Está en la página 1de 26

Cultural Capital, Gender, and School Success: The Role of Habitus

Author(s): Susan A. Dumais


Source: Sociology of Education, Vol. 75, No. 1 (Jan., 2002), pp. 44-68
Published by: American Sociological Association
Stable URL: http://www.jstor.org/stable/3090253
Accessed: 17/02/2009 10:18

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=asa.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to
Sociology of Education.

http://www.jstor.org
I __.

Cultural Capital, Gender, and School


Success:
The Role of Habitus

Susan A. Dumais
HarvardUniversity
Studies of the effects of cultural capital on the educational success of male and
female students have reached coLtiadictory concusions, and few studies have
considered the role that habitus plays in educational outcomes. This article ana-
lyzes the cultural participation of eighth-grade boys and girls and presents a
model that includes a measure for habitus. Through a detailed analysis of cultur-
al practices that have typically been grouped together as a single scale, the author
found that female and higher-SESstudents are more likely to participate in cul-
tural activities. In addition, in both standard ordinary least squares and fixed
school-effects models, she found that cultural capital has a positive, significant
effect on the grades of female students, both with and without controlling for
Bourdieu'snotion of habitus. For male students, the effect is weaker and present
only in the fixed-effects models. Habitus itself has a strong effect for both male
and female students in all models. The author argues that traditional gender
stereotypes play a role in the lack of cultural participation by male students and
that female students may be more encouraged to make use of their cultural cap-
ital to succeed in school.
L~~~~~~~~~ ~ ~ ~ ~ ~ ~ ~
ourdieu (1973:80) wrote: acquire cultural capital, a student must
have the abilityto receive and internalizeit.
By doing awaywith giving explicitly
to everyone what it implicitly Although schools require that students
demandsof everyone,the educationalsys- have this ability, they do not provide it for
tem demandsof everyonealikethat they them; rather, the acquisition of cultural
have what it does not give. This consists capital and consequent access to academic
mainly of linguisticand culturalcompe- rewards depend on the cultural capital
tence and that relationshipof familiarity passed down by the family, which, in turn,
with culturewhich can only be produced is largely dependent on social class.
by familyupbringingwhen it transmitsthe Culturalcapital, then, is comprised of "lin-
dominantculture.
guistic and cultural competence" and a
Bourdieu's theory of social reproduction broad knowledge of culture that belongs
and cultural capital posits that the culture to members of the upper classes and is
of the dominant class is transmitted and found much less frequently among the
rewarded by the educational system. To lower classes. Differences in culturalcapital

Sociology of Education 2002, Vol 75 (January): 44-68 44


Cultural Capital,Gender,
Cultural Capital, Gender,and
and School
School Success
Success 45
45

are reinforced by an educational system that in the same social class, their habitus may be
prefers these styles, leaving most members of quite different, on the basis of their socializa-
the lower classes with little hope of achieving tion and the views they form of the opportu-
social mobility. nity structure available to them. Forexample,
Of all Bourdieu'sconcepts, cultural capital while both male and female students choose
has become the most popular and thorough- business over any other college major,
ly examined by sociologists of education. women are less likely than men to get
Most studies have attempted to determine degrees in engineering, the physical sciences,
whether Bourdieu'snotion of social reproduc- and computer science, all of which have been
tion, based on the intergenerationaltransmis- traditionallyseen as male disciplines (Yupinet
sion of cultural capital, is valid, and, if so, the al. 2000).3 Jacobs (1995) noted that 30 per-
role that the educational system plays. cent of women would have to change their
However, by focusing on cultural capital, college majorsfor women to be distributed in
these studies have ignored another crucial the same manner as men. Moreover, Jacobs
component of Bourdieu's theoretical model (1996) showed that at every level of educa-
of practice: habitus. Habitus, or one's view of tion, women earn less than men; for example,
the world and one's place in it, is an impor- when both women and men have four years
tant consideration in trying to understand of college, women earn .71 of what men
how students navigate their way through the earn. Thus, despite advances for women over
educational system. Studying cultural capital the past few decades, there are many struc-
while ignoring habitus leaves Bourdieu'sthe- tural constraints to women's progress, and
oretical framework incomplete in its practical gender socialization continues to shape girls'
application. It is necessary to consider both lives.
one's resources (capital) and the orientation Evenwith these obstacles, female students
one has toward using those resources (habi- excel in school; they repeat grades less often
tus) to implement the model of practice in do than male students, have higher gradua-
the educational field in the way that Bourdieu tion rates from high school, and are more
intended. likely to attain bachelor's degrees (Jacobs
It is particularlyimportant to consider the 1996). This puzzle was addressed by
functions of both cultural capital and habitus Mickelson(1989), who noted that female stu-
when studying gender differences in school- dents' achievement in school seems anom-
ing. Jacobs (1996) noted that much of the alous because the rewards associated with
mainstream research in the sociology of edu- schooling are still lower than those for male
cation tends to ignore women, even though students. Determining the effects of both
gender is one of the main stratifyingfactors in habitus and cultural capital on students'
society. Bourdieu has been criticized for giv- grades will help sociologists understand this
ing short shriftto factors other than class that problem.
distinguish people from one another (Hall The goals of this article are to join cultural
1992).1 Indeed, many of Bourdieu's works capital and habitus in a model of educational
imply that gender is a secondary characteris- success, as Bourdieu intended; to determine
tic to social class.2 whether cultural capital and habitus play sig-
Recent research has considered other strat- nificant roles in educational success; and,
ifying factors by addressing the role of race, in more specifically,to determine whether one's
addition to class, in relation to culturalcapital gender, in addition to one's socioeconomic
(Farkas1996; Farkaset al. 1990; Kalmijnand status (SES), leads to different benefits from
Kraaykamp1996; Lareau and Horvat 1999). cultural capital and habitus in terms of edu-
Gender, however, has been a neglected area, cational outcomes. The following two sec-
even though boys and girls have different tions set up the theoretical frameworkfor the
experiences in the schooling system that may article. First, Bourdieu's broader theoretical
lead to their different positions in society. framework, including his notions of habitus,
Although one may argue that boys and girls field, and practice, is discussed, and gender is
receive the same cultural training if they are placed within this theoretical framework.
46 46 Dumais
Duma~~~~~~~is

Next, a review of research on cultural capital Habitus is one's disposition, which influences
and school success is presented, with a focus the actions that one takes; it can even be
on research on gender differences. manifested in one's physical demeanor, such
as the way one carries oneself or walks. It is
generated by one's place in the social struc-
CULTURALCAPITALAND ture; by internalizingthe social structure and
one's place in it, one comes to determine
GENDER what is possible and what is not possible for
one's life and develops aspirations and prac-
Bourdieu's General Theoretical tices accordingly. This internalization takes
Framework place during early childhood and is a primar-
ily unconscious process.5
Although culturalcapitalis an importantpartof The consequences of the development of
Bourdieu'stheory of social reproduction, it is habitus are large: Bourdieu argued that the
only one component of his theoretical frame- reproduction of the social structure results
work. In Distinction,Bourdieu(1984) noted that from the habitus of individuals. On the basis
capital, habitus, and field all work together to of the class position they were born into, peo-
generate practice,or social action. ple develop ideas about their individual
The field, the setting in which practices potential; for example, those in the working
take place, is "a network, or a configuration, class tend to believe that they will remain in
of objective relations between positions" the working class. These beliefs are then
(Bourdieu and Wacquant 1992:97). Fieldsare externalized into actions that lead to the
spaces in which dominant and subordinate reproduction of the class structure.
groups struggle for control over resources; Overall,then, one's practices or actions are
each field is related to one or more types of the result of one's habitus and capital within
capital. Indeed, Bourdieu and Wacquant a given field. In terms of schooling, Bourdieu
argued that capital does not exist or function and Wacquant (1992) described the school
except in relation to a field. system as a field. Doing one's schoolwork and
Culturalcapital is one of several forms of attempting to get good grades are kinds of
capital that Bourdieu(1997) described. Along practices in this field. Within the educational
with economic, social, and symbolic capital, field, the most valuable form of capital is cul-
culturalcapitalserves as a power resource,or a tural capital: "academic success is directly
way for groups to remaindominant or gain sta- dependent upon cultural capital and on the
tus. Bourdieudistinguishedamong three forms inclination to invest in the academic market"
of cultural capital: objectified cultural capital, (Bourdieu 1973:96). Both Bourdieu (1984)
which refersto objects that requirespecial cul- and DiMaggio and Useem (1978) found that
turalabilitiesto appreciate,such as worksof art; within the dominant classes, teachers have
institutionalizedculturalcapital,which refersto the most culturalcapital, value it, and tend to
educational credentials and the credentialing rewardstudents who possess it. Childrenwho
system; and embodied culturalcapital,which is have more cultural capital (having been
the disposition to appreciate and understand exposed to it from birth in their upper mid-
culturalgoods. It is this third form that most dle- and upper-class families) feel more com-
researchershave tried to operationalizein their fortable in school, communicate easily with
studies by showing students' interest in music teachers, and are therefore more likely to do
or art. Institutionalizedculturalcapitaldevelops well in school (De Graaf, De Graaf, and
as a result of one's having embodied cultural Kraaykamp2000). Lower-class students, on
capital and successfully converting it via the the other hand, find the school environment
educational system. To appropriate and use different from their home environment and
objectified culturalcapital, one needs embod- lack the capital necessary to fit in as well as
ied culturalcapital (Bourdieu1997).4 the higher-SESstudents.6
In addition to capital and field, another Habitus also plays a large role in students'
important concept for Bourdieu is habitus. success in school, but with few exceptions
Cultural Capital, Gender, and C C
School Success 47

(McClelland 1990; Reay 1995), it has been In every relationshipbetween educational


largely ignored by past research in the sociol- capital and a given practice,one sees the
effectsof the dispositionsassociatedwithgen-
ogy of education.7 Students' decisions to
invest in their education, study hard, and go der which help to determinethe logic of the
reconversionof inheritedcapitalinto educa-
to college depend on students' place in the
tionalcapital,that is, the "choice"of the type
class system and their expectations of of educationalcapitalwhichwill be obtained
whether people from that class tend to be fromthe same initialcapital,moreoften liter-
successful academically (Swartz 1997). aryfor girls,moreoften scientificfor boys.
Bourdieu (1973) argued that one's habitus
In other words, one's habitus, determined
develops in relation to how much cultural
capital one has; a person from the lower class by the availableopportunity structureor field,
is aware that people from that class tend to shapes the type of class-based capital that
have little cultural capital and that without men and women have, resulting in gendered
cultural capital, they are unlikely to succeed forms of cultural capital (Laberge 1995).
Moreover, one's habitus is also gendered as a
educationally. Therefore, lower-class students
tend to self-select themselves out of the col- result of the possibilities available to each
lege-going track on the basis of their views of group. In her study of the occupational
what is possible and what is not. On the other expectations of men and women, McClelland
hand, exceptional students from the lower (1990) stated that men who have profession-
class may see the accumulation of cultural al ambitions are following their habitus, while
women with the same ambitions are violating
capital as a way to overcome the obstacles what their traditional habitus may be dictat-
that are typical for those in their class posi-
tion. ing. Thus, women may be less likelyto aspire
to do well in school or to go far in school.
The foregoing description addressed class
Robinson and Garnier(1985) also noted that
differences in cultural capital, habitus, and
while men and women may begin their lives
schooling. Gender differences, however, are with similar class positions and cultural capi-
also relevant. Bourdieu (1984:107) stated
tal, the social reproduction process functions
that "sexual properties are as inseparable in such a way that women attain less-privi-
from class properties as the yellowness of a
leged positions than men do.
lemon is from its acidity." As one moves up In the United States, many educational
the social hierarchy,he noted, there are fewer
inequalities between men and women seem
differences between men and women, and to be disappearing. However, as noted at the
the division of labor is not as strong. Part of
beginning of this article, women are still at a
the traditionaldivision of labor is that women
disadvantage. For example, even when they
are the ones who are knowledgeable about have majored in the same subject in college,
art and literature;in the upper classes, how- women do not fare as well as men in the
ever, both men and women are educated in labor market; the median starting salary for
these matters. female business majorsis $4,000 less than it is
According to Bourdieu (1984), cultural for male business majors (Yupin et al. 2000).
capital is important for women for two rea- Mickelson (1989) offered four possible
sons: Women make use of the capital for hypotheses to explain why female students
acquiring husbands, and women play the key would continue to do well in school despite
role of transmitting the cultural capital to these obstacles. The differential reference-
their children. Men are more inclined to use group hypothesis states that although female
cultural capital for educational qualifications students are aware of male advantage, they
and for getting jobs. These different uses of base their successes on the successes of other
cultural capital imply that social actions take women. The Pollyanna hypothesis states that
place in different fields for men and women, girls assume that gender inequality is a thing
with different forms of capital serving as cur- of the past, while the social powerlessness
rency. Bourdieu (1984:105) hinted at this hypothesis states that girls assume the oppo-
possibility: site, believing that education is the only way
48
48 Dumais
Dumais

to attain high-status husbands. The sex-role inition, one can conceive of culturalcapital as
socialization hypothesis argues that girls do arbitrary, something established by one
well in school because they have been social- group to prevent another from gaining
ized to be good and follow orders. Each power. With regard to education, the capital
hypothesis depicts a different habitus that itself does not alter the students in any mean-
young women may possess. Empirical ingful way but, rather, serves as a "signal"to
research is needed to determine how habitus, teachers (who are predominantly middle class
in addition to capital, may affect the perfor- and tend to have a high regard for culture)
mance of female students. First,however, it is that they will relate better to these students
useful to review the research that has already and should promote them (DiMaggio 1982,
been conducted on cultural capital and Farkaset al. 1990).
schooling. It is also reasonable to argue that cultural
capital increases a student's competence or
Past Research on Cultural Capital ability to the point where the student is more
successful in school without preferentialtreat-
As mentioned earlier,the concept of habitus ment by teachers. This argument is promoted
has been omitted from nearly all studies of by the National Endowment for the Arts(NEA
culturalcapital and school outcomes; cultural Fact Sheet 2001), which noted that students
capital, on the other hand, has been studied who have studied the arts for four years score
extensively. Despite the heavy focus on cul- an average of 89 points higher on the SAT
tural capital, no real consensus has been than do students who have not. It is also
reached among educational researchers apparent in Shaw's (1999) notion of the
regarding an operationalization of cultural "Mozart effect," the belief that listening to
capital. Lamontand Lareau(1988) noted that classical music leads children to become more
even Bourdieu,who coined the term, was not intelligent. Studies that have focused on such
entirely clear about its meaning; thus, cultur- activities as music lessons or reading
al capital has been operationalized in a num- (Aschaffenberg and Maas 1997; De Graaf et
ber of different ways. Mohr and DiMaggio al. 2000) are examples of this interpretation
(1995:168) gave the following definition: "By of cultural capital.
culturalcapital, we refer to prestigious tastes, Finally,some studies have used measures
objects, or styles validated by centers of cul- that seem more obviously reflective of SES,
tural authority, which maintain and dissemi- such as whether the parents provide a specif-
nate societal standards of value and serve col- ic place to study (Teachman 1987), or
lectively to clarify and periodically revise the noncognitive traits, such as the way one car-
cultural currency." This definition is vague ries himself or herself and communicates with
enough to allow a wide number of interpre- others (Farkaset al. 1990). These measures
tations, all of which connect to Bourdieu's support the argument that higher-SES stu-
concept on at least some level. dents do better in school, but do not really
The default has been to define cultural address the process of social exclusion
capital according to the types of variables described in the first definition of cultural
availablein the data set that is being analyzed capital presented earlier.
and the researchers'ideas about how cultural There have also been conflicting conclu-
capital functions. The result is three main cat- sions regarding the effects of cultural capital
egories in which cultural capital has been on educational outcomes, in whatever form it
defined. The first and most common is to use has been operationalized. Some have found
variables that indicate participation in cultur- strong support for a cultural capital effect on
al activities, such as visiting museums or dependent variables like grades and years of
going to the ballet; some studies have education attained (DiMaggio 1982;
focused on these activities at the student level DiMaggio and Mohr 1985; Farkaset al. 1990;
(Katsillisand Rubinson 1990), while others Kalmijnand Kraaykamp1996). Others have
have focused on the parents' activities found that particulartypes of activities, such
(Kalmijnand Kraaykamp1996). With this def- as taking long-term lessons or reading, matter
Cultural Capital,Gender,
Cultural Capital, Gender,and
and School
SchoolSuccess
Success 49

more than others for these outcomes any family background variables (including
(Aschaffenberg and Maas 1997; De Graaf et father's occupation, parental education,
al. 2000). Still others have found no support income, and household cultural climate).
for a cultural capital effect (Katsillis and Aschaffenburg and Maas (1997) used data
Rubinson 1990; Robinson and Garnier1985). from the Survey of Public Participationin the
In the literature,then, there is no consen- Artsfrom 1982, 1985, and 1992, focusing on
sus on what culturalcapital means, whether it training in high cultural activities. Although
has an effect, and what that effect is. they noted that females are more likely than
Furthermore, the literature tends to ignore males to participate in cultural classes, they
group differences, particularlygender differ- found, unlike DiMaggio (1982), that with
ences. Much of the research has included regard to making educational transitions, cul-
gender as a control variable but has not dis- tural capital benefits male and female stu-
cussed it in any detail (see De Graaf et al. dents equally. They suggested that the
2000; Lareau1989; Lareauand Horvat 1999; greater number of educational opportunities
Roscigno and Ainsworth-Darnell 1999). In that exist for girls today may be one reason
studies that have discussed gender, the find- why culturalcapital is no longer more impor-
ings have not been consistent. For example, tant for them than it is for boys, as it was in
Katsillisand Rubinson(1990) did not find cul- the early 1960s. However, since they used a
tural capital effects for either sex, but different dependent variable than did
Teachman (1987) found that cultural capital DiMaggio, it is difficult to say whether the
has a positive and significant effect for the gender differences in culturalcapital have dis-
educational attainment of females but not for appeared.
males. Thus, we are left with inconsistent conclu-
The most thorough analyses of the rela- sions in the broader study of cultural capital,
tionship between gender and cultural capital particularlywhen considering gender. A more
to date have been by DiMaggio and Mohr complete empirical consideration of
(DiMaggio 1982; DiMaggio and Mohr 1985; Bourdieu's theoretical framework may help
Mohr and DiMaggio 1995). Using data from sociologists understand more fully the roles of
ProjectTalent (a large data set of high school cultural capital and habitus in educational
students from 1960), DiMaggio (1982) found outcomes and how these roles may vary by
that male students have low levels of interest gender.
in cultural activities, while female students
have moderately higher levels. DiMaggio also
found that female students from high-status ANALYSIS
families have the greatest return from having
cultural capital, a phenomenon he labeled Research on cultural capital and school suc-
the "cultural reproduction" model, but that cess has been inconclusive. It is not clear
male students tend to follow the "cultural whether cultural capital actually affects edu-
mobility" model, benefiting more from cul- cational outcomes and whether and why the
tural capital if they are from lower-status process of converting cultural capital into
groups. DiMaggio's explanation of the gen- educational success may differ for male and
der difference is that high-status females are female students. Furthermore, Bourdieu's
culturally prescribed into the activities that concept of habitus has not been included in
would give them cultural capital, while high- research on cultural capital and school suc-
status males tend to rebel against their par- cess, leading to an incomplete picture of how
ents' cultural values; he further argued that Bourdieu'smodel functions.
women may use cultural capital as a way to In my study, I attempted to improve on
attract high-status men. Using the same data the previous research in three ways. First,
set, Mohr and DiMaggio (1995) analyzed the most previous studies have created a cultural
intergenerationaltransmissionof culturalcap- capital index, rather than considered individ-
ital and found that gender has a greater effect ual activities separately, so it is not clear
on the possession of cultural capital than do whether one activity is more relevant than
50 Dumais
I
others. In my study, I analyzed participationmy main interest, I restricted the sample to
rates in several cultural activities by gender
white eighth graders; thus, further research
and SES before I created a cultural capital on the interaction of race and gender is nec-
variable, to determine if any specific activity is
essary.A sample weight was used throughout
preferred by higher-SESstudents or by boys the analyses to compensate for unequal prob-
or girls. abilities of selection into the sample and for
Second, most research on cultural capitalnonresponse.9
and schooling has not addressed differences The data collected for Project Talent and
that may exist between schools. Schools in used by DiMaggio (1982) contain one of the
rural regions or poor districts may not have most comprehensive sets of measures for cul-
the same access to cultural resources as do tural capital. While the cultural variables in
schools in other areas. Through the use of NELSwere clearlycollected with the intention
pooled within-school regression estimates, I of measuring cultural capital, they are not as
comprehensive as those of ProjectTalent. An
separated individual effects from school-level
resource-deprivationeffects. ideal set of measures for cultural capital may
include such things as the student's eloquence
Most important, by including a variablefor
students' occupational expectations, I opera-and his or her comfort in the school environ-
ment. However, the NELSdata set, like many
tionalized the concept of habitus to develop a
more complete model of Bourdieu's theory other data sets that have been used in studies
than has been evident in previous research. of culturalcapital,does not include such a wide
Habitus should have a strong and positive set of variables for cultural capital. It is true,
effect on the grades of both boys and girls. however, that the measures mentioned could
The effect of habitus may be stronger for also be seen as components of one's SES,which
male students because (according to gender- the NELSdata set has a measure of. Therefore,
socializationtheory) success and achievement by using a limited definition of cultural capi-
tal-participation in the arts-I make a stronger
should be part of their worldview, and achiev-
ing good grades can be seen as a step in thattest for whether there is actuallyan effect.
direction. For female students, the connec- I constructed culturalcapital variablesfrom
questions in the NELSparents' questionnaire.
tion is not as simple; as Mickelson(1989) out-
lined, habitus may have a strong effect for aOne set of questions asked: "Do you or your
number of reasons, ranging from a desire to eighth grader take part in any of the follow-
attract men to a belief that inequality for ing activities?" with the activities listed as
women no longer exists. "borrowing books from the public library,"
If cultural capital is as important as "attending concerts or other musical events,"
Bourdieu (1973) argued it is for educational and "going to art museums." Another set of
outcomes, then a cultural capital effect questions was related to whether students
should be present even after habitus is addedtook lessons in high cultural activities: "Has
to the model, and it should be present for your eighth grader ever taken classes outside
of school in one of the following activities?"
both boys and girls. If the effect is bigger for
with the activities listed as "artclasses outside
one group or the other, it will be important to
explain how and why this difference occurs of school," "music classes outside of school,"
and whether new indicators should be devel- and "dance classes outside of school."10
oped that take gender differences into The total number of students' cultural
account. activities asked about in NELS,then, is six. In
some cases, these activities are studied sepa-
Data and Variables rately in the following analyses. Culturalcapi-
tal in this study is simply the sum of the num-
The data I used were from the first panel of ber of activities in which the student partici-
the National Education Longitudinal Study pates. For each student, the cultural capital
(NELS),a survey based on a nationally repre- variable can range from zero (participating in
sentative sample of 24,599 eighth-grade no activities) to six (participating in all the
respondents in 1988.8 Because gender was activities listed).
CulturalCapital,Gender,and SchoolSuccess Si
51

It is extremely difficult to represent one's Table 1 shows the means and standard
habitus, or worldview, in a single variable, or deviations for the main variables. A correla-
even a large set of variables. However, one tion matrix of all the variables used in this
component of habitus is one's beliefs about study is presented in the appendix.12
the future. McClelland (1990) operational-
ized habitus as students' occupational aspira-
tions, particularlywhether or not they aspired RESULTS
to upper-white-collar jobs. I do the same
here, with the dummy variable students'
expectations representing whether or not the CulturalParticipation and
student said that he or she expected to have Occupational Expectations
one of the following occupations at age 30:
professional, managerial, or business; busi- Table 2 shows the cultural participation levels
ness owner; or science or engineering.l1 of the boys and girls in general and by SES.
The variable SES is a composite measure Overall, the girls outnumbered the boys in
developed by NELS,using data from the par- every activity, with differences as great as 30
ent's questionnaire (or, if missing, from the percentage points (dance lessons) and as
student's questionnaire), incorporating moth- small as 1 percentage point (art lessons). For
er's and father's educational levels, mother's both the boys and girls, the most popular
and father's occupations, and family income. activity was going to the library for non-
For ability, the standardized scores from the school-related reading: 80 percent of the
cognitive tests in mathematics, reading, sci- boys and 85 percent of the girls did this activ-
ence, and history that were given to the base- ity. Going to concerts and art museums were
year students were averaged. The variable the next most popular activities for both the
GPA(grade point average) is an average of boys and the girls. It appears, then, that one-
the self-reported grades for each student in time activities were more popular than
English, mathematics, science, and social lessons, which involve more of a time com-
studies on a scale from .5 (mostly below D) to mitment. Nevertheless, more than a third of
4 (mostly A's), thus approximating the stan- the girls took music or dance lessons, and 47
dard 0-4.0 GPA system that many schools percent of them took at least one kind of les-
use. Students were asked to report their son. The male students were much less likely
grades from the sixth grade until the time of to participate in these activities, with only 25
the survey. percent taking any kind of lesson.
Table 1. Means and Standard Deviations for VariablesUsed in the Analyses

Variable Mean SD

Sex (0=male, 1=female) .499 .500


SES .010 .715
GPA 2.936 .761
Ability
(averageof 4 test scores) 51.735 8.711
Student'soccupationalexpectations
(0 = non-upperwhite collar,1 = upper white collar) .411 .492
Takeart lessons (0 = no, 1 = yes) .081 .273
Takemusic lessons (0 = no, 1 = yes) .276 .447
Takedance lessons (0 = no, 1 = yes) .182 .386
Go to the public library(0 = no, 1 = yes) .824 .381
Go to concerts (0 = no, 1 = yes) .672 .470
Visitart museums(0 = no, 1 = yes) .389 .488
Culturalcapital(range = 0-6) 2.228 1.448
Participatein sports(0 = no, 1 = yes) .706 .455

Note: The sample size variedby the question asked, rangingfrom 14,341 to 16,321.
Source:NationalEducationLongitudinalStudy,1988.
52
52 Dumais
Dumals

Table 2. Percentage of Eighth Graders Participatingin Different CulturalActivities, by Gender and SES

Activity AllBoys AllGirls

Artlessons 7.51 8.73


Musiclessons 22.06 33.21
Dance lessons 3.37 32.93
Any type of lesson 24.52 46.62
Libraryvisits 79.62 85.23
Concerts 60.31 74.07
Artmuseums 37.28 40.54
Mean numberof activities 1.94 2.52

Activity Low-SESBoys Low-SESGirls

Artlessons 3.20 3.01


Musiclessons 8.91 14.16
Dance lessons 1.79 14.10
Anytype of lesson 11.04 21.90
Libraryvisits 67.59 74.66
Concerts 46.75 61.17
Artmuseums 24.62 25.00
Mean numberof activities 1.36 1.72

Activity High-SESBoys High-SESGirls

Artlessons 16.28 18.35


Musiclessons 40.82 58.89
Dance lessons 6.47 54.99
Anytype of lesson 44.54 73.13
Libraryvisits 90.77 93.52
Concerts 75.15 86.22
Artmuseums 53.38 61.35
Mean numberof activities 2.63 3.45

Note: The size of the sample variedin responseto each question. The N for all boys rangesfrom 7,160 to
7,488; for low-SESboys, from 1,726 to 1,765; and for high-SESboys, from 1,785 to 1,912. Forall girls,the N
rangesfrom 7,112 to 7,464; for low-SESgirls,from 1,826 to 1,878; and for high-SESgirls,from 1,726 to 1,836.
Source:NationalEducationLongitudinalStudy,1988.

The patterns are slightly different when lessons and attending concerts-both of
SES is taken into consideration. In this case, which had a 12 percentage point difference.
low-SESstudents were defined as those in the As was true for all eighth graders, the low-SES
bottom quartile, and high-SESstudents were students participated in onetime activities
those in the top quartile. (As a point of refer- more frequently than in lessons, with 11 per-
ence, in the lowest SES quartile, the median cent of the boys and 22 percent of the girls
years of father's education were 12, and the participating in any type of lesson.
median family income was $17,500; in the In the highest SES quartile, there was a
highest SES quartile, the median years of strikinggap in the number of boys versus girls
father's education were 18, and the median who took dance lessons: 55 percent of the
family income was $62,500).13 In the low-SES girls but only 7 percent of the boys. There
group, the male-female differences were not was also a sizable gap between boys and girls
as large as they were for eighth graders over- who took music lessons. Seventy-five percent
all. Indeed, for two activities (art lessons and of the girls, compared to 45 percent of the
visiting art museums), the level of male and boys, took any type of lesson.
female participation was nearly equal. The There was also a gender difference for
largest gap between male and female partici- overall, aggregate levels of cultural participa-
pation occurred for two activities-dance tion. For the eighth graders as a whole, the
CulturalCapital,
Cultural Capital,Gender, and School
Gender, and SchoolSuccess
Success 53
53

male students were more likely to participate able, be willing and able to pay for these
in 0, 1, or 2 activities, while the female stu- activities, and make the time commitment
dents were more likely to participate in 3 or necessary to maintain such long-standing cul-
more activities. The average number of activ- tural training. All these circumstances are
ities for the boys and girls were 1.94 and more likelyto occur in high-SESfamilies.
2.52, respectively. In the low-SES group, the Compared to finding reasons for SESdiffer-
girls were still more likely than the boys to ences in cultural participation, however, it is
participate in more activities, but the differ- harderto explain why males and females with
ence was not as strong; the average number the same amount of economic resources
of activities for the boys and girls were 1.36 would participate at such different levels. It
and 1.72, respectively. In the high-SESgroup, may be that parents, who have been social-
the gap was bigger, with the girls, on aver- ized to believe that certain activities are more
age, participating in 3.45 activities, and the appropriate for one gender than the other,
boys participating in 2.63. The high-SESgirls encourage their daughters to participate in
(8 percent) were also more likely than the cultural activities but do not put the same
high-SESboys (2 percent) to participate in all pressureon their sons; this argument is similar
six activities. to the one made by DiMaggio (1982) on the
In comparing the high-SES and low-SES basis of his findings with the Project Talent
groups, it is clear that the SESgap was greater data. As Mohr and DiMaggio (1995) contend-
than the gender gap for cultural participa- ed, it is most often the mother who passes
tion. For example, high-SES boys and girls down culturaleducation to her children;if the
were more than twice as likely to visit art mother has been socialized to believe that cul-
museums than were their low-SES counter- tural activities are in the female sphere, she
parts. Within the SES groups, however, the may be more inclined to teach her daughter,
difference between boys' and girls' art muse- not her son. Furthermore, a number of
um attendance was 0 percentage points for researchers (see, e.g., Sadker and Sadker
low-SESstudents and 8 percentage points for 1994; Thorne 1993) have found that despite
high-SES students. Even going to the public the women's movement, strong gender
library,the most popular activity for all stu- stereotypes still exist among children and per-
dents, was far less likely to occur for low-SES vade the school environment. Particularlydur-
students than for high-SESstudents. ing puberty and earlyyoung adulthood, males
From Table 2, one can conclude that rates may have a heightened sensitivity to main-
of cultural participation vary by gender as taining gender boundaries and a fear of get-
well as by SES.14Girls and high-SESchildren ting involved in anything that would be con-
were both more likelyto participate in each of sidered feminine. The extremely low propor-
the activities taken singly and had higher tion of male students in this study who took
aggregate rates of participation. In asking dance lessons provides some evidence for this
why SES is so important, one must consider speculation.
the role that parents play in a child's cultural On the other hand, participation in sports,
participation.15In many cases, it is the par- a more traditionallymasculine activity, is held
ents who hold the key to children's cultural in high regard by male students; Sadker and
participation by paying for lessons; providing Sadker(1994) found that middle school boys'
transportation to and from classes; or, as sports were the most prestigious school-spon-
Bourdieu (1984) argued, by demonstrating sored activity. In the present sample of eighth
an interest in culture to the children at home. graders, 53 percent of the boys participated
This is particularlythe case for the children in in school varsity sports (defined as sports
this study, who were only in the eighth grade played against teams from other schools), 46
and unable to drive to classes or cultural percent of the boys participated in intramural
places. Since these activities all take place out- sports (defined as sports played against a
side the structured school system, parents team in the same school), and 46 percent of
need to take the initiative. That is, they need the boys participated in non-school team
to be aware of the potential activities avail- sports, indicating that sports activities are
54 Dumais

more popular than all cultural activities though there is some support for Mickelson's
except going to the libraryand to concerts. (1989) Pollyanna hypothesis: Despite being
Conversely,44 percent of the girls participat- more encouraged than their male counter-
ed in varsity sports, 39 percent of the girls parts to participate in cultural activities,
participated in intramuralsports, and 33 per- female students' perceptions of the opportu-
cent of the girls participated in non-school nities available to them and of the roles they
team sports, indicating that the girls were just are expected to fill in society have changed
as likely to play sports as they were to take for the better over time. How this habitus,
music or dance lessons and more likely to along with cultural capital, factors into edu-
play sports than to take art lessons. As others cational success is analyzed next.
have documented (Lorber 1994; Thorne
1993), the stigma of being called a "tomboy" CulturalParticipation,
is not as strong as that of being called a
Gender,and Grades
"sissy."
Are these SES and gender differences as In the previous section, I reported that cultur-
strong when a component of students' habi- al participation and habitus vary by gender
tus-aspirations for a white-collar job-is and SESin interpretableways. In this section,
considered? Table 3 shows that a larger per- I test whether arts participation is a form of
centage of the girls (47 percent) than the capital that gives students educational advan-
boys (35 percent) overall aspired to white- tages and whether these advantages vary by
collar jobs and that the difference was partic- gender. In addition, I test whether one com-
ularly striking within the lowest SESquartile, ponent of habitus-students' beliefs about
where only 20 percent of the boys and 36 their abilityto obtain prestigious jobs-has an
percent of the girls had upper-white-collar effect on their grades.
occupational aspirations.The gender gap was Bourdieu (1973) contended that the
not as large within the highest SES quartile, school system rewards students who have
where 56 percent of the boys and 59 percent cultural capital; the ultimate reward is in the
of girls had such aspirations. As was true for form of educational credentials (institutional-
the rates of cultural participation, it appears ized cultural capital), but along the way, stu-
that differences in SEShave a stronger impact dents with higher cultural capital receive bet-
than do differences between boys and girls, ter grades and more attention and feedback.
but that gender differences are still present. Following DiMaggio's (1982) argument that
Although the finding regarding SESdiffer- teachers favor students with more cultural
ences in aspirations is similar to that of capital and thus give them higher grades, the
McClelland's study of students from the dependent variable in the following analyses
National Longitudinal Survey of the High was the students' GPAs.16It is true that girls in
School Class of 1972 (1990), the finding every SES group have higher GPAsthan do
regarding gender is different: McClelland boys and that students in the higher SES
found that boys, who made up only half the group have higher GPAs than those in the
sample, constituted three-fourths of the lower group. The question is, How much of
group who had high expectations. It seems as this difference is due to cultural capital and
Table 3. Eighth GradersWho Wanted a ProfessionalJob at Age 30, by Gender and SES

SES Boys Girls

% N % N

All 35.27 7,882 46.98 7,984


Low-SES 20.06 1,891 36.00 2,031
High-SES 56.14 2,031 59.36 1,982

Source:NationalEducationLongitudinalStudy,1988.
CulturalCapital,
Cultural Capital,Gender,
Gender,and
and School
SchoolSuccess
Success 55
55

habitus? (See the correlation matrices in the dents' habitus-the desire for an upper-
appendix for the correlations between GPA, white-collar profession-also has a positive
SES, and cultural participation for boys and and significant effect, second only to ability.
girls.) Model 3 is presented in the next column
Separate models were run by gender.17 with metric coefficients. As can be seen, if a
Standardized variables for SES, cognitive skill male student has high occupational aspira-
(the average of four tests in reading, math, tions, his GPA increases by .196 points.
history, and science), and the number of cul- Having a habitus that foresees success in a
tural activities were used. Robust standard career in one's future is clearly influential on
errorswere used to adjust for the clustering of one's grades. In contrast, if a male student
students within schools (Huber 1967), and, as participates in a cultural activity, his GPA
mentioned earlier,a sample weight was used. increases by only .013 points, an insignificant
Table 4 shows the results of the ordinary amount. Overall, then, for boys, no isolated
least squares (OLS) regressions on GPA for cultural capital effect on grades was found
boys and girls.18 In the first model, control after controlling for SES,ability,and the desire
variables that have been used in previous for an upper-white-collarprofession.
studies of cultural capital (DiMaggio 1982; For girls, the metric coefficient for cultural
Katsillisand Rubinson 1990)-SES and abili- capital is .028. This means that for each addi-
ty-were run alone. One sees that ability is an tional cultural activity a student participates
important factor for the grades of both boys in, her GPAwill increase by .03. Despite the
and girls; it is three times larger than SESfor statistical significance of the result, this is
boys and more than five times largerthan SES actually not a particularlylarge increase. Fora
for girls. Clearly,ability has a strong associa- student's grade to change from a B to a B+,
tion with students' grades. In Model 2, the for example, a change in GPA of .3, or 10
cultural capital variable was included along times that of the effect of cultural capital,
with SES and ability. While cultural capital is would be needed. In other words, despite the
significant for boys, it is only about one-fifth appearance of a cultural capital effect for
the size of the SEScoefficient. Ability remains female students, the actual change in grades
the strongest of the independent variables,at is minimal, and to make the claim that cul-
more than 10 times the size of the cultural tural capital is critical for academic success
capital variable. These results are different would certainly be an overstatement. As is
from those of DiMaggio (1982), who found mentioned in note 17, although the models
that the culturalcapital variable had an effect in Table 4 indicate that culturalcapital has an
of the same order of magnitude as ability.The effect on female students'-but not on male
result for girls in Model 2 differs from the students'-grades, I cannot reject the null
regression that was run for boys. Although hypothesis that there is no difference in the
the culturalcapital coefficient is small in com- effects of cultural capital on grades for male
parison to ability (only about one-ninth the and female students.
size), it is relativelysimilar in size to SES,and Clearly, the more significant relationship
it is still statisticallysignificant.19 for both boys and girls is between habitus (in
In Model 3, the habitus variable, reflecting the form of occupational expectations) and
whether a student had high occupational grades; regardless of gender, high expecta-
aspirations, was added. For boys, the cultural tions lead to higher grades. A student's
capital variable is reduced even further and expectations are developed from what he or
rendered insignificant. For girls, the result is she has experienced in the past and believes
quite different from the one for boys: While is likely to happen for people from his or her
ability is certainly the largest of the variables particular background. For girls, a student's
(nearly 13 times as large as cultural capital), disposition or attitude about whether she will
cultural capital continues to have a separate have a professional job outweighs whether
association with students' grades and one she participates in culturalactivities. Forboys,
that is almost the same size as SES.20The vari- although participating in cultural activities
able that represents a component of the stu- does not appear to have an effect on grades,
Table 4. OLSEstimatesof the Effectsof CulturalCapital on GPA

Model 1 Model 2 Model 3

Males Females Males Females Males Females

Cultural .021* .045*** .019 .042**


Capital (.010) (.009) (.010) (.009)

SES .122*** .073*** .116*** .057*** .096*** .049**


(.010) (.010) (.010) (.010) (.010) (.010)

Ability .378*** .421*** .376*** .41 3*** .358*** .398**


(.009) (.009) (.009) (.009) (.009) (.010)

Student's
Occupational .196*** .145**
Expectations (.018) (.016)

Intercept 2.888*** 3.062*** 2.892*** 3.051 *** 2.820*** 2.982**


(.010) (.011) (.010) (.011) (.013) (.014)

R2 .323 .344 .324 .347 .336 .356

N 7,738 7,776 7,738 7,776 7,503 7,622

Note: For models 1-3, the variables for cultural capital, SES, and ability were standardized. Sample weights and robust standard er
*= ** = *** =
significant at .05, significant at .01, significant at .001.
Source:NationalEducationLongitudinalStudy,1988.
Cultural
Cultural Capital, Gender,and
Capital, Gender, SchoolSuccess
and School Success 57
57

a student's disposition toward obtaining a In Model 1, the patterns are the same for
prestigious career does. Swartz (1997) noted both boys and girls as in the OLS models,
that in large differentiated societies like the although the coefficients are slightly larger
United States, where there is not as strong a than the coefficients from the OLSmodels. In
dominant culture as there is in France,cultur- Model 2, when cultural capital is introduced,
al capital (when defined as knowledge of and it is about one-seventeenth of the size of abil-
participation in highbrow artistic activities) ity for boys and one-seventh the size of abili-
may not be as useful a concept. In the United ty for girls. For both boys and girls, cultural
States, then, it may not be so much whether capital, ability, and SES all have a positive
one participates in cultural activities, but impact on students' grades. In Model 3 for
whether one has the habitus that leads one to boys, cultural capital has a significant effect
expect an upper-white-collar career, that even when SES,ability, and students' expec-
affects educational success and, in the case of tations are controlled for. This finding con-
social class, perpetuates the existing stratifica- trasts with that for the OLS model, in which
tion structure. the cultural capital variable was not signifi-
cant. Therefore, after controlling for school-
Fixed-Effects Models level differences, culturalcapital does have an
impact on male students' grades. Still, it
The results of the OLS regressions should be should be noted that the level of significance
interpreted with caution. For example, the for culturalcapital is not as high as it is for the
regressions did not control for the availability other variables and that the cultural capital
of cultural resources by urban or rural area, coefficient is actually not that much larger
nor did they account for other differences than it was in the OLS models, where it was
between schools. Some schools may grade not significant. For girls, the cultural capital
more easily than others, and mean SESvaries variable is largerthan it was in the OLSmodel
by school. Many studies on the effect of cul- and much larger than the coefficient in the
tural capital on schooling have failed to take male regressions. SESis larger relative to cul-
account of these differences (DiMaggio tural capital than it was in the OLS model,
1982) or have controlled for only a few of the and the coefficient for students' occupational
variables that may be different across schools expectations is slightly smaller than it was in
(Roscigno and Ainsworth-Darnell1999). the OLS model. Interpretingthe metric coef-
It is important to control for school-level ficients for Model 3, one sees that even par-
variables. For example, although an SESvari- ticipating in all the cultural activities would
able is included in the OLSregression,this vari- not raise a male student's GPAby a full grade.
able controls only for the SES of the student, According to the metric coefficients for girls,
not that of the school. Schools from higher SES the actual effect of cultural capital is also not
regions may be more competitive than schools that large; a female student's grade would
from lower SES regions and may thus have not benefit enormously from participating in
higher grading standards; this difference is an additional cultural activity. A combination
ignored in the standardOLSregressionbecause of high expectations (habitus) and high cul-
a school-level SESvariableis omitted. Because tural participation, however, would have an
within-school SESis heterogeneous, an individ- impact on students' grades.
ual's SESmay not adequately measure school- There are two additional items that should
level SES.To account for these variationsacross be noted. First,the variable for habitus-stu-
schools, then, a fixed-effects model is used to dents' occupational expectations-remains at
rule out school-level differences (for an expla- about the same strength as it did in the OLS
nation of fixed-effect equations, see Aaronson regression, indicating that one's habitus con-
1998; Greene 2000). The resultsof these mod- tinues to be an important factor in determin-
els control for differences between schools, ing one's GPA. Second, the SES variable is
such as region, SES, and the percentage of larger in fixed-effects Model 3 than it was in
minoritystudents. Table 5 presents the results OLS Model 3. In other words, after control-
for boys and girls.21 ling for school-level SES,one's personal SESis
Table 5. Pooled Within-schoolRegression Estimates of the Effects of CulturalCapital on GPA

Model 1 Model 2 Model 3

Males Females Males Females Males Females

CulturalCapital .027* .059*** .025* .057**


(.011) (.010) (.011) (.009)

SES .153*** .105*** .146*** .085*** .127*** .078**


(.012) (.011) (.012) (.011) (.012) (.012)

Ability .396*** .433*** .393*** .424*** .376*** .411**


(.010) (.011) (.010) (.011) (.010) (.011)

Student'sOccupational .194*** .130**


Expectations (.020) (.017)

Intercept 2.892*** 3.067*** 2.897*** 3.053*** 2.824*** 2.990**


(.001) (.001) (.002) (.003) (.008) (.009)

N 7,738 7,776 7,738 7,776 7,503 7,622

Note: Formodels 1-3, the variablesfor culturalcapital,SES,and abilitywere standardized.Sampleweights and robuststandarder
*= ** =
significant at .05, significant at .01, *** = significant at .001.
Source:NationalEducationLongitudinalStudy,1988.
Cultural Capital,Gender,
Cultural Capital, Gender,and
and School Success
School Success 59

extremely important for influencing one's boys and girls, aspiring to an upper-white-
grades, with lower-SES students being at a collar job-the measure for habitus used
disadvantage. here-remains significant.
When between-school differences are con-
trolled, then, the result for cultural participa-
tion is significant for boys, but not to the DISCUSSION
degree to which it is significant for girls; fur-
thermore, the difference in the effect that cul- Cultural capital does affect educational out-
tural capital has on boys and girls is statisti-
comes, but in a limited way. Other variables-
cally significant. Looking both between and including habitus-are more important to
across all schools, however, one sees that the
grades than is cultural capital, which even at
benefit of cultural capital for boys does not its strongest will not raise a student's grade
exist, as shown in the OLS models. Habitus noticeably. Inherent ability is the dominant
matters for both genders, and the effect is in factor in influencing a student's grades, fol-
the same direction for both boys and girls: lowed by a student's habitus and socioeco-
Having high occupational expectations con- nomic background. Cultural capital plays a
tributes to a student's receiving higher lesser role overall and, at times, plays no role
grades. for boys.
While cultural participation may not be all An examination of cultural participation
that beneficial to male students' grades, there rates showed that while SESaffects participa-
may be other activities that have a positive tion more than gender, the gender differ-
effect. As was mentioned earlier, many male ences are still sizable. Girls receive better
students participate in sports, which are seen returns to their cultural participationthan do
as more socially acceptable than cultural boys, but it is far from obvious why this dif-
activities. Researchershave argued that sports ference exists. With regard to girls' higher cul-
are associated with boys' academic success tural participation, the eighth grade is a time
(Coleman 1961; Rehberg and Schaefer when boys and girls are becoming more
1968). Indeed, one could argue that many of aware of their gender roles and of what is
the skillsthat one acquires from taking cultur- socially appropriate for a male or female.
al lessons (being disciplined, managing both Studies ranging from Coleman's Adolescent
schoolwork and outside activities) could also Society (1961) to more recent work (Adler,
be obtained from participating in sports. Kless, and Adler 1992) have found that the
Therefore, a set of analyses was run to see if boys who are most popular in school are the
within boys' and girls' educational fields, par- ones who have high athletic ability. In con-
ticipation in sports serves as a form of capital trast, a boy who shows an interest in theater
that helps students achieve higher grades. or dance may be likelyto have his peers ques-
Table 6 shows the results for both the OLS tion his masculinity.
and fixed-effects models, with a dummy vari- Following this gender-socializationreason-
able for whether the student participates in ing, it may be that male students who do have
any sports activities (including intramural, cultural capital downplay it or disguise it, for
school varsity, and team sports outside fear of being seen as "sissies."In other words,
school). within their field, their cultural capital may
The results show that participation in actuallybe a liabilityin the eyes of their peers or
sports has a positive and significant effect on teachers, discouraging boys from translating
the grades of both male and female students, their cultural knowledge skills into success in
both with and without controlling for the classroom.Another explanationfor the dif-
between-school effects. Forgirls, culturalcap- ference could lie in the female students, rather
ital continues to have a strong impact, but than the male students. A number of studies
participation in sports is also significant. For (see Sadkerand Sadker1994 for a reviewof the
boys, participation in sports is clearly more literature)have found that teachers tend to pay
important than is cultural participation, but more attention to male students than to female
culturalcapital still has a weak effect. Forboth students and that male students are more vocal
Table 6. Estimatesof the Effectsof CulturalCapital and Sports on GPA

Model 3
Model 2 OLSwith
Model 1 Fixed Metric
OLS Effects Coefficients

Males Females Males Females Males Females

CulturalCapital .027* .037*** .034** .049*** .018* .024***


(.011) (.009) (.012) (.010) (.007) (.006)

SES .080*** .047*** .109*** .075*** .108*** .063***


(.011) (.011) (.013) (.012) (.015) (.015)

SportsParticipation .232*** .129*** .230*** .148*** .232*** .129***


(.022) (.018) (.025) (.020) (.022) (.018)

Ability .357*** .397*** .380*** .412*** .040*** .045***


(.009) (.010) (.010) (.011) (.001) (.001)

Student'sOccupationalExpectations .194*** . 139*** .192*** .121*** .194*** .139***


(.018) (.016) (.021) (.017) (.018) (.016)

Intercept 2.645*** 2.903*** 2.650*** 2.900*** .489*** .504***


(.021) (.018) (.019) (.017) (.060) (.059)

R2 .359 .366 .359 .366

N 6,794 7,092 6,794 7,092 6,794 7,092

Note: For Models 1 and 2, the variables for cultural capital, SES, and ability were standardized. Sample weights and robust standar
* = significant at .05, ** = significant at .01, *** = significant at .001.
Source:NationalEducationLongitudinalStudy,1988.
Cultural Capital,Gender,
Cultural Capital, Gender,and
and School
School Success
Success 61
61

and active in the classroom. Possessingcultural These perceptions about the opportunity
capitalmay help female students become more structureand what is necessaryto succeed are
visible to teachers or may give them the sense all part of one's habitus. With regardto oppor-
of mastery and confidence necessary to tunities for differentgender and racialgroups,
become more assertive in the classroom. Bakerand PerkinsJones (1993:92) stated:"One
Gender-socialization theory (Kaufman and type of performance will add more to some
Richardson1982; Mickelson1989) argues that individuals'life chances and less to others." In
girlswant to win social approval,whereas boys other words, women may have a habitus that
focus more on mastery.Girlsmay thus empha- has been structured by society in such a way
size their cultural capital in the hopes that that they believe that educational and occupa-
teachers will like them or consider them good tional success are possible, but it is also true
students. Conversely, boys are considered to that they will encounter obstacles as they pro-
have a more oppositional stance toward teach- ceed. Parents of female students may realize
ers and authorityfigures in school (Sadkerand that their daughters will have to put in more
Sadker1994) and hence would be less likelyto effort than males to secure the same positions.
displaytheir culturalcapitalfor the purposes of Acquiring and using cultural capital may be
winning their teachers' approval. seen as a necessarystep along the way to help
Although Bourdieu (1984) and DiMaggio overcome some of the obstacles that being
(1982) argued that the main benefit of cul- female poses.22
tural capital for females is the greater ability Finally,it may well be that the typical cul-
to attract successful husbands, the roles of tural capital variablethat has been used in the
women in American society have changed sociology of education needs to be reevaluat-
greatly since the 1960s, when both ed. Perhaps participation in cultural activities
researchers' data sets were collected. Since is not an adequate or appropriate measure of
women are in the labor market for longer cultural capital in modern American society.
periods and in better occupational positions The gender differences found here would
than they were 40 years ago, they may need imply that, at the least, researchers need to
to invest more at the beginning to reach a consider some different and more complex
level of success that is comparable to that of components of cultural capital than simple
men. Thus, men may not need cultural capi- exposure to high culture. The work of Lareau
tal to succeed (and it may even be seen in a (1987, 1989) is a good starting point for
negative light because of its "effeminate" other ways of conceptualizing culturalcapital.
nature) in the educational and occupational Because neither DiMaggio (1982) nor
markets, but for women, it may be a factor Aschaffenberg and Maas (1997) included a
that gives them an edge over the competi- variable for habitus in their analyses, the
tion, making them more confident and results found here may not be directly com-
noticeable to teachers. This argument is in parable to the results of their studies. In addi-
line with the findings of DiMaggio's study tion, while DiMaggio's (1982) dependent
(1982), in which cultural capital within the variable, like the dependent variable here,
higher SES group had more of a positive was grades, Aschaffenberg and Maas (1997)
impact on grades for female students than for used educational transitions as their depen-
male students. That is, high-SESmales, who, dent variable, making a comparison with
by their position in the class structure, are the their study more difficult. It is also somewhat
most likely to be educationally and occupa- difficult to compare studies when, as was
tionally successful, do not need cultural capi- mentioned previously, each study seems to
tal as much as do females in the same class have used a different set of variables that
position. Persell, Catsambis, and Cookson were labeled cultural capital. Furthermore,in
(1992) also contended that it is necessary for restricting the sample to white students to
women to have both more of the same assets focus on gender differences, I did not address
that men have and additional assets that men racialdifferences.
lack to reach the same level of attainment as It should be noted that the role of habitus
men. may not be as clear-cutas was presented here;
62
62 Dumais
Dumais

it is difficult to determine whether students ly, this work was available only in French. His
develop a belief that they can obtain a presti- MasculineDomination(2001), translated into
gious occupation because they are already English,treats women's oppression as analyt-
doing well in school, or if the belief is what ically independent from class and describes
motivates them to strivefor good grades. This masculine domination as a prime example of
study is reallya first attempt to operationalize symbolic violence.
the concept of habitusalongside the concept of 2. One interesting interpretation by
cultural capital.23 There are certainly many McCall(1992:844) is that "secondary"can be
other factorsthat constitute a student's habitus interpreted to mean either not as important
than occupational expectations, and judging or that "gender as a principle of division is
from the significancethat this one small partof secondary because it is hidden and it is hid-
habitus has on grades, it will be importantto den because it appears to be universal and
develop measures for these other factors in natural." This second interpretation gives
future research. much more weight to gender as a stratifying
There are several other directions for future factor in society.
research. It would be helpful to have data for 3. As Yupin et al. (2000) reported, 22 per-
siblings, so that cultural activity levels could cent of males and 17 percent of females
be compared for boys and girls in the same major in business, 12 percent of males and 2
family. Also, as was stated earlier,the interac- percent of females major in engineering, 2
tion of gender and race has not yet been percent of males and 1 percent of females
studied in relation to cultural capital and major in the physical sciences, and 4 percent
schooling. Furthermore,data need to be col- of males and 1 percent of females major in
lected that will allow researchers to distin- computer science. Looked at another way, in
guish between Bourdieu'sconcept of habitus 1990, women received only 14 percent of the
as disposition and expectations for the future engineering degrees that were awarded that
and the more conventional status attainment year (Jacobs 1996).
(Sewell, Haller,and Portes 1969) and rational- 4. Most commonly, the argument is that a
choice concepts of expectations. high-SESgroup has the culturalcapital (in the
It is clear that we researchersmust be more form of highbrow possessions and activities)
careful in using and operationalizing the term and prevents low-SESpeople from ascending
culturalcapital and more willing to consider the social hierarchy.However,the power of any
the inclusion of habitus in our models. In the kindof culturalcapitalis dependent on the field
United States, cultural activities do not seem in which the actors exist. In a differentfield, it is
to affect all students the same way, leading to possible that the (broadly defined) cultural
the reproduction of society. As much as there knowledge possessed by the lower-SESgroup
is a class dynamic at work in social reproduc- would be more powerful than any knowledge
tion, there is also a gender dynamic, in which of high art (Swartz1997).
boys are expected to behave one way and 5. To some, this has always seemed to be
girls another. By their very presence, such the metaphysical process of early socializa-
complex effects render simplistic culturalcap- tion.
ital explanations less powerful. Future 6. Even lower-class students who manage
research must take account of the other com- to accumulate cultural capital in school and
ponents of Bourdieu's framework-habitus advance successfully through the school sys-
and field-to test more accurately the pres- tem are easy to distinguish from their upper-
ence and effect of culturalcapital for different class peers because their cultural capital is
groups in the American school system. more "scholastic"and conservative than that
of those who were exposed to a wider variety
of culturalresources in their homes (Bourdieu
NOTES 1984).
7. Furthermore,the studies that have dis-
1. Bourdieu has written more extensively cussed habitus (McClelland 1990; Reay 1995)
and explicitly about gender, but until recent- have not discussed cultural capital.
Cultural Capital,
Cultural Gender, and School Success
CapitalGenderandShoolSucessI
63

8. Questionnaires are available at the stu- 14. Although this study focused on stu-
dent, parent, teacher, and school levels; for dents who were white and in the eighth
this study, both the student and parent ques- grade, the findings regarding the importance
tionnaires were examined. The parent ques- of gender in cultural participation are consis-
tionnaire was administered to the parent or tent with the findings of Mohr and DiMaggio
guardian who was self-selected as the "most (1995) and with Yupin et al.'s (2000) study
knowledgeable about the student's educa- on cultural participation:Among high school
tional activities and related behaviors." seniors in 1995, 46 percent of the female stu-
9. The weight used in the analyses is called dents, compared to 35 percent of the male
BYQWT.According to the electronic code- students, participated in music or other per-
book for the NELSdata set (U.S. Department forming arts activities, while 63 percent of the
of Education 1988), "BYQWTwas calculated male students and 49 percent of the female
from the design weight (RAWWT)for the stu- students participated in sports (Yupin et al.
dent questionnaire, adjusted for the fact that 2000:50).
some of the selected students did not com- 15. Informationis available for three activ-
plete the questionnaire. RAWWTis the recip- ities for parents: attending concerts, going to
rocal of the conditional selection probability the library, and visiting art museums. The
for the student, given that the school was only activity that is affected by the gender of
selected into the base year sample, multiplied the child is concert attendance, an activity
by his or her school's design weight." that parents of girls are more likelyto engage
10. The parents responded to all these in, regardless of SES (the difference is the
questions with yes or no; therefore, it was not greatest in the lowest quartile, with 50 per-
possible to develop a measure of how fre- cent of the girls' parents and 42 percent of
quently the students participated in these the boys' parents attending concerts).
activities. Students' activity levels are highly correlated
11. In addition to these three categories, with parents' activity levels (.706, p < .001).
students could choose from the following 16. The NELSeighth-grade teacher's ques-
occupations: craftsperson, farmer, housewife, tionnaire does not have any useful questions
laborer, military/police, technical, sales/cleri- about teachers' perceptions of students that
cal, service worker, other, not working, or could serve as indicators of cultural capital.
don't know. These kinds of questions would be informa-
12. It should also be noted that 87 percent tive in understanding how cultural capital
of the 16,321 students in this sample (eighth translates into grades, and it would make
graders in 1988) attended public schools, 8 sense to include such questions in future sur-
percent attended Catholic schools, 1 percent veys.
attended private schools affiliated with the 17. A pooled regression was also run for
National Association of Independent Schools, both sexes together, with a dummy variable
and 4 percent attended other private schools. for female and variablesfor the interaction of
These figures are comparable with national female and SES,culturalcapital, and so forth.
statistics; according to the National Center In the final OLS model, the t-statistic for
for Education Statistics, 88 percent of stu- female*culturalcapital was 1.626, which does
dents were enrolled in public schools in 1985 not allow me to reject the null hypothesis that
(see NCESFast Facts2001). there is no difference between boys and girls
13. Although only the low- and high-SES in the effect of cultural capital. However, the
quartilesare discussed here, the patter for the female*occupational expectations variable
middle two quartilesis what would be expect- was significant at the .05 level, indicating that
ed; the second quartilehad higher participation there are differences between boys and girls
rates than the firstquartile,and the third quar- in this regard. All the other variables in the
tile had higher participationrates than the sec- model were also significant.
ond quartile (but lower than the fourth quar- 18. All these analyses were also run with
tile). In all the quartiles,the participationrates Englishgrades alone (which could be argued
for girlswere higher than those for boys. to be more directly related to cultural capital
64 Dumais
Dumais

than such courses as science or math) as the largerthan that for SES(and thus it would be
dependent variable. The results were less likely for cultural capital to appear as a
extremely similarto those in which GPAis the significant variable).
dependent variable;culturalcapital had a sig- 21. Again, a pooled regression was also
nificant effect for girls in the final model, but run for both sexes together, with a dummy
not for boys. variablefor female and variablesfor the inter-
19. DiMaggio (1982) included a variable action of female and SES,culturalcapital, and
for ability (the score on a vocabulary achieve- so forth. In this case, all the variables in the
ment test) in his models. Katsillis and model, including female*cultural capital and
Rubinson (1990) included a measure for pre-
female*occupational expectations, were sig-
vious achievement in their model of cultural
nificant, indicating that after school effects
capital's effects on GPAand found that even are controlled, there are real gender differ-
when previous achievement is dropped from
ences in the effects of cultural capital and
the model, cultural capital does not have a
habitus on grades.
significant effect. However, one could argue 22. On the other hand, it may be that par-
that the ability variable is, in part, cultural
ents find it easier to convince their daughters
capital that has been accumulated in the past
and converted into cognitive ability. When to participate in cultural activities and take
the ability variable is left out of the final long-term lessons than to convince their sons.
model here, the variables for SES, cultural Furtherresearch is necessary to determine the
roles of gender socializationversus opportuni-
capital, and habitus are all significant at the
.001 level for both boys and girls; however, ties in the social structure in relation to gen-
the R-squaredfor the model drops to .167 for der and cultural activities.
boys and .158 for girls. 23. Although McClelland (1990) opera-
20. This finding is quite striking, consider- tionalized habitus, she did not do so in the
ing that the measurement error for the cul- context of Bourdieu's broader framework,
tural capital variablewould be expected to be with the notions of cultural capital and field.
APPENDIX

CorrelationMatrix of VariablesUsed in the Study

1 2 3 4 5 6 7 8 9 10
AllStudents(N = 11,795)
1. Grades 1.000
2. Ability 0.565 1.000
3. Occupationalexpectations 0.274 0.253 1.000
4. SES 0.331 0.425 0.219 1.000
5. Culturalcapital 0.257 0.290 0.153 0.403 1.000
6. Artclasses 0.088 0.131 0.024 0.178 0.437 1.000
7. Musicclasses 0.216 0.252 0.115 0.312 0.616 0.218 1.000
8. Dance classes 0.148 0.102 0.104 0.204 0.530 0.165 0.257 1.000
9. Libraryvisits 0.161 0.192 0.090 0.194 0.493 0.074 0.129 0.106 1.000
10. Going to concerts 0.132 0.137 0.078 0.210 0.627 0.112 0.243 0.167 0.190 1.000
11. Museumvisits 0.115 0.159 0.090 0.248 0.629 0.194 0.177 0.148 0.197 0.255
12. Sportsparticipation 0.162 0.103 0.085 0.142 0.050 0.009 0.035 0.014 0.041 0.050

MaleStudents(N = 5,803)
1. Grades 1.000
2. Ability 0.557 1.000
3. Occupationalexpectations 0.288 0.273 1.000
4. SES 0.346 0.413 0.275 1.000
5. Culturalcapital 0.195 0.258 0.121 0.375 1.000
6. Artclasses 0.074 0.117 0.009 0.166 0.418 1.000
7. Musicclasses 0.156 0.218 0.099 0.273 0.588 0.213 1.000
8. Dance classes 0.026 0.045 0.063 0.098 0.313 0.099 0.179 1.000
9. Libraryvisits 0.160 0.193 0.099 0.200 0.517 0.068 0.100 0.049 1.000
10. Going to concerts 0.087 0.112 0.048 0.219 0.661 0.112 0.234 0.097 0.172 1.000
11. Museumvisits 0.103 0.132 0.070 0.227 0.657 0.166 0.174 0.099 0.201 0.265
12. Sportsparticipation 0.198 0.094 0.091 0.132 0.033 -0.028 0.035 0.014 0.042 0.061
APPENDIX (CONTINUED)

CorrelationMatrix of VariablesUsed in the Study

1 2 3 4 5 6 7 8 9 10

FemaleStudents(N = 5,992)

1. Grades 1.000
2. Ability 0.585 1.000
3. Occupationalexpectations 0.244 0.242 1.000
4. SES 0.327 0.438 0.178 1.000
5. Culturalcapital 0.287 0.347 0.140 0.463 1.000
6. Artclasses 0.096 0.147 0.030 0.190 0.462 1.000
7. Musicclasses 0.254 0.295 0.106 0.357 0.625 0.218 1.000
8. Danceclasses 0.176 0.167 0.074 0.311 0.593 0.206 0.269 1.000
9. Libraryvisits 0.150 0.195 0.065 0.193 0.473 0.077 0.145 0.115 1.000
10. Goingto concerts 0.157 0.176 0.078 0.216 0.584 0.108 0.228 0.154 0.195 1.000
11. Museumvisits 0.121 0.188 0.102 0.270 0.623 0.218 0.174 0.178 0.191 0.240
12. Sportsparticipation 0.156 0.109 0.107 0.146 0.114 0.043 0.083 0.085 0.058 0.076

Source:NationalEducationLongitudinalStudy,1988.
CulturalCapital,
Cultural Capital,Gender,
Gender,and
and School
SchoolSuccess
Success 67

REFERENCES Exposureto the Arts in America." Theoryand


Society5:141-59.
Aaronson, Daniel. 1998. "Using Sibling Data to Farkas, George. 1996. Human Capital or Cultural
Estimate the Impact of Neighborhoods on Capital? Ethnicityand Poverty Groups in an
Children's Educational Outcomes." journal of Urban School District. New York: Aldine de
HumanResources33:915-46. Gruyter.
Adler, PatriciaA., Steven J. Kless, and Peter Adler. Farkas,George, Robert P. Grobe, Daniel Sheehan,
1992. "Socialization to Gender Roles: and Yuan Shuan. 1990. "Cultural Resources
Popularity Among Elementary School Boys and School Success: Gender, Ethnicity, and
and Girls."Sociologyof Education65:169-87. Poverty Groups Within an Urban School
Aschaffenburg, Karen, and Ineke Maas. 1997. District."AmericanSociologicalReview55:127-
"Culturaland EducationalCareers."American 42.
SociologicalReview62:573-87. Greene, William H. 2000. EconometricAnalysis.
Baker,David P., and Deborah PerkinsJones. 1993. Upper Saddle River,NJ:Prentice Hall.
"Creating Gender Equality: Cross-national Hall, John R. 1992. "The Capital(s) of Cultures:A
Gender Stratification and Mathematical Nonholistic Approach to Status Situations,
Performance." Sociology of Education66:91- Class, Gender, and Ethnicity. Pp. 257-85 in
103. CultivatingDifferences:SymbolicBoundariesand
Bourdieu, Pierre. 1973. "Cultural Reproduction the Making of Inequality,edited by Michele
and Social Reproduction." Pp. 71-112 in Lamont and Marcel Fournier. Chicago:
Knowledge, Education, and Cultural Change: Universityof Chicago Press.
Papersin the Sociologyof Education,edited by Huber, P. 1. 1967. "The Behavior of Maximum
RichardBrown. London: Tavistock. Likelihood Estimates Under Non-Standard
-. 1984. Distinction: A Social Critique of the Conditions." Pp. 221-33 in Proceedingsof the
Judgementof Taste. Cambridge, MA: Harvard Fifth Berkeley Symposium on Mathematical
UniversityPress. Statistics and Probability.Berkeley: University
-. 1997. "The Forms of Capital." Pp. 46-58 in of CaliforniaPress.
Education:Culture,Economy,and Society,edit- Jacobs, Jerry. 1995. "Gender and Academic
ed by A. H. Halsey, Hugh Lauder, Phillip Specialties: Trends Among College Degree
Brown, and Amy Stuart Wells. Oxford, Recipientsin the 1980s." Sociologyof Education
England:Oxford UniversityPress. 68:81-98.
-. 2001. Masculine Domination, translated by - 1996. "Gender Inequality and Higher
Richard Nice. Stanford, CA: Stanford Education." Annual Review of Sociology
UniversityPress. 22:153-85.
Bourdieu, Pierre, and Loic J. D. Wacquant. 1992. Kalmijn,Matthijs, and Gerbert Kraaykamp.1996.
An Invitation to ReflexiveSociology. Chicago: "Race, Cultural Capital, and Schooling: An
Universityof Chicago Press. Analysis of Trends in the United States."
Coleman, James S. 1961. The Adolescent Society. Sociologyof Education69:22-34.
New York:Free Press of Glencoe. Katsillis, John, and Richard Rubinson. 1990.
De Graaf, Nan Dirk,Paul M. De Graaf,and Gerbert "CulturalCapital, Student Achievement, and
Kraaykamp. 2000."Parental Cultural Capital Educational Reproduction: The Case of
and Educational Attainment in the Greece." AmericanSociologicalReview55:270-
Netherlands: A Refinement of the Cultural 79.
Capital Perspective."Sociology of Education Kaufman, Debra R., and Barbara L. Richardson.
73:92-111. 1982. Achievement and Women: Challenging
DiMaggio, Paul. 1982. "Cultural Capital and TheAssumptions.New York:Free Press.
School Success: The Impact of Status Culture Laberge, Suzanne. 1995. "Towardan Integration
Participation on the Grades of U.S. High of Gender into Bourdieu'sConcept of Cultural
School Students." AmericanSociologicalReview Capital."Sociologyof SportJournal12:132-46.
47:189-201. Lamont, Michele, and Annette Lareau. 1988.
DiMaggio, Paul, and John Mohr. 1985. "Cultural "Cultural Capital: Allusions, Gaps and
Capital, Educational Attainment, and Marital Glissandos in Recent Theoretical
Selection." American Journal of Sociology Developments." SociologicalTheory6:153-68.
90:1231-61. Lareau,Annette. 1987. "Social Class Differences in
DiMaggio, Paul, and Michael Useem. 1978. "Social Family-School Relationships:The Importance
Class and Arts Consumption: The Origins and of Cultural Capital." Sociology of Education
Consequences of Class Differences in 60:73-85.
68
68 Dumais

-1989. Home Advantage: Social Class and Robinson, RobertV., and MauriceA. Garnier.1985.
Parental Interventionin ElementaryEducation. "ClassReproductionAmong Men and Women
London: FalmerPress. in France: Reproduction Theory on its Home
Lareau, Annette, and Erin McNamara Horvat. Ground." American Journal of Sociology
1999. "Moments of Social Inclusion and 91:250-80.
Exclusion:Race, Class, and CulturalCapital in Roscigno, Vincent J., and James W. Ainsworth-
Family-School Relationships." Sociology of Darnell. 1999. "Race, Cultural Capital, and
Education72:37-53. Educational Resources: Persistent Inequalities
Lorber, Judith. 1994. Paradoxes of Gender. New and Achievement Returns." Sociology of
Haven, CT:Yale UniversityPress Education72:158-78.
McCall, Leslie. 1992. "Does Gender Fit? Bourdieu, Sadker, Myra, and David Sadker. 1994. FailingAt
Feminism, and Conceptions of Social Order." Fairness:How America's Schools Cheat Girls.
Theoryand Society21:837-67. New York:Charles Scribner'sSons.
McClelland, Katherine. 1990. "Cumulative Sewell, William H., Archibald O. Haller, and
Disadvantage Among the Highly Ambitious." AlejandroPortes. 1969. "The Educationaland
Sociologyof Education63:102-21. Early Occupational Attainment Process."
Mickelson,RoslynArlin.1989. "WhyDoes JaneRead AmericanSociologicalReview34:82-92.
and Write So Well?The Anomaly of Women's Shaw, Gordon. 1999. KeepingMozart in Mind. San
Achievement."Sociologyof Education62:47-63. Diego, CA:Academic Press.
Mohr, John, and Paul DiMaggio. 1995. "The Swartz, David. 1997. Culture and Power: The
Intergenerational Transmission of Cultural Sociology of Pierre Bourdieu. Chicago:
Capital." Researchin Social Stratificationand Universityof Chicago Press.
Mobility14:167-99. Teachman, Jay D. 1987. "Family Background,
NCES Fast Facts: Public and Private School Educational Resources, and Educational
Enrollment. 2001. Available on-line at Attainment." American Sociological Review
http://nces.ed.gov/fastfacts/display.asp?id=65 52:548-57.
NEAFact Sheet: ArtsEducation.2001. Availableon- Thorne, Barrie.1993. GenderPlay:Girlsand Boysin
line at http:// arts.endow.gov/learn/Facts/- School. New Brunswick,NJ:RutgersUniversity
Artsed.html Press.
Persell, Caroline Hodges, Sophia Catsambis, and U.S. Department of Education, National Center for
Peter W. Cookson, Jr.1992. "DifferentialAsset Education Statistics. National Education
Conversion: Class and Gendered Pathways to Longitudinal Study, 1988 [Computer file].
Selective Colleges." Sociology of Education Chicago: National Opinion Research Center
65:208-25. [producer], 1989: Ann Arbor,Ml: Inter-univer-
Reay, Diane. 1995. "'They Employ Cleaners to Do sity Consortium for Political and Social
That':Habitusin the PrimaryClassroom."British Research[distributor],1990.
journalof Sociologyof Education16:353-71. Yupin, Bae, Susan Choy, Claire Geddes, Jennifer
Rehberg, RichardA., and Walter E. Schaefer. 1968. Sable, and Thomas Snyder. 2000.Educational
"Participationin InterscholasticAthletics and Equityfor Girlsand Women(NCES2000-030).
College Expectation." American Journal of Washington, DC: U.S. Government Printing
Sociology73:732-40. Office.

Susan A. Dumais, MA, is a doctoral candidate, Department of Sociology, Harvard University,


Cambridge,Massachusetts.Her mail fields of interest are social stratification,sociology of educa-
tion, and sociologyof culture.She is currentlycompletingher dissertationon the educationalattain-
ment of white working-classstudents.

The author is grateful for the comments of Shyon Baumann, GabriellaGonzalez, Devon Johnson,
MonicaMcDermott,Stephen Morgan,and the late Aage S0rensen. Thisresearchwas supportedby
a grant from the AmericanEducationalResearchAssociation, which receives funds for its AERA
Grants Programfrom the National Center of EducationStatistics and the Office of Educational
Researchand Improvements(U.S. Departmentof Education),and the National ScienceFoundation
underNSFGrantRED-9452861. Theopinionsexpressedhere reflectthose of the author and do not
necessarilyreflect those of the granting agencies. Direct all correspondenceto Susan A. Dumais,
Department of Sociology, Harvard University,473 WilliamJames Hall, 33 KirklandStreet,
Cambridge,MA02138; e-mail: dumais@wjh.harvard.edu.

También podría gustarte