Healthy Women, Healthy Men, and Healthy Adults: An Evaluation of Gender Role Stereotypes in the Twenty-first Century

Abstract

An important question often asked when counselors-in-training read textbook discussion of gender role stereotypes, especially of older work such as the classic study by the Brovermans and their colleagues, is “Haven’t these biases been eliminated or at least reduced?” The current study was designed to replicate the work of the Brovermans and their colleagues to answer that specific question and to determine how current counselors-in-training perceive healthy adult women, healthy adult men, and healthy adults. As in the prior research, initial ratings of the social desirability of traditional gender role stereotypes were conducted, and the findings showed many similarities to past research. That investigation was followed by a modified Stereotype Questionnaire, based on the original work of Rosenkrantz, Vogel, Bee, I. Broverman, and D. M. Broverman (1968). Healthy adult women were found to be significantly different from healthy adult men as well as from healthy adults. In addition, the results suggest that there have been changes in counselors’ perceptions of healthy adults. Counselors-in-training were found to hold two standards for mental health—one for women and another for men.

Full text

ORIGINAL ARTICLE
Healthy Women, Healthy Men, and Healthy Adults:
An Evaluation of Gender Role Stereotypes
in the Twenty-first Century
Susan Rachael Seem &M. Diane Clark
Published online: 8 November 2006
#Springer Science + Business Media, Inc. 2006
Abstract An important question often asked when coun-
selors-in-training read textbook discussion of gender role
stereotypes, especially of older work such as the classic
study by the Brovermans and their colleagues, is “Haven’t
these biases been eliminated or at least reduced?”The
current study was designed to replicate the work of the
Brovermans and their colleagues to answer that specific
question and to determine how current counselors-in-
training perceive healthy adult women, healthy adult men,
and healthy adults. As in the prior research, initial ratings of
the social desirability of traditional gender role stereotypes
were conducted, and the findings showed many similarities
to past research. That investigation was followed by a
modified Stereotype Questionnaire, based on the original
work of Rosenkrantz, Vogel, Bee, I. Broverman, and D. M.
Broverman (1968). Healthy adult women were found to be
significantly different from healthy adult men as well as
from healthy adults. In addition, the results suggest that
there have been changes in counselors’perceptions of
healthy adults. Counselors-in-training were found to hold
two standards for mental health—one for women and
another for men.
Keywords Gender roles .Stereotypes .
Counselors-in-training
Relationships among self-concept, psychological health, and
gender stereotypes emerged in the work of Rosenkrantz,
Vogel, Bee, I. Broverman, & D. M. Broverman, (1968). In
order to examine these relationships, Rosenkrantz et al.
investigated first the social desirability of traits used to
describe men and women and then developed the Stereo-
type Questionnaire. The initial listing of socially desirable
items was completed by having two classes of undergrad-
uate students write down the characteristics that they
believed differentiated men and women. Any item listed
more than once was included in the Stereotype Questionnaire,
which consisted of 122 items that were arranged along a Likert
scale in a bipolar format (e.g., “not at all aggressive”–“very
aggressive”). An interesting component of this 7-point Likert
scale is that there were ten potential points between each of the
Likert anchor-numbers for a total of 60 possible points.
Rosenkrantz et al. then tested the Stereotype Questionnaire
with a second independent sample of undergraduates to
determine which pole of each of the 122 bipolar items
represented the socially desirable one.
The second study of Rosenkrantz et al. (1968) focused
on determining the gender role stereotypes for men and for
women and comparing those stereotypes to the ideal of a
healthy adult. Their procedure asked all participants to
“imagine that you are going to meet a person for the first time
and the only thing you know in advance is that this person is
an adult (man)”(p. 288)/adult woman. A within group design
was used and participants were asked to rate each of the items
to the extent that it characterized: (a) a healthy adult man, (b) a
healthy adult woman, and (c) themselves (given the assump-
tion that the later rating would represent a healthy adult, sex
unspecified). Presentation of the healthy adult man and
healthy adult woman instructions was counterbalanced and
was always followed by the rating of a healthy adult. The
ratings were used to determine which traits represented gender
Sex Roles (2006) 55:247–258
DOI 10.1007/s11199-006-9077-0
S. R. Seem (*)
Department of Counselor Education,
State University of New York College at Brockport,
350 New Campus Drive, Brockport,
NY 14420, USA
e-mail: sseem@brockport.edu
M. D. Clark
Gallaudet University, Washington, DC, USA

role stereotypes by comparing the characteristics for a healthy
adult man and for a healthy adult women to the characteristics
for a healthy adult. Results indicated that college students,
regardless of sex, agreed on the gender role stereotypes of men
and women. They additionally concluded that the self-concepts
of male and female participants followed these traditional
gender role stereotypes meaning that women “presumably,...
also hold negative values of their worth relative to men”
(p. 293). Finally, they found a larger number of socially
desirable characteristics and behaviors were stereotypically
connected with masculinity than with femininity.
In an extension of the study of Rosenkrantz et al. (1968),
Broverman and colleagues, (I. K. Broverman, D. M.
Broverman, Clarkson, Rosenkrantz, & Vogel, 1970;I.
K. Broverman, Vogel, D. M. Broverman, Clarkson, &
Rosenkrantz, 1972) investigated female and male clinicians’
(i.e., psychologists, psychiatrists, and social workers) judg-
ments of the mental health of women and men. As in the
earlier work of Rosenkrantz et al., the Stereotype Question-
naire was used, but this time with a between groups design. In
their description of the instrument, Broverman et al. (1970)
reported use of the questionnaire of Rosenkrantz et al. with
122 bipolar items. Their example in the 1970 article, however,
did not include a Likert scale; instead it was posed as a forced
choice format (“not at all aggressive”–“aggressive”).
Broverman et al. (1970) hypothesized that (a) clinical
judgments of the characteristics of a healthy, mature
individual would differ as a function of the sex of the
person judged and that (b) behavioral characteristics that
were regarded as healthy for an adult, sex unspecified,
would be more often regarded as healthy for men than for
women (i.e., following cultural stereotypes of gender
differences). Two types of scores were calculated: health
scores and agreement scores. The health scores were based
upon the assumption that traits selected for healthy adults
would reflect the definition of mental health for all
individuals. These health scores were determined by the
pole selected by 75% or more of the participants who were
given the healthy adult instructions for each of the 122
items. The masculinity agreement score was based on the
percentage of participants who selected the socially
desirable pole with the healthy adult man instructions and
the femininity agreement score was based on the percentage
who selected the socially desirable pole with the healthy
adult woman instructions.
The results of Broverman et al. (1970) paralleled the
findings of Rosenkrantz et al. (1968) and showed that
clinicians’judgments of adult men’s mental health did not
differ significantly from their judgments of healthy adults,
sex unspecified, whereas their judgments of adult women’s
mental health did differ significantly from judgments of
healthy adult men and healthy adults sex unspecified. This
determination was based upon the large overlap among
health scores and masculinity agreement scores in contrast
to the limited overlap among health scores and femininity
agreement scores. These differences corresponded with
cultural stereotypes of men and women that were widely
held during that time. Healthy adult women were described
as different from both healthy adult men and healthy adults
in that they were more “submissive,”“less independent,”“less
aggressive,”“less competitive,”“more easily influenced,”
“more emotional,”and “less objective.”The authors conclud-
ed that a double standard of mental health existed for women,
in that for a woman to be seen as mentally healthy she must be
feminine and not adult-like (i.e., not like a man). Both female
and male clinicians implicitly supported this standard.
Furthermore, the authors discovered that clinicians when
given instructions to “describe healthy, socially competent
women”(p. 2) used fewer traits than when they were given
instructions to”describe healthy, socially competent men”
(p. 2). The traits found for healthy women were often viewed
as less socially desirable than the traits listed for healthy men.
In an attempt to evaluate any changes in mental health
professionals’gender role stereotypes and clinical judg-
ments, Phillips and Gilroy (1985) used a shortened version
of the Likert scale Stereotype Questionnaire created by
Rosenkrantz et al. (1968) in a between groups design. As in
the Broverman et al. (1970) study, the participants were
psychiatrists, social workers, and psychologists. Similar to
Broverman et al., they found no significant differences in
ratings of health for adult men, adult women, and adults,
sex unspecified, that were related to the clinician’s sex. In
contrast to the results of Broverman et al., Philips and
Gilroy found high levels of overlap when they compared
the health scores to both the masculinity agreement scores
and the femininity agreement scores. Philips and Gilroy
challenged the conclusion of Broverman et al. regarding a
double standard of mental health, and argued that the
conclusion of Broverman et al. was based on a statistical
artifact related to their use of a forced choice format.
Furthermore, Philips and Gilroy found no significant
relationship of the social desirability of traits and traditional
gender role stereotypes, and they postulated that most traits
were viewed as socially desirable for both sexes. Phillips
and Gilroy concluded that either the Broverman et al.
results were due to this statistical artifact or that their own
results were due to the fact that progress had been made by
clinicians in the reduction of their gender role stereotyping
in clinical judgments.
In another replication of Broverman et al. (1970),
Widiger and Settle (1987) addressed methodological con-
cerns in the classic design by focusing on the imbalanced
ratio of traditionally masculine valued to traditionally
feminine valued items as well as the use of bipolar
adjectives. Widiger and Settle argued that feminine traits
are not the opposite of masculine traits; they found that if
248 Sex Roles (2006) 55:247–258

healthy adult men were rated high on an item, it did not
necessarily mean that healthy adult women were not rated
relatively high on that same item. Widiger and Settle,
therefore, focused on developing unipolar items—i.e., each
item was paired with its negative, not its opposite (e.g.,
“very independent”versus “not at all independent”rather
than “independent”versus “dependent”). Furthermore, 72
additional items were created and rated for social desirabil-
ity by college students in order to correct for the imbalance
of traditionally masculine valued items in Broverman et al.,
who had a ratio of 71% traditionally masculine valued
versus 39% traditionally feminine valued items.
Given these changes, Widiger and Settle (1987) con-
tended that the conclusions of the Broverman et al. (1970)
study were a result of a statistical artifact that was due to
the ratio of traditionally masculine valued to traditionally
feminine valued items on the dependent measure. They
argued that because most of the socially desirable poles in
the Broverman et al. study “concerned a stereotypically
masculine characteristic, it is not surprising that the mean
masculinity health score was close to the adult health score
and greater than the femininity health score”(p. 464). The
results, they reasoned, depended upon how the analysis was
done. By altering the ratio of traditionally masculine valued
to traditionally feminine valued items in three analyses (i.e.,
first analysis: 27 traditionally feminine valued items versus
11 traditionally masculine valued items, second analysis: 27
traditionally masculine valued items versus 11 female-valued
items, and third analysis: an equal number of female- and
traditionally masculine valued items), the authors found
gender bias against women, gender bias against men, and no
gender bias, respectively, based solely on statistical measures.
In a recent partial replication of the classic study by
Rosenkrantz et al. (1968), Nesbitt and Penn (2000)surveyed
community college students using Stereotype Questionnaire
of Rosenkrantz et al. in order to investigate whether gender
role stereotypes had changed in the 30 plus years since the
original study. Although gendered traits such as expressing
and experiencing affect still differentiated healthy adult
women and healthy adult men, gender role stereotypes had
changed in that traits such as “logical,”“ambitious,”“direct,”
and “the ability to separate ideas from feelings”no longer
differentiated women from men.
In conclusion, some of the prior research on the impact
of gender role stereotyping on clinical judgement showed
bias against women, whereas the results of other studies
suggested that the dependent measure had actually elicited
the bias. Thus, the results are inconclusive. Despite
criticism of the methodology of the Broverman et al. (1970)
investigation (Phillips & Gilroy, 1985; Widiger & Settle,
1987) and the fact that it was conducted over 30 years ago,
that study is still cited in college textbooks (see, for example,
Crawford & Unger, 2000;Lerner,2002). Students consis-
tently ask whether the conclusions of Broverman et al. are
still true. Given this fact, it seems important to reexamine
gender role stereotyping by future counselors in the twenty-
first century.
Objectives
In the present research study, as in the prior research
discussed above, two studies were conducted: one to assess
social desirability, and the other to investigate gender role
stereotypes with the Stereotype Questionnaire. The objec-
tive of Study 1 was to evaluate the current social desirability
of the items to be used in Study 2 on the Stereotype
Questionnaire. Therefore in Study 1, participants were asked
to pick the socially desirable pole of each item. Study 2 had
two objectives. The first objective was to evaluate current
gender role stereotypes held by counselors-in-training re-
garding healthy adult women, healthy adult men, and healthy
adults, sex unspecified, using the Stereotype Questionnaire.
The second objective was to determine whether gender role
stereotypes for men and women have changed since the
studies of Broverman et al. (1970,1972).
Study 1
Method
Participants The participants in the social desirability study
were 89 undergraduate, introductory psychology students
enrolled in a comprehensive university in the northeastern
United States. They were 47 women and 42 men with a
mean age of 21.8 years (range—18 to 29; mode = 18).
Ninety-eight percent of the participants were European
Americans. The other 2% included one African American
and one Asian American. These percentages represent the
demographics of the student population at this university.
Participants were given extra credit for their participation.
Measure A Social Desirability Questionnaire was created
that consisted of the 62 unipolar items later used in the
Stereotype Questionnaire. This questionnaire was more
closely aligned to the Widiger and Settle (1987) question-
naire than to the Broverman et al. (1970,1972) question-
naire, which was in a bipolar format. Several new items
were developed for the present study to correct for the 14
bipolar items that remained in Widiger and Settle’s version.
Therefore, for the current study, each of these 14 items was
split into two unipolar items for a total of 28 new items. For
example, “strong”versus “weak”became “very strong”
versus “not at all strong”and “very weak”versus “not at all
weak.”Pilot testing eliminated seven of the new items
Sex Roles (2006) 55:247–258 249

because participants did not indicate a socially desirable
characteristic for those items. The Stereotype Questionnaire
for this study consisted of 62 unipolar pairs (see Table 1for
a complete list of all unipolar pairs). The items were
presented as pairs in two columns, and participants were
directed to circle one item from each pair that represented
the trait that was more socially desirable. Items were
counterbalanced so that one-half of the historically, socially
desirable items were presented in the left-hand column,
whereas the other one-half were presented in the right-hand
column.
Analytic plan The social desirability of each of the 62
unipolar items was determined by counting the number of
times each description (i.e., the left hand side—“very
sympathetic”versus the right hand side—“not at all sympa-
thetic”) was selected. This social desirability determination
was summed across participants’responses for each item and
recorded as a percentage. A zscore was used as the criterion
to determine if participants noted a socially desirable pole
between the two descriptions for each item (i.e., if one
description was selected significantly more often than the
other one). In order to declare one description of an item as
socially desirable, a zscore above 1.65 was required. This z
value was established by the zformula (Nobserved−N
expected/the square root of Nexpected) to obtain the
required percent. A percentage of 60.03% was determined
as follows Nobserved ¼89 2 þ1:65 square root of 89½¼
=
ð
60:03Þand rounded up to 61% such that p<0.05 (two-tailed).
Results and Discussion
Five of the 62 pairs did not have social desirability scores
above 61% for either description of the item. These items
included: “not at all uncomfortable when people express
emotions,”“not at all dependent,”“acts on logic rather than
feelings,”“very strong need for security,”and “not at all
passive.”(See Table 1). As can be seen in Table 1, the
remaining 57 pairs had social desirability scores above
61%.
Participants viewed in traditional ways all but one of the
items from past research. “Very subjective”versus “not at all
subjective”differed from past research in that historically, the
socially desirable masculine pole of “not at all subjective”
was replaced by the new socially desirable feminine pole
“of very subjective.”All of the remaining traits continued to
be viewed as they have been historically.
The items that did not obtain 61% social desirability
percentages include two that were historically feminine
items (“not at all uncomfortable when people express
emotions”and “very strong need for security”) and two that
were historically masculine items (“not at all dependent”and
“acts on logic rather than feelings”). The remaining item that
did not obtain a 61% score was a new item (“very passive”)
that in prior research had been paired with “active.”It appears
that “passive”is not seen as a socially desirable trait.
In general, the results show that all participants had a
clear understanding of cultural values regarding what is
considered socially desirable by our society (most items
received socially desirable scores above 80%). These items
are highly similar to those found by Rosenkrantz et al.
(1968) and Broverman et al. (1970,1972).
Study 2
Method
Participants The participants in Study 2 were 121 students
enrolled in two masters’level counseling programs in the
northeastern United States. The mean age was 28.4 years
(range—21 to 48; mode=23). There were 65 (54%) women
and 56 (46%) men; 112 were European Americans, six were
African Americans, and three were Asian Americans. This
represents the demographics of the student bodies in both
universities. Students were enrolled in four degree programs:
masters of science in education (61%), masters of science
(25%), masters of arts (10%), and certificate of advanced
study (4%). These students were recruited from classes
taught by the authors as well as by other members of their
departments. Independent graduate students (who were not
in the authors’classes) gave the surveys to the students who
chose to participate; all surveys were returned to the
respective university department’soffice.
Measure Items for the Stereotype Questionnaire were devel-
oped in Study 1.SeeTable1for a complete list of the items.
The items were arranged on a Likert scale in a unipolar
format (e.g., “not at all aggressive”–“very aggressive”), and
anchored on a 1 to 7 scale as in Rosenkrantz et al. (1968).
Procedure Each participant received a package that includ-
ed the three Stereotype Questionnaires, as in the Rosenkrantz
et al. (1968) within group design. Directions specified that
participants were to complete each survey in the order
presented. In addition, participants were informed that,
although each survey was the same, it would be preceded
by a different set of instructions. They were warned to read
each new set of instructions carefully before completing the
questionnaire. All participants were given the following
instructions: “Imagine that you are going to meet a person
for the first time and the only thing you know in advance is
that the person is a woman (man/adult). On the following
questionnaire, think of a normal man (woman/adult), and then
indicate on each item the pole to which a mature, healthy,
250 Sex Roles (2006) 55:247–258

Table 1 Results of the social desirability questionnaire with socially desirable items presented in the left column.
Social desirability scores
Item Percent
agreement
Item Percent
agreement
Unipolar items
Traditionally masculine items
Controls emotion* 90 Does not control emotions 10
Decisive* 84 Indecisive 13
Very worldly* 85 Not at all worldly 13
Very firm* 87 Not at all firm 13
Controls self under stress* 92 Becomes upset under stress 8
Very direct* 90 Not at all direct 10
Very active* 96 Not at all active 4
Very adventurous* 94 Not at all adventurous 6
Not at all sneaky*♦71 Very sneaky 29
Heroic* 88 Not heroic 11
Knows the way of the world* 96 Does not know the way of the world 4
Very Strong* 95 Not at all strong 4
Easily able to separate feelings from ideas* 83 Unable to separate feelings from ideas 16
Very independent* 98 Not at all independent 2
Relies on self* 85 Relies on others 11
Enjoys a challenge very much* 93 Does not enjoy a challenge very much 7
Not at all weak*♦94 Very weak 6
Very industrious* 93 Not at all industrious 7
Can make decisions easily* 92 Has difficulty making decisions 8
Very daring* 95 Not at all daring 4
Not at all afraid to take risks* 93 Very afraid to take risks 7
Very brave* 92 Not at all brave 8
Not at all excitable in a minor crisis* 72 Very excitable in a minor crisis 28
Has a strong will* 98 Does not have a strong will 2
Very authoritative* 67 Not at all authoritative 32
Very competitive* 84 Not at all competitive 16
Very objective*♦73 Not at all objective 26
Unipolar pairs
Traditionally feminine items
Very concerned about others* 94 Not at all concerned about others 6
Very neat in habits* 92 Not at all neat in habits 8
Enjoys art and literature very much* 81 Does not enjoy art and literature very much 18
Very sensitive* 87 Not at all sensitive 13
Very talkative* 87 Not at all talkative 13
Very gentle* 92 Not at all gentle 8
Very careful* 96 Very careless 4
Very affectionate* 87 Not at all affectionate 13
Not at all harsh*♦83 Very harsh 17
Very idealistic* 78 Not at all idealistic 21
Very sympathetic* 90 Not at all sympathetic 10
Very sentimental* 90 Not at all sentimental 10
Very charitable* 92 Not at all charitable 8
Very understanding of others* 98 Not at all understanding of others 2
Very creative* 94 Not at all creative 6
Very emotional* 65 Not at all emotional 34
Very warm in relations with others* 95 Not at all warm in relations with others 5
Very aware of feelings of others* 95 Not at all aware of feelings of others 5
Very considerate of others* 96 Very inconsiderate of others 4
Very interested in own appearance* 83 Not at all interested in own appearance 16
Very subjective* 72 Not at all subjective 27
Very compassionate* 95 Not at all compassionate 5
Not at all cold*♦96 Very cold 2
Sex Roles (2006) 55:247–258 251

socially competent woman (man/adult) would be closer.
Indicate your degree of conviction by placing a slash through
the point on the scale at which you feel this man (woman/
adult) would best be described by the trait under consider-
ation.”The order of presentation of the healthy adult woman
and a healthy adult man Stereotype Questionnaires was
counterbalanced, and the healthy adult always followed these.
Analytic plan The analytical plan included three types of
analyses. First, we calculated item agreement scores. Next,
binary, stereotypic items were determined. Then, continu-
ous, total, gender stereotypic scores were calculated and
used in a repeated measure MANOVA with paired t-test
post-hocs. Each of these coding schemes are described
below.
The 7-point Likert scale used in the Stereotype Ques-
tionnaire to present the unipolar items was analyzed as a
binary code to create item agreement scores. Because no
neutral (i.e., a score of 4) scores were obtained, scores
below the neutral rating of 4 (1 through 3) were recoded as
1. Scores above the neutral rating of 4 (5 through 7) were
recoded as 2. Individual item agreement scores then were
calculated by determining the percentage of participants
who selected the socially desirable item within each of the
62 pairs. These poles were labeled either traditionally
feminine or traditionally masculine, based on prior re-
search. Therefore, six sets of item agreement scores were
calculated for each socially desirable item, one for female
participants and another for male participants, on each of
the three Stereotype Questionnaires—healthy adult woman,
healthy adult man, and healthy adult. Specifically, each of
these six sets had 62 item agreement scores for a total of
372 item agreement scores across the six sets.
Binary stereotypic items were determined by combining
men’s and women’s item agreement scores on each of the
62 pairs across the three Stereotype Questionnaires. Then,
groups of stereotypic items were developed by organizing
the item agreement scores discussed above into lists, one
for men, one for women, and one for adults. Items were
selected for inclusion if 70% or more of the participants had
selected the socially desirable pole for that item. Seventy
percent was used rather than the 75% in earlier work as
examination of the current data showed this percentage to
be a natural break point and this cut-score was more
stringent than the 61% used in Study 1. (See Study 1 for a
discussion related to the determination of percentage cut-
offs using zscores.)
Then, the continuous data originally collected for each
item was selected for all stereotypic items. The participants’
responses on these items were then averaged together to
create six total gender stereotypic scores, two each for a
healthy adult woman (i.e., one each for men and women),
two each for a healthy adult man, and two each for a healthy
adult. Scores like these total gender stereotypic scores were
calculated by both Broverman et al. (1970,1972) and
Widiger and Settle (1987).
The six total gender stereotypic scores of the healthy adult
woman, healthy adult man, and healthy adult, described
above, were analyzed in a repeated measure MANOVA. Sex
was used as a between participant factor. Planned comparisons
Table 1 Continued
Social desirability scores
Item Percent
agreement
Item Percent
agreement
Very soothing* 94 Not at all soothing 5
Very home oriented*♦77 Not at all home oriented 23
Not at all rough*♦71 Very rough 28
Easily expresses tender feelings* 88 Does not express tender feelings at all 12
Very soft*♦72 Not at all soft 28
Very loving* 98 Not at all loving 2
Very forgiving* 95 Not at all forgiving 5
Unipolar pairs
Non-significant agreement scores
Not at all uncomfortable when people express
emotion (F)
55 Very uncomfortable when people express emotions 45
Not at all dependent (M) 52 Very dependent 48
Acts on logic rather than feeling (M) 60 Acts on feeling rather than logic 39
Very strong need for security (F) 49 Very little need for security 51
Very passive (F) 49 Not at all passive 51
*Socially desirable pole; ♦new item based on the non-socially desirable pole in prior research
FHistorically/traditionally feminine item, Mhistorically/traditionally masculine item
252 Sex Roles (2006) 55:247–258

with paired t-tests were performed to compare a healthy adult
woman to a healthy adult, a healthy adult man to a healthy
adult, and a healthy adult woman to a healthy adult man.
Results
Item agreement scores showed that there were 35 tradition-
ally feminine items and 27 traditionally masculine items
across the Stereotype Questionnaires. Item agreement
scores for a healthy adult woman were generally above
50% on both the traditionally feminine and masculine items
with a few notable exceptions. Within this view of a healthy
adult woman, female participants’item agreement scores
were less than 50% on only two traditionally feminine
items—“subjective”and “very strong need for security.”
Here, female participants did not view these items as
descriptive of a healthy adult woman. Male participants
tended to report higher item agreement scores on these
traditionally feminine items than did their female counter-
parts. One additional traditionally feminine item agreement
score for men was below 50% (“not at all weak”), whereas
the comparable female participants’item agreement score
was higher (60%). In terms of traditionally masculine items,
female participants had 12 item agreement scores less than
50% for a healthy adult woman; whereas male participants
had14items(seeTable2). Therefore, participants,
regardless of their sex, endorsed many traditionally mascu-
line items at relatively high levels (above 50%) as character-
istic of a healthy adult woman (15 by female and 13 by male
participants; see Table 2—healthy adult women—masculine
pole socially desirable).
Examination of the data for a healthy adult man revealed
that participants reported higher overall scores on the
traditionally masculine items than on the traditionally
feminine items (see Table 2). On the traditionally masculine
items, only four items had agreement scores below 50%,
and only female participants reported three of those scores.
In contrast to these relatively high traditionally masculine
item agreement scores, only five traditionally feminine
items had agreement scores above 50% for a healthy adult
man. These items included “idealistic,”“not at all weak,”
“interested in own appearance,”“loving,”and “acts on
feelings rather than logic.”Participants’sex differences were
also reflected in higher item agreement scores by female
participants than by male participants on traditionally
feminine items (i.e., only women reported item agreement
scores above 70% when they evaluated a healthy adult man
on traditionally feminine items).
The healthy adult item agreement scores tended to be
more equivalently endorsed by female and male participants
for both traditionally feminine and traditionally masculine
items. Items were more likely to have agreement scores
below 40% if they were traditionally feminine items. Seven
items were reported below 40% by female participants and
seven items by male participants; see Table 2. On the other
hand, items on the healthy adult questionnaire were more
likely to have item agreement scores above 60% if they were
traditionally masculine items. Here 11 items were reported
by female participants and 15 by male participants. Male
participants had high item agreement scores on these
traditionally masculine items. Their female counterparts also
had high item agreement scores on these items, except for the
items “competitive”(52%), “adventurous”(52%), and
“firm”(52%).
Stereotypic items The perception of a healthy adult woman
included 21 stereotypic items (see Table 3for a complete
list). All but three of these items were from the traditionally
feminine item pool. The three items not from the traditionally
feminine item pool included “strong,”“independent,”and
“enjoys a challenge.”Therefore, the healthy adult woman
description included 18 traditionally feminine items as well
as three traditionally masculine items.
The healthy adult man description included 15 stereo-
typic items (see Table 3for a complete list). All of these
items were from the traditionally masculine pool. Here the
healthy adult man had fewer stereotypic items than did the
healthy adult woman.
The healthy adult description included five stereotypic
items that reached the 70% agreement criteria (see Table 3
for a complete list). One of these items was from the
traditionally feminine pool (“understanding of others”), and
the remaining four items were from the traditionally mascu-
line pool. The healthy adult had many fewer endorsed items
than did either the healthy adult woman or the healthy adult
man.
MANOVA and paired t-tests A repeated measures MAN
OVA was performed to compare the total gender stereotypic
scores. Participant sex was the between subject factor. Both
female participants (SD=0.26) and male participants (SD =
0.50) had means of 3.97 on the total gender stereotypic
scores for the healthy adult women. For the healthy adult
man the total gender stereotypic scores showed sex differ-
ences; female participants had a mean score of 4.21 (SD =
0.25) and male participants had a mean score of 4.34 (SD =
0.49). Total gender stereotypic scores for the healthy adult
also were the same (mean score of 4.25) for female
participants (SD=0.34) and male participants (SD = 0.37).
There was a significant effect for total gender stereotypic
scores, F(1, 86)=30.81, p0.0001, but the interaction
between these scores and participant sex was not signifi-
cant, F(1, 86)=0.03, p0.87. To evaluate the main effect of
the total gender stereotypic scores, paired t-tests were
performed. These analyses showed significant differences
Sex Roles (2006) 55:247–258 253

Table 2 Individual percent item agreement scores for the three Stereotype Questionnaires presented separately by participant sex.
Healthy adult woman Healthy adult man Healthy adult
Participant sex Female (%) Male (%) Female (%) Male (%) Female (%) Male (%)
Feminine pole—traditionally socially desirable
Idealistic 64 62 52 59 59 55
Sympathetic 74 79 31 26 59 55
Sentimental 71 75 24 32 44 50
Charitable 74 74 46 44 63 53
Understanding of others 83 81 41 42 71 71
Creative 69 73 43 44 57 62
Not at all weak 60 47 75 66 59 55
Emotional 64 69 13 17 29 30
Warm in relations with others 84 87 40 42 60 60
Considerate of others 81 87 48 51 72 57
Interested in own appearance 76 94 57 59 64 78
Subjective 44 53 35 42 35 31
Concerned about others 86 77 41 51 70 64
Neat in habits 67 84 40 30 47 54
Enjoys art and literature 65 80 34 33 45 51
Sensitive 81 90 26 35 54 63
Talkative 68 70 32 35 48 64
Gentle 75 82 39 32 48 47
Careful 66 72 40 46 50 69
Affectionate 80 93 45 40 52 52
Not at all sneaky 48 49 34 36 41 46
Compassionate 87 87 43 43 69 61
Soothing 81 82 44 40 59 50
Home oriented 52 54 34 25 53 48
Easily expresses tender feelings 56 66 32 27 44 36
Soft 50 64 23 14 32 29
Loving 89 85 55 61 59 61
Forgiving 69 74 43 47 47 58
Acts on feelings rather than logic 49 66 72 62 17 13
Not at all uncomfortable when people express emotions 67 58 55 54 44 35
Aware of the feelings of others 86 80 31 30 60 57
Not at all cold 77 72 45 38 61 44
Not at all harsh 60 54 23 15 30 25
Not at all rough 52 54 21 7 37 26
Very strong need for security 44 54 28 38 39 43
Masculine pole—traditionally socially desirable
Knows the way of the world 59 76 73 72 70 76
Strong 79 67 72 85 75 76
Easily able to separate feelings from ideas 60 49 39 42 55 52
Independent 72 76 84 85 75 73
Relies on self 49 53 69 74 57 61
Enjoys a challenge 76 75 81 83 73 75
Industrious 60 53 79 76 65 63
Can make decisions easily 48 44 75 63 66 66
Daring 32 41 75 79 54 63
Controls emotions 29 26 65 78 38 44
Decisive 52 47 74 74 70 61
Worldly 52 44 62 63 66 65
Firm 43 53 73 76 52 62
Controls self under stress 44 36 66 64 58 58
Direct 54 58 77 78 65 64
Active 59 68 68 77 54 36
Adventurous 51 56 72 87 52 64
Heroic 54 40 62 62 62 51
254 Sex Roles (2006) 55:247–258

between a healthy adult woman and a healthy adult, t=
−5.72, df=92, p0.0001, and between a healthy adult
woman and a healthy adult man, t=−5.22, df = 89, p0.0001.
There was no significant difference between a healthy adult
man and a healthy adult, t=0.51, df= 93, p0.61.
Discussion
Several interesting results emerged when perceptions of a
healthy adult woman, a healthy adult man, and a healthy adult
were compared. First, individuals endorsed more traits for a
healthy adult woman than was true in the past. The increase in
the number of traits used to describe a healthy adult woman
suggested that current gender role stereotypes have changed,
but maybe not for the better. Now women are expected not
only to be nice and nurturing (traditional expectations) but
also to demonstrate traits in the traditionally masculine area of
competency (De Lisi & Soundranayagam, 1990;Nesbitt&
Penn, 2000).
In addition, in this study, a healthy adult woman had
more endorsed traits than either a healthy adult man or a
healthy adult. Men’s item agreement scores for traditionally
feminine items on the healthy adult women’s stereotypic
items were generally higher than those reported by women.
Men continue to appear to expect women to be more
traditionally feminine than women themselves believe is
appropriate. It is interesting that both sexes agree that
descriptions of a healthy adult woman include behaviors
that are considered traditionally masculine. For example,
high consistency occurred between the sexes on the traits
“independent”and “enjoys a challenge.”However, wom-
en’s item agreement score on “strong”was much higher
than that reported by men.
In contrast to the changes in gender role stereotypes for a
healthy adult woman, perceptions of a healthy adult man have
showed little change since the early work of Rosenkrantz
et al. (1968). Men are still expected to display traditionally
masculine characteristics and behaviors. In contrast to the
healthy adult woman, who now is described in both
traditionally feminine and masculine terms, the healthy adult
man is still perceived solely in traditionally masculine terms.
The soft, sensitive man may play well in the movies—but
our findings suggest that counselors-in-training would find
such a man less than healthy.
Most of the few traits used to describe a healthy adult
continue to be traditionally masculine. Given that most item
agreement scores were around the 50% mark, it is possible
that a healthy adult is perceived to be one who is neither
strongly feminine nor strongly masculine, but rather more
androgynous. On the other hand, it may be that there is no
clear consensus regarding how to describe a healthy adult.
Future researchers need to clarify these results to determine
whether or not it is possible to conceive of a healthy adult
outside of the constraints of traditional gender roles or
whether this finding was related to participant bias. It is
possible that given the within subjects design of the study,
that participants had “guessed”what responses were desired
and thereby formed their answers based on these assump-
tions. Future research could check these issues by having an
independent sample complete only the section for a healthy
adult, sex unspecified.
The most important finding is that a healthy adult
woman continues to be significantly different than a healthy
adult man and a healthy adult, sex unspecified. Yet again, a
healthy adult man was not found to be significantly
different from a healthy adult, sex unspecified. Stereotypic
items between a healthy adult woman and a healthy adult
man overlap only on the three traditionally masculine items
of “strong,”“independent,”and “enjoys a challenge.”
Otherwise, the remaining 17 stereotypic items for a healthy
adult woman continued to be traditionally feminine,
whereas the remaining 12 items for a healthy adult man
continued to be traditionally masculine. A healthy adult
woman also continued to be significantly different from a
healthy adult whereas a healthy adult man was not found to be
Table 2 Continued
Healthy adult woman Healthy adult man Healthy adult
Participant sex Female (%) Male (%) Female (%) Male (%) Female (%) Male (%)
Not at all afraid to take risks 44 40 78 68 50 58
Brave 57 46 73 74 49 58
Not at all excitable in a minor crisis 33 33 60 61 41 48
Has a strong will 67 71 79 78 67 68
Authoritative 48 33 71 70 45 42
Competitive 53 48 82 85 52 62
Objective 47 43 48 60 42 51
Not at all dependent 36 47 23 60 48 46
Not at all passive 35 31 65 52 8 13
Sex Roles (2006) 55:247–258 255

significantly different from this healthy adult. Future research-
ers should investigate these findings further. The fact that the
number of stereotypic items for a healthy adult woman was
higher than the number for a healthy adult man may have
caused this finding due to a statistical bias. Different results
might occur if the numbers of stereotypic items were equal
as discussed by Widiger and Settle (1987), who claimed
that this difference was a statistical bias, based on the num-
ber of socially desirable traits included on the Stereotype
Questionnaire.
Our findings also suggest that the gender role stereo-
types used by counselors-in-training appear not to have
changed much since the early 1970s. Gender role stereo-
types remain different for a healthy adult man and a healthy
adult woman, as well as different for a healthy adult woman
and a healthy adult, sex unspecified. On the other hand, the
results are not as clear-cut as those found in the earlier work
of Rosenkrantz et al. (1968) and Broverman et al. (1970,
1972). The current results do not suggest that counselors-
in-training view a healthy adult man as the “gold standard.”
Rather, the results suggest that they hold two different
standards of mental health—one for healthy women and
one for healthy men. These two standards are highly
traditional in terms of gender role stereotypes. The most
obvious change in these stereotypes is the addition of
several competency traits in the description of a healthy
adult woman.
General Discussion
To conclude, several changes have occurred in gender role
stereotyping over the past 35 plus years. The characteriza-
tion found for healthy adults was sketchy at best, but
included mostly traditionally masculine traits. These traits
appear to be the core descriptors of mental health, as three
of the five traits (“strong,”“independent,”and “enjoys a
challenge”) were included among the stereotypic items for
men and women, as well as for an adult, sex unspecified.
The other two traits were not shown to overlap between
healthy men and healthy women. One, “understanding of
others,”was shown to be stereotypically feminine, whereas
Table 3 Stereotypic items included in the healthy adult woman, healthy adult man and healthy adult stereotypic scores.
Healthy adult woman—stereotypic items Healthy adult man—stereotypic items
Item Percent Item Percent
Sympathetic 77 Knows the way of the world 72
Sentimental 70 Strong 77
Charitable 78 Independent 86
Understanding of others 83 Enjoys a challenge 83
Warm in relations to others 86 Industrious 75
Considerate of others 86 Daring 76
Interested in own appearance 83 Decisive 71
Concerned about others 81 Firm 75
Neat in habits 75 Direct 79
Enjoys art and literature 72 Adventurous 81
Sensitive 83 Not at all afraid to take risks 74
Gentle 78 Brave 74
Affectionate 85 Has a strong will 79
Compassionate 87 Authoritative 70
Soothing 82 Competitive 83
Loving 87
Independent 71
Enjoys a challenge 73
Aware of the feelings of others 78
Strong 71
Not at all cold 75
Healthy adult—stereotypic items
Item Percent
Understanding of others 71
Knows the way of the world 73
Strong 76
Independent 76
Enjoys a challenge 74
256 Sex Roles (2006) 55:247–258

the other, “knows the way of the world,”appeared to be a
stereotypically masculine item. It is important that mental
health is no longer solely considered synonymous with
masculine traits, as one traditionally feminine trait is now
viewed as mentally healthy. Additional research focused on
healthy adults might help to clarify the current standard for
mental health.
Another important finding is the fact that the feminine
gender role stereotype has changed, but the masculine gender
role stereotype remains consistent with earlier research.
Women now are expected to possess both nurturing and
competency traits, which suggest that acceptable behaviors
for women are broader and more androgynous. In contrast,
men’s traditional gender role expectations continue to focus
solely on competency traits and overlap with most of the
characteristics of a healthy adult. But, even though the gender
role stereotype for women has changed, a healthy adult
woman was still found to be significantly different from both a
healthy adult man and a healthy adult, sex unspecified.
Social desirability ratings also have not changed over the
last 30 plus years. What are classified as socially desirable
traits seem to be deeply, culturally engrained. Given the
inflexibility of the social desirability of traits, characteristics,
and behaviors, the definition of femininity and masculinity
needs to become more androgynous in order to create one
standard of mental health. This change would help to
eliminate the double standard of mental health for women
and men.
Although no differences were found when in comparisons
of previous research with between and within group designs
(Broverman et al., 1970,1972; Nesbitt & Penn, 2000;
Phillips & Gilroy, 1985; Rosenkrantz et al., 1968; Widiger
& Settle, 1987), replication of Study 2 with a between
subjects design would help to confirm the results found
here. Another methodological issue here is the interrelation-
ships among the many items. Many of these items were
endorsed at fairly high levels but did not reach the criterion
level we set. Examination of how these traits cluster
together might reveal the interrelationships among items.
One additional methodological issue remains, and it is
related to the development of items for the questionnaire.
Four items on the Stereotype Questionnaire were not
effectively changed into unipolar items: “decisive”–“indeci-
sive,”“relies on self”–“relies on others,”“very careful”–“v-
ery careless,”and “acts on logic rather than feeling”–“acts on
feeling rather than logic.”A socially desirable pole was not
found for the last item. If two items were created from this
one item, such that they were worded as “acts on logic rather
than feeling”–“does not act on logic rather than feeling”and
“acts on feeling rather than logic”–“does not act on feelings
rather than logic”the first item might have been rated as
socially desirable. “Decisive”maintained it affiliation with
a healthy adult man, but the other two items were not
included as a part of the stereotypic scores on any of the
Stereotype Questionnaires formats (i.e., healthy adult
woman, healthy adult man, or healthy adult). It is possible
that, if these items were reworded, differences might
appear. That we overlooked these bipolar opposites while
overtly attempting to eliminate them demonstrates how
deeply ingrained and covert our own stereotypes can
become. Future researchers should correctly assign these
four items to unipolar formats.
Finally, due to the transparency of the study, participants
may have responded in socially desirable ways. Hiding the
intent of the study might reveal different results. However,
participants still endorsed traditionally socially desirable
traits and gender role stereotypes for both women and men.
This finding suggests that, although participants may be
conscious of gender role stereotypes, they still may hold
implicit stereotypes that were revealed in this study. The
wording in the directions for the Stereotype Questionnaire
may also impact these implicit stereotypes. Participants are
asked to think of a “normal”adult (women/man) who is
“mature, healthy, (and) socially competent.”Here “normal”
and “healthy/socially competent”were not separately
presented, which might suggest that there is a “typical”
adult (woman/man). If one then assumes that there is a
typical individual, one might have a tendency to employ
stereotypes based on the instructions (Hoffman & Borders,
2001). This confound should be investigated to check
whether or not these types of directions can influence the
task that was employed in this research.
Future researchers need to investigate whether or not the
different mental health standards for women and men are
explicitly perceived or only implicitly held by counselors-
in-training. Future counselors may be unaware that they
hold two different standards of mental health and, there-
fore, would be expected to deny that they would counsel
women differently than men. If future counselors implicitly
hold these different standards, course work needs to be
focused explicitly on bringing this issue into conscious
awareness.
Acknowledgements We thank Mathew Winter for all his help in
earlier presentations, data entry and discussions regarding potential
outcomes. In addition, we thank Adrian Tomer and the anonymous
reviewers for many insightful comments.
References
Broverman, I. K., Broverman, D. M., Clarkson, F. E., Rosenkrantz, P. S.,
& Vogel, S. R. (1970). Sex-role stereotypes and clinical judgments
of mental health. Journal of Consulting and Clinical Psychology,
34,1–7.
Broverman, I. K., Vogel, S. R., Broverman, D. M., Clarkson, F. E., &
Rosenkrantz, P. S. (1972). Sex-role stereotypes: A current
appraisal. Journal of Social Issues, 28,59–78.
Sex Roles (2006) 55:247–258 257

Crawford, M., & Unger, R. (2000). Women and gender: A feminist
psychology (3rd ed.) New York: McGraw-Hill.
De Lisi, R., & Soundranayagam, L. (1990). The conceptual structure of
sex role stereotyping in college students. Sex Roles, 23, 593–611.
Hoffman, R. M., & Borders, L. D. (2001). Twenty-five years after the
Bem Sex-Role Inventory: A reassessment and new issues
regarding classification variability. Measurement and Evaluation
in Counseling and Development, 34,39–55.
Lerner, R. M. (Ed.). (2002). Concepts and theories of human
development (3rd ed.). Mahwah, New Jersey: Erlbaum.
Nesbitt, M. N., & Penn, N. E. (2000). Gender stereotypes after thirty
years: A replication of Rosenkrantz et al. (1968). Psychological
Reports, 87, 493–511.
Phillips, R. D., & Gilroy, F. D. (1985). Sex-role stereotypes and
clinical judgments of mental health: The Brovermans’findings
reexamined. Sex Roles, 12, 179–193.
Rosenkrantz, P., Vogel, S., Bee, H., Broverman, I., & Broverman, D. M.
(1968). Sex-role stereotypes and self-concepts in college students.
Journal of Consulting and Clinical Psychology, 32, 287–295.
Widiger, T. A., & Settle, S. A. (1987). Broverman et al. revisited: An
artifactual sex bias. Journal of Personality and Social Psychol-
ogy, 53, 463–469.
258 Sex Roles (2006) 55:247–258

Reproducedwithpermissionofthecopyrightowner.Furtherreproductionprohibitedwithoutpermission.