Hull 2009

THE CHANGING RELATIONSHIP BETWEEN FAMILY
SIZE AND EDUCATIONAL ATTAINMENT OVER THE

COURSE OF SOCIOECONOMIC DEVELOPMENT:
EVIDENCE FROM INDONESIA*
VIDA MARALANI
Many studies from developed countries show a negative correlation between family size and chil-
dren’s schooling, while results from developing countries show this association ranging from positive
to neutral to negative, depending on the context. The body of evidence suggests that this relationship
changes as a society develops, but this theory has been difcult to assess because the existing evidence
requires comparisons across countries with different social structures and at different levels of devel-
opment. The world’s fourth most populous nation in 2007, Indonesia has developed rapidly in recent
decades. This context provides the opportunity to study these relationships within the same rapidly
developing setting to see if and how these associations change. Results show that in urban areas, the
association between family size and children’s schooling was positive for older cohorts but negative
for more recent cohorts. Models using instrumental variables to address the potential endogeneity of
fertility conrm these results. In contrast, rural areas show no signicant association between family
size and children’s schooling for any cohort. These ndings show how the relationship between family
size and children’s schooling can differ within the same country and change over time as contextual
factors evolve with socioeconomic development.
N umerous studies of educational attainment in the United States have shown that school-
ing is negatively correlated with sibship size. That is, children with fewer brothers and sis-
ters obtain more schooling than those with more siblings. This negative association exists
for many different measures of children’s human capital, including completed schooling,
standardized test scores, and grades, and holds even after family socioeconomic character-
istics are controlled (Blake 1989; Featherman and Hauser 1978; see Steelman et al. 2002
for a comprehensive review of this literature). In the sociological literature, this nding is
often explained using an argument of nite resources: parents have limited time, money,
and patience to devote to the education of their children, and those with fewer children can
invest more per child. This theory of resource dilution ts well with the classic notion of the
quality-quantity trade-off in family economics (Becker 1991; Becker and Tomes 1976).1
In recent years, two lines of research have called into question this seemingly robust
negative relationship between family size and children’s schooling. First, some scholars
have argued that this nding might be biased or spurious. If this negative association is
explained by factors such as unobserved family characteristics that are not controlled in
the analysis, then our understanding of the true relationship between these variables may
*Vida Maralani, Department of Sociology, Yale University, P.O. Box 208265, New Haven, CT 06520-8265;
e-mail: [email protected]. I gratefully acknowledge the William and Flora Hewlett Foundation, the National
Institute of Child Health and Human Development, and the Robert Wood Johnson Foundation’s Health & Society
Scholars program for their nancial support. I owe many thanks to Elizabeth Frankenberg, Jinyong Han, Robert
Mare, William Mason, Douglas McKee, Judith Seltzer, Duncan Thomas, Donald Treiman, and three anonymous
reviewers for their valuable comments and advice.
1. The competing “conuence theory” suggests that additional children lower a family’s average intellectual
environment, leaving children in larger sibships and those with higher birth order worse off (Zajonc and Markus
1975). Empirical analyses, however, have not found much support for this theory (Hauser and Sewell 1985; Reth-
erford and Sewell 1991).
Demography, Volume 45-Number 3, August 2008: 693–717 693
694 Demography, Volume 45-Number 3, August 2008
be erroneous (Guo and VanWey 1999b). Moreover, few studies have addressed the concern
that fertility and children’s schooling are jointly determined, despite evidence suggesting
this is the case in some settings or sociohistorical periods (Axinn 1993; Caldwell, Reddy,
and Caldwell 1985; Wolfe and Behrman 1984). Indeed, the very notion of a quality-quantity
trade-off suggests that these processes are interrelated in complex ways that contradict the
basic assumptions of regression analysis or xed-effects models. Using data from Norway
and Israel, for example, recent research in economics has shown that once fertility and
child outcomes are modeled jointly, these variables display no meaningful association
(Angrist, Lavy, and Schlosser 2006; Black, Devereux, and Salvanes 2005). Qian (2005),
in contrast, found mixed results using a similar approach with data from China. She found
that an increase from zero to one sibling has a positive effect on children’s schooling, while
an increase from one to two siblings has a negative effect. Further research that addresses
these issues, especially in a range of countries, will advance our understanding of the extent
of such potential bias.
Second, a growing literature on the nature of these relationships in the developing
world shows that the negative correlation found so consistently in more developed coun-
tries is not necessarily generalizeable. Instead, the association between family size and
children’s outcomes varies greatly by time and place and ranges from negative to positive,
depending on the specic context (see Buchmann and Hannum 2001 for an overview of
this literature). A few examples sufce to show the range. The evidence from Thailand
and Brazil suggests a negative association between family size and educational attainment
(Knodel, Havanon, and Sittitrai 1990; Psacharopoulos and Arriagada 1989). Results from
Vietnam show that the association is negative only for families with six or more children,
and effects are modest when other family characteristics are controlled (Anh et al. 1998).
The evidence from Botswana and Kenya, on the other hand, suggests the reverse is true:
educational attainment has a positive relationship with family size (Chernichovsky 1985;
Gomes 1984). Even within the same country, studies show that patterns differ by subgroup.
Among Israeli Jews, for example, family size has a negative association with educational
attainment. Among Israeli Muslims, who are less advantaged socioeconomically, live in
less urban settings, have extended kinship networks, and have much higher fertility rates,
family size and educational attainment are not associated (Shavit and Pierce 1991; but see
Angrist et al. 2006).2
This array of evidence suggests that the relationship between family size and educa-
tional attainment is likely related to a society’s level of development, modes of production,
and access to schooling, which in turn shape the relative inuence of the family on the
schooling of children (Desai 1995; King 1987; Lloyd 1994). In certain contexts or at certain
stages of development, having more siblings to share household and labor market work may
provide children with more resources for schooling. In some settings, the quality-quantity
trade-off may not hold, and the desire to have better-educated children may not neces-
sarily lead parents to choose smaller families (Gomes 1984; Mueller 1984). These macro
socioeconomic mechanisms relating family size and children’s schooling might include the
availability of schools, transportation and communication infrastructure, and participation
in labor-intensive production such as agriculture. These mechanisms might be most appar-
ent when comparing less developed rural contexts with more developed urban ones.
Furthermore, context-specic factors such as family organization and cultural roles
determine wealth ows between parents and children, whether the burden of child rear-
ing is limited to the nuclear family or extended across broader kin networks, whether and
how much school-aged children work inside and outside the home, and whether these
factors change as societies develop or as overall levels of educational attainment increase
2. Except for Angrist et al. (2006), all of these studies assumed that fertility is exogenous to children’s
schooling.
Family Size and Educational Attainment With Socioeconomic Development 695
(Caldwell et al. 1985). In societies where parents bear most of the cost of schooling and
where the costs are high, we might expect a negative relationship between family size and
educational attainment. In societies with extended kinship networks and lower schooling
costs, the relationship may be neutral or positive (Lloyd and Blanc 1996). The effects of
family size on children’s schooling may also differ within families by factors such age,
sex, or birth order of children (Lloyd and Gage-Brandon 1994; Parish and Willis 1993).
Context-specic mechanisms relating family size to children’s schooling might include
nancial relationships within families, norms about how much schooling children should
complete, parents’ preferences by child sex or birth order, labor market returns to schooling,
and the costs of schooling.
The evidence also highlights the interdependence of family size, family structure, and
educational attainment. Family size and educational attainment are likely to be jointly de-
termined, at least to some degree, with families choosing the level of fertility that is likely
to produce children with the preferred level of education for a given family, context, or
society. The relationship between family size and children’s educational attainment can
have demographic feedbacks as well. Small families may raise educational attainment,
which in turn may lower fertility in the next generation. Moreover, if the effect of family
size grows more negative or positive over time, these aggregate demographic relationships
may intensify or accelerate.3
Over time, the relationship between family size and children’s schooling might change
as development brings changes in income, consumption, urbanization, migration, educa-
tional opportunities, gender roles, and family organization. Government policies to invest
in women’s schooling change women’s relative position in families and society and update
norms about schooling girls. Of course, these updated norms also double the number of
children families must educate relative to when only sons were schooled. Industrializa-
tion, modernization, and an expanding formal job sector increase the returns and hence
the demand for schooling. But higher levels of schooling may be cost prohibitive despite
families’ growing expectations for their children’s education. Other changing benets of
education might include more labor market opportunities for women and improvements in
health and home production, which could also impact the relationship between family size
and children’s education.
Socioeconomic development brings increased differentiation between urban and rural
areas, which might translate to differences in the association between family characteristics
and children’s schooling by geographical location. Urbanization improves access to schools
by improving transportation and communication infrastructure and allowing closer proxim-
ity to this infrastructure. As the formal job sector grows, urban residents have better access
to these jobs, which require more investment in schooling. Although city dwellers might be
richer on average, costs of living are also higher there, making the marginal expense of an
additional child different between urban and rural areas. The agrarian and informal econo-
my of rural areas, in contrast, might allow families more exibility in resource sharing and
balancing home, work, and school responsibilities. Educational opportunities, however,
might be quite limited in rural areas, and the formal sector more difcult to enter.
Taken together, the existing evidence from developed and developing nations suggests
that there is no axiomatic relationship between family size and children’s schooling. Rather,
this relationship varies by context and by levels of modernization and development. Much
of the existing evidence, however, requires us to compare across different cultures and
socioeconomic settings to determine the role of development in shaping the relationship
between family size and children’s schooling. Although it may be tempting to use an ar-
ray of cross-sectional evidence from across the world to infer something about the role of
3. See Preston (1976) for a precise demonstration of the relationship between the average number of children
ever born for a cohort of women and the average sibship size of the offspring of those women.
socioeconomic development, family organization, or government policies, comparing across

different cultures and socioeconomic contexts can be difcult and misleading (Thornton
2001). Comparisons across time within the same country help with this problem.
Although numerous studies have examined the changing association between family
size and children’s schooling across cohorts in the United States and Europe (de Graaf and
Huinink 1992; Hauser and Featherman 1976; Kuo and Hauser 1995), few analysts have
examined cohort trends in a developing country. The few studies that considered changes
within the same developing country generally used only two cohorts (e.g., Hermalin, Selt-
zer, and Lin 1982; Pong 1997; Sudha 1997; see Parish and Willis 1993 for an exception).
These offer only a limited view of the development process. One way to improve our un-
derstanding of these relationships is to study these associations across many different birth
cohorts within one rapidly developing country. Indonesia offers such a context.
Indonesia, the world’s fourth most populous country in 2007, has experienced enormous
socioeconomic growth in recent decades. Between 1970 and the late 1990s, Indonesia had
one of the fastest-growing economies in the world. Over this period, the infant mortality and
total fertility rate (TFR) declined dramatically, and life expectancy increased by almost 20
years (Indonesia Central Bureau of Statistics 2003; UNICEF 2000). There were also large
gains in educational attainment, especially for women. Overall, participation in agriculture
declined and industry grew. This setting provides a unique opportunity to examine how the
relationship between family size and educational attainment changes as a society develops.
This study examines the association between family size/composition and educational
attainment for four cohorts of Indonesians born between 1948 and 1991. To preview the
results, large families are associated with signicantly better educational outcomes for the
oldest urban birth cohort but are associated with signicantly poorer educational outcomes
in more recent urban cohorts. Instrumental variables analysis shows that this negative rela-
tionship in the recent urban cohorts remains even in models that address the possibility that
fertility and children’s schooling are jointly determined. In rural areas, in contrast, there
is little evidence of a signicant association between family size and children’s schooling
for any cohort.
THE INDONESIAN CONTEXT
Indonesia is an archipelago nation of thousands of islands, hundreds of ethnicities, and
nearly 235 million inhabitants in 2007. Following 300 years of Dutch colonization and
several years of Japanese occupation during World War II, Indonesia gained independence
in 1949. During the 1950s and 1960s, most Indonesians lived in rural areas, and poverty
was endemic. The analysis starts with the 1948–1957 birth cohort, whose members were
school-aged from the mid-1950s to the late 1960s. The Indonesian economy was largely
agricultural in 1960, with about 75% of the labor force engaged in agriculture and 85%
living in rural areas (UNICEF and Government of Indonesia 1988; Walton 1985). This was
a period marked by severe economic strain, hyperination, and infant mortality rates of
about 150 per 1,000 births (Hugo et al. 1987). Average life expectancy at birth was about
41 years, the real per capita gross national product (GNP) was $125 (USD), and only 27%
of women ages 20 to 24 were literate (Hull 1987; Jones 1990).
A military coup in 1965 brought a shift in political power and a new focus on domestic
development. Starting in 1967, the economy began to grow rapidly as the nation began to
export oil and natural gas and attract foreign investment (Walton 1985). The 1958–1967
cohort was born during this time of transition and was school-aged just as the Indonesian
economy began a period of massive growth. Between 1970 and the late 1990s, Indonesia
went from being one of the poorest countries in the world to one at a middle level of
socioeconomic development.
Indonesia used windfall prots from rising oil prices in the 1970s to nance an exten-
sive educational expansion at the primary school level. From 1974 to 1979, the government
built more than 61,000 primary schools and abolished tuition at public primary schools
(Duo 2001; Oey-Gardiner 1997). Primary school enrollments rose from 60% in 1974 to
94% by 1984 (Jones and Hull 1997). Meanwhile, the dramatic pace of socioeconomic de-
velopment continued. Between 1970 and 1980, the infant mortality rate fell from 118 to 98
per 1,000 births, life expectancy rose from 47 to 53 years, and female literacy grew from
47% to 66% (Firman 1997; Iskandar 1997; Jones 1990; Jones and Hull 1997). By the early
1980s, only 55% of the labor force participated in agriculture, and real per capita GNP had
risen to $447 (USD) (Hull 1987; Walton 1985). The 1968–1977 birth cohort was school-
aged at the height of this period of growth and development.
After a period of turbulent oil prices and economic adjustment, the economy continued
to grow in the 1980s. By the early 1990s, the infant mortality rate had declined to 66 per
1,000, female literacy was 93%, primary school enrollment had risen to 98%, and 31% of
Indonesia’s population lived in urban settings (Cobbe and Boediono 1993; Jones and Hull
1997). The youngest birth cohort, born between 1978 and 1981, was school-aged in the late
1980s and 1990s after more than 20 years of sustained growth and development.
These four birth cohorts straddle another important part of Indonesia’s development
experience. Historically in Indonesia, fertility was quite high and exhibited an inverted U-
shaped pattern with women’s educational attainment (women with some primary or middle
school experience had higher fertility than women with no schooling or with 12 or more
years). This pattern was stronger in urban than in rural areas. Also, average fertility was
higher in urban areas, and fertility exhibited a positive relationship with household income
(Cobbe and Boediono 1993; Hull and Hull 1977).
The 1970s and 1980s marked the implementation of Indonesia’s state-sponsored family
planning program (Gertler and Molyneaux 1994). The program distributed contraceptives
and educated women on how to use them, promoted two-child families, and encouraged
women to postpone marriage. The program aimed to both increase contraceptive supply
and spur contraceptive demand by changing fertility preferences and norms. Concomitant
with sustained economic growth and improvements in women’s socioeconomic position,
the program was hugely successful. Between 1970 and the early 1990s, TFR declined
from 5.6 to 2.9 children per woman (Gertler and Molyneaux 1994; Jones and Hull 1997).
Women’s reports of their ideal family sizes also declined from an average of 4.2 in 1972 to
about 3.2 in 1987 (Jones 1990). By the 1980s, the inverted U-shaped pattern of fertility by
women’s schooling began to atten out and disappear, especially in urban areas. Moreover,
the positive correlation between household wealth and fertility declined, and a negative cor-
relation emerged. The two oldest cohorts (1948–1957 and 1958–1967) were born before the
implementation of the family planning program, and the more recent cohorts (1968–1977
and 1978–1981) were born after.
The formal school system in Indonesia has four basic levels. Primary school covers
grades 1 to 6; junior secondary school covers grades 7 to 9; senior secondary school cov-
ers grades 10 to 12; and postsecondary covers all years from 13 onward. National nal
examinations separate each level of formal schooling. Urban-rural disparities in schooling
were large in the 1960s and 1970s and have narrowed in more recent decades. Still, en-
rollment ratios after age 12 in rural areas in the 1990s are similar to those in urban areas
in the 1970s (Oey-Gardiner 1997: gure 8.8). Although primary schooling has become
widespread throughout Indonesia, higher educational levels remain quite expensive, and
enrollment differences between urban and rural areas are substantial at these higher school
levels (Oey-Gardiner 1997).
DATA AND METHODS

This analysis is based on the 1993 and 1997 waves of the Indonesian Family Life Survey
(IFLS), a comprehensive longitudinal socioeconomic and health survey. The survey rep-
resents an area that includes 83% of Indonesia’s population (see Frankenberg and Thomas
2000 for detailed documentation on the IFLS). The sample used in this study includes
approximately 3,200 families with more than 13,000 living children. All variables are
measured as of 1997, except for those respondents who died between 1993 and 1997 or
households not found in 1997 (about 5% of the 1993 sample). For these cases, I use the
information provided in 1993, except for the schooling of respondents ages 19 or younger,
which I leave missing because it is likely to be censored and systematically underreported
when compared with the 1997 data. Overall, the data are quite complete, and the variables
used in the analyses include little missing data.
For each ever-married female respondent (Indonesia is a society with nearly universal
marriage and essentially no nonmarital fertility), I assemble a full count of all children,
alive or dead, coresident or living elsewhere. I also identify the schooling of each woman’s
husband, whether current or former. These women’s children are the units of analysis.
The woman and her husband are the parents, and their children constitute the sibship.4 For
mothers, fathers, and children, I measure educational attainment by years of completed
schooling. I control for mother’s age in all models and report robust standard errors that
correct for clustering of multiple children born to the same woman. I also control for prov-
ince of residence, a key sampling stratum, in all statistical models.
Because urban and rural areas are quite different in Indonesia, especially in terms of
resources and infrastructure, I estimate separate models by rural status. To identify whether
a child grew up in an urban or rural location, I use the mother’s detailed migration history to
determine each child’s region of residence when he or she was age 12. In their retrospective
reports, women reported whether each of their places of residence was a village, small town,
or big city. I classify villages as rural, and small towns and big cities as urban areas.5
I assemble four birth cohorts spanning ages 16 to 49 in 1997. The rst three cohorts are
restricted to families containing mothers who are at least 41 years old and children 20 years
and older in order to measure completed fertility and completed schooling. This sample of
children, born between 1948 and 1977, provides information on three broad cohorts: those
born in 1948–1957, 1958–1967, and 1968–1977. I examine the relationship between family
size and completed education by cohort and urban-rural status using ordered probit models
with completed education classied in categories as the dependent variable and family size
and household characteristics as independent variables.6 Children’s education is categorized
in the most basic levels of the Indonesian educational system: none, primary (1–6 years),
junior secondary (7–9 years), senior secondary (10–12 years), and postsecondary (13 or
more years). To these three birth cohorts I add a fourth cohort of young children, those
born between 1978 and 1981 (ages 16 to 19 in 1997). Here, there are no restrictions on the
mother’s age, and fertility is not completed in many cases. Because many members of this
most recent birth cohort have not completed their schooling, I use continuation beyond
primary school as the measure of educational attainment for this cohort.
Educational attainment is correlated with household wealth, especially in developing
countries (Chernichovsky and Meesook 1985; Filmer and Pritchett 1999; Hull and Hull
1977). The analyses shown below do not control for household wealth because the data do
not include an appropriate measure of family wealth for most individuals used in this study,
namely a measure of family wealth when the individuals were school-aged. The IFLS does
collect extensive information about assets, consumption, and expenditures at the time of the
4. I use the terms sibship size and family size interchangeably. These terms mean all of a child’s brothers and
sisters plus the child himself/herself. When I count a child’s siblings, I use the sibship size minus one.
5. About two-thirds of Indonesia’s total population and 70% of its urban population reside on Java. Java and
the outer islands have a mix of urban and rural populations, although the proportion urban differs across islands.
In the 1990s, Java was about 40% urban compared with an average of 26% for the other areas (Firman 1997).
6. About 7% of those in the 1968–1977 cohort were enrolled in school in 1997 (concentrated at ages 20 to
23). More than 90% of these respondents had completed grade 12 or higher, with a median and mode of 15 years of
schooling. Therefore, nearly all are assigned to the highest education category despite their censored schooling.
survey, but most individuals included in this analysis were adults at that point (they were
aged 16 to 49 in 1997). Thus, the measure of wealth comes after respondents’ schooling is
completed and is likely endogenous to their educational attainment. Because using current
wealth to predict previous schooling is problematic, I report results from models that do
not include this measure. Mother’s and father’s schooling, which serve as noisy proxies for
family wealth (Hull and Hull 1977), are controlled in all models.
Sensitivity tests conrm that the results shown below are robust to omitted measures of
household wealth.7 For respondents in the youngest birth cohort, who were aged 12 to 15
in 1993, including 1993 log per capita expenditures of their mothers’ household does not
change the results in either the exogenous or instrumented models. The results for the other
cohorts are also quite consistent. For these older cohorts, excluding individuals who coreside
with their mothers and are not enrolled in school and including 1993 log per capita expen-
ditures for their mothers’ household also produces quite similar results despite the reduced
sample size. Moreover, the overall pattern of results—the association between family size
and children’s schooling going from positive to neutral to negative across cohorts—remains
unchanged with or without a measure of family wealth.
Another methodological issue involves the effect of mortality on the sample. Devel-
oping countries generally have high rates of infant and child mortality. This means that a
full count of a woman’s live births may not represent the actual number of living siblings
a child had while growing up. Nor is it obvious which count represents the correct mea-
sure of sibship size. Infants who die soon after birth may not compete with older children
for resources that relate to schooling. But sick infants may require much of their parents’
attention, leaving older children with increased responsibilities within the household and
less time for school. In the results below, I use a count of living siblings as the measure
of family size. This measure facilitates comparisons across these Indonesian cohorts,
which have experienced steady declines in infant mortality. The full count of all live births
would implicitly embed this changing pattern of infant mortality in the construct. In these
Indonesian cohorts, including controls for the number of siblings who have died does not
change the estimated effects of family size or any of the other independent variables in
the analysis. In other settings, however, these different measures of family size may give
different results.
Selective mortality may be a potential problem in at least two other ways. First, the
oldest cohort (born 1948–1957) may be positively selected because their mothers, who are
on average in their 60s in 1997, survived long enough to appear in the IFLS sample. If the
relationship between family size and children’s schooling differs by mother’s schooling,
and mothers with less schooling are systematically censored from the sample, then the
results may be biased. Fortunately, analyses conrm that there are no meaningful interac-
tions between family size effects and women’s schooling for the oldest cohort, which is the
cohort most likely to suffer from selective survival of mothers. Second, the IFLS does not
collect information on the schooling of children who died more than 12 months before the
survey, or detailed pregnancy histories from women over age 49. This means that children
identied as the “oldest” may not be, strictly speaking, rstborn if the women had any chil-
dren born earlier who had died. Thus, measures of being the oldest boy or girl in a family
represent being the oldest living child at the time of the survey.
Analyzing the relationship between family size and children’s schooling also involves
grappling with the difcult problems of unobserved heterogeneity and joint determination.
If parents have characteristics not captured by the included covariates that are correlated
with family size and children’s schooling, or if parents choose their family size with an eye
to how much schooling they would like to provide their children, then standard regressions
7. I am grateful to Duncan Thomas for letting me borrow his coding of the IFLS household expenditure data
for these sensitivity tests. An appendix describing these sensitivity tests is available from the author.
of the type presented here provide biased estimates of the effect of family size on children’s
education. Although most of the research on the effects of family size on children’s out-
comes assumes that fertility is exogenous to children’s outcomes (see DeGraff and Bilsbor-
row 2003 for a partial review of the literature in this regard), some recent research has tried
to address these concerns through the use of xed-effects designs or by using twinning as
a “natural experiment” that causes an exogenous increase in fertility.
Fixed-effects models net out a constant individual or group-level factor that is unob-
served and is usually a component of the error term (see, e.g., Guo and VanWey 1999b,
which used repeated observations on individuals, or Parish and Willis 1993, which used
differences between siblings).8 The xed-effect approach assumes that the unobserved char-
acteristics are not related to the dependent and independent variables in complex ways. In
particular, the unobserved effect is assumed to be additive, meaning that the effect of the
unobserved factor does not differ across different levels of the other regressors. To use a
concrete example, this means assuming that the effect of parents’ unobserved preferences
for child’s schooling does not differ by child’s sex or birth order. The xed-effect approach
also assumes that the effect of the unobserved factor on the dependent variable does not
change over time.
The xed-effect approach, however, does nothing to resolve concerns about more
complex relationships such as “trade-offs” or jointly determined processes. In this case,
the analyst might model both processes together (as in a simultaneous equation model with
exclusion restrictions) or rely on an experiment that provides exogenous variation on one
of the jointly determined processes without affecting the other one. Researchers in econom-
ics have utilized twinning as an instrument to measure the exogenous effects of fertility
on outcomes such as labor force participation (Angrist and Evans 1998; Jacobsen, Pearce,
and Rosenbloom 1999; Rosenzweig and Wolpin 1980a), women’s economic outcomes
(Bronars and Grogger 1994), and children’s schooling (Angrist et al. 2006; Black et al.
2005; Rosenzweig and Wolpin 1980b). The instrumental variables approach and the use of
twins in particular have their own set of shortcomings.9 But there is also a major practical
problem with using twins as instruments: twinning is a rare event and one that is difcult
to capture in sample surveys. In this paper, I use an alternative instrument that relies on
women’s report of miscarriages, which are more common events and are widely available
in survey data. I use this approach to test the robustness of the results to assumptions about
joint determination. I discuss the approach in more detail below.
RESULTS
Changing Patterns Across Cohorts
Table 1 presents sample means and proportions for the full sample and each birth cohort.
Families are relatively large in this Indonesian sample. In the rst three birth cohorts,
children have an average of 4.5 living siblings. The average number of children ever born
ranges from about 7 in the oldest cohort to about 6.4 in the 1968–1977 cohort. Family
sizes are smaller in the most recent cohort, but this is also the cohort in which some of the
8. See Downey et al. (1999), Guo and VanWey (1999a), and Phillips (1999) for a discussion of the relative
merits of Guo and VanWey’s approach.
9. Rosenzweig and Wolpin (2000) have provided an excellent discussion of the limitations of using twins
as instruments. Rosenzweig and Wolpin (1980a) pointed out that there are biological differences between twins
themselves, and between twins and nontwins. Twins have substantially lower gestational age, lower birth weight,
and more complicated deliveries. Twins might therefore be less healthy, on average, than singletons. And while
twinning might represent the addition of an extra child in the short term, it is not equivalent to adding an addi-
tional child to a woman’s completed fertility. The evidence shows that women adjust their subsequent fertility to
compensate for the unexpected extra child. The completed fertility of women who have twins is substantially less
than an average of one child higher, and this difference has been declining over time and varies between groups
(Bronars and Grogger 1994; Jacobsen, Pearce, and Rosenbloom 1999).
Table 1. Sample Means and Proportions: IFLS

Born Born Born Born All Sample
1948–1957 1958–1967 1968–1977 1978–1981 Children
Number of Living Siblings 4.6 4.6 4.4 3.8 4.3
(0.07) (0.04) (0.03) (0.04) (0.02)
Number of Siblings Who Died 1.4 1.2 1.0 0.7 1.0
(0.05) (0.03) (0.02) (0.03) (0.01)
Average Number of Children
Ever Born (row 1 + row 2 + 1) 7.0 6.8 6.4 5.5 6.3
Family Size (living children)
1 0.03 0.03 0.02 0.02 0.02
2 0.06 0.06 0.05 0.10 0.06
3 0.11 0.11 0.13 0.18 0.13
4 0.14 0.14 0.18 0.20 0.17
5 0.14 0.16 0.18 0.17 0.17
6 0.17 0.17 0.16 0.13 0.16
7 0.11 0.13 0.12 0.09 0.11
8 0.11 0.10 0.09 0.06 0.09
9 0.07 0.07 0.05 0.04 0.06
10 or more 0.05 0.04 0.03 0.02 0.03
Child’s Education (years) 6.5 7.1 8.5 8.3 7.8
(0.13) (0.09) (0.07) (0.06) (0.04)
Child’s Educational Level
None 0.14 0.11 0.05 0.02 0.07
Elementary school 0.52 0.49 0.41 0.33 0.43
Junior secondary school 0.11 0.11 0.17 0.28 0.17
Senior secondary school 0.15 0.19 0.27 0.34 0.25
Postsecondary 0.07 0.09 0.10 0.02 0.08
Gender Gap in Education (M – F) 1.8 1.1 0.8 0.2 0.9
Mother’s Education (years) 1.7 2.3 3.3 4.0 3.0
(0.08) (0.06) (0.05) (0.08) (0.03)
Father’s Education (years) 3.6 4.0 4.9 5.5 4.6
(0.11) (0.07) (0.06) (0.09) (0.04)
Mother’s Age 67.1 60.6 52.5 43.8 54.7
(0.22) (0.15) (0.12) (0.17) (0.11)
Child’s Age 43.4 33.9 24.5 17.5 28.0
(0.08) (0.05) (0.04) (0.02) (0.09)
Rural Residence at Age 12 0.73 0.73 0.72 0.70 0.72
Oldest Boy in Family 0.29 0.25 0.20 0.23 0.23
Oldest Girl in Family 0.31 0.26 0.22 0.24 0.25
Child Under Age 6 in Sibship 0.00 0.00 0.05 0.21 0.06
Number of Observations 1,612 3,632 5,329 2,770 13,343
Notes: Numbers in parentheses are standard errors. Descriptive data are weighted using sample probability weights. Because
schooling and family size are not completed for those in the 1978–1981 cohort, these means are not directly comparable to the
other three cohorts.
Table 2. Distribution of Children’s Education (years of school completed) by Cohort: IFLS

Percentiles
___________________________________________
10 25 50 75 90 Mean N
Urban 1948–1957 3 6 9 12 15 9.3 559
Urban 1958–1967 3 6 12 12 17 10.0 1,203
Urban 1968–1977 6 9 12 12 15 10.9 1,911
Rural 1948–1957 0 2 6 6 12 5.4 1,053

Rural 1958–1967 0 3 6 9 12 6.0 2,429
Rural 1968–1977 3 6 6 12 12 7.6 3,418
Notes: N = 10,573. The table includes only cohorts with completed education. Descriptive data are weighted using sample
probability weights.
mothers are still in their peak childbearing years. In general, one- and two-child families are
uncommon in Indonesia. In the rst three birth cohorts, families with three to eight children
are relatively common. Despite a clear trend toward smaller families over time, the range
across cohorts of the number of children born to Indonesian parents is quite wide and fairly
high for the years covered by these data.
Average levels of schooling have increased steadily across birth cohorts.10 The pro-
portion of those with no schooling decreased steadily, while the proportion of those with
senior and postsecondary schooling increased in the cohorts with completed schooling
(1948–1977). At the same time, differences in educational attainment by sex changed from
a 1.8-year advantage for males in the oldest cohort to parity in the most recent cohort.11
Table 2 shows the distribution of children’s schooling by cohort and place of residence
for the rst three birth cohorts. For those living in urban areas at age 12, average levels of
schooling increased by about 1.6 years across the three birth cohorts (from about 9.3 years
to 10.9 years). The median level of schooling increased from completing junior secondary
to completing senior secondary, and the top of the distribution also increased.12 Most of the
educational gains in the urban cohorts occurred at the bottom of the education distribution.
The bottom decile went from having only some primary schooling to completing that level,
and the 25th percentile went from completing primary to completing junior secondary.
In contrast, those who grew up in rural areas have substantially lower levels of average
attainment and a more disadvantaged distribution of schooling across all cohorts. Average
attainments are three to four years lower here than in the urban cohorts. The median level
of schooling remained stable at about six years across all rural cohorts, although the inter-
quartile range shifted from straddling the primary grades to a range that includes junior and
senior secondary grades. For the bottom rural decile, the oor increased from no schooling
to an average of three years. Those in the top decile in rural areas achieved relatively high
attainments across all cohorts. Even in the top decile, however, few individuals who grew
up in rural areas progressed beyond senior secondary.
10. Means for the most recent cohort (1978–1981) are not directly comparable to the other three cohorts
because schooling and family size are not completed for this younger group.
11. Because schooling is censored for this cohort, it is impossible to know whether the gender gap will remain
closed once this cohort completes its schooling. If boys have higher continuation probabilities into the higher levels
of schooling, which the data suggest is still true for this recent IFLS cohort, then a small gender gap favoring boys
may emerge by the time the cohort completes its schooling.
12. About 7% of those in the 1968–1977 cohort are still enrolled in school, and it is likely that average at-
tainment in the top decile will be higher than shown once their schooling is completed.
Figure 1. Educational Attainment by Family Size in Indonesia, Adults Ages 20–49: IFLS (N =
10,573)
12
11
10
9
8
Years of Schooling
7
6
5
4
3
2
1
0
0/1 2 3 4 5 6 7 8 9 10 11 12 13
Number of Siblings
Figure 1 shows completed schooling by number of siblings for those born between
1948 and 1977 (ages 20 to 49 in 1997). Respondents with zero or one sibling, a rare sib-
ship size in this sample, have about one year less schooling than those with two to seven
siblings. Those with eight or more siblings complete more schooling on average. Remark-
ably, children’s education in the cross section displays little relationship with family size
for those with two to seven siblings, a substantial range by most demographic standards.
Even more surprisingly, the cross-sectional data show a positive relationship between very
large families and children’s schooling.
This cross-sectional view, however, hides substantial variation by cohort and place
of residence. Figure 2 shows the same sample as Figure 1, stratied by birth cohort and
urban-rural residence. The top panel provides a detailed specication of number of sib-
lings. The bottom panel collapses adjacent categories to provide a more tractable group-
ing for analysis. Comparing the two graphs shows that the more parsimonious grouping
stays true to the trends apparent in the more detailed representation. Figure 2 also high-
lights the dramatic increase in schooling in Indonesia over time and the large differences
in schooling by urban-rural residence.13
In the rst two urban birth cohorts (1948–1957 and 1958–1967) those with four or
fewer siblings have lower educational attainment than those with ve or more siblings. In
the 1948–1957 cohort, for example, those with ve or six siblings have an average of 1.5
13. None of the means for the 1968–1977 rural cohort are signicantly different from each other. In all other
cohorts, the mean for the 7+ category is signicantly different from the mean for the 0–2 and 3–4 categories. The
means for the 0–2 and 3–4 categories are not signicantly different.
Figure 2. Educational Attainment by Family Size and Birth Cohort, Adults Ages 20–49: IFLS (N =
10,573)
12
11
10
9
Years of Education
8
7
6
5
4
3
2
1
0
1948–1957 1958–1967 1968–1977 1948–1957 1958–1967 1968–1977
Urban Rural
Number of Siblings
0/1 2 3 4 5 6 7 8 9+
12
11
10
9
Years of Education
8
7
6
5
4
3
2
1
0
1948–1957 1958–1967 1968–1977 1948–1957 1958–1967 1968–1977
Urban Rural
Number of Siblings
0–2 3–4 5–6 7+
more years of schooling than those with fewer siblings, while those with seven or more
siblings have about 2.5 more years of schooling, on average. This relationship is quite dif-
ferent in the 1968–1977 urban birth cohort. This cohort shows large gains in average levels
of schooling overall and has a negative relationship between family size and children’s
schooling. Each larger category of siblings is associated with lower levels of average
schooling, with a difference of about 1.2 years in average attainment between the smallest
and largest sibship categories.
In the rst two rural cohorts, there is a positive relationship between family size and
completed schooling. In the 1948–1957 rural cohort, those from the largest sibships have
the highest levels of schooling, and this association is monotonically positive for those in
the 1958–1967 rural cohort. In contrast, the most recent rural cohort exhibits no relation-
ship between family size and educational attainment, a quite different pattern than has
emerged for its urban counterpart. The patterns in Figure 2 suggest that the overall asso-
ciation between family size and children’s schooling has been changing across Indonesian
cohorts and differs by geographical area. The offsetting patterns in Figure 2 also explain
the apparent lack of a pattern when family size and schooling are examined in the pooled
cross section (as in Figure 1).
Table 3 shows results from multivariate models that examine the relationship between
family size and completed education, controlling for other family characteristics.14 For each
cohort and place of residence, I show both linear and categorical estimates of the effects
of family size. All models control for mother’s and father’s education, child’s sex, whether
the child was the oldest child in the family, child’s age, mother’s age, and the regional
province in which the household was located. Parents’ education, child’s sex, and birth
order are mediators that have been shown relevant in other studies of these relationships.
The remaining variables control for sample composition.15 Models 3.1 and 3.2, shown in the
rst two columns, give estimates for the oldest urban birth cohort. For this cohort, the as-
sociation between family size and children’s schooling is positive. Holding all else constant
at the sample mean, an additional sibling increases the probability of being in the highest
education category by about .01. This positive association disappears in the 1958–1967
urban cohort (Models 3.3 and 3.4). In this urban cohort, there is no meaningful relationship
between family size and children’s schooling net of family and individual characteristics.
For the 1968–1977 urban cohort, however, a negative association emerges (Models 3.5 and
3.6). For this birth cohort, an additional sibling decreases the probability of being in the
highest education category by about .013. The results are similar for both the linear and
categorical specications of family size. Over a span of three birth cohorts, the association
between family size and children’s schooling goes from positive to neutral to negative for
those who lived in urban areas at age 12.
In these urban birth cohorts, daughters are disadvantaged in the oldest cohort but not
in the more recent ones—a sign of the closing gender gap in schooling at least in urban
areas. At the same time, being the oldest girl in a family is a disadvantage in more recent
urban cohorts but not in the oldest cohort. This pattern might reect emerging inequality
within the family or capture a transitory period effect during a time of rapid school expan-
sion. Girls who were born at the very beginning of this cohort may not have enjoyed the
full benet of the massive school expansion of this period. But the difference between the
estimate for being the oldest boy (which is not signicant) and being the oldest girl suggests
14. The results shown in Tables 3 and 4 assume that fertility and children’s schooling are not jointly
determined.
15. The remainder of the analysis focuses on the additive effects of family size on children’s schooling both
to draw out the major contours of these relationships across cohorts and to provide tractable models for the instru-
mental variables analysis. This simplication ignores two interactions that are descriptively interesting but do not
change the substantive patterns shown here. First, in all three urban cohorts, the effect of family size is diminished
or made more negative when fathers are highly educated. This interaction is not signicant in the rst two rural
cohorts and is marginally signicant in the 1968–1977 rural cohort. Second, in some cohorts, fathers with more
schooling provide an extra boost to their daughters’ education. Other interactions, such as family size by the sex
of the index child or family size by mother’s schooling, show no meaningful patterns across cohorts.
Table 3. Ordered Probit Models Predicting Completed Education, by Cohort and Rural Residence: IFLS
706
Urban Urban Urban Rural Rural Rural

1948–1957
_________________ 1958–1967
_________________ 1968–1977
_________________ 1948–1957
_________________ 1958–1967
_________________ 1968–1977
_________________
Model Model Model Model Model Model Model Model Model Model Model Model
3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12
Number of Siblings 0.064* 0.010 –0.052* 0.025 0.005 –0.012
(linear, 0–14) (0.025) (0.020) (0.017) (0.020) (0.015) (0.013)
Number of Siblings
(categories)
(ref. = 7 or more)
0–2 –0.363† 0.016 0.323* –0.150 –0.037 0.105
(0.188) (0.137) (0.121) (0.136) (0.103) (0.089)
3–4 –0.441* –0.184 0.198† –0.223† –0.024 0.010
(0.193) (0.116) (0.113) (0.127) (0.096) (0.082)
5–6 –0.408* –0.068 0.188† –0.184 –0.097 –0.031
(0.156) (0.119) (0.110) (0.136) (0.095) (0.085)
Mother’s Education 0.071* 0.073* 0.106* 0.107* 0.129* 0.129* 0.136* 0.139* 0.116* 0.117* 0.116* 0.116*
(0.024) (0.023) (0.012) (0.012) (0.012) (0.012) (0.023) (0.024) (0.014) (0.014) (0.010) (0.011)
Father’s Education 0.077* 0.080* 0.109* 0.109* 0.090* 0.091* 0.126* 0.124* 0.100* 0.100* 0.106* 0.107*
(0.018) (0.018) (0.012) (0.012) (0.010) (0.010) (0.017) (0.017) (0.010) (0.010) (0.009) (0.009)
Female –0.478* –0.489* –0.113 –0.125 –0.083 –0.087 –0.533* –0.538* –0.385* –0.383* –0.168* –0.169*
(0.160) (0.161) (0.092) (0.092) (0.071) (0.071) (0.126) (0.125) (0.071) (0.072) (0.053) (0.053)
Oldest Boy 0.041 0.012 –0.004 –0.011 –0.113 –0.098 0.128 0.112 –0.139* –0.145* –0.029 –0.042
(0.148) (0.150) (0.098) (0.099) (0.077) (0.076) (0.101) (0.101) (0.066) (0.067) (0.057) (0.057)
Oldest Girl 0.058 0.022 –0.261* –0.261* –0.140† –0.120 0.042 0.034 –0.079 –0.087 –0.178* –0.192*
(0.126) (0.131) (0.094) (0.091) (0.078) (0.078) (0.105) (0.105) (0.066) (0.066) (0.051) (0.051)
Log-Likelihood –701– –699– –1,415– –1,412– –2,145– –2,146– –1,132– –1,131– –2,825– –2,823– –3,970– –3,968–
N 559 1,203 1,911 1,053 2,429 3,418
Notes: Total N = 10,573. Robust standard errors are shown in parentheses. Cut-point parameters are not shown. All models control for child’s age, mother’s age, and province of residence (coef-
ficients not shown).
†
p < .10; *p < .05
Demography, Volume 45-Number 3, August 2008
an alternative possibility. Parents’ attitudes toward schooling girls may have evolved over
this period to the benet of younger girls.16
Models 3.7 to 3.12 show the results for the three rural birth cohorts. In these rural
cohorts, there is no signicant association between family size and children’s schooling
once other family characteristics are controlled. The categorical specication for the oldest
rural birth cohort (Model 3.8) suggests a marginally signicant negative association of hav-
ing three or four siblings relative to having seven or more. The coefcients, however, are
not jointly signicant, and the linear term is also not signicant. This marginally positive
association between larger families and children’s schooling in the oldest rural cohort is
consistent with the relationship found in the oldest urban cohort. Indeed, the overall pattern
of results for the rural cohorts, although smaller in magnitude and not signicant, is similar
to that found in the urban cohorts.17 I discuss possible explanations for these urban-rural
differences in more detail below.
Daughters complete less schooling than sons across all rural cohorts, although this
disadvantage diminishes over time. The marginal effect of being female on the probability
of being in the highest education category shrinks from –.027 to –.01. Nonetheless, women
have not made the same educational gains relative to men in rural areas as they have in
urban settings, another example of differential associations by levels of socioeconomic
development (in this case by urban-rural status within the same country). The association
between being the oldest child and completed schooling shows two interesting patterns. The
coefcient for being the oldest boy is negative in the middle rural cohort, and the coefcient
for being the oldest girl is negative in the younger rural cohort. These patterns suggest that
as schooling became more widely available in Indonesia, boys rst beneted more than
girls, and younger children beneted more than older children. In Indonesia, however, these
patterns appear to be related to the availability of schools rather than strong preferences by
birth order. This pattern does not exist for older boys in urban areas and dissipates for older
girls by the 1968–1977 urban cohort.
Overall, these ndings reproduce the trends shown in Figure 2 in a multivariate frame-
work, particularly for the urban birth cohorts. Historically in Indonesia, children from large
families living in urban areas obtained more schooling than those from smaller families,
holding other family characteristics constant. In recent urban cohorts, this positive asso-
ciation between large families and children’s schooling has disappeared, and a negative
association has emerged in the most recent cohort. In rural areas, there is no signicant
association between family size and educational attainment in any cohort once other fac-
tors are controlled.
The results in Table 4 extend the analysis to the most recent cohort of children avail-
able in the 1997 IFLS, those ages 16 to 19 in 1997. Because many members of this cohort
have not completed their schooling, these analyses rely on continuation beyond primary
school rather than completed education. About three-fourths of IFLS children ages 6 to 19
were enrolled in school in 1997. In both rural and urban areas, at least 95% of children ages
8 to 12 were enrolled in school. After age 12, enrollment by age drops faster in rural areas,
with about three-fourths of urban children still enrolled at age 16 compared with about half
of rural children age 16.18 While Indonesia has achieved nearly universal primary school
16. Overall, however, this association should be interpreted cautiously. Children in the most recent cohort
are more likely to have siblings who survive and who are captured in the survey. Therefore, children identied as
the “oldest” in these cohorts are less likely to be positively selected than those identied as the oldest child in the
oldest urban cohort (1948–1957) when both mothers and children were older and mortality rates were higher.
17. In a joint model, the two-way interactions of family size and cohort and rural status and cohort are sig-
nicant. The two-way interaction of siblings and rural status is not signicant. The three-way interaction of family
size, cohort, and rural status is also not signicant (models not shown).
18. Only 4% of children ages 7 to 19 worked while enrolled in school, with the proportions slightly higher
in rural than in urban areas (4.7% versus 2.3%).
Table 4. Binary Probit Models Predicting School Enrollment and Continuation Beyond Primary
School for the 1978–1981 Cohort (ages 16–19 in 1997): IFLS
Completed Junior Secondary School
___________________________________ Entered Senior Secondary School
___________________________________
Urban
_________________ Rural
_________________ Urban
_________________ Rural
_________________
Model Model Model Model Model Model Model Model
4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8
Number of Siblings –0.087* –0.017 –0.089* –0.026
(linear) (0.032) (0.023) (0.030) (0.026)
Number of Siblings
(ref. = 5 or more)
0–2 0.457* 0.195† 0.315* 0.212†
(0.166) (0.114) (0.143) (0.122)
3–4 0.196 0.132 0.198 0.132
(0.127) (0.093) (0.122) (0.098)
Mother’s Education 0.135* 0.135* 0.120* 0.121* 0.085* 0.086* 0.115* 0.116*
(0.020) (0.020) (0.015) (0.015) (0.017) (0.017) (0.015) (0.015)
Father’s Education 0.081* 0.081* 0.123* 0.122* 0.087* 0.086* 0.117* 0.116*
(0.016) (0.016) (0.012) (0.012) (0.015) (0.015) (0.013) (0.013)
Female 0.344* 0.332* –0.064 –0.062 0.213† 0.204† –0.002 –0.001
(0.131) (0.131) (0.092) (0.092) (0.119) (0.119) (0.101) (0.101)
Oldest Boy in Family 0.081 0.063 0.126 0.100 0.215 0.239† –0.022 –0.043
(0.155) (0.160) (0.105) (0.106) (0.135) (0.135) (0.113) (0.112)
Oldest Girl in Family –0.066 –0.062 0.012 –0.017 0.040 0.077 –0.213† –0.235*
(0.156) (0.157) (0.103) (0.103) (0.136) (0.135) (0.115) (0.115)
Child Under Age 6 0.173 0.162 –0.131 –0.099 0.177 0.127 –0.142 –0.118
in Sibship (0.157) (0.162) (0.105) (0.102) (0.147) (0.145) (0.111) (0.109)
Constant –0.937* –1.332* –1.149* –1.377* –1.561* –1.869* –1.589* –1.825*
(0.480) (0.511) (0.311) (0.348) (0.444) (0.478) (0.340) (0.379)
Log-Likelihood –441- –441- –932- –930- –548- –550- –765- –764-
N 1,059 1,059 1,711 1,711 1,059 1,059 1,711 1,711
Notes: All models control for child’s age in single years, mother’s age, and province of residence (coefficients not shown).
Sibling categories in models 4.4 and 4.8 are not jointly significantly different that zero.
†
p < .10; *p < .05
enrollment, enrollment proportions drop rapidly at ages associated with secondary school
and differ considerably between urban and rural areas.
Table 4 shows results from binary probit models that examine the likelihood of com-
pleting junior secondary schooling (completing grade 9) and entering senior secondary
(completing at least grade 10) by urban-rural residence. The models also include a measure
for having a child under age 6 in the sibship. This measure is intended to capture whether
having a very young sibling interferes with school investment for older siblings. Models
are restricted to children ages 16 to 19 at the time of the survey so that only those who are
old enough to have made these transitions are considered.19 Because families are smaller in
19. These samples and transitions are not conditioned on completing at least grade 8 or 9. Results do not
change if the sample is restricted to only those who made the previous transitions.
this cohort, the categorical version of family size is capped at ve or more children (rather
than at seven or more, as in Table 3) for this birth cohort.
In the 1978–1981 urban birth cohort, larger families are associated with a lower likeli-
hood of completing junior secondary schooling (Model 4.1). With covariates held at the
sample means, an additional living sibling is associated with a .02 decrease in the probabil-
ity of entering junior secondary school. Model 4.2 replicates this nding with a categorical
version of number of siblings. Having two or fewer siblings (versus ve or more) reduces
the likelihood of transitioning from primary to junior secondary school. In contrast, family
size and the likelihood of completing junior secondary school are not signicantly associ-
ated for children living in rural areas at age 12 (Models 4.3 and 4.4). The results are similar
for the likelihood of entering senior secondary school.20 With other family and individual
characteristics held constant, the association of family size and entering senior secondary
school is negative and signicant in urban areas (Models 4.5 and 4.6) but not signicant in
rural areas (Models 4.7 and 4.8). With covariates held at the sample means, an additional
living sibling reduces the predicted probability of entering senior secondary school by
about .03 in urban areas. Having a young child in the household is not associated with the
likelihood of making either school transition. This result does not differ by the sex of the
index child (interaction models not shown).
Differences by sex and birth order are consistent with the patterns shown in Table 3.
For urban cohorts, differences by birth order have disappeared, and an advantage for girls
has emerged. This advantage likely reects girls’ lower rates of school failure and grade
retention rather than any emerging sex preference. Indeed, by senior secondary school, this
female advantage has become only marginally signicant in urban areas. For rural areas,
disadvantages by sex and birth order have largely disappeared, reecting growing school
opportunities for all children and particularly for girls in these areas (albeit only at these
lower levels of schooling). The one exception to this pattern is that older rural girls are still
marginally less likely to enter senior secondary school relative to their younger siblings.
Testing for the Potential Endogeneity of Fertility
The analyses above provide ample evidence for a correlation between family size and
children’s schooling but do not address the concern that these processes may be jointly deter-
mined. To assess the sensitivity of the results to assumptions about the exogeneity of fertility,
I use instrumental variables analysis (IV) as an alternative estimation strategy (Wooldridge
2002). One way to instrument completed fertility is to nd a physiological (rather than be-
havioral) trait that is exogenous and heterogeneously distributed among individuals, which
affects completed fertility but is independent of preferences for family size. As discussed
above, twinning has been the most widely utilized instrument of this type in previous work.
A few studies, however, have also used miscarriages to instrument fertility (Hotz, McElroy,
and Sanders 2005; Hotz, Mullin, and Sanders 1997). Once maternal age and behaviors such
as smoking and drinking are controlled, miscarriages represent an exogenous and random
shock to a woman’s fertility (Kline, Stein, and Susser 1989; Leridon 1977; Porter and Hook
1980). Miscarriages are involuntary, spontaneous fetal deaths that reduce the number of
conceptions that result in live births, and therefore represent lost fertility exposure time
(Bongaarts and Potter 1983; Casterline 1989; Leridon 1977). As described by Casterline
(1989:81–82), “[a]ccepted estimates of overall spontaneous loss rates of 20 percent of rec-
ognized pregnancies (Bongaarts and Potter, 1983) thus imply two and one-half months of
time lost involuntarily for every live birth, on average. In most societies this represents a
reduction in fertility of 5–10 percent from levels expected in the absence of pregnancy loss.”
Net of age and behavior, women who experience many miscarriages may fail to achieve their
20. The sibling categories in Models 4.4 and 4.8 are not jointly signicantly different from zero.
desired family size despite their preferences. Factors such as socioeconomic and nutrition
status are not associated with miscarriage rates (Kline et al. 1989).21
Using miscarriages to instrument fertility is a different approach than using twins.
Miscarriages represent an exogenous constraint on fertility, while twins represent an ex-
ogenous “extra” child. Women who experience many miscarriages may end up with lower
fertility than they had hoped for, while women with twins may end up with more children
than desired. Ideally, one would test the robustness of the results to both these approaches.
However, this is untenable in most sample surveys because twinning is a rare event. Mis-
carriages, in contrast, are a more common event and are recorded in many demographic
surveys. When appropriate, miscarriages offer an alternative approach to measuring exog-
enous shocks to fertility.
Using the number of miscarriages a woman experiences as an instrument for completed
fertility does not sufciently control for underlying parental preferences about family size,
which may be correlated with preferences about children’s schooling (see Rosenzweig
and Wolpin 1980a for a related discussion regarding twins). Those who desire very large
families may have many pregnancies and therefore experience more miscarriages because
of their increased exposure to the risk of miscarriage (Kline et al. 1989). To address this, I
adjust number of miscarriages by regressing it on total number of pregnancies, pregnancies
squared, and pregnancies cubed, using a Poisson model. I then use the residual from this
regression, which is orthogonal to number of pregnancies, along with the residual squared
and cubed as the instruments. This adjusts the count of miscarriages for underlying family
size preferences in a exible and nonlinear way.22
Two important limitations to using miscarriages to instrument fertility include women’s
knowledge of having had a miscarriage in the rst place and recall error in the reporting of
miscarriages. Women’s knowledge and recall about miscarriages may vary by their level of
schooling and urban-rural residence, and both variables are controlled in all models. Women
may also forget to report miscarriages that occurred long ago as these events lose salience
in their mental record of their reproductive history (Casterline 1989; Panis and Lillard
1994). In the current analysis, the issue of recall error seems particularly problematic for
the older cohorts. I exclude these cohorts from the IV analyses and focus only on testing
the robustness of the negative association between family size and children’s schooling in
the most recent cohorts. The endogeneity of fertility is also more likely to be relevant for
more recent cohorts, who experienced the widespread distribution of modern contraception
after the 1970s.
The analyses in the previous section use ordered probit models because this specica-
tion offers a exible and less parametric way to capture educational attainment. Measured
in years of schooling, educational attainment is inherently discrete and ordered and is well-
suited to this statistical modeling strategy. But combining IV with these types of nonlinear
models is quite complex and generally not well identied. The best-developed IV methods
use two-stage least squares (2SLS) or binary probit models. Although, in general, ordered
21. About three-fourths of women in the IFLS sample reported no miscarriages. At the high end, about 7%
reported that more than 20% of their total number of pregnancies ended in miscarriage. A possible confounding
factor is if women report abortions, which are legally prohibited in Indonesia except to save a woman’s life, as
miscarriages. Women who miscarry may also try to compensate for this lost fertility or treat the children they do
have differently than women who never miscarry. Recall error is another important problem that I discuss in more
detail later.
22. There is no evidence that the risk of miscarriage increases with pregnancy number (Kline et al. 1989;
Leridon 1976). Indeed, in this sample, the average proportion of pregnancies ending in miscarriage is similar across
different family sizes. Also, the number of living siblings is correlated but not collinear with either miscarriages
or the proportion of pregnancies that end in miscarriage (W = .13 and –.02 respectively). In this sample, women’s
education is weakly positively correlated with miscarriages (W = .14), and this variable is controlled in all models.
Although using whether a woman’s rst birth resulted in a miscarriage would be an even better IV strategy (Rosen-
zweig and Wolpin 2000), the IFLS sample is too small to support this approach, especially for older cohorts.
Table 5. Summary of Instrumental Variables Results: IFLS

Urban Urban Urban Rural Rural Rural
1968–1977, 1978–1981, 1978–1981, 1968–1977, 1978–1981, 1978–1981,
Years of Complete Enter Years of Complete Enter
School Junior Senior School Junior Senior
Completed Secondary Secondary Completed Secondary Secondary
(linear) (0/1) (0/1) (linear) (0/1) (0/1)
Number of Siblings, –0.138* –0.087* –0.089* –0.017 –0.017 –0.026
Treated as Exogenous (0.050) (0.032) (0.030) (0.040) (0.023) (0.026)
Number of Siblings, –0.260† –0.110 –0.143† –0.167 –0.070 –0.054
Instrumented (0.144) (0.085) (0.079) (0.117) (0.049) (0.055)
F (IV first stage) 18.18 –– –– 24.78 –– ––
Partial R 2
(IV first stage) 0.146 –– –– 0.130 –– ––

N 1,911 1,059 1,059 3,418 1,711 1,711
Notes: The exogenous regressions for years of school completed are estimated using OLS, and those for school attendance
are estimated using probit models. The exogenous OLS results are available from the author. The exogenous probit models
are shown in Table 4. The IV coefficients are estimated using 2SLS for years of school completed and maximum likelihood IV
probit for school attendance. The instrument is described in the text. The full IV results are available from the author.
†
p < .10; *p < .05
probit and ordinary least squares (OLS) models do not necessarily produce the same re-
sults, for the case of educational attainment in this IFLS sample, the substantive results are
the same for both methods. Therefore, I estimate the IV models using 2SLS and IV probit
models. Instruments that are too weakly correlated with the endogenous regressor pose
another methodological problem for IV (Bound, Jaeger, and Baker 1995; Hall, Rudebusch,
and Wilcox 1996; Shea 1997; Staiger and Stock 1997), but standard metrics show that the
instruments used here are sufciently strong. The IV models shown in Table 5 all have
rst-stage F statistics larger than 12. The 2SLS models also have high values for the rst-
stage partial R2. 23
Table 5 summarizes the IV results. The rst row shows estimates for number of siblings
treated exogenously from OLS models corresponding to Table 3 (models not shown) and
the probit models shown in Table 4. All models control for child’s and mother’s age and
therefore mother’s age at birth. The models do not control for smoking or drinking because
these behaviors are rare among women in Indonesia. Less than 5% of IFLS women reported
having ever smoked, and controlling for this behavior does not change the results shown.
Alcohol consumption is even more uncommon in this predominantly Muslim country, and
the IFLS does not include this measure.
Overall, the IV models show the same patterns as the models that treat fertility as
exogenous, but the coefcients are not estimated as precisely. For the 1968–1977 urban
birth cohort, the instrumented effect of number of siblings is negative and marginally sig-
nicant, suggesting that children with more siblings obtain less schooling in more recent
urban cohorts. The effect of family size is also negative and marginally signicant for
the model predicting entry into senior secondary school. In the most recent urban birth
cohort, children with more siblings are less likely to make this transition. For those co-
horts in which the concern of the endogeneity of fertility and a spurious negative effect of
family size on children’s schooling is most likely, the IV results support the results of the
23. The exogenous OLS results corresponding to the ordered probit analyses shown in Table 3 and the full set
of IV results are omitted due to space limitations. These are available from the author as an appendix.
analyses that treat fertility as exogenously determined. The IV results lack power for the
two recent rural cohorts. Here, none of the models yield a signicant relationship between
family size and children’s schooling, although the patterns are suggestive and similar to
those in the earlier models.
DISCUSSION AND SUMMARY
In Indonesia, the relationship between family size and children’s schooling is not uniformly
positive or negative. Rather, there are important differences by cohort and urban-rural
residence. For rural cohorts, there is little evidence of a statistically signicant association
between family size and children’s schooling, a nding that is consistent in all birth cohorts.
In contrast, in urban areas, this association evolved from positive to neutral to negative
over a span of 30 years. For the 1948–1957 urban cohort, larger families were associated
with more schooling. This positive relationship diminished over time, and a negative rela-
tionship developed in the 1968–1977 cohort and remains for the most recent birth cohort
(1978–1981). Although the estimated effects are smaller and not statistically signicant,
this evolving pattern is also apparent across the rural cohorts, suggesting that as rural areas
continue to develop these associations may change.
In both rural and urban areas, differences in schooling by sex and birth order also
changed over time. In the oldest urban cohort and in all rural cohorts, girls obtained less
schooling than boys. This disadvantage, however, has disappeared across cohorts in urban
areas and diminished over time in rural ones. Similarly, disadvantages associated with be-
ing the oldest child also diminished and disappeared across cohorts. These results highlight
the Indonesian government’s success in expanding schooling opportunities, especially for
girls. The speed with which sex and birth order differences in schooling have narrowed
reects Indonesia’s relatively egalitarian family organization, at least relative to other East
Asian countries.
Standard analyses of the relationship between family size and children’s schooling
make strong assumptions about the relationship between family size and parents’ prefer-
ences for children’s schooling. Although IV analysis requires its own set of strong assump-
tions, the results presented here suggest that, at least in the most recent urban cohorts, the
negative relationship between family size and children’s schooling is not sensitive to as-
sumptions about the exogeneity of fertility. Ideally, one would conrm this result by using
other instruments, such as twins, but household surveys like the IFLS do not have large
enough samples to support this type of analysis, especially for a cohort design like the one
used here. Recent studies using twin data from developed countries found no relationship
between family size and children’s schooling once these variables were modeled jointly.
Rosenzweig and Wolpin (1980a), in contrast, found a negative relationship using data from
India, and Qian (2005) found mixed results using data from China. Few other studies have
explored this question in a developing setting.
This pattern of associations changing from positive to neutral to negative in urban
areas suggests that changes in socioeconomic and demographic conditions brought by de-
velopment can alter the ways in which families benet or impinge on children. Differences
between more socioeconomically developed urban areas and less developed rural ones
provide additional support for this explanation. The contrast between urban and rural areas,
for example, serves as another measure of differences in transportation and communication
infrastructure, school opportunities, and labor market opportunities.
Although testing specic mechanisms explaining this changing relationship is beyond
the scope of the current study (and the data are limited in this regard), rapid development
in Indonesia has changed numerous aspects of the economy, family organization, and edu-
cational opportunities, offering some candidate explanations for the patterns shown above.
For the oldest urban cohort, it seems likely that widely shared preferences for larger families
combined with better family resources, such as education and occupation and much better
accessibility of schools, offered those in urban areas the ability both to have more children
and to provide those children with some schooling. Moreover, in this older cohort, average
attainments were still relatively low such that the cost of schooling itself was low. It also
seems likely that schooling offered better payoffs for urban residents than for the largely
agrarian-based residents of rural areas.
Between 1970 and 1980, primary schooling became nearly universal in urban areas
and grew by nearly 30 percentage points in rural areas (Oey-Gardiner 1997). Given this
infusion of schooling at the primary level, it seems likely that families’ educational aspira-
tions grew. But the expansion in educational infrastructure was substantially slower at the
secondary level. The per-student cost of secondary schooling remains about 3 times that
of primary schooling, while the per-student cost of tertiary schooling is about 13 times as
high (UNESCO 1999). Parents, rather than other family members or broader kin, pay the
family’s portion of these costs, which can be sizeable especially for poorer families. Even
before the economic crisis of 1997–1998 (an experience that affected subsequent waves of
the IFLS but not the samples used here), Indonesia could not nance the expansion in post-
primary schooling that the nation hoped to achieve (World Bank 1998). As the educational
distribution continues to shift upward, parents with more children may be unable to afford
the cost of postprimary schooling, at least for some of their children. At the same time,
changing norms about schooling girls increases the number of children families may want
to educate relative to older cohorts.
Differences between urban and rural areas might be explained by several factors. Be-
fore the 1970s, parents in rural areas had limited options for schooling children. As Lloyd
(1994:9) argued, some threshold level of development is necessary before family size can
even inuence children’s schooling. In an area with no schools, having many or few sib-
lings is irrelevant for children’s schooling. Also, the local agrarian economy meant that the
returns to schooling were likely low relative to those living in urban areas. In more recent
years, primary schooling has become widely available, free, and compulsory. Although
most children now complete primary school, the education distribution in rural areas has
yet to shift up to the more expensive secondary level. It remains an open question whether
higher schooling costs, changing labor market conditions, change or stability in family or-
ganization, and growing migration rates will produce similar or different patterns in rural
areas in the future.
These results and candidate explanations are not meant to argue that socioeconomic
development inevitably produces family and school patterns that are negatively correlated,
whether causally or because those who value education highly will have fewer children.
Indeed, one of the most important lessons of this literature has been that patterns differ
greatly by context. These results instead speak to the diversity of possible relationships and
to the interplay of macro socioeconomic factors, government policies, norms and prefer-
ences, and familial and local mechanisms in producing patterns of stratication by place
and family structure.
REFERENCES
Angrist, J.D. and W.N. Evans. 1998. “Children and Their Parents’ Labor Supply: Evidence From
Exogenous Variation in Family Size.” American Economic Review 88:450–77.
Angrist, J.D., V. Lavy, and A. Schlosser. 2006. “Multiple Experiments for the Causal Link Between
the Quantity and Quality of Children.” Working Paper 06-26. Department of Economics, Mas-
sachusetts Institute of Technology. Available online at http://ssrn.com/abstract=931948.
Anh, T.S., J. Knodel, D. Lam, and J. Friedman. 1998. “Family Size and Children’s Education in
Vietnam.” Demography 35:57–70.
Axinn, W.G. 1993. “The Effects of Children’s Schooling on Fertility Limitation.” Population Studies
47:481–93.
Becker, G.S. 1991. A Treatise on the Family, 2nd ed. Cambridge, MA: Harvard University Press.
Becker, G.S. and N. Tomes. 1976. “Child Endowments and the Quantity and Quality of Children.”
Journal of Political Economy 84(4, Part II):S143–S162.
Black, S.E., P. Devereux, and K. Salvanes. 2005. “The More the Merrier? The Effect of Family Size
and Birth Order on Children’s Education.” Quarterly Journal of Economics 120:669–700.
Blake, J. 1989. Family Size and Achievement. Berkeley, CA: University of California Press.
Bongaarts, J. and R.G. Potter. 1983. Fertility, Biology, and Behavior: An Analysis of the Proximate
Determinants. New York: Academy Press.
Bound, J., D.A. Jaeger, and R. Baker. 1995. “Problems With Instrumental Variables Estimation When
the Correlation Between the Instruments and the Endogenous Explanatory Variables Is Weak.”
Journal of the American Statistical Association 90:443–50.
Bronars, S.G. and J. Grogger. 1994. “The Economic Consequences of Unwed Motherhood: Using
Twin Births as a Natural Experiment.” American Economic Review 84:1141–56.
Buchmann, C. and E. Hannum. 2001. “Education and Stratication in Developing Countries: A Re-
view of Theories and Research.” Annual Review of Sociology 27:77–102.
Caldwell, J.C., P.H. Reddy, and P. Caldwell. 1985. “Educational Transition in Rural South India.”
Population and Development Review 11:29–51.
Casterline, J.B. 1989. “Collecting Data on Pregnancy Loss: A Review of Evidence From the World
Fertility Survey.” Studies in Family Planning 20:81–95.
Chernichovsky, D. 1985. “Socioeconomic and Demographic Aspects of School Enrollment and At-
tendance in Rural Botswana.” Economic Development and Cultural Change 33:319–32.
Chernichovsky, D. and O.A. Meesook. 1985. “School Enrollment in Indonesia.” World Bank Staff
Working Paper Number 746. World Bank, Washington, DC.
Cobbe, J. and Boediono. 1993. “Education, Demographics, the Labor Market, and Development:
Indonesia in the Process of Transition?” Journal of Asian and African Studies 28:1–29.
DeGraff, D.S. and R.E. Bilsborrow. 2003. “Children’s School Enrollment and Time at Work in the
Philippines.” Journal of Developing Areas 37:127–58.
de Graff, P.M. and J.J. Huinink. 1992. “Trends in Measured and Unmeasured Effects of Family Back-
ground on Educational Attainment and Occupational Status in the Federal Republic of Germany.”
Social Science Research 21:84–112.
Desai, S. 1995. “When Are Children From Large Families Disadvantaged? Evidence From Cross-
National Studies.” Population Studies 49:195–210.
Downey, D.B., B. Powell, L.C. Steelman, and S. Pribesh. 1999. “Much Ado About Siblings: Change
Models, Sibship Size, and Intellectual Development: Comment on Guo and VanWey.” American
Sociological Review 64:193–98.
Duo, E. 2001. “Schooling and the Labor Market Consequences of School Construction in Indonesia:
Evidence From an Unusual Policy Experiment.” American Economic Review 90:795–813.
Featherman, D.L. and R.M. Hauser. 1978. Opportunity and Change. New York: Academic.
Filmer, D. and L. Pritchett. 1999. “The Effect of Household Wealth on Educational Attainment: Evi-
dence From 35 Countries.” Population and Development Review 25:85–120.
Firman, T. 1997. “Patterns and Trends of Urbanization: A Reection of Regional Disparity.”
Pp. 101–17 in Indonesia Assessment: Population and Human Resources, edited by G.W. Jones and
T.H. Hull. Singapore: Institute of Southeast Asian Studies.
Frankenberg, E. and D. Thomas. 2000. “The Indonesian Family Life Survey (IFLS): Study Design
and Results From Waves 1 and 2.” RAND DRU-2238/1-NIA/NICHD.
Gertler, P.J. and J.W. Molyneaux. 1994. “How Economic Development and Family Planning Pro-
grams Combined to Reduce Indonesian Fertility.” Demography 31:33–63.
Gomes, M. 1984. “Family Size and Educational Attainment in Kenya.” Population and Development
Review 10:647–60.
Guo, G. and L.K. VanWey. 1999a. “The Effects of Closely Spaced and Widely Spaced Sibship Size
on Intellectual Development: Reply to Phillips and to Downey et al.” American Sociological
Review 64:199–206.
———.1999b. “Sibship Size and Intellectual Development: Is the Relationship Causal?” American
Sociological Review 64:169–87.
Hall, A.R., G.D. Rudebusch, and D.W. Wilcox. 1996. “Judging Instrument Relevance in Instrumental
Variables Estimation.” International Economic Review 37:283–98.
Hauser, R.M. and D.L. Featherman. 1976. “Equality of Schooling: Trends and Prospects.” Sociology
of Education 49:99–120.
Hauser, R.M. and W.H. Sewell. 1985. “Birth Order and Educational Attainment in Full Sibships.”
American Educational Research Journal 22:1–23.
Hermalin, A., J.A. Seltzer, and C. Lin. 1982. “Transitions in the Effect of Family Size on Female
Educational Attainment: The Case of Taiwan.” Comparative Education Review 26:254–70.
Hotz, V.J., S. McElroy, and S. Sanders. 2005. “Teenage Childbearing and Its Life Cycle Consequenc-
es: Exploiting a Very Natural Experiment.” Journal of Human Resources 40:683–715.
Hotz, V.J., C.H. Mullin, and S.G. Sanders. 1997. “Bounding Causal Effects Using Data From a
Contaminated Natural Experiment: Analyzing the Effects of Teenage Childbearing.” Review of
Economic Studies 64:575–603.
Hugo, G.J., T.H. Hull, V.J. Hull, and G.W. Jones. 1987. The Demographic Dimension in Indonesian
Development. Singapore: Oxford University Press.
Hull, T.H. 1987. “Fertility Decline in Indonesia: An Institutionalist Interpretation.” International
Family Planning Perspectives 12:90–95.
Hull, T.H. and V.J. Hull. 1977. “The Relation of Economic Class and Fertility: An Analysis of Some
Indonesian Data.” Population Studies 31:43–57.
Indonesia Central Bureau of Statistics (BPS). 2003. “Population Statistics.” Available online at http://
www.bps.go.id/sector/population.
Iskandar, M.B. 1997. “Health and Mortality.” Pp. 205–31 in Indonesia Assessment: Population and
Human Resources, edited by G.W. Jones and T.H. Hull. Singapore: Institute of Southeast Asian
Studies.
Jacobsen, J.P., J.W. Pearce III, and J.L. Rosenbloom. 1999. “The Effects of Childbearing on Married
Women’s Labor Supply and Earnings: Using Twin Births as a Natural Experiment.” Journal of
Human Resources 34:449–74.
Jones, G.W. 1990. “Fertility Transitions Among Malay Populations of Southeast Asia: Puzzles of
Interpretation.” Population and Development Review 16:507–37.
Jones, G.W. and T.H. Hull, eds. 1997. Indonesia Assessment: Population and Human Resources.
Singapore: Institute of Southeast Asian Studies.
King, E. 1987. “The Effect of Family Size on Family Welfare: What Do We Know?” Pp. 373–411
in Population Growth and Economic Development: Issues and Evidence, edited by D.G. Johnson
and R.D. Lee. Madison, WI: The University of Wisconsin Press.
Kline, J., Z. Stein, and M. Susser. 1989. Conception to Birth: Epidemiology of Prenatal Development.
New York: Oxford University Press.
Knodel, J., N. Havanon, and W. Sittitrai. 1990. “Family Size and the Education of Children in the
Context of Rapid Fertility Decline.” Population and Development Review 16:31–62.
Kuo, H.D. and R.M. Hauser. 1995. “Trends in Family Effects on the Education of Black and White
Brothers.” Sociology of Education 68:136–60.
Leridon, H. 1976. “Facts and Artifacts in the Study of Intra-uterine Mortality: A Reconsideration
From Pregnancy Histories.” Population Studies 30:319–35.
———. 1977. Human Fertility: The Basic Concepts. Translated by J.F. Helzner. Chicago: The Uni-
versity of Chicago Press.
Lloyd, C.B. 1994. “Investing in the Next Generation: The Implications of High Fertility at the Level
of the Family.” Pp. 181–202 in Population and Development: Old Debates and New Conclusions,
edited by R. Cassen. New Brunswick, NJ: Transaction Publishers.
Lloyd, C.B. and A.K. Blanc. 1996. “Children’s Schooling in sub-Saharan Africa: The Role of Fathers,
Mothers, and Others.” Population and Development Review 22:265–98.
Lloyd, C.B. and A.J. Gage-Brandon. 1994. “High Fertility and Children’s Schooling in Ghana: Sex Dif-
ferences in Parental Contributions and Educational Outcomes.” Population Studies 48:293–306.
Mueller, E. 1984. “Income, Aspirations, and Fertility in Rural Areas of Less Developed Countries.”
Pp. 121–50 in Rural Development and Human Fertility, edited by W.A. Schutjer and C. Shannon
Stokes. New York: Macmillan Publishing Company.
Oey-Gardiner, M. 1997. “Educational Developments, Achievements and Challenges.” Pp. 135–66 in
Indonesia Assessment: Population and Human Resources, edited by G.W. Jones and T.H. Hull.
Singapore: Institute of Southeast Asian Studies.
Panis, C.W.A. and L.A. Lillard. 1994. “Health Inputs and Child Mortality: Malaysia.” Journal of
Health Economics 13:455–89.
Parish, W.L. and R.J. Willis. 1993. “Daughters, Education, and Family Budgets Taiwan Experiences.”
Journal of Human Resources 28:863–98.
Phillips, M. 1999. “Sibship Size and Academic Achievement: What We Now Know and What We Still
Need to Know: Comment on Guo and VanWey.” American Sociological Review 64:188–92.
Pong, S. 1997. “Sibship Size and Educational Attainment in Peninsular Malaysia: Do Policies Mat-
ter?” Sociological Perspectives 40:227–42.
Porter, I.H. and E.B. Hook. 1980. Human Embryonic and Fetal Death. New York: Academic Press.
Preston, S.H. 1976. “Family Sizes of Children and Family Sizes of Women.” Demography 13:105–14.
Psacharopoulos, G. and A.M. Arriagada. 1989. “The Determinants of Early Age Human Capital
Formation: Evidence From Brazil.” Economic Development and Cultural Change 37:683–708.
Qian, N. 2005. “Quantity-Quality: The Positive Effect of Family Size on School Enrollment in
China.” Working paper. Department of Economics, Brown University.
Retherford, R.D. and W.H. Sewell. 1991. “Birth Order and Intelligence: Further Tests of the Conu-
ence Model.” American Sociological Review 56:141–58.
Rosenzweig, M.R. and K.I. Wolpin. 1980a. “Testing the Quantity-Quality Fertility Model: The Use
of Twins as a Natural Experiment.” Econometrica 48:227–40.
———. 1980b. “Life-Cycle Labor Supply and Fertility: Causal Inferences From Household Models.”
Journal of Political Economy 88:328–48.
———. 2000. “Natural ‘Natural Experiments’ in Economics.” Journal of Economic Literature
38:827–74.
Shavit, Y. and J.L. Pierce. 1991. “Sibship Size and Educational Attainment in Nuclear and Extended
Families: Arabs and Jews in Israel.” American Sociological Review 56:321–30.
Shea, J. 1997. “Instrument Relevance in Multivariate Linear Models: A Simple Measure.” Review of
Economics and Statistics 79:348–52.
Staiger, D. and J.H. Stock. 1997. “Instrumental Variables Regression With Weak Instruments.”
Econometrica 65:557–86.
Steelman, L.C., B. Powell, R. Werum, and S. Carter. 2002. “Reconsidering the Effects of Sibling
Conguration: Recent Advances and Challenges.” Annual Review of Sociology 28:243–69.
Sudha, S. 1997. “Family Size, Sex Composition and Children’s Education: Ethnic Differences Over
Development in Peninsular Malaysia.” Population Studies 51:139–51.
Thornton, A. 2001. “The Developmental Paradigm, Reading History Sideways, and Family Change.”
Demography 38:449–46.
UNESCO Institute for Statistics, Organization for Economic Co-operation and Development. 1999.
“Financing Education—Investments and Returns: Analysis of the World Education Indicators
2002.” Paris: UNESCO Publishing.
UNICEF. 2000. “The State of the World’s Children.” Available online at http://www.unicef.org
/sowc00.
UNICEF and the Government of Indonesia. 1988. “Situation Analysis of Children and Women in
Indonesia.” Jakarta: UNICEF.
Walton, J. 1985. “Industrial Development and Trade.” Pp. 43–65 in Understanding Indonesia,
edited by L. Palmier. Aldershot, Hampshire (England), and Brookeld, VT: Gower Publishing
Company.
Wolfe, B.L. and J.R. Behrman. 1984. “Who is Schooled in Developing Countries? The Roles of
Income, Parental Schooling, Sex, Residence, and Family Size.” Economics of Education Review
3:231–45.
Wooldridge, J.M. 2002. Econometric Analysis of Cross Section and Panel Data. Cambridge, MA:
MIT Press.
World Bank. 1998. “Education in Indonesia: From Crisis to Recovery.” Report No. 18651-IND.
World Bank, Washington, DC.
Zajonc, R.B. and G.B. Marcus. 1975. “Birth Order and Intellectual Development.” Psychological
Review 82:74–88.

Hull 2009

Uploaded by

Copyright:

Available Formats

Hull 2009

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hull 2009

Uploaded by

Copyright:

Available Formats

THE CHANGING RELATIONSHIP BETWEEN FAMILY

SIZE AND EDUCATIONAL ATTAINMENT OVER THE

socioeconomic development, family organization, or government policies, comparing across

DATA AND METHODS

Table 1. Sample Means and Proportions: IFLS

Table 2. Distribution of Children’s Education (years of school completed) by Cohort: IFLS

Rural 1948–1957 0 2 6 6 12 5.4 1,053

0–2 3–4 5–6 7+

Urban Urban Urban Rural Rural Rural

Table 5. Summary of Instrumental Variables Results: IFLS

(IV ﬁrst stage) 0.146 –– –– 0.130 –– ––

You might also like