be resolved. For example, if the medical research involves relatively safe interventions such as fish oil pills or transcranial direct current stimulation (tDCS), then the above-mentioned worries about prisoners’ ability to provide free consent may be less pressing than if the risks of side-effects were greater. Furthermore, the validity of consent also seems to depend partly on whether the prison conditions are humane or excessively harsh. In addition, problems also arise in this context from within penal theories. Ironically, transparency about what factors will be taken into account in decisions about early release, which would presumably be encouraged by communication/reform theories, could raise the possibility that the prisoner was motivated to participate in the research programme by a desire for early release rather than altruism, thus making it harder to determine whether participation is a genuine sign of reform.
Article
Hostility has been identified as a risk factor in predicting offending. Some research has conceptualized hostility as a facet of mental health. One factor that has been identified to be protective against the effects of mental health on behavioral outcomes is social support. Despite this, prior research has not considered social support in this manner in moderating the relationship between hostility and offending. This study utilized data from all 11 waves of the Pathways to Desistance study. Mixed-effects regression models were used to test for the direct effects of hostility on offending outcomes and determine whether social support moderated these relationships. Results indicated that greater hostility predicted increased offending variety for both outcomes. Lower levels of social support also predicted increased offending scores for both outcomes. No significant moderation effect was observed for either offending outcome. Implications are discussed.
Article
Full-text available
Considerable evidence has amassed in studies of both nonoffender and offender samples that demonstrates both that sexual aggression is determined by a multiplicity of variables and that convicted sexual offenders are markedly heterogeneous (Knight, Rosenberg, & Schneider, 1985; Malamuth, 1986). Attempts both to identify sexually coercive men in normal samples and to assess etiology, concurrent adaptation, treatment efficacy, and recidivism for convicted sexually aggressive offenders have also suggested that the critical determining components of sexual aggression interact in complex ways. The purpose of this article is to survey both the offender and nonoffender sexual aggression research for evidence about which dimensions should be included in multivariate models that attempt to discriminate rapists from nonrapists, to identify subgroups among rapists, or to enhance the efficiency of dispositional decisions about these offenders.
Article
Full-text available
This chapter represents an attempt to integrate a widely disparate literature concerning factors which play a role in the etiology of sex offending and lead to its persistence. In particular we are concerned that most researchers seem to take a rather narrow perspective of this behavior, stressing their own preferred processes (i.e., psychological, biological, or sociological) to the virtual exclusion of others. We have previously emphasized the role of learning experiences (Marshall and Barbaree, 1984a), sociocultural factors (Marshall, 1984a), and biological processes (Marshall 1984b) in the etiology of rape, but this represents our first attempt at integration and the first time we have extended our theorizing to account for other sex offenses. We believe that a proper understanding of sex offending can only be attained when these diverse processes are seen as functionally interdependent.
Article
Full-text available
Of 483 inmates who were referred to a sex offender treatment program and followed for an average of 44 months of opportunity to reoffend, 38% were arrested for new violent or sexual offenses. Treated offenders were most frequently rearrested for sex offenses. Inmates judged unsuitable for treatment were rearrested less frequently, particularly for sex offenses. Inmates judged to not require treatment and those who refused treatment also had fewer rearrests for sex offenses than did treated participants, although they had more rearrests for violent offenses. After statistically controlling for the static variables that predicted reoffending, the treatment program was associated with a higher rate of sexual rearrests but had no effect on the composite variable, violent or sexual rearrests. Among treated offenders, clinical assessment of treatment gains was not significantly associated with recidivism.
Article
Full-text available
Recent research suggests that offenders have difficulties in successfully establishing adult relationships. The authors have developed an attachment-based model that relates offending styles and interpersonal goals, and they have provided some preliminary evidence for it at the level of attachment style and offender type. Here the authors examined offender type, attachment style, and some relevant interpersonal variables. Although only the offender-type differences related to anger, most of the attachment model's predictions were supported. Preoccupiedly and fearfully attached men were the most lonely, whereas fearfully and dismissingly attached men scored highest on fear of intimacy, anger expression, and anger suppression. Fearful men reported the greatest hostility toward women, whereas dismissing men were most accepting of rape myths. The relationship between attachment style and social competency issues appears more fundamental than that between offender type and social competency. As such, approaching the social dysfunctions associated with offending from an attachment perspective has considerable clinical utility.
Article
The aim of the current study was to investigate specific pathways in the offending process of extrafamilial sexual child molesters. Forty-four men who had committed at least one sexual offense against a nonfamilial prepubescent child were included in this study and were classified using cluster analysis. Subjects using the coercive pathway (n = 30) had generally used psychoactive substances before their offenses. Furthermore, they had molested a female victim without perceived vulnerability and whom they had already well known. These molesters had not planned their offense, which was of short duration (less than 15 min), and involved coital activities and coercion (verbal and/or physical). Subjects using the noncoercive pathway (n = 14) had generally used pornography and deviant sexual fantasies before their offenses. Moreover, they had molested a male victim, in whom they perceived a psychosocial vulnerability and who was not familiar to them. These molesters had planned their offense, which was of longer duration (more than 15 min) and involved noncoital activities without coercion. These two pathways were compared to the two pathways in the offending process identified by Ward and his colleagues.
Article
Risk assessment is an essential part of clinical practice. Each of the three aspects of risk (static, stable, and acute dynamic) are important at various points of contact between the man and the systems that are responsible for providing service. Dynamic factors, the typical treatment and supervision targets, have received less research attention than static factors. This paper examined the extent to which pretreatment, posttreatment and change scores were associated with reoffending among men incarcerated for sexually molesting. The results were generally supportive of change in prooffending attitudes as the key to not reoffending and suggested that the perspective-taking component of empathy and the use of fantasy may be important mechanisms. Affect scales generally failed to show any relationship with reoffending, outside decreases in trait and suppressed anger. Moreover, these data suggest that we could improve our assessments and treatment through increased sensitivity to offense pathways.
Article
This paper describes the construction and testing of a framework for dynamic risk assessment. A review of previous studies identified 4 domains into which dynamic risk factors for sexual offending seem to fall. These were sexual interests, distorted attitudes, socioaffective functioning, and self-management. Psychometric indicators for 3 of the domains were identified, and 2 studies are reported using these indicators to test the framework. Study 1 divided men serving a prison sentence for a sexual offense against a child into 2 groups-those with a previous conviction of this kind (Repeaters) and those for whom this was the only time they had been sentenced for such an offense (Current Only). The Repeaters were found to show more distorted attitudes, worse socioaffective functioning, and poorer self-management than did the Current Only group. Study 2 used a simple algorithm to combine these psychometric indicators into an overall “Deviance” classification. Reconviction data was obtained for offenders classified as high, moderate, or low on Deviance. Sexual reconviction was found to be monotonically associated with the Deviance classification. Logistic regression analysis showed that both static variables (Static-99) and the Deviance classification made independent contributions to prediction. It is suggested that risk assessment procedures should combine these 2 approaches.
Article
The Reaction Inventory was developed to measure anger, and the 76 items in the inventory were administered to 275 Ss. The estimate of the internal consistency coefficient was .95. The coefficient of concurrent validity with the Buss-Durkee Inventory for one subsample of 45 Ss was .52 and for another subsample of 138 Ss was .57. Factor analysis of the Reaction Inventory yielded 10 factors.
Article
The Buss-Durkee Inventory was administered to 95 male psychiatric patients. Each patient was rated for aggression by two nurses and a psychiatrist. The three sets of ratings had significant correlations with each other, but were not related significantly to the aggression and hostility subscales. A canonical variate analysis of the data also failed to yield significant results. Age was found to be correlated significantly with each set of ratings and to be unrelated to the subscale scores. This suggested that validation studies that use psychiatric patients should take account of the age and severity of illness of the Ss and that a more profitable design might involve samples other than psychiatric ones.