Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
Isolating a culture of son preference among
Armenian, Georgian, and Azeri Parents in Soviet-era
Russia∗
Matthias Schief
1
, Sonja Vogt
2
, Elena Churilova
3
, and Charles Efferson
2
1
Department of Economics, Brown University, USA
2Faculty of Business and Economics, University of Lausanne, Switzerland
3
International Laboratory for Population and Health, HSE University, Russia
Abstract
A basic hypothesis is that cultural evolutionary processes sustain differences
between groups, these differences have evolutionary relevance, and they would not
otherwise occur in a system without cultural transmission. The empirical challenge
is that groups vary for many reasons, and isolating the causal effects of culture
often requires appropriate data and a quasi -experimental approach to analysis. We
address this challenge with historical data from the final Soviet census of 1989, and
our analysis is an example of the epidemiological approach to identifying cultural
variation. We find that the fertility decisions of Armenian, Georgian, and Azeri
parents living in Soviet-era Russia were significantly more son-biased than those
of other ethnic groups in Russia. This bias for sons took the form of differential
stopping rules; families with sons stopped having chil - dren sooner than families
without sons. This finding suggests that the increase in sex ratios at birth in the
Caucasus, which began in the 1990s, reflects a cul- tural preference for sons that
predates the end of the Soviet Union. This result also supports one of the key
hypotheses of gene-culture coevolution, namely that cultural evolutionary processes
can support group-level differences in selection pressures that would not otherwise
occur in a system without culture.
JEL-Codes: J13, J16, Z1
Keywords: son preference, sex ratio, sex-selective abortions, Caucasus, gene-
culture coevolution
Media summary:
Fertility decisions of Armenian, Georgian, and Azeri parents
living in Soviet-era Russia reveal strong son bias.
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
2
∗
We thank Evgeny M. Andreev for kindly compiling the relevant data from the special sample micro-
data file of the Soviet Population Census of 1989. Sonja Vogt and Charles Efferson thank the Swiss
National Science Foundation (Grant Nrs. 100018 185417 and 100018 215540). Address correspondence
to matthias schief@brown.edu and charles.efferson@unil.ch.
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
I
Introduction
A seemingly innocuous hypothesis is that social learning and cultural transmission can
produce and maintain variation between groups in ongoing contact with each other
(Boyd and Richerson, 1985; Soltis et al., 1995; Richerson and Boyd, 2008; Henrich,
2017; Laland, 2017). The hypothesis seems innocuous if understood to mean simply that
cultures should vary. In reality, however, the idea is not necessarily innocuous. If cultures
vary specifically because of cultural evolutionary processes, genes and culture can
combine to generate evolutionary dynamics that would not otherwise be possible. Of
particular interest, cultural group selection becomes a possibility (Boyd et al., 2011;
Richerson et al., 2016; Francois et al., 2018; Henrich and Muthukrishna, 2021).
For present purposes, by cultural group selection we mean that two or more groups
exhibit systematically different, socially learned behaviours that lead to systematic,
group-level variation in fitness values for the individuals who comprise the groups. Here
we do not demonstrate or even test for cultural group selection of this sort. Indeed, our
analyses highlight just how challenging such an exercise would be, especially in terms
of drawing sound inferences about the causal role of culture. We do, however, take a
step in the direction of testing for this kind of cultural group selection in two ways. First,
we do not simply assume that observed differences in behaviour between groups is
cultural variation just because it seems like it probably should be cultural variation.
Rather, we work with data that attenuate, even if they cannot eliminate, the role of
confounding sources of variation. Second, we do not examine behavioural variation in
decision-making domains that have a vague relationship to fitness. Rather, we examine
family planning, specifically sex-biased fertility choices, and in this sense we focus on a
domain that should have a relatively direct link to realised fitness values.
A necessary but not sufficient condition for cultural group selection is that groups are
different from each other specifically because of cultural evolutionary processes. This is
easy to say but hard to pin down in terms of what it might actually mean for the
working empiricist (Lonati et al., 2024). Groups can be different for all sorts of reasons.
Isolating the cultural component can be challenging because cultural variation often
correlates strongly with other forms of group-level variation. For example, if the people
in one country behave differently from the people in another country (Hofstede, 2001;
Henrich et al., 2005; Bell et al., 2009; Muthukrishna et al., 2020), some or all of this
group-level variation may be cultural, but some or all of it may simply reflect facultative
responses to environmental variation (Lamba and Mace, 2011).
One straightforward work-around is what is sometimes called the “epidemiological”
approach to identifying cultural variation
(F
ern
´
andez,
2011; Lonati et al., 2024). The
basic idea is to compare people who have different cultural backgrounds but live in one
place under a common set of institutions and economic conditions.
By holding the
institutional and economic setting constant, one hopes to eliminate or at least attenuate
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
4
the effects of environmental variation as a confound, and recent research provides a number
of examples in different decision-making domains (Antecol, 2000; Fernandez, 2007;
F
ern
´
andez
and Fogli, 2009; Alesina and Giuliano, 2010; Luttmer and
Singhal, 2011; Enke, 2019; Almond et al., 2013; Galor and
O
¨
zak, 2016; Lowes et al.,
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
2017; Giuliano and Nunn, 2021).
Of course, some domains are more interesting than others in evolutionary terms. To
illustrate, imagine a study showing that children born in London to Spanish par- ents
sometimes say “la calle” but almost never “die Straße”, while children born in London
to German parents sometimes say “die Straße” but almost never “la calle”. Such a study
would provide a clear example of the epidemiological approach, and it might easily
convince us that the observed group-level variation is cultural. However, assuming la
calle and die Straße work equally well when speaking to Spanish speak- ers and German
speakers respectively, and assuming they work equally poorly when speaking to others,
we might also assume the observed variation is selectively neutral in gene-culture
coevolutionary terms. In particular, it would not and could not feed back to affect
survival and reproduction, the basic components of individual fitness.
Cultural variation in reproductive choices is at the opposite end of the spectrum,
and this is our starting point in the present paper. We use census data to examine
possible cultural variation in choices related to fertility, family planning, and parental
investment. Specifically, we use archival data from the final Soviet census of 1989 to
study son bias or daughter bias among different ethnic groups within the Soviet Union.
To attenuate environmental confounds, we limit attention to people who were all living
in Soviet-era Russia and thus shared a common economic and institutional setting.
This allows us to identify cultural differences in son bias across ethnic groups living in
Russia, foreshadowing the differential evolution of sex ratios in their ethnic homelands.
Thanks to the pioneering work of Guilmoto (2013), Guilmoto and
Duth
´
e
(2013),
and
Duth
´
e
et al. (2012), we know that, shortly after the dissolution of the Soviet
Union, sex ratios at birth (the number of boys born for every 100 girls) started to rise
in some successor states due to sex-selective abortion. The increase was strongest in
Armenia and its neighboring countries in the Caucasus, and Armenia now has one of
the most distorted sex ratios in the world (Schief et al., 2021). In other parts of the
former Soviet Union, in contrast, with the Baltics a clear example, the sex ratios at
birth remain undistorted.(United Nations, Department of Economic and Social Affairs,
Population Division, 2022)
What explains the observed rise in the sex ratio at birth in the Caucasus following the
end of the Soviet Union? An intriguing possibility is that heightened sex ratios in
the contemporary Caucasus reflect a deeply rooted cultural preference for sons that was
already in place before families started selectively aborting females. To investigate this
possibility we need to go further back to Soviet-era Russia and look for alterna- tive
signatures of a culturally evolved son bias. One candidate signature, especially
appropriate for census data, is differential stopping rules. A differential stopping rule
would mean, for example, that parents with two daughters and no sons are more likely
to continue having children than parents with two children, at least one of which is a son.
A culturally evolved son bias that manifests itself in this way is intriguing because it has
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
6
the potential, like sex-selective abortion, to impose a fitness cost. In particular, stopping
rules imply that some families may not optimize the quantity-quality tradeoff, a classic
topic in behavioural ecology (Lack et al., 1954; Stearns, 1992; Roff, 1993; Mul- der, 2000;
Walker et al., 2008; Lawson and Mace, 2011; Lawson and Borgerhoff Mulder,
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
2016), specifically because of their son-biased culture.
An alternative hypothesis is that the rise in sex ratios was driven by the emergence
of son bias, a recent cultural shift that for some reason occurred in the Caucasus, but not
elsewhere. Contemporary sex-selective abortion may still be dysfunctional, but the son-
biased culture in question would not be deep-rooted and would have evolved under
different circumstances. Finally, perhaps the people of the Caucasus did not have and
do not have a culturally evolved preference for sons at all. Instead, rapidly changing
economic and institutional conditions in the early 1990s may have simply created new
incentives favouring sons in the Caucasus but not, for example, in the Baltics. Sex allo-
cation is another classic topic in behavioural ecology, and in general we are quite happy
to think about differential investment in male versus female offspring without recourse
to the cultural evolution of preferences for one sex over another (Fisher, 1999; Trivers
and Willard, 1973; Leimar, 1996; Schindler et al., 2015). In such cases, differential in-
vestment simply reflects a facultative response to incentives, which we conceive broadly
to include any currency that might affect fitness in a sexually reproducing species.
Our data do not allow us to study fitness costs directly. They do, however, allow an
epidemiological approach to differential stopping rules before the emergence of distorted sex
ratios and by extension an examination of son bias in the Caucasus more broadly. Our
results suggest that son bias in the Caucasus, along with whatever distortions and fitness
costs it may or may not involve, is likely based on a deep-rooted preference for sons that
exists specifically because of cultural evolutionary processes.
II
Background
Identifying a culture of son bias can be challenging because such a culture does not
always translate into a male-biased sex ratio, and a male-biased sex ratio does not
necessarily require a culture of son bias. If skewed sex ratios at birth follow from a
culture of son bias, at least three conditions must hold (Lesthaeghe and Vanderhoeft,
2001; Guilmoto, 2009). First, parents must harbor culturally evolved son-biased fertility
preferences. Second, fertility rates must be low so that parents face a significant risk of
ending up without a son unless they actively engage in sex selection. Third, parents must
possess both the means and willingness to undertake sex selection. In particular, pre-
natal sex selection, which is the primary issue in the Caucasus, requires access to
obstetric sonography and some acceptance of abortion as a method of family planning
(Schief et al., 2021).
Before the 1990s, obstetric sonography was not widely available in the Caucasus.
Thus, even if a strong culture of son bias was in place, it could not have affected
population-level sex ratios at birth, and indeed sex ratios in Armenia, Georgia, and
Azerbaijan remained at their natural levels in Soviet times. Therefore, we cannot
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
8
compare historically undistorted sex ratios to contemporary sex ratios and conclude that
the change reveals a cultural shift toward son-biased fertility preferences.
Regarding the availability of medical ultrasound,
Duth
´
e
et al. (2012) report that
ultrasound technology began to appear in the Caucasus in the late 1980s and was fully
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
introduced in most hospitals by the mid 1990s. Hohmann et al. (2014) also report that
sex-screening methods became quickly available following the dissolution of the Soviet
Union. Michael et al. (2013) explains the limited availability of ultrasound machines in
the Soviet period as a lack of domestic manufacturing capacity coupled with Western
restrictions on exporting technology with military applications.
Without sex selection, son-biased parents can increase the probability of having a son
by using differential stopping rules. Put simply, if you have a son, and you are content
with the number of children you have, stop having children. If you do not have a son,
keep having children, and do so even if you would otherwise stop. While such a strategy
does not affect the sex ratio in the population, it does alter the distribution over family
compositions. For example, families with only daughters appear less often than
expected, and families in which the final child is a son appear more often than expected.
We can thus analyze historical fertility outcomes to document son preferences even when
sex ratios provide no insights.
Our strategy for inferring son preferences requires data on within-household fertility
outcomes as aggregate sex ratios are largely unaffected by stopping rule behaviour.
To be precise, fertility stopping can increase the aggregate sex ratio at birth only if
parents with a natural tendency to conceive sons have more children than parents with a
tendency to conceive daughters. Interestingly, continuing to have children until a son is
born means that parents with a natural propensity of conceiving sons would have fewer
children on average, thereby depressing rather than increasing the sex ratio at birth in the
population. In reality, heterogeneity in the natural propensity to conceive offspring of a
particular sex is too limited to imply significant effect for the aggregate sex ratio (Grech
et al., 2018).
What about the possibility that a biased sex ratio does not necessarily reflect a culture
favouring one sex over the other? Variation in the human sex ratio at birth can also
exist without explicit preferences and conscious sex selection by parents. For example,
a significant and growing body of evidence links maternal stress to the death of male
fetuses, which affects the sex ratio at birth in the wake of wars and other forms of social
upheaval (James and Grech, 2017). More broadly, the general theory of sex allocation,
of course, does not require culture. Whether considering the equilibrium sex ratio in the
population (Fisher, 1999) or facultative changes in parental investment based on parental
condition and the relative reproductive values of male versus female offspring (Trivers
and Willard, 1973; Leimar, 1996; Cronk, 2007; Schindler et al., 2015), the basic
evolutionary theory holds without culture. Especially telling, the explicit values in a
cultural group can favour one sex, while investments in terms of actual behaviour favour
the other sex, with the latter consistent with standard predictions under the Trivers-
Willard hypothesis (Cronk, 2017).
For this paper, we have compiled and analyzed a new data set that allows us to
grapple with the potentially complex and subtle relationship between a culture of son
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
10
bias and a son-biased sex ratio. Specifically, we have compiled archival data from the
1989 census of households living in the Russian Soviet Federative Socialist Republic.
The data allow us to compare the fertility outcomes of parents from different cultural
origins who lived in Soviet-era Russia and bore children during the late 1970s and the
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
1980s, a period predating the distortion of sex ratios in the Caucasus. By comparing
Armenians, Georgians, and Azeris to people from other ethnic groups, we aim to discern
whether modern-day disparities in sex ratios at birth across post-Soviet societies mirror
a recent divergence in son bias, or whether they reflect longstanding cultural differences that
were present before the Soviet collapse. Our data provide a way into this problem because
we can combine both the epidemiological approach and the ability to identify differential
stopping rules. The epidemiological approach increases confidence that we can isolate
cultural variation, while the ability to identify stopping rules is crucial because we know
that any son bias during the Soviet era did not take the form of sex-selective abortion.
Aside from general questions in gene-culture coevolutionary theory, our paper also
contributes to the literature on son preference in the Caucasus
(Duth
´
e
et al., 2012;
Guilmoto, 2013;
Me s l
´
e
et al., 2007; Michael et al., 2013; Schief et al., 2021). Previous
research has documented distorted sex ratios, discussed the role of sex-selective abor-
tions, and measured son preferences using questionnaire items or implicit association
tests. All the papers in this literature point to the increased availability of medical
ultrasound as an important part of any explanation of the increase in the sex ratio during
the 1990s.
The same literature, however, also raises the possibility that cultural evolution may
have been unfolding quickly at the time. For example, concurrent developments like the
decline of female political representation, deteriorating labour market conditions, or an
increase in violent conflicts may have caused an unprecedented preference for sons in
the newly independent states of the Caucasus. Simply noting that sex ratios at birth
started to increase dramatically in 1990s does not help to separate the possible
mechanisms. As explained above, this is a general problem associated with identifying
cultural variation either through time or across space. Culture often covaries with other
forms of variation among groups, and this confounded variation limits the scope for
isolating cultural variation in preferences, beliefs, and behaviours.
We examine a culture of son preference by limiting our comparisons to different
ethnic groups all living in Russia prior to the widespread availability of medical ul-
trasound. Thus, from a methodological perspective, our paper also adds to a growing
literature that attempts to identify the causal role of culture by observing the behaviour
of migrants who have different origins but live under a single set of institutions in a
common host country
(F
ern
´
andez
,
2011). In doing so, we provide evidence for a culture
of son preference among Armenian, Georgian, and Azeri parents that predates both
the expanded availability of medical ultrasound as well as the political and cultural
developments associated with the dissolution of the Soviet Union.
Importantly, our focus is on the kind of son bias that causes extreme distortions
of the sex ratio, the kinds of distortion that can only stem from active sex selection
on the part of the parents. For example, as we document in a previous paper (Schief
et al., 2021), conditional on the first two children being girls, the sex ratio among third-
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
12
born children in Armenia is a staggering 330 boys for every 100 girls. In stark contrast,
conditional on the first two children being boys, the sex ratio among third- born children
does not exceed the natural level of 105 boys for every 100 girls. Active
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
choices by parents who want to ensure the birth of a son, rather than environmental
factors, are required to explain these conditional sex ratios.
III
Data and Methods
Our data are a 5% random sample from the 1989 Census in the Russian Soviet Federa-
tive Socialist Republic. The unit of analysis in the census is a household. A household
is defined as a set of individuals who live together in a given apartment. If multiple
families share a communal apartment, they were counted as separate households. For
each household member, we observe their gender, date of birth, and ethnicity. Our 5%
random sample is the only remaining microdata from the 1989 Census. Access to these
data is strictly limited and requests for data access should be submitted to the Institute of
Demography at the Higher School of Economics in Moscow.
We observe the demographic characteristics of all family members who share a
household at the time of the census. However, our data do not include information on
family members who no longer live in the same household. Hence, the set of children
recorded in the census may not be the same as the set of children born to a given couple.
In particular, older children who have already moved out are not recorded. This is a
challenge as studying fertility outcomes requires that we observe all children born to any
given couple. We address this challenge by imposing a restrictive age limit on the oldest
child observed in the census, and we exclude all households with at least one child
exceeding this age limit.
We chose 14 as our cutoff age. With this cutoff, the child recorded in the census as
the oldest child in the family is in fact likely to be the first-born. To see this, consider
what would be required for a child to appear in our sample as the oldest child even though
the child was actually not first-born. Such an outcome would require that the interbirth
interval between the first- and second-born children was especially long or that the first-
born child moved out at an especially young age. For example, if the birth interval
between the first- and second-born child was less than 4 years, the first- born child would
have had to move out before the age of 18 for us to miscategorize the second-born child.
Alternatively, imagine a first-born child who moved out at the age of 23.
Miscategorizing the second child in this case would have required an interbirth interval
of at least 9 years. Because such scenarios are likely rare, we are confident that our
sample consists mostly of households for which we have information on all children. A
cutoff of 14 also means that we consider children born between 1974 and 1989. Finally,
the sample is restricted to families with less than 6 children.
Our goal is to examine the existence of a culture of son preference among Armenian,
Georgian, and Azeri parents that predates the rise in the sex ratios at birth in the
Caucasus. Evidence of historical son preference among Armenian, Georgian, and Azeri
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
14
parents, and the lack of such evidence for parents in other cultures, would suggest that
the rise in the sex ratio at birth in the Caucasus reflects deeply-rooted cultural preferences
rather than recent shifts towards son-biased fertility preferences. To produce such evidence,
we analyze fertility outcomes of parents in Soviet-era Russia and show
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
that Armenian, Georgian, and Azeri parents appear more son biased than Eastern
Europeans or Central Asians.
Comparing the fertility outcomes of parents within Soviet-era Russia has an impor-
tant advantage over comparing the fertility outcomes of parents across different Soviet
republics. To the extent that labour markets or local institutions in Armenia, Georgia,
and Azerbaijan differed from those in other Soviet republics in ways that incentivized
parents to have sons, it is conceivable that incentives, rather than a culture of son
preference, caused fertility outcomes to be more son-biased in the Caucasus. Com-
paring fertility outcomes among parents who all lived in Soviet-era Russia alleviates this
concern. This strategy of comparing individuals of different cultural backgrounds living
in the same institutional setting is called the epidemiological approach and has become
a standard tool to isolate the causal role of culture
(F
ern
´
andez
and Fogli, 2009;
F
ern
´
andez,
2011).
Papers relying on the epidemiological approach often study second-generation mi-
grants because first-generation migrants are likely more strongly selected on specific
traits and may not be representative of individuals in their country of origin. Our analyses
are based on all households recorded as living in Soviet-era Russia at the time of the
1989 census. Although our data will certainly also contain first-generation mi- grants,
many households will have lived in Russia for more than one generation. A limitation
of our data is the lack of information on the place of residence, education, or household
income, which unfortunately prevents us from statistically controlling for the socio-
economic characteristics of the parents in our sample.
III.1
Son preferences and the decision to have a third child
When parents decide whether or not to have another child, they take into account their
current family composition. For example, all things equal, parents with more children
are less likely to have another child compared to parents with fewer children because the
former are more likely to have reached their desired family size. In addition to the
number of existing children, parents with sex-biased fertility preferences also take into
account their sex composition. As a result, parents with a strong desire to have at least
one son are more likely to have a third child when their first two children are girls. Hence,
observed fertility outcomes reveal whether a significant proportion of parents follow
son-biased stopping rules.
We focus on the decision to have a third child. The decision to have a second child
is unlikely to reveal any son bias because the vast majority of parents in our context want
to have at least two children, independent of their sex. Investigating the fertility decisions
of parents with three or more children is problematic because large families are rare.
Moreover, most families with three children will already have a son. Son bias may
influence not only the decision whether to have another child, but also when to have it.
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
16
Specifically, a tendency of parents to invest more in sons may lead to extended intervals
between births after having a son, compared to shorter intervals following the birth of a
daughter (Mace and Sear, 1997).
We compute the odds of having three or more children at the time of the 1989 census
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
son
son
no son
no son
no son
no son
N
third child
/N
no third child
for different parents who are grouped according to the sex of their first two children. The
first group is made up of parents whose first two children are girls. The second group
are parents whose first two children are boys. And the third group are parents whose
first two children are of mixed sex.
Our analysis proceeds by comparing the odds for different groups. For example,
to see whether parents without a son are more likely have a third child compared to
parents who already have a son, we compute the log odds ratio,
(N
third child
/N
no third child
)
son son
where
N
third
child
is the number of households with at least three children and no son
among the first two children. Similarly,
N
no
third
child
is the number of households with-
out a third child and no son among the first two children,
N
third
child
is the number of
households with at least three children and at least one son among the first two chil- dren,
and
N
no
third
child
is the number of households without a third child but at least one son
among the first two children. If parents do not care whether their children are sons or
daughters, then their decisions to have another child depend exclusively on the number
of current children, and the log odds ratios should all be equal to 0. Deviations from zero
indicate a statistical association between the sex composition among the first two
children and the existence of a third child.
Our data allow us to compute the odds ratios by ethnicity. While the data dis- tinguish
between 176 different ethnicities, most have very few observations, and we restrict our
attention to the following: Russians, Estonians, Latvians, Lithuanians, Be- lorusians,
Ukrainians, Moldovans, Goergians, Armenians, Azeris, Turkmens, Uzbeks, Tajiks,
Kazakhs, and Kyrgyz. We further group these ethnicities into four regions based on their
homelands; Russia, Eastern Europe and the Baltics (Estonians, Lat- vians, Lithuanians,
Belorusians, Ukrainians, and Moldovans), Caucasus (Georgians, Armenians, and
Azeris), and Central Asia (Turkmens, Uzbeks, Tajiks, Kazakhs, and Kyrgyz).
IV
Results
Figure 1 shows the evolution of the sex ratio at birth in Russia, Eastern Europe, the
Caucasus, and Central Asia, as reported by the United Nations Population Division. The
clear outlier among these four regions is the Caucasus, which saw a dramatic increase in
the sex ratio at birth during the early 1990s. Against this backdrop, we now turn to
our data to ask whether parents of Armenian, Georgian, and Azeri origin living in the
Russian Soviet Federative Socialist Republic during the 1970s and 1980s similarly stand
out with their fertility choices.
Figure 2 shows point estimates for different log odds ratios along with 95% confi-
ln
, (1)
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
18
dence intervals. The black circle-shaped markers show log ratio of the odds of having a
third child for parents with no son relative to parents with at least one son among the first
two children. Similarly, the blue diamond-shaped markers show the log ratio of
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
Figure 1:
Sex ratios at birth (number of boys per 100 girls). Source: United Nations,
Department of Economic and Social Affairs, Population Division (2022). The sex ratio at
birth for a given region is computed as the average of the country-level sex ratios weighted
by population size.
the odds of having a third child for parents with no daughter relative to parents with
at least one daughter among the first two children. If parents have a preference for a
gender mix among their children, these log odds ratios may be larger than zero, even
if parents do not generally prefer sons over daughters (or vice versa). To isolate son
preference, the green triangle-shaped markers show the log ratio of the odds of having a
third child for parents with no son relative to parents with no daughters.
For Russians, all three log odds ratios are larger than 0, suggesting that Russian
parents have 1) a preference for at least one son, 2) a preference for at least one daughter,
and 3) a stronger preference for sons than for daughters. The inferred son preference
among Russian parents is relatively weak, however, with parents of two girls being only
moderately more likely to have a third child compared to parents of two boys. We
find very similar odds ratios for Eastern Europeans, although somewhat less precisely
estimated. The odds ratios for Central Asians are more suggestive of son preference,
although their confidence interval overlap with the Russian point estimates implying that
we cannot reject the hypothesis that Central Asians are no more son biased than Russians.
In contrast, Armenians, Georgians, and Azeris really stand out. The fertility outcomes
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
20
of parents from the Caucasus region reveal strong son bias, with parents whose first two
children were girls being 1.75 times more likely to have had a third child than parents
whose first two children were boys.
Figure 2 shows how the lack of a son among the first two children is associated
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
Figure 2:
Log odds ratios of having a third child by sex composition of first two children.
Confidence intervals are computed at the 95% level.
with the decision to have a third child. The same results can also be derived within a
regression framework. Table 1 reports the estimated coefficients from a logistic regres-
sion. The dependent variable is an indicator that equals one if the family has a third child
and zero otherwise. We regress this indicator on different variables describing the gender
composition among the first two children, interacted with ethnicity indicators. The
omitted category are Russian parents whose first two children are of mixed gender. This
regression analysis confirms that the lack of a son among the first two children is a strong
predictor of a third child among Armenian, Georgian, and Azeri households, and much
less so among other households. As before, we find that Russians have a preference for
a gender mix, evidenced by the higher odds of having a third child if the first two
children are of the same sex. Eastern Europeans and Central Asians are not statistically
different from Russians. In contrast, the interaction terms with the Caucasus dummy are
statistically significant at the 1% and 5% level, respectively. Ar- menian, Georgian, and
Azeri parents are more likely to have a third child if the first two children are girls and
less likely to have a third child if the first two children are boys.
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
Logistic regression
Dependent variable: Has a third child
No son among first two children 0.231∗∗∗
(0.0118)
No daughter among first two children 0.176∗∗∗
(0.0117)
No son among first two children
×
Eastern Europe
0.000
(0.0467)
No daughter among first two children × Eastern Europe -0.013
(0.0465)
No son among first two children × Caucasus 0.279∗∗
(0.0907)
No daughter among first two children × Caucasus -0.216∗
(0.0955)
No son among first two children
×
Central Asia
0.126
(0.0917)
No daughter among first two children × Central Asia -0.073
(0.0906)
Eastern Europe
0.069
∗
(0.0283)
Caucasus 0.662∗∗∗
(0.0553)
Central Asia 1.068∗∗∗
(0.0545)
Constant -1.792∗∗∗
(0.0072)
Observations 349677
LR chi2 (11) 1539.22
Standard errors in parentheses
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
What about the possibility that infant mortality affects our results? Deceased chil-
dren are not included in the census, implying that in our data the gender of the first
two children may not always correspond to the gender of the first two children born
in a given household. Nevertheless, infant mortality is extremely unlikely to drive our
results. Estimates put the rate of infant mortality in the Soviet Union during the
mid-1980s at about 32 in 1000 (Anderson and Silver, 1986). This number may mask
heterogeneity across different population subgroups. However, even if households faced
different mortality rates, reflecting for example differential access to medical care, the
gender composition within households would remain unaffected as long as the relative
mortality rates of boys versus girls is similar across households. Gender-biased care can
in principle distort the relative survival probabilities of boys versus girls, but there is no
evidence of gender-biased mortality rates in our context. In Table 2, we report the sex
ratios for the same set of ethnicities. As expected, the sex ratios are similar across groups,
and a common sex ratio of 103 boys for every 100 girls cannot be statistically rejected.
V
Discussion
Over the past decades, parents in Armenia, Georgia, and Azerbaijan have resorted to sex-
selective abortions to influence the sex composition of their children. As a result, the
number of boys born for every 100 girls has increased to levels far beyond the natural
sex ratio at birth. Heightened sex ratios first emerged in the 1990s following the
dissolution of the Soviet Union, which brought about important changes in the social and
economic structure of post-Soviet societies. What caused this increase in sex ratios? On
the one hand, the strengthening of national identities and religiosity, the sharp increase
in economic insecurity, the new reliance on male labour migration, and the diminished
representation of women in political position may all have occurred alongside a shift in
fertility preferences toward an increased valuation of sons over daughters. On the other
hand, the 1990s also saw a plummeting of fertility levels and a greatly expanded
availability of medical ultrasound so that parental preferences for sons would for the first
time be reflected in heightened sex ratios. Against this backdrop, it is unclear ex ante
whether the increasing sex ratios at birth during the 1990s reflected a contemporary
shift toward more son-biased fertility preferences or rather a newfound ability to meet
deeply-rooted cultural preferences for sons by preferentially aborting female fetuses.
Using microdata from the 1989 Census in the Russian Soviet Federative Socialist
Republic, we show that the fertility decisions of Armenian, Georgian, and Azeri parents
in the late 1970s and the 1980s were much more son-biased than the fertility decisions of
Russians, Eastern Europeans, and Central Asians. We interpret this finding as evidence
that the recent increase in the sex ratio at birth in the Caucasus is at least partially
reflecting deep-rooted cultural preferences for sons.
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
Russia
Eastern Europe
Caucasus
Central Asia
Sex Ratio
103.0
103.6
105.6
104.7
95% confidence interval
[102.7,103.4]
[102.0,105.1]
[101.7,109.6]
[100.7,108.9]
Number of HHs in sample
700,359
47,065
6,493
5,371
Table 2:
Summary statistics. The sex ratio is defined as the number of boys per 100 girls.
question would require mapping the prevalence of stopping behaviour before the advent
of ultrasound technology into the later evolution of the sex ratio when sex-selective
abortions are possible. Although achieving a precise mapping may not be possible, it
may still be helpful to consider a stylized model to analyze the potential impacts of
stopping rule behaviour on sex ratios in the presence of sex-selective abortions.
Assume that parents can have either two or three children. Parents are more likely to
have a third child if their first two children are girls, and they can choose to resort to sex
selection on the third child to ensure the birth of a son. We assume a natural sex ratio of
105 boys for every 100 girls. We further assume that half the parents resort to sex
selection on the third child to ensure a male birth, provided they do not already have a
son. These assumptions imply a conditional sex ratio of 310 boys for every 100 girls
- consistent with estimates of the conditional sex ratio among third-born children with
two older sisters in Armenia (Guilmoto, 2013; Schief et al., 2021). The free parameters
of this model are the conditional probabilities of having a third child given the gender of
the first two children.
We consider three scenarios for the conditional probabilities of having a third child
given the gender of the first two children. In the first scenario, the probability of having
a third child is 0.5 for all parents, irrespective of the gender of the first two children. The
second scenario introduces sex-biased fertility stopping. In our data, Armenian,
Georgian, and Azeri parents whose first two children are girls are 1.75 times more likely
to be recorded with a third child than parents from the same countries whose first two
children are boys. Motivated by these numbers, we increase the probability of having a
third child from .5 to .875 for parents without a son. As a consequence, average fertility
is also higher in scenario 2 than scenario 1. Finally, in the third scenario, we consider a
case without sex-biased fertility stopping but the same average fertility as in the second
scenario. To achieve this, we set the probability of having a third child to 0.59 for all
parents.
Table 3 reports the sex ratio at birth in each of these scenarios. Notably, the sex ratio
is largest in the second scenario where it reaches almost 114 boys for every 100 girls.
This sex ratio is consistent with estimates for Armenia during the late 1990s and
early 2000s. It therefore seems plausible that the magnitude of the son preferences
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
27
inferred from the observed fertility choices in our data can rationalize the increase in the
sex ratio during the 1990s following the introduction of sex selection technology in the
Caucasus. Admittedly, however, this stylized model needs to be interpreted with caution,
as it relies on strong assumptions. It is likely, for example, that the possibility
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
Scenarios I II III
Implied sex ratio at birth
110.0
113.6
110.7
Table 3:
Hypothetical sex ratios (number of boys per 100 girls) under different scenarios for
stopping rules.
of sex selection would also change the conditional probability of deciding to have third
child.
Interestingly, our results also support – tentatively – one of the key hypotheses of
gene-culture coevolution, namely the claim that cultural evolutionary processes can
reshape selection on genes in ways that would not otherwise occur (Richerson and Boyd,
2008). If distorted sex ratios are a manifestation of the same son-biased culture that we
identify in Soviet-era Russia, then gene-culture coevolutionary models suggest that the
culturally evolved choices behind these biased sex ratios can generate selection pressures
on the genes that influence the primary sex ratio at fertilization (Kumm et al., 1994;
Laland et al., 1995; Kumm and Feldman, 1997). In particular, as sex ratios at birth
started to climb in the 1990s, families were setting the stage for an excess of
marriageable men and a shortage of marriageable women in the 2020s. They were setting
the stage for a system out of equilibrium from the perspective of natural selection and
the associated primary sex ratio it favors.
Why do Armenia, Georgian, and Azeri parents have strong son preferences? While
our paper leaves this question ultimately unanswered, it provided helpful cues in the
quest to explain the phenomenon of son bias in the Caucasus. Whatever the origin of
son biased fertility preference in the Caucasus, our analyses strongly suggest that it has
little to do with the events associated with the dissolution of the Soviet Union. The
falling apart of an established economic and political order and the social upheaval that
went along with that surely affected preferences and norms and may have left a lasting
mark on the culture in these societies. However, strong son bias appears to predate
the 1990s, and the search for a satisfactory explanation should therefore be focused on
earlier history.
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
Acknowledgements
We thank Evgeny M. Andreev for kindly providing us with the required data from the
special sample micro-data file of the Soviet Population Census of 1989. We also thank
two anonymous referees and the editor, Ruth Mace, for their helpful suggestions.
Author Contributions
MS: conceptualisation, analysis, and writing. SV: conceptualisation. EC: data prepa-
ration. CE: conceptualisation and writing.
Financial Support
SV and CE thank the Swiss National Science Foundation (Grant Nrs. 100018 185417
and 100018 215540).
Conflicts of Interest
CE is a member of the editorial board of Evolutionary Human Sciences. Otherwise,
MS, SV, and EC declare none.
Research Transparency and Reproducibility
Access to the special sample micro-data file of the Soviet Population Census of 1989 is
restricted but can be requested.
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
References
Alesina, A. and Giuliano, P. (2010). The power of the family. Journal of Economic
Growth, 15:93–125.
Almond, D., Edlund, L., and Milligan, K. (2013). Son preference and the persistence of
culture: evidence from South and East Asian immigrants to Canada. Population and
Development Review, 39(1):75–95.
Anderson, B. A. and Silver, B. D. (1986). Infant mortality in the soviet union: regional
differences and measurement issues. Population and development review, pages 705–
738.
Antecol, H. (2000). An examination of cross-country differences in the gender gap in
labor force participation rates. Labour Economics, 7(4):409–426.
Bell, A. V., Richerson, P. J., and McElreath, R. (2009). Culture rather than genes
provides greater scope for the evolution of large-scale human prosociality. Proceedings
of the National Academy of Sciences, 106(42):17671–17674.
Boyd, R. and Richerson, P. J. (1985). Culture and the Evolutionary Process. Chicago:
University of Chicago Press.
Boyd, R., Richerson, P. J., and Henrich, J. (2011). Rapid cultural adaptation can facil-
itate the evolution of large-scale cooperation. Behavioral Ecology and Sociobiology,
65(3):431–444.
Cronk, L. (2007). Boy or girl: gender preferences from a darwinian point of view.
Reproductive Biomedicine Online, 15:23–32.
Cronk, L. (2017). Culture’s influence on behavior: Steps toward a theory. Evolutionary
Behavioral Sciences, 11(1):36.
Duth
´
e,
G.,
Mesl
´
e,
F., Vallin, J., Badurashvili, I., and Kuyumjyan, K. (2012). High sex
ratios at birth in the Caucasus: Modern technology to satisfy old desires. Population
and Development Review, 38(3):487–501.
Enke, B. (2019). Kinship, cooperation, and the evolution of moral systems. The Quar- terly
Journal of Economics, 134(2):953–1019.
Fernandez, R. (2007). Women, work, and culture. Journal of the European Economic
Association, 5(2-3):305–332.
F
ern
´
andez,
R. (2011). Does culture matter? Handbook of social economics, 1:481–510.
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
Fisher, R. A. (1999). The Genetical Theory of Natural Selection: A Complete Variorum
Edition. Oxford University Press.
Francois, P., Fujiwara, T., and Van Ypersele, T. (2018). The origins of human prosocial-
ity: Cultural group selection in the workplace and the laboratory. Science Advances,
4(9):eaat2201.
Galor, O. and
O
¨
zak,
O
¨
. (2016). The agricultural origins of time preference. American
Economic Review, 106(10):3064–3103.
Giuliano, P. and Nunn, N. (2021). Understanding cultural persistence and change. The
Review of Economic Studies, 88(4):1541–1581.
Grech, V., James, W. H., and Lauri, J. (2018). On stopping rules and the sex ratio at
birth. Early Human Development, 127:15–20.
Guilmoto, C. Z. (2009). The sex ratio transition in Asia. Population and Development
Review, 35(3):519–549.
Guilmoto, C. Z. (2013). Sex imbalances at birth in Armenia: Demographic evidence
and analysis. Report prepared for the United Nations Population Fund, Yerevan:
UNFPA.
Guilmoto, C. Z. and
Duth
´
e,
G. (2013).
Masculinization of births in eastern europe.
Population & Societies, 11:1–4.
Henrich, J. (2017). The Secret of Our Success: How Culture is Driving Human Evo-
lution, Domesticating Our Species, and Making Us Smarter. Princeton University
Press.
Henrich, J., Boyd, R., Bowles, S., Camerer, C., Fehr, E., Gintis, H., McElreath, R.,
Alvard, M., Barr, A., Ensminger, J., Henrich, N. S., Hill, K., Gil-White, F., Gurven,
M., Marlowe, F. W., Patton, J. Q., and Tracer, D. (2005). “Economic man” in cross-
cultural perspective: behavioral experiments in 15 small-scale societies. Behavioral
and Brain Sciences, 28(6):795–815.
Henrich, J. and Muthukrishna, M. (2021). The origins and psychology of human coop-
eration. Annual Review of Psychology, 72:207–240.
Hofstede, G. (2001). Culture’s Consequences: Comparing Values, Behaviors, Institu-
tions and Organizations Across Nations. Thousand Oaks, CA: Sage, 2nd edition.
Hohmann, S. A.,
Lef
`
evr e,
C. A., and Garenne, M. L. (2014). A framework for analyzing
sex-selective abortion: the example of changing sex ratios in Southern Caucasus.
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
33
International Journal of Women’s Health, pages 889–897.
James, W. H. and Grech, V. (2017). A review of the established and suspected causes
of variations in human sex ratio at birth. Early human development, 109:50–56.
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
Kumm, J. and Feldman, M. W. (1997). Gene-culture coevolution and sex ratios: Ii. sex-
chromosomal distorters and cultural preferences for offspring sex. Theoretical Population
Biology, 52(1):1–15.
Kumm, J., Laland, K. N., and Feldman, M. W. (1994). Gene-culture coevolution and
sex-ratios: The effects of infanticide, sex-selective abortion, sex selection, and sex-
biased parental investment on the evolution of sex ratios. Theoretical Population
Biology, 46(3):249–278.
Lack, D. et al. (1954). The natural regulation of animal numbers. The Natural Regu-
lation of Animal Numbers.
Laland, K. N. (2017). Darwin’s Unfinished Symphony. Princeton University Press.
Laland, K. N., Kumm, J., and Feldman, M. W. (1995). Gene-culture coevolutionary
theory: A test case. Current Anthropology, 36(1):131–156.
Lamba, S. and Mace, R. (2011). Demography and ecology drive variation in cooper-
ation across human populations. Proceedings of the National Academy of Sciences,
108(35):14426–14430.
Lawson, D. W. and Borgerhoff Mulder, M. (2016). The offspring quantity–quality trade-
off and human fertility variation. Philosophical Transactions of the Royal Society B:
Biological Sciences, 371(1692):20150145.
Lawson, D. W. and Mace, R. (2011). Parental investment and the optimization of human
family size. Philosophical Transactions of the Royal Society of London B: Biological
Sciences, 366(1563):333–343.
Leimar, O. (1996). Life-history analysis of the trivers and willard sex-ratio problem.
Behavioral Ecology, 7(3):316–325.
Lesthaeghe, R. and Vanderhoeft, C. (2001). Ready, willing and able: a conceptu-
alization of transitions to new behavioral forms. Diffusion processes and fertility
transition: Selected perspectives, 1:240–264.
Lonati, S., Lalive, R., and Efferson, C. (2024). Identifying culture as cause: challenges
and opportunities. Evolutionary Human Sciences, page TBD.
Lowes, S., Nunn, N., Robinson, J. A., and Weigel, J. L. (2017). The evolution of culture
and institutions:
Evidence from the Kuba Kingdom. Econometrica, 85(4):1065–1091.
Luttmer, E. F. P. and Singhal, M. (2011). Culture, context, and the taste for redistri-
bution. American Economic Journal: Economic Policy, 3(1):157–179.
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
Mesl
´
e,
F., Vallin, J., and Badurashvili, I. (2007). A sharp increase in sex ratio at
birth in the Caucasus. why? how. Watering the neighbour’s garden: The growing
demographic female deficit in Asia, pages 73–88.
Michael, M., King, L., Guo, L., McKee, M., Richardson, E., and Stuckler, D. (2013).
The mystery of missing female children in the Caucasus: an analysis of sex ratios by
birth order. International perspectives on sexual and reproductive health, pages 97–
102.
Mulder, M. B. (2000). Optimizing offspring: the quantity–quality tradeoff in agropas-
toral kipsigis. Evolution and Human Behavior, 21(6):391–410.
Muthukrishna, M., Bell, A. V., Henrich, J., Curtin, C. M., Gedranovich, A., McIn- erney,
J., and Thue, B. (2020). Beyond Western, Educated, Industrial, Rich, and Democratic
(WEIRD) psychology: Measuring and mapping scales of cultural and psychological
distance. Psychological Science, 31(6):678–701.
Richerson, P., Baldini, R., Bell, A. V., Demps, K., Frost, K., Hillis, V., Mathew, S.,
Newton, E. K., Naar, N., Newson, L., and et al. (2016). Cultural group selection plays
an essential role in explaining human cooperation: A sketch of the evidence.
Behavioral and Brain Sciences, 39.
Richerson, P. J. and Boyd, R. (2008). Not by genes alone: How culture transformed
human evolution. University of Chicago press.
Roff, D. (1993). Evolution of Life Histories: Theory and Analysis. Springer Science &
Business Media.
Schief, M., Vogt, S., and Efferson, C. (2021).
Investigating the structure of son bias in
Armenia with novel measures of individual preferences. Demography, 58(5):1737– 1764.
Schindler, S., Gaillard, J.-M.,
Gru
¨
ning,
A., Neuhaus, P., Traill, L. W., Tuljapurkar, S.,
and Coulson, T. (2015). Sex-specific demography and generalization of the trivers–
willard theory. Nature, 526(7572):249–252.
Soltis, J., Boyd, R., and Richerson, P. J. (1995). Can group-functional behaviors evolve
by cultural group selection?: an empirical test. Current Anthropology, 36(3):473–494.
Stearns, S. C. (1992). The Evolution of Life Histories, volume 249. Oxford: Oxford
University Press.
Trivers, R. L. and Willard, D. E. (1973). Natural selection of parental ability to vary the
sex ratio of offspring. Science, 179(4068):90–92.
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
This is an Accepted Manuscript for Evolutionary Human Sciences. This version may be subject to change during
the production process.
DOI: 10.1017/ehs.2024.9
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original
work is unaltered and is properly cited. The written permission of Cambridge University Press must
be obtained for commercial re-use or in order to create a derivative work.
Walker, R. S., Gurven, M., Burger, O., and Hamilton, M. J. (2008). The trade-off
between number and size of offspring in humans and other primates. Proceedings of
the Royal Society B: Biological Sciences, 275(1636):827–834.
https://doi.org/10.1017/ehs.2024.9 Published online by Cambridge University Press
