Professional Documents
Culture Documents
To cite this article: Eli Coleman, Michael Miner, Fred Ohlerking, Nancy
Raymond (2001) Compulsive Sexual Behavior Inventory: A Preliminary Study of
Reliability and Validity, Journal of Sex & Marital Therapy, 27:4, 325-332, DOI:
10.1080/009262301317081070
Taylor & Francis makes every effort to ensure the accuracy of all
the information (the “Content”) contained in the publications on our
platform. However, Taylor & Francis, our agents, and our licensors
make no representations or warranties whatsoever as to the accuracy,
completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of
the authors, and are not the views of or endorsed by Taylor & Francis.
The accuracy of the Content should not be relied upon and should be
independently verified with primary sources of information. Taylor and
Francis shall not be liable for any losses, actions, claims, proceedings,
demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in
relation to or arising out of the use of the Content.
This article may be used for research, teaching, and private study
purposes. Any substantial or systematic reproduction, redistribution,
reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access
and use can be found at http://www.tandfonline.com/page/terms-and-
conditions
Downloaded by [Central Michigan University] at 11:33 02 December 2014
Journal of Sex & Marital Therapy, 27:325–332, 2001
Copyright © 2001 Brunner-Routledge
0092-623X/01 $12.00 + .00
Eli Coleman, Michael Miner, Fred Ohlerking, and Nancy Raymond are affiliated with the
Program in Human Sexuality, University of Minnesota Medical School. A grant from the
Minnesota Department of Corrections and funds from the University of Minnesota Medical
School, Departments of Psychiatry and Family Practice supported this research. The authors
would like to thank Michael Ross, Ph.D., MPH, Ross Crosby, and Deb Finstad for their
assistance in data analysis. The first author thanks Philip Colgan, Ph.D., for his assistance in
developing the original items from which the scale was eventually derived .
Address correspondence to Eli Coleman, 1300 S. 2nd St. #180, Minneapolis, MN 55454,
USA. E-mail: colem001@tc.umn.edu
325
326 E. Coleman et al.
ing, and (d) create health risks (Coleman, 1991). CSB can be divided into
two basic types, paraphilic and nonparaphilic. Paraphilias are nonnormative
behavior involving recurrent and distressing fantasies. Nonparaphilic CSB
falls under the category of sexual disorder not otherwise specified, with the
description of “distress about a pattern of repeated sexual relationships in-
volving a succession of lovers who are experienced by the individual only as
things to be used.” (p. 538; Nonparaphilic CSB involves normative sexual
behavior that is engaged in an excessive and compulsive manner; Coleman,
1991).
There have been a few attempts to develop scales to identify individuals
with CSB (e.g., Carnes, 1991; Kafka, 1997; Kalichman, Adair, Rompa, Multhauf,
Johnson & Kelly, 1994; Kalichman & Rompa, 1995), with various limitations
in methodology and abilities to generalize.
The previous attempts to develop a scale of CSB failed to incorporate all
of the major components of this clinical phenomenon. Therefore, the first
author developed a scale that consisted of items related to sexual control
and items measuring various aspects of behavior that have been found clini-
cally to be associated with CSB—both paraphilic and nonparaphilic types.
This scale was developed to create a standardized, reliable, and valid
tool to assist clinicians in distinguishing individuals who meet a threshold of
clinical pathology of CSB as well as to offer a tool to use in future studies of
this phenomenon. We hypothesized that the paraphilic and nonparaphilic
CSB groups would not differ significantly in terms of this measure of CSB but
would differ significantly from control subjects.
METHODS
Participants
For the paraphilic group, we elected to study a group of individuals who
had a diagnosis of pedophilia and who were recruited from sexual-offender
treatment programs. All participants were evaluated by one of the authors
for the presence or absence of pedophilia.
Compulsive Behavior Inventory 327
Thirty-five men were enrolled in the study. The average age was 36
years (Caucasian: 31 [89%]; African American: 3 [9%]; Other: 1 [2%]). Thirteen
(37%) were not working because of restriction of residential treatment; 8
(23%) were laborers; 5 (14%) were technical workers; 4 (11%) worked for a
small business; 2 (6%) were clerical workers; and 3 (9%) were unemployed.
Twenty-four (69%) had completed high school; 6 (17%) had a college de-
gree; and 5 (14%) had not completed high school.
Participants with nonparaphilic CSB were recruited through advertise-
ments in local newspapers. Potential participants that met the following cri-
teria were enrolled in the study:
Downloaded by [Central Michigan University] at 11:33 02 December 2014
Fifteen male participants met criteria and were enrolled in the study (All
participants were evaluated by one of the authors. The average age of the
participants was 38 years (Caucasian: 14 [93%]; African American: 1 [7%]).
Four (27%) identified themselves as technical workers, four (27%) as clerical
workers, 3 (20%) as professionals or executives. Of the remaining 4 partici-
pants, 2 were unemployed, 1 worked in small business, and 1 was retired.
Six (40%) had a college degree, 6 (40%) had partial college training, 2 (13%)
had graduate or professional degrees, and 1 participant had not completed
high school.
Control participants also were recruited through newspaper advertise-
ments. Those reporting prior diagnosis of psychological disorders were ex-
cluded. Forty-two male controls were enrolled in the study. The average age
of the 42 participants was 43 years (Caucasian: 40 [95%]; African American: 1;
Other: 1). Thirty (71%) described themselves as being professionals or ex-
ecutives, and 5 (12%) described themselves as clerical workers. Of the re-
maining 7 participants, 3 were technical workers, 2 were laborers, 1 was
unemployed, and 1 participant’s employment data was not collected. Eigh-
teen (43%) had completed high school, 13 (31%) had a college degree, and
11 (26%) had some postgraduate training.
Procedures
This inventory consisted of 42 items related to sexual control and various
aspects of behavior that have been found clinically to be associated with
CSB—both paraphilic and nonparaphilic types. Participants were asked to
328 E. Coleman et al.
Statistical Analysis
A factor analysis was performed using the principal components method for
determining factor solution, followed by a varimax rotation. Factor retention
was determined using the scree criteria, and items with factor loadings ex-
ceeding 0.60 on the rotated factors were retained for the final scale. The
reliability of the retained factors then were tested using Cronbach’s alpha.
Downloaded by [Central Michigan University] at 11:33 02 December 2014
Then the data were exposed to a linear discriminant function analysis that
explored the ability to separate the CSB individuals from the controls. Fi-
nally, three ANOVAs were performed to explore mean differences for each
of the study groups on the identified subscales.
RESULTS
Indications of Validity
The initial test of the validity of the above three-factor scale was to expose
the data to a linear discriminant function analysis. We tested the ability of the
scale to distinguish between groups considered to have issues of CSB from
those who did not. Thus, the pedophile and nonparaphilic sexual compulsives
were combined into one group (sexual compulsives). The classification ma-
Compulsive Behavior Inventory 329
How often have you had trouble controlling your sexual urges? .85
Have you felt unable to control your sexual behavior? .85
How often have you used sex to deal with worries or problems in your life? .84
How often have you felt guilty or shameful about aspects of your sexual
behavior? .83
How often have you concealed or hidden your sexual behavior from others? .82
How often have you been unable to control your sexual feelings? .82
How often have you made pledges or promises to change or alter your sexual
Downloaded by [Central Michigan University] at 11:33 02 December 2014
behavior? .81
How often have your sexual thoughts or behaviors interfered with the formation
of friendships? .72
How often have you developed excuses and reasons to justify your
sexual behavior? .78
How often have you missed opportunities for productive and enhancing activities
because of your sexual activity? .78
How often have your sexual activities caused financial problems for you? .69
How often have your felt emotionally distant when you were engaging in sex
with others? .68
How often have you had sex or masturbated more than you wanted to? .67
Factor 2: Abuse
Eigenvalue = 4.26
Factor 3: Violence
Eigenvalue = 2.46
Have you ever hit, kicked, punched, slapped, thrown, chocked, restrained,
or beaten any of your sexual partners? .83
Have you given others physical pain for sexual pleasure? .75
In fighting, have you been hit, kicked, punched, slapped, thrown, chocked,
restrained, or beaten by your current or most recent partner? .74
Have you received physical pain for sexual pleasure? .68
Have you received money to have sex? .65
Have you been forced to have sex with your husband, wife, or lover? .65
Have you been watched masturbating or having sex without giving permission? .63
330 E. Coleman et al.
trix indicated that 92% of the cases were correctly classified, with one of the
normal controls incorrectly classified as compulsive and six compulsives
classified as controls.
In order to further explore the validity of this scale, the three subscales
were used as dependent variables in ANOVAs exploring group differences
between the three study groups. A significant effect for group was found for
the control subscale, F3,99 = 33.2. p < 0.0001. Pairwise comparisons indicated
that pedophiles scored significantly lower than the other two groups on this
subscale (see Table 2), whereas none of the other groups differed from one
another. Finally, there was a significant overall effect for the violence subscale,
F2, 89 = 4.48, p < 0.02. Pairwise comparisons indicated that controls were
Downloaded by [Central Michigan University] at 11:33 02 December 2014
significantly higher than pedophiles, but that compulsives did not differ from
either controls or pedophiles.
DISCUSSION
controls was only partially confirmed. Clearly, the overall scale correctly
classified 92% of the overall sample into CSB and control. The two CSB
groups differed from controls on the basis of the control subscale but not in
terms of the abuse or violence subscales of the CSB Inventory. The majority
of the variance of the scale was accounted for by the control subscale. When
we examine the subscales and the kinds of behaviors that the pedophiles
and nonparaphilic CSB participants have been engaged in, it makes sense
that the nonparaphilic CSB group is more similar to the controls.
Although our data offer evidence of criterion-related validity, they are
postdictive in nature. We administered the CSB Inventory and used test scores
to determine membership in intact groups. This method appears appropriate
in constructing a test to be used for diagnostic purposes.
This scale was developed on a small and select sample of men. Al-
though the sample size is well within the range recommended for multivari-
ate analysis of number of variables plus 50 (Harris, 1985), the more common
convention is 5 to 10 subjects per variable. The small sample sizes and meth-
ods of recruitment limit generalizability of this study. Certain factors, includ-
ing disparities in education between the samples, may explain some of the
differences between groups, although this seems unlikely because of the
nature of the differences. Furthermore, the paraphilic sample was rather
homogeneous, at least with respect to behavior, whereas the nonparaphilic
group was more heterogeneous. There clearly is a need for cross-validation
with large, more diverse, and independent samples to replicate our factor
structure, refine the instrument, and further explore its validity.
REFERENCES
Harris, R. J. (1985). A primer of multivariate statistics (2nd ed.). Belmont, CA: Duxbury
Press.
Kafka, M. P. (1997). Hypersexual desire in males: An operational definition and
clinical implications for males with paraphilias and paraphilia-related disorders.
Archives of Sexual Behavior, 26, 505–526.
Kalichman, S. C. & Rompa, D. (1995). Sexual sensation seeking and sexual
compulsivity scales: Relability, validity, and predicting HIV Risk Behavior. Jour-
nal of Personality Assessment, 65, 586–601.
Kalichman, S. C., Adair, V., Rompa, D., Multhauf, K., Johnson, J., & Kelly, J. (1994).
Sexual sensation seeking: Scale development and predicting AIDS-risk behav-
ior among homosexually active men. Journal of Personality Assessment, 62,
385–397.
Downloaded by [Central Michigan University] at 11:33 02 December 2014