Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

ARTICLE

JOURNAL
10.1177/0886260503256654
Ménard et al.
OF/ GENDER
INTERPERSONAL
DIFFERENCES
VIOLENCE
IN SEXUAL
/ October
HARASSMENT
2003

Gender Differences in
Sexual Harassment and Coercion
in College Students
Developmental, Individual, and
Situational Determinants

KIM S. MÉNARD
Pennsylvania State University
GORDON C. NAGAYAMA HALL
University of Oregon
AMBER H. PHUNG
MARIAN F. ERIAN GHEBRIAL
LYNETTE MARTIN
Pennsylvania State University

Differences in male (N = 148) and female (N = 278) college students’use of sexually


harassing and coercive behaviors were investigated. Men were twice as likely to be
sexually harassing and 3 times more likely to be sexually coercive as women. Among
men, sexual harassment was predicted by child sexual abuse, hostility, adversarial
heterosexual beliefs, and alcohol expectancy, with the latter mediating the effects of
aggression. Sexual coercion was predicted by adult sexual victimization and alcohol
expectancy, with alcohol expectancy again mediating the effect of aggression.
Among women, sexual harassment was predicted by adult sexual victimization,
adversarial heterosexual beliefs, aggression, and alcohol expectancy, with aggres-
sion mediating the effect of adversarial heterosexual beliefs and alcohol expectancy
mediating the effect of aggression. Sexual coercion was predicted by a hostile per-
sonality, which mediated the effects of both child and adult sexual victimization.
These findings suggest both gender similarities and differences in determinants of
sexual aggression.

Keywords: sexual coercion; sexual harassment; gender differences

Although more frequently perpetrated by men, sexual aggression is a


ubiquitous social problem also committed by women (Hogben, Byrne, &
Hamburger, 1996; Russell & Oswald, 2001; Struckman-Johnson &
JOURNAL OF INTERPERSONAL VIOLENCE, Vol. 18 No. 10, October 2003 1222-1239
DOI: 10.1177/0886260503256654
© 2003 Sage Publications
1222
Ménard et al. / GENDER DIFFERENCES IN SEXUAL HARASSMENT 1223

Struckman-Johnson, 1994). Sexual aggression encompasses a wide contin-


uum of behaviors from using harassment, to coercion, to physical force to
obtain sex from an unwilling partner. Furthermore, a review of 120 studies
found that sexually harassing and coercive behaviors are more prevalent than
more physically violent forms of sexual aggression (Spitzberg, 1999).
Although the correlates of male sexual aggression are well-documented (cf.
Malamuth, 1998), there is a paucity of research into why women behave in
sexually harassing and coercive manners (but cf. Perry, Schmidtke, & Kulik,
1998; Russell & Oswald, 2001). Moreover, despite the fact that these behav-
iors are conceptualized as similar but occupying different places along the
severity continuum (Groth, 1979; Pryor, 1987; Till, 1980) and that similar
determinants seem to predict both behaviors (Malamuth, 1986; Pryor, 1987),
research on these two forms of sexual aggression has been conducted inde-
pendently of one another. Thus, it is not known whether the same mediation
processes predict these two sexually aggressive behaviors, or whether these
processes vary by gender. This study seeks to address these questions by
investigating predictors of sexually harassing and sexually coercive behavior
separately for both men and women.
It is only recently that researchers of sexual harassment and coercion have
investigated these behaviors in women as well as men (Anderson, 1998;
Perry et al., 1998; Russell & Oswald, 2001; Shea, 1998). The few studies that
have examined sexual harassment by women have produced very different
prevalence rates. Thus, it is difficult to draw conclusions regarding the preva-
lence of this behavior among women. For instance, the American Associa-
tion of University Women (2001), which interviewed a nationally representa-
tive sample of 2,064 eighth through 11th graders, found that 57% of boys and
50% of girls reported harassing their peers. Yet in the most recent U.S. Merit
Systems Protection Board (USMSPB) survey, almost all (93%) of the 44% of
women who were sexually harassed were assaulted by men and a little more
than half (65%) of the 19% of men who were sexually harassed were harassed
by women, suggesting that men are more likely to harass than women
(USMSPB, 1995). Research comparing men’s versus women’s sexual coer-
civeness indicates that men are about twice as likely as women to coerce part-
ners into unwanted sexual activity (Hogben et al., 1996; O’Sullivan, Byers, &
Finkelman, 1998). Because men more frequently than women sexually

Authors’ Note: The authors thank D. Wayne Osgood for statistical assistance on this project.
They also thank D. Wayne Osgood and Aaron L. Pincus for helpful comments on an earlier draft
of this article. In addition, they thank Shelly Gibson and Sarah Jarusinski for their assistance
with data collection. This research was supported by National Institute of Mental Health Grant
No. R01 MH58726.
1224 JOURNAL OF INTERPERSONAL VIOLENCE / October 2003

harass and coerce others, few studies have examined the determinants of
women aggressors or have proffered explanations as to why women commit
these behaviors.
Because of its employment context, most theories of sexual harassment
necessarily focus on organizational hierarchies conducive to sexual harass-
ment. Although elements of an organization’s structure predicts sexual
harassment (Fitzgerald, Drasgow, Hulin, Gelfand, & Magley, 1997),
research demonstrates that peers, more frequently than superiors, perpetrate
most sexual harassment (USMSPB, 1981, 1988, 1995), suggesting that con-
text alone cannot explain the behavior. Indeed, some (Vaux, 1993) have ques-
tioned the need to consider the work context, suggesting that harassers may
just be rapists at their place of employment. Guided by the work of Malamuth
(1981) on rape proclivity, Pryor (1987) found that certain individual charac-
teristics increase sexual harassment proclivities. Specifically, and consistent
with Malamuth’s findings, men who have difficulty being empathic, are
higher in authoritarianism, hold adversarial beliefs, adhere to traditional sex
roles, report a likelihood to rape, and can mask their motives by situational
excuses had a greater proclivity to sexually harass others. Based on these
findings, Pryor and his colleagues (Pryor, La Vite, & Stoller, 1993) proposed
a person/situation model, wherein sexual harassment is more likely to occur
when person and situational factors are present in combination. That is, those
with a high proclivity to sexually harass are more likely to do so when the sit-
uational context condones or excuses their behavior.
Similarly, researchers investigating sexual aggression, including coer-
cion, have also acknowledged the need for multifactorial models to explain
this behavior (Hall & Hirschman, 1991; Malamuth, 1986, 1998). They too
contend that developmental, individual, and situational factors do not in and
of themselves produce sexual aggression, but rather the cumulative effects of
all three increase the risk for the behavior. For example, work by Malamuth
and his colleagues (Malamuth, Sockloskie, Koss, & Tanaka, 1991), using
structural equation modeling, found that development antecedents including
childhood victimization result in individual differences, including hostility,
which in combination increased the risk for sexually aggressive behavior
among male college students. Although this model did not include situational
factors like alcohol, research consistently finds a link between alcohol use
or expectancies and sexual aggression (Abbey, McAuslan, & Ross, 1998;
Abbey, Zawacki, Buck, Clinton, & McAuslan, 2001; Muehlenhard & Linton,
1987; Wilson, Calhoun, & McNair, 2002). These links may operate through a
number of pathways. First, alcohol is believed to be an aphrodisiac that
increases sexual desire (Crowe & George, 1989). Second, those who drink
heavily may use intoxication as an excuse for socially unacceptable behavior,
Ménard et al. / GENDER DIFFERENCES IN SEXUAL HARASSMENT 1225

including sexual aggression (Kanin, 1984). Third, personality characteris-


tics, such as impulsivity, may cause both greater alcohol consumption and
greater likelihood to sexually aggress (Seto & Barbaree, 1997). Combining
these factors, Craig (1990) proposed an interactive situational model that
incorporates elements of the individual and the situation. According to her
model, the history and disposition of the male allow him to identify situations
that provide opportunities to be sexually coercive. Thus, developmental,
individual, and situational factors combined explain sexually coercive
behavior.
In sum, although these theoretical models seem to explain men’s sexually
harassing and coercive behavior, whether these models will elucidate wom-
en’s sexually harassing and coercive behavior remains an empirical question.

The Present Model and Hypotheses

Based on prior research, we hypothesize that developmental, individual,


and situational determinants will have direct, additive effects on sexual
harassing and coercive behavior. Consistent with Pryor and his colleagues’
(1993) person/situation model and Craig’s (1990) interactive situational
model, we also hypothesize indirect effects: specifically, that individual dif-
ferences will mediate the relationship between developmental determinants
(i.e., child and adult sexual victimization) and sexual harassment and coer-
cion. Furthermore, alcohol expectancies will mediate the relationship
between individual determinants and sexual aggression. Finally, with regard
to gender, the greater prevalence of sexual aggression among men as com-
pared to women implies that gender differences exist. However, because of
the paucity of research on female sexual aggressors, we do not make any spe-
cific hypotheses but rather predict that there will be gender differences in pre-
dictive models of sexual harassment and sexual coercion.

METHOD

Participants
The participants were 426 undergraduate students (278 women and 148
men) at a large public university in the northeastern United States. Similar to
the undergraduate population at this institution, most participants were
young (mean age = 20.3 years; SD = 2.2), single (93.7%), heterosexual
(98.3%) Caucasians (88.3%).
1226 JOURNAL OF INTERPERSONAL VIOLENCE / October 2003

Procedure

Participants recruited from psychology and sociology classes received


extra credit for taking part in the research. They were informed that the
research involved completing questionnaires about their sexual attitudes and
experiences. Measures were presented in a fixed order because questions
regarding harassment and coercion were judged to be more likely to influ-
ence responses to victimization, personality, and attitudes than the reverse.
Participants completed the questionnaires in groups of 13 to 74. The testing
rooms were large (seating capacity approximately 180) to ensure participants
had privacy to complete the survey. Participants signed consent forms indi-
cating that they understood their participation was anonymous, confidential,
and could be discontinued without penalty.

Measures
Participants completed a battery of self-report measures selected to reflect
sexually harassing and coercive behavior, which served as the dependent
variables, and child as well as adult sexual victimization, personality, atti-
tudes, and alcohol expectancies as the independent variables.

Child sexual abuse. Child sexual abuse was measured using the Child Sex-
ual Abuse subscale of the 38-item Childhood Abuse Trauma Scale (Sanders &
Becker-Lausen, 1995). Participants rated each statement (e.g., “Did you have
a traumatic sexual experience as a child or teenager?”) on a 5-point, Likert-
type scale from 1 (never) to 5 (always). Items were summed to form a com-
posite referred to as “child sexual abuse” (alpha = .80).

Adult sexual victimization. Adult sexual victimization was measured


using 8 items from the Sexual Behavior Scale (Anderson, 1998) modified to
reflect sexual victimization (e.g., how many times they have attempted to
have sexual contact with someone by using physical force). The items were
dichotomized (coded 0 = absent and 1 = present) and summed to form a com-
posite referred to as “adult sexual victimization” (alpha = .83).

Personality. Personality traits were measured using the 64-item Interper-


sonal Adjective Scales (IAS) (Wiggins, 1995). Participants indicated how
accurately each of the adjectives described them as a person using an 8-point,
Likert-type scale ranging from 1 (extremely inaccurate) to 8 (extremely accu-
rate). In this study, we used the two fundamental trait dimensions of domi-
nance and love in our analyses. A positive or high score on love is indicative
Ménard et al. / GENDER DIFFERENCES IN SEXUAL HARASSMENT 1227

of a nurturing interpersonal style, whereas a low or negative score on this


dimension indicates a hostile manner of relating to others. A positive or high
score on dominance reflects a dominant interpersonal style, whereas a low or
negative score on this domain reflects a submissive interpersonal style. The
alpha reliabilities for this sample were .94 for both dominance and love.

Adversarial heterosexual beliefs. Participants’beliefs that sexual relation-


ships are antagonistic were measured using the 15-item Adversarial Hetero-
sexual Belief Scale (Lonsway & Fitzgerald, 1995). Participants rated each
statement (e.g., “In all societies it is inevitable that one sex is dominant”)
using a 5-point, Likert-type scale ranging from 1 (strongly disagree) to 5
(strongly agree). Items were summed to form a scale referred to as
“adversarial beliefs” (alpha = .87).

Nonsexual aggression. Participants’ level of nonsexual aggression was


measured using 26 items from the Aggression Questionnaire (Buss & Perry,
1992). Participants rated each item (e.g., “I have trouble controlling my tem-
per”) on a 5-point, Likert-type scale ranging from 1 (extremely uncharacter-
istic of me) to 5 (extremely characteristic of me). Item ratings were summed
to form a composite score referred to as “aggression” (alpha = .87).

Alcohol expectancies. Alcohol expectancies were measured using the 34-


item Alcohol Expectancy Questionnaire (Leigh & Stacy, 1993). Each item
(e.g., “When I drink alcohol, I get into fights”) was rated using a 6-point,
Likert-type scale from 1 (no chance) to 6 (certain to happen). We used the
sum of the items score to form this variable (alpha = .93).

Sexual harassment. Sexual harassment was measured using the 18 items


from the Sexual Experience Questionnaire (Fitzgerald, Gelfand, & Drasgow,
1995) modified to reflect perpetrating behavior. That is, we asked, “How
often have you offered better treatment to someone if they were sexually
cooperative?” Participants rated each statement using a 5-point, Likert-type
scale from 1 (never) to 5 (most of the time). Items were then summed to form a
composite referred to as “sexual harassment” (alpha = .91).

Sexual coercion. Eight items (e.g., “How many times have you attempted
to have sexual contact with someone by getting them drunk or stoned?”) were
selected from the Sexual Behavior Scale (Anderson, 1998) to form our mea-
sure of sexually coercive behavior. Responses were dichotomized (coded 0 =
absent and 1 = present) and summed to form a composite score referred to as
“sexual coercion” (alpha = .70).
1228 JOURNAL OF INTERPERSONAL VIOLENCE / October 2003

Statistical Analyses

Because we wanted to investigate whether men and women differed in


their use of sexually harassing or coercive behavior, we first compared male
and female participants’ scores on the dependent and the predictor variables
using a student’s t test. Next, Fisher’s r to z transformations were calculated
on the bivariate correlations of the predictor variables with both sexual
harassment and sexual coercion in order to conduct z tests to determine
whether significant gender differences existed at the bivariate level. Next, we
conducted hierarchical regression analyses to examine the predictive power
of each of our constructs of interest (Aiken & West, 1991). We analyzed the
same hierarchical models separately for men and women to determine
whether the predictive processes of sexually harassing and coercive behav-
iors differ by gender. Finally, we completed mediation tests (Clogg, Petkova,
& Haritou, 1995) following the addition of each variable block to the regres-
sion equation to determine whether significant mediation took place (Baron
& Kenny, 1986).

RESULTS

Comparison of Men’s and Women’s Scores


As seen in Table 1 and consistent with prior research, men had signifi-
cantly higher scores on both the sexual harassment and sexual coercion
scales. Men’s mean scores on the sexual harassment scale were almost dou-
ble that of women (men = 14.58, women = 7.69). Similarly, men’s average
score on sexual coercion was 3.5 times higher than that of women (men = .78,
women = .22). In addition, approximately 3 times more men (49%) than
women (16%) acquiesced to any sexual coercion items. Women had signifi-
cantly higher scores on the adult sexual victimization scale (men = .88,
women = 1.41) and on both the dominance and love dimensions of the IAS
(men dominance = –.01, women = .19; men love = –.78, women = .31, respec-
tively). However, men had higher scores than women on the Adversarial
Heterosexual Belief Scale (men = 30.08, women = 25.25) and the Aggression
Questionnaire (men = 68.73, women = 61.60). There were no significant gen-
der differences on the child sexual abuse or the alcohol expectancies scale.
In addition to comparing men and women’s mean responses using t tests,
Fisher’s r to z transformations were also calculated to determine whether
there were significant gender differences in the prediction of the two depen-
dent variables at the bivariate level. As seen in Table 2, the correlation
TABLE 1: Mean Comparisons of Male and Female Participants

Males Females
Mean SD N Mean SD N t test

Dependent variables
Sexual coercion .78 1.03 142 .22 .62 267 6.95***
Sexual harassment 14.58 8.36 148 7.69 5.44 274 10.16***
Independent variables
Child sexual abuse .64 1.88 148 .75 2.23 247 .65
Adult sexual victimization .88 .88 139 1.41 1.41 269 3.55***
IAS—Dominance –.01 .78 121 .19 .92 220 2.09*
IAS—Love –.78 1.20 113 .31 1.07 216 8.53***
Adversarial heterosexual belief 30.08 7.72 148 25.25 6.40 278 6.90***
Aggression Questionnaire 68.73 14.94 148 61.60 16.78 274 4.33***
Alcohol Expectancy Scale 89.81 89.81 141 86.90 86.90 268 1.29

NOTE: IAS = Interpersonal Adjective Scale.


*p < .05. **p < .01. ***p < .001.
1229
1230 JOURNAL OF INTERPERSONAL VIOLENCE / October 2003

TABLE 2: Zero-Order Correlations of Predictor Variables With Sexual Harassment


and Sexual Coercion.

Sexual Harassment Sexual Coercion


Variables Men Women Men Women

Child sexual abuse .36*** .06 .15 .19***


Adult sexual victimization .42*** .30*** .64*** .24***
IAS—Dominance –.07 .09 –.08 .13
IAS—Love –.54*** –.25*** –.31*** –.15*
Adversarial heterosexual beliefs .48*** .32*** .22** .15*
Aggression .43*** .43*** .33** .21***
Alcohol expectancy .31*** .21*** .40*** .18**

NOTE: IAS = Interpersonal Adjective Scale. Within each dependent variable, paired correla-
tions in bold are significantly different from one another based on Fisher’s r to z transformations
(p < .05).
*p < .05. **p < .01. ***p < .001.

between sexual harassment and child sexual abuse was significantly greater
(z = 3.14, p .05) for men (r = .36) than for women (r = .06). The correlations
between sexual harassment and the love dimension of the IAS were also sig-
nificantly different for men and women. The magnitude of correlation for
men (r = –.54) was larger than for women (r = –.25), suggesting that the rela-
tionship between harassment and hostility is greater for men than for women.
None of the remaining correlations differed significantly by gender.
With regard to sexual coercion, the correlation between adult sexual vic-
timization and sexual coercion was significantly different (z = 4.92, p < .001)
across gender, with a higher correlation found among men (r = .64) than
women (r = .24). In addition, the correlation between sexual coercion and
alcohol expectancies was significantly higher for men (r = .40) than for
women (r = .18; z = 2.33, p < .01). None of the remaining correlations with
sexual coercion differed significantly by gender.

Multivariate Analyses

We conducted a series of hierarchical regression analyses to predict two


behaviors: (a) sexual harassment and (b) sexual coercion. Ordinary least
squares (OLS) and tobit regression analyses to predict sexual harassment and
sexual coercion, respectively, used six hierarchical models in which variables
were entered in the following order: child sexual abuse, adult sexual vic-
timization, personality traits love and dominance (entered as one block),
adversarial heterosexual beliefs, aggression, and alcohol expectancies. This
order of entry was used to predict both dependent variables for men only and
Ménard et al. / GENDER DIFFERENCES IN SEXUAL HARASSMENT 1231

then for women only. Variables were entered in this order to reflect our
hypotheses that developmental factors, individual traits, and alcohol expec-
tancies would increase the risk of sexual harassment and coercion. In addi-
tion, individual factors would mediate the relationship between developmen-
tal factors and sexual harassment and coercion, and alcohol expectancies
would mediate the relationship between individual factors and sexual harass-
ment and coercion.
Because sexual harassment data was normally distributed, OLS regres-
sion, completed with SPSS, was used for these analyses. However, because
the sexual coercion data was skewed (i.e., skewness = 2.5 and kurtosis = 11.1
for entire sample), tobit regression (Tobin, 1958), completed with LIMDEP
(Greene, 1995) was used in the prediction of this construct. It is inappropriate
to use OLS regression with skewed data as it violates the assumption of
homogeneity of variance and may produce out-of-range predicted values (cf.
Osgood, Finken, & McMorris, 2002). In tobit regression, scores that fall
above or below a particular cutoff value are censored, whereas those that are
on the other side of the threshold are treated as continuous variables. That is,
tobit is a maximum-likelihood estimator that assumes a normal distribution
for cases above the threshold and a probit probability model for cases below
the threshold. To improve the distribution, the natural logarithm was taken of
the sexual coercion scores.

Sexual harassment—men. Sexual harassment served as the dependent


variable in this series of regressions examining male participants’ behavior.
As seen in Table 3, child sexual abuse, hostility, adversarial heterosexual
beliefs, and alcohol expectancies significantly predicted sexually harassing
behavior. That is, for men, as child sexual abuse, hostility, adversarial beliefs,
and alcohol expectancy scores increase, so too does sexual harassing behav-
ior. Only one of the predicted mediations was significant however. Alcohol
expectancy significantly mediated the effect of aggression on men’ sexual
harassment (t = 3.55, p < .001). The final model accounted for 49% of the
variance based on the adjusted R2, F(7, 97) = 15.29, p < .001.

Sexual harassment—women. In this series of regressions, women’s


harassing behavior served as the dependent variable. Adult sexual victimiza-
tion, adversarial heterosexual beliefs, aggression, and alcohol expectancy
significantly predicted sexual harassment. That is, as women’s scores on sex-
ual victimization, adversarial heterosexual beliefs, aggression, and alcohol
expectancies increase, their level of sexually harassing behavior increases as
well. With regard to the mediation, there were two significant mediating
effects. First, the mediating effect of aggression on adversarial heterosexual
1232

TABLE 3: OLS and Tobit Hierarchical Regression Analysis for Variables Predicting Sexual Harassment and Sexual Coercion, Respectively

Sexual Harassment—Men Sexual Harassment—Women Sexual Coercion—Men Sexual Coercion—Women


(N = 104) (N = 205) (N = 105) (N = 206)
Variables B SE B β B SE B β B SE B B/SE B SE B B/SE

Step 1
Child sex abuse 1.715 .362 .42*** .104 .148 .05 .041 .033 1.25 .050 .020 2.55**
Step 2
Child sex abuse 1.479 .363 .37*** –.026 .144 –.01 –.014 .029 –.47 .042 .020 2.14*
Adult sex victim 1.777 .671 .24** 1.008 .220 .31*** .324 .052 6.26 .081 .038 2.12*
Step 3
Child sex abuse 1.456 .324 .36*** –.073 .142 –.03 –.016 .030 –.54 .034 .019 1.76
Adult sex victim .605 .624 .08 .871 .223 .27*** .302 .054 5.65 .056 .037 1.51
IAS—Dominance .940 .819 .09 .319 .364 .06 .006 .075 .08 .133 .067 1.98*
IAS—Love –3.247 .568 –.47*** –1.072 .315 –.23*** –.070 .052 –1.36 –.121 .061 –2.00*
Step 4
Child sex abuse 1.345 .315 .33*** –.103 .137 –.05 –.018 .029 –.62 .034 .019 1.76
Adult sex victim .606 .603 .08 .809 .216 .25*** .301 .053 5.72 .059 .038 1.58
IAS—Dominance .972 .792 .09 .419 .354 .08 .005 .075 .065 .125 .068 1.85
IAS—Love –2.594 .595 –.37*** –.824 .312 –.18** –.046 .054 –.85 .128 .061 2.09*
Adversarial heterosexual
beliefs .241 .085 .23** .196 .052 .25*** .009 .007 1.25 –.009 .010 –.93
Step 5
Child sex abuse 1.375 .308 .34*** –.147 .135 –.07 –.016 .028 –.57 .034 .020 1.77
Adult sex victim .482 .592 .06 .705 .213 .22*** .292 .051 5.72 .060 .038 1.59
IAS—Dominance .518 .798 .05 .317 .346 .06 –.038 .075 –.50 .126 .068 1.85
IAS—Love –2.216 .604 –.32*** –.302 .342 –.06 –.009 .055 –.17 –.134 .068 –1.95*
Adversarial heterosexual
beliefs .161 .090 .16 .132 .054 .17* .003 .008 .352 –.008 .010 –.80
Aggression .116 .050 .21* .081 .024 .27*** .010 .004 2.15 –.001 .004 –.18
Step 6
Child sex abuse 1.470 .303 .36*** –.148 .133 –.07 –.017 .030 –.57 .034 .019 1.74
Adult sex victim .044 .602 .01 .616 .214 .19** .262 .052 5.07 .048 .038 1.27
IAS—Dominance .445 .778 .04 .310 .342 .06 –.037 .075 –.50 .122 .068 1.80
IAS—Love –1.908 .601 –.28** –.417 .341 –.09 .009 .544 .17 –.164 .073 –2.25*
Adversarial heterosexual
beliefs .192 .088 .19* .108 .055 .14* .005 .008 .60 .012 .011 –1.14
Aggression .085 .050 .15 .073 .024 .24** .007 .005 1.52 .002 .004 –.43
Alcohol expectancy .084 .034 .21* .040 .016 .16* .007 .003 2.04 .007 .004 1.92*
2 2 2 2
NOTE: IAS = Interpersonal Adjective Scale. Sexual Harassment Women: R = .00 for Step 1; ∆R = .09 for Step 2; ∆R = .15 for Step 3; ∆R = .20 for Step 4;
∆R2 = .25 for Step 5; ∆R2 = .27 for Step 6. Sexual Harassment Men: R2 = .18 for Step 1; ∆R2 = .23 for Step 2; ∆R2 = .42 for Step 3; ∆R2 = .47 for Step 4; ∆R2 = .50
for Step 5; ∆R2 = .53 for Step 6. Sexual Coercion Women: Pseudo R2 = .11 for Step 1; ∆R2 = .14 for Step 2; ∆R2 = .21 for Step 3; ∆R2 = .21 for Step 4; ∆R2 = .21
for Step 5; ∆R2 = .23 for Step 6. Sexual Coercion Men: Pseudo R2 = .03 for Step 1; ∆R2 = .41 for Step 2; ∆R2 = .44 for Step 3; ∆R2 = .45 for Step 4; ∆R2 = .49 for
Step 5; ∆R2 = .51 for Step 6.
*p < .05. **p < .01., ***p < .001.
1233
1234 JOURNAL OF INTERPERSONAL VIOLENCE / October 2003

beliefs was significant (t = 9.17, p < .001). Second, the mediating effect of
alcohol expectancy on aggression was significant (t = 4.55, p < .001). The
final model accounted for 24% of the variance according to the adjusted R2,
F(7, 198) = 19.45, p < .001. It is noteworthy that the same predictor variables
explain only half the variance in harassing behavior among women as they do
among men.

Sexual coercion—men. Sexual coercion served as the dependent variable


in this series of regressions. Adult sexual victimization and alcohol expec-
tancy significantly predicted sexual coercion. That is, men sexually vic-
timized in adulthood and those who scored high on the alcohol expectan-
cies scale are more likely to be sexually coercive. With regard to
mediation, only the effect of alcohol expectancies on aggression was signifi-
cant (t = 2.24, p > .001). Based on the calculation of the pseudo R2, the final
model accounted for approximately 51% of the variance.1

Sexual coercion—women. In this series of regressions, women’s sexually


coercive behavior served as the dependent variable. Only the love dimension
of the IAS significantly predicted sexually coercive behavior among women.
That is, as scores on the love dimension decrease, indicating a hostile inter-
personal style, sexually coercive behavior increases. With regard to media-
tion, women’s hostility mediated the relationships of both child (t = 3.49, p <
.001) and adult sexual victimization (t = 4.17, p < .001) to sexual coercion.
The final model explained approximately 23% of the variance. Again, it is
noteworthy that the predictor variables explained less than half the variance
among women as they explained among men.

DISCUSSION

Consistent with our first hypothesis and prior literature, developmental


factors, individual traits, and alcohol expectancies predicted men’s and
women’s sexual harassing and coercive behavior. However, we obtained
mixed results regarding our second hypothesis relating to mediation. Finally,
although there were many similarities across gender, we also found impor-
tant differences primarily in the mediation process.

Determinants of Sexual Harassment and Coercion


Prior sexual victimization, hostility, adversarial beliefs, aggression, and
alcohol expectancies predicted sexual harassment and sexual coercion. How-
Ménard et al. / GENDER DIFFERENCES IN SEXUAL HARASSMENT 1235

ever, we anticipated that similar models would predict both types of sexually
aggressive behaviors, but this was not the case. Fewer variables predicted
sexual coercion than predicted sexual harassment. For instance, adversarial
beliefs were predictive of sexual harassment but not sexual coercion. Socio-
cultural beliefs may play a larger role in sexual harassment, because it is the
less violent behavior and as such, there may be fewer social sanctions against
it. Indeed, some individuals (i.e., all male groups) may even condone sexu-
ally harassing behavior (Allison & Wrightsman, 1993). Future research
should investigate the possibility that different variables predict sexually
aggressive behavior depending on where along a severity continuum the acts
under investigation fall.

Mediation

Although we hypothesized that individual factors would mediate the


effects of developmental factors and that alcohol expectancies would medi-
ate the effects of individual factors, results regarding both types of mediation
were mixed. First, contrary to research on victimization outcomes that finds
an association between victimization and individual development (e.g.,
Miller & Lisak, 1999), support for this hypothesis was found only for female
coercers. It is possible that this hypothesis was not supported because our
high functioning college sample precluded more severe cases of sexual vic-
timization and offending. For instance, it is noteworthy that we also did not
find significant gender differences in child sexual victimization, a result that
is contrary to prior literature (Finkelhor & Ormrod, 2001).
Second, results provide partial support for the mediating effects of alcohol
expectancies on individual differences. Consistent with both Pryor et al.’s
(1993) person/situation model and Craig’s (1990) interactive situational
model, alcohol expectancies mediated the relationship between individual
determinants (aggression) and sexually aggressive behavior for all but the
female sexual coercers. Thus, those who are sexually harassing or coercive
may use alcohol expectancies to cognitively restructure the situation to one
that is favorable to sexual aggression. This notion is consistent with research
on the effects of alcohol use and expectancies on sexual coercion. For
instance, work by George and Marlatt (1986) found that male subjects who
had not consumed alcohol, but believed they had, appeared to use alcohol as
“both the impetus and alibi for inappropriate behavior” (p. 157). Similarly, in
their investigation of the link between alcohol consumption, expectancies,
and sexual coercion, Wilson and her colleagues (2002) found that the rela-
tionship between alcohol consumption and sexual coercion was most promi-
nent when the men held high alcohol expectancies regarding sexual behavior.
1236 JOURNAL OF INTERPERSONAL VIOLENCE / October 2003

Our failure to find support for this hypothesis among female coercers may
be due to our small sample size (Cohen, 1992). Alternatively, female coercers
may be a unique group whose behavior does not adhere to findings estab-
lished using male sexual aggressors. However, prior to drawing any conclu-
sions regarding the mediation effects of alcohol expectancies, these results
need to be replicated with a larger, more diverse sample.

Gender Differences

The predictive models for sexual coercion produced the greatest gender
differences. Indeed, z tests (Brame, Paternoster, Mazerolle, & Piquero, 1998)
(not shown) of the gender differences in the regression coefficients across
both dependent variables found no significant differences in sexual harass-
ment and only one gender difference in coercion. That is, when predicting
sexual coercion, the regression coefficients for adult sexual victimization dif-
fered significantly by gender (z = 3.32; p < .001). Although similar models
predict sexual harassment and sexual coercion, there are some interesting dif-
ferences with regard to mediation. Among men, aggression mediated by
alcohol expectancies is a recurrent feature of sexually aggressive behavior,
whereas among women, hostility is a determining factor. Although hostility
and aggression demonstrate the importance of interpersonal style in the pre-
diction of sexual aggression, the role of alcohol expectancies is noteworthy as
well.
Because this sample consisted of college students, the results are not rep-
resentative of all individuals who commit sexual harassment or coercion.
About half of Americans attend college (U.S. Bureau of Statistics, 1999).
College students are among the age group most at risk for sexual victimiza-
tion and perpetration (Bureau of Justice Statistics, 2001). Thus, the current
sample has some generalizability to the population of interest. However, cau-
tion is needed when generalizing from the present sample to those who have
been identified by the criminal justice system, as the latter group is likely to
engage in more extreme or violent behavior.
The overall results have two policy implications: First, if we hope to
reduce sexual aggression, we must provide child and adult sexual assault vic-
tims with adequate treatment so that they do not attempt to master their vic-
timization through the perpetration of others. Second, the belief that alcohol
use can change one’s behavior increases the risk of sexual perpetration and,
consequently, victimization. Thus, greater promotion of alcohol education,
including sobriety, may be helpful in reducing sexual aggression among col-
lege students.
Ménard et al. / GENDER DIFFERENCES IN SEXUAL HARASSMENT 1237

NOTE

2
1. Tobit does not automatically produce an R ; however, an estimate of this value can be cal-
culated using sigma.

REFERENCES

Abbey, A., McAuslan, P., & Ross, L. T. (1998). Sexual assault perpetration by college men: The
role of alcohol, misperception of sexual intent, and sexual beliefs and experiences. Journal
of Social and Clinical Psychology, 17, 167-195.
Abbey, A., Zawacki, T., Buck, P. O., Clinton, A. M., & McAuslan, P. (2001). Alcohol and sexual
assault. Alcohol Research and Health, 25, 43-51.
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions.
Newbury Park, CA: Sage.
Allison, J. A., & Wrightsman, L. S. (1993). Rape: The misunderstood crime. Newbury Park, CA:
Sage.
American Association of University Women. (2001). Hostile hallways: Bullying, teasing, and
sexual harassment in school. Washington, DC: AAUW Educational Foundation.
Anderson, P. B. (1998). Women’s motives for sexual initiation and aggression. In P. B. Anderson
& C. Struckman-Johnson (Eds.), Sexually aggressive women: Current perspectives and con-
troversies (pp. 79-93). New York: Guilford.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psy-
chological research: Conceptual, strategic, and statistical considerations. Journal of Person-
ality and Social Psychology, 51, 1173-1182.
Brame, R., Paternoster, R., Mazerolle, P., & Piquero, A. (1998). Testing for the equality of maxi-
mum-likelihood regression coefficients between two independent equations. Journal of
Quantitative Criminology, 14, 245-261.
Bureau of Justice Statistics. (2001). Criminal victimization in the United States, 1999. Washing-
ton, DC: U.S. Department of Justice.
Buss, A. H., & Perry, M. (1992). The Aggression Questionnaire. Journal of Personality and
Social Psychology, 63, 452-459.
Clogg, C. C., Petkova, E., & Haritou, A. (1995). Statistical methods for comparing regression
coefficients between models. American Journal of Sociology, 100, 1261-1293.
Cohen, J. (1992). A power primer. Psychological Bulletin, 12, 155-159.
Craig, M. E. (1990). Coercive sexuality in dating relationships: A situational model. Clinical
Psychology Review, 10, 395-423.
Crowe, L., & George, W. (1989). Alcohol and human sexuality: Review and integration. Psycho-
logical Bulletin, 105, 374-386.
Finkelhor, D., & Ormrod, R. (2001). Child abuse reported to the police. Washington, DC: Office
of Juvenile Justice and Delinquency Prevention.
Fitzgerald, L. F., Drasgow, F., Hulin, C. L., Gelfand, M. J., & Magley, V. J. (1997). The anteced-
ents and consequences of sexual harassment in organizations: A test of an integrated model.
Journal of Applied Psychology, 82, 578-589.
Fitzgerald, L. F., Gelfand, M. J., & Drasgow, F. (1995). Measuring sexual harassment: Theoreti-
cal and psychometric advances. Basic and Applied Social Psychology, 17, 425-445.
1238 JOURNAL OF INTERPERSONAL VIOLENCE / October 2003

George, W. H., & Marlatt, G. A. (1986). The effects of alcohol and anger on interest in violence,
erotica, and deviance. Journal of Abnormal Psychology, 95, 150-158.
Greene, W. H. (1995). LIMDEP: Version 7 users manual. Plainview, NY: Econometrics Soft-
ware Inc.
Groth, A. N. (1979). Men who rape: Psychology of the offender. New York: Plenum.
Hall, G. C. N., & Hirschman, R. (1991). Toward a theory of sexual aggression: A quadripartite
model. Journal of Consulting and Clinical Psychology, 59, 662-669.
Hogben, M., Byrne, D., & Hamburger, M. E. (1996). Coercive heterosexual sexuality in dating
relationships of college students: Implications of differential male-female experiences.
Journal of Psychology and Human Sexuality, 8, 69-78.
Kanin, E. J. (1984). Date rape: Unofficial criminals and victims. Victimology, 9, 95-108.
Leigh, B. C., & Stacy, A. W. (1993). Alcohol outcome expectancies: Scale construction and pre-
dictive utility in higher order confirmatory models. Psychological Assessment, 5, 216-229.
Lonsway, K. A., & Fitzgerald, L. F. (1995). Attitudinal antecedents of rape myth acceptance: A
theoretical and empirical reexamination. Journal of Personality and Social Psychology, 68,
704-711.
Malamuth, N. M. (1981). Rape proclivities among males. Journal of Social Issues, 37, 138-157.
Malamuth, N. M. (1986). Predictors of naturalistic sexual aggression. Journal of Personality and
Social Psychology, 50, 953-962.
Malamuth, N. M. (1998). The confluence model as an organizational framework for research on
sexually aggressive men: Risk moderators, imagined aggression, and pornography con-
sumption. In R. G. Geen & E. Donnerstein (Eds.), Human aggression: Theories, research,
and implications for social science (pp. 229-245). San Diego, CA: Academic Press.
Malamuth, N. M., Sockloskie, J. S., Koss, M. P., & Tanaka, J. S. (1991). Characteristics of
aggressors against women: Testing a model using a national sample of college students.
Journal of Consulting and Clinical Psychology, 59, 670-681.
Miller, P. M., & Lisak, D. (1999). Association between childhood abuse and personality disorder
symptoms in college males. Journal of Interpersonal Violence, 14, 642-656.
Muelhenhard, C. L., & Linton, M. A. (1987). Date rape and sexual aggression in dating situa-
tions: Incidence and risk factors. Journal of Consulting Psychology, 34, 186-196.
Osgood, D. W., Finken, L. L., & McMorris, B. J., (2002). Analyzing multiple-item measures of
crime and deviance II: Tobit regression. Quantitative Criminology, 18, 319-347.
O’Sullivan, L. F., Byers, E. S., & Finkelman, L. (1998). A comparison of male and female col-
lege students’experience of sexual coercion. Psychology of Women Quarterly, 22, 177-195.
Perry, E. L., Schmidtke, J. M., & Kulik. C. T. (1998). Propensity to sexually harass: An explora-
tion of gender differences. Sex Roles, 38, 443-460.
Pryor, J. B. (1987). Sexual harassment proclivities in men. Sex Roles, 17, 269-290.
Pryor, J. B., La Vite, C. M., & Stoller, L. M. (1993). A social psychological analysis of sexual
harassment: The person/situation interaction. Journal of Vocational Behavior, 42, 68-83.
Russell, B. L., & Oswald, D. L. (2001). Strategies and dispositional correlates of sexual coercion
perpetrated by women: An exploratory investigation. Sex Roles, 45, 103-115.
Sanders, B., & Becker-Lausen, E. (1995). The measurement of psychological maltreatment:
Early data on the Child Abuse and Trauma Scale. Child Abuse and Neglect, 19, 315-323.
Seto, M. C., & Barbaree, H. E. (1997). Sexual aggression as antisocial behavior: A developmen-
tal model. In D. M. Stoff, J. Breiling, & J. D. Maser (Eds.), Handbook of antisocial behavior.
New York: John Wiley.
Shea, M. E. C. (1998). When the tables are turned: Verbal sexual coercion among college
women. In P. B. Anderson & C. Struckman-Johnson (Eds.), Sexually aggressive women:
Current perspectives and controversies (pp. 94-104). New York: Guilford.
Ménard et al. / GENDER DIFFERENCES IN SEXUAL HARASSMENT 1239

Spitzberg, B. H. (1999). An analysis of empirical estimates of sexual aggression victimization


and perpetration. Violence and Victims, 14, 241-260.
Struckman-Johnson, C., & Struckman-Johnson, D. (1994). Men pressured and forced into sex-
ual experience. Archives of Sexual Behavior, 23, 93-114.
Till, F. J. (1980). Sexual harassment: A report on the sexual harassment of students. Washington,
DC: National Advisory Council on Women’s Educational Programs.
Tobin, J. (1958). Estimation of relationship for limited dependent variables. Econometrica, 26,
24-36.
U.S. Bureau of Statistics. (1999). Statistical abstract of the United States. Washington, DC:
Government Printing Office.
U.S. Merit Systems Protection Board. (1981). Sexual harassment in the federal workplace: Is it a
problem? Washington, DC: Government Printing Office.
U.S. Merit Systems Protection Board. (1988). Sexual harassment in the federal workplace: An
update. Washington, DC: Government Printing Office.
U.S. Merit Systems Protection Board. (1995). Sexual harassment in the federal workplace:
Trends, progress, continuing challenges. Washington, DC: Government Printing Office.
Vaux, A. (1993). Paradigmatic assumptions in sexual harassment research: Being guided with-
out being misled. Journal of Vocational Behavior, 42, 116-135.
Wiggins, J. (1995). Interpersonal Adjective Scales professional manual. Odessa, FL: Psycho-
logical Assessment Resources.
Wilson, A. E., Calhoun, K. S., & McNair, L. D. (2002). Alcohol consumption and expectancies
among sexually coercive college men. Journal of Interpersonal Violence, 17, 1145-1159.

Kim S. Ménard was a doctoral candidate in the Crime, Law, and Justice Program in the
Department of Sociology at Pennsylvania State University when this study was com-
pleted. She is now an assistant professor in the Department of Sociology at San Jose
State University. Her primary research focus is criminal victimization, with a secondary
focus on sexual assault.

Gordon C. Nagayama Hall is a professor of clinical psychology at the University of Ore-


gon. His primary research focus is sexual aggression, with a secondary focus on cultural
variation in sexual aggression.

Amber H. Phung is a doctoral candidate in clinical psychology at Pennsylvania State


University.

Marian F. Erian Ghebrial is a doctoral candidate in clinical psychology at Pennsylvania


State University.

Lynette Martin is a doctoral candidate in the Crime, Law, and Justice Program in the
Department of Sociology at Pennsylvania State University.

You might also like