The Drive for Muscularity and Masculinity: Testing the Associations Among Gender-Role Traits, Behaviors, Attitudes, and Conflict

Psychology of Men & Masculinity
Copyright 2005 by the Educational Publishing Foundation
2005, Vol. 6, No. 2, 83–94
1524-9220/05/$12.00
DOI: 10.1037/1524-9220.6.2.83
The Drive for Muscularity and Masculinity: Testing the
Associations Among Gender-Role Traits, Behaviors, Attitudes, and
Con?ict
Donald R. McCreary
Deborah M. Saucier
BrockUniversity
University of Saskatchewan
Will H. Courtenay
McLean Hospital, Harvard Medical School and Men’s Health Consulting
Prior qualitative research has suggested that people assume muscular men are more
masculine. This assumption was tested quantitatively in 2 studies. In Study 1, men and
women completed measures of gender-role traits and behaviors, whereas in Study 2,
men completed measures of gender-role con?ict and traditional attitudes about men.
Study 1 revealed a correlation between self-rated male-typed traits and behaviors, with
a need to be more muscular for both men and women. In Study 2, men with more
traditional attitudes about men also wanted to be more muscular; men who wanted to
be more muscular were experiencing con?ict with regard to society’s expectations that
they be successful, powerful, and competitive, and they reported that ?nding a balance
between workand leisure is dif?cult.
Until recently, research exploring gender dif-
women re?ects being small and thin, the social
ferences in body image concerns and their out-
standard for men re?ects being big and muscu-
comes has been based solely on perceptions of
lar, what Mishkind, Rodin, Silberstein, and
adiposity (i.e., body fat). This research has led
Striegel-Moore (1986) refered to as the “mus-
to the belief that because men are less con-
cular mesomorphic” shape.
cerned or dissatis?ed than women with their
There is a growing amount of research dem-
degree of adiposity, are less likely than women
onstrating the importance of being muscular in
to be dieting to lose weight, and rarely experi-
both adolescent males (e.g., Jones, 2001; Mc-
ence clinical disorders associated with body im-
Creary & Sasse, 2000, 2002; O’Dea & Raw-
age (e.g., anorexia and bulimia nervosa), they
storne, 2001) and adult men (e.g., Fisher, Dunn,
are relatively happy with their bodies (e.g.,
& Thompson, 2002; Jacobi & Cash, 1994; La-
Feingold & Mazzella, 1998; Garner, Olmstead,
vine, Sweeney, & Wagner, 1999; Phillips &
& Polivy, 1983; Muth & Cash, 1997). This
Diaz, 1997). For example, research has shown
belief, however, does not take into consider-
that many adolescent boys are engaged in resis-
ation the fact that adiposity is not equally im-
tance training activities to gain muscle mass
portant for men and women. That is, whereas
(Ricciardelli & McCabe, 2003). Among adoles-
the social standard of bodily attractiveness for
cent boys, a higher drive for muscularity is
associated with poorer self-esteem and more
symptoms of depression (McCreary & Sasse,
Donald R. McCreary, Department of Psychology, Brock
2000). Pope et al. (2000) have shown that, when
University, St. Catharines, Ontario, Canada; Deborah M.
comparing men’s actual degree of muscularity
Saucier,
Department
of
Psychology,
University
of
to their perceived degree of muscularity, men
Saskatchewan, Saskatoon, Saskatchewan, Canada; Will H.
signi?cantly underestimated their percentage of
Courtenay, McLean Hospital, Harvard Medical School and
Men’s Health Consulting, Berkeley, California.
muscle mass. Pope et al. (2000) also showed
Correspondence concerning this article should be ad-
that men’s ideal body size represents an average
dressed to Donald R. McCreary, who is now at the Stress
increase of 28 pounds (12.7 kg) of muscle and
and Coping Group, Defence R&D Canada—Toronto, 1133
that men feel women are most attracted to a
Sheppard Avenue West, P.O. Box 2000, Toronto, Ontario
M3M 3B9, Canada. E-mail: Don.McCreary@drdc-rddc.gc.ca
body shape that is, on average, 30 pounds (13.6
83

---

84
MCCREARY, SAUCIER, AND COURTENAY
kg) heavier in muscle than their actual size.
Toward Women Scale (Spence & Helmreich,
When asked where they would like to be more
1978) and the Male Role Norms Scale (Thomp-
muscular, men and boys typically want larger
son & Pleck, 1986; Thompson, Pleck, & Fer-
pectorals, biceps, and shoulders (Drewnowski
rera, 1992). Other frequently used measures of
& Yee, 1987; Huenemann, Shapiro, Hampton,
gender-role socialization include the Gender
& Mitchell, 1966; Moore, 1990).
Role Con?ict Scale (GRCS; O’Neil, Helms,
Some researchers (e.g., Grogan & Richards,
Gable, David, & Wrightsman, 1986), the Mas-
2002; Klein, 1993; Weinke, 1998) have sug-
culine Gender Role Stress Scale (Eisler & Skid-
gested that one of the main reasons muscularity
more, 1987), and the Feminine Gender Role
is so important to men and boys is that it is
Stress Scale (Gillespie & Eisler, 1992; see
linked to perceptions of their masculinity; in
Beere, 1990, for more examples of how gender-
other words, the more muscular a man or boy is,
role socialization can be measured). Thus, re-
the more masculine he is. One might also sug-
searchers examining the relationships between
gest that boys and men who are not muscular
the drive for muscularity and perceptions of
(e.g., those who are endomorphic or those who
masculinity and femininity need to be very spe-
are ectomorphic) may see themselves as more
ci?c about the gender-role dimensions they are
feminine. For example, Grogan and Richards
measuring.
(2002) have suggested that boys and men who
In addition to studying the associations
diet run the riskof being perceived as feminine,
among the drive for muscularity and both male-
in part because dieting is perceived to be a
and female-typed aspects of gender-role social-
feminine-typed behavior.
ization, researchers also need to consider
One way to test these hypotheses is to explore
whether those associations vary as a function of
the associations among masculinity, femininity,
gender. McCreary and his colleagues (Mc-
and the drive for muscularity. On the basis of
Creary, Newcomb, & Sadava, 1999; McCreary
?ndings from the qualitative research, we pre-
& Sasse, 2000) refer to this notion as “differ-
dicted that masculinity would be positively as-
ential salience” and have argued that the asso-
sociated with the drive for muscularity, whereas
ciations between gender-typed dimensions and
femininity would be negatively associated with
various outcome variables (e.g., self-esteem, al-
the drive for muscularity. However, there is an
cohol dependence) should be moderated by
important limitation inherent in this prediction:
gender. They believe that, even though men and
Masculinity and femininity are global, higher
women internalize many gender-typed dimen-
order constructs and cannot be measured di-
sions in the same way (i.e., factor analyses have
rectly (Spence, 1984). Psychology has over-
shown that there are few, if any, gender differ-
come this problem by developing indices of
ences in the factor structures of many gender-
speci?c dimensions of masculinity and feminin-
role measures; e.g., McCreary et al., 1998),
ity. Personality trait measures, such as the Per-
differential social pressures to conform to gen-
sonal Attributes Questionnaire (PAQ; Spence &
der-role expectations are stronger for men than
Helmreich, 1978), the Extended PAQ (EPAQ;
for women (i.e., failure to conform to gender-
Spence, Helmreich, & Holahan, 1979), and the
role norms or expectations tends to be punished
Bem Sex Role Inventory (Bem, 1974), are used
more in men than in women; McCreary, 1994).
to assess the degree to which men and women
They have demonstrated this differential sa-
have internalized the gender stereotypic person-
lience in two studies: McCreary, Newcomb, and
ality traits of agency–instrumentality (i.e., male
Sadava (1999) showed that the associations be-
stereotypic traits) and communion– expressivity
tween various dimensions of the male gender
(i.e., female stereotypic traits). Engaging in
role and both alcohol use and alcohol problems
gender-typed behaviors is assessed using the
were different for men and women, and Mc-
Sex Role Behavior Scale (SRBS; Orlofsky,
Creary and Sasse (2000) revealed that the drive
1981) and its short form companion (Orlofsky
for muscularity was associated with lower self-
& O’Heron, 1987). The degree to which people
esteem and more depressive symptoms in boys
have adopted traditional versus contemporary
but not in girls (McCreary & Sasse, 2000). In
or liberal views about men, women, and the
the present context, the notion of differential
relationships between men and women is as-
salience would predict that the association be-
sessed using measures such as the Attitudes
tween gender-role constructs and the drive for

---

DRIVE FOR MUSCULARITY
85
muscularity should be stronger for men than for
(very much like me) to 6 (not at all like me). All
women.
items are reverse coded so that higher scores
The purpose of the two studies presented
re?ect a greater drive for muscularity. Sample
herein is to test the associations between several
items include, “I wish I were more muscular”
dimensions of masculinity, femininity, and the
and “I feel guilty if I miss a weight training
drive for muscularity. In Study 1, a sample of
session.” McCreary and Sasse (2000, 2002)
college-aged men and women completed mea-
have shown that the DMS has good construct
sures of both gender-typed traits and gender-
validity (in the form of face validity), conver-
typed behaviors; in Study 2, a group of young
gent validity, and discriminant validity. Recent
men completed a measure of male-oriented gen-
factor analytic workhas shown that the DMS
der-role con?ict, as well as a measure of tradi-
has a two-factor, lower order structure for men,
tional attitudes about men. In both instances, the
representing the attitudinal and behavioral
degree to which those gender-typed dimensions
items. For both men and women, however, a
predicted the drive for muscularity was exam-
single higher order DMS factor emerged (Mc-
ined; in Study 1, the extent to which those
Creary, Sasse, Saucier, & Dorsch, 2004). Thus,
relationships varied as a function of the partic-
when comparing men and women, researchers
ipants’ gender also was assessed.
should average over the DMS items to create a
single DMS score. Cronbach’s alphas were .91
Study 1
(men) and .83 (women) in the current sample.
The EPAQ.
The EPAQ is a 40-item mea-
Method
sure of gender-typed personality traits and rep-
resents an extension of the PAQ (Spence &
Participants and Procedure
Helmreich, 1978). Although the EPAQ is com-
posed of six subscales, problems with internal
A total of 157 men and 343 women between
consistency (Spence et al., 1979) have led to
the ages of 17 and 78 years (M
20 years, with
only three being used on a regular basis:
95% of the sample between 17 and 24 years)
Agency (mastery-oriented traits that are socially
participated in the current study. All were re-
desirable for both men and women to possess
cruited during a mass testing session for an
but are more stereotypically associated with
introductory psychology subject pool at a large
men; 8 items), Communion (interpersonally ori-
midwestern Canadian university. Seventy-four
ented traits that are desirable for both men and
percent of the men and 79% of the women were
women to possess but are more stereotypically
in their ?rst year of college. Ten percent of the
associated with women; 8 items), and Unmit-
men and 28% of the women reported being on
igated Agency (being focused on the self to the
a diet to lose weight, and 17% of the men and
exclusion of others; socially undesirable for
1% of the women reported being on a diet to
both men and women to possess but more ste-
gain weight. No other demographic or health-
reotypically associated with men; 8 items). The
related information was collected.
other three subscales (i.e., Masculinity–Femi-
ninity, Unmitigated Communion, and Verbal
Measures
Aggressiveness) were not used in this study.
Items on the EPAQ are scored on a 5-point
All participants completed the Drive for
Likert-type scale, with higher scores represent-
Muscularity Scale (DMS; McCreary & Sasse,
ing greater degrees of internalization of each
2000) and two commonly used measures of
trait. The alphas for the present study were .74,
gender-role socialization: the EPAQ (Spence et
.80, and .71, respectively (men), and .73, .70,
al., 1979) and the Short-Form SRBS (Orlofsky
and .62, respectively (women). The low level of
& O’Heron, 1987). Each of these measures is
internal consistency of the Unmitigated Agency
described below.
scale in women is problematic in that it reduces
The DMS.
The DMS is a 15-item measure
the power of that scale to detect signi?cant
of the extent to which people desire to have a
associations with smaller effect sizes.
more muscular body. Items on the DMS repre-
Short-Form SRBS.
The SRBS is the only
sent a combination of attitudes and behaviors
available measure of the extent to which people
and are scored on a 6-point scale ranging from 1
act in gender-typed ways. The 96 items from the

---

86
MCCREARY, SAUCIER, AND COURTENAY
short-form version of the SRBS are grouped
Results
into three higher order factors of 32 items each:
Male-Typed Behaviors (equally desirable for
To explore the associations between the drive
men and women but more stereotypic of men),
for muscularity and self-reported gender-typed
Female-Typed Behaviors (equally desirable for
traits and behaviors, as well as whether those
men and women but more stereotypic of
associations varied as a function of the partici-
women), and Sex-Speci?c Behaviors (more de-
pant’s gender, we performed a series of moder-
sirable and stereotypic of men or women). All
ating variable hierarchical regression analyses.
items are rated on a 5-point Likert-type scale.
The results of these analyses are presented be-
The items within each of the Male- and Female-
low, following an overview of the various scale
Typed Behaviors subscales are averaged into
descriptive statistics.
their respective scores, with higher values being
indicative of engaging in greater degrees of
Descriptive Statistics
behavior in those domains. Scoring of the Sex-
Speci?c Behaviors dimension is somewhat dif-
Table 1 reports the means and standard
ferent, however. This subscale is composed of
deviations for the DMS and all gender-role
items that are either male (16 items) or female
measures, separately for men and women.
speci?c (16 items). The latter items are reverse
Independent sample t tests were used to ex-
coded, and the two sets of items are averaged
plore gender differences among the variables.
together to create the overall subscale score,
To control for the increased probability of
which is then interpreted in the male-typed di-
making a Type I error, we applied a Bonfer-
rection. In the present study, the SRBS demon-
roni correction to all p values (.05/7 tests
strated lower than expected levels of reliability,
.007). As Table 1 shows, all mean differences
with men’s Cronbach alpha coef?cients being
were statistically signi?cant. Men reported
.71, .76, and .56 for the three scales, respec-
signi?cantly higher levels of the drive for
tively, and women’s Cronbach alpha coef?-
muscularity and both agentic and unmitigat-
cients being .63, .68, and .56, respectively. As
edly agentic personality traits; they also acted
we noted earlier, these lower than expected al-
in signi?cantly more male-typed and sex-spe-
pha coef?cients decrease the power of the anal-
ci?c ways. Women, in contrast, reported sig-
yses to detect signi?cant associations with
ni?cantly higher levels of communal traits
smaller effect sizes.
and acted in more female-typed ways.
Table 1
Means and Standard Deviations for the Drive for Muscularity Scale (DMS) and
Six Indices ofGender-Role Socialization
Men
Women
(n
159)
(n
349)
Scale
M
SD
M
SD
t(506)
DMS
2.69
0.95
1.99
0.61
10.01*
EPAQ
Agency
3.64
0.60
3.40
0.58
4.25*
Communion
3.68
0.59
4.10
0.46
8.58*
Unmitigated Agency
2.71
0.58
2.47
0.51
4.75*
SRBS
Male-Typed Behaviors
3.25
0.43
2.83
0.35
11.74*
Female-Typed Behaviors
2.62
0.37
3.17
0.36
15.86*
Sex-Speci?c Behaviors
3.51
0.26
2.66
0.28
32.46*
Note.
EPAQ
Extended Personal Attributes Questionnaire; SRBS
Sex Role Behavior
Scale.
* p
.001.

---

DRIVE FOR MUSCULARITY
87
Regression Analyses
Gender-role behavior analysis.
The addi-
tion of the four main effects (Male- and Female-
Because gender-role traits and behaviors are
Typed Behaviors, Sex-Speci?c Behaviors, par-
conceptually different from one another, we
ticipant’s gender) in Step 1 explained a signif-
performed two separate moderating variable hi-
icant amount of variance in DMS scores,
F(4,
erarchical regressions: one with the three trait
503)
42.46, p
.01 (adjusted R2
25.2%).
variables as the main predictors and one with
Univariate t tests revealed that three beta values
the three behavioral variables as the predictors.
were signi?cantly different from zero. As with
In both analyses, the main effects for the gen-
the ?rst regression analysis, participant’s gender
der-role dimensions and participant’s gender
was a signi?cant predictor (
–.19), t(3)
(men
1, women
2) were entered into the
–2.75, p
.01. The direction of the standard-
equation in Step 1. In Step 2, the centered
ized coef?cient means that men scored higher
Gender
Gender Role interactions were en-
than women on the DMS, after partialing out
tered (see Cohen, Cohen, West, & Aiken, 2002,
the effects of the three gender-typed behavior
for a discussion of moderating variable regres-
domains. Both the SRBS Male-Typed and Sex-
sion analysis and the use of centering). If the
Speci?c Behavior scales were signi?cantly and
addition of the centered interaction terms in
positively associated with DMS scores, (
Step 2 accounts for a signi?cant increase in the
.27), t(3)
4.99, p
.01, and (
.16),
amount of variance explained in the criterion,
t(3)
1.99, p
.05, respectively, after partial-
then there is a signi?cant Gender
Gender
Role moderating effect. Univariate t tests ex-
ing out the effects of the other variables. Thus,
plore which main effect and interaction beta
those who acted in more male-valued and male
values are signi?cant. The dependent variable
sex-speci?c ways tended to have higher DMS
for each analysis was the respondent’s score on
scores. The centered interaction terms did not
the DMS.1
add a signi?cant amount of variance to the
Gender-role trait analysis.
When the four
prediction of the criterion,
F(3, 500)
0.89,
main effects (Agency, Communion, Unmiti-
p
.05 ( R2
0.4%), suggesting that the
gated Agency, participant’s gender) were en-
relationships between these gender-typed be-
tered in Step 1, they predicted a signi?cant
haviors and DMS scores did not vary as a func-
amount of variance in DMS scores,
F(4,
tion of participant’s gender.
508)
28.48, p
.01 (adjusted R2
17.8%).
Univariate t tests revealed that two beta values
Discussion
were signi?cantly different from zero. Partici-
pant’s gender was negatively related to DMS
To determine the extent to which the drive for
scores (
–.38), t(3)
– 8.59, p
.01,
muscularity was predicted by gender-typed
meaning that men scored signi?cantly higher
traits and behaviors, two separate hierarchical
than women, after partialing out the effects of
moderating variable regression analyses were
the three gender-role dimensions. Unmitigated
performed. In the ?rst analysis, gender, gender-
Agency (i.e., being more concerned with the
typed traits, and their interaction terms were the
self than with others) scores were associated
predictors, whereas in the second analysis, gen-
with DMS scores (
.15), t(3)
3.19, p
der, gender-typed behaviors, and their interac-
.01, after controlling for participant’s gender
and the other two gender-role dimensions. The
tion terms were the independent variables. In
direction of the effect means that those who
both analyses, the participants’ drive for mus-
were higher in unmitigated agency had higher
cularity served as the criterion.
levels of the drive for muscularity. The addition
of the centered interaction terms in Step 2 did
1 McCreary et al. (2004) noted that Item 10 in the DMS
not add a signi?cant amount of variance to the
(“I thinkabout taking anabolic steroids”) did not load on
prediction of the criterion,
F(3, 500)
2.34,
any factor for either men or women and should not be
p
.05 ( R2
1.1%), suggesting that the
included in the DMS calculation. We included the item in
relationships between these three gender-typed
these analyses because the manuscript in which the recom-
mendation was made had not yet been published. However,
personality traits and DMS scores do not vary as
we also ran the analyses with Item 10 removed from the
a function of participant’s gender.
DMS score, and the results were almost identical.

---

88
MCCREARY, SAUCIER, AND COURTENAY
The results from these two analyses revealed
have identi?ed four dimensions in which this
four important ?ndings. First, in both analyses,
gender-based con?ict occurs: (a) success,
men scored higher than women on the DMS,
power, and competition; (b) restrictive sexual
after controlling for gender-typed traits or be-
and affectionate behavior between men; (c)
haviors. This is important because it shows that
restrictive emotionality; and (d) con?ict be-
the gender differences in DMS scores observed
tween workand family relations. Research
here (see Table 1) and in other research (e.g.,
has shown that men who experience con?ict
McCreary & Sasse, 2000; McCreary et al.,
in one or more of these dimensions tend to
2004) are fairly robust and are not mediated by
have lower self-esteem, lower levels of inti-
these two aspects of gender-role socialization.
macy, and higher levels of anxiety and de-
Second, the ?ndings revealed that three of the
pression (O’Neil et al., 1995). Men with high
four male-typed gender-role dimensions (Un-
gender-role con?ict also are less likely to seek
mitigated Agency, Male-Typed Behaviors, Male
help for problems with psychological well-
Sex-Speci?c Behaviors) all were positively as-
being (O’Neil et al., 1995), are more likely to
sociated with the drive for muscularity. Third,
engage in multiple health-riskbehaviors (Mc-
the lackof association between DMS scores and
feminine-typed traits and behaviors indicates
Creary & Courtenay, 2003), and tend to con-
that low levels of the drive for muscularity are
form more to male gender-role norms (Ma-
not associated with greater degrees of feminin-
haliket al., 2003).
ity (at least as measured by these two dimen-
The degree to which people have adopted
sions of feminine gender-role socialization).
traditional attitudes about men (also referred to
These two points have theoretical signi?cance
as male role norms or masculinity ideology)
because they show that it is higher levels of the
also is a signi?cant aspect of gender-role social-
male gender role, and not lower levels of the
ization. Traditional male role norms emphasize
female gender role, that are most closely asso-
men’s power, success, and toughness, as well as
ciated with the desire to be more muscular. The
the avoidance of anything deemed feminine.
fourth important ?nding was that the associa-
Adult and adolescent males who have adopted
tions between the drive for muscularity and
traditional attitudes about men have been shown
gender-typed traits and behaviors were similar
to engage in a wide range of risky health be-
for men and women. In other words, there was
haviors for which incidence rates are higher for
no differential salience in these relationships.
men; these include increased use of alcohol
The reasons for this are unclear at this point,
(McCreary et al., 1999), sedatives, and tranquil-
especially after ?nding a Gender
Gender
izers (Snell, Belk, & Hawkins, 1987); a greater
Role interaction in other studies (i.e., McCreary
likelihood of being sexually active; a greater
et al., 1999; McCreary & Sasse, 2000).
likelihood of having tricked or forced someone
Thus, it appears as though the association
to have sex with them; a greater likelihood of
between self-perceptions of masculinity and
engaging in unprotected sex and having had
muscularity that emerged from the earlier
more sexual partners; and a greater likelihood of
qualitative research (i.e., Grogan & Richards,
having been arrested (Pleck, Sonenstein, & Ku,
2002; Klein, 1993; Weinke, 1998) can be
1993, 1994). Traditional beliefs about men also
quanti?ed. However, gender-typed traits and
have been found to predict greater involvement
behaviors are not the only aspects of mascu-
linity that can be measured. In Study 2, we
in multiple high-riskbehaviors over time (Cour-
explore the relationships between the drive
tenay, 1998).
for muscularity and two other masculine gen-
Thus, Study 2 extends the ?ndings from
der-role
constructs:
gender-role
con?ict
Study 1 to these two dimensions of male gen-
(O’Neil et al., 1986) and traditional attitudes
der-role socialization. It was predicted that
about men. Gender-role con?ict is a construct
DMS scores would be positively associated
that explores the cognitive, emotional, uncon-
with gender-role con?ict and traditional atti-
scious, and behavioral problems men experi-
tudes about men. However, because the gender-
ence as they attempt to conform to society’s
role con?ict paradigm is appropriate only for
expectations of how men should be (O’Neil,
men, gender-role salience was not examined in
Good, & Holmes, 1995). O’Neil et al. (1986)
Study 2, as it was in Study 1.

---

DRIVE FOR MUSCULARITY
89
Study 2
restrictions, devaluation, or negative conse-
quences, both for one’s self or for other men
Method
(O’Neil et al., 1986, 1995). The GRCS con-
tains 37 items that are grouped into four sub-
Participants and Procedure
scales: (a) Success, Power, and Competition (13
items), which measures concern about personal
A total of 527 men (aged 17 to 54 years;
achievement and gaining dominance and supe-
Mdn
20 years, with only 1.3% of the sample
riority over others; (b) Restrictive Emotionality
being older than 22 years) were recruited from
(10 items), which assesses dif?culty and fears
an all-male college in the midwestern United
about emotional self-disclosure and verbal ex-
States. The participants reported being of
pression of feelings; (c) Restrictive Affectionate
mostly European American ethnicity (92%),
Behavior Between Men (8 items), which exam-
heterosexual (97%), and not married (84%), and
ines discomfort associated with verbal and
most reported living on campus (90%). Thirty
physical expressions of caring toward other
percent were in their ?rst year of college, 29%
men; and (d) Con?ict Between Workand Fam-
were in their sophomore year, 20% were in their
ily (6 items), which measures dif?culty and
junior year, and 21% were in their senior year.
dissatisfaction with the balancing of school or
Participants were recruited via an e-mail sent
workand family relations. Each item is rated on
to all students by staff at the university health
a 6-point scale, ranging from 1 (strongly dis-
clinic. The e-mail invited the students to partic-
agree) to 6 (strongly agree). Scores for each of
ipate in an anonymous online survey of their
the subscales were created by averaging their
health and health behaviors. The voluntary na-
respective items. Higher scores indicate a
ture of the survey was emphasized, as was the
higher degree of con?ict in each area. Previous
student’s right to withdraw participation at any
psychometric analyses of the GRCS have
time. After completing the survey, participants
shown it to be both reliable and valid (O’Neil et
were given the option of ?lling in a separate
al., 1995). In the present study, the alpha reli-
Web page with their name and contact informa-
ability coef?cients were .90, .90, .89, and .85,
tion to win a prize for a store voucher. Partici-
for each subscale, respectively.
pants were noti?ed that the data from the survey
Traditional attitudes about men.
To assess
and the identi?cation information from the form
traditional attitudes about men, the following
they completed to register for the prize went to
?ve items were included in the survey: (a) Do
separate computer servers and would not be
you believe that taking risks that are sometimes
combined. Of the 1,850 students enrolled at the
dangerous is part of what it means to be a man
college at that time, a total of 527 (29%) com-
and part of what distinguishes men from
pleted usable surveys.
women? (b) As a man, how important is it for
you to be self-suf?cient and always to try to
Measures
handle problems on your own? (c) As a man,
how important is it for you to be physically
In addition to the DMS, participants com-
strong and tough? (d) As a man, how important
pleted the GRCS (O’Neil et al., 1986) and a
is it for you to control your emotions and never
series of items that assessed traditional attitudes
to reveal sadness or vulnerability? and (e) As a
about men and men’s role in society. Other
man, how important is it for you not to engage
questions on the survey assessed health and
in activities that you thinkothers might consider
health riskbehaviors but are not reported here.
feminine? These items are similar to those used
The DMS.
The DMS (McCreary & Sasse,
on preexisting, albeit larger, measures of mas-
2000) was described thoroughly in Study 1.
culinity ideology (e.g., Male Role Norms Scale;
However, it should be noted that the DMS’s
Thompson & Pleck, 1986). Survey size restric-
degree of internal consistency was .89 in this
tions precluded our using one of these larger
sample.
measures. Each item was scored on a 4-point
The GRCS.
The GRCS is a commonly used
scale, from 1 (not at all true) to 4 (very true;
questionnaire designed to assess the psycholog-
Item 1) and 1 (not at all important) to 4 (very
ical con?ict men experience when rigid, sexist,
important; Items 2–5). All items were averaged
or restrictive gender roles result in personal
into a single index, and higher scores indicate a

---

90
MCCREARY, SAUCIER, AND COURTENAY
more traditional view of how men should be.
two dimensions of the GRCS and the measure
The alpha reliability estimate for the masculine
of traditional attitudes about men increased, so
ideology measure was .75.
did DMS scores.
Results
General Discussion
Descriptive Statistics
Two studies revealed signi?cant associations
Table 2 contains the means and standard de-
between the drive for muscularity and percep-
viations for the DMS, the four GRCS subscales,
tions of several aspects of masculine gender-
as well as the measure of traditional male
role socialization: unmitigatedly agentic per-
attitudes.
sonality traits, male-typed and sex-speci?c be-
haviors, traditional attitudes about men, and
Regression Analysis
gender-role con?ict in the areas of success,
power, and competition, as well as workand
To determine the extent to which the four
leisure con?ict. In all cases, DMS scores were
GRCS subscales and the measure of traditional
positively associated with the indices of mascu-
attitudes about men are predictive of young
linity. Thus, in Study 1, both men and women
men’s desire to be more muscular than they are
associated male-typed traits and behaviors with
currently, we performed a standard multiple re-
a need to be more muscular, whereas, in
gression analysis. Because there were no hy-
Study 2, men who endorsed more traditional
potheses about order of entry, all ?ve predictors
gender-typed beliefs wanted to be more muscu-
were entered at the same time. The predictors
lar. Finally, men in Study 2 who wanted to be
explained a signi?cant amount of variability in
more muscular experienced greater levels of
DMS scores, F(5, 501)
27.87, p
.01 (ad-
gender-role con?ict with respect to (a) society’s
justed R2
21.0%). An examination of the
expectations that they be successful, powerful,
individual standardized coef?cients revealed
and competitive and (b) balancing workand
that three betas were signi?cantly different from
leisure.
zero: the GRCS Success, Power, and Competi-
These ?ndings replicate and extend the lim-
tion subscale (
.26), t(4)
5.28, p
.01;
ited qualitative research that explored men’s
the GRCS Con?ict Between Workand Family
perceptions that muscular men are more mascu-
subscale (
.15), t(4)
3.44, p
.01; and
line and that men feel more masculine if they
the traditional attitudes about men scale (
gain more muscularity. However, as noted ear-
.14), t(4)
2.98, p
.01. The direction of all
lier, the main limitation of this qualitative re-
three beta values was positive: As scores on the
search is that although many people have an
intuitive understanding of their degree of mas-
culinity, this intuitive notion cannot be mea-
sured directly. Thus, the ?ndings from these two
Table 2
studies extend the existing literature by deter-
Means and Standard Deviations for the Drive for
mining which aspects of masculinity are
Muscularity Scale, Gender Role Con?ict Scale
uniquely associated with the drive for muscu-
(GRCS) Subscales, and a Measure ofTraditional
larity. Future research might explore the asso-
Masculine Ideology in a Sample ofYoung Men
(N
527)
ciations between masculinity and muscularity
even further. For example, whereas we chose to
Scale
M
SD
use a short single measure of traditional atti-
Drive for Muscularity Scale
2.42
0.79
tudes about men, there are other larger measures
GRCS Success, Power, and
of this construct that assess various subdimen-
Competition
3.59
0.88
sions of masculinity attitudes (e.g., toughness,
GRCS Restrictive Emotionality
2.97
0.97
violence or aggression, antifemininity, rejection
GRCS Restrictive Affectionate
Behavior Between Men
2.99
1.04
of homosexuals, attitudes toward sex). These
GRCS Con?ict Between Work
more comprehensive scales include the Male
and Family
3.47
1.05
Role Norms Scale (Thompson & Pleck, 1986),
Masculine ideology
2.28
0.57
the Brannon Masculinity Scale (Brannon &

---

DRIVE FOR MUSCULARITY
91
Juni, 1984), and the Male Role Norms Inven-
for muscularity the same across all age groups,
tory (Levant & Fischer, 1998). Other dimen-
or could the associations be stronger or weaker
sions of masculinity that can be explored in-
at speci?c times in our lives (or among speci?c
clude dominance (McCreary & Rhodes, 2001;
age cohorts)? For example, do people perceive a
Sidanius, Liu, Shaw, & Pratto, 1994), hyper-
stronger association between muscularity and
masculinity (Burk, Burkhart, & Sikorski, 2004;
masculinity in their earlier years than in their
Mosher & Sirkin, 1984), and conformity to
later years? Could the strength of these associ-
masculine norms (Mahaliket al., 2003). This
ations be linked to developmental tasks, in the
latter aspect of masculinity is new and espe-
same way that other researchers have linked
cially promising because it assesses the extent
developmental tasks with life span gender-role
to which men feel compelled to conform to
development (e.g., Gutmann, 1975; McCreary,
several aspects of the male gender role (e.g.,
1990; Sinnott, 1986)? Second, do the associa-
winning, emotional control, risktaking, power
tions between masculinity and muscularity vary
over women, being a playboy, self-reliance, pri-
across ethnic groups? One problem with testing
macy of work, pursuit of status, disdain for
this question is that it assumes that the various
homosexuals), as opposed to the stress and
dimensions of masculinity and the drive for
strain they experience as a result of acting in
muscularity mean the same thing to all ethnic
either stereotypically male-appropriate or male-
groups, which may not be the case (Doss &
inappropriate ways. Initial research suggests
Hopkins, 1998). If our measures of masculinity
that both the Conformity to Masculine Norms
and the drive for muscularity are equally valid
Inventory’s (CMNI) Winning subscale and the
in other ethnic groups, then multigroup compar-
overall CMNI score are positively correlated
isons across cultures should be conducted.
with DMS scores (Mahaliket al., 2003). How-
A third limitation stems from the low reli-
ever, the overall sample size for the DMS por-
ability coef?cients of some of the measures
tion of the CMNI study was small and pre-
used in Study 1. The alpha reliability coef?-
cluded the use of multivariate analyses.
cients for the women’s responses to the Unmit-
There are limitations to the present studies.
igated Agency scale and all three dimensions of
First, both studies were correlational in nature,
the SRBS, as well as men’s responses to the
and only the presence and direction of signi?-
Sex-Speci?c Behaviors SRBS subscale, were
cant relationships can be determined. Future
below .70. The immediate impact of this reduc-
research might explore the direction of causality
tion in internal consistency is to reduce the
in the associations between the drive for mus-
scales’ power to detect signi?cant associations.
cularity and perceptions of masculinity. For ex-
In other words, smaller effects that would have
ample, does an increase in muscularity cause an
been found to be statistically signi?cant if the
increase in perceived masculinity, and, if so,
scales’ alphas were higher did not reach the
which dimensions of masculinity are most
threshold for statistical signi?cance. This may
strongly in?uenced by this potential causal re-
have had implications for the predicted Gen-
lationship? Or, could increases in perceived
der
Gender Role interactions in Study 1.
masculinity (e.g., from changes in other aspects
These interactions were predicted based on the
of a person’s life) cause an increased desire to
notion of differential salience but failed to
become more muscular? Or, could the associa-
emerge. Perhaps with more reliable measures,
tion be cyclical? Experimental research is
which would increase the power of the analyses,
needed to address these questions.
the interactions would have been observed. De-
A second limitation is the restricted nature of
veloping new measures of these constructs may
the samples, especially with regard to age and
be required.
ethnicity. The samples used in these two studies
In summary, two studies demonstrated that
were predominantly college aged (18 –22 years)
the drive for muscularity is positively associ-
and from a European American ethnic back-
ated with several aspects of masculinity: unmit-
ground.2 This leads to several possible ques-
tions that can be examined empirically in future
2 Although ethnicity was not assessed in Study 1, the
research. First, are the associations between
student population at that university is mostly from a Eu-
these dimensions of masculinity and the drive
ropean American background.

---

92
MCCREARY, SAUCIER, AND COURTENAY
igated agency, male-typed and sex-speci?c be-
bulimia. International Journal ofEating Disor-
haviors, traditional attitudes about men, and two
ders, 2, 15–34.
aspects of gender-role con?ict (success, power,
Gillespie, B. L., & Eisler, R. M. (1992). Develop-
and competition and workand leisure con?ict).
ment of the Feminine Gender Role Stress Scale: A
These ?ndings reinforce those from qualitative
cognitive– behavioral measure of stress, appraisal,
and coping for women. Behavior Modi?cation, 16,
research in which men stated their belief that
426 – 438.
those with a greater degree of muscularity are
Grogan, S., & Richards, H. (2002). Body image:
more masculine and that gaining muscularity
Focus groups with boys and men. Men and Mas-
increases one’s masculinity.
culinities, 4, 219 –232.
Gutmann, D. (1975). Parenthood: A key to the com-
parative study of the life cycle. In N. Datan & L.
References
Ginsberg (Eds.), Life-span developmental psychol-
ogy: Normative life crises (pp. 167–187). New
Beere, C. A. (1990). Gender roles: A handbook of
York: Academic Press.
tests and measures. New York: Greenwood Press.
Huenemann, R. L., Shapiro, L. R., Hampton, M. C.,
Bem, S. L. (1974). The measurement of psycholog-
& Mitchell, B. (1966). A longitudinal study of
ical androgyny. Journal ofConsulting and Clinical
gross body composition and body con?rmation
Psychology, 42, 152–162.
and their association with food and activity in a
Brannon, R., & Juni, S. (1984). A scale for measuring
teen-age population. American Journal ofClinical
attitudes towards masculinity. JSAS Catalogue of
Nutrition, 18, 325–338.
Selected Documents in Psychology, 14, 6.
Jacobi, L., & Cash, T. F. (1994). In pursuit of the
Burk, L. R., Burkhart, B. R., & Sikorski, J. F. (2004).
perfect appearance: Discrepancies among self-
Construction and preliminary validation of the Au-
ideal perceptions of multiple physical attributes.
burn Differential Masculinity Inventory. Psychol-
Journal ofApplied Social Psychology, 24, 379 –
ogy ofMen and Masculinity, 5, 4 –17.
396.
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S.
Jones, D. C. (2001). Social comparison and body
(2002). Applied multiple regression/correlation
image: Attractiveness comparisons to models and
analysis for the behavioral sciences (3rd ed.).
peers among adolescent boys and girls. Sex
Mahwah, NJ: Erlbaum.
Roles, 45, 645– 664.
Courtenay, W. H. (1998). Better to die than cry? A
Klein, A. M. (1993). Little big men: Body building
longitudinal and constructionist study of masculin-
subculture and gender construction. Albany, NY:
ity and the health riskbehavior of young American
State University of New YorkPress.
men. (Doctoral dissertation, University of Califor-
Lavine, H., Sweeney, D., & Wagner, S. H. (1999).
nia at Berkeley, 1998). Dissertation Abstracts In-
Depicting women as sex objects in television ad-
ternational, 59, 232.
vertising: Effects on body dissatisfaction. Person-
Doss, B. D., & Hopkins, J. R. (1998). The Multicul-
ality and Social Psychology Bulletin, 25, 1049 –
tural Masculinity Ideology Scale: Validation from
1058.
three cultural perspectives. Sex Roles, 38, 719 –
Levant, R. F., & Fischer, J. (1998). The Male
741.
Drewnowski, A., & Yee, D. K. (1987). Men and body
Role Norms Inventory. In C. M. Davis, W. H.
image: Are males satis?ed with their body weight?
Yarber, R. Bauserman, G. Schreer, & S. L.
Psychosomatic Medicine, 49, 626 – 634.
Davis (Eds.), Sexuality-related measures: A
Eisler, R. M., & Skidmore, J. R. (1987). Masculine
compendium (2nd ed., pp. 469 – 472). Newbury
gender role stress: Scale development and compo-
Park, CA: Sage.
nent factors in the appraisal of stressful situations.
Mahalik, J. R., Locke, B. D., Ludlow, L. H., Diemer,
Behavior Modi?cation, 11, 123–136.
M. A., Scott, R. P. J., Gottfried, M., & Freitas, G.
Feingold, A., & Mazzella, R. (1998). Gender differ-
(2003). Development of the Conformity to Mas-
ences in body image are increasing. Psychological
culine Norms Inventory. Psychology ofMen and
Science, 9, 190 –195.
Masculinity, 4, 3–25.
Fisher, E., Dunn, M., & Thompson, J. K. (2002).
McCreary, D. R. (1990). Self-perceptions of life-
Social comparison and body image: An investiga-
span gender role development. International
tion of body comparison processes using multidi-
Journal ofAging and Human Development, 31,
mensional scaling. Journal ofSocial and Clinical
135–146.
Psychology, 21, 566 –579.
McCreary, D. R. (1994). The male role and avoiding
Garner, D. M., Olmstead, M. P., & Polivy, J. (1983).
femininity. Sex Roles, 31, 517–531.
Development and validation of a multidimensional
McCreary, D. R., & Courtenay, W. H. (2003, Au-
eating disorder inventory for anorexia nervosa and
gust). Gender role con?ict predicts engaging in

---