Gender, Social Change, and Educational Attainment
Author(s): Ann M. Beutel and William G. Axinn
Source: Economic Development and Cultural Change , Vol. 51, No. 1 (October 2002), pp. 109134
Published by: The University of Chicago Press
Stable URL: https://www.jstor.org/stable/10.1086/345517
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access
to Economic Development and Cultural Change
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Gender, Social Change, and Educational Attainment*
Ann M. Beutel
University of Oklahoma
William G. Axinn
University of Michigan
Sociological research has focused much attention on processes of educational
attainment. In part, this is because of direct benefits thought to result from
education and, in part, because educational attainment is considered a key step
in other processes of attainment, such as occupational attainment.1 Because
education gives individuals opportunities to achieve status mobility, the links
between ascribed dimensions of status, such as gender, race, and education,
have always drawn sociologists’ attention. The spread of mass education constitutes a fundamental social transformation and a watershed in attainment processes because it opens up previously unavailable status mobility routes. Rarely,
however, do we have an opportunity to examine directly the relationship between
ascribed dimensions of status and educational attainment during the very beginning of the spread of mass education. In this article, we use a unique set of
measures from a setting in the midst of the spread of education to examine both
the impact of gender on processes of educational attainment and the ways in
which community-level social change attenuates the impact of gender on
education.
This article tests several key hypotheses regarding this fundamental social
transformation. First, we investigate hypotheses regarding the impact of gender
on educational attainment during the spread of mass education. Although a large
body of previous research documents important gender differences in educational attainment from settings in which education is already widespread, little
is known about the connections between gender and specific dimensions of the
educational attainment process, such as enrollment and drop-out rates, at the
onset of universal education. Second, we test hypotheses regarding the impact
of social changes at the community (i.e., local) level on individual educational
attainment. These hypotheses predict that macro-level changes in educational,
employment, and consumption opportunities will increase school attendance.
䉷 2002 by The University of Chicago. All rights reserved.
0013-0079/2003/5101-0005$10.00
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
110
Economic Development and Cultural Change
Third, we test key hypotheses regarding the impact of local-level social change
on the individual-level relationship between gender and educational attainment.
Our framework for studying this overall social transformation uses the social
organization of the family as the key intervening link between macro-level
social change and the impact of gender on educational attainment.
The setting on which we focus is the multiethnic, rural population of
south-central Nepal. Rural Nepal provides an ideal setting for our study because secular education was entirely unavailable prior to the early 1950s but
has become universal within the lifetimes of current residents. Simultaneously,
the proliferation of education has been accompanied by dramatic social
changes in the availability of wage labor employment, the spread of markets,
and the distribution of transportation infrastructure. Historically, status attainment in this setting was in large part determined by caste and gender,
whereas now these social changes have drastically transformed attainment
processes. With detailed community-level measures of social change and
equally detailed individual-level life histories of educational experience, we
can now thoroughly examine the relationships among gender, social change,
and educational attainment.
I. Theoretical Framework
A. Gender and Educational Attainment
A considerable amount of previous research documents important differences
in educational attainment between men and women. For example, research
conducted in the United States during the 1960s and 1970s showed that women
tended to achieve lower levels of both educational attainment and intergenerational occupational mobility than men.2 Furthermore, this body of research
found parental characteristics to be important determinants of educational
attainment for both men and women but that boys perceived higher parental
encouragement for education than girls.3 In addition, those studies of attainment processes that included factors particularly relevant to women’s adult
family roles, such as dating frequency, age at first marriage, and age at first
birth, found significant effects on educational outcomes for females but not
for males.4 This evidence suggests that gender differences in educational attainment may arise because of gender differentiation in adult roles and the
emphasis on family-related roles for women.5 In combination with actual or
perceived role conflict between the pursuit of the family role and the pursuit
of education and career roles, the emphasis on family roles for women may
lead them to abandon schooling or their parents to discourage them from
educational pursuits.6
Research from agricultural societies has also shown the important effects
of gender on educational participation and attainment, with girls much less
likely to be enrolled in school at any point in time and achieving lower levels
of educational attainment than boys.7 Consistent with the findings of status
attainment research in the United States, this suggests that gender role differentiation has important effects on girls’ education. In agrarian societies
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Ann M. Beutel and William G. Axinn
111
where female roles are defined largely in terms of home and family, parents
are likely to have significant incentives to keep their daughters at home and
out of school.8 Rural Nepal is certainly such a society, and the need for girls’
labor on the farm and in the household has been cited as a major reason for
not sending girls to school or for sending girls to school for relatively short
periods of time regardless of the level of household wealth.9 Thus, even though
parents may desire some amount of schooling for daughters as a means of
enhancing their marriage prospects, perceived conflict between women’s roles
in the family and their educational and work roles motivates parents to discontinue their daughters’ schooling at a significantly earlier stage than is true
for their sons.10 Empirically, therefore, we expect parents to be less likely to
enroll daughters in school than sons and more likely to take their already
enrolled daughters out of school than sons.
In settings such as rural Nepal, the consequences of perceived genderspecific role conflict between family roles and work roles can be exacerbated
by gender differences in marriage patterns and parental dependence on children
for old-age security. In the South Asian region, most parents depend on income
provisions from their children for financial support in old age.11 Furthermore,
in North India and Nepal, daughters generally leave their parents’ household
at marriage to reside with their parents-in-law.12 As a result, parents are much
more inclined to invest in their sons’ human capital than in their daughters’
human capital, as investments in daughters are generally believed to benefit
in-laws, not natal families.
Moreover, although wage labor opportunities have been spreading steadily through South Asia, males are much more likely to take wage labor jobs
than females, and many believe these money earning opportunities are relatively unavailable to women.13 This further reduces parental motivation to
educate daughters. Overall, therefore, these setting-specific characteristics of
rural Nepal reinforce our predicted gender difference effects on educational
enrollment and attainment so that we expect gender differences to be particularly great in our study.
B. Social Change and Educational Attainment
Several dimensions of social change are likely to have an impact on educational attainment. First, the proliferation of schools is itself likely to promote
enrollment and attainment. As schools become increasingly available, the costs
of sending children to school will decline, and parents will find it easier to
enroll their children. Likewise, the longer schools remain nearby, the more
common school attendance is likely to become.
Second, the spread of wage labor employment opportunities is also likely
to stimulate greater educational attainment. School enrollment allows individuals to invest in their human capital in order to increase their chances of
obtaining a wage labor job and mobility among jobs.14 This is particularly
likely to be true in South Asia because the British system of formal education,
adopted throughout much of the region, was based on principles of certifying
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
112
Economic Development and Cultural Change
individuals for specific jobs, especially governmental jobs. As a result, as
wage labor employment opportunities proliferate, motivation to enroll in
school is expected to increase as a means of securing those jobs. Likewise,
we expect motivation to educate children to increase with longer exposure to
wage labor jobs nearby.
Third, the spread of markets is likely to increase school enrollment and
educational attainment. Here, the mechanisms are indirect. As markets spread
throughout rural Nepal, goods and services become more widely available,
but only to those who have the money to purchase those goods and services.
Thus, the spread of markets is expected to increase the demand for money,
which will encourage individuals to pursue wage labor jobs and higher wages
among those jobs, and the desire to obtain higher paying jobs is expected to
motivate educational attainment.15 Because these mechanisms are indirect, lags
may occur before the spread of markets stimulates greater educational attainment. Nevertheless, the longer markets are present nearby, the greater are the
chances that parents will send their children to school and keep them there.
Fourth, the spread of improved transportation infrastructure is also likely
to increase educational attainment. This is because improved transportation
infrastructure facilitates access to schools, wage labor employment opportunities,
and markets. By increasing access to these other social institutions, each of
which is expected to motivate educational attainment by itself, improved transportation infrastructure is also likely to increase school enrollments.
The social changes in the availability of schools, wage labor jobs, markets,
and transportation are all likely to increase the chances that parents will enroll
and keep their children in school. However, in order to understand the impact
of these same social changes on gender differences in the likelihood that parents
send their children to and keep them in school, we must first consider the impact
of these same social changes on the social organization of family life. This is
because we believe the social organization of family roles is a fundamental
determinant of gender differences in educational attainment.
C. Social Change, the Social Organization of Families, and Gender
Difference in Education
We use the family mode of social organization framework to consider the
impact of social change on the social organization of families. Our choice of
this framework is rooted in the premise that improved transportation and
communication, monetization of the economy, and population growth increase
the division of labor in society. Emile Durkheim argued that these factors
increase the numbers of people who interact with one another, or the “moral
density” of society.16 He mentioned three specific mechanisms that predicate
the division of labor: population concentration, the formation and development
of cities, and improved communication and transportation.17 This framework
recognizes that changes in transportation, communication, and monetization
can reorganize not only production but a wide variety of other social activities
as well, which has a great impact upon the social organization of families.
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Ann M. Beutel and William G. Axinn
113
The family mode of social organization framework considers a wide
array of social changes and their potential influence on individuals in families.18 The framework builds on previous research that focused exclusively on
the family mode of production and extends to family modes of social organization across a variety of domains: consumption, residence, recreation, protection, socialization, procreation, and production.19 Historically, most of these
activities of daily living were organized within the family.20 Yet, as social
changes created new nonfamily institutions to organize these activities, they
increasingly took place outside the family. No society is expected to be completely organized inside or outside of families, but the nonfamily and family
modes of social organization are two ideal types that aid our understanding
of social change and the family.
Social changes, such as the spread of nonfamily schools, nonfamily employment opportunities, nonfamily consumption opportunities (markets), and
nonfamily transportation, alter the organization of social life so more of social
life takes place outside the family. As activities of daily living increasingly
take place outside the home and away from the family, the structure of social
interactions changes and alters social relationships with both family members
and others outside the family. This reorganization of family life is the key
link between macro-level social changes and micro-level changes in adult
roles and perceived role conflict between family roles and nonfamily roles.
As nonfamily social institutions spread and more of social life takes place
outside the family, family-related adult roles are likely to become a less
significant feature of adult roles. With regard to educational attainment, the
consequence is quite likely reduced gender differences in school enrollment.
Again, time lags are likely to be a key feature of this impact of new nonfamily
institutions on gender differences in education, but the longer these new nonfamily institutions are part of daily life, the smaller the gender differences we
expect to find in the propensity of parents to send their sons and daughters
to school and keep them there. In a highly gender-stratified setting such as
rural Nepal, the impact of new nonfamily institutions may not be sufficient
to eliminate gender differences in enrollment and attainment, but they are
likely to significantly reduce them.
Thus, in this setting, as education begins to spread, we expect sons to
be more likely to attend and to stay in school than daughters. The social
changes in the spread of wage labor employment, markets, and transportation
that accompany the spread of schools will themselves increase enrollment.
However, the spread of these same nonfamily institutions is likely to produce
a fundamental change in the organization of daily social life that reduces the
conflict between adult family roles and adult work roles. As a result, these
same social changes are eventually likely to reduce gender differences in
educational attainment. Thus, in this setting, the spread of new nonfamily
social institutions may be the key to the success of mass education in providing
a new route to status attainment that is independent of key ascribed statuses,
such as gender.
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
114
Economic Development and Cultural Change
II. The Study
A. Setting
The setting for our study is Western Chitwan Valley in south-central Nepal.
Nepal is an ethnically diverse country of approximately 22 million people
and is located in the central Himalayas. The country spans approximately 500
miles in length and only about 60 miles north to south. Chitwan is 100 miles
southwest of the capital city, Kathmandu. The population of Chitwan in 1996
(the year of our data collection) was approximately 414,261.
Until the 1950s, Western Chitwan Valley was covered with virgin forests
and infested with malaria-carrying mosquitoes. In the mid-1950s, the Nepalese
government (with assistance from the United States) began a program to clear
much of the forests, eradicate malaria, and distribute land to settlers from the
higher Himalayas, who were attracted to the area because of the rich soil and
flat terrain of the valley. Individuals from a variety of religious and ethnic
groups, which differ widely on a range of attitudes, beliefs, and behaviors,
came to inhabit the region. High-caste Hindus form the largest religious-ethnic
group in the valley, but large numbers of low-caste Hindus, Newar, Tibetoburmese, and indigenous ethnic groups reside there as well.
Chitwan was relatively isolated from the rest of the country until the late
1970s when major roads were built, linking the largest town of Narayanghat
to the eastern, northern, and western portions of the country. As a result of
the new roadways, Narayanghat became a transportation hub within the country, and a variety of government offices, businesses, and employment opportunities developed in Narayanghat and spread from Narayanghat throughout Western Chitwan Valley. Thus, over the course of approximately 40 years,
the valley changed greatly from being relatively isolated to being the location
of a vast array of nonfamily institutions. As a result, the social organization
of daily life in Chitwan has shifted from virtually all social activities being
organized within families to many social activities becoming organized outside
of the family. The focus of our investigation here is the spread of mass
education.
Formal education was virtually nonexistent in Nepal until the 1950s when
the British system of formal schooling was introduced.21 That system was
based on the principles that all children, regardless of social background,
ethnicity, or gender, have a right to an education and that the national government should ensure equal access to education.22 Despite this, large numbers
of Nepalese children do not attend school. In 1990, the combined primary
and secondary gross enrollment ratio in Nepal (number in school/number in
school eligible age group) was 39 for females and 80 for males.23 This contrasts
with a combined primary and secondary gross enrollment ratio in the United
States of 99 for females and 99 for males.24 Data on school enrollment for
the Chitwan district only are unavailable. However, the higher adult literacy
rate in Chitwan (49.46%) as compared with that of the entire country (36.75%)
suggests that residents of Chitwan have had comparatively high educational
opportunities.25 The data we analyze here demonstrate a dramatic increase
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Ann M. Beutel and William G. Axinn
115
Fig. 1.—Number of schools and students over time
over time in enrollment in school (described below). This increase was so
dramatic, in fact, that, by 1996, 100% of the children ages 5 and 6 in our
study area had been to school at least for one day.
The spread of formal education in Western Chitwan paralleled the spread
of education throughout the rest of Nepal. The first school was established
in 1954, and both the number of schools and the number of students enrolled
in schools increased dramatically in the decades that followed. Through a
combination of archival, ethnographic, and survey methods, we gathered a
detailed set of histories for all the schools that had ever existed in the study
area between 1946 and 1995.26 Figure 1 displays data from those histories,
documenting the growth of both the number of schools and the number of
students enrolled in those schools between 1955 and 1995. Clearly, the spread
of education in the area has been dramatic, reaching a total of 123 schools
and 43,785 enrolled students by 1995.
Gender differences in education have been equally dramatic in this setting. In Nepal in general, girls are much more likely than boys to not be
enrolled in school: of the children not enrolled in schools, approximately twothirds are girls. Girls are also more likely to repeat grades and drop out of
school.27 Girls’ participation in schooling has lagged behind boys’ despite a
variety of government programs intended to promote their involvement in
education.28 In Chitwan, however, the gender gap in schooling has narrowed
much more quickly than in most other regions in Nepal. As figure 2 shows,
only 6.7% of Chitwan students in 1954 were girls, but by 1995, 47.4% of
students were girls. Our interviews with parents in Chitwan also indicate that,
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
116
Economic Development and Cultural Change
Fig. 2.—Percentage of total students who are female over time
by 1996, parents placed a high value on educating both their sons and their
daughters. In response to survey questions (described below), 93% of respondents rated attending college as very important for their sons and 89%
rated attending college as very important for their daughters. In addition, there
is relatively little variation in the importance placed on college education for
daughters by gender of the parent: in the survey, 93% of male respondents
and 84% of female respondents rated sending their daughters to college as
very important.
One of the key features that makes Chitwan different from other parts
of rural Nepal has been the tremendous proliferation of nonfamily organizations and services throughout the area. In addition to collecting histories of
schools, we used neighborhood history calendars to collect histories of exposure to new nonfamily organizations and services for a sample of 171
neighborhoods in Western Chitwan Valley.29 These histories allow us to calculate the number of minutes from each neighborhood to the nearest school,
employer, market, or transportation service, independently for each calendar
year, as the spread of these services brings them much closer to each neighborhood. Figure 3, which displays changes over time in the mean minutes by
foot to the nearest school, employment, market, and transportation service,
illustrates how dramatic changes in nonfamily organizations and services in
western Chitwan have been. Although the distance from the average neighborhood in western Chitwan to the nearest nonfamily service was well over
an hour in the 1950s, by the 1990s the average neighborhood in western
Chitwan had several different nonfamily services and organizations within a
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Ann M. Beutel and William G. Axinn
117
Fig. 3.—Mean minutes by foot to nearest nonfamily service or organization over
time.
20-minute walk. The dramatic change in the proliferation of nonfamily services and organizations (described in fig. 3) happened over the same span of
years as the dramatic rise in the proportion of students who were female
(described in fig. 2). In what follows, we use detailed survey data on individual
children’s schooling experiences to examine the extent to which these two
changes are linked.
B. Survey Data
Our statistical analyses of the impact of gender and social change on educational attainment use data from the Chitwan Valley Family Study (CVFS).
The CVFS collected data from neighborhoods, households, and individuals.
Neighborhoods served as the primary sampling unit and were defined as
clusters of 5 to 15 households. In these small neighborhood clusters, residents
interact with one another on a daily basis. Neighborhoods were selected
through an equal probability, systematic sample of all neighborhood clusters
in Western Chitwan Valley, with the resultant sample consisting of 171 neighborhoods. Within each neighborhood, every individual between the ages of
15 and 59 was interviewed in 1996. Nonresident spouses of residents also
were interviewed to provide comprehensive data from both husbands and
wives. Using life history calendars, detailed measures of life events, such as
marriage, migration, schooling, and childbearing, were collected.30 For each
respondent, these life history calendars also were used to collect a complete
history of the respondent’s children’s schooling experiences. The survey in-
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
118
Economic Development and Cultural Change
terviews enjoyed a very high response rate of 97%, yielding life history
measures from 5,272 individuals.
C. Measures
Education. Our statistical analyses of the impact of gender and social change
on educational attainment focus on two key dimensions of the educational
attainment process: children’s entry into school and children’s dropping out of
school. To simplify the analyses and to insure comparable measurement is
available for all those included in the study, we focus on the schooling experiences of firstborn children. The measures of these children’s schooling experiences come from interviews with their mothers. The sample of mothers we
include in our analysis are those women who were between the ages of 25 and
54 at the time of the interview and had at least one child (N p 1,159).
In the analysis of school entrance, the dependent variable measures
whether a woman’s firstborn child entered school between the ages of 3 and
12. This variable is operationalized as the hazard of entering school, or the
probability of entering school in any year given the child has not already
entered school. In the analysis of school exit, which was limited to those
firstborn children who did enter school between ages 3 and 12, the dependent
variable measures whether they ever exited school by age 15. This variable
also is operationalized as a hazard rate—here, the hazard of first exit from
schooling. We choose age 15 as the upper limit because it is unlikely that
children would have left school because of graduation by that age.
Gender and social change. Two key independent variables are child’s
gender and social change. We measure gender of the child with a dichotomous
variable, coded 1 if the child is male and 0 if the child is female.
As discussed above, the dimensions of social change on which we focus
are the proliferation of nonfamily schools, nonfamily wage labor employment
opportunities, nonfamily markets, and nonfamily transportation services. We
used neighborhood history calendar measures to operationalize our measurement of the spread of these nonfamily organizations and services.31 These
calendars measured the number of minutes’ walk from the neighborhood to
the nearest school, employment opportunity, market, or transportation service.
Schools were defined as the nearest location (but not necessarily a physical
building) of nonfamily instruction aimed at children and youth. Employment
opportunities referred to the nearest employer who employs 10 or more individuals for pay. Markets were defined as the nearest location of two or more
shops where goods and services are sold for money. Transportation services
referred to the nearest location where a resident could board a public motorized
vehicle and ride for a fee.
The measures of these distances employed no boundaries, so all distances
were possible. Our aim in the present analyses is to assess the impact of
nearby nonfamily organizations and services, so we focus on nonfamily organizations and services within 10 minutes’ walk of the neighborhood.32 Because our aim is also to assess the impact of length of exposure to these
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Ann M. Beutel and William G. Axinn
119
nonfamily organizations and services, we constructed variables that measured
the number of years that schools, employment opportunities, markets, or transportation services were within a 10-minute walk of the neighborhood.
To test our hypotheses regarding the impact of social change on the
relationship between gender and educational attainment, we include an interaction term for the number of years of each nonfamily service or organization
(schools, employment opportunities, markets, transportation services) and gender of the child. These variables, discussed in greater detail below, allow us
to examine whether gender differences in educational attainment diminish
with increased exposure to nonfamily services and organizations.
Controls. A good deal of previous research demonstrates that parental
social background, particularly education, influences children’s educational
and occupational attainment.33 Parental aspirations for children and self-aspirations both play key roles in transmitting the effects of parental education
on their children’s educational attainment.34 In fact, these intergenerational
effects on children’s schooling may be particularly strong at the beginning of
the spread of mass education. This is because in settings where schooling is
inaccessible for the majority of individuals, intergenerational mobility is constrained and individual attainment depends more on family background.35
Parental schooling experiences also may have an important impact on the
gender gap in educational attainment. Research has shown that parents’ education has an important effect on daughter’s schooling.36 This is in part
because parents who have been formally educated are more likely to place a
high value on education for daughters as well as sons than are parents who
have not been formally educated.37
To control for these important intergenerational effects, our multivariate
models include measures of mother’s father’s education and mother’s education. Mother’s father’s education is measured as a dichotomous variable,
coded 1 if the mother’s father had any formal schooling. Mother’s education
is measured as the number of years of schooling she received up to 1 year
before the analysis interval begins—that is, until her firstborn was 2 years of
age in the school entrance analysis and until 1 year before her firstborn went
to school in the school exit analysis.38
Ethnicity may influence educational attainment because religious or ethnic groups often differ in the value they place on education. For example, in
the United States, research has found Jewish identity to have a much larger
effect (via parental aspirations and self-aspirations) than other religious identities on educational attainment.39 Similarly, research in Nepal has shown that
children who are high-caste Hindu or Newar—religious-ethnic groups that
place a high value on education in this setting—are more likely than others
to be enrolled in school and to attain high levels of educational attainment.40
Therefore, our multivariate models also include controls for mother’s religious-ethnic identity: high-caste Hindu, low-caste Hindu, Hill Tibetoburmese,
Newar, or Terai Tibetoburmese. High-caste Hindu serves as the reference
group in the analysis. Historically, members of that group have had the greatest
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
120
Economic Development and Cultural Change
access to educational and other opportunities and, as a result, have been the
most highly educated. Identity in each of the other groups is coded as a
dichotomy. Newars, who practice a mixture of Hinduism and Buddhism, are
another group who have enjoyed both access to and participation in formal
schooling. The Hill Tibetoburmese, who mainly follow the principles of Buddhism, have not been characterized by the same high level of devotion to
education as the high-caste Hindus or the Newars. In contrast with these
groups, low-caste Hindus have had many fewer educational opportunities.
Finally, the Terai Tibetoburmese, a group indigenous to the region, has had
much less access to educational and other opportunities than have other groups.
Because a key aim is to assess the impact of social change at the community level on educational attainment at the individual level, we also control
for the migration experiences of the mothers. These migration experiences
take individuals outside of the communities we have measured, possibly exposing them to a different set of community characteristics. We include two
controls for migration. The first is a non–time varying measure of any migration experience (i.e., any residence outside the neighborhood) between the
time of the mother’s birth until 1 year before the analysis interval began—that
is, until her firstborn was 2 years of age in the entrance analysis and 1 year
before her firstborn entered school in the exit analysis. The second is a timevarying measure of whether the respondent is outside the neighborhood during
the analysis interval—starting when her firstborn was 3 years of age in the
entrance analysis and the year her firstborn started school in the exit analysis.41
Finally, our multivariate models control for the mother’s birth cohort
using three dummy variables: 1962–71 (ages 25–34 at the time of the survey),
1952–61 (ages 35–44 at the time of the survey), and 1942–51 (ages 45–54
at the time of the survey). Each respondent is coded 1 on one of these variables
and 0 on the two other variables. The youngest birth cohort serves as the
reference group in the analysis. These controls help to insure that differences
across birth cohorts are not responsible for the pattern of relationships we
observe among our other measures.
Descriptive statistics for the measures of social change, gender, and nontime varying controls are shown in table 1. Table 1 does not describe the
measures of the hazard of entrance into school, exit out of school, or timevarying covariates because the values of these variables change by year in
our dynamic person-year analysis (described below).
D. Analysis Strategy
We use event history methods to model each mother’s firstborn child’s risk
of entering or exiting school. Because the life history calendars collected
measures precise to the year, we treat person-years as the unit of exposure to
risk and estimate our models using discrete-time methods. Thus, person-years
of exposure to the risk of entering or exiting school is the unit of analysis.
As noted earlier, firstborns between 3 and 12 years of age are at risk of school
entrance, and firstborns who have entered school and are between the ages
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Ann M. Beutel and William G. Axinn
121
TABLE 1
Descriptive Statistics for Measures Used in Analyses
Mean
Gender of child (1 p male)
Number of years service or
organization within a
10-minute walk until
child age 2:
School
Employment opportunities
Market
Transportation
Maternal characteristics:
Mother’s father has any
schooling
Mother’s years of education
Mother ever move (until
child age 2)
Mother’s religious-ethnicity:
High-caste Hindu
Low-caste Hindu
Hill Tibetoburmese
Newar
Terai Tibetoburmese
Mother’s birth cohort:
Born in 1962–71
Born in 1952–61
Born in 1942–51
Standard
Deviation
Minimum
Maximum
.49
.50
11.79
4.52
8.19
6.41
11.16
9.40
10.35
9.56
0
0
0
0
42
43
43
40
.22
2.24
.41
4.08
0
21
.99
.11
.48
.11
.18
.07
.17
.43
.35
.22
of 3 and 15 are at risk of school exit. We estimate the models using logistic
regression, which can be represented by the equation:
ln
( )
p
p a ⫹ S (Bk)(Xk) ,
1⫺p
where p is the yearly probability of entering or exiting school, p/1⫺ p is the
odds of school entrance or exit occurring, a is a constant term, Bk represents
the effects parameters of the explanatory variables, and Xk represents the
explanatory variables in the model. Using person-years of exposure to risk
as the unit of analysis increases the sample size, but this procedure does not
deflate standard errors and provides appropriate tests of statistical
significance.42
We transform the logistic coefficients by exponentiating them, so that
the coefficients we show in the tables are multiplicative effects on the odds
of entering or exiting school in any one year. A coefficient of 1.00 indicates
no effect on the odds, a coefficient greater than 1.00 indicates a positive effect
on the odds, and a coefficient less than 1.00 indicates a negative effect on
the odds. Because the odds of entering or exiting school in any given year
are quite low, these odds are quite similar to rates. Therefore, we will some-
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
122
Economic Development and Cultural Change
times describe the coefficients we present as effects on the rates of entering
or exiting school.
III. Results
A. School Entrance
Table 2 presents the results of our analyses of the effects of gender and locallevel social change on school entrance. For each neighborhood service or
organization, there are two columns of results. The first column shows the
results for a model without any interaction terms, and the second column
shows the results for a model that includes an interaction term for the nonfamily service or organization and gender of the child. Looking first at the
results for models without interaction effects, we see that gender has large
effects on school entrance that are of a similar magnitude across models,
indicating that boys enter school at higher rates than girls. For example, in
the first model of school presented in table 2, we see that boys enter school
at rates 51% higher annually than girls. Thus, our hypothesis regarding the
effect of gender on educational attainment is supported.
We also find support for our hypothesis regarding the effect of locallevel social change on educational attainment. All of the nonfamily service
and organization variables have significant effects on rates of entering school.
Each year that a school has been within a 10-minute walk of the neighborhood
increases the annual rate of school entrance by 1%, although the effect of
school is only marginally significant at the .10 level; each year that an employment opportunity has been within a 10-minute walk increases the annual
rate of school entrance by 2%; each year that a market has been within a 10minute walk increases the annual rate of school entrance by 3%; and each
year that transportation service has been within a 10-minute walk increases
the annual rate of school entrance by 3%. The rate of sending a child to
school, then, is much higher in neighborhoods where nonfamily services and
organizations have been nearby for a long period of time than in neighborhoods
where nonfamily services and organizations have been nearby for a short
period of time. For example, the difference between living in a neighborhood
that has had a market within a 10-minute walk for 25 years and living in a
neighborhood that has had a market within a 10-minute walk for 5 years
results in an 81% increase in the annual rate of entering school (1.03 20 p
1.81). Thus, the proliferation of nonfamily institutions, as measured by number
of years that nonfamily services and organizations have been within a 10minute walk of the neighborhood, increases the likelihood that parents will
send their (firstborn) child to school.43
The effects of social background, as measured by maternal characteristics,
are similar across the first set of models. Consistent with prior evidence on
educational attainment, both of the measures of education are significant,
indicating that (firstborn) children whose maternal grandfather and mother
have been formally educated are more likely to enter school. Across the first
set of models, the fixed measure of mother’s migration has no effect while
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Ann M. Beutel and William G. Axinn
123
Fig. 4.—Predicted equations for rates of school entrance for boys and girls by
presence of markets.
the time-varying measure of migration has a statistically significant effect on
the rate of school entrance. Turning to ethnicity, we see that, relative to highcaste Hindus, children of Newar women have higher rates of school entrance
and children of low-caste Hindus, Hill Tibetoburmese, and Terai Tibetoburmese women have lower rates of school entrance. These findings support the
idea suggested earlier that historical religious-ethnic group differences in the
value placed on education and access to education and other opportunities
influences the likelihood of sending children to school. Finally, mother’s birth
cohort is also significant. The earlier a mother’s birth cohort, the less likely
it is that her firstborn child enters school.
In the second set of models, we add an interaction term for the nonfamily
service or organization and gender of the child. As noted earlier, the interaction
term is used to test our hypothesis that, as new nonfamily services and organizations spread, the influence of gender on educational attainment will be
reduced. In fact, this is what we find. In each of the models, the interaction
term has a statistically significant effect, although it is only marginal (P !
.10) in the model of employment opportunity. These coefficients, which all
have values less than 1.00, indicate that the longer a nonfamily service or
organization has been nearby, the higher the rate of school entrance will be
for girls. We illustrate the relationship between gender and local-level social
change in figure 4. As an example, we plot the predicted equations by gender
of the child when a market has been within a 10-minute walk of the neighborhood for 1 year, 5 years, 15 years, and 25 years for a child who was 5
years of age at school entrance and whose mother had the following characteristics: her father had been formally educated, she had received 3 years
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Discrete-Time Hazards Estimates of the Effects of Gender, Nonfamily Services and Organizations, and Maternal
Characteristics on School Entrances
School
(1)
124
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
TABLE 2
Gender of child (1 p male)
Community characteristics:
No. of years service
or organization within
a 10-minute walk
No. of years service
or organization within
a 10-minute walk #
gender of child
Maternal characteristics:
Mother’s father’s education
Mother’s education
Mother’s migration
(fixed)
Mother’s migration
(time-varying)
Employment
(2)
(1)
(2)
Market
(1)
Transportation
(2)
(1)
(2)
1.51***
(4.65)
2.01***
(5.41)
1.51***
(4.63)
1.60***
(4.78)
1.51***
(4.60)
1.78***
(5.10)
1.50***
(4.55)
1.79***
(5.47)
1.01⫹
(1.53)
1.02***
(3.16)
1.02***
(3.22)
1.03***
(3.32)
1.03***
(5.71)
1.04***
(5.82)
1.03***
(4.84)
1.04***
(5.65)
.99⫹
(⫺1.40)
.98**
(⫺3.05)
1.54***
(3.72)
1.13***
(8.28)
.73
(⫺.80)
.77**
(⫺2.63)
1.55***
(3.75)
1.13***
(8.30)
.74
(⫺.75)
.77**
(⫺2.67)
1.51***
(3.53)
1.12***
(8.05)
.73
(⫺.82)
.74**
(⫺3.06)
1.50***
(3.46)
1.12***
(7.88)
.74
(⫺.76)
.74**
(⫺3.05)
.98**
(⫺2.39)
1.51***
(3.52)
1.11***
(7.04)
.69
(⫺.96)
.72***
(⫺3.28)
1.50***
(3.45)
1.11***
(6.89)
.70
(⫺.90)
.72***
(⫺3.29)
.97**
(⫺3.04)
1.45***
(3.19)
1.12***
(7.79)
.70
(⫺.89)
.74***
(⫺3.11)
1.44**
(3.10)
1.12***
(7.76)
.73
(⫺.79)
.73***
(⫺3.16)
Mother’s religious-ethnicity
(reference p high-caste
Hindu):
Low-caste Hindu
Newar
Terai Tibetoburmese
Mother’s birth cohort
(reference p 1962–71):
Born in 1952–61
Born in 1942–51
125
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Hill Tibetoburmese
x2
Degrees of freedom
Person-years (N)
.53***
(⫺4.32)
.79*
(⫺1.91)
1.63**
(2.61)
.38***
(⫺7.32)
.51***
(⫺4.46)
.78*
(⫺2.00)
1.62**
(2.57)
.37***
(⫺7.47)
.52***
(⫺4.38)
.77*
(⫺2.15)
1.51*
(2.18)
.39***
(⫺7.21)
.51***
(⫺4.46)
.76*
(⫺2.23)
1.48*
(2.05)
.38***
(⫺7.29)
.49***
(⫺4.79)
.75*
(⫺2.32)
1.39*
(1.73)
.35***
(⫺7.94)
.48***
(⫺4.94)
.74**
(⫺2.41)
1.34⫹
(1.54)
.34***
(⫺8.11)
.55***
(⫺4.03)
.76*
(⫺2.18)
1.55*
(2.33)
.39***
(⫺7.18)
.53***
(⫺4.19)
.75**
(⫺2.37)
1.49*
(2.11)
.38***
(⫺7.35)
.63***
(⫺4.22)
.41***
(⫺6.67)
1,571.75
21
4,809
.63***
(⫺4.23)
.41***
(⫺6.76)
1,581.11
22
4,809
.63***
(⫺4.29)
.42***
(⫺6.87)
1,579.74
21
4,809
.63***
(⫺4.31)
.41***
(⫺6.93)
1,581.68
22
4,809
.67***
(⫺3.62)
.48***
(⫺5.62)
1,602.09
21
4,809
.67***
(⫺3.67)
.47***
(⫺5.66)
1,607.80
22
4,809
.65***
(⫺3.96)
.44***
(⫺6.28)
1,592.60
21
4,809
.64***
(⫺4.06)
.44***
(⫺6.34)
1,601.88
22
4,809
Note.—Models also control for age. Estimates are presented as odd ratios, and t-values are given in parentheses.
⫹
P ! .10, one-tailed test.
* P ! .05, one-tailed test.
** P ! .01, one-tailed test.
*** P ! .001, one tailed test.
126
Economic Development and Cultural Change
of formal schooling prior to the start of the analysis interval, she had lived
outside the neighborhood prior to the analysis interval but not did live outside
the neighborhood during the analysis interval, she was Hill Tibetoburmese,
and she was born between 1952 and 1961.44 Figure 4 shows that for both
boys and girls, the longer a market has been within a 10-minute walk, the
greater the likelihood of entering school. However, the number of years a
market has been nearby has a stronger effect on entering school for girls than
for boys. As markets proliferate (i.e., as the number of years that a market
has been within a 10-minute walk increases), the gender gap in school entrance
declines because the rate of school entrance for girls rises faster than the rate
of school entrance for boys. The interaction terms with school, employment,
and transportation service, respectively, have similar effects and can be interpreted in the same way: the spread of each of these has a stronger effect
on the rate of school entry for girls than for boys. An examination of the
models by gender subgroups confirms this interpretation. In these models
(results not shown), the effect of the nonfamily service or organization is
either much weaker or insignificant for boys than for girls. Finally, it should
be noted that adding an interaction term to the models has virtually no effect
on the other coefficients in the models, except for gender, which has a stronger
effect in the models with interaction terms than the models without them.
B. School Exit
Table 3 shows the results of the models of school exit. This table is organized
the same way as table 2. We first examine the results for the models without
interaction terms. These results provide additional support for our hypothesis
that gender influences educational attainment. Gender has a large effect in
each of the models, with rates of school exit much lower among boys than
girls. Across the first set of models, being male reduces the odds of exiting
school in any year by approximately 40%.
However, the models of school exit provide only limited support for our
hypothesis that local-level social change influences individual-level educational attainment. The number of years that a school has been within a 10minute walk is significant (P ! .05 ), reducing the odds of school exit by 2%
per year, and the number of years that a market has been within a 10-minute
walk is marginally significant (P ! .10 ), reducing the odds of school exit by
1% per year. However, the number of years that either an employment opportunity or transportation has been within a 10-minute walk is not statistically
significant. Nevertheless, the difference in rates of school exit when nonfamily
services and organizations have been nearby for a long period of time versus
a short period of time is substantively interesting. For example, the difference
between living in a neighborhood that has had a school within a 10-minute
walk for 25 years and living in a neighborhood that has had a school within
a 10-minute walk for 5 years results in a 33% decrease in the annual rate of
exiting school (.98 20 p .67).
In comparison to the models of school entrance, the family’s educational
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Ann M. Beutel and William G. Axinn
127
background is somewhat less important in predicting school exit. Across the
first set of models, the measure of mother’s father’s education is not significant.
The measure of mother’s education, however, has a significant effect in each
of the models. Each year of formal schooling a mother has reduces the odds
of school exit. As in the models of school entrance, the fixed measure of
mother’s migration is not significant. The time-varying measure of migration
does have a significant effect in each model, indicating a greater likelihood
that (firstborn) children will exit school by age 15, but this effect is marginal.
The pattern of effects for religious-ethnic group on school exit mirrors those
for school entrance. That is, relative to high-caste Hindus, low-caste Hindus,
Hill Tibetoburmese, and Terai Tibetoburmese have higher rates of school exit
and Newars have lower rates of school exit. Finally, mother’s birth cohort
has a large effect. Women in the two oldest cohorts were much more likely
to have firstborn children leave school by age 15 than women in the youngest
cohort.
Turning to the models that include an interaction term for the nonfamily
service or organization and gender of the child, we find little evidence that
social change influences the relationship. Only in the model that examines
transportation is the interaction term significant. The coefficient of 1.04 in
this model indicates that the longer transportation service has been nearby,
the less likely girls are to exit school. An examination of the model by gender
subgroups (results not shown) indicates that transportation service has a significant negative effect on school exit for girls, but no significant effect on
school exit for boys. But overall, the second set of models provides little
evidence that local-level social change influences the relationship between
gender and school exit.
IV. Discussion and Conclusions
In keeping with the findings of much previous research, our study shows that
gender has an important impact on educational attainment. We find that, unlike
boys, girls are much less likely to enter school and much more likely to exit
school once they have entered. In contrast to much previous research, however,
we also examine whether social change at the community (i.e., local) level
influences individual-level educational attainment and whether communitylevel social change influences the relationship between gender and individual
educational attainment. While some previous studies have considered the relationship between social change and gender differences in schooling, more
often than not these studies have used measures of social change at the national
or regional levels. Although this research is useful, it is difficult to understand
how changes examined at those levels of analysis influence individual-level
decisions to send boys and girls to school. Studies that have examined social
characteristics at the community level, the level that is most relevant to individual lives, have used static measures of community characteristics, which
makes it difficult to discern the time ordering of the relationship between
social change and educational behaviors. The availability of unique data con-
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Discrete-Time Hazards Estimates of the Effects of Gender, Nonfamily Services and Organizations, and Maternal
Characteristics on School Exits
School
128
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
TABLE 3
Gender of child (1 p male)
Community characteristics:
No. of years service
or organization within
a 10-minute walk
No. of years service
or organization within
a 10-minute walk #
gender of child
Maternal characteristics:
Mother’s father’s education
Mother’s education
Mother’s migration
(fixed)
Mother’s migration
(time-varying)
Employment
Market
Transportation
(1)
(2)
(1)
(2)
(1)
(2)
(1)
(2)
.58***
(⫺3.81)
.51***
(⫺3.46)
.59***
(⫺3.70)
.56***
(⫺3.74)
.59***
(⫺3.74)
.53***
(⫺3.65)
.59***
(⫺3.71)
.52***
(⫺4.05)
.98*
(⫺1.89)
.98*
(⫺2.06)
.99
(⫺1.15)
.98⫹
(⫺1.38)
.99⫹
(⫺1.39)
.98
(⫺1.70)
1.02
(1.00)
.84
(⫺.90)
.84***
(⫺4.27)
1.07
(.11)
1.26⫹
(1.38)
.83
(⫺.93)
.84***
(⫺4.24)
1.08
(.12)
1.28⫹
(1.44)
1.02
(.86)
.85
(⫺.82)
.84***
(⫺4.20)
1.08
(.12)
1.30⫹
(1.57)
.85
(⫺.81)
.84***
(⫺4.15)
1.04
(.06)
1.31⫹
(1.60)
.99
(⫺.56)
1.02
(1.05)
.86
(⫺.77)
.84***
(⫺4.15)
1.04
(.06)
1.30⫹
(1.56)
.86
(⫺.75)
.84***
(⫺4.10)
1.04
(.06)
1.31⫹
(1.60)
.98⫹
(⫺1.46)
1.04*
(1.69)
.85
(⫺.79)
.84***
(⫺4.22)
1.08
(.12)
1.28⫹
(1.46)
.86
(⫺.78)
.84***
(⫺4.19)
1.03
(.04)
1.28⫹
(1.48)
Mother’s religiousethnicity
(reference p highcaste Hindu):
Low-caste Hindu
Newar
Terai Tibetoburmese
Mother’s birth cohort
(reference p 1962–71):
Born in 1952–61
Born in 1942–51
129
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Hill Tibetoburmese
2
x
Degrees of freedom
Person-years (N)
3.53***
(5.95)
1.54*
(2.31)
.46*
(⫺1.96)
3.60***
(6.30)
3.54***
(5.96)
1.54*
(2.30)
.45*
(⫺1.99)
3.61***
(6.31)
3.54***
(5.96)
1.55**
(2.34)
.46*
(⫺1.90)
3.49***
(6.13)
3.59***
(6.01)
1.56**
(2.36)
.47*
(⫺1.89)
3.52***
(6.17)
3.65***
(6.07)
1.55**
(2.34)
.47*
(⫺1.84)
3.64***
(6.34)
3.71***
(6.12)
1.56**
(2.37)
.48*
(⫺1.82)
3.65***
(6.34)
3.50***
(5.90)
1.54*
(2.28)
.45*
(⫺1.99)
3.51***
(6.16)
3.55***
(5.95)
1.56**
(2.36)
.45*
(⫺1.96)
3.55***
(6.21)
1.71**
(2.50)
2.77***
(4.37)
260.73
23
7,170
1.70**
(2.48)
2.79***
(4.40)
261.72
24
7,170
1.79**
(2.73)
3.01***
(4.83)
258.48
23
7,170
1.79**
(2.74)
3.06***
(4.87)
259.22
24
7,170
1.77**
(2.69)
2.90***
(4.61)
259.09
23
7,170
1.78**
(2.71)
2.95***
(4.66)
260.18
24
7,170
1.81**
(2.78)
3.05***
(4.84)
257.41
23
7,170
1.83**
(2.82)
3.16***
(4.95)
260.26
24
7,170
Note.—Models also control for time. Estimates are presented as odd ratios, and t values are given in parentheses.
⫹
P ! .10, one-tailed test.
* P ! .05, one-tailed test.
** P ! .01, one-tailed test.
*** P ! .001, one tailed test.
130
Economic Development and Cultural Change
taining dynamic measures of community characteristics and individual experiences has allowed us to establish the temporal ordering of communitylevel changes and individual behaviors and, in turn, carefully examine the
relationships among gender, social change, and educational attainment.
Our results show how examining changes in community characteristics
over time enhances our understanding of the relationships among gender, social
change, and educational attainment. We find that local-level social change,
specifically the spread of nonfamily services and organizations, influences
both boys’ and girls’ educational attainment. Moreover, we also find evidence
that the spread of nonfamily services and organizations has greater effects on
entering school for girls than for boys. This important difference produces a
decline in the gender gap in school entry. It seems likely that as the presence
and proximity of nonfamily services and organizations increases within a
community, gender norms and gender role attitudes change. However, it is
important to note that we found little evidence that the spread of nonfamily
services and organizations has an impact on the gender gap in school exit.
Despite the spread of nonfamily services and organizations, a large gender
difference in dropping out of school remains. It may be that gender norms
and gender role attitudes first change in ways that promote girls’ entrance
into school but not necessarily their continued enrollment in school. In fact,
dropping out of school may be a particularly important component of gender
differences in attainment as girls age, mature, and near involvement in familybuilding behaviors, such as marriage and childbearing, that may conflict with
continued schooling.
Along these lines, other data from our survey suggest that the reasons
for leaving school are linked to gender roles. Among the mothers in the survey
whose children attended school and then quit (before the survey), a similar
percentage of male and female children left school because they had failed
(24.5% of male children and 24.2% of female children, respectively), but
other reasons for leaving school differed considerably by gender. Only 2.4%
of male children left because they married, but 32.7% of female children left
school for that reason. In contrast, 16.0% of male children left school because
of a job, but only 0.7% of female children exited school for that reason. Thus,
family formation events appear to be a much more common reason for truncating educational attainment among females in this setting, and work events
appear to be a much more common reason for discontinuing education among
males. So, even in communities in which female school entry rivals male
school entry, family formation events, and societal expectations regarding
women’s behaviors and roles within their families still lead females to truncate
their education earlier than males.
This particular set of findings suggests that our understanding of the
relationships among gender, social change, and educational attainment would
be further enhanced by using direct and dynamic measures of gender norms
and gender role attitudes within communities. Karen Oppenheim Mason has
suggested that such measures would be useful for understanding the dynamics
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Ann M. Beutel and William G. Axinn
131
of change in women’s fertility and mortality.45 They undoubtedly would be
useful for understanding the dynamics of change in girls’ and women’s educational participation and attainment as well.
Notes
* This article was written while the first author was funded by a National Institute
of Child Health and Human Development postdoctoral traineeship through the Population Studies Center at the University of Michigan (no. HD07339). We wish to thank
the National Institute of Child Health and Human Development for their generous
support of both the data collection and the data analyses reported in this manuscript
(no. R01 HD32912). We also wish to thank Jennifer Barber, Dirgha Ghimire, Lisa
Pearce, and the staff of the Population and Ecology Research Laboratory for their
assistance in gathering the data used in these analyses; Cathy Sun and Lisa Neidert
for their work in preparing data files; and Scott Yabiku for his help conducting the
analyses. Any errors or omissions remain solely our responsibility.
1. Peter M. Blau and Otis D. Duncan, The American Occupational Structure
(New York: Wiley & Sons, 1967); Otis D. Duncan, David L. Featherman, and Beverly
Duncan, Socioeconomic Background and Achievement (New York: Seminar, 1972);
William H. Sewell and Robert M. Hauser, Education, Occupation, and Earnings:
Achievement in the Early Career (New York: Academic Press, 1975); James Coleman,
Foundations of Social Theory (Cambridge, Mass.: Harvard University Press, 1990);
Gary S. Becker, Human Capital: A Theoretical and Empirical Analysis, with Special
Reference to Education, 3d ed. (Chicago: University of Chicago Press, 1993).
2. William H. Sewell and Vimal P. Shah, “Socioeconomic Status, Intelligence,
and the Attainment of Higher Education,” Sociology of Education 40, no. 1 (Winter
1967): 1–23, and “Parents’ Education and Children’s Educational Aspirations and
Achievements,” American Sociological Review 33, no. 2 (April 1968): 191–209; Karl
L. Alexander and Bruce K. Eckland, “Sex Differences in the Educational Attainment
Process,” American Sociological Review 39, no. 5 (October 1974): 668–82; Margaret
Mooney Marini, “The Transition to Adulthood: Sex Differences in Educational Attainment and Age at Marriage,” American Sociological Review 43, no. 4 (August
1978): 483–507, and “Sex Differences in the Process of Occupational Attainment: A
Closer Look,” Social Science Research 9, no. 4 (December 1980): 307–61. However,
in the United States, patterns of educational attainment have changed in recent years
such that a higher percentage of young women than young men earn bachelor’s and
master’s degrees (U.S. Bureau of the Census, “Educational Attainment in the United
States,” Current Population Reports, Series P20-528 [Washington, D.C.: U.S. Government Printing Office, 2000]).
3. Sewell and Shah, “Parents’ Education and Children’s Educational Aspirations
and Achievements,” and “Social Class, Parental Encouragement, and Educational Aspirations,” American Journal of Sociology 73, no. 5 (March 1968): 559–72. Jerry A.
Jacobs (“Gender Inequality and Higher Education,” in Annual Review of Sociology,
ed. John Hagan [Palo Alto, Calif.: Annual Reviews, 1996], pp. 153–85), citing research
by Jere R. Berhman, Robert A. Pollak, and Paul Taubman (“Do Parents Favor Boys?”
International Economic Review 27, no. 1 [February 1986]: 33–54) and Robert M.
Hauser and Hsiang-Hui Daphne Kuo (“Does the Gender Composition of Sibships
Affect Educational Attainment?” Working Paper no. 95-06 [Center for Demography
and Ecology, University of Wisconsin—Madison, July 1997]), notes that, in the contemporary United States, the similar levels of educational attainment among young
adult males and females suggest that parents’ educational aspirations do not differ by
gender of the child nowadays.
4. Marini, “The Transition to Adulthood, ” and “Sex Differences in the Process
of Occupational Attainment.”
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
132
Economic Development and Cultural Change
5. Arland Thornton, Duane F. Alwin, and Donald Camburn, “Causes and Consequences of Sex-Role Attitudes and Attitude Change,” American Sociological Review
48, no. 2 (April 1983): 211–27; S. Philip Morgan and Linda J. Waite, “Parenthood
and the Attitudes of Young Adults,” American Sociological Review 52, no. 4 (August
1987): 541–47.
6. Marini, “The Transition to Adulthood”; Ronald R. Rindfuss, C. Gray Swicegood, and Rachel Rosenfeld, “Disorder in the Life Course: How Common Is It, and
Does It Matter?” American Sociological Review 52, no. 6 (December 1987): 785–801.
7. Audrey Chapman Smock, Women’s Education in Developing Countries: Opportunities and Outcomes (New York: Praeger, 1981); Dean T. Jamison and Marlaine
E. Lockheed, “Participation in Schooling: Determinants and Learning Outcomes in
Nepal,” Economic Development and Cultural Change 35, no. 2 (January 1987):
279–306; Elizabeth King, Educating Girls and Women: Investing in Development
(Washington, D.C.: World Bank, 1990); M. Anne Hill and Elizabeth M. King,
“Women’s Education in Developing Countries: An Overview,” in Women’s Education
in Developing Countries: Barriers, Benefits, and Policies, ed. Elizabeth M. King and
M. Anne Hill (Baltimore: Johns Hopkins University Press, 1993), pp. 1–50; United
Nations Development Programme, Human Development Report, 1999 (New York:
Oxford University Press, 1999).
8. Jamison and Lockheed; Nelly P. Stromquist, “Determinants of Educational
Participation and Achievement of Women in the Third World: A Review of the Evidence and a Theoretical Critique,” Review of Educational Research 59, no. 2 (Summer
1989): 143–83.
9. Meena Acharya and Lynn Bennett, The Rural Women of Nepal: An Aggregate
Analysis and Summary of Eight Village Studies (Kathmandu: Centre for Economic
Development and Administration, 1981); Shrtii Shakti, Women, Development, Democracy: A Study of the Socio-Economic Change in the Status of Women in Nepal,
1981–1993, prepared for USAID, DANIDA, and CCO (Kathmandu: Shtrii Shakti,
1995).
10. Jacqueline A. Ashby, “Equity and Discrimination among Children: Schooling
Decisions in Rural Nepal,” Comparative Education Review 29, no. 1 (February 1985):
68–80; Stromquist; Shahrukh R. Khan, “South Asia,” in King and Hill, eds. (n. 7
above), pp. 211–46.
11. Mead Cain, “Risk, Fertility, and Family Planning in a Bangladesh Village,”
Studies in Family Planning 11, no. 6 (June 1980): 219–23, “Risk and Insurance:
Perspectives on Fertility and Agrarian Change in India and Bangladesh,” Population
and Development Review 7, no. 3 (September 1981): 435–74, and “Perspectives on
Family and Fertility in Developing Countries,” Population Studies 36, no. 2 (July
1982): 159–75; John C. Caldwell, Theory of Fertility Decline (New York: Academic
Press, 1982).
12. Tim Dyson and Mick Moore, “On Kinship Structure, Female Autonomy, and
Demographic Behavior in India,” Population and Development Review 9, no. 1 (March
1983): 35–60; Ashby; John C. Caldwell, P. H. Reddy, and Pat Caldwell, “Educational
Transition in Rural South India,” Population and Development Review 11, no. 1 (March
1985): 29–51; Jamison and Lockheed.
13. Ashby; Caldwell, Reddy, and Caldwell; Hill and King, “Women’s Education
in Developing Countries.”
14. Blau and Duncan; Coleman; Becker (all in n. 1 above).
15. Adam Smith, An Inquiry into the Nature and Causes of the Wealth of Nations
(London: Strahan & Cadell, 1776); Coleman.
16. Emile Durkheim, The Division of Labor in Society (1933; reprint, New York:
Free Press, 1984), p. 257.
17. Ibid., pp. 257–58.
18. Arland Thornton and Thomas E. Fricke, “Social Change and the Family:
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
Ann M. Beutel and William G. Axinn
133
Comparative Perspectives from the West, China, and South Asia,” Sociological Forum
2, no. 2 (Fall 1987): 746–79; Arland Thornton and Hui-Seng Lin, Social Change and
the Family in Taiwan (Chicago: University of Chicago Press, 1994).
19. Caldwell; Ron Lesthaeghe and Chris Wilson, “Modes of Production, Secularization, and the Pace of Fertility Decline in Western Europe, 1870–1930,” in The
Decline of Fertility in Europe, ed. Ansley Coale and Susan Cotts Watkins (Princeton,
N.J.: Princeton University Press, 1986), pp. 261–92.
20. William F. Ogburn and Meyer F. Nimkoff, Technology and the Changing
Family (Boston: Houghton Mifflin, 1955); Thornton and Fricke.
21. Prior to the 1950s, education that did take place outside the family was largely
in the form of religious training at Buddhist monasteries or Sanskrit schools. In addition, some Nepalis were formally educated at Indian schools (Dor Bahadur Bista,
Fatalism and Development: Nepal’s Struggle for Modernization [Calcutta: Orient
Longman, 1991]). But for the vast majority of Nepalis, education took place within
the family.
22. W. D. Halls, “United Kingdom: System of Education,” in The International
Encyclopedia of Education, 2d ed., ed. Torsten Husén and T. Neville Postlethwaite
(New York: Pergamon, 1994), 11:6515–23; Kim P. Sebaly, “Nepal,” in World Education Encyclopedia, ed. George Thomas Kurian (New York: Facts on File, 1988), 2:
904–10. However, the structure of primary and secondary levels of education in Nepal
has changed several times (T. R. Khaniya and M. A. Kiernan, “Nepal: System of
Education,” in The International Encyclopedia of Education, 2d ed., ed. Torsten Husén
and T. Neville Postlethwaite [New York: Pergamon, 1994], 7:4060–67). Currently, the
primary level is made up of grades 1–5, the lower secondary level of grades 6–8, the
secondary level of grades 9–10, and the higher secondary level—a relatively recent
innovation—of grades 11–12. Students may enter the higher secondary level if they
successfully pass the school leaving certificate (SLC) examination at the end of grade
10. Ten-year school graduates can earn a university degree after 5 years of study, while
12-year school graduates can earn one after 3 years of study (Nepal South Asia Centre,
Nepal Human Development Report, 1998 [Kathmandu: Nepal South Asia Centre,
1998]).
23. United Nations, The World’s Women, 1995: Trends and Statistics (New York:
United Nations, 1995).
24. Ibid.
25. Nepal South Asia Centre.
26. These methods are described in detail elsewhere—see William G. Axinn,
Jennifer S. Barber, and Dirgha J. Ghimire, “The Neighborhood History Calendar: A
Data Collection Method Designed for Dynamic Multilevel Modeling,” in Sociological
Methodology, ed. Adrian E. Raftery (Cambridge, Mass.: Blackwell, 1997), pp. 355–92.
27. Nepal South Asia Centre.
28. Shakti (n. 9 above).
29. Axinn, Barber, and Ghimire.
30. For a detailed description, see William G. Axinn, Lisa D. Pearce, and Dirgha
J. Ghimire, “Innovations in Life History Calendar Applications,” Social Science Research 28, no. 3 (September 1999): 243–64.
31. Axinn, Barber, and Ghimire.
32. Of course, with these data, calculation of any other distance threshold is
equally straightforward. Therefore, we reestimated our models using a variety of different distance thresholds. Use of other distances, such as the number of years that
nonfamily organizations and services were within a 5-minute walk, did not appreciably
affect our results.
33. For example, Blau and Duncan (n. 1 above).
34. William H. Sewell, Archibald O. Haller, and Alejandro Portes, “The Educational and Early Occupational Attainment Process,” American Sociological Review
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms
134
Economic Development and Cultural Change
34, no. 1 (February 1969): 82–92; William H. Sewell, Archibald O. Haller, and George
W. Ohlendorf, “The Educational and Early Occupational Status Attainment Process:
Replication and Revision,” American Sociological Review 35, no. 6 (December 1970):
1014–27; Sewell and Hauser (n. 1 above).
35. Nan Lin and Daniel Yauger, “The Process of Occupational Status Attainment:
A Preliminary Cross-National Comparison,” American Journal of Sociology 81, no.
3 (November 1975): 543–62; Jonathan Kelly, Robert V. Robinson, and Herbert S.
Klein, “A Theory of Social Mobility with Data on Status Attainment in a Peasant
Society,” in Research in Social Stratification and Mobility: A Research Annual, ed.
Donald J. Treiman and Robert V. Robinson (Greenwich, Conn.: JAI, 1981), pp. 27–66;
Donald J. Treiman and Kin-Bor Yip, “Educational and Occupational Attainments in
21 Countries,” in Cross-National Research in Sociology, ed. Melvin L. Kohn (Newbury
Park, Calif.: Sage, 1989), pp. 373–94.
36. For a summary of studies, see Hill and King, “Women’s Education in Developing Countries” (n. 7. above).
37. Ibid.
38. We also estimated models containing a second measure of mother’s education,
which was a time-varying measure of the number of years of education she received
during the analysis interval—i.e., starting from when her firstborn was 3 years of age
in the school entrance analysis and starting from the year her firstborn entered school
in the exit analysis. Models containing the time-varying measure of mother’s education
produced results virtually identical to the ones reported here.
39. Robin Stryker, “Religio-ethnic Effects on Attainments in the Early Career,”
American Sociological Review 46, no. 2 (April 1981): 212–31.
40. Ashby (n. 10 above); Jamison and Lockheed (n. 7 above).
41. Alternative controls for migration, such as deleting respondents who experienced any migration or deleting those person-years when migration occurred, produced estimates nearly identical to those presented here.
42. Paul D. Allison, “Discrete-Time Methods for the Analysis of Event Histories,”
in Sociological Methodology, ed. Samuel Leinhardt (San Francisco: Jossey-Bass,
1982), pp. 61–98, and Event History Analysis: Regression for Longitudinal Event Data
(Beverly Hills, Calif.: Sage, 1984); Trond Petersen, “Estimating Fully Parametric Hazard Rate Models with Time-Dependent Covariates: Use of Maximum Likelihood,”
Sociological Methods and Research 14, no. 3 (February 1986): 219–46, and “The
Statistical Analysis of Event Histories,” Sociological Methods and Research 19, no.
3 (February 1991): 270–323.
43. Although there are substantial positive correlations among measures of community change over time, we also estimated models of school entrance and exit including all community-level measures of social change in a single model. As one
might expect, we found that the effects of the nonfamily services and organizations
were substantially attenuated when all were included in the same model. However,
because our aim is to evaluate the total impact of each dimension of social change,
we present separate models for each measure of social change in the tables.
44. The negative values on the y-axis of fig. 4 are an artifact of the values of
control variables used to calculate the predicted values displayed in fig. 4. The predicted
values themselves have no substantive interpretation—we display fig. 4 to highlight
the difference in slopes between the male and female effects.
45. Karen Oppenheim Mason, “Conceptualizing and Measuring Women’s Status”
(paper presented at the annual meeting of the Population Association of America,
Miami, April 1994).
This content downloaded from
144.122.207.25 on Tue, 03 Jan 2023 15:22:15 UTC
All use subject to https://about.jstor.org/terms