With the Help of Kin?
With the Help of Kin?
0 Hilde Bras is Professor of Sociology of Consumption and Households in the Department of Social Sciences at Wageningen University. She is currently leading the research project The Power of the Family: Family Influences on Long-Term Fertility Decline in Europe, 1850-1920. Her research interests focus on family factors as determinants of health and demographic outcomes
Relatives play an important role in human reproduction according to evolutionary theories of reproductive behavior, but previous empirical studies show large differences in the effects of kin on fertility outcomes. In our paper we examine the effect of co-resident kin and non-kin on the length of birth intervals over the reproductive life course of Dutch women born between 1842 and 1920. We estimate Cox proportional hazard models for parity progression based on the presence of kin and non-kin in the household while controlling for a large number of individual and community-level characteristics. We find that couples living with their brothers experienced shorter birth intervals whereas couples residing with a widowed father had relatively longer birth intervals. The effects of these types of kin on reproduction were most pronounced up to the birth of the fifth child, but not thereafter. We found no effect for mothers or other types of kin. Co-resident kin and non-kin may play an important role in human reproduction (Johow and Voland 2012; Sear and Coall 2011; Tymicki 2004). Because of the relatively short interval between successive births and the long period during which newborns are dependent on others for their nutrition, parents behave as cooperative breeders (Hrdy 2005, 2008). This means that parents rely on other people-alloparents-who provide assistance in the form of care or resources, thereby helping to raise offspring and enabling parents to increase their reproductive outcomes (Hrdy 1999, 2009; Kramer 2010). According to Hamilton's rule, kin assist in producing and raising offspring because of the indirect fitness benefit that this cooperative behavior yields (Hamilton 1964a, b). Recent empirical studies, many of which cover observations from contemporary and historical pre-transition societies, find that reproductive outcomes are indeed associated with the availability of kin assistance (Pollet et al. 2007; Voland and Beise
Kin; Co-residence; Birth spacing; Inclusive fitness; The Netherlands
2002). Research suggests that among the most important caregivers are the couples
parents (Hawkes et al. 1997, 1998; Tymicki 2004); their children, referred to as
helpers-at-the-nest (Crognier et al. 2001; Kramer 2005); and the siblings of the
couple (i.e., aunts and uncles of the newborn child) (Draper and Hames 2000; Feng
et al. 2010; Hill and Hurtado 2009; Sear et al. 2003; Sear and Mace 2008).
However, empirical results vary because different categories of kin do not always
influence fertility in the same manner (Sear and Coall 2011). These differential
outcomes suggest that the effects of kin help on fertility may be contingent on specific
local conditions and economic factors (Hames and Draper 2004) or that the effects of
kin may vary over the life course of women as they progress from one birth to the next.
Yet, few studies have systematically investigated the differential effects of kin on
reproductive outcomes over the reproductive life course. Moreover, thus far only some
studies have taken account of the influence of a wider group of kin and non-relatives
(Bereczkei 1998; Lyngstad and Prskawetz 2010). The large amount of attention that has
been given to parental influence has left the role of co-resident siblings, cousins, and
non-kin underexplored (Nath et al. 2000; Snopkowski and Sear 2013; Voland and Beise
2002). One of the main reasons for the omission of the wider group of kin and other
coresidents in many studies is the scarcity of sources that encompass detailed information
on both the reproductive behavior of the couple and changes in the presence of kin and
non-kin within the household over longer periods of time.
In this study, we exploit data on changes and differences in the composition of Dutch
households from the second half of the nineteenth century until the beginning of the
twentieth century in order to examine the effect of co-resident (non-)kin on the length
of birth intervals over the reproductive life course. Focusing on sequential fertility
outcomes over the life courses of the women in our sample allows for a more accurate
investigation of kin effects than examining kin influences on total fertility or starting
and stopping behavior. We use a rich data source, the Historical Sample of the
Netherlands, which enables us to uncover the direction, magnitude, and significance
of the effects of co-resident kin and non-kin on the reproductive careers of Dutch
women born between 1842 and 1920. We develop our hypotheses on kin effects on
fertility using insights from inclusive fitness theory.
Inclusive fitness theory is derived from evolutionary biology and concerns the
natural selection of traits, such as altruism, which increase the genetic success of an
organism (Hamilton 1964a, b; Hrdy 2008; Mace 2014). The main assumption of
inclusive fitness theory is that humansas do all speciesstrive to allocate their
resources, including support, knowledge, and time, in such a way that they maximize
their inclusive fitness, expressed as the number of kin weighted by the relative presence
of ones genes, or alleles (Hamilton 1964a, b; Hrdy 2008, 2009). In general, inclusive
fitness theory thus suggests that people are driven to increase the fertility of their lineal
and collateral kin in order to ensure the persistence of their genes in future generations.
As long as the marginal benefits and costs of such assistance are in equilibrium,
alloparental caregiving is likely to positively affect the number and survival chances
of a persons relatives and thus confers an indirect fitness benefit (Grafen 1984). It
follows that if couples can rely on close kin members for support, they are more likely
to raise more and/or better-quality children than couples who are not helped by kin
(Kaptijn et al. 2010; Kramer 2010; Salmon and Shackelford 2008; Schaffnit and Sear
2014). Inclusive fitness theory has thus been invoked to help understand the relatively
long post-generative life span of women. The grandmother hypothesis suggests that
even though postmenopausal mothers no longer reproduce, they can still contribute to
their inclusive fitness by providing resources or care to their children and
grandchildren, thereby enabling them to increase their fertility or child survival rates
(Hawkes 2003; Hawkes et al. 1997, 1998). The grandmother hypothesis has received
strong empirical support (Sear and Coall 2011), although the pronatal effects of the
couples parents may differ between paternal and maternal parents (cf. Euler and
Weitzel 1996; Pollet et al. 2007; Sear et al. 2003; Strassmann and Garrard 2011;
Voland and Beise 2002) or are conditional on the familys social class (Johow and
We extend our analysis of kin influence beyond the couples parents by also
taking into account the effects of the presence of the couples siblings, other
relatives, and household members who have no genetic relationship to the
couple. Based on inclusive fitness theory, we depart from the broad hypothesis
that all close kin will behave cooperatively and exert a positive influence on
reproduction. We posit that the higher the degree of genetic relatedness, the more
stimuli people have to increase their relatives fertility. Genetically close relatives
(e.g., the couples parents or siblings) are expected to behave more
cooperativelyleading to shorter birth intervalsthan genetically more distant relatives
(e.g., the couples aunts and uncles). In addition, uncertainty about the degree
or nonexistence of genetic relatedness lowers the likelihood of cooperative
behavior. The confidence of paternity hypothesis predicts that investments in
grandchildren are lower if they are related through sons than through daughters
(see Strassmann and Garrard 2011). Moreover, the incurred fitness benefit of
cooperative behavior toward kin is possibly lower for fathers of the wife or
husband, as there may be a component of uncertainty regarding the genetic bond
with their offspring. For this reason, we hypothesize that mothers of the
couplein particular, maternal mothershave a more pronounced positive effect on
a womans fertility than fathers of the couple have (Sear and Mace 2008).
Accordingly, living with both parents or having a widowed mother will be
associated with shorter birth intervals compared with living with a widowed
father. The couples siblings are hypothesized to have a positive effect on
reproductive outcomes because of their genetic relation to the couple
(Bereczkei 1998; Feng et al. 2010). Finally, household members with no genetic
relation to the couple are, from an evolutionary perspective, less likely to affect
reproductive outcomes because their fitness is not affected by this behavior.
Although the presence of kin is hypothesized to be associated with shorter
birth intervals in general, the effects of kin on fertility may vary over the
reproductive life span of the women in our sample. Kin effects are hypothesized
to be stronger for the early parities than for later parities for three reasons. First,
drawing on Lyngstad and Prskawetz (2010), the first, formative years of the
family mark a unique transition in the life course of the young couple.
Alloparental support in the early years could be important in compensating for
inexperience among new parents, learning about the nutritional needs of
newborns, or supplying the extra resources needed to feed an additional mouth.
Second, local cultural norms concerning kinship and fertility may affect the
likelihood of living together with kin before the couple is able to establish their
own household. Skinner refers to this particular process as a launching-pad
family system in which living with parents is common in the first few years
after the couple is married, and the newly formed couple establishes their own
independent household only after this initial co-resident phase (Skinner 1997).
When the couple lives with the parents for longer periods of time, economic
conditions or health concerns of the elderly parents, who are likely less able to
provide support, may be the main motivation for kin co-residence (Pebley and
Stupp 1987). Lastly, women who do not constrain their reproduction are not
likely to be affected by the presence of others who are in a position to provide
pronatal support. The intrinsic motivation or biologically heritable specific
fecundity of these women is by itself enough to ensure short birth intervals while
the presence of kin likely does not affect their fertility at all.
In the next section, we discuss our sample, measurements, and methods.
Subsequently, we present the results of our event history analysis showing to what
extent the presence of particular kin and non-kin in the household was related to the
timing of subsequent births. In the final section, we discuss our findings in light of the
recent literature, our hypotheses, and the data and methods used.
Data, Measurements, and Methods
Co-Residence with Kin in The Netherlands
The composition and size of nineteenth-century Dutch households varied
considerably across regions and over time (Bras et al. 2010a; Kok and Mandemakers
2009; Kok et al. 2011). Kin co-residence in the Netherlands during the period of
our analysis was primarily driven by altruistic motives to help kin, in particular
those who were in need of help, and by rational motives, in particular in the
eastern regions where co-residence was associated with inheritance practices
(Kok and Mandemakers 2010). In her study on the dynamics of family structure
in the textile town of Tilburg, Janssens (1993:84) concluded that poverty was not
the sole reason for kin co-residence during the formative years of the family life
cycle. In many cases, co-residence with immediate family occurred because it
was the most practical option, for example, following a failed migration, the
death of a parent, or because of the contributions kin could make to the
household budget. The 1899 census indicates that 5.1 million Dutch individuals
lived in about one million households, with household sizes ranging from 4.6 to
4.9 persons in the northwestern coastal provinces to 5.1 to 5.5 persons in the
eastern provinces of Limburg, Gelderland, Overijssel, and Drenthe (Central
Bureau of Statistics Statline, statline.cbs.nl). In the northwestern provinces, the
nuclear family, or neolocal household formation, was the norm and most couples
did not live together with their parents (Van der Woude 1970). Only in urban
centers such as Amsterdam, Rotterdam, and Haarlem were households larger on
average. The prevalence of co-resident kin was highest in the eastern provinces,
and households in those regions were also much more likely to include persons
who were not genetically related to the couple, such as boarders, servants, or
lodgers (Kok and Mandemakers 2010). In terms of household size, the other
provinces ranged somewhere in between (Fig. 1). Over time, co-residence with
kin became less common in the Netherlands, in particular in the cities from the
early twentieth century onward (Bras et al. 2010a).
The data used in the analysis were obtained from the Historical Sample of the
Netherlands (HSN, release 2007, www.iisg.nl/hsn). The HSN is a representative,
nationwide random sample of about 78,000 individuals (called Research Persons)
born in the Netherlands between 1812 and 1922 (Mandemakers 2002, 2004). The main
sources of the HSN are civil certificates and municipal population registers, which were
established by royal decree on December 22, 1849. As of January 1, 1850, all
municipalities began to keep population registers, based on the census of 1849, on a
dynamic, continuous basis. Since 1861, all Dutch citizens are obligated to report events
for recording in the population registers. Professional civil service workers were hired
to maintain the registers. The main advantage of the HSN for our study is that people
are followed from the cradle to the grave; individual life histories are not censored
when individuals moved to another place in the Netherlands because their migration is
recorded in the population registers. In addition, the date and place of birth or death,
Fig. 1 Number of persons per municipality, divided by the number of households per municipality. Data:
Census 1899 (statline.cbs.nl). Cartography: NLGIS map of the Netherlands 1899 (Boonstra 2007)
marital status, sex, religion, occupation, and relation to the head of the household are
recorded for all members of the household. The exact date when a household member
entered or left, due to birth, death, or migration, is known, including a reference to the
place of origin or destination. From these registers, life courses have been reconstituted
until 1939 when the registers were replaced by a system of personal cards. While the
quality of the HSN is high and the observations in principle cover the entire country,
there are some limitations to its use (Bras 2014; Knotter and Meijer 1995; Van Poppel
et al. 2012; Vulsma 1988).1
First, maintenance of the population registers required all municipalities to
continuously update the records, and when a person moved from one place to
another, information had to be copied from other registers. This was not always
done accurately (Knotter and Meijer 1995; Vulsma 1988). Second, since the
nationwide registration of vital events started on January 1, 1850, the HSN does
not include many complete life histories of women born before this date.
Consequently, data from the 1860s and 1870s will mostly cover women born in
the 1850s who had their first child at a relatively young age, causing a downward
bias in the age at first birth for the earliest cohort. However, for the purpose of our
analysis, this bias does not alter our conclusions on the effects of kin on the length
of the interval between births. Third, the HSN is in continuous development, and
data before 1883 has only been digitized for three provinces (Zeeland, Utrecht,
and Friesland) and one city (Rotterdam). From 1883 on, the HSN has national
coverage, and information is available for all 11 provinces. This gives the
analytical sample we extracted from the HSN an urban bias and explains why the
percentage of farmers and farmworkers (21.6%) is relatively small (Bras 2014;
Van Poppel et al. 2012). Fourth, owing to the nature of the registers, our
observations of kin are limited to those residing with the couple and exclude relatives
living outside of, but in close vicinity to the couples house. This implies that the
frequency of contact with others might have been higher in reality than is
observed based on the presence of kin and non-kin in the household.
Nevertheless, we assume that the presence of co-resident kin allows for frequent
contact and is therefore a good and reliable measure of the availability of kin
(Feng et al. 2010; Morgan and Rindfuss 1984; Schaffnit and Sear 2014;
Snopkowski and Sear 2013; Tsay and Chu 2005).
For our analytical sample, we have selected couples (N = 2628) who entered a
first marriage and for whom complete information is available about (1) the
couples reproductive career until the woman is 50 years old and the husband is
still alive and (2) the presence of kin and non-kin in the household. We focus only
on the waiting time from the first until the second marital birth and on subsequent
birth intervals but exclude the interval between marriage and first birth. The
interval between marriage and first birth was often relatively short and in many
cases not affected by kin presence, but instead by cultural norms which prescribed
that the conception of the first child should follow not too long after marriage
1 See also Janssens (1993:64), who estimates that around 0.2 % of all live births are not listed in the population
registers. Many of these children passed away very soon after birth. Although their omission produces an
upward bias in the average length of birth intervals, it is unlikely to greatly affect our conclusions given the
relatively small number of missing observations.
(Liefbroer and De Jong Gierveld 1995). Focusing on the interval between first and
second or subsequent births reduces the possible influences of such cultural
norms. Our analytical sample includes 2331 couples who had at least two children
and 8052 closed birth intervals connecting to childbirths that took place between
1869 and 1939.
Our outcome variable is the duration of the interval between first and subsequent live,
marital birthsi.e., parity two and upmeasured in months. The birth interval is a
useful indicator of fertility and is widely used in the literature on reproduction (e.g.,
Nath et al. 2000; Van Bavel 2004; Van Bavel and Kok 2004, 2010; Van Poppel et al.
2012). As in most nineteenth- and early-twentieth-century northwest European
societies, fertility within marriage was generally regulated by prolonged breastfeeding,
periodical or complete abstinence, coitus interruptus, and, to a lesser extent, abortion
(Santow 1995). Birth spacing was part of the set of means to regulate fertility and was
usually motivated by a desire to control family size. Life-cycle models of reproductive
behavior indicate that variations in household income or expenditures resulting from
the birth of children may produce an imbalanced ratio between consumers and
producers in terms of household income or labor, resulting in economic stress and
thereby necessitating fertility regulation (see, e.g., Alter 1988; Heckman and Walker
1990; Hotz et al. 1997). Underlying our hypotheses is the assumption that, from an
evolutionary perspective, shorter birth intervals and an increase in child survival rates
increase the fitness of women and the inclusive fitness of kin (Morgan and King 2001).
However, reproduction is energetically costly to women and could jeopardize their
health as well as that of their offspring, leading to a trade-off between quantity and
quality of children (Borgerhoff Mulder 2000; Conde-Agudelo et al. 2006, 2007; Palloni
and Millman 1986). Therefore, optimal birth intervals may be of intermediate length to
ensure a sufficient number of offspring without exposing women and their offspring to
Table 1 reports the mean length of closed birth intervals by parity and for all
births, as well as the total number of living children by birth cohort. Our sample
covers the early period of fertility decline, between 1890 and 1920; the Dutch
fertility transition was completed between 1920 and 1940 or even later in some
regions (see Bras 2014). The average number of living children per woman
declined over time, most strikingly since the 18811890 birth cohort. In line with
Van Poppel et al. (2012), the mean length of birth intervals in our sample
increased most significantly after the last quarter of the nineteenth century, from
26.4 months for women born before 1860 to 28.8 months for women born in 1901
or later. As Table 1 shows, this was in particular due to an increase in the length of
the early parities, suggesting that spacing was increasingly used as a strategy to
control childbirth, although stopping altogether (rather than delaying) was the
main cause of Dutch fertility decline (Van Poppel et al. 2012). When we compare
birth intervals over the life course, we observe that earlier cohorts (e.g., before
1860) experienced birth intervals that were increasing in length over the
reproductive life span. In contrast, women born after 1900 had increasingly shorter
birth intervals over their life courses and stopped having children sooner.
Table 1 Mean length of birth intervals in months by parity and mean total number of children by birth cohort
of the wife
Birth interval by parity (in months)
a Excluding the interval between marriage and first birth
b Number of living children observed at last birth (standard deviation in parentheses)
Source: Historical Sample of the Netherlands, release 2007
We include the presence of one or more of the following kin and non-kin types as
independent variables (all are relative to the couple): (1) parents; (2) siblings; (3) other
relatives, such as cousins, aunts, and uncles; and (4) non-kin (servants, boarders, and
lodgers). Since couples may live only part of their reproductive lives together with
others, all kin and non-kin are coded as time-varying dummy variables with value 0
(not present) or 1 (at least one person of that particular kin type is present in the
household of the couple). Table 2 provides descriptive information about the different
variables used in the models, by parity. Co-residence with kin occurred frequently in
the Netherlands, but during the life course of the couples in our sample, many relatives
left the household. There were hardly any differences in the proportion of couples
living with relatives of the husband or with relatives of the wife. Around a quarter of all
couples lived with both parents of either the husband or the wife during the second
parity (the interval between first and second birth). Over the life course, the share of
couples living with both parents or with widowed parents of either spouse decreased to
less than 10% and less than 5%, respectively. Co-residence with siblings was as
common as living with parents. Here too, we observe no difference between relatives
of the husband and relatives of the wife in terms of the share of households where these
kin types lived, although co-residence with siblings of the wife occurred slightly more
frequently in the highest parities (from the birth of the fifth child on). Co-residence with
other types of kin (e.g., aunts or uncles of the couple) was less common and occurred in
less than 5% of all cases for the first parity and decreased to around 1% for the highest
parities. In our sample, co-residence with non-kin, such as servants or boarders, did not
occur frequently. Servants were present in around 1 of all households and boarders
were present in only a handful of cases. The share of households with non-kin did not
vary significantly over the life course of the women in our sample.
We control for several community-level characteristics. Based on population size
and the percentage of the population working in agriculture, we indicate whether the
household was located in a rural or urban setting. Owing to the overrepresentation of
Descriptive statistics for variables used in the analysis, by parity
Table 2 (continued)
a Percentages rounded to the nearest tenth, unless otherwise stated
b Anova F test for difference between parity, significance thresholds: p<0.1, * p<0.05, ** p<0.01, ***
Source: Historical Sample of The Netherlands, release 2007
urban areas in the HSN, in particular for the earlier cohorts, more than half of all
households are categorized as urban. Regional differences in social norms and attitudes
toward kinship are captured by a categorical variable indicating the family system in
the region of the household. Family systems are connected to the composition of
households, the strength of kin ties, the inheritance of property, and norms and values
regarding family relations and life course events (Bras et al. 2010a; Hilevych and
Rotering 2013; Kok et al. 2011; Todd 1985). The northwestern coastal provinces,
where partible inheritance was practiced and kin ties were relatively weak compared
with other regions, are coded as nuclear family systems. The southeastern provinces are
coded as stem family systems because of the occurrence of impartible inheritance, or
Anerbenrecht, and the specific customs with regard to co-residence in which young
couples married in and became part of the parental household. The remaining
provinces are coded as intermediate family systems (Bras and Van Tilburg 2007).
At the individual level, we include the birth cohort of the wife to account for the
general trend of fertility decline and increasing lengths of birth intervals over time, as
discussed above. A womans age is one of the main determinants of fecundity and coital
frequency, and thus connected to the duration of parity progression (Van Bavel and Kok
2004). In order to control for age effects, we include the age of the wife and the
duration of marriage at childbirth. Both are expected to be associated with increasing
birth intervals. Larger power differences between husband and wife have been linked to
increased reproductive success (Bereczkei and Csanaky 1996; Voland and Engel 1990),
and therefore we include categories for the age difference between husband and wife,
coded as 0 (husband 0 to 5 years older), 1 (husband more than 5 years older), and 2
(wife older), as a crude proxy for power distance between spouses. We control in all
models for the total number of children born, including deceased children. The
premature death of the previous child may induce a replacement effect and thus may
shorten the time to conception of the next child (Derosas 2006; Knodel 1982; Van
Bavel and Kok 2004, 2010). In addition, since breastfeeding delays the return to
ovulation, a childs survival somewhat decreases a womans chance of becoming
pregnant (Santow 1987). To control for these effects, we control for the death of the
previous child within 8 months after birth or after 8 months since birth.
The HSN includes information on the religion of both husband and wife, which
allows for coding all combinations of religious denominations as categorical dummy
variables. The following categories are discerned, following Van Bavel and Kok (2004,
2005): liberal Protestants, orthodox Protestants, Catholics, mixed, and other. The
first category, liberal Protestants, includes the majority of moderate and liberal schools
in the Dutch Reformed Church and the liberal Protestant churches, such as Mennonites,
Lutherans, and Remonstrants. When both spouses fall under this category, the couple is
classified as liberal Protestant. The second category, orthodox Protestants, contains
couples in which one or both spouses were members of the Calvinist church or
belonged to the orthodox denomination in the Reformed Church. The third category,
Catholic, is composed of couples in which both spouses were members of any Catholic
denomination, such as Roman Catholics, Old Catholics, and Free Catholics. The
fourth category, mixed, comprises couples where one spouse was Catholic and
the other liberal or orthodox Protestant. The last category, other, contains couples
who were Jewish, who belonged to a liberal secessionist denomination, for whom
no religious affiliation was specified, or who had no religion (Bras et al. 2010b;
Bras 2014). Because orthodox Protestant and Catholic denominations were more
stringent in following doctrine and were generally more likely to reject modern
forms of birth control, we presume birth intervals for these groups to have been
shorter compared with more liberal or moderate denominations (Van Bavel and
The social class of the household is based on the occupation of the husband as
registered in the marriage certificate. If the occupation was missing, it was taken from
the population registers. These occupations are coded using the HISCO classification
system, a catalogue of historical occupations that corresponds to the International
Standard Classification of Occupations. The HISCO codes are then categorized
according to the HISCLASS scheme, following Van Leeuwen et al. (2004), into the
following categories: (1) higher managers and professionals, (2) lower managers and
professionals, including clerks and salesman, (3) foremen and skilled laborers, (4)
farmers and fisherman, (5) semi-skilled laborers, (6) unskilled laborers and farm
laborers, and (7) unknown occupation. Previous research has shown that the middle
and upper classes were the first to postpone childbirths, whereas birth intervals among
farmers decreased between 1890 and 1920. Over time, however, the length of birth
intervals converged among all social classes (Bras 2014).
We use an event history approach to examine whether the lengths of womens closed
birth intervals were associated with the presence of kin and non-kin in their households
(Cleves et al. 2010). Event history analysis, also known as survival or duration analysis,
models the effects of covariates on the time until the occurrence of a particular event.
The chance of the event occurring in the next period is expressed as a coefficient that is
dependent on the shape and height of the baseline hazard function. Since the
composition of the household continuously changes as people move in and out, our kin
covariates are time-varying. The ability of survival analysis to accommodate this type
of data makes it a very useful technique. We focus on the effects of kin on the length of
birth intervals and model the effects of kin on the transition from the first living child to
the next birth, from the second living child to the next birth, and so on. We fit Cox
proportional hazard models to examine the effect of the presence of different types of
household members on parity progression risk. The Cox proportional hazard models
take the following form;
where h(t|xj) denotes the hazard rate, or the chance of having a next birth, in period t for
the specified vector of time-varying covariates, h0(t) is a non-negative and unspecified
baseline hazard function that varies arbitrarily over time and is not dependent on the
covariates Xj, and x is a vector of unknown regression coefficients to be estimated
from the data using maximum likelihood (Cleves et al. 2010). Given that we consider
the effect of kin on time until next birth using a sample of closed birth intervals, the
cumulative hazard rate increases over time until it is equal to one and all women have
given birth. The Cox proportional hazards model allows for estimating the relative
hazard rate of women in different groupsfor example, living with or without
particular types of kin. Both groups have the same baseline hazard at time t, but the
magnitude of the hazard is multiplied by the exponentiated regression coefficient of
each group. Time is measured in months, and a coefficient larger than zero denotes a
higher chance of giving birth in period t, or in other words, a shorter birth interval.
The Cox model assumes that the estimated hazard ratios are proportional to each
other. The assumption of proportionality was tested by examining the Schoenfeld
residuals and the proportionality of the log-log plot of the survival rate against the
log of time for each variable. These pre-analysis tests indicated that the baseline hazard
should be allowed to vary between religions for all models, which ensures that we can
still provide reliable estimates for the effects of kin on the hazard rate of childbirth in
the period t, although the effect of religion is left unspecified. For parity six and up, we
estimate one model for all births after the fifth birth and control for net parity, which is
the total number of living children at time t. In order to ensure proportionality of
hazards in this model, we included additional time-varying effects for marriage duration
and the death of the previous infant if they survived until they were at least 8 months of
Another important assumption of the Cox model is that the risk of parity progression
in the sample is randomly distributed across observations. However, since birth
intervals are by their nature clustered on the level of the couple, it is likely that parity
progression risks are not completely random and the hazard ratio may thus be
conditional on the individual frailty of each couple (Cleves et al. 2010). When specifying
multiple parities or when a couple has experienced the loss of a previous child, and thus
would be observed two or more times in the analysis of a particular parity progression
rate, this may be a cause for concern. Testing for the significance of the estimated frailty
variance revealed that it was not necessary to include an individual frailty component
for each couple. Nevertheless, a robust estimation of variance is recommended given
the clustered nature of our observations (Lin and Wei 1989). Goodness of fit was
evaluated by examining the Cox-Snell residuals against the Nelson-Aalen cumulative
hazard function (Cleves et al. 2010; Cox and Snell 1968). Our analysis is robust for
variations in sample size across both geographical areas and time.
Table 3 provides parameter estimates for the multivariate Cox proportional hazards
model by parity. Column 1 contains the results for the time until the birth of a next
child, since the birth of the previous child, for all couples who had one living child.
Column 2 provides estimates for the length of the birth interval for all couples who had
two living children, and so on. The births of children after the fifth living child (parity
six and up) are grouped in one Cox proportional hazards model in which the number of
living children is controlled for. All models include control variables for unreported
household and community-level characteristics.
Our results indicate that for later-order births, from parity six and up, co-residence
with kin was not significantly associated with longer or shorter birth intervals. For
lower parities, living with both parents of either the husband or the wife did not affect
the time until next birth, although living with the husbands parents had a small
delaying effect on the birth of the fourth child. In contrast, living with a widowed
father of either spouse significantly reduced parity transition rates. The birth intervals of
women living with a widowed father were about twice as long as birth intervals of
women who did not live with a widowed father.2 Conversely, living with a widowed
mother in the household did not significantly affect time until next birth, with the
exception of the fourth parity in the case of a husbands widowed mother.
The length of time between births was at least 20% shorter for women who lived
with at least one brother or brother-in-law, compared with women who did not live with
a brother or brother-in-law in the same household. The positive effect of the presence of
the husbands brother was particularly high for women experiencing the transition from
the third to the fourth child, for whom the parity transition rate was almost twice as
large as for other parities. Strikingly, in contrast to living with brothers, living with
sisters of either of the spouses did not significantly affect parity transition rates.
Other types of co-resident kin, such as aunts or uncles, had no systematic effect on
the duration of the transition to the next birth. The presence of non-kin did affect birth
intervals, but their effect was not stable over the reproductive life course. The presence
of servants affected only the waiting time until the birth of the third child and the birth
intervals of fifth and later-order children. Co-resident boarders and lodgers had a
significant delaying effect on the birth interval for the third parity, but shortened the
2 The exponentiated coefficients represent the chance, or hazard, of a birth occurring in period t relative to the
reference category. Here e0.7 implies a hazard rate of 0.5, or in other words, a doubling of the length of the
Table 3 Estimated coefficients for the effects of kin presence in the household on the likelihood of second or
later-order marital births by parity
Cox proportional hazards models, coefficients reported. Cluster robust standard errors, adjusted for
dependence among births of the same couple, stratified on religion, Breslow approximation for tied survival times. A
positive sign indicates a shorter birth interval for the associated covariate. All models control for occupation,
religion, urbanization, community characteristics, age, and spousal age gap (coefficients for those variables are
available from authors upon request)
Source: Historical Sample of The Netherlands, release 2007
p<0.1, * p<0.05, ** p<0.01, *** p<0.001, not present
a Reference: none present
b Square root of marriage duration after 18 years
c Linear interaction with analysis time after 24 months
d Reference: 18711880
birth interval for the fourth and fifth parity. Keeping in mind the low number of
observations of non-kin in our sample, these findings should be interpreted with care.3
The time until next birth was longer for couples who were married for a longer
period of time, as was expected. However, for women who had at least six children, we
observe that parity transition rates increased after 18 years of marriage compared with
women who were married for less than 18 years. Crude parity indicates the total
number of children that have been born to a woman. As would be expected, women
who had experienced a larger number of pregnancies had shorter intervals between
births. Similar to findings by Van Bavel and Kok (2004, 2010), parity transition rates of
women whose last child died before it was 8 months old were significantly higher than
those of women whose last child survived. The death of the previous infant at 8 months
of age or later delayed the birth of a next child, although after 24 months the chance of
having a next birth was significantly higher. In line with the observations presented in
Table 1, birth intervals were longer for women born in later cohorts compared with
those born in the reference period 18711880.
In human behavior, as in the behavior of other animals, the provision of support to
genetically related kin is expected to confer an indirect advantage in terms of inclusive
fitness (Hamilton 1964a, b; Hrdy 2008, 2009; Gurven et al. 2001). It follows that
couples who can rely on close kin members for support are more likely to raise more or
better-quality children than couples who are not surrounded by close kin (Kramer 2010;
Salmon and Shackelford 2008). In this study we have investigated whether the
presence of co-resident kin and non-kin affected the length of birth intervals for 2628
Dutch women born between 1842 and 1920. Our point of departure was the broad
hypothesis that, on the basis of inclusive fitness theory, all close kin members would
exert a positive influence on reproductive outcomes by enabling the women in our
sample to have shorter intervals between births. The effects of kin on fertility were
expected to be positively associated with the strength of the genetic bond between kin,
whereas genetically more distant kin would have a minor effect on reproduction.
Furthermore, kin influences were hypothesized to be stronger during the first, formative
years of the family when alloparental support could compensate for the inexperience
among new parents learning about the nutritional needs of newborns or for the extra
work needed to feed an additional mouth.
Using continuous-time data on household composition as a proxy for cooperative
behavior, we find that co-resident kin had different effects on fertility at different stages
of the reproductive life span of women. The effects of kin were not significant for
higher parities. This finding may provide a partial explanation for the variations in kin
effects on fertility that have been observed in the literature in which only measures of
complete fertility outcomes are taken into account (see Sear and Coall 2011 for an
extensive overview of the literature). However, without knowledge of the distribution
of resources and care among household members over time, it is difficult to infer from
our data why the effects of kin were only significant in the early parities. As Lyngstad
3 Our findings concerning the influence of kin are the same when non-kin are excluded from our analyses.
and Prskawetz (2010) argue in their study of Swedish sibling pairs born in the
midtwentieth century, the decrease of kin influences might be attributable to uncertainties
around the process of entering parenthood, but possibly also to changes in both the
different roles of kin within the household and their ability to provide the couple with
any form of support.
Our findings suggest that co-resident kin did not affect reproductive outcomes of
Dutch women in a uniform way. In contrast to other empirical findings, we find that
parity progression rates were not significantly affected by the presence of widowed
mothers or both parents of either spouse (Hawkes et al. 1998; Hawkes 2003; Pollet
et al. 2007; Voland and Beise 2002). The absence of a positive grandmother effect
regarding the length of birth intervals is also observed in other studies (e.g., Hill and
Hurtado 2009). However, Dutch women who lived with their widowed father or the
widowed father of their husband experienced significantly longer birth intervals than
women living without a widowed father. Whereas some studies have shown no effect
of fathers on fitness outcomes (Borgerhoff Mulder 2007; Sear and Coall 2011), in
others a negative effect of fathers on their daughters reproductive behavior has been
observed, in particular in relation to offspring survival chances (Kemkes-Grottenthaler
2005). Our findings concerning the delaying effect of co-resident widowed fathers on
parity progression provide support for the confidence of paternity hypothesis, which
suggests that uncertainty over genetic relatedness will lower the extent of cooperative
behavior to offspring (Strassmann and Garrard 2011). However, on its own this
hypothesis has received little empirical support in explaining differences in the
influence of parents on demographic outcomes (Euler and Weitzel 1996; Pashos and
McBurney 2008). The delaying effect of widowed fathers might also be attributable
to the notion that fathers consumed a relatively large share of the couples resources,
especially care, for themselves while providing the couple with little support or few
pronatal incentives (see also Kemkes-Grottenthaler 2005). Recent empirical studies
show that kin effects are indeed modified by conflicts over resources (Borgerhoff
Mulder 2007; Schaffnit and Sear 2014). With regard to siblings, our findings indicate
that the presence of brothers, but not the presence of sisters, was positively associated
with parity progression. This observation suggests that the additional resources that
brothers brought into the household had enabling effects on couples reproductive
outcomes (Becker 1981; Becker and Barro 1988; Feng et al. 2010).
Our study contributes to the growing literature on empirical approaches to
evolutionary theories of demographic behavior. The findings presented here raise
further questions concerning the role and position of kin members within the
household as well as the extent of their cooperative behavior, such as provisioning
of care or contributions to household income, which is difficult to infer from kin
presence alone (see, e.g., Schaffnit and Sear 2014). In addition, owing to the
nature of the HSN data, our observations are limited to household members, but
kin living outside the household may also have affected the womens reproductive
careers (Johow and Voland 2012). These issues further complicate the connection
between the assumptions on which our hypotheses are based and our findings.
Although people may receive fitness benefits from higher reproductive outcomes
of their kin, shorter birth intervals are not by definition in the womans interest
and in fact may lower the quality of offspring (Borgerhoff Mulder 2000;
CondeAgudelo et al. 2006, 2007; Palloni and Millman 1986). Nonetheless, our findings
do clearly indicate that the presence of widowed fathers and brothers affected
parity progression rates, leading to the conclusion that reproductive outcomes
were subject to the distribution of resources and care within the household.
Inclusive fitness theory enables us to understand the motives underlying the
behavior of household members toward genetically related others, but actual
demographic outcomes are determined by the specific historical, social,
economic, and spatial conditions of the household, as well as maternal health and the
extent of cooperative behavior of kin that enables women to give birth. Future
research on the interaction between wealth and kin influence, or differences in
cultural norms concerning kinship and reproduction, which lie beyond the scope
of this study, may further illuminate variations in the influence of kin on
Acknowledgments We thank the participants of the session on Family Systems, Family Relations and
Fertility at the European Social Science History Conference 2014 for their valuable comments and insights.
We also wish to thank the International Institute for Social History (IISH) for providing access to the Historical
Sample of the Netherlands and Anglique Janssens for reading an earlier version of the paper. This study was
supported by a VIDI Innovational Research Grant to Hilde Bras from the Netherlands Organization for
Scientific Research (NWO), for the research project entitled The Power of the Family: Family Influences on
Long-Term Fertility Decline in Europe, 18502010 (contract grant number 452-10-013).
Open Access This article is distributed under the terms of the Creative Commons Attribution License which
permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source
Paul Rotering received a BA in history and an MSc in economics from Radboud University, the Netherlands,
and is currently a PhD candidate in historical demography at the Department of Social Sciences at Wageningen
University, the Netherlands. His research focuses on the role of the family in explaining long-term differences
in fertility decline. His interests include novel methods for quantitative analysis of rich and large-scale
Alter , G. ( 1988 ). Family and the female life course: the women of Verviers , Belgium, 1849 - 1880 . Madison: University of Wisconsin Press.
Becker , G. S. ( 1981 ). A treatise on the family . Cambridge: Harvard University Press.
Becker , G. S. , & Barro , R. J. ( 1988 ). A reformulation of the economic theory of fertility . The Quarterly Journal of Economics , 103 ( 1 ), 1 - 25 .
Bereczkei , T. ( 1998 ). Kinship network, direct childcare, and fertility among Hungarians and Gypsies . Evolution and Human Behavior , 19 ( 5 ), 283 - 298 .
Bereczkei , T. , & Csanaky , A. ( 1996 ). Mate choice, marital success, and reproduction in a modern society . Ethology and Sociobiology , 17 ( 1 ), 17 - 35 .
Boonstra , O.W.A. ( 2007 ). NLGIS shapefiles . Data Archiving and Networked Services (DANS) (urn:nbn:nl:ui: 13-wsh-wv7).
Borgerhoff Mulder , M. ( 2000 ). Optimizing offspring: the quantity-quality tradeoff in agropastoral Kipsigis . Evolution and Human Behavior , 21 ( 6 ), 391 - 410 .
Borgerhoff Mulder , M. ( 2007 ). Hamilton's rule and kin competition: the Kipsigis case . Evolution and Human Behavior , 28 ( 5 ), 299 - 312 .
Bras , H. ( 2014 ). Structural and diffusion effects in the Dutch fertility transition , 1870 - 1940 . Demographic Research, 30 ( 5 ), 151 - 186 .
Bras , H. , Kok , J. , & Mandemakers , K. ( 2010a ). Sibship size and status attainment across contexts: evidence from the Netherlands , 1840 - 1925 . Demographic Research, 10 ( 4 ), 73 - 104 .
Bras , H. , Liefbroer , A. , & Elzinga , C. ( 2010b ). Standardization of pathways to adulthood? An analysis of Dutch cohorts born between 1850 and 1900 . Demography, 47 ( 4 ), 1013 - 1034 .
Bras , H. , & Van Tilburg , T. ( 2007 ). Kinship and social networks: a regional analysis of sibling relationships in twentieth-century Netherlands . Journal of Family History , 32 ( 3 ), 296 - 322 .
Cleves , M. , Gould , W. , Gutierrez , R. , & Marchenko , Y. ( 2010 ). An introduction to survival analysis using Stata (3rd ed.). College Station , TX: Stata Press.
Conde-Agudelo , A. , Rosas-Bermdez , A. , & Kafury-Goeta , A. ( 2006 ). Birth spacing and risk of adverse perinatal outcomes: a meta-analysis . Journal of the American Medical Association , 295 ( 15 ), 1809 - 1823 .
Conde-Agudelo , A. , Rosas-Bermdez , A. , & Kafury-Goeta , A. ( 2007 ). Effects of birth spacing on maternal health: a systematic review . American Journal of Obstetrics and Gynecology , 196 ( 4 ), 297 - 308 .
Cox , D. R. , & Snell , E. J. ( 1968 ). A general definition of residuals . Journal of the Royal Statistical Society , Series B, 30 ( 2 ), 248 - 275 .
Crognier , E. , Baali , A. , & Hilali , M. ( 2001 ). Do helpers at the nest increase their parents' reproductive success ? American Journal of Human Biology , 13 , 365 - 373 .
Derosas , R. ( 2006 ). Between identity and assimilation: Jewish fertility in nineteenth-century Venice . In R. Derosas & F. Van Poppel (Eds.), Religion and the Decline of Fertility in the Western World (pp . 177 - 206 ). Dordrecht: Springer.
Draper , P. , & Hames , R. ( 2000 ). Birth order, sibling investment, and fertility among Ju/'hoansi (!Kung) . Human Nature , 11 ( 2 ), 117 - 156 .
Euler , H. , & Weitzel , B. ( 1996 ). Discriminative grandparental solicitude as reproductive strategy . Human Nature , 7 ( 1 ), 39 - 59 .
Feng , W. , Lee , J. Z. , Tsuya , N. O. , & Kurosu , S. ( 2010 ). Household organization, co-resident Kin, and reproduction . In N. O. Tsuya , W. Feng , G. Alter , & J. Z. Lee (Eds.), Prudence and pressure: reproduction and human agency in Europe and Asia, 1700 - 1900 (pp. 67 - 96 ). Cambridge: MIT Press.
Grafen , A. ( 1984 ). Natural selection, kin selection and group selection . In J. R. Krebs & N. B. Davies (Eds.), Behavioural ecology (pp. 62 - 84 ). Oxford: Blackwell.
Gurven , M. , Allen-Arave , W. , Hill , K. , & Hurtado , A. M. ( 2001 ). Reservation food sharing among the Ache of Paraguay . Human Nature , 12 ( 4 ), 273 - 297 .
Hames , R. , & Draper , P. ( 2004 ). Women's work, child care, and helpers-at-the-nest in a hunter-gatherer society . Human Nature , 15 ( 4 ), 319 - 341 .
Hamilton , W. D. ( 1964a ). The genetical evolution of social behaviour . I. Journal of Theoretical Biology , 7 ( 1 ), 1 - 16 .
Hamilton , W. D. ( 1964b ). The genetical evolution of social behaviour . II. Journal of Theoretical Biology , 7 ( 1 ), 17 - 52 .
Hawkes , K. ( 2003 ). Grandmothers and the evolution of human longevity . American Journal of Human Biology , 15 ( 3 ), 380 - 400 .
Hawkes , K. , O'Connell , J. F. , & Blurton Jones , N. G. ( 1997 ). Hadza women's time allocation, offspring provisioning, and the evolution of long postmenopausal life spans . Current Anthropology , 38 ( 4 ), 551 - 577 .
Hawkes , K. , O'Connell , J. F. , Jones , N. G. B. , Alvarez , H. , & Charnov , E. L. ( 1998 ). Grandmothering, menopause, and the evolution of human life histories . Proceedings of the National Academy of Sciences , 95 ( 3 ), 1336 - 1339 .
Heckman , J. J. , & Walker , J. R. ( 1990 ). The relationship between wages and income and the timing and spacing of births: evidence from Swedish longitudinal data . Econometrica , 58 ( 6 ), 1411 - 1441 .
Hilevych , Y. , & Rotering , P. P. P. ( 2013 ). Moederschap en sociale netwerken in Oekrane , 1955 - 1965 . In K. Matthijs, P. Puschmann , & H. Bras (Eds.), Gender in/en demografie, Jaarboek demografie 2013 (pp. 209 - 231 ). Leuven/Den Haag: Acco.
Hill , K. , & Hurtado , A. M. ( 2009 ). Cooperative breeding in South American hunter-gatherers . Proceedings of the Royal Society B: Biological Sciences , 276 ( 1674 ), 3863 - 3870 .
Hotz , V. J. , Klerman , J. A. , & Willis , R. J. ( 1997 ). The economics of fertility in developed countries . In M. R. Rosenzweig & O. Stark (Eds.), Handbook of population and family economics (pp. 275 - 347 ). Amsterdam: Elsevier.
Hrdy , S. ( 1999 ). Mother Nature: a history of mothers, infants, and natural selection . New York : Pantheon Books.
Hrdy , S. ( 2005 ). Evolutionary context of human development: the cooperative breeding model . In C. S. Carter, L. Ahnert , K. E. Grossmann , S. Hrdy , M. E. Lamb , S. W. Porges , et al. (Eds.), Attachment and bonding: a new synthesis (pp. 9 - 32 ). Cambridge: MIT Press.
Hrdy , S. ( 2008 ). Evolutionary context of human development: The cooperative breeding model . In C. Salmon & T. Shackelford (Eds.), Family relationships: An evolutionary perspective (pp. 39 - 68 ). Oxford: Oxford University Press.
Hrdy , S. ( 2009 ). Mothers and others: The evolutionary origins of mutual understanding . Cambridge: Harvard University Press.
Janssens , A. ( 1993 ). Family and social change: The household as a process in an industrializing community . Cambridge : Cambridge University Press.
Johow , J. , & Voland , E. ( 2012 ). Conditional grandmother effects on age at marriage, age at first birth, and completed fertility of daughters and daughters-in-law in historical Krummhrn . Human Nature , 23 ( 3 ), 341 - 359 .
Kaptijn , R. , Thomese , F. , Tilburg , T. , & Liefbroer , A. ( 2010 ). How grandparents matter . Human Nature , 21 ( 4 ), 393 - 405 .
Kemkes-Grottenthaler , A. ( 2005 ). Of grandmothers, grandfathers and wicked step-grandparents . Differential impact of paternal grandparents on grandoffspring survival . Historical Social Research/Historische Sozialforschung , 219 - 239 .
Knodel , J. ( 1982 ). Child mortality and reproductive behaviour in German village populations in the past: a micro-level analysis of the replacement effect . Population Studies , 36 ( 2 ), 177 - 200 .
Knotter , A. , & Meijer , A. C. ( 1995 ). De gemeentelijke bevolkingsregisters , 1850 - 1920 . Den Haag: Instituut voor Nederlandse Geschiedenis .
Kok , J. , & Mandemakers , K. ( 2009 ). Je zoudt maar last van mij hebben . Verwanten in het Nederlands huishouden , 1860 - 1940 . Tijdschrift voor Sociale en Economische Geschiedenis , 6 ( 4 ), 139 - 165 .
Kok , J. , & Mandemakers , K. ( 2010 ). A life-course approach to co-residence in the Netherlands , 1850 - 1940 . Continuity and Change , 25 ( 02 ), 285 - 312 .
Kok , J. , Vandezande , M. , & Mandemakers , K. ( 2011 ). Household structure, resource allocation and child wellbeing. A comparison across family systems . Tijdschrift voor Sociale en Economische Geschiedenis , 8 ( 4 ), 76 - 101 .
Kramer , K. L. ( 2005 ). Children's help and the pace of reproduction: cooperative breeding in humans . Evolutionary Anthropology: Issues, News, and Reviews , 14 ( 6 ), 224 - 237 .
Kramer , K. L. ( 2010 ). Cooperative breeding and its significance to the demographic success of humans . Annual Review of Anthropology , 39 ( 1 ), 417 - 436 .
Liefbroer , A. , & De Jong Gierveld , J. ( 1995 ). Standardization and individualization: The transition from youth to adulthood among cohorts born between 1903 and 1965 . In H. Brekel & F. Deven (Eds.), Population and Family in the Low Countries 1994 . Dordrecht: Springer Netherlands.
Lin , D. Y. , & Wei , L. J. ( 1989 ). The robust inference for the cox proportional hazards model . Journal of the American Statistical Association , 84 ( 408 ), 1074 - 1078 .
Lyngstad , T. , & Prskawetz , A. ( 2010 ). Do siblings' fertility decisions influence each other ? Demography , 47 ( 4 ), 923 - 934 .
Mace , R. ( 2014 ). When not to have another baby: an evolutionary approach to low fertility . Demographic Research , 30 ( 37 ), 1074 - 1096 .
Mandemakers , K. ( 2002 ). Building life course datasets from population registers by the Historical Sample of the Netherlands (HSN) . History and Computing , 14 ( 1-2 ), 87 - 107 .
Mandemakers , K. ( 2004 ). De Historische Steekproef Nederlandse bevolking (HSN) en het project Life Courses in Context . Bevolking en gezin, 33 , 91 - 114 .
Morgan , S. P. , & King , R. ( 2001 ). Why have children in the 21st century? Biological predisposition, social coercion, rational choice . European Journal of Population , 17 ( 1 ), 3 - 20 .
Morgan , S. P. , & Rindfuss , R. R. ( 1984 ). Household structure and the tempo of family formation in comparative perspective . Population Studies , 38 ( 1 ), 129 - 139 .
Nath , D. C. , Leonetti , D. L. , & Steele , M. S. ( 2000 ). Analysis of birth intervals in a non-contracepting Indian population: an evolutionary ecological approach . Journal of Biosocial Science , 32 ( 03 ), 343 - 354 .
Palloni , A. , & Millman , S. ( 1986 ). Effects of inter-birth intervals and breastfeeding on infant and early childhood mortality . Population Studies , 40 ( 2 ), 215 - 236 .
Pashos , A. , & McBurney , D. ( 2008 ). Kin relationships and the caregiving biases of grandparents , aunts, and uncles. Human Nature , 19 ( 3 ), 311 - 330 .
Pebley , A. R ., & Stupp , P. W. ( 1987 ). Reproductive patterns and child mortality in Guatemala . Demography, 24 ( 1 ), 43 - 60 .
Pollet , T. , Nettle , D. , & Nelissen , M. ( 2007 ). Maternal grandmothers do go the extra mile: factoring distance and lineage into differential contact with grandchildren . Evolutionary Psychology , 5 ( 4 ), 832 - 843 .
Salmon , C. , & Shackelford , T. (Eds.). ( 2008 ). Family relationships, an evolutionary perspective . Oxford: Oxford University Press.
Santow , G. ( 1987 ). Reassessing the contraceptive effect of breastfeeding . Population Studies , 41 ( 1 ), 147 - 160 .
Santow , G. ( 1995 ). Coitus interruptus and the control of natural fertility . Population Studies , 49 ( 1 ), 19 - 43 .
Schaffnit , S. B. , & Sear , R. ( 2014 ). Wealth modifies relationships between kin and women's fertility in highincome countries . Behavioral Ecology , 25 ( 4 ), 834 - 842 .
Sear , R. , & Coall , D. ( 2011 ). How much does family matter? Cooperative breeding and the demographic transition . Population and Development Review , 37 , 81 - 112 .
Sear , R. , & Mace , R. ( 2008 ). Who keeps children alive? A review of the effects of kin on child survival . Evolution and Human Behavior , 29 ( 1 ), 1 - 18 .
Sear , R. , Mace , R. , & McGregor , I. A. ( 2003 ). The effects of kin on female fertility in rural Gambia . Evolution and Human Behavior , 24 ( 1 ), 25 - 42 .
Skinner , G. W. ( 1997 ). Family systems and demographic processes . In D. I. Kertzer & T. Fricke (Eds.), Anthropological demography (pp. 53 - 95 ). Chicago and London: The University of Chicago Press.
Snopkowski , K. , & Sear , R. ( 2013 ). Kin influences on fertility in Thailand: effects and mechanisms . Evolution and Human Behavior , 34 ( 2 ), 130 - 138 .
Strassmann , B. , & Garrard , W. ( 2011 ). Alternatives to the Grandmother Hypothesis . Human Nature , 22 ( 1-2 ), 201 - 222 .
Todd , E. ( 1985 ). The explanation of an ideology . Oxford: Basil Blackwell.
Tsay , W.-J. , & Chu , C. Y. C. ( 2005 ). The pattern of birth spacing during Taiwan's demographic transition . Journal of Population Economics , 18 ( 2 ), 323 - 336 .
Tymicki , K. ( 2004 ). Kin influence on female reproductive behaviour: the evidence from reconstitution of the Bejsce parish registers, 18th to 20th centuries, Poland . American Journal of Human Biology , 16 , 508 - 522 .
van Bavel , J. ( 2004 ). Diffusion effects in the European fertility transition: historical evidence from within a Belgian town (1846-1910) . European Journal of Population/Revue Europenne de Dmographie , 20 ( 1 ), 63 - 85 .
Van Bavel , J. , & Kok , J. ( 2004 ). Birth spacing in the Netherlands . The effects of family composition, occupation and religion on birth intervals , 1820 - 1885 . European Journal of Population/Revue Europenne de Dmographie , 20 ( 2 ), 119 - 140 .
Van Bavel , J. , & Kok , J. ( 2005 ). The role of religion in the Dutch fertility transition: starting, spacing and stopping in the heart of the Netherlands , 1845 - 1945 . Continuity and Change , 20 ( 2 ), 247 - 263 .
Van Bavel , J. , & Kok , J. ( 2010 ). A mixed effects model of birth spacing for pre-transition populations: evidence of deliberate fertility control from nineteenth century Netherlands . The History of the Family , 15 ( 2 ), 125 - 138 .
Van der Woude , A. ( 1970 ). De omvang en samenstelling van de huishouding in Nederland in het verleden . A.A.G. Bijdragen, ( 15 ), 202 - 241 .
Van Leeuwen , M. H. , Maas , I. , & Miles , A. ( 2004 ). Creating a historical international standard classification of occupations: an exercise in multinational interdisciplinary cooperation . Historical Methods: A Journal of Quantitative and Interdisciplinary History , 37 ( 4 ), 186 - 197 .
Van Poppel , F. , Reher , D. , Sanz-Gimeno , A. , Sanchez-Dominguez , M. , & Beekink , E. ( 2012 ). Mortality decline and reproductive change during the Dutch demographic transition: revisiting a traditional debate with new data . Demographic Research , 27 ( 11 ), 299 - 338 .
Voland , E. , & Engel , C. ( 1990 ). Female choice in humans: a conditional mate selection strategy of the Krummhrn women (Germany , 1720 - 1874 ). Ethology, 84 ( 2 ), 144 - 154 .
Voland , E. , & Beise , J. ( 2002 ). Opposite effects of maternal and paternal grandmothers on infant survival in historical Krummhrn . Behavioral Ecology and Sociobiology , 52 ( 6 ), 435 - 443 .
Vulsma , R. F. ( 1988 ). Burgerlijke stand en bevolkingsregister . Den Haag: Centraal Bureau voor Genealogie.