See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/21598484

The proportions of heterosexual and homosexual pedophiles among sex

offenders against children: An exploratory study

Article in Journal of Sex and Marital Therapy · February 1992

DOI: 10.1080/00926239208404356 · Source: PubMed

CITATIONS READS

37 57,094

2 authors, including:

Robin J Wilson
McMaster University
76 PUBLICATIONS 1,917 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Circles South East View project

Passport to Independence View project

All content following this page was uploaded by Robin J Wilson on 25 April 2015.

The user has requested enhancement of the downloaded file.

This article was downloaded by: [University of Toronto]
On: 15 December 2010
Access details: Access Details: [subscription number 911810122]
Publisher Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-
41 Mortimer Street, London W1T 3JH, UK

Journal of Sex & Marital Therapy

Publication details, including instructions for authors and subscription information:
http://www.informaworld.com/smpp/title~content=t713723519

The proportions of heterosexual and homosexual pedophiles among sex

offenders against children: An exploratory study
Kurt Freunda; Robin J. Watsona
a
Department of Behavioural Sexology, Clarke Institute of Psvchiatrv,

To cite this Article Freund, Kurt and Watson, Robin J.(1992) 'The proportions of heterosexual and homosexual pedophiles
among sex offenders against children: An exploratory study', Journal of Sex & Marital Therapy, 18: 1, 34 — 43
To link to this Article: DOI: 10.1080/00926239208404356
URL: http://dx.doi.org/10.1080/00926239208404356

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial or
systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or
distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss,
actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly
or indirectly in connection with or arising out of the use of this material.
 The Proportions of Heterosexual and
Homosexual Pedophiles Among
Sex Offenders Against Children:
An Exploratory Study
K U R T FREUND mid ROBIN J WATSON
Downloaded By: [University of Toronto] At: 20:52 15 December 2010

Preiiiou.s investigations have indicated that the r d i o of ,sex offenders

aguinst female children us. offenders against mule children is upproxi-
mately 2:l, ruhile the ratio of gynephiles to androphiles among the
gene&l population is approximately 20:l. The present study investi-
gated whether the etiologV oj.preferred partner sex nmong pedophiles is
relatecl to tile etiology ofjmjerred partner sex among males preferring
adult partners. Using ptiallometric test sensitivities to calculate the pro-
portio" of ti-ue pedophiles among ~ U T ~ groups ~ I L ~of sex offenders
aguinst children, and taking into consideration previously reported
meun numbers of uictims per offender group, the ratio of' heterosexuul
to Iiomosexital pPdophiles iuas culczclnted to be npproxiniately I 1 :I.Tfiis
sugge.sts thut the resulting proportion of true pedophiles arnong persons
uiith (1 homosexual erotic deuelopnient is greuter thnn that i n persons
uiho deuelop heterosexuully. This, of course, zuould not indicate that
anclrophilic males t i m e a &wt~I propensity to offend against children,.

As part o f an ongoing investigation of potential epidemiological connec-

tions between gynephilia (an erotic preference for physically mature fe-
males), androphilia (an erotic prcfcrence for physically mature males),
and heterosexual and homosexual pedophilia. the following study is an
exploration of differences between heterosexual and homosexual pedo-
philia. T h e reco<nition o f pedophilia as a specific disorder is attributable
t o Kraf ft-Ebing? howe\Ter, some authors continue to use this term to
''
denote a variety of dif'terent clinical pictures.', In females, pedophilia
is either very rare' o r virtually nonexistent.

Journal of Sex & Marital Therapy, Vol. 18, No. 1, Spring 1992 6 BrunnerIMazel, Inc.
34
 T h e revised third edition of the Diagnostic and Statistical Manual of
Mental Disorders5 lists the essential feature of pedophilia as “recurrent,
intense, sexual urges and sexually arousing fantasies of at least six months
duration, involving sexual activity with a prepubescent child” (p. 284).
Such fantasies are indeed one of the foremost characteristics of pedo-
philia; however, relying o n self-reports of fantasies is of little diagnostic
value since pedophiles rarely admit their true erotic fantasies. I n the
following, the term “prepubescent,” used to denote the period of child-
hood, is defined as ending at age 11. This age limit was adopted from
Gebhard, Gagnon, Pomeroy, and Christensen.6
Phallometric studies (using continuous recording of penile volume
changes at presentation of pictures of nude female and male children
and adults) have demonstrated that a majority of sex offenders who
victimized two or more children (unrelated to them) respond more to
Downloaded By: [University of Toronto] At: 20:52 15 December 2010

movies a n d slides of nude children than to pictures of nude physically

mature person^.^-^ They erotically prefer children to adults.
T h e definition of pedophilia as an erotic preference and the assessment
of this preference by means of the phallometric test make possible diag-
nosing a much larger proportion of sex offenders against children as
pedophiles (with sufficient specificity) than usin these patients’ self-re-
ports. This has been demonstrated previous1y.lf ” For this reason, the
present study used the phallometric test as the differential diagnostic
indicator of pedophilia.
I n an earlier study,” of which the present study is a continuation, the
incidence of (male) sex offenders against female children was compared
with the incidence of such offenders against male children. T h e propor-
tions of these two offender groups was approximately 2: 1 , omitting indi-
viduals who offended against children of both sexes. This was in
agreement with earlier findings6*Is Furthermore, a recent literature
’ ~ involved 17 additional studies on sex offend-
search by C a m e r ~ n ,which
ers against children, listed the ratio of victimized female to male children
in the majority of cases also as approximately 2: 1. This differs substan-
tially from the ratio of gynephiles to androphiles (among males who
erotically prefer physically mature partners), which is approximately
20:1.15- *
T h e discrepancy between the proportions of sex offenders against fe-
male children and such offenders against male children o n the one hand,
and of gynephiles and androphiles o n the other, was unexpectedly large.
Therefore, we supposed this difference to be robust enough to indicate,
even at this superficial level of information, to warrant further research
in this direction.
Abel, Becker, Mittelman, Cunningham-Rathner, Rouleau, a n d Mur-
phy” have demonstrated that nonincestuous offenders against male chil-
dren have many more victims than such offenders against female chil-
dren. According to Table 2 of that study, the mean number of victims
of nonincestuous offenders against female children was 19.8, while that
of such offenders against male children was 150.2. It has been convinc-
ingly argued that the much larger number of victims of offenders against
 36 JOILT?T(I/ (4 Sex &f iClo~-ttnlTtirrtipy, Vol. 18, No. 1, Spring 1992

male child]-en increases the risk of these offenders being caught."'. I'
However, the rate of the increase of this risk with the number of victims
is unknown.
T h e tendency of' the group of offenders against male children toward
a greater number of victims suggests that this group may contain a larger
proportion of true pedophiles than the group of offenders against female
children. 3'his is in agreement with the results of an earlier study.''
In addition to determining the specificity of the phallometric test of
pedophilia, this study assessed the test's sensitivity using groups of nonin-
C C S offenders
~ who victimized t w o or more children as index cases. These
multiple victimizer groups can be supposed to consist of a large majority
of true pedophiles. O u r clinical estimate was that 90% of these offenders
against female children and 95% of these offenders against male children
Downloaded By: [University of Toronto] At: 20:52 15 December 2010

were true pedophiles. Under this supposition, the proportion of individ-

uals phallometrically diagnosed as pedophiles could be used as an indica-
tor of' the sensitivity of the phallometric method. This in turn made
possible calculation of an estimate of true pedophiles in groups of sex
offenders of differing compositions a n d among offenders against only
one child. O n the basis of similar estimates, the present investigation
converted the proportions of sex offenders against female children vs.
such offenders against male children into a n estimate of the proportion
of heterosexual vs. homosexual pedophiles.

METHOD

Subjects were 463 nonpsychotic sex offenders against female or male

6-1 1 -year-old children (omitting those who had offended against chil-
dren of both sexes), aged 18-55 years inclusive. O f the 273 offenders
against female children, 79 were accused of or charged with offenses
against only one child, 116 for offenses against t w o or more children,
and 78 for offenses against at least one female child and at least one
female pubescent. Of the 192 offenders against male children, 50 were
accused of or charged with an offense against only one child, 69 were
accused of or charged with offenses against at least t w o children, and 73
were accused of or charged with offenses against at least one male child
and at least one male pubescent. Table 1 demonstrates mean age, median
education, a n d median Hollingshead-Kedlich index of occupational sta-
tus.'l All individuals in the study were referred to us by psychiatrists who
were assessing them either for lawyers (the majority of these subjects) o r
within the prison system. A few cases w e r e referred to us by the Chil-
dren's Aid Society o r by probation officers. Almost all of the individuals
were white. T h e overlap of the prescnt sample with the earlier sample''
is demonstrated in Table 1.
Following the completion of a study o n incest," we n o longer saw
clients who had offended against female children within only a daughter
 Proportions of Heterosexual and Homosexud pedophiles 37

TABLE 1
Group Characteristics
Offenders Against Female Children Ot'fcnders Against Male Children

Characteristic single multiple ch + p u b single multiple ch + pub

N 79 116 78 50 69 73
M e a n age* 32.7.'." 36.1' 37.3' 31 3'L,'L.'1) 34,6'l>L
SD age 10.2 9.2 9.4 8.9 10.0 Y.ti
Median education 8<Ed<12 8 < E d < 1 2 8<Ed<I:! 8<E:d<12 Il<Ed<l2 8<Ed<l2
Median occupation** USL USL. USL SSL USI. SSI.
N for occupation 64 84 60 45 59 44
Mentally h a n d i c a p p e d 8 10 3 8 19 8
Partial a d m i t t e r s 9 11 18 3 6 5
Full a d m i t t e r s 1 21 19 12 20 44
Number in F & W, 1991 20 29 I1 7 14 9
Downloaded By: [University of Toronto] At: 20:52 15 December 2010

Nofr:Pairs of figures sharing like suprascripts are not significantly different at /I < .05. Single = charged for o r
accused of victirniring one child: multiple = charged for- or accused of victimizing two or more childi-en; ch +
puh = chai-ged for or accused of victimi~ingat least one child o d a t least one pubescent. SSL = semi-skilled
labor: USL = unskilled labor. N for occupation = numbei- of individuals for whom occupational statlib was
known. Partial adrnitters = individuals who admit to having some sexual fantasies about minors; Full admitters =
individuals who admit to a preponderance of sexual fantasies about minors. Number in F & W. 1991 = number
of subjects included previoiisly in Frrund & Watson (1991)."
'F(5.459) = 4.61. p < .001
**liollingshead & Redlich"

or stepdaughter relationship (the "sole-incest offenders"). For this rea-

son, we did not have an analogous group of sole-incest offenders, and it
was necessary to add an imaginary group of such offenders to the present
sample of 273 nonincest offenders against female children. T h e size of
this imaginary group was derived in accordance with the proportion
reported in Abel et a1.20T h e procedure is described below.

Procedure

Construction of the Group of Sole-Incest Offenders Against Female Children.

In the study by Abel et aL2"there were 224 individuals who had offended
against female children in a nonincestuous context and 159 incestuous
offenders against female chldren. There was, however, an overlap. O f
the 224 individuals with nonincestuous offenses, 78 also had incestuous
offenses. When subtracted from the group of incestuous offenders, this
left a total of 81 sole-incestuous offenders. This is 36.2% (81 of 224) of
the number of the offenders other than sole-incestuous (against female
children). Therefore, we added 98 (approximately 36% imaginary sole-
incestuous offenders to our sample of the 273 offenders against female
children in a context other than sole-incest. T h e few incest offenders
against male children who had no nonincest offenses have always been
treated in our studies as nonincest cases; therefore, no imaginary sole-
incest group of offenders against males was necessary.
Conversion of Porportions of Offenders Against Female Children us. Sex Of-
fenders Against Male Children into Estimates of Proportions of Heterosexual us.
Homosexual Pedophiles. This conversion was carried out in t w o ways: 1)
 according to the proportions of pedophiles in the various offender
groups, arrived at in the previous study,’’ and 2) by recalculating the
sensitivities of the phallometric test for heterosexual and homosexual
pedophilia according to data in the present study, and using these esti-
mates to calculate proportions of heterosexual and homosexual pedo-
philes among sex Offenders against children. In these calculations two
formulas (see formulas 1 and 2 below) were used.
In the study cited above as well as in the present study, sensitivities of
the phallometric method for heterosexual and homosexual pedophilia
were assesscd according to formula 1: SPTZ.F = a / ( b x c), in which a is
the number of individuals diagnosed as having an erotic preference for
minors, b is the total number of individuals who received a valid phallo-
metric diagnosis, and c is the clinical estimate of the number of individuals
in the pertinent niultivictini offender group (stated above as 90% for
Downloaded By: [University of Toronto] At: 20:52 15 December 2010

multivictim offenders against female children a n d 95% for niultivictim

offenders against male children).
T h e expected proportion (c’) of pedophiles in the various other groups
was estimated according to formula 2: c ’ = (L/ ( b x S P R S ) . In this expres-
sion seus is the sensitivity calculated for the pertinent multivictim offender
group (see above). A more detailed description of the diagnostic proce-
dures is given in the paper cited above, a n d a full technical description
is available on request.
T h e statistical comparison of’ the offender groups in regard to mean
age was by analysis of variance and Tukey’s honesty significant difference
tests; the comparisons in regard to education a n d occupational status
were by corresponding nonparametric tests (i.e., Kruskall-Wallis and
M an n -W hit n e y U -test ) .

RESULTS

Table 1 lists the mean age, median education, and the occupational status
of the groups of offenders. T h e offenders against at least one female
child and at least one female pubescent were significantly older than the
offenders with accusations or charges involving only one male or female
child, and offenders with charges involving at least t w o male children.
T h e offenders against at least t w o female children were significantly
older than the offenders with accusations or charges involving only one
male child. T h e r e were n o further significant differences with respect to
age, and no significant differences k t h respect to either other variable.
Table 2 demonstrates the number of offenders against female children,
offenders against male children, a n d the number a n d proportion of het-
erosexual and homosexual pedophiles in the various offender groups.
T h e ratio of offenders against female children to offenders against male
children was 1 .Y3: 1. However, the earlier cited finding of Abel et al.20of
a 7.6 times larger mean number of victims of offenders against male
children than the mean number of victims of offenders against fcmale
children also has to be considered. If the risk of an offender being caught
call be expected to increase proportionally and linearly with each victim,
 Proportions of Hetero.texua1 and Homosexucd Pedophiles 39

TABLE 2
Estimated Proportion of Pedophiles
Using Estimates from Using New Sample
1991 Study Estimates
Total
N % N Q N
Offenders Against
Female Children
Single victim 79 44.5 35 61.6 48
Multiple victims 116 90.0 104 90.0 I04
Ch + Pub victims 78 74.6 58 81.5 63
Incest offenders 98 25.0 24 25.0 24
Total 37 1 nla 22 1 nla 239
Offenders Against
Downloaded By: [University of Toronto] At: 20:52 15 December 2010

Male Children
Single victim 50 86.7 43 76.4 38
Multiple victims 69 95.0 65 95.0 65
Ch + Pub victims 73 86.7 63 100.0 73
Total 192 nla 171 nla 176

Nole. Numbers of heterosexual and homosexual pedophiles were calculated using the estimated incidence pi-opoi -
lions from Freund and Watson” and from data in the current study. Ch +Pub victims = oftenders against at
least one child under I 1 ond at least one pubescen~aged 12-14.

the ratio of the offenders against female children vs. offenders against
male children would have to be recalculated to 14.7: I (1.93 X 7.66).
When incorporating the proportions of true pedophiles, this ratio
changes. T h e sensitivities of the phallometric test of pedophilia calcu-
lated in the earlier study (cited above) by means of formula 1 were
78.2% for heterosexual pedophilia and 88.6% for homosexual pedo-
philia. T h e calculation of the sensitivities on the basis of data in
the present sample rendered similar results:76.7% for heterosexual
pedophilia and 84.9 % for homosexual pedophilia.
In o u r study,” the calculated proportions of pedophiles in these two
groups (formula 2, see Method), were 24.6% pedophiles among the
sole-incest offenders against only one female child ( n = 21) and 25.6%
pedophiles among the sole-incest offenders against t w o or more female
children (n = 74).
I n the earlier study, no estimates for groups of victimizers of only
o n e male child and of victimizers of at least o n e male child and at
least one male pubescent could be made because the numbers were
too small. Therefore, the present study used the same estimate for
victimizers of one male child or one male pubescent from the earlier
study to estimate the number of pedophiles in the group of offenders
with charges or accusations for victimizing one male child a n d the
group of offenders against at least one male child and at least one
male pubescent.
If assessed according to conversion method 1 (see above), the ratio
of heterosexual vs. homosexual pedophiles in the present sample was
1.35:l. If assessed according to conversion method 2, this ratio was
 1.44: 1. When considering the 7.6 factor of Abel et al., the two ratios
of heterosexual vs. homosexual pedophiles, according to methods 1
and 2 above, would increase to 10.26:1, and 10.94:l.

DISCUSSION

T h e present study is part of a project that addresses potential etiological

differences and connections between gynephilia, androphilia, a n d het-
erosexual and hornosexual pedophilia. An earlier study assessed the
ratio of the incidences of sex offenders against female children vs. sex
offenders against male children.” This ratio (approximately 2: 1) dif-
fered drastically f‘rom the ratio of‘ gynephiles vs. androphiles (approxi-
mately 20: 1 ) .
Because not all sex offenders against children are pedophiles, the
Downloaded By: [University of Toronto] At: 20:52 15 December 2010

present study investigated whether this discrepancy reflected an analo-

gous discrepancy between the ratio of the incidence of heterosexual
and homosexual pedophiles a n d the ratio of gynephiles vs. androphiles.
T h e assessment of “true” pedophilia was carried out by means of a
phallometric test procedure validated for this purpose.’ T h e present
study converted a ratio of offenders against female children vs. offend-
ers against male children into the ratio of heterosexual vs. homosexual
pedophiles .
T h e study also attempted to take into account the argument of Abel
et al.’!’. 2o that the much larger number of victims of offenders against
male children increases the risk of these offenders being caught. Abel
et al. did not estimate this risk increase (represented by the number
of offenders caught) as a function of the reported number of victims.
However, the numerous social factors entering the relationship between
the number of victims a n d the risk of being apprehended at present
permit only a tentative assessment. For this reason, the present study
used an estimate for this increase that, on the basis of clinical experi-
ence, appeared to be realistic. ’This correction brought the ratio of
heterosexual vs. homosexual pedophiles to approximately 1 1: 1.
There does not appear to be any indication that the discussed
differences between offender groups were d u e to differences in eth-
nicity, age, education, or occupational status. An interesting finding
was that in contrast to nonincest cases, the proportions of pedophiles
among incestuous single victimizers and multiple victimizers of female
children appear to be equal, which is in agreement with the notion
that a majority of these incest offenders are not pedophiles or, in
other words, that in incest offenses against female children causes
other than pedophilia are in the foreground.
T h e question might be asked why, at a stage where in this area of
research there is still a dearth of knowledge a n d where the few available
methods are still very crude, efforts are being made to test the validity
of clinical or epidemiological impressions. It appears, however, that
this should not be a reason to leave observed asymmetries unexplored,
as long as the methods used in such explorations are sufficiently clear
 Proportions of Heterosexual and Homosexual Pedobhiles 41

to make possible relatively unambiguous replication. This holds true

in particular for etiologically oriented research into pedophilia where
the existing etiological theories are basically untestable. T h e reason is,
as demonstrated by Garland and Dougher,2” that the observational
basis of these theories is either highly unreliable or n ~ n e x i s t e n t . ~ ~ . *;
At the present state of our knowledge in regard to the paraphilias,
the method of “strong inference” inductive approach26 used in the
more developed sciences (e.g., molecular biology) can, however, only
be crudely emulated. This may be accomplished by devising optimal
crucial comparisons and multiple hypotheses that could make possible
(in the present case very preliminary) simultaneous refutation of more
than one hypothesis, serving to maximally narrow the number of the
still unrefuted possibilities as much as possible.
Seen from this point of view, the present study may be interpreted
Downloaded By: [University of Toronto] At: 20:52 15 December 2010

as testing simultaneously three alternative hypotheses: 1) heterosexual

pedophilia is equally related etiologically to gynephilia as homosexual
pedophilia is related to androphilia; 2) with a heterosexual develop-
ment, the resulting proportion of pedophiles is greater than with a
homosexual development; 3) with a homosexual development, the
resulting proportion of pedophiles is greater than with a heterosexual
development.
T h e typical ratios for heterosexual vs. homosexual pedophilia, which
would be expected for each of these three hypotheses if valid, are
20: 1, a ratio notably larger than 20: 1, and a ratio notably smaller than
20: 1. T h e possibility that there is no etiological relationship whatsoever
between pedophilic gender preference and gynephilia or androphilia
cannot be excluded. Barring other factors, the most convincing support
of this possibility would be a ratio of 1 : l .
Support was given by these results to hypothesis 3, according to
which a homosexual development notably often does not result in
androphilia but in homosexual pedophilia.’* Hypothesis 2 was weak-
ened most; also weakened was hypothesis 1. This, of course, should
not be understood as saying that androphiles may have a greater
propensity to offend against children than d o gynephiles, a myth
refuted in an earlier
T h e present results may be very temporary, either because subse-
quent estimates based on much greater numbers of subjects may be
different, or because better methods of calculating these estimates may
be found available in the near future and may not lead to the
same conclusions as the present investigation. Nonetheless, the study
reported above does not appear to be redundant as a first basis
of orientation and challenge for constructing etiological theories of
pedophilia.

REFERENCES

1 . Krafft-Ebing K von: Psychopathia sexualis: u medico-forerisic study. New York, Pioneer,

1950/orig 1886.
 32 .Journal of Sex &? ~VlnrilalTherapy, Vol. 18, No. I , Spi-uig I992
~~

2. Ames M , Houston D: Legal. social, a n d biological definitions of pedophilia. Arch

Sex Behuu 19:333-342, 3990.
3 . Araji S. Finkelhor D: Explanations o f pedophilia: review of empirical research.
Bull A m u Arud P.rychiat Laio 13:17-37, 1985.
4. Finkelhor L): Child sexual abuse: new theon‘ miid mearch. New York, Free Press,
1984.
5 . A ni e r ica 11 Psychiatric Association : Diognos l i c a r t d sl&ticri I rrtcinuri 1 of vientol disorders
(3rd ed, rev). Washington. DC. APA, 1987.
6. Gebhard P, Gagnon J , Pomeroy W, Christenson C: Sex c4feuder.r: an analy.szs of
lypeJ. New York, H a r p e r & Row, 1965.
7. Freund K : Diagnosing heterosexual pedophilia by means of a test of sexual
interest. Behau Res Ther 3:229-234, 1965.
8. Freund K: Diagnosing homo- o r heterosexuality a n d erotic age-prcfcrcnce by
means of a psychophysiological test. Beliazp Krs Ther 5:209-228, 1967.
9. Quiiisey \’, Steinnian C , Bergerson S, Holmes T: Penile circumference, skin
Downloaded By: [University of Toronto] At: 20:52 15 December 2010

conductance, a n d ranking responses o f child molesters a n d “normals” to sexual

a n d nonsexual visual stimuli. B e h m Ther 6:213-219, 1975.
10. Freund K, Blanchard R: Phallometric diagnosis o f pedophilia. J Coi~siiltClin Psychol
57:1-6, 1989.
1 1 . Freund I(, Watson RJ: Assessment of the sensitivity a n d specificity.of a phallometric
test: a n update of “Phallometric diagnosis of pedophilia. Psychol Assess 3:254-260,
1991.
12. Freund K , Heasman G, Racansky lC, Glancy G: Pedophilia a n d heterosexuality
vs. homosexuality. J Sex M u ~ i t a lTtier 1 0 : 193-200, 1984.
13. Mohr W, T u r n e r RE, Jerry MB: Pedophilin atid exhibitiotiisrn. University of l o r o n t o
Press, 1964.
14. Cameron P: Homosexual molestation of children: sexual interaction of teacher
a n d pupil. Psycho1 Rep 57: 1227-1236, 1985.
15. Gebhard P: Incidence of overt homosexuality. In J M Liningood (ed), I n d l u t e uf
mellta/ health task force on homoc-exuciiifv. Rockville, MD. National lnstitiite of Mental
Health, 1972.
16. Hirschfeld M: Die Honiosexiiulilal des hlatiries und des Weibes [Homosexuality in men
a n d women] (2nd ed). Berlin, Marcus, 1920.
17. Kirisey AC, Pomeroy WB, Martin CE: S e x u d behavior in the human mule. Philadel-
phia, Saunders, 1948.
18. Whitam FL,: Culturally invariable properties of male homosexuality: tentative
conclusions from cross-cultural research. Arch Sex Behnu 12:207-226. 1983.
19. Abel C;, Becker J , Mittelman M, Cunningham-Rathner , J , Rouleau J , Murphy W:
Self-reported sex crimes of nonincarcerated paraphiliacs. J Interpersonal Violence
2:3-25, 1987.
20. Abel G , Becker J . Cunningharn-Rathner J , Mittelman M, Rouleau J : Multiple
paraphilic diagnoses a m o n g sex offenders. Bull Atner Acad P.cychial Law 16: 153-168,
1988.
21 Hollingshead A, Redlich F: Sociul class aizd rrterital Llln~.ss: a community study. New
Y o r k , Wilcv, 19%.
‘,‘,
LL. Freund K.’Watson R, Dickey R: Sex offenses perpctratctl by m e n who are not
peclophiles. ,/ Sex Kes, in press.
23. Garl;rnd K, Dougher M: The abusecl/abuser hypothesis of‘ child abuse: a critical
review of theorv a n d research. In J Feiernian (ed), Pedophiba: bio.tocirr1 dzrrteiision.r.
Ne\v Y o r k . Springer-Verlag, 1990.
24. Hindrnan ,I: Research disputes assumptions about child rnolesters. NDAA Bull
7: 1-3. l‘.)88.
 Proporlions of Helerosexual mid Homosexual Pedophilec 43

25. Freund, K , Watson, R , Dickey, R: Does sexual abuse in childhood cause petlo-
philia?: An exploratory study. Arch Sex Behav 19:557-568, 1990.
26. Platt ,J: Strong inference. Science 146:347-353, 1964.
27. Freund, K , Watson, R, Rienzo, D: Heterosexuality, homosexuality, and erotic age
preference. J Sex Res 26:107-117, 1989.
Downloaded By: [University of Toronto] At: 20:52 15 December 2010

View publication stats