The Spanish Journal of Psychology
Copyright 2008 by The Spanish Journal of Psychology
2008, Vol. 11, No. 2, 609-617
ISSN 1138-7416
Evolution of Gender Stereotypes in Spain: Traits and Roles
Mercedes López-Sáez, J. Francisco Morales and Ana Lisbona
Universidad Nacional de Educación a Distancia (Spain)
The aim of this study is twofold: to determine whether (and how) gender stereotypes
have changed over time through a comparison of two different sets of data collected in
1993 (N = 1255) and 2001 (N = 1255) from a representative sample of the Spanish
population, and to examine the relation between gender traits and roles and its stability
over time. In addition, special attention is paid to the psychometric properties of the
measures of gender traits and roles used in the study. The content of gender stereotypes
was found to remain stable over the target period of time, confirming the classical typology
(a higher assignment of expressive-communal traits to women and of instrumental-agentic
traits to men). The structure of the gender-role questionnaire allows us to distinguish
between family-role and work-role stereotyping. Gender-role stereotyping shows a marked
decline between 1993 and 2001, a result that contrasts with the stability of trait-role
stereotyping. The fact that a very low correlation is observed at the two time points
between these two components of gender stereotyping strongly suggests their independence.
Keywords: gender stereotypes, instrumental traits, expressive traits, role stereotypes
En este estudio se persigue un doble objetivo: comprobar la evolución de los estereotipos
de género en dos aplicaciones, efectuadas en 1993 (N = 1255) y 2001 (N = 1255), ambas
con muestras representativas de la población española, y analizar la relación que existe
entre rasgos y roles, examinando si se han producido cambios en esa relación con el
paso del tiempo. Asimismo, se analizan las propiedades psicométricas de las medidas
de rasgo y de rol empleadas. Los resultados muestran que el contenido de los estereotipos
de rasgo no se ha modificado, confirmándose la clásica tipología en la que se asignan
más rasgos expresivo-comunales a las mujeres que a los hombres y, por el contrario,
más rasgos instrumental-agentes a los hombres. La estructura del cuestionario de roles,
permite diferenciar entre estereotipia de rol familiar y estereotipia de rol laboral. Al comparar
la evolución de los estereotipos en este periodo, se observa un descenso en la estereotipia
de la población española, más acusada en el componente de rol que en el de rasgo. La
correlación entre ambos componentes es muy baja o no significativa, sin que se observen
cambios importantes en la evolución de esa relación. Estos resultados parecen corroborar
la independencia entre los componentes de rasgo y de rol.
Palabras clave: estereotipos de género, rasgos instrumentales, rasgos expresivos,
estereotipia de rol de género
The sampling of this investigation was financed by the Instituto de la Mujer.
Correspondence concerning this article should be addressed to the authors: Facultad de Psicología, UNED, Departamento de Psicología
Social y de las Organizaciones, C/ Juan del Rosal, 10, 28040 Madrid (Spain). E-mails: mlopez@psi.uned.es; folco_and@yahoo.es;
amlisbona@psi.uned.es Translation: Virginia Navascués Howard
How to cite the authors of this article: López-Sáez, M., Morales, J.F. and Lisbona, A.
609

---

610
LÓPEZ-SÁEZ, MORALES, AND LISBONA
In 1992, thanks to the initiative of the Instituto de la
or to leadership (Eagly & Karau, 2002; Killen, López-Zafra,
Mujer [Women’s Institute], various specialists elaborated a
& Eagly, 2006; Powell, Butterfield, & Parent, 2002; Sczesny,
system of indicators to appraise the situation of Spanish
Bosak, Neff, & Schyns, 2004), to mention some examples
women in comparison to men. Among the indicators were
from recent research, For the investigators of gender
included gender stereotypes. This study, in its key aspects,
inequalities, the following question is unavoidable: If reality
is based on the proposal of gender stereotyping indicators
changes, do stereotypes change?
of Morales and López-Sáez (1993, 1994). The idea behind
Gender stereotypes and social reality. Initially,
the inclusion of these subjective measures is that the
investigations of gender stereotypes focused on the
internalization of stereotypes affects behavior and is the
differentiation of two kinds of traits: expressive-communal
origin of gender differences. One of the main goals of this
traits, associated with femininity, and the instrumental-
system of indicators was periodic assessment, in order to
agentic, associated with masculinity (Bem, 1974; Spence,
obtain a diagnosis of the situation at a concrete point in time
1993; Spence & Buckner, 2000; Spence & Helmreich, 1978).
and to determine the evolution over time. In this article,
At the beginning of the 1980s, within the framework of the
changes in gender stereotyping between the first
schema theories, a series of approaches were developed that
administration of the indicators—1993—and the last one
deal with the relation between masculine and feminine
with available data for the time being—2001—are verified.
personality traits and gender stereotyped behaviors. These
Within the system of indicators, two components were
theories are based on identifying the social prescriptions of
used to measure gender stereotyping: personality traits and
femininity and masculinity. Sandra Bem (1974) developed
roles (Morales & López-Sáez, 1993, 1994). Trait stereotypes
a measurement of gender identity based on personality traits,
reflect psychological characteristics that are more frequently
the Bem Sex Role Inventory (BSRI), choosing the traits that
attributed to men or women, whereas role stereotyping
were significantly more desirable for women than for men
focuses on beliefs about the activities considered more
and those that were significantly more desirable for men
appropriate for men and for women. The theoretical
than for women. The BSRI measures individual differences
framework that supports the choice of this measurement of
in the internalization of the prescriptions of femininity and
indicators is integrated within the relation that various
masculinity. The self-schema gender theory of Bem (1981)
authors have established between trait and role gender
predicts differences between people who identify with the
stereotypes (Eagly & Wood, 1991, 1999; Eagly, Wood &
traditional traits of their sex—that is, women with high
Diekman, 2000; Wood & Eagly, 2002). The social changes
identification with femininity and men with high
in women’s roles throughout the 20th century are
identification with masculinity—and the rest of the people,
indisputable. Currently the cultural models tend to equate
when processing gender-related information. Only such sex-
both genders, and the roles of men and women are very
typified people would process gender information, even self-
similar, especially in the professional area. It seems logical
related information, schematically. From a similar approach,
to think that such social changes will affect the beliefs about
for Markus and collaborators, behavior would depend on
the characteristics of both genders. Specifically, in Spain,
the type of stimulus, and people’s sex would not affect it
for over more than a decade, more women than men are
(Markus, Crane, Bernstein, & Siladi, 1982): People who
studying university careers. However, the inequalities persist:
highly identify with femininity are schematic when
the percentage of women who study architecture and
processing information associated with feminine aspects,
engineering only reaches 27%; among university graduates,
whereas people who highly identify with masculinity will
the number of unemployed women is twice that of men,
only be schematic when processing information related to
and the number of women hired at the highest level of
masculine features.
university professors does not even reach 14%. There is
The theories of gender schemas are strictly cognitive,
only 2.8% of women on the boards of administration of
as they focus on the way people process information about
the main Spanish companies. Moreover, women’s salary is
gender as a function of their identity, without taking into
27.2% less than men’s for the same job, and women who
account the way the context may affect such processing.
work outside of the home devote more than twice the time
Although the social gender identity is relatively stable, and
to housework than men do. In posts of political power,
men tend to identify more with the traditionally masculine
despite the advances in equality, women only have 29% of
stereotypes and women with the feminine ones, people’s
the representation in all the city halls, Parliaments, and
self-perception varies as a function of the situation in which
Senate (Source: Instituto de la Mujer-INI [Women’s
they find themselves. Echebarria & González (1999), using
Institute-National Institute of Statistics]).
a version of the BSRI, found changes in the pattern of
Gender studies reveal the persistence of inequalities
gender identity as a function of whether or not the context
between men and women and these differences can largely
was public or private: In a professional situation, men and
be partially explained by the persistence of stereotypes
women identified more with instrumental traits, whereas in
related to work (Alter & Seta, 2005; Glick, Larsen, Johnson,
a private context, both groups identified more with
& Branstiter, 2005; Selwyn, 2007; White & White, 2006)
expressive characteristics.

---

EVOLUTION OF GENDER STEREOTYPES IN SPAIN
611
Not only does the proximate context affect gender
is organized around the subtypes of both social categories.
differences. The role of social structures is crucial for the
In the perception of people, categories other than gender—
maintenance or change of men-women differences. There
family and professional role, affiliation groups, ideology,
is mutual influence between beliefs and structures, so that
physical appearance, or personality—may be used to classify
society changes when citizens’ beliefs change, but also,
people (Carpenter & Trentham, 1998, 2001). Moreover, the
changes in society contribute to changing beliefs. Therefore,
characteristics of the situation may enhance the use of certain
a relation is established between sexist beliefs and structural
subtypes when perceiving other people or oneself (Glick et
inequalities between men and women. For example,
al., 2005; Echebarria & González, 1999).
regarding sexist attitudes, Glick et al. (2000) found that the
Investigation of the subtypes of men and women has
countries that scored high in sexist attitudes had the worst
found some consistent patterns (Deaux, Winton, Crowley,
indicators of women’s equality. Along these lines, in a
& Lewis, 1985; Six & Eckes, 1991). Among the women’s
representative sample of the population of Galicia (Spain),
subtypes are included professionals, housewives, athletes,
Glick, Lameiras, and Rodríguez-Castro (2002) confirmed
and sexy women, and among men’s, macho-men, business
the influence on sexist attitudes of two very important
men, athletes, and family men. This division suggests than
institutions in any culture: the educational system and
a person’s gender affects the categorization made when they
religion. These authors found a negative correlation between
are classified in a subtype. Some investigations on subtypes
the educational level and hostile or benevolent sexism, and
support the idea that the supraordinal category of gender,
a positive one between Catholic religiosity and benevolent
as a salient characteristic, has primacy in people’s perception
sexism. This empirical evidence, found in correlational
(Brewer & Lui, 1989; Stangor, Lynch, Duan, & Glas, 1992).
works, was corroborated in an experimental study by Jost
Although classifying people according to subtypes can be
and Kay (2005) that showed that the activation of gender
more informative than doing so by broad categories such
stereotypes contributes to justify system inequalities in gender
as gender, when knowledge about the person is minimal or
roles.
the characteristics of the situation equate men and women
From a clearly psychosocial approach, the social role
as belonging to a same subtype (i.e., professionals or
theory postulates a relation between roles and stereotypes,
parents), dimensions such as gender are used to classify
so that perceptors attribute to the people who play a certain
people. A woman is more closely associated with feminine
role, specific internal dispositions that are coherent with
characteristics and, therefore, does not match a role
that role (Eagly & Wood, 1991, 1999; Eagly et al., 2000;
associated with masculine characteristics, because she will
Wood & Eagly, 2002). According to this proposal, the
be perceived as having fewer traits related to the competence
origin of descriptive gender stereotypes lies in the
demanded for a traditionally masculine task. This explains
inference of correspondence between what men and
the differences maintained between men and women despite
women do and their internal dispositions. The social
the formal equalities, in essential facets such as work, salary,
distribution of roles has traditionally assigned tasks that
couple violence, choice of career, power, or attention to
require instrumental-agentic characteristics to men and
dependent people. It has been confirmed that gender
tasks that demand expressive-communal qualities to
stereotypes play an important role in the expectations of
women. Therefore, from this theoretical approach one
success and satisfaction in the selection of a job applicant,
would expect a high relation between role and trait
depending on whether the post is traditionally feminine or
stereotypes and, as role assignation becomes more
masculine (see Alter & Seta, 2005; also and using a
egalitarian, the psychological differences attributed to men
qualitative methodology, discourse of women in traditionally
and women should gradually disappear.
masculine posts is analyzed by López-Sáez, Lisbona, and
But the cross-cultural study of Williams and Best (1990)
Sáinz, 2004).
does not corroborate that social changes lead to a change
The present study has a dual purpose: (a) to compare
in the stereotypes of men’s and women’s characteristics.
the results obtained in the measurement of gender stereotypes
These authors found a high consensus among countries in
in the last two applications, effected in 1993 and 2001, both
trait stereotypes that reproduced the two classic dimensions:
with representative samples of the Spanish population, in
expressiveness associated with women and instrumentality
order to determine how they have evolved during this period;
with men. According to their data, the differences among
and (b) to study in depth the relation between traits and
countries was more closely related to the system of values
roles, to determine whether changes in this relation have
or religious beliefs than to structural aspects that reflected
occurred over time. The psychometric properties of the
real changes, such as women’s education or their rates of
measurements proposed as indicators by Morales and López-
activity.
Sáez (1993, 1994) will also be analyzed. Although these
Content of gender stereotypes and the perception of
measuring instruments have been used periodically, the
people. The measurement of gender stereotypes poses the
results of the psychometric assessment had not been
problem of whether there is an abstract representation of
published, although they had been included in the reports
men and women in general, or whether this representation
sent to the Women’s Institute.

---

612
LÓPEZ-SÁEZ, MORALES, AND LISBONA
Method
less than 0, it is stereotypically feminine. The DR is
conceived as a stereotyping measurement that allows one
Participants
to predict the probability of assigning a characteristic of one
group to another (Allen, 1996; McCauley, Stitt, & Segal,
We compared two representative samples of the Spanish
1980). With this kind of measurement, some of the problems
population, over 18 years old, obtained 8 years apart: in
associated with general measurements of gender stereotypes
1993 and 2001. The type of sampling was polystage,
are overcome. If the proportion of women who have a trait
stratified cluster according to the habitat where they resided
is twice that of the men, in the absence of other specific
in the first stage and to the home in the second stage. In
data about each person, the probability of assigning that
the last stage, the people from each home were selected to
characteristic more to a woman than to a man will be twice
obtain the sampling quotas of sex, age, and occupation
as high in any context.
corresponding to each sample section. The size of each of
Roles. The measurement of role stereotyping is based
the samples was as follows: 653 women and 602 men in
on a selection of items from the Questionnaire of Sexual
1993; 640 women and 615 men in 2001.
Role Ideology (Moya, Navas, & Gómez, 1991). In the 1993
application, 15 items from this questionnaire were used
Instruments and Procedure
(López-Sáez & Morales, 1995). After exploratory factor
analysis, the items that met the criterion of factor loadings
We used a questionnaire that included measurements of
over .50 were selected. Thus, in the 2001 application, a 9-
trait and role stereotyping, in addition to other indicators,
item scale was used (see Table 2 in the Results section).
and sociodemographic characteristics. Data were collected
Participants were asked to rate their degree of agreement,
in the interviewees’ homes, by professionals who were duly
ranging from 1 (disagree completely) to 5 (completely agree).
trained for this purpose.
Personality traits. To measure trait stereotyping, in two
Statistical Analyses
sections of the questionnaire, participants were asked to
indicate the percentage of men and women in general who
A t-test was used to compare each DR with 0. The
possess each one of the 17 traits (see Table 1 in the Results
evolution of the differences between the two years was
section). These items were selected in three successive
verified by means of ANOVA if the variances were
phases. From the traits of the BSRI of Bem (1974), the 14
homogeneous (Levene’s test nonsignificant), and with a
items with the most gender stereotyping in two different
nonparametric test, Wilcoxon’s W, if this assumption was
samples were selected (Morales & López-Sáez, 1994). In
not met. The reliability of the scales was calculated by means
addition, 6 negative items were included, which were
of Guttman’s split halves if the items were a ratio
extracted from the review of other works. Of the 20 traits
measurement, and with Cronbach’s alpha if the items were
selected in 1993, the results showed that 17 were
in scalar format. We used Cohen’s test to compare
stereotypical, and they were selected for subsequent
correlations. We performed confirmatory factor analysis using
applications.
Amos 5.0 on the data from the 2001 sample to verify the fit
Trait stereotyping was calculated following the procedure
to a model of the dimensions of trait and role stereotyping.
of Martin (1987). For each trait and each participant, a
diagnostic ratio (DR) was calculated by dividing the
percentage of men to whom the trait was attributed by the
Results
percentage of women to whom it was attributed, and
subsequently performing a mathematical transformation to
Trait Stereotyping
equate the range of the original ratio measurements (OR).
In each case, if the quotient obtained was higher than or
Firstly, we analyzed the content of the trait stereotypes
equal to 1, it was transformed by subtracting 1 (DR = OR
in 2001, verifying whether the DR of each characteristic
- 1). If the quotient was less than 1, the inverse of the
was significantly different from 0.
original ratio was subtracted from 1 (DR = 1–1/OR). With
The content of the beliefs about which characteristics are
this transformation, the absence of stereotyping receives a
possessed by men more than by women had not changed
score of 0, positive scores indicate that the trait is attributed
substantially in the interval analyzed. In the two samples, in
more to men, and negative scores indicate that it is attributed
all the traits, the stereotypes that assign more instrumental and
more to women. Once the transformation is performed, the
agentic characteristics to men than to women were maintained.
DR used as an indicator is the mean DR for the sample. A
Therefore, the content of masculine trait stereotyping observed
mean DR of 0 indicates that in that population, there is no
in 1993 still occurred in 2001 (see Table 1). The differences
stereotyping of that trait. If the mean DR is significantly
between the two years was verified by means of Wilcoxon’s
higher than 0 (p of the t statistic), it can be concluded that
W. Regarding the changes produced between 1993 and 2001,
the trait is stereotypically masculine. If it is significantly
in masculine trait stereotyping, the decrease in all the traits

---

EVOLUTION OF GENDER STEREOTYPES IN SPAIN
613
was noteworthy, except for “aggressive,” which increased (p
absolute value of r (.54) allowed us to verify the structure
< .05 in all cases), and for the traits “hard-hearted,” “strong
of a conjoint model of two correlated dimensions, those
personality,” and “athletic,” which did not undergo significant
corresponding to the masculine and feminine items. This
changes. Regarding personality traits associated more with
model presented the following goodness-of-fit indexes: GFI
women than with men, in both samples, traditional feminine
= .95, AGF = .94, RMSEA = .05. Therefore, it can be
trait stereotyping was repeated, with more expressive-communal
concluded that, both theoretically and empirically, the items
characteristics assigned to women. When comparing the results
of this questionnaire are adequate to measure the two classic
of 1993 and 2001, the increase in the strength of the stereotypes
dimensions of trait stereotyping in Spanish populations.
in the traits of “tender,” “compassionate,” “warm,” and
Once the appropriateness of the questionnaire had been
“affectionate” was notable (p < .05 in all cases); “submissive”
verified, for subsequent analyses, we calculated a global
did not change, and stereotyping in the rest decreased (p <
mean of masculine trait stereotyping (MTS) and in feminine
.05 in all cases).
trait stereotyping (FTS), by adding the DRs of the items
Confirmatory factor analysis was performed to verify
defined as masculine or feminine, respectively, and weighting
the psychometric fit to a two-dimensional model:
as a function of the number of items. High scores in MTS
instrumentality-agency and expressiveness-communality.
and low scores in FTS (because of the negative sign of these
The estimation method used was maximum likelihood, which
DRs) indicate that the participants stereotype traditionally.
is more adequate than unweighted squared minimums for
We also obtained a global mean of trait stereotyping
large sample sizes (Ximénez & García, 2005). The model
(GTS), by calculating the mean of MTS and FTS (in
based on DRs of the 9 masculine items presented good fit
absolute values). High scores in this measure indicate more
indicators: goodness of fit index (GFI = .99), adjusted
stereotyped responses in both kinds of traits. The 1993 mean
goodness of fit index (AGFI = .97), and the root mean
of the masculine traits (mean MTS = 1.27, SD = 3.6;
square error of approximation (RMSEA = 05). The indicator
Guttman = .57) was significantly higher (z = 2.59, p < .01)
based on chi-square was not taken into account as it is not
than the 2001 mean (mean MTS = .95, SD = 2.4; Guttman
adequate for samples of over 1000 cases. Likewise, the fit
= .65). A decrease in the global FTS was also observed,
of the model based on the 8 feminine items was also
because the 1993 mean (mean FTS = .91, SD = 2.48,
confirmed: GFI = .98, AGFI = .95, RMSEA = .06. The
Guttman = .74) was significantly more stereotyped (z = 1.09,
Table 1
Diagnostic Ratios (DR) and t-Values comparing Stereotypically Masculine and Feminine Traits Measured in 1993 and 2001
1993
2001
Traits DR
t
DR
t
Masculine Traits
Acts like a leader
1.47
8.31**
1.07
10.78**
Adventurous
2.76
9.79**
1.48
8.69**
Egotistic
.66
4.27**
.56
3.92**
Individualist
.65
3.78**
.59
3.20**
Hard-hearted
1.36
6.50**
.84
6.85**
Aggressive
1.48
7.25**
1.67
7.30**
Strong personality
.22
1.92*
.30
2.50*
Athletic
1.87
7.27**
1.46
8.69**
Feminine Traits
Submissive
–1.24
5.23**
–1.42
6.50**
Loves children
–.56
3.67**
–.53
6.00**
Soft-hearted
–.62
5.00**
–.85
6.00**
Understanding
–.46
5.56**
–.37
7.00**
Compassionate
–.37
2.77**
–.44
3.91**
Sensitive
–.36
2.47**
–.39
12.67**
Cries easily
–4.01
10.56**
–2.85
9.34**
Warm
–.30
10.00**
–.53
3.86**
Affectionate
–.26
3.13**
–.56
4.00**
*p < .05. **p < .01.

---

614
LÓPEZ-SÁEZ, MORALES, AND LISBONA
p < .04) than the 2001 mean (mean FTS = .87, SD = 2.36,
according to the loadings obtained from the 2001 data. The
Guttman = .60). However, in the GTS, no significant
first factor extracted was called Family Role Stereotyping
differences were observed between 1993 (mean GTS = 2.17,
and the second one Work Role Stereotyping. These two
SD = 3.6, Guttman = .63) and 2001 (mean GTS = 1.84, SD
factors represent different dimensions of gender stereotyping,
= 3.93, Guttman = .72). The reliability coefficients of the
with theoretically different contents. Therefore, in addition
scales were satisfactory and, in general, an improvement
to the total of the scale, we considered each one of the two
was observed in the 2001 application.
dimensions as subscales of the measurement of stereotyping.
These results show that the global content of the trait
The reliability of each scale was acceptable or good, and
stereotypes has not changed substantially. However, it is
an improvement was observed in the 2001 sample compared
noteworthy that, in all the stereotypically masculine traits
to the 1993 sample. Confirmatory factor analysis allowed
except for “aggressive,” the DR decreased from 1993 to
us to determine that the structure of the two dimensions,
2001, which indicates that in our society, there is a tendency
Family Role and Work Role, which were correlated (r =
to perceive men and women as more equal in the
.62), corresponded to a model with good fit indicators: GFI
characteristics of the instrumental-agentic dimension. This
= .97, AGFI = .95, RMSEA = .06.
tendency was not observed in stereotypically feminine traits,
According to the one-factor within-subject ANOVA,
as they decreased only in three traits and four traits increased.
Family Role Stereotyping was higher than Work Role
Stereotyping (see Table 3), both in the 1993 sample, F(1,
Role Stereotyping
1179) = 1543, p < .0001, and in the 2001 sample, F(1, 213)
= 1673, p < .0001. When comparing 1993 with 2001
To explore the structure of this scale, we conducted an
(Wilcoxon’s W), we observed that stereotyping in the total
exploratory factor analysis with principal component method
scale decreased (z = 7.82, p < .0001), as well as in the
and varimax rotation. The factor structure was the same in
Family Role (z = 6.61, p < .0001) or the Work Role
both samples, although the loadings varied in the two
subscales (z = 10.56, p < .0001). To sum up these results,
applications. Conjointly, the two factors explain 33% of the
in role stereotyping, the changes are important, especially
variance of 1993 and 48.6% of that of 2001. Table 2 displays
those referring to aspects related to work roles, but the
the results of these analyses, with the items arranged
changes in family role stereotyping are less intense.
Table 2
Factor Structure of the Role Gender Stereotyping Scale
Factors and Items
1993 2001
FACTOR 1: FAMILY ROLE: % explained variance
23.7%
39.0%
(Cronbach ? of subscale)
(.75)
(.81)
Factor loadings
If a child gets sick and both parents work, it is generally better for the mother to ask for time off at work
to care for the child.
.67
.76
It is better for a woman to try to achieve security by encouraging her husband at work rather than to get
ahead of him in her own career.
.66
.66
It is more important for a woman than for a man to be a virgin until she marries.
.65
.61
I think it is much more disagreeable for a woman to swear and use vulgar language than it is for a man to do so.
.60
.59
It’s natural for men and women to perform different tasks.
.59
.58
Maternity is the greatest source of satisfaction a woman can have.
.61
.44
FACTOR 2: WORK ROLE: % explained variance
9.7%
9.6%
(Cronbach ? of subscale)
(.53)
(.80)
When a high percentage of women begin to have access to a profession, it tends to lose social prestige.
.72
.76
The woman who limits herself to her profession tends to adopt masculine traits and behaviors.
.56
.72
At work, women do not usually have original ideas because they are too concerned about their labor security.
.54
.70
Total Role: % explained variance
33.4%
48.6%
(Cronbach ? of total scale)
(.76)
(.84)

---

EVOLUTION OF GENDER STEREOTYPES IN SPAIN
615
Table 3
Discussion
Evolution of Role Stereotyping. Means and Standard
Deviations (in Brackets)
The consistency of the results obtained in the
measurement of stereotyping, both of traits and roles, in two
1993 2001
applications at different moments in time reveals the
Total Scale
2.92 (.79)
2.64 (.90)
usefulness of the instruments employed to measure
Family Role
3.32 (.96)
3.03 (1.04)
stereotyping in the Spanish population, both because of their
Work Role
2.16 (.85)
1.84 (.96)
psychometric properties and their theoretical validity.
Regarding the content of trait stereotypes in Spanish samples,
confirmatory factor analysis ratifies the validity of the
classical dimensions of expression-communality, attributed
Relation between Trait Stereotyping and Role
more to women, and instrumentality-agency, attributed more
Stereotyping
to men, as well as the currency of these stereotypes. This
content has not changed substantially over eight years. Our
In Table 4 can be seen the correlations between the
results corroborate those found in a recent study carried out
different subscales of trait stereotyping (absolute values)
by Spence and Buckner (2000), using items from the
and role stereotyping. The upper part of the matrix displays
Personal Attributes Questionnaire (PAQ; Spence &
the data from the 1993 sample; the lower part, those
Helmreich, 1978) and the BSRI (Bem, 1974). These authors
corresponding to 2001, in boldface. There were no significant
found that the dimensions of trait stereotyping
correlations between the components of trait and role
(instrumentality associated with men and expressiveness
stereotyping in the 1993 sample, and the magnitude was
with women) that were used to elaborate the PAQ and the
very low (never higher than .12) in the 2001 sample. The
BSRI in the 1970s were maintained in 1996. Moreover, our
correlation found in the total trait and role stereotyping of
data show that society evolves to equate men and women
1993 (r = .04) was not significantly different (z = 1.5, p <
in the traits traditionally considered “masculine,” but not so
.13) from that of 2001 (r = .10). According to these results,
much in those traditionally considered “feminine.” No doubt,
we could not corroborate the hypothesis of the relation
the fact that women have adopted the characteristics
between trait stereotyping and role stereotyping, and although
associated with the instrumental-agentic dimension to a
the correlation tended to increase, the difference was not
greater extent than men have assumed the related expressive-
significant.
communal qualities has influenced this change and, therefore,
However, the correlations between the subscales of the
this aspect of trait stereotyping has decreased.
same component—feminine and masculine traits, or family
Regarding roles, gender stereotyping is higher in the
and work roles—were much higher (never less than .35).
features linked to the family—where, according to our data,
The association established between feminine and masculine
the role assigned to women is still very traditional—than in
traits in 1993 (r = .37) was not significantly different (z =
aspects related to paid work. It seems that this family facet
0.55, p < .29) from that found in 2001 (r = .35). The
is the most acute role stereotyping phenomenon, and it is
correlation between role stereotyping—family and work—
also the most immovable. However, the remarkable decrease
in 2001 (r = .50) was significantly higher (z = 3.28, p <
in role stereotyping between 1993 and 2001 is noteworthy,
.001) than the one obtained in 1993 (r = .39).
especially with regard to work roles.
Table 4
Correlations among Trait and Role Stereotyping in 1993 and 2001 (in Boldface)
1993
Masculine traits
Feminine traits
Total traits
Family role
Work role
Total roles
Masculine traits —
.37**
.89**
.04
.04
.05
Feminine traits
.35** —
.75**
.01
.03
.01
Total traits
.83**
.81** —
.03
.05
.04
Family role
.09**
.05
.08** —
.39**
.94**
Work role
.10**
.08**
.12**
.50** —
.67**
Total roles
.10**
.06*
.10**
.95**
.74** —
2001
* p < .05. ** p < .01.

---

616
LÓPEZ-SÁEZ, MORALES, AND LISBONA
The lack of a relation between trait stereotyping and role
and negative aspects in the content of both of them. The
stereotyping found in the 1993 sample (r = .04) was significant
stereotype of women, tender and oriented toward
in 2001 (r = .10), although the magnitude was very low, and
interpersonal relations, is very positive, but high-status groups
the difference between these two moments in time was not
are not associated with such communal characteristics;
statistically significant. Correlations between feminine and
however, they are associated with the instrumental
masculine trait stereotypes were found, which shows that
characteristics that are typical of masculine stereotypes,
people who stereotype instrumental traitstraditionally, assigning
whereas as low-status groups are associated with communal
more of these traits to men, also do so with the expressive
characteristics, but not with instrumental ones, as occurs
traits, assigning more of them to women. The correlations
with women. Thus, the assignation of expressive traits to
between the two dimensions of role stereotyping were also
women and instrumental traits to men would provide a
high in both samples. These results coincide with those
psychological explanation to the distribution of roles and
obtained in various samples of university students by Spence
would contribute to maintaining them. This justification of
and Buckner (2000), whose correlations between instrumental
roles contributes to legitimizing the inequalities of the system
and expressive stereotypes ranged between .26 and .43.
(Burgess & Borgida, 1999; Jost & Kay, 2005).
However, the correlations found by these same authors between
Our study was planned with the purposes of describing
stereotypes and attitudes were significant only in the case of
stereotyping in the Spanish population, assessing the possible
masculine trait stereotyping and only in men. According to
changes that have occurred in an eight-year interval, and
their conclusions, this kind of results lends strength to the idea
determining the validity of the instruments employed.
of independence between the different components of gender
Therefore, from our data, we cannot reach any conclusions
stereotypes and supports the multidimensional models in the
that allow us to determine why we found no relation between
analysis of gender (see Spence, 1993).
trait stereotypes and role stereotypes. Future research should
Our results point in the same direction as those obtained
study this aspect in depth to seek possible explanations.
by Deaux and Lewis (1984) and other investigators (see
Burgess & Borgida, 1999) concerning the independence of
the diverse components of stereotypes (physical traits,
References
personality traits, roles, or occupations), and they seem to
point to processes that are related to stereotyping, but
Allen, B.P. (1996). African Americans’ and European Americans’
different. In this sense, our data do not support the
mutual attributions: Adjective generation technique (AGT)
theoretical proposal of Eagly and collaborators, because
stereotyping. Journal of Applied Social Psychology, 26, 884-912.
their postulates indicate that such a relation should exist
Alter, R.J., & Seta, C.E. (2005). Compensation for inconsistencies:
(Diekman & Eagly, 2000; Eagly et al., 2000).
The effects of stereotype strength on expectations of applicants’
For Williams and Best (1990), trait stereotyping justifies
job success and satisfaction. Sex Roles, 53, 79-87.
and explains role stereotyping and, therefore, beliefs about
Bem, S.L. (1974). The measurement of psychological androgyny.
personality characteristics should be more resistant to change.
Journal of Consulting and Clinical Psychology, 42, 155-162.
This would explain why there is no relation between the
Bem, S.L. (1981). Gender schema theory: A cognitive account of
new roles performed by men and women and the traits
sex typing. Psychological Review, 88, 354-364.
attributed to them, as the results of our study seem to
Betancor, V., Rodríguez, A., Rodríguez, R. Leyens, J.P., & Quiles,
corroborate. Various authors have argued in favor of the
M.N. (2005). El efecto del estatus en la atribución de las
role of the differentiation of gender to maintain the status
dimensiones estereotípicas de sociabilidad y competencia.
quo. The model of the content of stereotypes (Glick & Fiske,
Psicothema, 17, 297-302.
1999) considers that sociability and competence are central
Brewer, M. B., & Lui, L. N. (1989). The primacy of age and sex in
dimensions in the perception of others. The content of these
the structure of person categories. Social Cognition, 7, 262-274.
dimensions is very similar to the central dimensions of
Burgess, D., & Borgida, E. (1999). Who women are, who women
gender stereotyping: expressiveness and instrumentality.
should be: Descriptive and prescriptive gender stereotypes and sex
According to these authors, men and high-status groups are
discrimination. Psychology, Public Policy, and Law, 5, 665-692.
perceived as being competent and instrumental, but as being
Carpenter, S., & Trentham, S (1998). Subtypes of women and men:
low in the dimensions of sociability and expressiveness. The
A new taxonomy and an exploratory categorical analysis.
opposite occurs with women and low-status groups, which
Journal of Social Behavior and Personality, 13, 679-696.
are perceived as high in sociability and low in competence.
Carpenter, S., & Trentham, S. (2001). Should we take “gender”
In fact, low-status groups, for instance, women, are
out of gender subtypes? The effects of gender, evaluative
considered by the members of their own group as being
valence and context on the organization of person subtypes.
high in sociability and low in competence, in comparison
Sex Roles, 45, 455-480.
to high-status groups (Betancor, Rodríguez, Rodríguez,
Deaux, K., & Lewis, L.L. (1984). Structure of gender stereotypes:
Leyens & Quiles, 2005). Moreover, the traits attributed to
Interrelationship among components and gender labels. Journal
men and women are complementary and there are positive
of Personality and Social Psychology, 46, 991-1004.

---

EVOLUTION OF GENDER STEREOTYPES IN SPAIN
617
Deaux, K., Winton, W., Crowley, M., & Lewis, LL. (1985). Level
Markus, H., Crane, M. Bernstein, S., & Siladi, M. (1982). Self-schemas
of categorization and content of gender stereotypes. Social
and gender. Journal of Personality and Social Psychology, 42, 38-50.
Cognition, 3, 145-167.
Martin, C.L. (1987). A ratio measure of sex stereotyping. Journal
Diekman, A.B., & Eagly, A.H. (2000). Stereotypes as dynamic
of Personality and Social Psychology, 52, 489-499.
constructs: Women and men of the past, present and future.
McCauley, C., Stitt, C.L., & Segal, M. (1980). Stereotyping: From
Personality and Social Psychology Bulletin, 26, 1171-1188.
prejudice to prediction. Psychological Bulletin, 87, 195-208.
Eagly A.H., & Karau, S. J. (2002). Role congruity theory of prejudice
Morales J.F., & López-Sáez, M. (1993). Bases para la construcción
toward female leaders. Psychological Review, 109, 573-598.
de un sistema de indicadores sociales de estereotipia de género.
Eagly, A. H., & Wood, W. (1991). Explaining sex differences in
Psicothema, 5, 123-132.
social behavior: A meta-analytic perspective. Personality and
Morales, J.F., & López-Sáez, M. (1994). Estereotipos de género y
Social Psychology Bulletin, 17, 306-315.
valores. In M. Alvaro (Ed.), Propuesta de un sistema de
Eagly, A.H., & Wood, W. (1999). The origins of sex differences
indicadores sociales de igualdad entre géneros (pp. 375-400).
in human behavior. Evolved dispositions versus social roles.
Madrid: Instituto de la Mujer.
American Psychologist, 54, 408-423.
Moya, M.C., Navas, M.S., & Gómez, C. (1991, September). Escala
Eagly, A.H., Wood, W., & Diekman, A.B. (2000). Social Role
sobre la Ideología del Rol Sexual. Santiago de Compostela,
theory of sex differences and similarities: A current appraisal.
Spain: Libro de comunicaciones del III Congreso Nacional de
In T. Eckes & H.M. Trautner (Eds.), The developmental social
Psicología Social, 1, 554-566.
psychology of gender (pp. 123-174). Mahwah, NJ: Erlbaum.
Powell, G.N., Butterfield, D.A., & Parent, J.D. (2002). Gender and
Echebarria, A., & González, J.L. (1999). The impact of context on
managerial stereotypes: Have the times changed? Journal of
gender social identities. European Journal of Social Psychology,
Management, 28, 177-193.
29, 287-304.
Sczesny, S., Bosak, J., Neff, D., & Schyns, B. (2004). Gender
Glick, P., & Fiske, S.T. (1999). Sexisms and other “isms”:
stereotypes and the attribution of leadership traits: A cross-
Interdependence, status and the ambivalent content of
cultural comparison. Sex Roles, 51, 631-645.
stereotypes. In W.B. Swann, J.H. Langlois, & L.A. Gilebert
Selwyn, N. (2007). E-learning or she-learning? Exploring students’
(Eds.), Sexism and stereotypes in modern society (pp. 193-221)
gendered perceptions of education technology. British Journal
Washington, DC: American Psychological Association.
of Educational Technology, 38, 744-746.
Glick, P., Fiske, S.T., Mladinic, A., Saiz, J., Abrams, D., Masser,
Six, B., & Eckes, T. (1991). A closer look at the complex structure
B., et al. (2000). Beyond prejudice as simple antipathy: Hostile
of gender stereotypes. Sex Roles, 2, 457-471.
and benevolent sexism across cultures. Journal of Personality
Spence, J.T. (1993). Gender-related traits and gender ideology:
and Social Psychology. 79, 763-775.
Evidence for a multifactorial theory. Journal of Personality
Glick, P., Lameiras, M., & Rodríguez-Castro, Y. (2002). Education
and Social Psychology, 64, 624-635.
and Catholic religiosity as predictors of hostile and benevolent
Spence, J.T., & Buckner, C.E. (2000). Instrumental and expressive
sexisms toward women and men. Sex Roles, 47, 433-441.
traits, trait stereotypes, and sexist attitudes. Psychology of
Glick, P., Larsen, S., Johnson, C., & Branstiter, H. (2005).
Women Quarterly, 24, 44-62.
Evaluations of sexy women in low- and high-status jobs.
Spence, J.T., & Helmreich, R.L. (1978). Masculinity and femininity:
Psychology of Women Quarterly, 29, 389-395.
Their psychological dimensions, correlates and antecedents.
Instituto de la Mujer – Instituto Nacional de Estadística. Mujeres
Austin, TX: University of Texas Press.
y hombres en España 2006. Retrieved in 2007 from
Stangor, C., Lynch, L., Duan, C., & Glas, B. (1992). Categorization
http://www.mtas.es/mujer/mujeres/estud_inves/index.htm
of individuals on the basis of multiple social features. Journal
Jost, J.T., & Kay, A.C. (2005). Exposure to benevolent sexism and
of Personality and Social Psychology, 62, 207-218.
complementary gender stereotypes: Consequences for specific
White, M. J., & White, G. B. (2006). Implicit and explicit
and diffuse forms of system justification. Journal of Personality
occupational gender stereotypes. Sex Roles, 55, 259-266.
and Social Psychology, 88, 498-509.
Williams, J.E., & Best, D.L. (1990). Measuring sex stereotyping:
Killen, L.A., López-Zafra, E, & Eagly, A.H. (2006). Envisioning
A multination study. Newbury Park, CA: Sage.
oneself as a leader: Comparisons of women and men in Spain
Wood, W., & Eagly, A. H. (2002). A cross-cultural analysis of the
and the United States. Psychology of Women Quarterly, 30,
behavior of women and men: Implications for the origins of
313-322.
sex differences. Psychological Bulletin, 128, 699-727.
López-Sáez, M., Lisbona, A., & Sáinz, M. (2004). Mujeres
Ximénez, M.C., & García, A.G. (2005). Comparación de los
ingenieras: percepciones sobre su vida profesional. Revista de
métodos de estimación de máxima verosimilitud y mínimos
Psicología General y Aplicada, 57, 161-180.
cuadrados no ponderados en el análisis factorial confirmatorio
López-Sáez, M., & Morales, J.F. (1995). Gender stereotypes in the
mediante simulación Monte Carlo. Psichotema, 17, 528-535.
Spanish population: Looking toward the future. In L. Amâncio
& C. Nogueira (Eds.), Gender, management and science (pp.
Received April 26, 2007
151-168). Braga, Portugal: Instituto de Educaçâo e Psicologia,
Revision received January 8, 2008
Universidade do Minho.
Accepted January 19, 2008

---