Language at Work: Children's Gendered Interpretations of Occupational Titles

by Lynn S. Liben, Rebecca S. Bigler, Holleen R. Krogh
Language at Work: Children's Gendered Interpretations of Occupational Titles
Lynn S. Liben, Rebecca S. Bigler, Holleen R. Krogh
Child Development
Start Page: 
End Page: 
Select license: 
Select License

Child Development, May/Iune 2002, Volume 73, Number 3, Pages 810-828
Language at Work:
Children's Gendered Interpretations of Occupational Titles
Lynn S. Liben, Rebecca S. Bigler, and Holleen R. Krogh
A large literature has shown that children's beliefs and aspirations about occupations reflect cultural gender
stereotypes. One channel that may create or sustain occupational stereotypes is language. Two studies were
designed to examine whether children interpret occupational titles as gender specific or gender neutral. In
Study I, children (6- to Ll-year-olds, N = 64) were asked directly if various job titles could be used for both
men and women doing the job. In Study 2, children (6- to lO-year-olds, N = 51) were shown pictures of men
and women engaged in job activities and asked which one(s) showed someone who could be called a(n)_.
Titles were linguistically unmarked for gender (e.g., doctor), strongly marked (e.g., policeman), or weakly
marked (e.g., postmaster). Marked titles were given in masculine and feminine forms. Findings reinforced past
work showing that marked titles are exclusionary, revealed that some children harbor confusions about even
unmarked titles, and demonstrated the mediating role of individual differences in attitudes. Implications for
the changing lexicon and for educational programs are discussed.
As captured by Freud's insistence on the dual importance
of lieben und arbeiten (love and work, see Erikson,
1968, p. 136), the world of work is central to human
life. An important research goal is thus to
understand how children's beliefs about and interests
in occupations develop. These beliefs are influenced
by many factors, including (among others) the way
that various kinds of workers are depicted in the media,
the occupations pursued by their parents or other significant
adults, and the experiences encountered in
career-education programs. In turn, the ways that
these factors operate are strongly affected by cultural
views of the relation between gender and work.
Despite the fact that increasing numbers of men
and women have been entering jobs historically associated
with the other sex, many occupations in American
culture remain strongly "gendered" (e.g., Phillips
& Imhoff, 1997; Williams, 1993). There is considerable
research to show that even young children associate
different occupations with men and women, and aspire
to occupations that are consistent with the cultural
stereotypes for their own sex (e.g., Archer, 1984; Bigler
& Liben, 1990; Carter & Levy, 1988; Helwig, 1998;
Liben, Bigler, & Krogh, 2001; O'Keefe & Hyde, 1983;
Serbin, Powlishta, & Gulko, 1993; Signorella, 1987,
1999; Signorella, Bigler, & Liben, 1993; Stangor &
Ruble, 1989;Stockard & McGee, 1990;Tremaine, Schau,
& Busch, 1982).
A wealth of research has identified experiences
that are relevant to establishing and maintaining
these gendered beliefs (for a recent review, see Ruble
& Martin, 1998). First, because many occupations are
still, in fact, predominated by men or women, children
remain more likely to encounter someone of the
traditional sex in certain jobs (e.g., more often seeing
male police officers and female nurses than the reverse).
Second, and again despite some recent change,
a sex imbalance remains in the occupational portrayals
of men and women in books, film, and television (e.g.,
Calvert & Huston, 1987; Coltrane & Adams, 1997;
Gettys & Cann, 1981; Purcell & Stewart, 1990; Signorielli
& Bacue, 1999; Tognoli, Pullen, & Lieber, 1994).
Third, many adults and peers continue to make explicit
statements about what constitutes men's or
women's work. Taken together, these and related factors
lead children to develop gender schemata (e.g., Bem,
1974; Martin & Halverson, 1981) about what jobs are
and are not appropriate for each sex. Once schemata
are in place, they filter incoming information, making
it difficult for subsequent experiences to modify stereotypic
beliefs (e.g., Bigler & Liben, 1990; Cordua,
McGraw, & Drabman, 1979; Liben & Signorella, 1980,
1993; Martin & Halverson, 1983).
The work described in this article addressed another
potentially important factor in creating and sustaining
children's beliefs about the gendered nature
of occupations-language. Language may be a particularly
important channel because it surrounds children
from birth and because it is pervasive in family,
school, and media contexts.
There are already demonstrations that the language
used to refer to occupations has a powerful im-
© 2002 by the Society for Research in Child Development, Inc.
All rights reserved. 0009-3920/2002/7303-0010
pact on people's beliefs about and interests in jobs. A
classic article by Bem and Bem (1973) showed that
high school seniors' interests in jobs were strongly affected
by the specific form of job titles used in helpwanted
advertisements. The effect was particularly
strong for females: when jobs were labeled with a
masculine suffix (e.g., "lineman"), only 5% expressed
an interest in the job; when the job title was renamed
with a feminine suffix ("linewoman"), 45% expressed
interest. For males, the parallel contrast was 30% versus
65%. In more recent work, McConnell and Fazio
(1996) have shown that job titles also affect the inferences
people make about the characteristics of the
people who hold the named jobs.
Findings from studies like these, as well as political
efforts associated with the feminist movement of the
last quarter of the twentieth century (e.g., Brownmiller,
1999), have led to prescriptive changes in the
English language as used in American culture. With
few exceptions, it is now illegal to advertise for positions
by sex; the Publication Manual of the American
Psychological Association (American Psychological Association,
APA, 1994) asks authors to avoid using
gender-marked occupational titles; and the Dictionary
of Occupational Titles, published by the federal government
(U.S. Employment Service, 1998), uses only
gender-free titles. Although the assumption is generally
made that gender-free titles will reduce children's
gender stereotypes, the inverse is possible.
That is, as noted by Liben and Signorella (1993,p. 148),
gender-free titles used during this transitional period
in society may "actually impede progress toward
gender equity by ignoring the nontraditional nature
of the stimulus or situations." Furthermore, it is not
known whether children understand that genderunmarked
titles are indeed applicable to both men
and women, nor is it known under what circumstances
children believe that marked titles limit the range of
application. Several features of the surrounding linguistic
environment led us to predict that children
would find the range of application puzzling.
First, despite the prescriptive changes described
above, many people continue to use only traditional
masculine forms of job titles, perhaps reflecting their
ignorance of recommended changes, intentional rejection
of these changes, or habit. Furthermore, traditional
materials remain widely available. For example,
many teachers post traditional images of "community
helpers" on their bulletin boards and label them with
traditional language; and parents, teachers, and librarians
continue to promote classic children's books, such
as Curious George (Rey, 1941/1998), that are replete
with such language.
Second, even individuals who use contemporary
Liben, Bigler, and Krogh 811
language may offer varied and inconsistent models.
Some speakers use masculine and feminine versions
of traditional titles to discriminate between men and
women (e.g., using "fireman" for men and "firewoman"
for women), others use traditional titles for
men but gender-neutral titles for women (e.g., "fireman"
for men and "firefighter" for women), and still
others use gender-neutral titles for all workers (e.g.,
"firefighter" for both).
Third, there is fast-paced change in English itself.
For example, although some gender-specific titles
have long been available (e.g., "waiter" and "waitress"
or "actor" and "actress"), others have entered
the lexicon only recently (e.g., "congressman" and
Fourth, one relatively common type of occupational
title is ambiguous with respect to genderthose
ending with the suffix "master." These titles
may be understood as gender specific if the suffix is
taken to mean a male youth (as in "Master John"), but
gender neutral if the suffix is taken to mean skill (as in
"mastery"). In light of these alternatives, it is perhaps
not surprising that there is considerable and often
heated debate about the necessity or acceptability of
creating and using feminine versions of "master" titles
(e.g., concertmistress, postmistress, or webmistress;
see Safire, 1999). Furthermore, even when the suffix
"mistress" is substituted for the suffix "master," the denoted
jobs mayor may not be understood as the same.
A quotation posted at an exhibit of the National Museum
of Woman in the Arts in Washington, D.C.
makes the point well: "The unspoken assumption in
our language [is] that art is created by men. The reverential
term 'Old Master' has no meaningful equivalent.
When cast in its feminine form, 'Old Mistress,'
the connotation is altogether different" (Gabhart &
Broun, 1972, p. 6). Dictionary definitions are of little
help in clarifying whether titles ending with "master"
are intended to carry gender-specific meaning. For
example, in the RandomHouseDictionary of the English
Language (Flexner, 1987), schoolmaster is defined as "a
man who presides over or teaches in a school" (p. 1278),
whereas ringmaster is defined as "one in charge of the
performances in the ring of a circus" (p. 1235). This
same dictionary provides gender-neutral definitions
for both postmaster and ringmaster, but includes the
feminine counterpart postmistress but not ringmistress.
Further evidence of inconsistency derives from
comparing definitions of the same word across dictionaries,
from examining earlier and later editions of
the same dictionary, and from consulting spell-check
programs of word processing packages.
Faced with this linguistic context, children may
well be confused about the relation between occupa812
Child Development
tional titles and implied sex of workers. At one extreme
is the possibility that children have come to believe
that all occupational titles carry meaning about
the sex of the person performing that occupation. At
the other extreme is the possibility that children have
come to believe that occupational titles are never informative
about the worker's sex. Between these extremes
lies the possibility that children believe that
some, but not all, titles specify the worker's sex, perhaps
varying systematically with the precise form of
the job title, qualities of the job, and / or characteristics
of the work setting.
The present research was designed first, to explore
whether children interpret gender-neutral occupational
titles as applicable to both males and females,
and second, to investigate what children believe to be
the range of application of occupational titles that do
contain some form of gender marking. In Study I,
children were asked directly about whether particular
job titles could be used for someone doing that job,
irrespective of whether the worker was a man or a
woman. In Study 2, children were shown pictures of
men and women performing job activities, and were
asked whether a particular job title could be used for
the depicted person, and if not, what job title would
be appropriate. As explained in more detail in the introduction
to Study 2, hypotheses concerning the conditions
that might be especially conducive to confusions
about the gender-specific nature of job titles
were also explored. In both studies, we tested children
ranging in age from about 6 to about 11 years.
This age was selected because it covers a period in
which there is considerable knowledge about the cultural
stereotyping of occupations, but variability in
the degree to which individual children endorse
those cultural stereotypes (Signorella, 1987).
The major purpose of the first study was to collect
data on children's understanding of the link between
occupational titles and the sex of the worker by asking
children directly whether they thought that certain
job titles could be applied to both men and
women doing the named job. These questions were
asked about two major categories of job titles. The
first included occupational titles that traditionally
have had no linguistic gender marking of any kind,
referred to in this article as unmarked. These are titles
that were already fully consistent with nonsexist
writing requirements of the APA Publication Manual,
and needed no modification to be appropriate for the
contemporary Dictionary of Occupational Titles. Note
that even though the job titles in this category are not
marked linguistically for gender, some of the jobs to
which these titles refer are gender stereotyped by
American culture. Indeed, we intentionally included
jobs that are culturally stereotyped as masculine, feminine,
and neutral among our unmarked items (e.g.,
"doctor," "librarian," and"artist," respectively). With
respect to unmarked titles, our goal was to determine
whether children would understand that these job titles
are applicable to both men and women, an understanding
that is implicitly assumed in the way that
these titles are used in educational, vocational, and
research settings.
The second major category-marked-included
occupational titles that contain some linguistic component
that might be viewed as specifying sex. Included
were two subcategories. The first, strong
marked, included titles with the suffix of man (e.g.,
"policeman" or "fireman"). We categorized these as
strong marked because of the strongly gendered interpretation
of the morpheme "man." The second
subcategory, weak marked, included job titles with
the suffix "master" (e.g., "postmaster"). We categorized
these as weak marked because (as discussed in
detail earlier) "master" may refer to either a young
male or to skill. Also included in the weak-marked
category were titles ending in "-er" or "-or" and having
feminine counterpart titles in common use. We
categorized these titles as weak marked because the
availability of a common feminine counterpart title
may endow the "er / or" title with a masculine interpretation,
while at the same time the"er / or" ending
alone does not unambiguously specify sex (e.g., compare
"waiter / waitress" to "teacher").
Overall, we expected that children would be more
likely to assign job titles to both men and women
(hereafter referred to as "both" responses) for unmarked
than for marked titles. Further, within the category
of marked titles, we hypothesized that children
would give more "both" responses for weak-marked
titles than for strong-marked titles. In addition, we
predicted that children would give more "both" responses
to masculine versions of marked titles than to
feminine versions of marked titles created by replacing
"man" by "woman," "master" by "mistress," and
"er / or" by "ress." (We had initially planned to examine
an additional contrast between titles in which the
canonical form of the occupational title was masculine
versus feminine. The English language and
American culture offered too few of the latter occupational
titles to enable us to do so, however, and thus
all data and examples are based on items in which the
canonical form is masculine.)
Finally, this investigation was designed to test
whether children's responses would differ in relation
to their own gender-stereotyped attitudes. Based on
prior work that has shown the impact of individual
differences in the strength of gender-stereotyped attitudes
(e.g., Signorella, 1987), we predicted that children
with less stereotyped attitudes would give more
"both" responses.
Children were recruited by sending letters to parents
in after-school and summer programs in midwestern
and southwestern cities. The sample included
children whose parents had returned signed
permission slips and who themselves had agreed to
participate. The final sample consisted of 34boys and 30
girls, divided into younger (6- to 8-year-olds, M == 7.3
years, SD == .8 years; n == 35) and older (9- to 11-yearolds,
M == 9.5 years, SD == .6 years; n == 29). Although
formal information on race and ethnicity was not collected,
the programs drew almost exclusively from
White, middle-class families, thus requiting caution
with respect to generalization to other populations.
Occupational titles were selected by drawing from
previous occupational research, by using computer
dictionaries to search for all words with desired suffixes
(e.g., by entering "*master" into spell-check programs),
and by consulting general-purpose dictionaries
and dictionaries of occupations. The job titles
included occupations that prior research (or new
adult ratings when necessary) had shown to be culturally
viewed (in the United States) as masculine,
feminine, or neutral. In addition, we included six occupational
titles (chandler, ginner, higgler, limner,
milliner, and nose) that earlier work had shown to be
unknown to children of this age (Liben et al., 2001).
These titles were used to provide an opportunity to
test whether children understood the instructions.
Two lists of occupational titles were prepared. Both
lists contained all of the unmarked items. In addition,
List A contained half of the marked items in masculine
versions and the remaining half in feminine versions.
List B contained the gender-reversed versions
of each item. Titles were presented in a single, randomly
determined order with the exception that the
lists intentionally began with three unmarked titles
representing culturally neutral, masculine, and feminine
occupations. The two lists contained one or the
other gender version of each title in the same order.
Thus List A began musician, banker, secretary, ac-
Liben, Bigler, and Krogh 813
tress, businessman whereas List B began musician,
banker, secretary, actor, businesswoman.
Procedure and Instructions
Children were tested individually by a female interviewer.
After establishing rapport, the tester explained
that the study was designed to learn about
the way in which job titles can be used. Instructions
emphasized explicitly that the focus was on the acceptable
use of the word rather than on who usually
does or should perform the job, and on the way they
thought that the words are used in our language,
rather than on how they themselves used the words.
Children were reminded: "Remember, we are not trying
to find out who usually does these jobs, or about
who should do these jobs. Instead, we are trying to
find out whether a man or a woman who is doing
some job can be called by the label or word. It doesn't
matter if you or other people think some other label is
better. The question we are asking is whether it is okay
to use this word for a man or a woman who is doing
the job." Children were warned that the list would
contain some very unusual jobs as well as familiar
ones, and that if they had never heard of the job to
simply tell the interviewer so that the item could be
skipped. Four practice items were then given. The
practice items were selected to provide the full range
of responses in a way that would not bias responding
on the occupations themselves. Children were asked
to try first some words that not everyone thinks of as
regular jobs but that would be easy to categorize:
"mother" (used for women only), "father" (used for
men only), and "parent" (used for both men and
women). The final practice item was "cartographer,"
which had been shown in pilot testing to be unknown
to children (hence allowing practice on the "don't
know" response).
After the tester was satisfied that the children understood
the directions, she read each occupational title
aloud, and asked whether that title could be used for
both men and women doing these jobs, only for men,
or only for women. Throughout testing, children
were reminded that the question was about the use of
the word, and not about who usually does these jobs
or who they thought should do these jobs. Children
were also reminded that if they did not know the job,
they should just say so.
Gender Attitude Measure
Children's levels of gender stereotyping were assessed
by asking whether each of 25 activities (e.g.,
814 Child Development
building a fort, baking cookies) should be done by
boys, by girls, or by both boys and girls. The word
"should" was selected for these questions based on
past work (Signorella et al., 1993) that suggested that
"should" questions are particularly well suited for
tapping children's attitudes (rather than knowledge).
The activities were taken from the short form of the
activity stereotyping subscale of the Children's Occupational,
Activity, and Trait-Attitude Measure
(COAT-AM) described in Bigler, Liben, and Yekel
(1992). Included were 10 masculine, 10 feminine, and
5 neutral activities, presented in random order. The
activities subscale of the COAT was used to avoid any
overlap with occupations that were the central focus
of the current research. Following past procedures
used for these attitudinal measures (Signorella et al.,
1993), the dependent measure was the number of
masculine and feminine items assigned to the "both"
category so that a higher score indicated a lower level
of gender stereotyping.
Children were divided into high- and low-stereotype
groups on the basis of a median split of the number
of "both" responses, median, 11; Cronbach's a, .85.
Younger and older children were distributed across
both high- and low-stereotype groups (with the correlation
between chronological age and COAT score
being only .0036). Thus, this was an excellent sample
for evaluating the independent contributions of both
gender attitudes and chronological age.
As a first step, to ascertain that children had understood
and followed instructions, we began by considering
children's responses to the obscure job titles. For
these items, evidence that instructions had been understood
would be provided by children saying that
they did not know the job. Virtually all of the responses
to obscure items were of this kind, suggesting that children
were able to monitor their own knowledge of job
titles adequately and respond appropriately. Given
this finding, we also dropped from further analyses
all additional items that elicited "don't know" responses
from 75% or more of the children. The final
list of occupational titles used for analyses is given in
the Appendix, divided into the item categories explained
earlier. In addition, any items unknown to a
particular child were also omitted for that child.
Toadjust for the fact that there were therefore some
differences in the numbers of items used for different
children, as well as for the fact that not all categories
contained the identical number of exemplars, the dependent
measure used in all analyses was the percentage
(rather than raw number) of "both" responses.
(Given the use of percentage data in this and
in the remaining studies, we also routinely conducted
analyses after arcsine transformations. Except for one
finding that changed from p == .03 to P == .06, results
were identical, and thus only analyses on untransformed
data are reported.) Responses other than
"both" were not broken down further because when
non-both responses were given, there was virtually no
further variability. For example, non-both responses to
"policeman" or "construction worker" were virtually
always to "men only," whereas non-both responses
to "policewoman" or "nurse" were virtually
always to "women only." Below we describe analyses
and findings for the linguistically unmarked occupational
titles, followed by analyses and findings for
the linguistically marked titles.
Linguistically Unmarked Occupational Titles
The percentage of "both" responses to unmarked
titles served as the dependent measure. These data
were analyzed by a four-way analysis of variance
(ANOVA) of 2 (age) X 2 (participant sex) X 2 (participant
stereotyping) X 3 (cultural stereotype of the job
to which the title referred: neutral, masculine, or feminine).
Of these, the first three were between-subjects
factors and the last was a within-subjects factor. Here
and throughout the remainder of this article, the significance
of differences between means were analyzed
using Newman-Keuls follow-up tests, with effects
reported as significant when p < .05.
Before describing the results of this and additional
analyses, it is important to note that the use ofANOVA
and, particularly the use of dichotomous groups for
age and level of stereotyping would be problematic if
(1) age and level of stereotyping were highly correlated
within the sample, and (2) small cell sizes were
responsible for failing to reject the null hypothesis. To
evaluate the potential problem, multiple hierarchical
regression analyses were conducted in which age and
level of stereotyping were included as predictors of
children's responses on each of the item types (e.g.,
unmarked, marked). Sample size precluded the inclusion
of interaction terms. Results indicated that even
after age was entered into the model, level of stereotyping
was a significant predictor of responding for
virtually every case that was found to be significant in
the ANOVAs reported below (and vice versa for age).
In other words, age could not explain the variance in
responding that was associated with level of stereotyping.
Further, although the regression analyses included
the entire sample, there was little evidence
that the increased power affected the pattern of results.
In only one case (out of 28) did the regression
model reveal a significant effect that was not also evident
in the ANOVA. Given these findings, only the
ANOVA results are reported in detail.
The four-way ANOVA revealed a significant main
effect of participant level of stereotyping, F(l, 56) =
4.51, P = .038, with low-stereotype children giving a
higher percentage of "both" responses than highstereotype
children, Ms (50s) = 77 (20) versus 67 (19),
respectively. Also significant was the main effect of
the cultural stereotype of the job, F(2, 112) = 54.98,
P< .001, with titles for culturally neutral jobs eliciting
a higher percentage of "both" responses than titles of
jobs that are culturally stereotyped as masculine or
feminine, Ms (50s) = 92 (16), 68 (27), and 70 (19), respectively,
the latter two not differing significantly
from one another. None of the other main effects or
interactions were significant.
Linguistically Marked Occupational Titles
The percentage of "both" responses to marked titles
served as the dependent variable in a five-way ANOVA
in which between-subjects factors were again participant
age, sex, and level of stereotyping (low or high);
and the two within-subjects factors were suffix gender
(masculine, feminine) and suffix strength (strong,
Analyses revealed two significant two-way interactions.
The first was the interaction of participant
age and suffix gender, F(l, 37) = 4.55, P = .040. Both
older and younger children gave a higher percentage
of "both" responses to masculine than to feminine job
titles, with the difference significant in the older
group, Ms (5Ds) = 62 (22) versus 30 (29), respectively,
but not in the younger group, Ms (5Ds) = 54 (30) versus
43 (39), respectively.I
The second significant interaction was between
suffix strength and suffix gender, F(l, 37) = 6.81, P =
.013. Children gave a significantly higher percentage
of "both" responses to weak masculine job titles,
1We excluded data from children who had inadequate numbers
of responses on anyone of the item categories (most commonly
the weak suffix feminine items), thus explaining why
there were fewer dfs for analyses of marked than for unmarked
titles. For each of the four item categories taken separately, we
also compared responses of children who were retained versus
excluded from the full analysis. In three of the four cases, there
were no significant differences between the two groups. In the
fourth (weak suffix, masculine items), the participants who had
been excluded from the full analysis had a smaller percentage of
"both" responses. The contrasts across item types were similar,
however, and thus the list wise deletion of data appears to have
had little effect on the pattern of findings reported in the text.
Liben, Bigler, and Krogh 815
M (50) = 85 (26) than to strong masculine, weak feminine,
and strong feminine titles, Ms (50s) = 47 (34),
47 (43), and 30 (44), respectively. The latter three did
not differ significantly from one another.
Two main effects were subsumed by the higher order
interaction just described, including a main effect
for suffix gender, F(l, 37) = 29.75,P< .001,with masculine
titles receiving a higher number of "both" responses
than feminine titles, Ms (5Ds) = 57 (27) versus
38 (36), respectively; and a main effect for suffix
strength, F(l, 37) = 14.72, P< .001, with weak suffixes
eliciting a higher percentage of "both" responses than
strong suffixes, Ms (5Ds) = 63 (29) versus 40 (34),
Finally, a significant main effect was found for participant's
level of stereotyping, F(l, 37) = 4.83, P =
.034, with low-stereotype children giving a higher
percentage of "both" responses than high-stereotype
children, Ms (50s) = 53 (31) and 42 (22), respectively.
At the most general level, the patterns of responses
suggest that children are, indeed, sensitive to the linguistic
forms of job titles, and that these sensitivities
differ in relation to participant variables. First, the
data bear on the question of whether children interpret
what are presumptively gender-neutral (unmarked)
job titles as being universally applicable to
both men and women. Children did not uniformly
believe that these job titles could be applied to people
doing the job without regard to the worker's biological
sex. Over 30% of responses given by the highstereotype
children were denials that the title could
be applied to a worker of the other sex. Although the
high-stereotype children gave significantly more of
these inflexible responses than did the low-stereotype
children, even among the latter group, roughly 25%
of responses were of this kind. The import of these
numbers is even more dramatic once it is recognized
that these percentages include responses to job titles
for culturally neutral occupations that elicited "both"
responses at an extremely high level (92%). Taken together,
the data from the unmarked titles are consistent
with the hypothesis that children do not universally
understand that gender-neutral titles are applicable
to both men and women.
The findings from the marked titles are also consistent
with the initial hypotheses about title variables.
As predicted, the nature of the suffix affected children's
judgments about the range of applicability of
the job title. Titles with weak masculine suffixes (such
as waiter or postmaster) elicited significantly more
"both" responses than any other type of title, and at a
816 Child Development
high rate (85%). These data might well be used to support
the argument that it is not necessary to create
feminine counterpart titles such as webmistress
(Safire, 1999). Alternatively, it could be argued that
even 15% is unacceptable when there are potentially
serious consequences for children's occupational aspirations
and educational decisions.
Data from the marked titles with strong, masculine
suffixes showed far lower acceptance rates. That
is, consistent with earlier research showing that
people infer male referents from masculine pronouns
and "-man" job titles (e.g., Bern & Bern, 1973; Hyde,
1984; Martyna, 1978, 1980), children commonly (although
not universally) denied that strong masculine
titles could be used for women workers. Interestingly,
there were no significant differences among the
acceptance rates for strong masculine, strong feminine,
and weak feminine titles. These data suggest
that the marking of the title (in the linguistic sense of
marking; see Clark, 1973) is accomplished by strong
masculine suffixes and by any feminine marker
The data from the marked titles also showed effects
of two participant variables. First, there was a
significant main effect of participant level of stereotyping.
Similar to the result from the unmarked titles,
low-stereotype children gave significantly more
"both" responses than did high-stereotype children.
These data support the hypothesis that children
with less stereotyped attitudes are more likely to accept
the egalitarian use of words. Second, there was
a significant effect of age in interaction with suffix
gender. Specifically, although both younger and
older children were more likely to accept titles with
masculine than feminine suffixes as applicable to
both men and women, the difference was significant
only in the older participants. This finding suggests
that with age, children become increasingly sensitive
to the differential power of masculine and feminine
In summary, the overall response patterns suggest
that, as a group, children are sensitive to the genderspecific
intent of some job titles. However, at the same
time that the data suggest that children give appropriate
interpretations of many gender-marked versions
of job titles, the data also suggest that children
harbor at least some important confusions. It is particularly
noteworthy that the analyses of responses to
the unmarked titles demonstrate that some children
do not understand that gender-neutral occupational
titles are universally applicable to both men and
women. Study 2 was designed to examine whether
similar conclusions would be reached with a different
The findings from Study 1 are consistent with the hypothesis
that children have at least some mistaken beliefs
about the gender-specific meaning of unmarked
titles and generate further questions about what conditions
may be conducive to developing these beliefs.
For Study 2, we developed a new methodology that
could be used to collect converging data relevant to
Study 1 while exploring some additional issues. First,
pictorial stimuli for culturally masculine, feminine, and
neutral jobs were constructed, and parallel pictures with
male and female workers were prepared. Weasked children
whether certain occupational titles could be used
for the depicted workers, and if the offered title was rejected,
what the depicted worker would be called. We
also asked children to explain the job duties of the depicted
workers. As in Study 1, both unmarked and
marked job titles were included, with marked titles
varying by suffix gender and suffix strength.
In addition to examining the influence of titles
themselves, we also designed the study to test a hypothesis
about the circumstances under which children
would be likely to develop erroneous beliefs
about the gender specificity of job titles. Specifically,
we hypothesized that children would be especially
likely to think that two different, unmarked occupational
titles are meant to differentiate male from female
workers (rather than to differentiate between
two different occupations) when two conditions are
met. The first is that the two jobs share many features
(e.g., are conducted in a similar setting, have overlapping
job duties, and have similar functions at some
higher level of analysis). For convenience, this factor
is hereafter referred to as setting, with pairs that come
from a shared setting contrasted with pairs that come
from a distinct setting. The second condition is that
each of the two jobs carries a different cultural gender
stereotype. For convenience, this factor is hereafter referred
to as cultural stereotype of the pair, with those
pairs in which each job comes from a different stereotype
(i.e., one masculine, one feminine) contrasted
with those pairs in which each job is from the same
stereotype (i.e., both masculine or both feminine).
An example of a shared-setting, different-stereotype
pair is doctor / nurse. People in these jobs work in the
same locations (e.g., a hospital), perform some of
the same actions (e.g., take medical histories), and in
some sense perform the same function (provide health
care). The two jobs are differentially stereotyped insofar
as American culture views doctor as masculine and
nurse as feminine. An example of a shared-setting,
same-stereotype pair is train conductor / train engineer.
People in these jobs likewise share work settings,
actions, and higher level functions (both work
near locomotives, inspect train parts, and keep the
trains running), but both jobs are culturally stereotyped
as masculine. We hypothesized that children
would be more likely to attribute gender-specific
meaning to unmarked job titles under the former conditions.
This hypothesis was based on the reasoning
that when job differences are perceptually subtle
while biological sex is perceptually salient, the child
would be likely to infer, erroneously, that different job
titles must refer to the sex of the worker. To test this
hypothesis, we asked children whether job titles
could be applied to workers pictorially depicted in
jobs that did or did not have shared-setting counterparts,
and that did or did not cross gender stereotypes
of the culture.
As in Study I, jobs with gender-marked as well as
unmarked titles were included. With respect to
marked job titles, we hypothesized that children would
be more willing to extend masculine versions of
marked titles to women than the reverse (as in Study
1); that this difference would be more pronounced for
weak-marked than strong-marked titles (again, as in
Study 1); and that when children rejected either masculine
or feminine forms, they would typically replace
them with the gender-reversed job title (e.g., if
the title "policewoman" were rejected for a man pictured
performing police like duties, the child would
offer "policeman" as a preferred job title).
With respect to job duties, we expected more erroneous
beliefs that job duties differ for men and
women depicted in the same occupation when there
are not common masculine and feminine versions of
job titles. For example, we expected that children
would be likely to believe that men and women depicted
as doctors have different job duties than to believe
that men and women depicted as salespeople
have different job duties. In the latter case, common
gender-linked job titles with the same root exist
("salesman" and "saleswoman"). When such masculine
and feminine forms of job titles are available,
there is no reason for the child to impose biological
meaning on the job activities. The shared occupational
root (here "sales") should also suggest the similarity
of the jobs. When job duties were said to differ,
we expected that the higher status duties would be attributed
to men.
Here we overview the method used in Study 2. We
showed 6- to 10-year-old children pictorial representations
of men and women engaged in various jobrelated
activities, and asked whether each depicted
person could be called by a particular occupational
title. In instances in which the child rejected a title for
a particular picture, we asked what the depicted per-
Liben, Bigler, and Krogh 817
son would be called. Given that it is logically possible
that people with two different job titles might be
thought to perform the same or different duties and,
similarly, that people with the same job title might be
thought to perform the same or different duties, we
also asked children what they thought the people depicted
in the pictures did in their occupations. By
counterbalancing whether a man or woman appeared
in the identical scenes, we could be certain that any
differences in responses about job titles or job duties
were attributable to the sex of the depicted person,
rather than to other details about the pictures. By
drawing from occupations that varied with respect to
whether they have shared-setting counterparts, and
whether they represent both masculine and feminine
cultural gender stereotypes, we could test our hypothesis
about the conditions conducive to developing
erroneous beliefs that unmarked job titles are gender
specific. By using different job titles, we could
examine the role of the linguistic forms of titles, thus
providing converging evidence related to the results
from Study 1.
Participants were 6- to 10-year-old children who
were attending summer school and after-school programs
in midwestern and southwestern cities whose
parents had returned signed permission slips and
who themselves had agreed to participate. As in
Study I, the programs draw largely from White, middleclass
families, thus requiring caution with respect to
the scope of generalization. Children were tested individually
by a female research assistant who tape recorded
sessions for later transcription. Of the 62 children
interviewed, recording difficulties or inaudible
tapes reduced the final sample to 51. Included were
26 boys and 25 girls, divided into younger (M == 6.6
years, SD == .6; n == 28) and older (M == 8.9 years, SD ==
.7; n == 23) groups. Although individual differences in
gender attitudes would have been of interest, constraints
on the time we were permitted to interview
children prevented the administration of an individual
stereotyping endorsement measure.
Overview. The interviewer began by explaining to
children that she was going to be asking some questions
about what pictures shown on cards might be
called. In each case, there were two pictures per card.
The initial pictures and instructions were designed to
818 Child Development
introduce children to the format of the task, and to ensure
that children understood that on anyone card
there could be two pictures of things that have (1) different
labels and different functions, (2) the same label
and the same function, or (3) the same label but a different
function. Instructions also presented the possibility
that a different card might contain a picture that shared
a label and / or function with a picture on another card.
Once this variety was understood in relation to objects,
it was extended to people for the data collection proper.
Practice. Children were first shown a set of three
oak-tag cards as illustrated in Figure 1. On each card,
two black and white pictures had been mounted. One
card contained a picture of a horn (trumpet) and a picture
of a pencil. The second card contained a picture
of a horn (deer horn) and a picture of a pen. The third
card contained two pictures of drums. Children were
asked "Can you find a card that has a picture of something
that you could call a drum?" Once children
picked a card and pointed to one of the pictures, the
experimenter held up that card and then asked: "Does
the other picture on this card show something that you
could call a drum?" After children responded affirmatively,
the experimenter asked whether they both did
the same thing or different things, and in response to
children's positive response, the tester summarized by
saying "So sometimes you may find a card that has
two pictures that you could call the same thing and
that do the same thing."
The first card was put down, and children were
asked to try another one: "Can you find a card that
has a picture of something you could call a horn?"
Once children selected a card and picture, the tester
held up the card and asked "Does the other picture
Figure 1 Stimulus cards used as part of the task instructions
in Study 2.
on this card also show something you could call a
horn?" Now the correct response was "No" becausedepending
on which card children had selected-the
other picture showed either a pen or a pencil. The tester
then asked: "Is there another card that shows something
you could call a horn?" Here the correct answer
was "Yes" because another card had a picture of another
kind of horn. The tester then cycled through the
question about whether the horns did the same thing,
and summarized by saying "So sometimes you may
find pictures that you could call the same thing. They
might be on the same card or on a different card. Sometimes
they might do the same thing-but sometimes
they can be called the same thing but do different
things." In the third example, children were asked to
find a card that had a picture of a pencil. Another card
had a picture of a pen, and was used to illustrate that
two pictures could be called something different (i.e.,
pen and pencil) but could nevertheless do the same
thing (i.e.,both write, although only one writes in ink).
Test procedure. The next phase began after the interviewer
was satisfied that children understood that pictures
could show things that could be called by the
same or different words and might have the same or
different functions (and that these might be on the
same or different cards). The interviewer told children
that the next cards would show pictures of people
doing various jobs and that we were interested in
learning (1) whether they thought people in the pictures
could be called by the same or different names,
and (2) whether they thought people in the jobs did the
same or different things. A procedure similar to that
used during practice was followed using 18 three-card
sets. Note that because each of the three cards contained
two pictures, every set displayed six pictures at
once. A sample three-card set is shown in Figure 2.
For each of the 18 three-card sets, children were
asked if they could "find a card that has a picture of
someone that you could call a(n) __" and then
asked to point to the picture on the card. The card was
held up, and children were asked if the other picture
on that same card also showed "someone you could
call a(n) __." If children responded yes, then they
were asked "Do they both do the same thing?"; if so,
"What do they do?," and if not, "Can you tell me
what is different about what they do?" If children rejected
the second picture on the card, they were asked
"What would you call this person?" and then asked
"Do they both do the same thing even though you
couldn't call them the same thing?" As before, followup
questions were asked to elicit what was the same
or different about what the depicted people did. If
children could not find a single exemplar of the
named occupation, the experimenter selected the apFigure
2 A sample three-card set of pictures used for Study 2.
Cards contain pictures designed to depict "artist," "doctor,"
and "nurse." Illustrated are cards for Set A. Those for Set B
showed identical scenes, but with male and female workers
reversed. Sets A and B were counterbalanced across children.
propriate card, pointed to one picture, and asked
"What would you call this one?" Following this entire
procedure, children were asked if there was"another
card with someone who could be called a(n) __?"
and the entire set of questions was repeated until
other cards were rejected, after which a new threecard
set was taken out. The 18 three-card sets were
given in a single, randomly determined order, and the
two pictures were glued on each card to reduce the
possibility of experimenter error. The spatial placements
of the cards on the desk were varied randomly
across participants.
Test materials. As explained above, each card contained
two pictures. One picture showed a man and
one showed a woman. Both people were shown engaged
in job activities that were drawn to depict the
identical occupation. The details of the activities and
backgrounds differed only slightly. For example, and
as illustrated in Figure 2, the two pictures on the"doctor"
card both showed doctors performing surgery, but
each was positioned differently in relation to the patient.
The two pictures on the "nurse" card both
showed nurses interacting with a patient, but one was
shown checking a patient's temperature and the other
was shown giving an injection. To ensure that these
slight activity and pictorial differences did not carry
unintended distinctions (e.g., in the status of the depicted
job activities), two versions of every picture pair
were prepared. Thus, half the children saw a card set in
which the man was depicted in a particular pose with
a particular background and the woman was depicted
in a different pose and background, and the other chil-
Liben, Bigler, and Krogh 819
dren saw the reversed versions. Card sets were not expected
to affect response patterns. For unmarked titles,
all children were asked about the identical titles. For
marked titles, half the children were asked about masculine
versions of the titles, whereas the remaining
children were asked about feminine versions of the titles.
This factor was manipulated between rather than
within subjects to avoid biasing children's responses to
questions about alternative job titles.
In all, 18 three-card sets depicted 54 different occupations.
Of these, 35 had titles that were linguistically
unmarked for gender. Of the unmarked items, 15
came from distinct job settings. The remaining 20 represented
10 shared-setting pairs. Half of these pairs
crossed cultural gender stereotypes (e.g., doctor /
nurse) and half did not (e.g., train conductor / train engineer).
Represented within the unmarked titles were
culturally masculine, feminine, and neutral occupations.
The remaining 19 occupations were jobs with
marked titles, of which 10 had strong suffix markings
(e.g., policeman/policewoman), and 9 had weak suffix
markings (e.g., waiter/waitress), with suffix gender
manipulated between children. From the perspective
of a complete design it would have been
preferable to have manipulated shared versus distinct
work settings within marked as well as within unmarked
titles. With the exception of policeman and
metermaid, however, we were unable to identify common
exemplars of jobs that had differently marked titles
but shared work settings, and thus setting type was
not systematically varied for marked titles. The full set
of occupational titles is given in the Appendix.
Dependent Measures
The procedure provided several different kinds of
dependent measures. One measure was whether children's
first picture selection was, in fact, an illustration
of someone in the named occupation (correct
identification). Another measure was whether children's
first selection was of the man or woman, which
we took as an index of which sex had the greater salience
for the named job (sex of first selection). A third
measure assessed whether children correctly accepted
the second person on the card as an illustration of the
named occupation (other-sex title extension). Also recorded
were verbatim responses to the open-ended
questions. Specifically, recorded were (1)job titles given
by children when they rejected the provided title for another
depicted worker (replacement titles) and (2) children's
explanations of what was the same or different
about the jobs being performed by people pictured (job
duties). Categories used to classify the open-ended
responses are described in the results section.
820 Child Development
Overview and Preliminary Analyses
As described earlier, we were interested in two broad
classes of occupational titles: those that have no linguistic
marking that might be taken to imply the sex of the
referent (unmarked) and those that do (marked). Within
unmarked titles there was an additional distinction:
whether the jobs came from shared- or distinct-work
settings. To check our assumption that these two kinds
of unmarked occupations did, indeed, function differently;
we first conducted a one-way ANOVA in which
job type (shared versus distinct setting) was a withinsubject
variable, and the dependent measure was the
percentage of correct identifications. As predicted, the
effect of job type was significant, F(l, 48) = 391.38, P<
.001, with a lower percentage of correct responses for
jobs from shared than from distinct settings, Ms (5Ds) =
52 (14)versus 97 (5).Given that the two job types elicited
different levels of accuracy, and given that each necessarily
contained different subtypes within them,
remaining analyses were conducted separately for
shared- and distinct-setting job types.
Thus, the sections that follow cover, in tum, analyses
of data on (1) jobs that have unmarked job titles and
come from shared settings, (2)jobs that have unmarked
titles and come from distinct job settings, and (3) jobs
that have marked titles. Finally, (4) data are presented
on children's responses to the open-ended questions
concerning replacement job titles and job duties.
As in Study I, percentages (rather than the raw
numbers) of "both" responses served as the dependent
measure; and, as before, analyses of transformed
data yielded comparable findings and thus are not
discussed further.
Linguistically Unmarked Occupational Titles:
Shared-Setting Jobs
Preliminary analyses. Given the large number of
potentially important between-subjects factors, a preliminary
ANOVA was run to evaluate effects of children's
sex and card set (e.g., whether the female nurse
was shown taking a patient's temperature and the
male nurse was shown giving an injection, or the reverse).
As anticipated, results indicated no significant
main effects or interactions involving either variable,
and thus data were pooled across these factors in the
remaining analyses.
Correct identifications. The percentage of correct
identifications served as the dependent variable in a 2 X
2 ANOVA in which the between-subjects factor was
age (younger versus older) and the within-subjects
factor was cultural stereotype of the pair (same stereotype
versus different stereotype). The only significant
effect was the main effect of age, F(l, 49) = 4.80, P<
.05, with older children giving a greater percentage of
correct identifications than younger children, Ms
(5Ds) = 56 (12) and 47 (16), respectively.
Inspection of the data indicated that for occupational
titles referring to shared-setting jobs, whenever children
failed to select one of the pictures on the intended
card, they virtually always selected one of the pictures
on the shared-setting card. For example, if children
erred when asked to find a picture of a "doctor," it was
virtually always because they selected a picture on the
nurse card. Indeed, X2 analyses indicated that neither
younger nor older children selected a picture on the intended
card at greater than chance (50%) levels. Thus,
children showed little ability to discriminate between
the occupational activities of shared-setting jobs.
It is critically important to note here that children's
difficulty in selecting the intended card posed no problem
for the core question of interest, namely whether
children would extend the named job title to the other
depiction on the same card. For example, consider a case
in which children were to find a card that had a picture
of someone who could be called a "doctor" and suppose
that children selected the card that had been intended to
depict "nurses." The critical question still remained
whether children were willing to extend the title "doctor"
to the second picture on the selected card. Thus for
purposes of data analysis, a picture on either of the two
shared-setting job cards was accepted as correct, and
then sex of first selections and other-sex title extensions
were examined as explained earlier.'
2 A possible explanation of children's difficulty in selecting the
intended card is that the pictorial depictions were insufficiently
distinct to allow viewers to discriminate between the jobs. In other
words, it may have been limitations in the depictions, rather than
limitations in children's understanding of occupational activities,
that accounted for children's poor performance. To test this notion,
52 college students (27 females, 25 males) were shown the identical
set of stimuli and, using an abbreviated version of the procedure,
were asked to select the named occupation from the shared-setting
occupational pairs. Adults were correct 75% of the time (significantly
above chance) compared with children who were correct
only 51% of the time (not above chance level). Although adults
were thus considerably more accurate than children, even they did
not perform at ceiling. This finding suggests either the inadequacy
of the images and / or the subtle and often visually imperceptible
nature of differences in job activities. As noted in the text, however,
this issue presents no difficulty for the current research because the
question about title extension applies equally well to either picture
set. That is, once having selected a given card, the question of importance
was whether children would extend the same occupational
title to the second picture on that same card. Again, by having
counterbalanced whether a man or woman appeared in each
particular pictorial context, we can be confident that the findings
reflected something about the participants' beliefs, rather than
something about the particular pictorial depictions.
Sex offirst selection. To evaluate the relative salience
of men and women as exemplars of occupations
in relation to the cultural gender stereotype of
the occupation, the percentage of time a picture of a
woman was selected first was the dependent measure
in a two-way ANOVA in which age (younger, older)
was a between-subjects factor and cultural stereotype
of the job (neutral, masculine, feminine) was a withinsubjects
factor. The main effect of cultural stereotype
of the job was significant, F(2, 94) == 4.59,P< .05.Followup
tests showed that pictures of women were selected
significantly more often for culturally feminine jobs
than for culturally neutral or culturally masculine
jobs, with the latter two not differing significantly
from one another, Ms (SDs) == 64 (32), 50 (35), and 47
(25), respectively.
Other-sex title extensions. Next examined was
whether children correctly extended the named occupational
title to the other-sex picture on the same
card. This was the most critical test of children's understanding
that an unmarked job title can be used
for someone of either sex. Children were also asked
whether the two persons performed the same or different
job activities. We had originally expected to analyze
children's responses concerning activities separately.
In cases in which children correctly extended
the title to the other-sex picture, however, they only
rarely (12 of 510 responses) asserted that the duties of
the two workers differed. Because there were so few
of these responses, activity responses were not analyzed
The data were analyzed by a two-way ANOVA in
which the dependent variable was the percentage of
time children correctly extended the title to the othersex
picture on the card, the between-subjects factor
was age, and the within-subjects factor was the cultural
stereotype of the pair. As predicted, the main effect
of pair type was significant, F(l, 47) == 12.44, P<
.001. Children correctly extended occupational titles to
the other-sex picture on the card more often with samestereotype
pairs than with different-stereotyped pairs,
Ms (SDs) == 98 (5) and 89 (19), respectively.
Responses to a second card. As another means of assessing
children's understanding of occupational
titles, children were asked "ls there another card
showing someone you could call a__?" The correct
answer to these questions was "no" because at most,
another card showed a related (shared-setting) job
rather than an identical occupation. Overall, however,
children stated that the card that was intended
to show the shared-setting job depicted someone who
could also be called by the target occupational title
88% of the time. This finding is consistent with children's
failure to discriminate between the shared-
Liben, Bigler, and Krogh 821
setting job depictions when asked the initial job identification
question discussed earlier (see also footnote
2). Taken together, these data suggest that the sharedsetting
job cards were generally perceived by children
as depicting the identical occupations. Patterns of the
remaining data on the second card (i.e., data on sex of
the first selection, other-sex title extensions, and responses
to the open-ended questions) were like those
already reported for the first card. Thus, to avoid redundancy,
these data are not presented.
Linguistically Unmarked Occupational Titles:
Distinct-Setting Jobs
Preliminary analyses. The next set of analyses concerned
children's understanding of titles for jobs that
do not have shared-setting counterparts and contain no
linguistic gender markings. Again, preliminary analyses
on children's sex and card set revealed neither main
effects nor interactions, and thus data were pooled
across these variables in the remaining analyses.
Correct identifications. The percentage of correct job
identifications was evaluated with anANOVA with a
between-subjects variable of age and a withinsubjects
variable of the cultural stereotype of the job
(neutral, masculine, feminine). Neither main effects
nor interactions were significant. Overall, the percentage
of correct responses was high, M == 97 (SD == 7).
Sexoffirst selection. The percentage of time that children
selected the picture of the woman worker first
was analyzed by an ANOVAwith a between-subjects
factor of age (younger, older), and a within-subjects
factor of cultural stereotype of the job (neutral, masculine,
feminine). The analysis revealed a main effect
of cultural stereotype of job, F(2, 98) == 8.88, P< .001,
with pictures of women selected significantly more
often for feminine than for neutral or masculine jobs,
with the latter two not differing significantly from
one another, Ms (SDs) == 69 (27), 59 (22), and 51 (24),
Other-sex title extensions. Children's willingness to
extend linguistically unmarked occupational titles
was analyzed with a two-way ANOVA in which the
dependent measure was the percentage of time that
children correctly applied the occupational label to
the other-sex individual portrayed in the same job.
Again, age (younger, older) was the between-subjects
variable and cultural stereotype of the job (neutral,
masculine, feminine) was the within-subjects variable.
As before, when children agreed that the label
could be extended, they only rarely (15 out of 765)
said that the other-sex worker performed different
duties, and thus these responses were not analyzed
822 Child Development
The ANOVA revealed neither a main effect nor an
interaction. Overall, the percentage of time that children
correctly extended the job title to the other-sex
depiction was high, M == 92 (SD =15).
Linguistically Marked Occupational Titles
Preliminary analyses. The next analyses concern
children's understanding of marked titles. Preliminary
analyses again revealed neither main effects nor
interactions for children's sex or card set, and thus
data were pooled across these variables in the analyses
reported below.
Correct identifications. The percentage of correct responses
was the dependent measure in an ANOVA in
which the between-subjects factors were age (younger,
older) and suffix gender (masculine or feminine), and
the within-subjects factor was suffix strength (strong
versus weak). Results indicated no significant main
effects or interactions. Children were highly accurate
in finding a correct depiction of the named occupation,
M == 97 (SD == 8).
Sex offirst selection. Again, the percentage of time
that children selected the picture of a woman first was
the dependent measure in a three-way ANOVA in
which the between-subjects factors were age (younger,
older) and suffix gender (masculine, feminine), and the
within-subjects factor was suffix strength (strong,
weak). Results showed a significant two-way interaction
between suffix strength and suffix gender, F(l, 46) ==
45.05,P < .001. Post hoc analyses indicated that children
chose a picture of a women first significantly
more often whenever the suffixes were feminine
rather than masculine, but that this effect was significantly
greater for strong suffixes, Ms (SDs) == 98 (4)
versus 28 (26), than for weak suffixes, Ms (SDs) == 58
(21) versus 41 (18), respectively. Also significant were
the subsumed main effects of suffix gender, F(l, 46) ==
12.70, P < .001, and suffix strength, F(l, 46) == 16.63,
P < .001.
Other-sex title extensions. The percentage of time
that children extended the occupational title to the
other-sex picture served as the dependent measure in
a three-way ANOVA in which the between-subjects
factors were age (younger, older) and suffix gender
(masculine, feminine), and the within-subjects factor
was suffix strength (strong, weak). Results showed a
significant two-way interaction of suffix gender and
suffix strength, F(l, 46) == 4.72, P< .05. Post hoc comparisons
indicated that children extended occupational
titles to the worker of the other sex more often
for titles with masculine than feminine suffixes, but
that the difference was significantly greater for strong
suffix titles, Ms (SDs) == 70 (30) versus 26 (36) than for
weak suffix titles, Ms (SDs) == 94 (13) versus 79 (3D),
respectively. Also significant were subsumed main effects
of suffix gender, F(l, 46) == 12.67, P < .001, and
suffix strength, F(l, 46) == 49.72, P< .001.
Replacement Occupational Titles and
Comparative Job Duties
Overview. As explained earlier, whenever children
rejected a given occupational title for the other-sex
worker on the same stimulus card, they were asked
what that worker would be called. In addition, irrespective
of whether children extended the given title
to the other-sex worker, they were asked if the two
workers shown on the card did the same or different
things, and to explain. The sections below describe
data on replacement job titles and job duties, divided
by item types as in prior analyses.
Unmarked occupational titles: Shared-setting jobs.
For unmarked titles referring to jobs from shared
work settings, there were 30 instances in which children
rejected the title for the other-sex worker. Of
these, 26 occurred in different-stereotype pairs (such
as doctor / nurse). As shown in Table I, the most common
type of replacement titles offered a different job
title taken from the same general field. Remaining responses
were divided among descriptions of the picture,
new but equivalent job titles, and titles naming a
different field. Interestingly, almost three fourths of
the 30 replacement titles served to avoid what would
have been (if the original title had been extended to
the other picture on the same card) a nontraditional
worker. In these cases, children's substitute titles rendered
the depictions as traditional (as judged by
adult ratings). Only 1 of the 30 created a counter stereotypic
worker in place of what would have been a
traditional one. The remaining replacement titles left
the original nature of the portrayal unchanged. In
over half the cases that children rejected the title for
the other-sex worker (63%), they nevertheless said
that the two workers performed the same duties.
Unmarked occupational titles: Distinct-setting jobs.
For unmarked titles referring to jobs from distinct settings,
there were 53 instances in which children rejected
the title for the other-sex worker. As shown in
Table 2, the most common replacement titles were
those that referred to a different occupational field
from that of the original title, and next most common
were descriptions of the worker's activity (rather
than actual job titles). Occasionally, children generated
a common but equivalent job title and sometimes
generated a completely novel occupational title.
As in shared-setting titles, the largest group of replacement
titles (27 of the 53) eliminated what would have
Liben, Bigler, and Krogh 823
Table 1 Unmarked Titles, Shared-Setting Jobs: Replacement Titles Given After Original Title Was
Rejected for Other-Sex Worker
Type of Replacement Nontraditional -. Traditionality Traditional -.
(e.g., original -. "new") Traditional Unchanged Nontraditional
Description of picture
(baker -. "person in the kitchen") 3 0
Same job
(television anchor -. "news reporter") 0 5 0
Same field, different job
(doctor -. "nurse") 16 1 0
Different field
(teacher -. "judge") 3 0 1
been a nontraditional worker, with the inverse occurring
only rarely (2 of the 53). As was the case for
shared-setting jobs, over half the time that the title
was rejected for the other-sex worker (58%), children
nevertheless said that the workers performed the
same duties.
Marked titles. For marked titles, there were 346 instances
in which children refused to extend the title to
the other-sex worker. As implied by the data reported
earlier on other-sex title extensions, the vast majority
of these (315 or 91%) were for strong suffix titles, occurring
far more frequently for titles that were initially
feminine (225 instances) than for those that
were initially masculine (90 instances). Irrespective of
the type of original title, the replacement titles offered
by children predominately (253 of the 346) matched
the alternative form of the title that had been prepared
for the study itself (e.g., substituting woman for
man, or master for mistress). Most of the remaining replacement
titles (65) retained the root occupation, but
substituted some other kind of gender-specific
marker (e.g., replacing "policewoman" with "police
guy"). Next most common (26) were replacement
titles that substituted an entirely new occupation (e.g.,
replacing "stewardess" with "pilot"), with only rare
(2) replacement titles that substituted a genderneutral
term for a gender-marked one (e.g., replacing
"fireman" with "firefighter").
At the most general level, the findings of Study 2
were similar to those of Study 1 in demonstrating that
children make gender-specific interpretations of job
titles under some circumstances. First, consistent with
past research and contemporary recommendations
concerning the advantages of using nonsexist language/
gender-specific interpretations were most common
for marked occupational titles. Furthermore, and
consistent with initial hypotheses and the results from
Study 1/ certain kinds of marked titles-those with
strong suffixes and feminine suffixes-were especially
likely to lead to restrictive interpretations.
The data from Study 2 on unmarked titles also
demonstrated that gender-specific interpretations can
occur even when occupational titles are meant to be
entirely silent about the worker's sex. In Study 2/ such
interpretations were less common than they had been
Table 2 Unmarked Titles, Distinct-Setting Jobs: Replacement Titles Given After Original Title
Was Rejected for Other-Sex Worker
Type of Replacement Nontraditional -. Traditionality Traditional -.
(e.g., original>- "new") Traditional Unchanged Nontraditional
Description of picture
(auto mechanic -. "ladyfixing a car") 8 4 1
Same job
(guitarist-. "bandmember") 0 9 0
Same field, different job
(secretary -. "office manager") 2 2 0
Different field
(librarian -. "companyworker") 17 4
Novel label
(farmer -. 'jarmtress") 5
824 Child Development
in Study 1. How best to interpret this difference is a
topic for further research. The least interesting possibility
is that the contrast simply stemmed from the
different set of job titles used in the two studies (a difference
necessitated by the goal in Study 2 of testing
our hypotheses about pair types).
A more interesting possibility is that the same underlying
gender-schematic processes affected the incidence
of other-sex extensions in both studies, but
that they worked in opposite directions under the
two methods. The method used in Study 1 might
have underestimated children's willingness to extend
titles to the other sex worker because children might
have had difficulty even imagining that someone of
the other sex performs the job. That is, the verbal
phrase alone ("a woman doing the job") may be insufficient
to allow respondents to imagine a woman
actually performing the named job, particularly for
respondents who had strongly gender-stereotyped
beliefs. The graphic depictions of men and women
engaged in job duties that were used in Study 2 may
have been more effective in establishing the premise
that someone of the other sex was actually performing
the named job. At the same time, the method used
in Study 2 might have overestimated children's willingness
to extend titles to other-sex workers because
some children might have failed to interpret the
graphic stimuli correctly. Much of our work (Liben &
Signorella, 1980, 1993; Signorella & Liben, 1984) as
well as that of others (e.g., Martin & Halverson, 1983;
Stangor & Ruble, 1989)has demonstrated that children
sometimes distort the sex of character when drawings
show people in nontraditional roles. In these cases of
distortion, stimuli become functionally consistent
with cultural gender stereotypes. So in the present
study, for example, some children may have agreed
that the label "doctor" could be extended to the
"other picture on this card" because they interpreted
that other picture simply as another exemplar of a
male doctor rather than-as intended-as a picture of
a female doctor.
Irrespective of the ultimate explanation for the
lower and, in an absolute sense, small confusion rate
in Study 2, it is noteworthy that confusions occurred
at all given that the titles were entirely unmarked for
gender, and given that men and women were depicted
in identical ways. Importantly, the contexts in
which errors occurred were not random, but instead
varied systematically and as predicted: there were significantly
fewer other-sex title extensions to sharedsetting
jobs when job pairs crossed cultural gender
stereotypes than when they did not. These data are
consistent with our original hypothesis that children
are more likely to think that unmarked job titles carry
meaning about biological sex when men and women
are stereotypically associated with different occupations
in the same setting. Again, however, an alternative
interpretation that must be considered is that the
difference reflects the particular materials that comprised
the stimulus sets. That is, as a consequence of
what jobs are in American culture and hence were
available for stimulus materials, jobs in the sharedsetting,
different-stereotype category may have
been overall more strongly gender stereotyped than
were jobs included among the shared-setting, samestereotype
category. Although this alternative interpretation
cannot be ruled out on the basis of current
data alone, it is made less viable by the data from the
distinct-setting jobs: the incidence of other-sex title extensions
did not vary in relation to whether the jobs
were culturally stereotyped as masculine, feminine, or
Children's replacement titles provided converging
evidence that at least under some conditions children
believe that unmarked occupational titles are sex specific.
Again, although the number of replacement titles
was not large enough to suggest an overwhelming
and pervasive problem, it was large enough to document
the existence of children's confusions, and to
show that these confusions were systematic rather
than random. Collapsing over the data shown in
Tables 1 and 2, there were 83 replacement titles (a number
that underestimates the phenomenon because, as
explained earlier, all second-card data were omitted
from the analyses to avoid redundancy). What is
striking in these data is that when children rejected an
unmarked occupational title for a worker of the other
sex, they were likely to provide a replacement title that
eliminated what would have been a nontraditional depiction.
The reverse occurred only very rarely.
As predicted, replacement titles given for the
marked titles generally matched the other-sex version
of the title that had been used for the study itself (e.g.,
replacing "policeman" with "policewoman"). This
was not invariably the case, however, as, for example,
the child who rejected "scoutmaster" for a woman,
offering the replacement title of "canoeist." Perhaps
children are already aware of the problem noted by
the curators of the "Old Masters and Old Mistresses"
exhibit cited in the Introduction.
Children's responses about job activities were also
consistent with the initial hypothesis that children
sometimes mistakenly believe that people in sharedsetting
jobs are called by different titles simply because
they are of a different biological sex. Specifically,
as described earlier, for the shared-setting jobs,
almost two thirds of the time that children refused to
extend a job title to the other-sex worker, they nevertheless
claimed that both workers performed the same
job duties. Such responses also occurred in the distinctsetting
jobs, but less often (just under half the time).
It was not possible to test the finer grained predictions
discussed in the introduction to Study 2 concerning
children's beliefs about job duties, because
children offered little detail about occupational
duties. Whether this reflects their impoverished knowledge
structure or instead reflects our failure to probe
with follow-up questions is unknown. It would, however,
be interesting if future work examined explicitly
whether children who are aware that both men and
women can hold the same occupational title (e.g.,
"doctor" or "business executive") nevertheless recognize
that men and women in these occupations may
have different responsibilities at a finer level of analysis
(e.g., being aware of sex differentials in surgeons
versus pediatricians, or in vice presidents versus chief
executive officers).
Taken together, the findings from these studies provide
evidence that although the workplace is becoming
more balanced with respect to distributions of
men and women, the psychological world of work remains
strongly gendered. At the broadest level, the
findings show that gender-related judgments about
jobs vary in relation to characteristics of occupational
titles. Although this relation cannot, alone, be used to
evaluate whether language plays a causal role in the development
of gender-related occupational beliefs, it
does provide an important foundation for justifying further
research on the role of language, and has potential
implications for language and educational practices.
From the perspective of trying to understand how
children's gender-related beliefs about occupations
develop, a particularly important need is for future
work to address the question of whether language
can itself create gendered beliefs about occupations.
This question could be tested by manipulating the
precise forms of the titles that are used to teach about
an occupation, and evaluating whether children's
subsequent job beliefs and interests are affected. It
would be important to consider whether effects, if observed,
are moderated by children's own initial gender
attitudes. We would predict that children with
more highly gender-stereotyped attitudes would be
more likely to be affected by strong-marked titles,
and would be likely to impose gendered meaning
even on gender-neutral titles. In short, the model that
we propose is a bidirectional one: language is thought
to influence children's beliefs about the gendered nature
of occupations while, simultaneously, children's
Liben, Bigler, and Krogh 825
existing gender attitudes are thought to shape their
interpretations of language. Support for the bidirectional
nature of effects comes from Study 1 insofar as
title effects were significantly moderated by individual
differences in participant's level of stereotyping.
Unfortunately, practical constraints prevented the
collection of individual difference measures in Study
2, and thus we can only speculate that the effects observed
were likely to have been disproportionately
accounted for by the more highly gender-stereotyped
From the perspective of recommendations about
language itself, the findings from the current studies
reinforce and extend conclusions from prior literature,
and raise some interesting and potentially important
cautions about the changing lexicon. First, the
findings of the present research join a substantial
body of work demonstrating that strong-marked occupational
titles are exclusionary. The present work
thus reinforces existing recommendations and rules
against using these titles in help-wanted advertisements,
educational materials, and other forms of prose.
To our knowledge, however, there is no comparable
research literature that has examined the interpretation
of what we have called weak-marked titles. The
present data suggest that masculine versions of weakmarked
titles are almost always accepted as applicable
to women as well as men. Findings from the sex-offirst-
selection measure in Study 2 suggest that even
their salience is not overwhelmingly skewed: pictures
of women workers were chosen first 40% of the time
for masculine, weak suffix titles.
These data raise an interesting problem about parallel
masculine and feminine versions of weak suffix
titles. One possibility is to conclude that what we
have called weak masculine suffixes (like "master" or
"-er") are really functionally neutral, and thus these
titles can be retained and used routinely for both men
and women. Another possibility is to routinely use
both masculine and feminine forms for generic referents,
much as we now routinely use "he or she." What
seems particularly problematic, however, is the current
condition in which sometimes parallel versions
are used, and sometimes they are not. For example, to
the extent that the title "postmistress" becomes increasingly
common, one might reasonably come to
infer that the title "postmaster" is restricted to male
referents only. By extension, one might increasingly
infer that other "master" words refer only to men,
even when there is no feminine counterpart title in
common use.
This reasoning raises a similar issue for unmarked
titles as well. In current parlance, many people mark
nontraditional instances of otherwise unmarked job
826 Child Development
titles with a leading feminine or masculine modifier
(e.g., "lady dentist" or "man nurse"). This occurred in
children's replacement titles in Study 2 and has been reported
elsewhere (Liben & Signorella, 1993; Rosenthal
& Chapman, 1982).By implication, these gender modifiers
mark the particular instance as atypical, and
imply that the canonical referent is a worker of the
unnamed sex. As in the case of postmistress considered
above, this linguistic form thus has the potential
outcome of suggesting to listeners that the meaning
of the unmodified title refers to the traditional sex
only. Repeated linguistic encounters like these seem
likely to exacerbate existing gender stereotypes about
jobs. A possible solution to this problem (albeit a radical
and unwieldy one) would be to routinely preface
every occupational title by a gender modifier. Both
gender modifiers would be used when the referent's
sex is unknown (e.g., "Always remember to obey the
man police officer or the woman police officer on
duty"). One gender modifier would be used when the
gender of the referent was known, but it would be included
irrespective of whether the instance was traditional
or nontraditional (e.g., not only "I saw a
woman dentist this morning," but also "I saw a man
dentist this morning").
Whether a radical suggestion like this is a good
one would require additional research and debate,
particularly because it would be repeatedly characterizing
people along a particular grouping variable
(biological sex). There have been forceful political
arguments against gratuitously drawing attention to
sex per se (see especially Bern, 1983, 1993), and empirical
data have shown that functional groupings
based on perceptually salient characteristics have
negative consequences for intergroup attitudes (Bigler,
1995; Bigler, Jones, & Lobliner, 1997). Of course,
even if the routine use of sex-specific modifiers were
judged beneficial, it is unrealistic to expect that their
use would be consistent. This reality makes it particularly
important to design educational programs
that explicitly teach children about the irrelevance of
sex and provide a nonrestrictive vocabulary for the
world of work.
In planning such interventions, it is important to
remember the bidirectionality of effects noted above
and observed in past work. That is, prior research on
schematic processing (e.g., Carter & Levy, 1988; Levy,
1989, 1994;Liben & Signorella, 1980; Martin & Halverson,
1981, 1983;Signorella & Liben, 1984) as well as the
findings from Study 1 on participant stereotyping,
suggest that highly stereotyped children are likely to
reinterpret nontraditional instances to make them
consistent with their own prior beliefs. This situation
makes it particularly important to provide highly explicit
and even didactic educational programs rather
than merely exposing children to nonstereotypic instances
and appropriate language. For example, as
suggested by children's greater willingness to extend
titles to the worker of the other sex in Study 2 than in
Study 1, it may be important to provide explicit pictorial
(or live) examples of both male and female
workers of a specific occupation, as well as to provide
explicit statements about the applicability of the occupational
title to workers of both sexes. It may also be
necessary to provide explicit corrective feedback as in
prior interventions for teaching children about the irrelevance
of gender for occupational choice (Bigler &
Liben, 1990, 1992).
From the perspective of individuals, participation
in interventions such as these may lead children to
consider jobs that would otherwise have been dismissed
as irrelevant. From the perspective of society
more broadly, such interventions might reduce the
number of children who develop into college students
or publishers' representatives who ask women
professors to take messages for their male colleagues
on the assumption that the former are secretaries for
the latter.
In summary, it is already well-established that children
should be exposed to both male and female role
models in the same occupations. The data reported
here suggest that it may also be important to modify
the language of work to expand the occupational
choices that children consider.
The authors appreciate the friendly cooperation of the
children, parents, and personnel of the Boys and Girls
Clubs of Central Minnesota and its director, Mark
Sakry; and Extend-A-Care in Austin, TXand its director,
John Combs. They thank Lecianna Jones, Kirsten
Mackler, Erica Michael, T. Anne Newman, Tony
North, Lisa Szechter, Gina Toumi, and Candice Yekel
for their help in various aspects of data collection and
entry. The authors are also gratefUl to Margaret Signorella
and two anonymous reviewers for their thoughtful
suggestions on earlier versions of this manuscript.
Corresponding author: Lynn S. Liben, Department of
Psychology, The Pennsylvania State University, 417
Moore Building, University Park, PA 16802; e-mail: Rebecca S. Bigler is at the University
of Texas at Austin; and Holleen R. Krogh is at the Mississippi
University for Women, Columbus.
Unmarked Occupational Titles. Masculine: banker, farmer,
lieutenant, professor, train engineer, dentist, architect, company
president, sergeant, barber, auto mechanic, doctor,
plumber, business executive, train conductor, truck driver,
construction worker, scientist. Feminine: secretary, nurse,
librarian, cheerleader, dental assistant, teacher, beautician,
fashion model, bank teller, baby-sitter, housekeeper. Neutral:
musician, cashier, pianist, writer, singer, artist.
Marked Occupational Titles. Strong suffix: businessman/
woman, fisherman/woman, garbageman/ woman, policeman/
woman, chairman/woman, yardman/woman, shipman/
woman, trainman/woman, mailman/woman, fireman/
woman, truckman/ woman, salesman/ woman. Weak
suffix: schoolmaster/ mistress, ringmaster / mistress, bandmaster
/ mistress, scoutmaster/ mistress, choirmaster / mistress,
waiter / waitress, steward / stewardess.
Unmarked Titles: Shared Setting. Different-stereotype pairs,
masculine / feminine: banker /bank teller, business executive
/ secretary, dentist / dental assistant, doctor / nurse, professor
/ teacher. Same-stereotype pairs, masculine: general/
lieutenant, train conductor/ train engineer. Same-stereotype
pairs, neutral: baker / cook, pianist/ composer, television
anchor / television reporter.
Unmarked Titles: Distinct Setting. Masculine: auto mechanic,
farmer, plumber, scientist, truck driver. Feminine:
baby-sitter, cheerleader, fashion model, gymnast, librarian.
Neutral: artist cashier, guitarist, singer, writer.
Marked Titles. Strong suffix: Coyboy / girl, choirboy /
girl, fireman/woman, fisherman/woman, garbageman/
woman, househusband/wife, mailman/woman, policeman/
woman, salesman/woman, washerman/woman. Weak
suffix: Postmaster / mistress, ringmaster / mistress, scoutmaster
/ mistress, actor / actress, janitor / janitress, sculptor /
sculptress, shepherd / shepherdess, steward / stewardess,
waiter / waitress.
American Psychological Association. (1994). Publication
manualof the AmericanPsychological Association (4th ed.).
Washington, DC: Author.
Archer, C. J. (1984). Children's attitudes toward sex-role division
in adult occupational roles. Sex Roles, 10, 1-9.
Bem, S. (1974). Gender schema theory: A cognitive account
of sex typing. Psychological Review, 88, 354-364.
Bem, S. L. (1983). Gender schema theory and its implications
for child development: Raising gender-aschematic
children in a gender-schematic society. Signs, 8, 598-616.
Bem, S. L. (1993). The lenses of gender: Transforming the debate
onsexual inequality. New Haven, CT:YaleUniversity Press.
Bem, S. L., & Bem, D. J. (1973). Does sex-biased advertising
"aid and abet" sex discrimination? Journal of Applied Social
Psychology, 3, 6-18.
Liben, Bigler, and Krogh 827
Bigler, R. S. (1995). The role of classification skill in moderating
environmental effects on children's gender stereotyping:
A study of the functional use of gender in the
classroom. Child Development, 66, 1072-1087.
Bigler, R. 5., Jones, L. C., & Lobliner, D. B. (1997). Social categorization
and the formation of intergroup attitudes in
children. Child Development, 68, 530-543.
Bigler, R. 5., & Liben, L. S. (1990). The role of attitudes and
intervention in gender-schematic processing. Child Development,
61, 1440-1452.
Bigler, R. 5., & Liben, L. S. (1992). Cognitive mechanisms in
children's gender stereotyping: Theoretical and educational
implications of a cognitive-based intervention.
Child Development, 63, 1351-1363.
Bigler, R. 5., Liben, L. 5., & Yekel, C. A. (1992, August). Developmental
patterns of gender-related beliefs: Beyond
unitary constructs and measures. Paper presented at the
annual meeting of the American Psychological Association,
Washington, DC.
Brownmiller, S. (1999). In our time: Memoir of a revolution.
New York: Random House.
Calvert, S. L., & Huston, A. C. (1987). Television and children's
gender schemata. In L. S. Liben & M. L. Signorella
(Eds.), Children's gender schemata (pp. 75-88). San Francisco,
CA: Jossey-Bass.
Carter, D. B., & Levy, G. (1988). Cognitive aspects of early
sex-role development: The influence of gender schemas
on preschoolers' memories and preferences for sextyped
toys and activities. ChildDevelopment, 59, 782-792.
Clark, H. H. (1973). Space, time, semantics, and the child. In
T. E. Moore (Ed.), Cognition and the acquisition of language
(pp. 27-63). New York: Academic Press.
Coltrane,S., & Adams, M. (1997).Work-family imagery and
gender stereotypes: Television and the reproduction of
difference. Journal of Vocational Behavior, 50, 323-347.
Cordua, G. D., McGraw, K. 0., & Drabman, R. S. (1979).
Doctor or nurse: Children's perceptions of sex typed occupations.
Child Development, 50, 590-593.
Erikson, E. (1968). Identity: Youth and crisis. New York:
Flexner, S. B. (Ed.). (1987). TheRandomHousedictionary of the
Englishlanguage (2nd ed.). New York: Random House.
Gabhart, A., & Broun, E. (1972). Old mistresses: Women artists
of the past. Walters Art Gallery Bulletin, 24(7).
Gettys, L. D., & Cann, A. (1981). Children's perceptions of
occupational sex stereotypes. Sex Roles, 7, 301-308.
Helwig, A. A. (1998). Gender-role stereotyping: Testing theory
with a longitudinal sample. Sex Roles, 38, 403-423.
Hyde, J. S. (1984). Children's understanding of sexist language.
Developmental Psychology, 20, 697-706.
Levy, G. D. (1989). Developmental and individual differences
in preschoolers' recognition memories: The influence
of gender schematization and verbal labeling of information.
Sex Roles, 21, 305-324.
Levy, G. D. (1994). High and low gender schematic children's
release from proactive interference. Sex Roles, 30,
Liben, L. 5., Bigler, R. 5., & Krogh, H. R. (2001). Pink and
blue collar jobs: Children's judgments of job status and
828 Child Development
job aspirations in relation to sex of worker. Journal of Experimental
ChildPsychology, 79,346-363.
Liben, L. S., & Signorella, M. L. (1980). Gender-related schemata
and constructive memory in children. Child Development,
51, 11-18.
Liben, L. S., & Signorella, M. L. (1993). Gender-schematic
processing in children: The role of initial interpretations
of stimuli. Developmental Psychology, 29, 141-149.
Martin, C. L., & Halverson, C. F. (1981). A schematicprocessing
model of sex typing and stereotyping in children.
Child Development, 52, 1119-1134.
Martin, C. L., & Halverson, C. F. (1983). The effects of sextyping
schemas on young children's memory. Child Development,
54, 563-574.
Martyna, W. (1978). What does "he" mean? Use of the generic
masculine. Journal of Communication, 28, 131-138.
Martyna, W. (1980). Beyond the "he/man" approach: The
case for nonsexist language. Signs,S, 482-493.
McConnell, A. R., & Fazio, R. H. (1996).Women as men and
people: Effects of gender-marked language. Personality
and Social Psychology Bulletin,22, 1004-1013.
O'Keefe, E. S. C., & Hyde, J. S. (1983). The development of
occupational sex-role stereotypes: The effects of gender
stability and age. Sex Roles, 39, 481-492.
Phillips, S. D., & Imhoff, A. R. (1997).Women and career development:
A decade of research. Annual Review of Psychology,
48, 31-59.
Purcell, P., & Stewart, L. (1990). Dick and Jane in 1989. Sex
Roles, 22, 177-185
Rey, H. A. (1998). CuriousGeorge. Boston: Houghton Mifflin.
(Original work published 1941)
Rosenthal, D. A., & Chapman, D. C. (1982). The lady spaceman:
Children's perceptions of sex-stereotyped occupations.
Sex Roles, 8, 959-965.
Ruble, D. N., & Martin, C. L. (1998). Gender development.
In N. Eisenberg (Ed.), W.Damon (Series Ed.), Handbook of
childpsychology: Vol. 3. Social, emotional and personality development(
5th ed., pp. 933-1016). New York: Wiley.
Safire, W. (1999, May 16). On language: Genderese. TheNew
York Times Magazine, 30-31.
Serbin, L. A., Powlishta, K. K., & Gulko, J. (1993). The development
of sex typing in middle childhood. Monographs
of the Society for Research in ChildDevelopment, 58(2, Serial
No. 232).
Signorella, M. L. (1987). Gender schemata: Individual differences
and context effects. In L. S. Liben & M. L. Signorella
(Eds.), New directions for child development. No. 38.
Children's gender schemata (pp. 23-37). San Francisco:
Signorella, M. L. (1999). Multidimensionality of gender
schemas: Implications for the development of genderrelated
characteristics. In W. B. Swann, [r., J. H. Langlois,
& L. A. Gilbert (Eds.), Sexismand stereotypes in modern society:
Thegender science ofJanet Taylor Spence (pp.107-126).
Washington, DC: American Psychological Association.
Signorella, M. L., Bigler, R. S., & Liben, L. S. (1993). Developmental
differences in children's gender schemata
about others: A meta-analytic review. Developmental Review,
13, 147-183.
Signorella, M. L., & Liben, L. S. (1984). Recall and reconstruction
of gender-related pictures: Effects of attitude,
task difficulty, and age. Child Development, 55, 393-405.
Signorielli, N., & Bacue, A. (1999). Recognition and respect:
A content analysis of prime-time television characters
across three decades. Sex Roles, 40, 527-544.
Stangor, C., & Ruble, D. F. (1989). Differential influences of
gender schemata and gender constancy on children's information
processing and behavior. Social Cognition, 7,
Stockard, J., & McGee, J. (1990). Children's occupational
preferences: The influence of sex and perceptions of occupational
characteristics. Journal of Vocational Behavior,
36, 287-303.
Tognoli, J., Pullen, J., & Lieber, J. (1994). The privilege of
place: Domestic and work locations of characters in children's
books. Children's Environments Quarterly, 11, 272280.
Tremaine, L. S., Schau, C. G., & Busch, J. W. (1982). Children's
occupational sex-typing. Sex Roles, 8, 691-710.
U.S. Employment Service. (1998). Dictionary of occupational
titles (6th ed.). Washington, DC: U.S. Government Printing
Williams, C. L. (Ed.). (1993). Doing "Women'sWork": Men in
nontraditional occupations. Newbury Park, CA: Sage.

  • Recommend Us