How Universal is the Negative Correlation between Education and Fertility?

Abstract

A negative correlation between education and fertility has been described with great regularity in modern societies. The present investigation examines the strength of this relationship with data from the 1990, 1995 and 2000 waves of the World Values Survey covering 78 countries with a combined sample size of up to 181,728 respondents. The negative correlation is present in nearly all countries, is stronger in females than males, is greater for educational level than for length of schooling, and is not mediated by personal wealth. It is strongest at relatively low levels of economic, social and cognitive development and becomes weaker in the most advanced societies. However, it is also less than maximal in the least developed countries. The relationship is strongest in Latin America and the Middle East, where the typical correlations for cohorts with completed fertility are -.31 for females and -.24 for males, and weakest in Protestant Europe, where average correlations are -.10 for females and -.01 for males. The negative relationship persists in the younger generation of advanced societies, who are reproducing under conditions of sub-replacement fertility.

Full text

205
Volume 33, Number 2, Summer 2008
How Universal is the Negative Correlation
between Education and Fertility?
Gerhard Meisenberg∗
Ross University School of Medicine, Dominica
A negative correlation between education and fertility has been
described with great regularity in modern societies. The present
investigation examines the strength of this relationship with data from the
1990, 1995 and 2000 waves of the World Values Survey covering 78
countries with a combined sample size of up to 181,728 respondents. The
negative correlation is present in nearly all countries, is stronger in females
than males, is greater for educational level than for length of schooling, and
is not mediated by personal wealth. It is strongest at relatively low levels of
economic, social and cognitive development and becomes weaker in the
most advanced societies. However, it is also less than maximal in the least
developed countries. The relationship is strongest in Latin America and the
Middle East, where the typical correlations for cohorts with completed
fertility are -.31 for females and -.24 for males, and weakest in Protestant
Europe, where average correlations are -.10 for females and -.01 for males.
The negative relationship persists in the younger generation of advanced
societies, who are reproducing under conditions of sub-replacement
fertility.
Kew Words: Education; Income; Fertility; World Values Survey; Dysgenics.
A negative relationship of fertility with measures of education has
been reported with great regularity, both in the less developed countries
and in the most advanced societies (Graff, 1979; Cao & Lutz, 2004; Goujon
& Lutz, 2004; Retherford & Luther, 1996; Weinberger, 1987). Similar
negative correlations have also been described between fertility and
psychometric intelligence (Lynn, 1996; Lynn & van Court, 2004;
Meisenberg et al., 2006; Neiss et al., 2002; Retherford & Sewell, 1988),
whereas the relationship between income and reproduction is more
variable (Coleman, 1990; Weeden et al., 2006). The relationships between
education, income and intelligence are stronger than the relationships of
any of these variables with fertility (Herrnstein & Murray, 1994; Irwing &
Lynn, 2006). Thus it is possible that relationships between schooling and
reproduction are not directly causal but are mediated by income or
∗ Address for correspondence: Department of Biochemistry, Ross University, Medical
School, Picard Estate, Dominica. Email Gmeisenberg@rossmed.edu.dm

206 Gerhard Meisenberg
The Journal of Social, Political and Economic Studies
intelligence.
The aim of the present study is not the delineation of causal paths, but a
comparative survey of the strength of the education-fertility relationship in
today’s world. Data are from 78 countries participating in the World
Values Survey. The specific questions examined are:
(1) Is a negative relationship universal across all countries and
world regions, or are there notable exceptions? Some
exceptions to the rule of a negative relationship of childbearing
with education and related constructs, such as intelligence and
social class, have been reported. In Sweden, for example,
reproductive output was positively related to income before
1920 (Edin & Hutchinson, 1935) and unrelated to intelligence
before 1970 (Nyström et al., 1990).
(2) Is the relationship stronger in females than males? Because of
the strong emphasis on female education and career prospects
in advanced postmodern societies, it is possible that the
relationship is more negative for females than males in these
societies but not necessarily in less developed countries, and
especially not in societies such as the Muslim Middle East in
which traditional gender roles persist to the present day.
(3) Is there a systematic relationship to the stage of the fertility
transition? According to Richard Lynn (1996), the relationship
is strongest during the fertility transition. It weakens in post-
transitional societies, in which contraceptive habits have
percolated through all social classes. This hypothesis predicts
an inverted U-shaped relationship between the total fertility
rate and the strength of the education-fertility correlation.
Data Sources and Methods
The World Values Survey
The World Values Survey has been performed in 5 waves so far, with
the results of the first 4 waves publicly available. Results of the 5th wave
are scheduled for release in 2009. The present study combines the results of
the second, third and fourth waves, for which data were collected around
1990, 1995 and 2000, respectively. Data for the 1st to 3rd waves are publicly
available for online analysis at http://nds.umdl.umich.edu. Data of the 4th
wave can be purchased on CD with Inglehart et al. (2004).

How Universal is the Negative Correlation between Education and Fertility? 207
Volume 33, Number 2, Summer 2008
Table 1
The relationship of number of children with educational degree and years in
school, expressed as standardized beta and controlled for age, for
different world regions. Only post-reproductive individuals (females 41-
75 years and males 45-75 years) are included. The average of the
country-level data is shown.
Children – Educ. Degree Children – Years School Sample Size
Region Females Males Females Males N Countries N Respondents
Protestant -.100 -.006 -.083 -.028 8 4160-5778
Europe
Catholic -.145 -.091 -.093 -.078 9 3442-6483
Europe
English- -.181 -.104 -.099 -.065 5-6 2247-3701
speaking
Excommunist -.216 -.139 -.200 -.113 22 10730-15881
Latin -.313 -.223 -.211 -.145 9-10 3789-5637
America
Middle East -.308 -.257 -.220 -.210 6-7 1252-2982
South Asia -.268 -.151 -.134 -.142 4 1311-1965
East Asia -.127 -.107 -.169 -.081 3-4 1113-1771
Asian -.255 -.194 -.202 -.100 2 526-688
Communist -.214 -.115 -.214 -.106 5 1147-1867
Africa
World -.213 -.137 -.166 -.108 73-77 33179-42663
For 78 countries, information is available about number of children
and either highest educational degree or the age at which the respondent
left school. Also a measure of relative income (compared to the average in
the country) is available, but there is no measure of intelligence.
However, the sample sizes are somewhat low (see Appendix). In
addition, many country samples are not very representative for the national
population, especially with respect to education. The correlations between
the average education of the samples in the World Values Survey
(averaged from highest educational degree and years of schooling) and the
average educational attainment in the country is only .615 for the old

208 Gerhard Meisenberg
The Journal of Social, Political and Economic Studies
generation and .399 for the young generation. Much of this is due to
oversampling of higher educational levels in some of the less developed
countries.
Calculation of the Education-Fertility Relationship
The respondents were divided into four groups: old females (age 41-
75), old males (age 45-75), young females (age 19-40) and young males
(age 20-44). These age limits were used to separate those of reproductive
age from those with completed or nearly completed childbearing. Within
each age-gender category, the relationship between education and fertility
was determined in regression models in which the number of children was
predicted with age and either highest degree or the age at which the
respondent left school. Because of limitations in the online analysis system,
Tables 1 and 2 and the appendix report the strength of the relationship as
the standardized beta, which is numerically very close to the part and
partial correlations. In the text, these values are referred to simply as
correlations. Controlling for age is necessary even in the old groups
because of cohort changes in fertility and education.
Other Data Sources
Several macrosocial indicators were investigated as possible
influences on the strength of the education-fertility relationship. As far as
possible, these indicators were taken from the time period in which the
cohorts under investigation were reproducing. They include:
Total fertility rate (TFR), averaged for the years 1960-1980 for the
older cohorts and 1980-1995 for the younger cohorts. Data are
from the World Development Indicators of the World Bank,
which can be purchased at:
http://publications.worldbank.org/subscriptions.
Average educational level of the adult population was from the
Barro-Lee data set for the average length of schooling of the
adult population, freely available at:
http://www.cid.harvard.edu/ciddata/barrolee/appendix_data_ta
bles.xls. Data were averaged from 1975 and 1985 for the old
groups and from 1980 to 1995 for the young groups. Missing
data were extrapolated from the arcsine-transformed adult

How Universal is the Negative Correlation between Education and Fertility? 209
Volume 33, Number 2, Summer 2008
literacy rate, average of 1990 and 2002, from the Human
Development Report of the United Nations, 2004 edition.
Logarithm of GDP (lgGDP) was averaged from 1955 to 1980 for the
old groups and from 1975 to 1995 for the young groups. Data
sources are the historical GDPs published by Angus Maddison
(http://www.ggdc.net/maddison/) and the World Development
Indicators of the World Bank.
Political freedom was the average of political rights and civil
liberties published by Freedom House at:
http://www.freedomhouse.org/research/freeworld, averaged for
the years 1972 to 1987 for the old groups and 1980 to 1995 for
the young groups.
Democracy was measured as Vanhanen’s Democracy Index,
available from the Finnish Social Science Data Archive at
http://www.fsd.uta.fi/english/data/catalogue/FSD1289/.
The index was averaged for the years 1960-1980 for the old
groups, and 1975-1990 for the young groups.
Results
Education-fertility relationship in different world regions
The relationship between education and the number of children was
negative in the large majority of countries. 301 of the 310 correlations in
the appendix have a negative sign, one is zero, and only 8 have a positive
sign. Positive signs are found for old males in Belgium, Finland, Latvia,
Sweden, Switzerland and Uganda; for old females in Malta; and for young
males in Estonia. These positive correlations were always small.
Tables 1 and 2 summarize the strength of the relationship between the
number of children and measures of education in different world regions.
The relationships were determined separately for each country, and then
averaged for all countries in the region. The regions were defined similar
to the “cultural provinces” in Inglehart et al. (2004). Greece was lumped
with “Catholic Europe”; “Middle East” included the predominantly
Muslim countries from Morocco to Pakistan; “South Asia” refers to South
and Southeast Asia, represented by India, Bangladesh, Indonesia and the
Philippines; “East Asia” is represented by Japan, South Korea, Taiwan and
Singapore, and “Communist Asia” by China and Vietnam; “Africa” refers
only to sub-Saharan Africa, represented by Nigeria, South Africa,

210 Gerhard Meisenberg
The Journal of Social, Political and Economic Studies
Tanzania, Uganda and Zimbabwe. 77 or 78 of the countries had data for
highest educational degree, and 73 or 74 countries had data for length of
schooling. Several observations can be made in Tables 1 and 2:
(1) Correlations are usually more negative for highest educational
degree than for length of schooling.
(2) Correlations are more negative in females than males. This
corroborates the results of many previous studies (reviewed in
Lynn, 1996).
(3) All regions show the negative relationship, but there are
substantial differences in its magnitude. In both age groups it is
strongly negative in Latin America and the Middle East, and
least negative in Protestant Europe.
(4) In Europe and the English-speaking countries, correlations are
more negative in the young than the old generation; and in the
ex-Communist countries and the Middle East they are less
negative in the young than the old generation.
Generally, in those countries in which the correlations are strongly
negative for females, they are also strongly negative for males. The
country-level correlation between the male and female correlations (old
and young groups combined) is .711 for highest educational degree and .552
for years in school. There is also continuity across generations in the size of
the negative relationship. The country-level correlation between the
education-fertility correlations of the old and young groups (males and
females combined) is .618 for educational degree and .395 for years in
school.
Changes with economic and social development
Tables 1 and 2 support the conclusion of Richard Lynn (1996) that
dysgenic fertility for education and intelligence peaks during the fertility
transition and becomes smaller when contraceptive habits diffuse from the
upper to the lower social classes. If advanced societies have a less negative
relationship between education and fertility, then we expect positive
correlations of development indicators such as GDP, political freedom and
the average educational level with the education-fertility relationship.
Table 3 shows that this is the case. In most cases the strongest relationships
are with the average educational level in the country and with the total

How Universal is the Negative Correlation between Education and Fertility? 211
Volume 33, Number 2, Summer 2008
fertility rate (TFR).
Table 2
The relationship of number of children with educational degree and years in
school, expressed as standardized beta and controlled for age, for
different world regions. Only individuals of reproductive age (females
19-40 years and males 20-44 years) are included. The average of the
country-level data is shown.
Children – Educ. Degree Children – Years School Sample Size
Region females males females males N countries N respondents
Protestant -.137 -.057 -.124 -.062 8 3606-4098
Europe
Catholic -.220 -.101 -.192 -.095 9 3451-4629
Europe
English- -.213 -.143 -.154 -.110 5-6 2272-2861
speaking
Excommunist -.174 -.082 -.203 -.089 22 11269-11887
Latin America -.314 -.154 -.183 -.150 10-11 5268-6183
Middle East -.230 -.212 -.194 -.149 6-7 3129-5072
South Asia -.239 -.194 -.190 -.152 4 2716-3626
East Asia -.170 -.111 -.163 -.092 3-4 1317-2017
Asian -.239 -.144 -.159 -.039 2 526-688
Communist
Africa -.255 -.152 -.138 -.113 5 2955-4058
World -.213 -.121 -.178 -.105 74-78 39125-42551
However, the dependence of the education-fertility relationship on
economic, intellectual and social development is not expected to be linear.
We rather expect that the negative education-fertility correlation first rises
and then falls as we move from very low to very high levels of development.
Figures 1 to 3 show that this is indeed the case in the old generation, with
males and females combined. In Figure 1 we see that the relationship is
most negative in country samples with an average of 3.5-4.8 children and
slightly less negative in samples with an average of more than 4.8 children.
In the low-fertility samples, the negative relationship is only half as strong
as in the samples with an average of 3.5-4.8 children. A similar relationship

212 Gerhard Meisenberg
The Journal of Social, Political and Economic Studies
(not shown) is observed when the total fertility rate (TFR) of the country,
rather than the average number of children of the respondents in the World
Values Survey, is the independent variable.
Table 3
Correlations of development indicators with the education-fertility
correlation. A positive sign means a less negative correlation between
highest educational degree and the number of children. N = 77
countries for the old and 78 for the young generation, except for
democracy, where N = 76 and 77. Correlations greater than .230 are
significant at p<.05.
Correlation Children – Educational Degree
Old Old Young Young
Females Males Females Males
IQ .427 .382 .335 .381
Education .439 .508 .453 .508
lgGDP .355 .402 .246 .374
Pol. Freedom .333 .306 .015 .139
Democracy .438 .390 .148 .213
TFR -.486 -.479 -.274 -.484
Figure 2 shows this relationship for the logarithm of gross domestic
product (GDP), and Figure 3 for the average IQ in the country. We see that
the education-fertility relationship is not quite as negative in the poorest
countries and those with the lowest IQs as it is in those at intermediate
levels of economic and cognitive development. The smallest negative
correlations are seen in prosperous countries with high average IQ.
However, the negative relationship persists even in these countries.
We can expect that with further economic and intellectual
development, in the near future the least developed countries will have the
strongest negative correlations. This trend is indeed seen in the young
cohort, where the negative education-fertility relationship is strongest in
the poorest countries (lgGDP<3.1). For fertility and IQ, the smaller size of
the negative correlation at the lowest levels of development is still
noticeable, but is smaller than in the old group (data not shown). At the
upper end of the development scale, the negative correlation persists in the

How Universal is the Negative Correlation between Education and Fertility? 213
Volume 33, Number 2, Summer 2008
young generation of the most advanced countries, where the education-
fertility relationship appears to be stabilizing at moderately negativ e
values.
Figure 1
The strength of the negative relationship (expressed as standardized β)
between number of children and highest educational degree at different
fertility levels of the interviewed samples, males aged 44-75 and females
aged 41-75 years combined. Sample sizes range from 5 countries (>4.8
children) to 25 countries (2.1-2.49 children).
Education versus income
Raw correlations between income and fertility are often negative, although
they vary by time, place and sample characteristics (Coleman, 1990; Simon,
1974). This raises the possibility that the fertility-reducing effect of
education is mediated by wealth. In one question of the World Values
Survey the respondents rate their household income on a 10-point scale
relative to the typical income in their country. It is therefore possible to
compare the relative importance of education and income.

214 Gerhard Meisenberg
The Journal of Social, Political and Economic Studies
Figure 2
The strength of the negative relationship (expressed as standardized β)
between number of children and highest educational degree depending
on log-transformed per-capita income, males aged 44-75 and females
aged 41-75 years combined. Sample sizes range from 5 countries
(lgGDP = 2.8-2.99) to 18 countries (lgGDP = 3.4-3.59).
Figure 4 shows the results of regression models predicting the number
of children with self-reported income and education (measured as the
average of highest educational degree and years in school) in the 2000
wave of the World Values Survey. Males and females are combined, and
sex, age and country are controlled. We see that even with income
controlled, education remains an independent negative predictor of fertility
in all world regions. Independent effects of income are small and variable.
In the cohort with completed fertility, the independent effect of income
tends to be positive in the advanced and relatively “egalitarian” societies
of Europe and East Asia.

How Universal is the Negative Correlation between Education and Fertility? 215
Volume 33, Number 2, Summer 2008
Figure 3
The strength of the negative relationship between number of children and
highest educational degree (expressed as standardized β) depending on
the average IQ in the country, males aged 44-75 and females aged 41-
75 years combined. Sample sizes range from 5 (IQ<80) to 25 (IQ =
95-99)
Independent effects of income are generally more negative than the
values in Table 4 when the analysis is limited to married people,
presumably because both family income and the number of children tend to
be higher for married people than for singles. It is not clear whether high
income reduces childbearing among married people, or whether
childbearing reduces family income. In the United States at least, working
mothers earn less than other women, possibly because they are mothers
(Budig & England, 2001). In the advanced world regions but not the less
developed regions, the independent effects of education tend to be less
negative (but not positive) when only married people are included. The
reason is that in most of the more advanced societies, the likelihood of
being married is reduced by higher education (data not shown).

216 Gerhard Meisenberg
The Journal of Social, Political and Economic Studies
Table 4:
Partial correlations of the number of children with education (Ch-Edu) and
income (Ch-Inc), controlling for age, sex, country, and income or
education, respectively.
Old cohort Young cohort
Ch-Edu Ch-Inc N Ch-Edu Ch-Inc N
Protestant -.060*** .078*** 3386 -.148*** .079*** 3143
Europe
Catholic -.129*** .055*** 4405 -.219*** .078*** 4385
Europe
English- -.120*** .029 3513 -.193*** -.078*** 3709
speaking
Excommunist -.167*** .023* 12666 -.145*** -.012 12493
Latin America -.250*** -.018 4583 -.255*** -.071*** 7859
Middle East -.252*** -.021 4092 -.215*** -.078*** 6884
South Asia -.233*** .010 1850 -.277*** -.024 3333
East Asia -.209*** .146*** 1844 -.171*** -.070** 2345
Asian -.232*** .031 815 -.208*** .059 1027
Communist
Africa -.088*** -.085** 1614 -.252*** -.011 4834
Discussion
Universality of the relationship
The main question of this study can be answered unambiguously: In
today’s world, the negative relationship between education and fertility is
universal. It is present in all world regions, in males and females, and in old
and young cohorts. However, its strength varies in predictable ways. It is
stronger in females than males, and is strongest during the fertility
transition, in nations that are still at relatively early stages of economic and
cognitive development.
Figures 1 to 3 show that the negative relationship is somewhat less
pronounced at the lowest levels of development. This confirms earlier
reports that the education-fertility relationship is not necessarily negative
and is sometimes even positive at the lowest levels of development (Castro
Martín, 1995; Jejeebhoy, 1995). Such positive relationships are now rare

How Universal is the Negative Correlation between Education and Fertility? 217
Volume 33, Number 2, Summer 2008
because there are hardly any “pre-transitional” or “natural-fertility”
populations left in the world.
Although the negative education-fertility association peaks at
relatively low levels of development, it persists in the most advanced
economies at a strength that is about 40% of its maximal strength at the
height of the fertility transition (Figures 1-3). This is in agreement with
earlier evidence for educational fertility differentials at all stages of the
fertility transition (Bongaarts, 2003).
A comparison of results for the old and young cohorts in Tables 1 and 2
shows that, if anything, the relationship is becoming stronger again in the
young cohorts of the most advanced countries, who are reproducing under
conditions of sub-replacement fertility. Part of this can be attributed to the
postponement of childbearing by educated women and men (Kemkes-
Grottenthaler, 2003; Martin, 2000; Neiss et al., 2002; Yang & Morgan,
2003), rather than a generational trend in the effect of education on lifetime
fertility. However, analyses in the United States have shown that the
negative education-fertility relationship in this country is either constant or
even slightly increasing over time (Mare, 1997; Retherford & Luther,
1996). In sub-Saharan Africa, on the other hand, the greater strength of the
education-fertility relationship in the young generation most likely signals
the progression of these countries into the fertility transition, rather than
the postponement of childbearing by educated people.
Some regional differences do exist. Regression models show that the
relationship is less negative than expected in Protestant Europe but more
negative than expected in the English-speaking countries. The reason for
this is not known. On the whole, however, the universality of the
phenomenon, and its predictability from “development indicators” such as
IQ and GDP, show that its major determinants are not culture-specific.
Also the greater effect of education on female than male reproduction
is universal. When we compare regions at similar stages of economic
development, there is a trend for the male-female difference to be smaller
in world regions with traditional gender roles than in those with more
liberated women. For example, the male-female difference in the effect of
education is greater in Latin America than the Muslim Middle East, and
greater in Europe and the English-speaking countries than in East Asia
(Tables 1 & 2).

218 Gerhard Meisenberg
The Journal of Social, Political and Economic Studies
Another unambiguous result is that the effect of education is not
mediated by income. If education reduced fertility only because educated
people are wealthier, and personal wealth reduced fertility, then Table 4
would show no independent effects of education but consistent negative
effects of income, which clearly is not the case. This result invalidates the
common perception among development strategists that economic
development is the best contraceptive (Rockefeller, 1978). The late Julian
Simon already noted: “Whereas without any or much subclassification, the
relationship of income to fertility is sometimes weakly positive and
sometimes negative, within very detailed subclassifications, the
relationship is almost invariably positive.” (Simon 1974).
Education has not only economic but (hopefully) also cognitive
consequences. This raises the possibility that the fertility-reducing effect of
education is either mediated by cognitive skills acquired in school, or that
pre-existing intelligence raises educational attainment and also lowers
fertility. However, several studies have shown that the relationship
between education and fertility is at least as strong as, and often slightly
stronger than, the relationship between intelligence and fertility. This is the
case for the General Social Survey in the United States (Parker, 2004).
However, in this survey the measure of intelligence consisted only of a ten-
item vocabulary test, and therefore education may well have been the more
accurate measure of the respondent’s intelligence.
The National Longitudinal Survey of Youth used a more accurate IQ
test (the AFQT). Here, the correlation of the number of children at age 39-
47 is -.177 with highest grade completed, -.170 with highest educational
degree, -.166 with IQ (measured between age 14 and age 22) and -.055 with
log-transformed family income (N = 6163). Another American study found
that effects of IQ on fertility are mediated mainly by education (Retherford
& Sewell, 1989). These observations suggest that it is not necessarily
“general” intelligence, but scholastic ability or interests that are inimical to
reproduction. At the country level (N = 187 countries), the correlation of
the total fertility rate is -.770 with log-transformed GDP, -.785 with the
average IQ in the country (Lynn & Vanhanen, 2006), and -.845 with a
composite measure of education (unpublished observations by the author).
A likely proximal mechanism is the effective use of contraception.
Even in the United States during the last third of the 20th century,

How Universal is the Negative Correlation between Education and Fertility? 219
Volume 33, Number 2, Summer 2008
contraception failure was still a major reason for fertility differentials by
income and education (Henshaw, 1998; Kost & Forrest, 1995; Schirm et al.,
1982). Also higher intelligence leads to more effective family planning and
a smaller number of unwanted pregnancies (Udry, 1978).
Opportunity costs are another possible mechanism. Time spent in
school conceivably competes with time available for children, especially for
women. Therefore lengthy schooling but not necessarily a high educational
degree should reduce fertility, but this is contradicted by Tables 1 and 2.
The results are more compatible with the idea that educational degrees buy
better career opportunities, which are seen as incompatible with
parenthood. Indeed, female labor force participation but not length of
schooling has been described as an independent predictor of the total
fertility rate at the country level (McClamroch, 1996). Thus the
contraceptive effect of female education is likely to be mediated by higher
female employment.
Other theories invoke a rational approach to life. At low levels of
development, birth rates are high because people do not even perceive
childbearing as being under their control (Knodel et al., 1984; van de
Walle, 1992). The adoption of a rational approach to life, which is fostered
by formal education, leads first to a verbally expressed desire for smaller
families, followed later by the increasingly effective use of contraception
(Bongaarts, 1997). Therefore differential fertility is greatest in countries
in which the rational approach to life has already infested the upper classes,
but has not yet trickled down to the lower classes.
Another line of reasoning starts with the observation that most people
prefer to interact with others who are similar to themselves (Rushton &
Bons, 2005). In consequence, children are more attractive for intellectually
unsophisticated people than for those with highly developed intellectual
interests. This has been proposed as one reason for the persistence of
educational fertility differentials even in the most advanced societies
(Meisenberg, 2007, p. 334/335).
The negative relationship between social success and reproductive
success in modern societies has puzzled some sociobiologists, who naively
expected this relationship to be universally positive (Abernethy, 1999;
Kanazawa, 2005; Vining, 1986; Weeden et al., 2006). However,
evolutionary theory actually predicts that humans do not have a strong

220 Gerhard Meisenberg
The Journal of Social, Political and Economic Studies
desire for children. Like other animals, ancestral humans did not know
about the mechanism of reproduction and could not use conscious choice to
adjust their reproductive behavior to their desire for children; and ev en
after they had learned about the connection between sexual intercourse and
childbirth, they did not consistently use this knowledge for family planning
until the 19th or 20th century. A strong desire for children would only have
impelled ancestral humans to snatch other people’s children, which would
not be adaptive in the Darwinian sense (Meisenberg, 2007, p. 88/89).
One consequence of a negative relationship between education and
reproduction is that less educated parents provide less favorable
environments for their children. The UNICEF publishes data on relative
child poverty, defined as the percentage of children growing up in
households with incomes below 50% of the national median income
(UNICEF Innocenti Research Centre, 2007). In the 26 countries for which
data are available, the correlation between child poverty and the education-
children correlation for young respondents in the World Values Survey is -
.618 (p = .001). Variations in the strength of the education-fertility
relationship explain at least 38% of the differences in relative child poverty
between countries.
In addition, the heritability of educational attainment (measured as
years of schooling) has been estimated as 60% for samples from several
countries (Vogler & Fulker, 1983), 68% for the United States (Rowe et al.,
1999) and 57-82% for Australia (Baker et al., 1996). Genetic effects on
years in school overlap partially with genetic effects on psychometric
intelligence (Rowe et al., 1999). Therefore the preferential reproduction of
less educated parents handicaps children not only through a less favorable
rearing environment, but also through genetic disadvantages. Unlike
environmental effects, genetic disadvantages are cumulative over
generations.
Finally, we must be aware that the negative relationship between
education and fertility is of very recent origin. For example, in the early 17th
century, wealthy Englishmen had about twice as many surviving children as
the poor. This was caused mainly by greater marital fertility of the rich.
Since the rich were more literate than the poor, there was also a positive
relationship between literacy and fertility (Clark, 2007).
Many other studies have found positive relationships of reproduction

How Universal is the Negative Correlation between Education and Fertility? 221
Volume 33, Number 2, Summer 2008
with wealth or occupational status in pre-industrial societies, not only in
Europe (Heckh, 1952; Røskaft et al., 1992; Stys, 1957/58; Weiss, 1990) but
also in East Asia (Lamson, 1935; Notestein, 1938; Yamamura, 1985) and
various small-scale societies (Casimir & Rao, 1995; Cronk, 1991; Turke &
Betzig, 1985). This is the reason why fertility often rises in the earliest
stages of socio-economic development, before plummeting with the spread
of formal education (Jejeebhoy, 1995; Nag, 1980). It was only during the
demographic transition of the late 19th century in Europe and North
America (and later elsewhere) that the “surviv al of the richest” gave way
to the “survival of the dumbest” that we see today (Stevenson, 1920).
References
Abernethy, Virginia D.
(1999) A Darwinian account of the fertility opportunity hypothesis. Population &
Environment 21: 119-148.
Baker, Laura A., S.A. Treloar, C.A. Reynolds, A.C. Heath and N.G. Martin
(1996) Genetics of educational attainment in Australian twins: sex differences and
secular changes. Behavior Genetics 26: 89-102.
Bongaarts, John
(1997) Trends in unwanted childbearing in the developing world. Studies in Family
Planning 28: 267-277.
Bongaarts, John
(2003) Completing the fertility transition in the developing world: the role of
educational differences and fertility preferences. Population Studies 57: 321-
336.
Budig, Michelle J. and P. England
(2001) The wage penalty for motherhood. American Sociological Review 66: 204-225.
Cao, Gui-Ying and W. Lutz
(2004) China’s future urban and rural population by level of education. In: W. Lutz,
W.C. Sanderson and S. Scherbov (eds): The End of World Population Growth
in the 21st Century. New Challenges for Human Capital Formation and
Sustainable Development, pp. 265-282. London: Earthscan.
Casimir, Michael J. and A. Rao
(1995) Prestige, possessions and progeny. Cultural goals and reproductive success
among the Bakkarwal. Human Nature 6: 241-272.
Castro Martín, Teresa
(1995) Women’s education and fertility: results from 26 Demographic and Health
Surveys. Studies in Family Planning 26: 187-202.
Clark, Gregory
(2007) A Farewell to Alms. A Brief Economic History of the World. Princeton:
Princeton University Press.
Coleman, David A.
(1990) The demography of social class. In: C.G.N. Mascie-Taylor (ed): Biosocial
Aspects of Social Class, pp. 59-117. Oxford: Oxford University Press.

222 Gerhard Meisenberg
The Journal of Social, Political and Economic Studies
Cronk, Lee
(1991) Wealth, status, and reproductive success among the Mukogodo of Kenya.
American Anthropologist 93: 345-360.
Edin, Karl Arvid and E.P. Hutchinson
(1935) Studies of Differential Fertility in Sweden. London: P.S. King & Son.
Goujon, Anne and W. Lutz
(2004) Future human capital: population projections by level of education. In: W.
Lutz, W.C. Sanderson and S. Scherbov (eds): The End of World Population
Growth in the 21st Century. New Challenges for Human Capital Formation and
Sustainable Development, pp. 121-157. London: Earthscan.
Graff, Harvey J.
(1979) Literacy, education, and fertility, past and present: a critical review.
Population and Development Review 5: 105-140.
Heckh, G.
(1952). Unterschiedliche Fortpflanzung ländlicher Sozialgruppen aus
Südwestdeutschland seit dem 17. Jahrhundert (Differential reproduction of
rural social groups in South-West Germany since the 17th century). Homo 3:
169-175.
Henshaw, Stanley K.
(1998) Unintended pregnancy in the United States. Family Planning Perspectives 30:
24-29.
Herrnstein, Richard and C. Murray
(1994) The Bell Curve. Intelligence and Class Structure in American Life. New York:
Free Press.
Inglehart, Ronald, M. Basáñez, J. Díez-Medrano, L. Halman and R. Luijkx
(2004). Human Beliefs and Values. A Cross-Cultural Sourcebook Based on the 1999-2002
Values Surveys. México: Siglo Editores.
Irwing, Paul and R. Lynn
(2006) The relation between childhood IQ and income in middle age. Journal of
Social, Political and Economic Studies 31: 191-196.
Jejeebhoy, Shireen J.
(1995) Women’s Education, Autonomy and Reproductive Behaviour: Experience from
Developing Countries. Oxford: Clarendon Press.
Kanazawa, Satoshi
(2005) An empirical test of a possible solution to “the central theoretical problem of
human sociobiology.” Journal of Cultural and Evolutionary Psychology 3: 255-
266.
Kemkes-Grottenthaler, Ariane
(2003) Postponing or rejecting parenthood? Results of a survey among female
academic professionals. Journal of Biosocial Science 35: 213-226.
Knodel, John, N. Havanon and A. Pramualratana
(1984) Fertility transition in Thailand: a qualitative analysis. Population and
Development Review 10: 297-328.
Kost, Kathryn and J.D. Forrest
(1995) Intention status of U.S. births in 1988: differences by mothers’ socioeconomic
and demographic characteristics. Family Planning Perspectives 27: 11-17.
Lamson, Herbert D.
(1935) Differential reproduction in China. Quarterly Review of Biology 10: 308-321.

How Universal is the Negative Correlation between Education and Fertility? 223
Volume 33, Number 2, Summer 2008
Lynn, Richard
(1996). Dysgenics. Genetic Deterioration in Modern Populations. Westport (CT):
Praeger.
Lynn, Richard and M. van Court
(2004) New evidence of dysgenic fertility for intelligence in the United States.
Intelligence 32: 193-201.
Lynn, Richard and T. Vanhanen
(2006) IQ & Global Inequality. Augusta (GA): Washington Summit.
Mare, Robert D.
(1997) Differential fertility, intergenerational educational mobility, and racial
inequality. CDE Working Papers, No. 97-03, retrieved March 2008 from:
http://www.ssc.wisc.edu/cde/cdewp/1997papers.htm
Martin, Steven P.
(2000) Diverging fertility among U.S. women who delay childbearing past age 30.
Demography 37: 523-533.
McClamroch, Kristi
(1996) Total fertility rate, women’s education, and women’s work: What are the
relationships? Population and Environment 18: 175-186.
Meisenberg, Gerhard, E.C. Lawless, E. Lambert and A. Newton
(2006) The social ecology of intelligence on a Caribbean island. Mankind Quarterly
46: 395-433.
Meisenberg, Gerhard
(2007) In God’s Image. The Natural History of Intelligence and Ethics. Brighton:
Book Guild.
Nag, Moni
(1980) How modernization can also increase fertility. Current Anthropology 21: 571-
587.
Neiss, Michelle, D.C. Rowe and J.L. Rodgers
(2002) Does education mediate the relationship between IQ and age of first birth? A
behavioural genetic analysis. Journal of Biosocial Science 34: 259-275.
Notestein, Frank W.
(1938) A demographic study of 38,256 rural families in China. Milbank Memorial
Fund Quarterly 16: 57-79.
Nyström, Sune, L.O. Bygren and D.R. Vining
(1991) Reproduction and level of intelligence. Scandinavian Journal of Social
Medicine 19: 187-189.
Parker, Macon Paul
(2004) Intelligence and dysgenic fertility: re-specification and reanalysis.
Chrestomathy: Annual Review of Undergraduate Research at the College of
Charleston 3: 167-181.
Retherford, Robert D. and W.H. Sewell
(1988) Intelligence and family size reconsidered. Social Biology 35: 1-40.
Retherford, Robert D. and W.H. Sewell
(1989) How intelligence affects fertility. Intelligence 13: 169-185.
Retherford, Robert D. and N.Y. Luther
(1996) Are fertility differentials by education converging in the United States?
Genus 52 (3-4): 13-37.

224 Gerhard Meisenberg
The Journal of Social, Political and Economic Studies
Rockefeller, John D.
(1978) Population growth: the role of the developed world. Population and
Development Review 4: 509-516.
Røskaft, Eivin, A. Wara and A. Viken
(1992) Reproductive success in relation to resource-access and parental age in a
small Norwegian farming parish during the period 1700-1900. Ethology and
Sociobiology 13: 443-461.
Rowe, David C., W.J. Vesterdal and J.L. Rodgers
(1999) Herrnstein’s syllogism: genetic and shared environmental influences on IQ,
education, and income. Intelligence 26: 405-423.
Rushton, J. Philippe and T.A. Bons
(2005) Mate choice and friendship in twins. Evidence for genetic similarity.
Psychological Science 16: 555-559.
Schirm, Allen L., J. Trussell, J. Menken and W.R. Grady
(1982) Contraceptive failure in the United States: the impact of social, economic
and demographic factors. Family Planning Perspectives 14: 68-75.
Simon, Julian
(1974) The Effects of Income on Fertility. Chapel Hill (NC): Population Center.
Stevenson, T.H.C.
(1920) The fertility of various social classes in England and Wales from the middle
of the nineteenth century to 1911. Journal of the Royal Statistical Society 83:
401-432.
Stys, W.
(1957/58) The influence of economic conditions on the fertility of peasant women.
Population Studies 11: 136-148.
Turke, Paul W. and L.L. Betzig
(1985) Those who can do: wealth, status, and reproductive success on Ifaluk.
Ethology and Sociobiology 6: 79-87.
Udry, J. Richard
(1978) Differential fertility by intelligence: the role of birth planning. Social Biology
25: 10-14.
UNICEF Innocenti Research Centre
(2007) Report Card 7. Child Poverty in Perspective: An Overview of Child Well-Being in
Rich Countries. New York: UNICEF. Available at http://www.unicef-
icdc.org/presscentre/presskit/reportcard7/rc7_eng.pdf.
van de Walle, Etienne
(1992) Fertility transition, conscious choice, and numeracy. Demography 29: 487-
502.
Vining, Daniel R.
(1986) Social versus reproductive success: the central theoretical problem of human
sociobiology. Behavioral and Brain Sciences 9: 167-216.
Vogler, George P. and D.W. Fulker
(1983) Familial resemblance for educational attainment. Behavior Genetics 13: 341-
354.
Weeden, Jason, M.J. Abrams, M.C. Green and J. Sabini
(2006) Do high-status people really have fewer children? Human Nature 17: 377-392

How Universal is the Negative Correlation between Education and Fertility? 225
Volume 33, Number 2, Summer 2008
Weinberger, Mary Beth
(1987) The relationship between women’s education and fertility: selected findings
from the World Fertility Surveys. International Family Planning Perspectives
13: 35-46.
Weiss, Volkmar
(1990) Social and demographic origins of the European proletariat. Mankind
Quarterly 31: 127-152.
Yamamura, Kozo
(1985) Samurai income and demographic change: the genealogies of Tokugawa
bannermen. In: S.B. Hanley and A.P. Wolf (eds): Family and Population in
East Asian History, pp. 62-80. Stanford: Stanford University Press.
Yang, Yang and S.P. Morgan
(2003) How big are educational and racial fertility differentials in the U.S.? Social
Biology 50: 167-187.
Appendix
The strength of the relationship between number of children and education
(averaged from highest degree and years in school), waves 2, 3 and 4 of the
World Values Survey combined, with age controlled. The strength of the
relationship is expressed as standardized β, which is numerically very close to the
part and partial correlations.
Old females Old males Young females Young males
Country st. β N st. β N st. β N st. β N
Albania -.266 227 -.278 179 -.196 262 -.085 280
Algeria -.357 165 -.347 152 -.154 151 -.197 161
Argentina -.132 771 -.077 536 -.282 629 -.176 583
Armenia -.129 806 -.149 559 -.025 836 -.050 868
Australia -.107 878 -.004 662 -.160 710 -.103 725
Austria -.111 854 -.083 573 -.115 555 -.033 498
Azerbaijan -.300 596 -.160 553 -.134 1031 -.047 1021
Bangladesh -.220 179 -.201 410 -.265 1093 -.199 1127
Belarus -.171 862 -.147 545 -.168 657 -.034 533
Belgium -.095 1167 .042 876 -.097 707 -.035 753
Bosnia -.205 508 -.163 494 -.186 558 -.097 664
Brazil -.272 676 -.151 637 -.186 947 -.200 792
Bulgaria -.297 815 -.256 621 -.326 510 -.195 501
Canada -.150 805 -.136 661 -.166 712 -.094 752
Chile -.214 727 -.195 504 -.223 840 -.113 794

226 Gerhard Meisenberg
The Journal of Social, Political and Economic Studies
China -.211 581 -.084 639 -.213 750 -.110 893
Colombia -.363 474 -.203 397 -.372 952 -.154 1135
Croatia -.146 562 -.054 456 -.234 394 -.092 381
Czech Rep. -.180 829 -.055 552 -.117 510 -.035 630
Denmark -.149 469 -.024 373 -.140 354 -.065 380
Domin. Rep. -.090 194 -.021 132
Egypt -.318 443 -.183 496 -.215 703 -.158 596
El Salvador -.370 218 -.228 211 -.386 409 -.202 322
Estonia -.094 839 -.013 485 -.135 579 .018 577
Finland -.074 578 .011 447 -.078 431 -.035 472
France -.241 621 -.121 517 -.202 551 -.091 475
Georgia -.075 1143 -.051 759 -.140 1194 -.041 1081
Germany -.090 1765 -.070 1393 -.071 1246 -.045 1021
Greece -.165 209 -.203 138 -.152 436 -.002 302
Hungary -.169 525 -.082 386 -.242 353 -.116 386
Iceland -.104 361 .000 264 -.208 368 -.062 416
India -.185 856 -.160 934 -.219 1520 -.212 1620
Indonesia -.233 325 -.142 234 -.215 137 -.209 246
Iran -.221 264 -.194 316 -.264 413 -.189 438
Ireland -.117 420 -.083 354 -.185 360 -.152 373
Italy -.117 983 -.152 741 -.242 600 -.164 635
Japan -.066 919 -.056 780 -.064 535 -.027 515
Jordan -.207 148 -.151 163 -.201 335 -.212 204
South Korea -.179 706 -.172 516 -.077 787 -.122 841
Latvia -.120 804 .005 536 -.082 598 -.044 533
Lithuania -.199 774 -.161 517 -.106 535 -.067 598
Luxembourg -.148 282 -.039 248 -.221 254 -.111 271
Macedonia -.303 409 -.154 424 -.305 549 -.139 563
Malta .047 256 -.131 199 -.198 223 -.115 249
Mexico -.221 610 -.207 519 -.275 925 -.176 946
Moldova -.172 493 -.163 334 -.155 408 -.032 463
Morocco -.203 336 -.330 227 -.173 756 -.150 783
Netherlands -.130 466 -.113 348 -.150 372 -.108 355
New Zealand -.153 335 -.044 241 -.135 241 -.133 249
Nigeria -.014 367 -.200 385 -.114 1607 -.073 1459
Norway -.098 777 -.083 618 -.195 614 -.028 664
Pakistan -.129 372 -.126 442 -.118 730 -.110 481
Peru -.323 414 -.213 347 -.260 744 -.190 680
Philippines -.172 372 -.091 286 -.154 599 -.070 639
Poland -.304 869 -.233 615 -.281 540 -.243 558
Portugal -.118 478 -.109 363 -.294 369 -.133 377
Puerto Rico -.240 253 -.130 116 -.164 177 -.188 115

How Universal is the Negative Correlation between Education and Fertility? 227
Volume 33, Number 2, Summer 2008
Romania -.324 597 -.244 460 -.291 341 -.192 405
Russia -.164 1739 -.091 1017 -.161 1351 -.100 1324
Serbia & M. -.214 1007 -.033 878 -.133 413 -.063 444
Singapore -.169 242 -.049 126 -.176 363 -.202 437
Slovakia -.278 440 -.053 327 -.196 403 -.156 406
Slovenia -.233 794 -.224 574 -.176 572 -.036 501
South Africa -.199 1008 -.194 664 -.188 1295 -.087 1552
Spain -.094 1714 -.049 1220 -.211 1142 -.158 1151
Sweden -.004 755 .035 653 -.117 427 -.089 496
Switzerland -.116 832 .058 638 -.026 614 -.038 624
Taiwan -.224 387 -.151 306 -.301 526 -.082 489
Tanzania -.258 155 -.132 213 -.268 320 -.154 402
Turkey -.374 1137 -.267 963 -.330 2046 -.228 2049
Uganda -.277 111 .011 78 -.209 382 -.208 368
Ukraine -.137 1198 -.131 713 -.129 775 -.054 661
United K. -.126 542 -.176 435 -.174 409 -.168 404
USA -.140 995 -.070 806 -.159 730 -.086 774
Uruguay -.259 595 -.192 383 -.284 375 -.121 275
Venezuela -.261 397 -.238 294 -.217 621 -.120 653
Vietnam -.244 235 -.228 194 -.170 240 -.044 248
Zimbabwe -.323 157 -.034 116 -.182 355 -.119 286