Exploratory Factor Analysis and Construct Validity of The Male Role Norms Inventory-Adolescent-revised (MRNI-A-r)

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

Psychology of Men & Masculinity © 2012 American Psychological Association

2012, Vol. 13, No. 4, 354 –366 1524-9220/12/$12.00 DOI: 10.1037/a0029102

Exploratory Factor Analysis and Construct Validity of the Male


Role Norms Inventory-Adolescent-revised (MRNI-A-r)
Ronald F. Levant, Baron K. Rogers, Bridgette Cruickshank, Thomas J. Rankin,
Britney A. Kurtz, Christina M. Rummell, Christine M. Williams, and
Alexander J. Colbow
The University of Akron

The Male Role Norms Inventory-Adolescent-revised (MRNI-A-r) was assessed for


This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

factor structure, convergent validity using the Adolescent Masculinity Ideology in


Relationships Scale, and discriminant validity using the Masculinity Scale of the Short
Form of Personal Attributes Questionnaire. Participants included 323 racially diverse
middle school boys and girls with a mean age of 12.8 years (SD ⫽ .92). Exploratory
factor analysis found a three factor structure: (1) Emotionally Detached Dominance; (2)
Toughness; and (3) Avoidance of Femininity. For the sample as a whole, the MRNI-A-r
and its factors had internal reliabilities ranging from .71 to .89. Some support was found
for convergent and discriminant validity. Boys endorsed traditional masculine norms to
a greater extent than did girls on the MRNI-A-r total scale and on factors 1 (Emotion-
ally Detached Dominance) and 2 (Toughness). African American boys endorsed
traditional masculine norms to a greater extent than did European American boys on the
MRNI-A-r total scale and factors 1 (Emotionally Detached Dominance) and 3 (Avoid-
ance of Femininity).

Keywords: masculinity ideology, male role norms, adolescents, measurement

Gender ideologies are beliefs about the roles gies—the central construct in the Strain Para-
and behaviors thought to be appropriate for digm.
either males or females. According to the Gen- Within this paradigm, masculinity ideology
der Role Strain Paradigm (Pleck, 1981, 1995), refers specifically to internalized beliefs regard-
these ideologies vary by culture. Within a given ing culturally defined standards or norms for
culture, the dominant gender ideologies influ- males’ roles and behaviors (Pleck, 1981, 1995).
ence how parents, teachers, and peers socialize Through social interactions resulting in rein-
children, and thus, how children think, feel, and forcement or punishment, masculinity ideology
behave in regard to gender-salient matters (Le- informs, encourages, and constrains boys and
vant, 1996a; Pleck, Sonenstein, & Ku, 1994b). men to conform to the prevailing male role
In the Gender Role Strain paradigm, gender norms by adopting certain socially sanctioned
roles and gendered behavior are thus thought to masculine behaviors and avoiding certain pro-
be the result of social learning and social influ- scribed behaviors (Levant, 1996a). Further,
ence processes, instructed by gender ideolo- since a given culture’s masculinity ideology is
also learned by females, it also influences wom-
en’s and girls’ ideas of how men and boys
This article was published Online First August 6, 2012. should behave.
Ronald F. Levant, Baron K. Rogers, Thomas J. Rankin, It is thus important to consider that both men
Britney A. Kurtz, Christina M. Rummell, Christine M. and women hold specific masculinity ideolo-
Williams, and Alexander J. Colbow, Department of Psy- gies. As Whorley and Addis (2006, p. 656)
chology, The University of Akron; Bridgette Cruickshank,
Department of Counseling, The University of Akron. noted: “when we study masculinity only in men
The authors gratefully acknowledge the assistance of Dr. we can easily fall prey to an implicit essential-
Sandra Spickard Prettyman, Steve Graef, and Amanda Beck ism by failing to distinguish . . . [sex and gen-
of the University of Akron. der]; thoroughly studying masculinity means
Correspondence concerning this article should be ad-
dressed to Ronald F. Levant, Department of Psychology, understanding how it operates in the lives of
The University of Akron, Akron, OH 44325-4301. E-mail: both men and women.” Accordingly, we must
levant@uakron.edu consider that males and females may have dif-
354
MRNI-A-R 355

ferent views regarding the appropriate norms Measuring the Masculinity Ideologies of
for male behavior. In some societies (e.g., China Adolescents
and Russia), men and women have been found
to be quite similar in their views of the norms To date, only a few studies have reported
for male behavior, with both genders endorsing attempts to measure the masculinity ideologies
the traditional masculine norms as the appropri- of adolescents. Pleck, Sonenstein, and Ku
ate norms for men’s behavior (Levant & Rich- (1993a; 1993b) developed the 8-item Male Role
mond, 2007). However, in the U.S., adult Attitude Scale (MRAS) to be used with older
women tend to reject the traditional norms for adolescent boys aged 15–19. Seven items were
men’s behavior to a much greater extent that do adapted from Thompson and Pleck’s (1986)
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

adult men (Levant & Richmond, 2007). These Male Role Norm Scale (MRNS), a 26-item ab-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

differences are thought to reflect the influence breviated version of the Brannon Masculinity
of feminism on U. S. women (and the lack, Scale, Short Form (Brannon & Juni, 1984)
thereof, on Chinese and Russian women), and which was designed for adults. Items most rel-
suggest that traditional masculinity ideology evant to the adolescent population were selected
may be a point of contention between the gen- and the wordings of items were simplified.
ders (Levant, 1996b, 1997). Given these gender Items were chosen to represent the three facto-
differences, it is important to study masculinity rial dimensions of the MRNS: Status, Tough-
ideologies (views regarding appropriate norms ness, and Anti-Femininity. Further, an addi-
for male behavior) in both males and females. tional item about sex, a topic left out of the
Researchers have created a variety of mea- MRNS, was also included. While such brief
sures to operationalize and investigate varia- scales are useful for population-based epidemi-
tions in, and correlates of, the endorsement of ological surveys (which is what the MRAS was
both traditional and nontraditional masculinity intended for), they do not allow the more fine-
ideologies (Thompson & Pleck, 1995; Thomp- grained analyses that a broader set of subscales
son, Pleck, & Ferrera, 1992). The Male Role would provide. Limitations of this instrument
Norms Inventory (MRNI) defined seven male include, thus, the small number of items which
role norms, or dimensions of traditional mascu- covers only a small set of the male role norms,
linity ideology: Avoidance of Femininity, Fear and also its low reliability (the coefficient alpha
and Hatred of Homosexuals, Self-Reliance, Ag- for the scale is .56 as reported by Pleck, Sonen-
gression, Achievement/Status, Non-Relational stein, & Ku, 1994a), and the fact it has not been
Attitudes toward Sexuality, and Restrictive developed to be administered to younger teens
Emotionality (Levant et al., 1992). The MRNI nor teenage girls.
was designed to survey both men and women; Chu, Porche, and Tolman (2005) developed
however, like most measures of masculinity the Adolescent Masculinity Ideology in Rela-
ideology, it was designed for use with adults. tionships Scale (AMIRS), a scale that assesses
Socialization processes begin at a young age attitudes and beliefs about appropriate behavior
and seem to have a great impact on the de- for males within interpersonal relationships.
velopment of gender roles and gender-typed Designed specifically for use with adolescent
behaviors for children and adolescents that boys (aged 12–18), the AMIRS was derived
continue into adulthood (Levant, 2001). Fur- from adolescent boys’ narratives about their
thermore, during early adolescence many perceptions and experiences of masculinity,
gender-related issues arise, such as the inten- particularly in their peer relationships. Interest-
sification of concerns about gender (Galam- ingly, the AMIRS was found to be negatively
bos, Almeida, & Peterson, 1990), interests in associated with self-esteem, which suggests the
sexuality, concerns about aggression, sub- problematic consequences that may be associ-
stance use, and career development, to name a ated with boys’ endorsement of traditional mas-
few. In addition, early adolescence is a time culinity ideology. Limitations of this instrument
of heightened conformity in many domains, include its focus on relationships, which covers
especially those related to gender. Hence, it only a subset of male role norms, and the fact
seems particularly important to study mascu- that it has not been designed to be administered
linity ideology in young adolescents. to teenage girls.
356 LEVANT ET AL.

An adolescent version of the MRNI has been line norms are less important among teens in
developed, the Male Role Norms Inventory- Scotland as compared to the U.S.
Adolescent (MRNI-A; Brown, 2002). The read-
ing and comprehension levels of the MRNI Revision of the MRNI-A
were adapted in order to create an instrument
developmentally appropriate for use with Given the limitations in subscale reliability
younger populations. Items for the adolescent and discriminant validity for boys, the MRNI-A
version were based on MRNI, but were changed was revised in the hopes of improving its reli-
in two major ways. First, items were adapted ability and validity. Using the American data set
and new items were created to represent ado- from Levant et al. (2008), the investigators an-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

lescent-specific contexts, for example, “It is ok alyzed all of the MRNI-A items to determine if
This document is copyrighted by the American Psychological Association or one of its allied publishers.

for a boy to ask for help fixing a flat tire on his removing the item would improve the coeffi-
bike,” and “Boys should not be afraid to go cient alpha of its respective subscales, and, as a
inside a ‘haunted house.’” We did include an result, removed two items—items which had
item related to work (“A boy must be able to also been judged to have problems in wording
make his own way in the world.”). We chose to (items 38 and 42). We next simplified the lan-
emphasize sports (six items), as sports is a cen- guage (e.g., changed “If necessary a boy should
tral theme in boy’s lives (Messner, 1992). We sacrifice friendships in order to get ahead” to “if
chose to not mention the school context, be- needed, a guy should stop being friends with
cause school for many boys is associated with a someone to be more popular”) and improved the
sense of failure (Pollack, 1998), and we did not age-appropriateness of the remaining items
(e.g., “boys” was changed to “guys” for all
want to arouse negative feelings associated with
items) while preserving the content of items.
failure. Second, while certainly relevant to the
We also reworded all negatively worded (re-
adolescent population, the scale developers de-
verse-scored) items to be positively worded, as
cided to not include the Fear and Hatred of
recommended by DiStefano and Motl (2006).
Homosexuals and Non-Relational Attitudes To-
These investigators demonstrated that, contrary
ward Sexuality subscales, because they believed to the view that a mix of positively- and nega-
that including such scales might create an in- tively-worded items prevents response bias,
surmountable obstacle to gaining approval to negatively-worded items on self-report surveys
administer the inventory in schools with these actually tap method effects. The result was the
items included. Hence the MRNI-A had five MRNI-A-revised (MRNI-A-r), a 41-item inven-
subscales: Restrictive Emotionality, Avoidance tory with the same five subscales as the
of Femininity, Aggression, Achievement/ MRNI-A.
Status, and Self-Reliance.
Data were collected from American (N ⫽ Purpose and Hypotheses of the Present
172) and Scottish (N ⫽ 264) adolescent boys Study
and girls to evaluate the scale reliability and
construct validity of the MRNI-A (Levant, The purpose of the present study was to as-
Graef, Smalley, Williams, & McMillan, 2008). sess the factor structure of the MRNI-A-r using
Results indicated that the MRNI-A showed exploratory factor analysis, assess its reliability
good overall internal consistency for the scale and convergent and discriminant validity
as a whole in both samples, but that the reliabil- among boys and girls, and to examine how
ity of the subscales ranged from just barely response patterns differ by gender and race/
adequate to poor. Evidence for convergent va- ethnicity.
lidity for the MRNI-A was found for both boys We first hypothesized that the five-factor
and girls, and evidence for discriminant validity structure for the MRNI-A-r would be supported,
was found for girls but not for boys. Consistent corresponding to the five subscales of the
with research on adults, females in both samples MRNI-A-r. We also planned to examine the
endorsed less traditional views of masculinity internal consistency reliability of the resultant
than did males. Scottish adolescents endorsed factors of the MRNI- A-r and of the total scale.
less traditional views of masculinity than did Second, we hypothesized that evidence
Americans, suggesting that traditional mascu- would be found for the convergent validity of
MRNI-A-R 357

the MRNI-A-r by the instrument showing a fensive adaptation of strongly endorsing the
significant strong correlation with another mea- norms of dominant masculinity, one well-
sure of traditional masculinity ideology, the Ad- known example of which is “Cool Pose” (Ma-
olescent Masculinity Ideology in Relationships jors & Billson, 1992).
Scale (AMIRS; Chu et al., 2005). Since the
AMIRS is scaled in the same direction of the Method
MRNI-A-r, and in both cases higher scores in-
dicate greater endorsement of traditional mas- Procedures
culinity ideology, we would expect to find a
positive correlation. The rationale for this hy- Permission was sought and granted to collect
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

pothesis is that both measures, consistent with data in three school systems from their respec-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

tenets of the Gender Role Strain Paradigm, take tive school administrations. Following approval
a normative approach to studying gender roles, by the University of Akron IRB, parental in-
and are designed to measure individuals’ atti- formed consent and child assent were obtained
tudes toward and beliefs about these roles. by research assistants going into the schools and
Third, we hypothesized that evidence would describing the study and the benefits of trying to
be found for the discriminant validity of the better understand masculinity and gender roles.
MRNI-A by the instrument showing only small Research assistants distributed the question-
positive correlations with a masculinity mea- naires to those participants whose parents con-
sure that views masculinity from a different sented and who assented to participate, which
theoretical perspective, one which conceptual- they completed during class time. Anonymity of
izes gender as a set of personality traits that are results was protected by the separation of in-
appropriate for each sex. For this we used the formed consent and identifying information
Masculinity Scale of the Short Form of the from questionnaire results at the time of collec-
Personal Attributes Questionnaire (PAQ-M; tion, such that no one is able to match personal
Spence & Helmreich, 1978). Our rationale for data to survey responses.
this hypothesis is that, although both the
PAQ-M and the MRNI-A-r are measures of Participants
masculinity, they approach it from different the- A racially diverse sample was sought. Hence,
oretical orientations, and hence would be un- data were gathered at three middle schools, one
likely to be highly correlated with each other. of which was predominantly white (N ⫽ 106),
The PAQ-M conceptualizes masculinity as a set one of which was mostly white but had a more
of sex-typed instrumental personality traits, mixed racial composition (N ⫽ 143), and one of
whereas the MRNI-A-r conceptualizes gender which was predominantly Black (N ⫽ 74).
as a set of culturally defined normative expec- Three hundred twenty-three students partici-
tations for behavior. Specifically, we hypothe- pated: 162 males, 157 females, and four who
sized that evidence for discriminant validity of did not indicate their gender. The students had a
the MRNI-A-r will be found by finding only mean age of 12.8 years (SD ⫽ .92). Two hun-
small-to-moderate positive correlations with the dred and 15 students (66.5%) listed their race/
PAQ-M for both boys and girls. ethnicity as white/European American, 61
Fourth, we hypothesized that the response (18.9%) as Black/African American, 19 (5.9%)
patterns would differ by gender and race/ as Bi/Multiracial, and nine (2.8%) or fewer as
ethnicity. Based on prior research on the MRNI either Latino(a)/Hispanic, American Indian,
(Levant & Richmond, 2007) and the MRNI-A Asian/Asian American, Pacific Islander/Inuit,
(Levant et al., 2008) we hypothesized that boys’ Other, or did not respond to this question.
scores would be more traditional than that of
girls, and that African American teens’ scores Measures
would be more traditional than that of European
American teens’ scores. These race/ethnicity Male Role Norms Inventory-Adolescent-
differences are postulated, in the theory of he- revised (MRNI-A-r). The MRNI-A-r, a re-
gemonic masculinity (Connell & Messer- vision of the Male Role Norms Inventory-
schmidt, 2005), to arise from the marginaliza- Adolescent (MRNI-A; Brown, 2002; Levant et
tion of males of color, and their resulting de- al., 2008) designed to measure individuals’ be-
358 LEVANT ET AL.

liefs about appropriate behavior for adolescent increase the sensitivity of the scale. As such, our
boys. The MRNI-A-r is a 41-item inventory in 9-item version resulted in a coefficient alpha of
which participants indicate the extent of their .72 for boys and .62 for girls.
agreement or disagreement with statements Masculinity Scale of the Short Form of
concerning beliefs about how boys ought to The Personal Attributes Questionnaire
behave on a 7-point Likert scale (where 1 ⫽ (PAQ-M; Spence & Helmreich, 1978). Dis-
Strongly Disagree to 7 ⫽ Strongly Agree) in criminant validity was assessed using the
which higher scores correspond to more tradi- PAQ-M. The PAQ-M is an 8-item scale which
tional views. The five subscales are: Avoidance is theoretically consistent with Gender Role
of Femininity (eight items, e.g., “A guy should Identity Paradigm (Pleck, 1981, 1995), and as-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

prefer football to sewing”), Self-Reliance (six sesses how strongly individuals rate themselves
This document is copyrighted by the American Psychological Association or one of its allied publishers.

items, e.g., “Guys should always be able to as having stereotypical masculine (instrumen-
figure out what they should do”), Aggression tality/agency) personality traits. Participants
(eight items, e.g., “If someone else starts it, a rate themselves on a 5-point scale which is
guy should be allowed to use violence to defend anchored by two dichotomous personality attri-
himself”), Achievement/Status (eight items, butes (e.g., “Not at all competitive/Very com-
e.g., “Guys should do whatever it takes to be petitive”). Items are scored so that higher scores
cool”), and Restrictive Emotionality (11 items, indicate higher stereotypical masculine (instru-
e.g., “Guys shouldn’t cry, especially in front of mentality/agency) personality traits. Cronbach
others”). coefficient alpha of .65 for PAQ-M has been
Adolescent Masculinity Ideology in Rela- shown in a prior study (Robitschek, 2003). For
tionships Scale (AMIRS; Chu et al., 2005). this study, the PAQ-M coefficient alpha was .70
The AMIRS is a 12-item inventory that assesses for boys and .62 for girls.
participants’ attitudes toward appropriate mas- Demographic information sheet. On a
culine behavior in interpersonal relationships, in demographic information sheet developed for
which participants indicate the extent of their this study, participants were asked to report
agreement or disagreement on a 4-point Likert their sex/gender, age, and race/ethnicity.
Scale (where 1 ⫽ disagree a lot to 4 ⫽ agree a
lot), with statements concerning these attitudes. Results
After recoding three reverse-scaled items,
higher scores correspond to more traditional Data Cleaning and Missing Values
views. A sample item is “It’s important for a
guy to act like nothing is wrong, even when Before conducting statistical analyses the
something is bothering him.” The scale devel- data were thoroughly screened to ensure the
opers reported a coefficient alpha for the scale accuracy of the data file. There were missing
averaged over several samples ⫽ .70, evidence data as some participants did not respond to
for convergent and concurrent validity though every item. The data set consisted of 58 items
correlations of the AMIRS with measures of (41 for the MRNI-A-r, eight for the PAQ-M,
normative masculinity, attitudes toward and nine for the AMIRS) which had been re-
women, and self-esteem, and evidence for dis- sponded to by 323 participants, for a total
criminant validity through a comparison of the of 18,734 data points, of which 147 were miss-
patterns of correlations of the AMIRS and a ing (0.8%), with the number of missing re-
measure of masculine personality traits with sponses per item ranging from 0 to 27. No
measures of normative masculinity, attitudes to- discernable patterns were found by visually in-
ward women, and self-esteem (Chu et al., specting the missing data, which suggested that
2005). We modified the AMIRS in two ways. they were missing at random. With a small
First, we removed two items related to sex and percentage of data points missing in a random
one item related to homosexuality because these pattern from a large data set, the problem is less
topics are sufficiently controversial that it serious, according to Tabachnick and Fidell
would be unlikely to gain approval for distribu- (2007). Nonetheless, we took a conservative
tion in a middle school. Second we changed the approach, and replaced missing values using
scaling from a 4-point scale to a 7-point scale SPSS-17’s Linear Trend at Point method. This
(1 ⫽ Disagree a lot to 7 ⫽ Agree a lot) to is a regression-based imputation method, in
MRNI-A-R 359

which other variables are used as IV’s to write puted for factors ranging from one to five. The
regression equations for the variable with miss- test involves determining the point at which the
ing data serving as DV. Cases with complete actual raw data eigenvalue is lower than either
data generate the regression equation. The equa- the mean or the 95th percentile eigenvalue.
tion is then used to predict the missing values There is some debate as to whether comparisons
for the incomplete cases. should be made to the mean or the 95th percen-
tile eigenvalue (Glorfeld, 1995). Some recom-
Test of Hypothesis 1 mend the latter, on the grounds that observed
eigenvalues greater than the 95th percentile are,
We hypothesized that exploratory factor in effect, significant at the 5% level (Hayton,
analysis of the MRNI-A-r would support the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Allen, & Scarpello, 2004). An inspection of


This document is copyrighted by the American Psychological Association or one of its allied publishers.

five subscale structure as represented by the five Table 1 indicates that when the number of fac-
subscales. Prior to the analysis, the suitability of tors moves from 3 to 4, the raw data eigenvalue
data for factor analysis was assessed. The sam- changes from being larger than both the average
ple size of 323 allows for 7.9 participants per and 95th percentile eigenvalue at 3 factors, to
item. Costello and Osborne (2005) found being smaller than both the average and 95th
that 62.3% of the factor analyses they surveyed percentile eigenvalue at 4 factors, thus indicat-
had a participants per item ratio in the range ing that a four factor structure best represents
of ⱖ 5:1 and ⱕ 10:1. Tabachnik and Fidell the dimensionality of the data.
(2007, p. 613) note that “As a general rule of An exploratory factor analysis of the 41 MRNI-
thumb, it is comforting to have at least 300 A-r items was carried out using Principal Axis
cases for factor analysis.” The Kaiser-Meyer- Factoring (PAF). We selected Principal Axis Fac-
Olkin value was .89, which exceeds the sug- toring (PAF) rather than Principal Components
gested value of .6 (Kaiser, 1974). Bartlett’s Test Analysis (PCA) because PAF analyzes only com-
of Sphericity (Bartlett, 1954) was statistically mon variance in its search for underlying latent
significant, again further supporting the fac- structure; in contrast, PCA analyzes common,
torability of the correlation matrix. unique, and error variance and is thus better
To determine the number of factors to ex- characterized as a data reduction technique than
tract, we noted Kahn’s (2006) critique of such a factor analytic one (Kahn, 2006). Given our
commonly used methods as Kaiser’s criterion goal of understanding the latent structure of the
(eigenvalues greater than 1) and Catell’s scree items and subscales comprising the MRNI-A-r,
test, and followed his recommendation to use PAF seemed the more appropriate technique.
parallel analysis, which was performed using Finally, we chose PAF rather than Maximum
O’Connor’s (2000) SPSS computer program. Likelihood because PAF does not assume mul-
One thousand randomized permutations of the tivariate normality (Fabrigar, Wegener, MacCa-
MRNI-A-r data were generated. Table 1 shows llum, & Strahan, 1999), which was important
the mean and 95th percentile eigenvalues de- given that some MRNI-A-r items evoke re-
rived from these extensive repetitions, as well sponses consistent with highly traditional mas-
as the actual eigenvalues computed from the culinity ideology, thus producing item skew
original raw data, all three of which were com- which would violate the assumption of multi-
variate normality. We fixed the number of fac-
tors to extract at four, and using Promax rota-
Table 1 tion with Kaiser Normalization. This oblique
Parallel Analysis of the MRNI-A-R Scores rotation criterion was chosen following
Using 1,000 Random Permutations of the Preacher and MacCallum (2003), who empha-
Raw Data
sized that orthogonal criteria “are rarely defen-
Mean random 95th percentile Raw data sible because factors are rarely if ever uncorre-
Factor eigenvalues eigenvalues eigenvalues lated in empirical studies” (p. 40).
1 1.78 1.93 10.12 The four factors accounted for 40.11% of the
2 1.69 1.66 2.95 variance. We set the minimum loading allow-
3 1.62 1.61 1.90 able to .35, which resulted in the removal of
4 1.56 1.57 1.47 eight items. In addition, four items were re-
5 1.51 1.50 1.32
moved because of cross-loading problems, fol-
360 LEVANT ET AL.

lowing Tabachnick’s and Fidell’s (2007) sug- drawn from the original Restrictive Emotional-
gestion that items that load .32 or greater on a ity subscale, six from the Achievement/Status
second factor should be removed. The fourth subscale, two from the Self-Reliance subscale,
factor was not viable, as it had only three items, and one from the Avoidance of Femininity sub-
one of which was eliminated because the load- scale. Based on the content of the items we
ing was less than .35, and two of which were named this factor “Emotionally Detached Dom-
eliminated because of cross-loading problems. inance.” This reflects the fact that the prepon-
The resulting 29-item 3-factor scale with factor derance (81%) of the items came from two
loadings, which accounts for 36.51% of the subscales, Restrictive Emotionality (reflecting
variance, is displayed in Table 2. emotional detachment) and Achievement/Status
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Factor 1, accounting for 24.68% of the vari- (reflecting dominance). These two themes are
This document is copyrighted by the American Psychological Association or one of its allied publishers.

ance, included 16 items, of which seven were also closely related, in that being emotionally

Table 2
MRNI-A-R Factors and Loadings From Principal Axis Factoring (Pattern Matrix)
Factor (Percent Variance Accounted for; Alpha)
Item Loading
Factor 1: Emotionally Detached Dominance (24.68%; ␣ ⫽ .89)
12. Guys should never tell others if they’re worried or afraid. (RE) .86
16. Guys shouldn’t ever show their feelings. (RE) .78
10. When in a group of guys and girls, guys should always make the final decision. (AS) .71
8. Guys shouldn’t cry, especially in front of others. (RE) .66
36. If a guy is in pain, it’s better for him to keep it to himself rather than to let people know. (RE) .64
26. Chores like doing the laundry or cooking aren’t for guys. (AF) .58
2. Guys should do whatever it takes to be cool. (AS) .52
33. Guys shouldn’t show fear. (RE) .51
28. Guys should not tell their friends they care about them. (RE) .49
14. A guy should win at any game he plays. (AS) .49
1. If needed, a guy should stop being friends with someone to be more popular. (AS) .45
5. A guy should never depend on someone else to help him. (SR) .44
11. It is not ok for a guy to ask for help fixing a flat tire on his bike. (SR) .44
30. It’s important to have the newest video game system. (AS) .43
34. When they’re sad or upset, guys should just “suck it up” and get over it. (RE) .41
20. In a group of guys and girls, it is up to the guys to get things organized and moving ahead. (AS) .37
Factor 2: Toughness (7.20%; ␣ ⫽ .71)
27. It’s important for a guy to be able to play it cool. (RE) .62
25. When the going gets tough, guys get tough. (A) .56
23. If someone else starts it, a guy should be allowed to use violence to defend himself. (A) .54
40. It’s important for guys to try hard to be the best. (AS) .50
17. A guy who can’t make up his mind will not be respected. (SR) .46
35. Boys should not throw baseballs “like a girl.” (AF) .39
39. A guy with no interest in adventure is not very cool. (A) .39
Factor 3: Avoidance of Femininity (4.63%; ␣ ⫽ .75)
32. Guys shouldn’t carry purses. (AF) .76
21. It is too girlish for a guy to wear make-up. (AF) .76
29. Guys should play with trucks rather than dolls. (AF) .64
18. Guys should not be allowed to wear skirts. (AF) .62
22. Sports like hockey and wrestling should be only played by boys. (AF) .40
4. A boy should prefer football to sewing. (AF) .40
Note. N ⫽ 162 males, 157 females, and four who did not indicate their gender. Capital letters in parentheses following
each item refer to the original subscales from which the item was drawn: AF ⫽ Avoidance of Femininity; SR ⫽
Self-Reliance; A ⫽ Aggression; AS ⫽ Achievement/Status; and RE ⫽ Restrictive Emotionality. We used the same numbers
for the items in the MRNI-A-r that we did for the corresponding items in MRNI-A (Levant et al., 2008), even though the
MRNI-A-r does not include items 38 and 42 from the MRNI-A.
MRNI-A-R 361

detached serves the purpose of being domi- “It’s important for guys to try hard to be the
nant—that is, it is hard to dominate someone if best.”
you are emotionally vulnerable to them. The Factor 3, accounting for 4.63% of the vari-
two items from the Self-Reliance subscale are ance, included six items, all of which were
related to the emotionally detached theme in drawn from the original Avoidance of Feminin-
that they reflect a denial of neediness. The one ity subscale. Based on the content of the items
item from the Avoidance of Femininity sub- we named this factor “Avoidance of Feminin-
scale related to the dominance theme in that it ity.”
reflects an unwillingness to do women’s work
which was beneath them. Internal Consistency Reliability
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Factor 2, accounting for 7.20% of the vari-


This document is copyrighted by the American Psychological Association or one of its allied publishers.

ance, included seven items, of which three were We assessed the internal consistency reliabil-
drawn from the original Aggression subscale, ity of the factors of the MRNI- A-r and the total
and one each from the Restrictive Emotionality, scale. Over the entire sample the alpha for the
Achievement/Status, Self- Reliance, and Avoid- total 29-item scale was .89, the alpha for fac-
ance of Femininity subscales. Based on the con- tor 1, Emotionally Detached Dominance, was
tent of the items we named this factor “Tough- .89, the alpha for factor 2, Toughness, was .71,
ness,” which reflects the fact that almost half and the alpha for factor 3, Avoidance of Femi-
(43%) of the items came from the Aggression ninity, was .75. Alphas for boys and girls on the
subscale, and the other four items (originating total scale and its factors, which range from .68
in various subscales) reflect aspects of tough- to .88, are shown in Table 3.
ness. Three items reflect an “absence of softness
Descriptive Statistics
or sentimentality,” which is part of the defini-
tion of toughness according to the Free Mer- Descriptive statistics of study variables, in-
riam-Webster Dictionary (Tough, 2010): “It’s cluding the three factors of the MRNI-A-r, are
important for a guy to be able to play it cool”; presented in Table 3.
“A guy who can’t make up his mind will not be
respected;” and “Boys should not throw base- Test of Hypothesis 2
balls ‘like a girl.’” One item reflects “uncom-
promising determination,” which is also part of We hypothesized that evidence would be
the definition of toughness according to the Free found for the convergent validity of the MRNI-
Merriam-Webster Dictionary (Tough, 2010): A-r by the instrument showing significant and

Table 3
Means, Standard Deviations, Cronbach Alphas, and F, p, and Eta2 Values of Boys and Girls Scores on
Study Variables
Boys Girls
Scale M SD ␣ M SD ␣ F p Eta2
MRNI-A-r 3.99 .91 .88 3.30 .80 .87 51.27 ⬍.001 .14
Factor 1 3.36 1.07 .87 2.44 .87 .86 69.35 ⬍.001 .18
Factor 2 4.56 1.07 .71 4.02 1.04 .68 21.58 ⬍.001 .06
Factor 3 5.00 1.39 .74 4.75 1.36 .75 2.62 n.s. .01
AMIRS 3.53 1.03 .72 3.04 .89 .64 20.43 ⬍.001 .06
PAQ-M 3.76 .64 .70 3.68 .57 .62 1.48 n.s. .01
Note. N ⫽ 162 males, 157 females, and four who did not indicate their gender. MRNI-A-r ⫽ Male Role Norms
Inventory—Adolescent-revised total factored 29-item scale. Scores for the MRNI-A-r and its factors ranged from 1–7, with
higher scores indicating greater endorsement of traditional masculinity ideology. Factor 1 ⫽ Factor 1 of the MRNI-A-r,
Emotionally Detached Dominance; Factor 2 ⫽ Factor 2 of the MRNI-A-r, Toughness l; Factor 3 ⫽ Factor 3 of the
MRNI-A-r, Avoidance of Femininity; AMIRS ⫽ the Adolescent Masculinity Ideology in Relationships Scale. Scores for
the AMIRS ranged from 1–7, with higher scores indicating greater endorsement of traditional masculinity ideology.
PAQ-M ⫽ Personal Attributes Questionnaire⫺Male. Scores for the PAQ-M ranged from 1–5, with higher scores indicating
greater self-reported masculine-valued or instrumental personality traits.
362 LEVANT ET AL.

Table 4 Femininity) did not show significant correla-


Bivariate Correlation Coefficients of Boys’ and tions with the AMIRS for either boys or girls.
Girls’ Scores on Study Variables
Scale 1 2 3 4 5 6 Test of Hypothesis 3
ⴱⴱ ⴱⴱ ⴱⴱ ⴱⴱ ⴱ
1. MRNI-A-r — .92 .77 .59 .62 .20
We hypothesized that evidence would be
2. Factor 1 .88ⴱⴱ — .61ⴱⴱ .30ⴱⴱ .69ⴱⴱ .16ⴱ
3. Factor 2 .78ⴱⴱ .58ⴱⴱ — .28ⴱⴱ .44ⴱⴱ .33ⴱⴱ found for the discriminant validity of the
4. Factor 3 .63ⴱⴱ .29ⴱⴱ .34ⴱⴱ — .15 .10 MRNI-A by finding small positive correlations
5. AMIRS .42ⴱⴱ .46ⴱⴱ .27ⴱⴱ .15 — .05 with the PAQ-M for both boys and girls. Bi-
6. PAQ-M .28ⴱⴱ .19ⴱ .34ⴱⴱ .16ⴱ .11 — variate correlation coefficients of boys’ and
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Note. N ⫽ 162 boys, 157 girls, and four who did not girls’ scores on study variables are displayed in
This document is copyrighted by the American Psychological Association or one of its allied publishers.

indicate their gender. Correlations for boys are above the Table 4. For boys, the MRNI-A-r total scale and
diagonal; correlations for girls are below the diagonal; factors 1 and 3 (Emotionally Detached Domi-
MRNI-A-r ⫽ Male Role Norms Inventory—Adolescent- nance and Avoidance of Femininity, respec-
revised. Factor 1 ⫽ Factor 1 of the MRNI-A-r, Emotionally
Detached Dominance; Factor 2 ⫽ Factor 2 of the MRNI-
tively) showed small positive correlations with
A-r, Toughness; Factor 3 ⫽ Factor 3 of the MRNI-A-r, the PAQ-M (.10 to .20), whereas factor 2
Avoidance of Femininity; AMIRS ⫽ the Adolescent Mas- (Toughness) showed a moderate positive corre-
culinity Ideology in Relationships Scale; PAQ-M ⫽ Per- lation with the PAQ-M (.33). For girls, only
sonal Attributes Questionnaire⫺Male. factors 1 and 3 (Emotionally Detached Domi-

p ⬍ .05, two-tailed. ⴱⴱ p ⬍ .01, two-tailed.
nance and Avoidance of Femininity, respec-
tively), showed small positive correlations (.16
to .19) with the PAQ-M, whereas the MRNI-A-r
strong positive correlations with the AMIRS. total scale and factor 2 (Toughness) showed
Bivariate correlation coefficients of boys’ and moderate positive correlations with the PAQ-M
girls’ scores on study variables are displayed in (.28 to .34).
Table 4. For both boys and girls, the MRNI-A-r
total scale and factors 1 and 2 (Emotionally Test of Hypothesis 4
Detached Dominance and Toughness, respec-
tively) showed significant and strong positive As shown in Table 3, significant differences
correlations with the AMIRS, although the cor- emerged between the boys and girls on the
relations were larger for boys (.44 to .69) than MRNI-A-r total scale and the first two factors
for girls (.27 to .46). Factor 3 (Avoidance of (Emotionally Detached Dominance and Tough-

Table 5
Means, Standard Deviations, and F, p, and Eta2 Values of African American and European American
Adolescents Scores on Study Variables
European
African Americans Americans
Scale M SD M SD F p Eta2
MRNI-A-r 3.94 .91 3.56 .85 9.27 .003 .03
Factor 1 3.31 1.15 2.77 .99 12.98 ⬍.001 .04
Factor 2 4.24 1.01 4.26 1.01 .01 n.s. .00
Factor 3 5.27 1.16 4.84 1.34 5.13 .024 .02
AMIRS 3.21 1.13 3.32 .91 .64 n.s. .00
PAQ-M 3.57 .71 3.74 .56 3.72 n.s. .01
Note. N ⫽ 61 African Americans and 215 European Americans. MRNI-A-r ⫽ Male Role Norms Inventory—Adolescent-
revised total factored 29-item scale. Scores for the MRNI-A-r and its factors ranged from 1–7, with higher scores indicating
greater endorsement of traditional masculinity ideology. Factor 1 ⫽ Factor 1 of the MRNI-A-r, Emotionally Detached
Dominance; Factor 2 ⫽ Factor 2 of the MRNI-A-r, Toughness; Factor 3 ⫽ Factor 3 of the MRNI-A-r, Avoidance of
Femininity; AMIRS ⫽ the Adolescent Masculinity Ideology in Relationships Scale. Scores for the AMIRS ranged from
1–7, with higher scores indicating greater endorsement of traditional masculinity ideology. PAQ-M ⫽ Personal Attributes
Questionnaire⫺Male. Scores for the PAQ-M ranged from 1–5, with higher scores indicating greater self-reported mascu-
line-valued or instrumental personality traits.
MRNI-A-R 363

ness, respectively), with boys endorsing tradi- racially diverse sample of middle school boys
tional masculinity ideology to a greater extent and girls. The hypothesized five-factor structure
than did girls (effect sizes ranged from .06 to for the MRNI-A-r was not supported, but in-
.18). No differences emerged between the boys stead our data supported a three-factor structure.
and girls on the third factor (Avoidance of Fem- The factors were Emotionally Detached Domi-
ininity), which both scored in a traditional di- nance, Toughness, and Avoidance of Feminin-
rection, suggesting that both boys and girls be- ity. Whereas the third factor drew exclusively
lieve to an equal extent that boys should avoid from one subscale, the first two factors drew
femininity. mainly from 2 or 1 subscales, respectively, but
Although this was not part of Hypothesis 4, it also included items from other subscales. The
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

is nonetheless of interest to observe that signif- fact that factors included items from several
This document is copyrighted by the American Psychological Association or one of its allied publishers.

icant differences also emerged between the boys subscales is not surprising, in that five subscales
and girls on the AMIRS, with boys endorsing resulted in three factors, and therefore items
traditional attitudes toward appropriate mascu- from the larger number of subscales were com-
line behavior in interpersonal relationships to a bined into the smaller number of factors. In
greater extent than did girls (effect size ⫽ .06). addition, the factor structure of the MRNI-A-r
No differences emerged between the boys and resembles that of the MRNS: Emotionally De-
girls on the PAQ-M, suggesting that boys and tached Dominance resembles the MRNS Status
girls self-report stereotypical masculine (instru- factor, Toughness is the same as the MRNS
mentality/agency) personality traits about Toughness factor, and Avoidance of Femininity
equally. is quite similar to the MRNS Anti-Femininity
A comparison between African American factor. Although the MRNS was developed for
and European American teens’ scores on the adults, the similarity in the two scales’ factor
study variables is displayed in Table 5. Signif- structures adds some confidence to the present
icant differences emerged between African results. Finally, the finding of only three factors
American and European American teens’ scores compared to the expected five suggests that
on the MRNI-A-r total scale and the first and teens may conceptualize gender norms in a less
third factors (Emotionally Detached Dominance differentiated way than do adults, resulting in
and Avoidance of Femininity, respectively), fewer distinctions and thus, fewer factors.
with African American teens endorsing tradi- The three factors of the MRNI-A-r and the
tional masculinity ideology to a greater extent total scale showed adequate reliability for the
than did European American teens (effect sizes most part in this sample, (␣’s ranged from .68 –
ranged from .02 to .04). No differences emerged .89). The marginal alpha of .68, for girls scores
on the second factor (Toughness), which both on factor 2, is of concern. However these results
groups scored in a traditional direction, suggest- are an improvement over the original MRNI-A,
ing that African American and European Amer- where the subscale alphas by gender ranged
ican teens believe to an equal extent that boys from .36 to .76, and five of 10 were under .70.
should be tough. No significant differences We found some evidence for the convergent
emerged on the AMIRS, suggesting that Afri- validity of the MRNI-A-r total scale and fac-
can American and European American teens tors 1 and 2 (Emotionally Detached Dominance
share similar attitudes toward appropriate mas- and Toughness, respectively) by their showing
culine behavior in interpersonal relationships, significant strong positive correlations with the
nor on the PAQ-M, suggesting that African AMIRS, with the evidence being stronger for
American and European American teens self- boys than for girls. An inspection of the items of
report stereotypical masculine (instrumentality/ the AMIRS shows that none tap the content of
agency) personality traits about equally. avoidance of femininity, perhaps explaining
why factor 3 (Avoidance of Femininity) did not
Discussion show significant correlations with the AMIRS
for either boys or girls. Further, although admit-
We set out to assess the factor structure of the tedly post hoc, the lack of correlation of factor 3
MRNI-A-r, its reliability and convergent and (Avoidance of Femininity) with the AMIRS
discriminant validity, and variations in response could be viewed as evidence of its discriminant
patterns by gender and race/ethnicity, with a validity. These results must be tempered by the
364 LEVANT ET AL.

marginal reliability for girls on the AMIRS in psychological and social indicators, such as
this study (.64), which may have resulted from self-esteem, school performance, family rela-
the removal of three items related to sex and tionships, and the like. In the long run, longitu-
homosexuality, changing the scaling of the dinal comparisons of individuals’ scores on the
measure, and the fact that we administered the adolescent and adult versions of the scale might
AMIRS to both boys and girls, even though it prove to be fruitful.
was envisioned for use with boys only by the Although participants in the present study
scale developers. were racially diverse, future research might ex-
We found some evidence for the discriminant amine additional populations to gather more
validity of the MRNI-A-r. For boys, the MRNI- evidence about the measure’s psychometric
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

A-r total scale and factors 1 and 3 showed small properties with other racial, ethnic-cultural, and
This document is copyrighted by the American Psychological Association or one of its allied publishers.

positive correlations with the PAQ-M, whereas SES groups, and also with various phases of
factor 2 showed a moderate positive correlation. adolescence.
For girls, only factors 1 and 3 showed small Finally, the Gender Role Strain posits that
positive correlations with the PAQ-M, whereas socialization processes begin at a young age and
the MRNI-A-r total scale and factor 2 showed impact the development of gender roles and
moderate positive correlations. Hence this hy- gender-typed behaviors for children and adoles-
pothesis received mixed support. Further, these cents that continue into adulthood (Levant,
results must be tempered by the marginal reli- 2001). Having an additional measure of tradi-
ability for girls (.62) on the PAQ-M in this tional masculinity ideology with good psycho-
study. This improves on the original MRNI-A, metric properties for early adolescents may help
which did not find any evidence for the discrim- us to understand in greater depth how these
inant validity for boys. socialization processes work at a time when
We think that the evidence presented in this gender concerns intensify.
study, which must still be regarded as prelimi-
nary, justifies the initial use of MRNI-A-r in
future studies. Investigators may choose to use References
the three factors or the total score, as best fits
the research questions they are investigating. Bartlett, M. S. (1954). A note on the multiplying
Some limitations of this study stem from the factors for various chi square approximations.
self-report nature of the measures, which intro- Journal of the Royal Statistical Society, 16, 296 –
duces the possibility of socially desirable re- 298.
sponding and does not allow a multimethod Brannon, R., & Juni, S. (1984). A scale for measuring
evaluation of the data (e.g., including interview- attitudes about masculinity. Psychological Docu-
ing), which is important in assessing construct ments, 14. (University Microfilms No. 2612).
Brown, J. (2002). Statistical analysis of a measure of
validity. In addition, the exploratory factor anal-
adolescent gender ideology, the Male Role Norms
ysis was not conducted separately for boys and Inventory-Adolescent: A pilot study. Unpublished
girls because there was not a large enough sam- doctoral directed study, Nova Southeastern Uni-
ple size to do so. Future research might under- versity, Fort Lauderdale, Florida.
take a measurement invariance analysis, to de- Chu, J. Y., Porche, M. V., & Tolman, D. L. (2005).
termine whether the factor structures are the The Adolescent Masculinity Ideology in Relation-
same for both genders. In addition, future re- ships Scale: Development and validation of a new
search might undertake a confirmatory factor measure for boys. Men and Masculinities, 8, 93–
analysis. 115. doi:10.1177/1097184X03257453
Future research on the discriminant validity Connell, R. W., & Messerschmidt, J. W. (2005).
of the MRNI-A-r might adopt the strategy used Hegemonic masculinity: Rethinking the concept.
Gender & Society, 19, 829 – 859. doi:10.1177/
by Chu et al. (2005), by comparing the pattern
0891243205278639
of correlations of the MRNI-A-r and the Costello, A. B., & Osborne, J. W. (2005). Best practices
PAQ-M with a range of other instruments. Fu- in exploratory factor analysis: Four recommenda-
ture research on the MRNI-A-r might also as- tions for getting the most from your analysis. Prac-
sess the convergent and discriminant validity of tical Assessment, Research & Evaluation, 10. Re-
the three factors, and examine how MRNI-A-r trieved from http://pareonline.net/getvn.asp?v⫽
total and factor scores correlate with various 10&n⫽7
MRNI-A-R 365

DiStefano, C., & Motl, R. W. (2006). Further inves- Levant, R. F., & Richmond, K. (2007). A review of
tigating method effects associated with negatively research on masculinity ideologies using the Male
worded items on self–report surveys. Structural Role Norms Inventory. Journal of Men’s Stud-
Equation Modeling, 13, 440 – 464. doi:10.1207/ ies, 15, 130 –146. doi:10.3149/jms.1502.130
s15328007sem1303_6 Majors, R., & Billson, J. M. (1992). Cool pose: The
Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., dilemmas of Black manhood in America. New
& Strahan, E. J. (1999). Evaluating the use of York, NY: Lexington Books.
exploratory factor analysis in psychological re- Messner, M. A. (1992). Power at play: Sports and
search. Psychological Methods, 4, 272–299. doi: the problem of masculinity. Boston, MA: Beacon
10.1037/1082-989X.4.3.272 Press.
Galambos, N. L., Almeida, D. M., & Peterson, A. C. O’Connor, B. P. (2000). SPSS and SAS programs for
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

(1990). Masculinity, femininity, and sex role atti- determining the number of components using par-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

tudes in early adolescence: Exploring gender in- allel analysis and Velicer’s MAP test. Behavior
tensification. Child Development, 61, 1905–1914. Research Methods, Instruments, & Computers, 32,
doi:10.2307/1130846 396 – 402. doi:10.3758/BF03200807
Glorfeld, L. W. (1995). An improvement on Horn’s Pleck, J. H. (1981). The myth of masculinity. Cam-
parallel analysis methodology for selecting the bridge, MA: MIT Press.
correct number of factors to retain. Educational Pleck, J. H. (1995). The gender role strain paradigm:
and Psychological Measurement, 55, 377–393. An update. In R. F. Levant & W. S. Pollack (Eds.),
doi:10.1177/0013164495055003002 A new psychology of men (pp. 11–32). New York,
Hayton, J. C., Allen, D. G., & Scarpello, V. (2004). NY: Basic Books.
Factor retention decisions in exploratory factor Pleck, J. H., Sonenstein, F. L., & Ku, L. C. (1993a).
analysis: A tutorial on parallel analysis. Organiza- Masculinity ideology and its correlates. In S. Os-
tional Research Methods, 7, 191–205. doi: kamp & M. Costanzo (Eds.), Gender issues in
10.1177/1094428104263675 contemporary society (pp. 85–110). Newbury
Kahn, J. H. (2006). Factor analysis in counseling
Park, CA: Sage.
psychology research, training, and practice: Prin-
Pleck, J. H., Sonenstein, F. L., & Ku, L. C. (1993b).
ciples, advances, and applications. The Counseling
Masculinity ideology: Its impact on adolescent
Psychologist, 34, 684 –718. doi:10.1177/
males’ heterosexual relationships. Journal of So-
0011000006286347
cial Issues, 49, 11–29. doi:10.1111/j.1540-
Kaiser, H. H. (1974). An index of factorial simplicity.
4560.1993.tb01166.x
Psychometrica, 39, 31–36. doi:10.1007/
BF02291575 Pleck, J. H., Sonenstein, F. L., & Ku, L. C. (1994a).
Levant, R. F. (1996a). The new psychology of men. Attitudes toward male roles: A discriminant valid-
Professional Psychology: Research and Prac- ity analysis. Sex Roles, 30, 481–501. doi:10.1007/
tice, 27, 259 –265. doi:10.1037/0735-7028.27.3 BF01420798
.259 Pleck, J. H., Sonenstein, F. L., & Ku, L. C. (1994b).
Levant, R. F. (1996b). The crisis of connection be- Problem behaviors and masculinity ideology in
tween men and women. Journal of Men’s Stud- adolescent males. In R. D. Ketterlinus & M. E.
ies, 5, 1–12. Lamb (Eds.), Adolescent problem behaviors: Is-
Levant, R. F. (1997). The masculinity crisis. Journal sues and research (pp. 165–186). Hillsdale, NJ:
of Men’s Studies, 5, 221–231. Lawrence Erlbaum.
Levant, R. F. (2001). Desperately seeking language: Pollack, W. S. (1998). Real boys. New York, NY:
Understanding, assessing and treating normative Random House.
male alexithymia. In G. R. Brooks and G. Good Preacher, K. J., & MacCallum, R. C. (2003). Repair-
(Eds), The new handbook of counseling and psy- ing Tom Swift’s electric factor analysis machine.
chotherapy for men (Vol. 1, pp. 424 – 443). San Understanding Statistics, 2, 13– 43. doi:10.1207/
Francisco, CA: Jossey-Bass. S15328031US0201_02
Levant, R. F., Graef, S. T., Smalley, K. B., Williams, Robitschek, C. (2003). Validity of Personal Growth
C., & McMillan, N. (2008). The evaluation of the Initiative Scale scores with a Mexican American
Male Role Norms Inventory-Adolescent (MRNI- college student population. Journal of Counseling
A). Thymos: Journal of Boyhood Studies, 2, 46 – Psychology, 50, 496 –502. doi:10.1037/0022-
59. doi:10.3149/thy.0201.46 0167.50.4.496
Levant, R. F., Hirsch, L., Celentano, E., Cozza, T., Spence, J. T., & Helmreich, R. L. (1978). Masculine
Hill, S., MacEachern, M., . . . Schnedeker, J. instrumentality and feminine expressiveness: Their
(1992). The male role: An investigation of con- relationships with sex role attitudes and behaviors.
temporary norms. Journal of Mental Health Coun- Psychology of Women Quarterly, 5, 147–163. doi:
seling, 14, 325–337. 10.1111/j.1471-6402.1980.tb00951.x
366 LEVANT ET AL.

Tabachnick, B. G., & Fidell, L. S. (2007). Using Tough. (2010). In Free Merriam-Webster Dictionary.
multivariate statistics (5th ed.). Boston, MA: Pear- Retrieved from http://mw1.merriam-webster.com/
son Education. dictionary/tough
Thompson, E. H., & Pleck, J. H. (1986). The structure of Whorley, M., & Addis, M. E. (2006). Ten years of
male norms. American Behavioral Scientist, 29, 531– research on the psychology of men and masculin-
543. doi:10.1177/000276486029005003 ity in the United States: Methodological trends and
Thompson, E. H., & Pleck, J. H. (1995). Masculinity critique. Sex Roles, 55, 649 – 658. doi:10.1007/
ideology: A review of research instrumentation on s11199-006-9120-1
men and masculinities. In R. F. Levant & W. S.
Pollack (Eds.), A new masculinity of men (pp.
129 –163). New York, NY: Basic Books.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Thompson, E. H., Pleck, J. H., & Ferrera, D. (1992).


This document is copyrighted by the American Psychological Association or one of its allied publishers.

Men and masculinities: Scales for masculinity ide- Received March 30, 2010
ology and other masculinity-related constructs. Sex Revision received January 30, 2012
Roles, 27, 573– 608. doi:10.1007/BF02651094 Accepted April 23, 2012 䡲

You might also like