Page 1
The Journal of Sex Research Vol. 26, No. 1, pp. 107-117 February, 1989
Heterosexuality, Homosexuality, and
Erotic Age Preference
KURT FREUND, M.D., D.Sc. ROBIN WATSON, B.Sc.
DOUGLAS RIENZO, B.Sc.
Department of Behavioural Sexology, Clarke Institute of Psychiatry
Heterosexual and homosexual males who erotically preferred physically
mature partners were compared with respect to the erotic impact of the
nonpreferred age groups (of the preferred gender) and the nonpreferred
gender. Erotic impact was assessed by phallometric test of erotic gender
and age preferences. This measures penile volume changes during the
presentation of potentially erotic stimuli. Homosexual males who pre-
ferred physically mature partners responded no more to male children
than heterosexual males who preferred physically mature partners
responded to female children. Although heterosexual males showed a
(slight) erotic aversion to the male stimuli, and homosexual males did
not show an aversion to the female stimuli, this difference was minimal.
KEY WORDS: penile responses, homosexuality, heterosexuality, age
preference
(I) The Erotic Impact of the Nonpreferred
Age Group of the Preferred Gender
An earlier study assessed the incidence of male sex offenders against
female children vs. such offenders against male children (Freund,
Heasman, Racansky, & Glancy, 1984). Approximately one-third of
these individuals had victimized boys and two-thirds had victimized
girls. This finding is consistent with the proportions reported in two
earlier studies (Gebhard, Gagnon, Pomeroy, & Christenson, 1965;
Mohr, Turner, & Jerry, 1974), Also, in a recent literature search
(Cameron, 1985) which examined 17 more studies on sex offenders
against children, the ratio of victimized female to male children was
approximately 2:1. Interestingly, this ratio differs substantially from
the ratio of gynephiles (men who erotically prefer physically mature
females) to androphiles (men who erotically prefer physically mature
males), which is at least 20 to 1 (Gebhard, 1972; Hirschfeld, 1920;
Kinsey, Pomeroy, & Martin, 1948; Whitam, 1983).
The authors wish to thank Cathy Spegg, B.S., M.B.A., of the Computer Services
Department at the Clarke Institute for her assistance with data organization and
statistical advice.
Requests for reprints should be sent to Dr. Kurt Freund, Department of Behavioural
Sexology, Clarke Institute of Psychiatry, 250 College Street, Toronto, Ontario, Canada,
M5T 1R8.
107
Page 2
108 FREUND, WATSON, AND RIENZO
Such issues led to two systematic investigations of the erotic appeal
of children for males who eroticaJly prefer physically mature persons.
In both investigations, the phallometric arousal test of gender and age
preferences was used. This test assesses erotic preferences hy
measurement of penile volume changes during the presentation of
visual and auditory stimuli representing various gender-age categories
(Freund, 1963, 1967; Freund, Diamant, 8c Pinkava, 1958),
The first study (Freund, McKnight, Langevin, & Cibiri, 1972)
assessed the erotic value of minors for gynephiles. The results indi-
cated that for a substantial proportion of gynephiles the erotic impact
of pubescent girls is quite strong, and that the female child also has
significant, albeit weak, erotic arousal potential.
The second study (Freund, Langevin, Cibiri, & Zajac, 1973) com-
pared gynephiles and androphiles on the erotic impact of children,
pubescents, and adults of the preferred as well as of the nonpreferred
gender. There were no differences between the two subject groups in
respect to responses to adults, pubescents, and to 6- to 8-year-old
children; however, responses of gynephiles to 8- to 11-year-old female
children turned out to be larger than responses of androphiles to male
children of that age. While this difference was significant, it was very
small.
Unfortunately, in these earlier studies there were no neutral stimuli.
Thus, comparison of gynephiles and androphiles (on responses to
potentially erotic stimulus categories) rests on the assumption that
these groups did not differ in penile responses to neutral stimuli. How-
ever, if this assumption is incorrect, the small difference between the
groups' responses to children of the preferred gender would not be
reliable. Therefore, the present study re-investigated this issue (with
new groups of gynephiles and androphiles), using a new version of the
phallometric test of age and gender preferences (Freund & Blanchard,
in press) and a stimulus set which included neutral stimuli
(landscapes).
(IIj The Erotic Impact of the Nonpreferred Gender
In contrast to gynephiles, a substantial number of androphiles claim
to be "bisexuals" and are often in heterosexual marriages. However,
Krafft-Ebing (1986/1965), Hirschfeld (1920), and Kronfeld (1923)
believe that self-labelled bisexuals are really androphiles. Consequent-
ly, an earlier study (Freund, Scher, Chan, & Ben-Aron, 1982) tested
this hypothesis by comparing self-labelled bisexuals with androphiles
who did not claim to be bisexuals (using both self-report scales and the
phallometric test). The outcome supported the hypothesis of the
Page 3
EROTIC AGE PREFERENCE 109
above-noted authors: while the self-professed bisexuais differed signifi-
cantly from the androphiles who did not claim to be bisexual on the
self-report scales, there was no such difference between the two groups
on phallometrically indicated bisexuality.
The demonstration that self-professed bisexuals are really andro-
philes, however, does not rule out the possibility that androphiles in
general may be more bisexual (i.e., demonstrate less erotic gender dif-
ferentiation) than gynephiles. This possibility could not be sufficiently
investigated by the study mentioned above (Freund et al., 1973)
because of the absence of neutral stimuli. That is, the inclusion of
neutral stimuli makes it possible to determine a baseline level, above
which erotic arousal may be said to be occurring and below which
erotic aversion may be said to be occurring. The results of the earlier
investigation could only demonstrate that neither gynephiles nor
androphiles have any gross aversion to the nonpreferred gender. A
more precise comparison of gynephiles and androphiles was the task of
the present study.
Method
Subjects
Two groups of paid volunteers, one gynephilic, the other androphilic,
and a group of (gynephilic) sex offenders against physically mature
females (without any offenses against minors) were selected by com-
puter program from the data bank of phallometric test results of the
sexology department of a psychiatric teaching hospital. All subjects
had undergone the most recent version of the phailometric test
(Freund & Blanchard, in press). The gynephilic volunteers were
recruited from a government placement office for the unemployed, and
the androphilic volunteers were recruited by an advertisement posted
in a gay bookstore. The sex offenders against physically mature
females consisted of patients with various expressions of a courtship
disorder (Freund, Scher, & Hucker, 1983), such as voyeurism, exhibi-
tionism, etc., who had never been charged with rape, and of men who
had been charged with rape. They were referred by psychiatrists
examining them on court order or at the request of the offenders'
lawyers J
^Every subject was told prior to the first test session that the test would measure his
erotic preferences and that he could withdraw from the test at any time. Also, subjects
were shown the volume sensor and told that blood circulation in their penis would be
measured. Additionally, volunteers were informed that the test result would remain
anonymous, and the offenders were told that the results would be reported to the
referring physician. After answering the subject's pertinent questions, a consent form
was given for signing.
Page 4
110 PREUND, WATSON, AND RIENZO
Those among the gynephilic volunteers and the sex offenders
against women whose test outcome was interpreted (by a computer
program) as indicating androphilia were excluded from the study as
were androphilic volunteers who were erroneously diagnosed as gyne-
philes. Individuals whose results showed signs of faking (Freund, Wat-
son, & Rienzo, 1988) were not included in the present investigation.2 Of
the remaining individuals, single test sessions were discarded if the
protocol indicated that a technical error had occurred or if the overall
test response for that session was insufficient as indicated by the "out-
put index" (see below).
The phallometric test was administered in two sessions. Table 1
shows number, mean age, and median education for each group. The
two sessions differ in respect to these numbers because (a) only valid
test sessions were included, and (b) when this test version was intro-
duced, we experimented initially with the number of times the dif-
ferent stimulus categories were presented. Results of subjects who
received only one test session in its final version were included only for
that session and not for the other in which the preliminary version was
used (in Table 1 these cases are listed as "Not given").
Procedure
Description of the test. Both phallometric test sessions included
pictures of nude physically mature and physically immature females
and males, successively presented one at a time on a screen and accom-
panied by taped narratives. A break separated the two sessions to
minimize fatigue effects (as had also been intended with the variation
of types of stimulus presentation—see below). In Session One, each of
27 trials lasted 28 seconds and consisted of two consecutive movie
strips each showing a different person of the same gender-age category
walking towards the viewer. In Session Two, each of 25 trials lasted 54
seconds during which slides of 3 nude persons were shown successively
such that 3 screens simultaneously showed a front view, a rear view,
and the genital region. In the course of a trial, no one screen showed
the same type of view (front, rear, genital region) twice. The narratives
accompanying Session One described the persons on the screen
engaging in neutral activities, such as swimming. In Session Two, the
narratives described imaginary sexual interaction between the tested
subject and the person on the screen.
^The signs of faking are; (a) a characteristic pattern of intentional movements in the
response records; (b) the highest mean response in either of the two test sessions or the
second highest mean response in the first session to neutral stimuli; and (c) the second
highest mean response (in either test session) to a stimulus category of the gender
opposite to the gender of the stimulus category most responded to, provided that at
least one of the two categories is of physically mature persons and that the difference
between the two means is not larger than 0.5 z-scores (see "Measurement").
Page 5
EROTIC AGE PREFERENCE 111
Table 1
Group Characteristics of Gynephiles and Androphiles
Session I
N
Mean age
S.D. age
Median Education
Session II
.V
Mean age
S.D. age
Median Education
Single sessions not used
Session I
Invalid
Not given
Session II
Invalid
Not given
Gyne Off
36
27.0
6.8
8 < Ed < 12
44
28.0
7.0
8 < Ed < 12
13
0
4
1
Gyne Vol
56
25.6
6.4
Ed = 12
78
26.4
6.8
Ed = 12
11
17
2
4
Andro Vol
28
26.7
5.2
Ed = 12
36
27.2
5.4
Ed = 12
4
6
0
2
Note. Gyne Off = gynephilic offenders, Gyne Vol = gynephilic volunteers, Andro
Vol = androphilic volunteers, Ed = educational level. 8 < Ed < 12 = more than 8
grades completed but less than 12, Ed = 12 = 12 grades completed. Single sessions not
used indicates the number of subjects in each group who provided only one usable test
session: Invalid = session not usable because of insufficient response (Low O.I.) or
technical error; Not given = preliminary version of the session administered, therefore
not used in the comparisons.
When the phallometric test of gender-age preference is administered
at the start of each session (after attachment of the volume sensor), the
subject is involved in a conversation on neutral topics until penile
volume does not diminish any more. After each trial, there is a waiting
period until penile volume returns to and oscillates around the baseline
levei before commencement of the next trial. Subjects are monitored
with a low light-level video camera to ensure that they are attending to
the visual stimuli. Nine categories of stimuli were presented in Session
One, including 4 age categories for each of the two genders and also
one category of sexually neutral movie clips. The 4 age categories
were: children of ages 5 to 8 and 8 to 11, pubescents of ages 12 to 15,
and physically mature persons. Stimuli were presented in 8 successive
blocks each consisting of 9 trials, one trial for each of the gender-age
categories and one for the sexually neutral category (landscapes), in
fixed random order. In Session Two, 4 gender-age categories, as well as
a category of neutral slides (also landscapes), were presented in 5 suc-
cessive blocks analogous to those in Session One. The younger
category of children and the pubescents were not included in Session
Page 6
112 FREUND, WATSON, AND RIENZO
Two. Session One was preceded by three habituation trials. Session
Two by two.
Measurement. For each trial, penile volume changes (in mL) were
measured in two ways: D-type scores, the largest deviation from initial
value, and A-type scores, the area under the plotted response curve.
The raw D-type scores were converted into standard scores, derived
from each subject's own D-type data only, for each session separately,
and the same operation was carried out with his A-type scores. These
D-type and A-type standard scores were subsequently combined
according to the formula (z + z )/2,
To avoid inclusion of test sessions in which there was virtually no
response to the stimuli, each subject's overall level of responding was
expressed for each session as an output index {hereafter O.I., Freund,
Chan, &, Coulthard, 1979; Freund & Blanchard, in press). This was the
mean of the 3 highest responses in raw D-scores, responses to neutrals
excluded. O.I. for Session One had to be at least 0.5 mL, while a
minimum of 1,0 mL was necessary in Session Two.^
For each gender-age category, a one-way ANOVA was carded out,
with groups as the independent variable and combined z-scores as the
dependent variable. Differences between pairs of groups were assessed
by Scheffe test.
Results
In neither of the test sessions was there a significant difference
between the groups in respect to O.I. (one-way ANOVA for Session
One: F < 1; d.f.: 2, 117; p = 0,66; for Session Two: F < 1; d,f.: 2, 155;
p = 0.76). There was also no significant difference between the groups
when all those subjects were included in this comparison who had been
excluded only because of low O.I.
Tables 2 and 3 show the mean response of each group to each of the
age categories of the preferred gender, minus the group's mean
response to neturals (both in z-scores). Tables 4 and 5 show the same
•''Diagnostic indicators. Diagnostic indicators were used in the present study to iden-
tify (and exclude) those individuals who had claimed to be gynephilic but were diag-
nosed as androphilic or who had claimed to be androphilic and were diagnosed as
gynephilic. These indicators include a gender preference index (greatest mean response
in 2-scores to any age category of females minus greatest mean response to any age
category of males) and two age preference indices, one for the diagnosis of pedophilia,
the other for the diagnosis of homosexual hebephiha. The pedophiha index is the mean
response to the category of physically mature males or females, whichever is larger,
minus the mean response to any of the gender-age categories of children, whichever is
largest. The homosexual hebephilia index is the mean response to physically mature
females or males, whichever is greater, minus the mean response to male pubescents
(this version of the test does not diagnose heterosexual hebephiSiaf. Iniormation about
cutting scores and further details of the diagnostic process are contained in Freund and
Blanchard (in pressf.
Page 7
EROTIC AGE PREFERENCE
113
Table 2
Relativized Penile Response to the Preferred Gender Session I
Age Category
6-8
M
S.E.
8-11
M
S.E.
12-15
M
S.E.
Adult
M
S.E.
Gyne Off
-.0028^
i.O973)
.3811^
(.1128)
.7533,
(.1048)
1.7497^
(.1239)
Subject Group
Gyne Vol
.0966^
(.0627)
.5102,
(.0863)
.6920,
(.0838)
1.7241,,
(.1063)
Andro Vol
-.0586^
(.0849)
.2239,
(.0932)
.9164,
(,1607)
1.56]ly
(.1739)
ANOVA+
F = 1.01
p = 0.37
F = 2.00
p = QA4
F = 1.01
p = 0,37
F < 1
p = 0.61
Note. The first number in each cell indicates the given group's mean response (M) to the
category in question minus its naean response to the sexuaUy neutral category. The
second number in the cell (in round brackets) indicates the standard error of this rela-
tivized mean {S.E.). Gyne Off = gynephilic offenders. Gyne Voi = gynephiiic volun-
teers. Andro Vol = androphilic volunteers.
+di = (2,il7i in every case.
Within each age category, means having the same subscript are not significantly dif-
ferent atp < 0.05 according to the Scheffe test.
Table 3
Relativized Penile Response to the Preferred Gender Session II
Age Category
8-11
M
S.E.
Adult
M
S.E.
Note. The first number in each cell indicates the given group's mean response {M) to the
category in question minus its mean response to the sexually neutral category. The
second number in the cell (in round brackets) indicates the standard error of this rela-
tivized mean iS.E.}. Gyne Off = gynephilic offenders. Gyne Vol = gynephilic volun-
teers. Andro Vol — androphilic volunteers.
+df = (2,155) in every case.
Gyne Off
.8880^
1.1113)
1.9055^
(.0925)
Subject Group
Gyne Vol
.9558.J
(.0824)
1.8786^
(.0627)
Andro Vol
,7125,
(,1299)
1.7192,
(.1207)
ANOVA+
F - 1.32
p = 0,27
F = 1.09
p = 0.34
Within each age category, means having the same subscript are not significantly dif-
ferent at j D < 0.05 according to the Scheffe test.
Page 8
114 FREUND, WATSON, AND HIENZO
Tabie 4
Relativized Penile Response to the Nonpreferred Gender — Session I
___ — _ _ Subject Group _________
Gyne Off Gyne Vol Age Category Andro Vol ANOVA+
6-8
M
S.E.
-.5283,
(.0741)
~-3491,j,
(.0774)'
-.1100^
(.1007)
F = 4.88
p =0.01
S'lO
M
S.E.
12-15
M
S.E.
--4117^
(.0929)
-•2418^5,
(.0791)'
.0593^,
(.0882)
F = 5.73
p = 0.00
-.3633,
(.1048)
--1095, t ^
(.0876)'
.0879^,
(.1002)
F = 4.28
p = 0.02
Adult
M
S.E.
Note. The first number in each cell indicates the given group's mean response {M) to the
category in question minus its mean response to the sexually neutral category. The
second number in the cell (in round brackets) indicates the standard error of this rela-
tivized mean {S.E.). Gyne Off = gynephilic offenders. Gyne Vol = gynephilic volun-
teers. Andro Vol = androphilic volunteers.
-•3144,
(.0770)
-.2813,
(.0936)
.1332
(.1295)
F = 4.90
p = 0.01
+ df = (2,117) in every case.
Within each age category, means having the same subscript are not significantly dif-
ferent atp < 0.05 according to the Scheffe test.
Table 5
Relativized Penile Response to the Nonpreferred Gender — Session U
_^ ^
Gyne Off
Subject Group
Gyne Vol
_^
Age Category Andro Vol
8-11
M
S.E.
Adult
M
S.E.
Note. The first number hi each cell indicates the given group's mean response (M) to the
category in question minus its mean response to the sexuaUy neutral category. The
second number in the cell (in round brackets) indicates the standard error of this rela-
tivized mean {S.E.). Gyne Off = gynephilic offenders. Gyne Vol = gynephilic volun-
teers. Andro Vol = androphilic volunteers.
+df = (2,155) in every case.
Within each age category, means having the same subscript are not significantly dif-
ferent atp < 0.05 according to the Scheffe test.
.2020,
(.0783)
.1913,
(.0657)
.2956,
(.0978)
F < 1
p = 0.64
.0111^
(.0692)
.0242^
(.0657)
.2394^
(.0905)
F = 2.31
p = 0-10
Page 9
EROTIC AGE PREFKRENCE 115
relativised mean scores for each of the age categories of the non-
preferred gender.
Tables 2 and 3 demonstrate that for none of the age categories of the
preferred gender did the three groups differ from each other in respect
to the difference between the mean penile response to that category
and the mean penile response to neutrals.
The androphiles, however, differed significantly from both gyne-
phiiic groups in respect to responses to mature persons of the non-
preferred gender (see Tables 4 and 5). The androphiles responded more
to physically mature females than the gynephiles responded to
physically mature males. The response of the gynephiles to physically
mature males was less than that to neutrals, indicating aversion,
whereas the response of androphiles to physically mature females was
positive.
In respect to the remaining age categories of the nonpreferred
gender, the androphiles differed significantly only from the sex
offenders against adult females, with the androphiles responding
relatively more than the sex offenders. The Scheffe test did not
differentiate in any category between the gynephilic volunteers and
either of the two remaining groups.
Discussion
The results of the present study suggest that the erotic attractive-
ness of male children (or pubescents) for androphiles is not greater
than the erotic attractiveness of female children (or pubescents) to
gynephiles. Thus, there must be another reason for the finding that the
proportion of sex offenders against male children among homosexual
men is substantially larger than the proportion of sex offenders
against female children among heterosexual men (Cameron, 1985).
Perhaps, the negative result of the present study supports an alterna-
tive hypothesis which focuses on homosexual pedophilia instead of
androphilia, namely, that the development of pedophilia is more close-
ly linked with homosexuality than with heterosexuality (Freund et al,
1984). However, since this explanation is still speculative, research
needs to continue in this direction. From a more practical point of
view, the negative finding in the comparison of gynephiles and andro-
philes (in respect to attractiveness of children of the preferred sex)
indicates a reformulation of notions about sexual offenses against
children. For example, those who blame androphiles for the relatively
larger incidence of sexual offenses against male children, compared to
the incidence of sexual offenses against female children, must come up
Page 10
116 FREUND, WATSON, AND RIENZO
with a reasonable explanation of why these offenses are not actually
perpetuated by pedophiles.
Both gynephilic groups showed an erotic aversion to males of all
ages, whereas androphiles showed an erotic aversion only to the
youngest age group of females. The interpretation that this is a
cultural phenomenon is supported by the fact that only the (gyne-
philic) sex offenders, who probably tried their best to have a favorable
test outcome, attained a convincing degree of aversion, responding
less to every age category of males than to neutrals when compared to
androphiles' responses to females. The gynephilic volunteers, on the
other hand, differed significantly in this direction from the androphilic
volunteers only with regard to responses to adults, and not with
regard to the remaining age-categories. This difference may therefore
be due to cultural taboos.
Finally, the question may be asked to what degree the phallometric
method is pertinent in this kind of analysis. The answer is that there
have been numerous studies on phallometric differentiation between
heterosexual pedophiles, homosexual pedophiles, gynephiles, and
androphiles (most of which are listed in Freund & Blanchard, in press).
These studies show that only rarely are sex offenders against male
children diagnosed as androphiles and that the phailometric diagnosis
of gynephilic and androphilic volunteers almost always corresponds to
their claimed erotic preference.
References
CAMERON, p. (1985). Homosexual molestation of children/Sexual interaction of teacher
and pupil. Psychological Reports, 57, 1227-1236.
FKEUND, K. (1963). A laboratory method for diagnosing predominance of homo- or
hetero-erotic interest in the male. Behaviour Research and Therapy, 1, 85-93.
FRE[JND, K. (1967). Diagnosing homo- or heterosexuality and erotic age-preference by
means of a psychophysioiogical test. Behaviour Research and Therapy, 5, 209-228.
FREUND, K., & BT.ANCHARD, R. (In press). Phallometric diagnosis of pedophilia. Journal
of Consulting and Clinical Psychology.
FREUND, K., CHAN, S., & COULTBARD, R. <1979). Phallonietric diagnosis with "Non-
admitters." Behaviour Research and Therapy, 17, 451-457.
FKEUND, K., DI.AMANT, J., & PINKAYA, V. (1958). On the validity and reliability of the
phalloplethysmographic diagnosis of some sexual deviations. Review of Czecho-
slovak Medicine, 4, 145-151.
FREUND, K.. HEASMAN, G., RACANSKY, I. G., & GLANCY, G. (1984). Pedophilia and
heterosexuality vs. homosexuality. Journal of Sex and Marital Therapy, 10,
193-200.
FKEUND, K.. LANGEVIN, R., CIBIRI, S., & ZAJAC, Y. (1973). Heterosexual aversion in
homosexual males. British Journal of Psychiatry, 122, 163-169.
FKEUND, K., MCKNIGHT, C. K., LANGEVIN, R., & CIBIRI, S. (1972). The female child as a
surrogate object. Archives of Sexual Behavior, 2, 119-133.
Page 11
EROTIC AGE PREFERENCE 117
FREUND, K., SCHKH, H., CHAN, S., & BE\-ARU.\, M. (1982). Experimental analyses of
pedophilia. Behaviuur Research and Therapy, 20, 105-112.
FREUMD, K., SCHEH, H., & HiicKER, S. (1983). The courtship disorders. Archives of
Sexual Behavior, 12, 369-379.
FKEUKD, K., WATSCJN. R.,& RIENZO, U.W'^^^). Signs of feigning in the phaliome trie test.
Behaviour Research and Therapy, 26, 105-112.
OEBHARD, P. H. (1972). Incideivce of overt homosexuality. In J, M, Liningood (Ed.).
Institute of Mental Health Task Force on Homosexuality (pp. 22-29J. Rockville,
MD: National Institute of Mental Health.
GEHHARD. P. H., GAGNON, J. H., POMKROY. W". B-, & C]!KIS'J>:NS<)\, C. U. (1965). Sex
offenders: An analysis of types. New York: Harper & Row.
HiRSCHFKLD, M. (1920). Die Homosexualitdt des Mannes imd des Weibes |Homo-
sexuality in men and womenl (2nd ed.). Berlin: L. Marcus.
KiNSEY, A. C, POMEROY, W. B., & MARTJN. C. E. (1948). Sexual behavior in the human
male. Philadelphia: Saunders.
KuAFPr-ERI\G, R. (1965). Psychopathia sexualis (Frankin S. Klaf, Trans.). New York:
Stein and Day. (Original work published 1886).
KRONFELD, A. (1923). Sexualpathologie [Sexual pathologyl. Leipzig: F. Deuticke.
MoHK, J. W"., TURNKR, R. E., & JERRY, M. B. (1964). Pedophilia and exhibitionism.
Toronto: University of Toronto Press.
WHTTAM, F. L. (1983). Culturally invariable properties of male homosexuality: Tenta-
tive conclusions from cross-cuitural research. Archives of Sexual Behavior. 12,
207-226.
Page 12
Download full-text 
