275
1. Introduction
The international and Czech literature examining
the causes of low fertility has identified a number of
factors as lying behind this demographic phenomenon.
Until recently, these were usually extracted
from data generated by research on female fertility.
Men themselves were rarely the subject of study, or
even mentioned, and, if they were, it was usually
in connection with male contraception and family
planning – and very often in the context of the less
developed countries.
This situation began to change in the second
half of the 1990s, since which time the issues of
male reproductive behaviour and male fertility
have gradually, and rightly, been coming into focus
of international experts in the area of population
studies – see e.g. Goldschneider – Kaufman, 1996;
Sonenstein et al., 1997; Green – Biddlecom, 2000;
Bledsoe – Lerner – Guyer, 2000; Toulemon, 2001;
Tölke – Diewald, 2003; Rotkirch – Basten, 2010; and
Zhang, 2011. The reason behind this is obvious:
reproduction and fertility result from behavioural
Beatrice Chromková Manea1)
– Ladislav Rabušic
Abstract
The present article tackles new empirical evidence in the study of male reproductive behaviour in the Czech
Republic. It builds on a previously published theoretical and methodological overview on the chosen topic.
We will try to answer the following questions: (1) What is the completed male fertility level and what preferences
do men have regarding family size? (2) What are the determinants of male fertility? For this purpose, we
use data from an ad-hoc survey conducted in 2011 on a representative sample of men aged 40–55. The results
are presented in a comparative perspective – male/female indicators. The article includes a description of the
main demographic indicators of fertility with a subchapter on the measurement of fertility ideals and preferences.
Further on, we present some of the covariates which determine the final level of male fertility and the
ideal number of children at both individual and societal levels. Some of the findings point to the lack of data
from official statistics necessary to calculate the main indicators of male fertility. However, these deficiencies
are partially removed by the results obtained from the analysis run on survey data.
Keywords: male fertility behaviour, completed fertility rates, ideal number
of children, fertility determinants Demografie, 2013, 55: 275–290
*)	The research reported in this article was supported by the Czech Science Foundation (GACR) under grant no. P404/11/0329
Male Reproductive Behaviour in the Czech Republic.
†)	The authors would like to thank two anonymous reviewers for bringing to their attention certain deficiencies in the text and
for the precious suggestions.
1)	Corresponding Author: Beatrice Chromková Manea, Faculty of Social Studies MU Brno, Office for Population Studies,
Joštova 10, Brno 602 00, e-mail: manea@fss.muni.cz
Male fertility in the Czech
Republic – first empirical
evidence*) †)
ArticleS2013 55 (4)
276
interaction between a man and a woman and researchers
seeking to comprehend this interaction
and to be able to model (and possibly predict) reproduction
should not ignore one of the elements
of this interaction.
This research initiative has also found reflection
in the Czech environment. Thanks to an overview
paper by Rabušic and Chromková Manea (2011) –
the first of its kind in the Czech Republic – certain
contextual attributes of male fertility studies have been
introduced into the discourse of Czech population
studies. However, what needs yet to be addressed are
the reasons why male fertility has hitherto rarely been
discussed in analyses of reproduction, and a clarification
of which indicators of male reproduction are
available and which, on the contrary, are not.
The aim of this article is to build on the forementioned
paper and to present the first empirical
findings about male fertility in the Czech Republic.
We concern ourselves with the following research
questions: (1) What is the completed male
fertility level and what preferences do men have regarding
family size? (2) What are the determinants
of male fertility? Where data are available and where
it makes sense, a comparison will be drawn between
men and women.
2. Method
If we want to gain insight into male fertility, we
should compare its patterns with those of female
fertility. The problem here is that we cannot generate
the same indicators for men as we can for women
from Czech statistical data. Births reports, which
form the basis for fertility statistics, do not provide
full information on all fathers – they are limited
to data on fathers of children born within marriage,
with data on men whose child was conceived
and born out of wedlock often missing.2)
Data are
therefore not available on the total number of children
born to men or on their age distribution. These
gaps in statistical recording make it impossible, for
example, to compare male and female age-specific
fertility rates or to compare completed fertility
and out-of-wedlock fertility rates in the male
and female populations.3)
Where it is not possible
to rely on official statistics, relevant data must
be obtained from quantitative surveys.
2)	Part of the children conceived before marriage (out of wedlock) is legitimised later by marriage (the so-called premarital conception).
However, Zeman (2007: 20) points out that during the communist period most of births out of wedlock were firstorder
births, while nowadays a considerable portion of births outside marriage are second-order births, which may indicate
the existence of functional cohabiting relationships replacing marriage, and not only premarital cohabitation with mostly firstorder
births, which were later legitimised by marriage.
3)	 Since 2007, the situation has somewhat improved as to the availability of data on fathers. Records concerning children
	 of single, divorced and widowed women have been expanded to also include analysis of the data on fathers. The problem,
	 though, is that these data on fathers is submitted, on a voluntary basis, by the mothers of the children. From the statistical
	 point of view, such data are likely to be incomplete. Analysis of the 2011 data shows that: ‘The highest proportion of births
	 with no father information provided was among third-born children and children born of higher birth order. In 2011, these
	 children accounted for 12.9% of the total number of births, and for 34.8% of children born out of wedlock. The lowest
	 proportion of children with no father information provided out of all live births of the given birth order was among
	 second-born children – 6.0% in 2011. In the group of children born out of wedlock, the relatively lowest proportion
	 of children with missing father data was found among first-born children (20.0% in 2011). The proportion of children with
	 no information on the father declines with increasing education of the mother. These children accounted for 37.3% of all
	 children born to women with primary school education, but only 2.5% of children born to university-educated mothers.
	 The proportion of children with no father information given also declines with the increasing age of the mother. In the group
	 of children born to women below the age of 20 father information was missing in the case of almost every other child in 2011
	 (49.1%), compared to only 4.7% in the group of children born to women aged 30–34 and 7.2% in the group of children whose
	 mother was 35 years old or older at the time of the birth (Vývoj, 2011: 20–21)’.
277
We will seek answers to our questions by drawing
on data from a representative quantitative survey
for the male population aged 40–55 (see Table
5 in appendix for the sample characteristics).4)
The survey was conducted in the second half
of the year 2011. Using a random sample method
and a standardised questionnaire, interviewers
from the Median agency, which carried out
the data collection, interviewed men between 40
and 55 years of age, which is the age relevant to our
research interests. Men in this age bracket represented
the sample unit. If the man lived in a household
with a female partner, whether married or
cohabiting, the woman was also interviewed (using
a specific questionnaire for women). In this way,
data was obtained from 800 couples: as mentioned
above, men were aged 40–55 years; the age range
of their partners was broader. If the man lived alone,
data was collected from him only – there were 450
such cases. If random sampling brought the interviewer
into a household of a woman living without
a male partner and aged 40–55 years, the woman
was interviewed based on the questionnaire version
for women. The number of these female respondents
was restricted to 450. In total, information
was obtained from 2,500 respondents (1,600 men
and women in couples, 450 men without a spouse/
partner and 450 women without a spouse/partner).
For the purposes of this paper, we will largely work
with the subsample of men and women aged 40+.
3. Results
3.1. Completed fertility
Let us now examine male reproductive behaviour
on the basis of the data collected in the special sociological
survey referred to above, titled ‘Male Reproductive
Behaviour’ and conducted at the end
of 2011. These data provide insight into the completed
male fertility level, and moreover, into the attitudes
of men concerning the number of children, conception
intentions and other characteristics that affect fertility.
A standard indicator of fertility levels is the completed
fertility rate. In Czech demographic statistics,
the completed fertility rate is not calculated for the
male population, and thus is also not published. Our
survey data make this calculation possible, as it might
be assumed that our male respondents, aged 40–55
years, have for the most part completed their reproduction
and are rarely likely to become fathers at this age.
The number of children reported by the male respondents
in the survey could be interpreted as their
completed number of children.5)
However, as the male
reproductive cycle is not limited by a fixed age ceiling,
and given the age range of the male respondents’
partners (see the paragraph below), a supplementary
question was included in the questionnaire asking
whether the respondent was still considering having
a child in the future and how many. We could then
calculate the hypothetical completed fertility rate
as the total number of children the respondent already
had (and had had) and the number of children
he would still like to have.
The age distribution of the female part of our sample
used for comparison with males differs from that
of their male counterparts and ranges between 20
and 65 years. As a quarter of the females in the sample
were aged 20–40 years, it could be assumed that their
reproductive cycle was not finished yet. We therefore
performed calculations of the completed fertility level
(in the case of women aged over 40 years), as well
as the hypothetical completed fertility level of women.
Figure 1 presents the structure of male or female
respondents aged 40 and over according to the number
of children they have. It shows that childlessness
was more common among men than among women.6)
The proportion of men and women who had one
child was the same (17%). Two was the most frequent
4)	We required the sample to be representative for the Czech male population aged 40–55. Stratified random sampling was carried
out in several steps by Median agency. The primary sampling units were determined on the basis of representativeness
of the size of residence and county. Sampling frame was based on all address in the Czech Republic, where strata were made
up of all geographic territorial units (size of residence and county).
5)	All children reported by the male respondents, including adopted children, stepchildren and deceased children were considered
in our analyses.
6)	The difference is statistically significant at the 0.01 level (according to a t-test).
Beatrice Chromková Manea – Ladislav Rabušic
Male Fertility in the Czech Republic – First Empirical Evidence
ArticleS2013 55 (4)
278
number of children both among men and women.
The proportion of women who had two children was
higher than that of men, the percentages being 55%
and 49%, respectively.7)
These findings are not particularly surprising.
An even higher proportion of childless men than
childless women might have been expected considering
that the first marriage rate is lower for men than
it is for women and, at the same time, the remarriage
rate among divorced men and women is about
the same.8)
The statistically significant higher proportion
of childless men than childless women is also
confirmed by data from another survey we conducted
in 2005 (unpublished results).9)
In this survey, 27%
of men compared to only 10% of women in the 35–39
age group were childless; 20% of men had one child
(24% of women), 42% of men had two children (52%
of women), and 11% of men had three children (14%
of women). Given that the respondents were born
in the period 1965–1985, they constitute one
of the first generations whose reproductive period
falls, more or less, under the new demographic regime.
The data above indicated that the total fertility rate
of Czech women would not be particularly staggering
in 5–6 years’ time, that is in 2010–2011.10)
Data from our survey thus indicate the following
pattern: men are childless statistically significantly
more often than women, and women have two children
statistically significantly more often than men.
Men and women who have one child do not statistically
differ, and neither do those who have three
or more children (according to T-test results).
However, men and women basically do not differ
as regards the total average number of children,
as shown in Table 1 (the row ‘completed fertility rate’).
On average, men had 1.75 children and women had
1.84 children. In the light of the divorce and remarriage
rates (but bearing in mind that the remarriage
7)	The difference is statistically significant at the 0.05 level.
8)	Total male first marriage rate based on first-nuptiality tables was 53.5% in 2011, while the female rate was 61% in 2011
(data from Demografická ročenka ČR 2011).
9)	This sociological survey was called ‘Marriage, Work and Family’ and the data were collected from a representative sample
of Czech women aged 20–40 and their male partners (N = 2,456 respondents) in 2005. Unpublished.
10)	Indeed, the total fertility rate was 1.49 in 2010 and 1.43 in 2011.
0
10
16.3
9.6
16.5 16.8
49.3
54.5
17.8 19.120
30
40
50
60
No children One children Two children Three and more children
Men Woman
Figure 1 Percentage distribution of respondents aged 40+ regardless of family status according
to the number of children they have (N=2321)
Source: Data set Male Reproductive Behaviour 2011; authors’ calculations.
279
rates for divorced men and women are similar), we
expected men to have higher completed fertility rates,
as they marry younger female partners with whom
have higher chances to have additional children
and have a higher average number of children than
women.11), 12)
Our expectation was not fulfilled, but
partially confirmed the results obtained by Pakosta
(2008), who found, on the one hand, that divorcing
before childbearing in the case of women increases
the likelihood that women give birth to only one child,
and, on the other hand, a large proportion of women
who divorced after having a child(ren) and remarry
then plan to have/have another child (they attempt
to fulfill their desired number of children not reached
in the previous marriage), which is reflected in a higher
completed fertility rate for remarried women compared
with women who never divorced.
If men aged 40–47 have been realised their expected
fertility (see Table 1), the level of completed
fertility would obviously have increased slightly, but
not enough to reach and outdo the rates for women.
Yet another finding can be drawn from Table 1.
The average number of children in the group
of male respondents aged 48–55 years is higher than
the average figures in the 40–47 age group, which
suggests that men have children even around the age
of fifty. The average number of children in the group
of women aged 48 years and over is also higher than
in the 40–47 age group.13)
This might be due
to the higher completed fertility rate of older cohorts,
which had children in the 1970s. This is valid for men
too (see Rychtaříková, 2004).
No differences could be found between the groups
regarding the hypothetical completed fertility rate either
(see Table 1), with a level of 1.83 children for men
and 1.90 for women. The hypothetical completed fertility
level is slightly higher than the actual completed
fertility rate in the group of men aged 40–47 years, indicating
that some men at this age were still planning
to have children. As regards women aged 40 years and
over, the actual and hypothetical completed fertility
do not de facto differ. Hence, the surveyed generation
of women were not planning to have children with
their spouses or partners past the age of 40.
These findings do not point to any substantial
differences in the reproductive behaviour of men
and women who, for the most part, fulfilled their
fertility aspirations in the period of so-called ‘Czech
11)	We also take into account conclusions by e.g. Bumpass (1984) and Loomis and Landale (1994) who claim that previous parity
may play an important role in the woman’s subsequent fertility behaviour, since it may reduce the probability of the woman
conceiving another child if she already has two or more children from a previous relationship.
12)	Divorced men who remarry are in most of the cases older than their female partners. For example, in 2011 20% of all remarried
men were on average 10–14 years older than their female partners (and about 20% of them 7–9 years older than women).
On the other hand, 17% of all women who remarried were on average one year older than their husbands (and about the same
proportion were 5–6 years older) – data from Demografická ročenka ČR 2011.
13)	One would think that the number of stepchildren might influence the level of the completed fertility rate. However, only 2.7%
of all women aged 40+ are stepmothers, while 7% of all men in our sample have step-children.
Table 1 Actual and hypothetical completed fertility by sex and age groups (respondents aged 40+)
Males total
(N=1,251)
Age 40–47
(N=613)
Age 48–55
(N=619)
Completed fertility rate (CFR) 1.75 1.65 1.85
Hypothetical completed fertility rate (HCFR) 1.83 1.80 1.86
Females total
(N=1,070)
Age 40–47
(N=585)
Age 48+
(N=507)
Completed fertility rate (CFR) 1.84 1.82 1.94
Hypothetical completed fertility rate (HCFR) 1.90 1.88 1.94
Source: Data set Male Reproductive Behaviour 2011; authors’ calculations.
Beatrice Chromková Manea – Ladislav Rabušic
Male Fertility in the Czech Republic – First Empirical Evidence
ArticleS2013 55 (4)
280
socialism’. Although it might therefore seem that
it is not necessary to concern ourselves with the issue
of male reproductive behaviour, we would not
agree with such a conclusion. We are convinced that
these analyses are meaningful – at the very least because
we now know more about male fertility than
we did before conducting our survey. The small differences
found could be attributed to a cohort effect,
since we surveyed the generation of respondents born
in the period 1956–1971 whose reproductive patterns
were rooted in the conditions of the old (that is socialist)
demographic regime, with all its peculiarities.14)
If we continue to study male reproduction and if we
examine different cohorts, we can expect to witness
very interesting diachronic developments brought
about by the differentiation and individualisation
of lifestyles of subsequent generations.
3.2 Family size preferences – the ideal and reality
In their pursuit to understand fertility rates, researchers
increasingly believe that it is necessary to also
explore individual preferences, motivations and intentions,
as these are important predictors affecting,
at the individual level (and also in the aggregate form
of the total fertility rate), the final number of children
men and women have. Preferences, motivations
and intentions are part of a person’s individual social
and psychological profile, but they are also
contextdependent – e.g. related to a person’s situation
in terms of economic circumstances or success
in the marriage/mating market, affected by social
norms (and stereotypes) and so on.
Demographers and sociologists have therefore introduced
further indicators, which help reveal various
aspects of fertility and reproductive behaviour:
the ideal number of children, the preferred number
of children, the intended/expected number of children
and the desired number of children, in order
to pinpoint the mechanisms that ultimately determine
the final number of children men and women have.
Miller and Pasta (1995a, 1995b) and Miller (1994)
claim, for example, that the desired and intended
number of children are key concepts for understanding,
and therefore also for predicting, fertility rates
and family size. They uphold the opinion that a distinction
needs to be made between desires/preferences
and intentions: desires/preferences are an expression
of what people want to do, while intentions refer
to what people plan to do. Desires/preferences are underlined
by attitudes, beliefs and motivations. Consequently,
they propose that distinct indicators should
be constructed for each of these categories.
In our survey, we used the variable ‘ideal number
of children’. As is generally known, what this variable
actually measures sensitively depends on how
the question is formulated and phrased. It may indeed
be formulated at the individual level (the number
of children the respondent would desire to have under
ideal life conditions) or at a general societal level
(the ideal number of children in general). Multiple
surveys have shown what we suspected, that
the reported ideal number of children at a general level
is higher on average than the reported ideal number
at the individual level. In addition, it has repeatedly
been shown that there exists a distinction between
the ideal number of children, the intended number
of children and the actual number of children.
Van Peer (2002) and van Peer and Rabušic (2008)
have confirmed that the individual norm governing
the number of children is set lower than the general
norm and that the actual (realised) fertility level
is below what is considered ideal at the individual level.
Let us now examine our survey findings
on male reproductive fertility from the point of view
of the ideal number of children. The questionnaire
asked about the ideal number of children at both
the individual and the general societal level. The ideal
number of children at the individual level was assessed
using the question: ‘If circumstances in your
life were entirely favourable how many children would
you like to have in total?’ The ideal number of children
at the general societal level was measured using
the question: ‘What do you think is the ideal number
of children in a family?’ These questions were asked
separately from each other during the interview
to avoid the so-called ‘halo effect’, whereby an answer
14)	As regards the family, it means the discourse establishing the family as the basis of the state and the social pressure on people
to have children.
281
to one question influences the answer to the following
question. The resulting percentage distribution of respondents
according to number of children and fertility
averages for each indicator are presented in Table 2.
It has again been confirmed that the perceived individual
and the societal ideal concerning the number
of children are on average higher than the actual average
number of children (see the last row in Table
2). The disparity between the two ideal numbers
and the reality is particularly marked in the case
of childlessness. Being childless is almost never seen
as an ideal (the value ranges between 1%
and 3%), while the reality is many times higher (16%
in the case of men and 11% in the case of women).
The one-child ideal points to interesting results: while
it is viewed as a societal and individual ideal by 13%
and 15% of men, respectively, and by 9% and 10%
of women, respectively, the reality is different, particularly
in the case of women: 18% of women stopped
at one child, which is almost double the ideal number,
and so did 17% of men – here the ideal only slightly
differs from the reality. This finding further confirms
our prognosis, formulated in 2007, that the proportion
of women with only one child (that is their final number
of children) will stand at about 20–25% around
the years 2020–2025 (for more details see Rabušic –
Chromková Manea (2007), where arguments supporting
this prognosis are presented).
Considerable differences can also be seen
in the case of the two-child model. Two children are
largely regarded – by 70% of men and 74% of women
– as a general societal ideal (that is as an ideal number
of children in a family). At the individual level,
the preference for two children is much weaker: expressed
by 58% of men and 62% of women. The reality
is quite different, even below the level of individual
preferences: 49% of men and 53% of women ultimately
had two children (the difference between the proportions
of the actual and the ideal number of children
is statistically significant).
Variations between the real number of children
and the ideal number of children at both
the individual and the societal level are also found for
the model 3+ children. This model is chosen to be ideal
at the individual level by 24% of all men, respectively
27% of women, while only 15% of all men and 16%
of all women consider it to be ideal at the general social
level. On the other hand, only approximately 18%
of men and women have 3 or more children.
As Table 2 shows, in many cases the real number
of children is lower than the ideal, desired number.
There is a variety of reasons why the declared ideal
numbers concerning family size differ from the actual
family size: postponing marriage or partnership
(and hence conception) until a later age, career
as a life priority, economic circumstances, the health
condition of one or both partners, and divorce. Also,
conflicting preferences between the spouses (partners)
concerning the number of children may play a role.15)
Last but not least, the Czech social climate that ‘prefers’
the two-child model might put some pressure on those
who prefer more than two children.
15)	For details on couple dynamics and its impact on fertility, see Chromková Manea – Fučík, 2007.
Table 2 Percentage distribution of respondents according to the ideal and the actual number
of children (all repondents – N=2,500)
Number of children
(% Col)
Social ideal Individual ideal Actual number of children
Males Females Males Females Males Females
No child 1.6 0.5 3.3 0.9 16.3 10.7
13.4 9.2 14.5 9.9 16.5 18.01 child
2 children 69.9 73.9 58.2 61.9 49.3 53.1
3+ children 15.0 16.3 24.0 27.2 17.8 18.2
Average number of children 2.00 2.07 2.10 2.22 1.69 1.79
Source: Data set Male Reproductive Behaviour 2011; authors’ calculations.
Note: The societal ideal was indicated by the question: ‘What do you think is the ideal number of children in a family?’ The individual ideal
was indicated by the question: ‘And if circumstances in your life were entirely favourable how many children would you like to have in total?’
Beatrice Chromková Manea – Ladislav Rabušic
Male Fertility in the Czech Republic – First Empirical Evidence
ArticleS2013 55 (4)
282
3.3. Variables in the context of male fertility
What are the characteristics men possess that determine
their completed fertility rate, on the one hand,
and their views of the ideal number of children, on the
other hand? Since our study is essentially an exploratory
analysis – inasmuch as the research topic is new
– we do not test any deductively derived hypotheses
and we will begin with a simple bivariate description.
We will ask to what extent the basic characteristics
of age, family status, education and level of religiosity
influence the variability of the dependent variables –
completed fertility rate, the ideal number of children
at the individual level and the ideal number of children
at the general societal level. Inspired by Catherine
Hakim’s (2000) preference theory, we have added
a synthetic indicator measuring preferences for work
and family, which should be a significant covariate
of fertility behaviour.16)
The results of bivariate analysis including
the correlation coefficients presented in Table 3
(see the columns referring to male respondents)
show that the impact of the selected characteristics
of the male respondents (except age, all are
in the form of categorical variables) on their completed
fertility rates, and the perceived ideal
number of children at the individual, as well
as the general societal level, is for the most part very
weak or none at all. Older male respondents have
on average a higher completed fertility rate than
younger ones, as well as a higher ideal number of children
at the individual and societal levels. Nevertheless,
certain categories of independent variables did show
some correlation. This applies to family status, where
those male respondents who were single had a significantly
lower completed fertility rate than the others
(having on average 0.29 children). The ideal number
of children for single men was also considerably lower
in comparison with the other family status categories
for men. Clearly, this is also true for education, where
a downward linear trend is apparent in Table 3, where
the highest completed fertility rate as well as the ideal
number of children, regardless of the level of measurement,
is found among men with primary education,
whereas at the opposite pole are men with university
education, who have on average the lowest completed
fertility rate and ideal number of children. Indeed,
the Kendall’s Tau coefficient of association between
completed education and the completed fertility rate
was –0.50 (sig. < 0.05) for men, while between completed
education and the ideal number of children
at the individual level was –0.45 (sig = 0.11) for men
too. However, the effect of education on views about
the ideal number of children is statistically insignificant
– except for male respondents with primary
school education who differed from the others
in terms of the ideal number of children they would
like to have, if circumstances were ideal: the average
was 2.58, which is the highest value of all
the characteristics shown in Table 3. No differences
were found – and this is a surprise – in relation to religious
beliefs: be it in terms of the male actual number
of children (measured as the completed fertility rate)
or their perceptions of the ideal number of children.
The ‘preference theory’ variable fulfilled, to some
extent, its function as a distinguishing feature for
the male population. The work-centred men have – in
line with Hakim’s theory – a significantly lower completed
fertility rate as compared to other preference
typology categories. Work-centred men also stated
the lowest values when answering the questions about
the ideal number of children at both the individual
and societal levels. As for the measurement of association
between preference typology and the dependent
variables (real and ideal number of children), the levels
of association as measured by the Gamma coefficient
come close to zero (–0.03) and are statistically insignificant
in the case of all the three dependent variables.
The following part of the paper will focus
on the results for women (see the columns referring
to female respondents in Table 3). As was the case
with men, women of a younger age (40–47 years) have
a lower completed fertility rate and ideal number
of children than older age category. And similarly
to their male counterparts, single female respondents
showed lower values than respondents in the other
family status categories in terms of their completed
fertility rate (on average 0.39 children) and perceptions
of the ideal number of children (on average 1.68
16)	As regards the family, it means the discourse establishing the family as the basis of the state and the social pressure on people
to have children.
283
and 1.77, respectively). Also in the case of female
respondents, there are differences between women
with different levels of completed education, where
primary educated women have on average the highest
completed fertility level (on average 2.54 children)
and the highest ideal number of children at both
the individual and societal levels (the highest value
is observed for the ideal number of children at individual
level, where women with primary education
reported on average 3 children). Female respondents
17)	To find the bivariate correlation between age and the three dependent variables, we used ‘age’ measured at the interval level.
Table 3 Average number of children (completed fertility rate), average ideal number of children at the individual
level and average ideal number of children at the general societal level by selected characteristics, coefficients
of association – males aged 40–55, females aged 40+, in 2011 (N=1,251 for males, N=1,070 for females)
Completed fertility rate
Ideal number of children
at the individual level
Ideal number of children
at the general societal level
Males Females Males Females Males Females
Age group
Pearson coefficient (sig.
level)17)
0.112
(sig.<0.00)
0.078
(sig.<0.01)
0.083
(sig.<0.00)
0.081
(sig.<0.01)
0.098
(sig.<0.00)
0.084
(sig.<0.01)
1.65 1.82 2.09 2.22 1.99 2.1040–47
48+ 1.85 1.94 2.11 2.33 2.09 2.14
Family status
Eta coefficient (sig. level) 0.530 0.403 0.294 0.156 0.259 0.201
Married 1.98 2.01 2.22 2.31 2.08 2.15
Widowed 1.91 2.06 2.25 2.32 2.13 2.14
Divorced 1.87 1.84 2.16 2.30 2.04 2.09
Single 0.29 0.39 1.41 1.68 1.65 1.77
Completed education
Kendall tau coefficient
(sig. level)
–0.500
(sig.<0.05)
–0.114
(sig.<0.00)
–0.450
(sig.<0.11)
–0.179
(sig.<0.00)
–0.004
(sig.<0.89)
–0.115
(sig.<0.00)
Primary 2.23 2.54 2.58 3.03 2.13 2.30
Lower secondary / vocational 1.78 1.93 2.19 2.35 2.03 2.19
Upper secondary 1.74 1.84 2.11 2.20 2.04 2.08
University 1.59 1.64 2.10 2.09 2.05 2.04
Religious belief
Eta coefficient (sig. level) 0.050 0.102 0.132 0.172 0.123 0.169
Religious person 1.85 2.03 2.34 2.49 2.16 2.27
Not a religious person 1.72 1.82 2.10 2.17 2.00 2.06
Preference typology18)
Gamma coefficient
(sig. level)
–0.030
(sig.<0.49)
0,205
(sig.<0.00)
–0.038
(sig.<0.43)
0.143
(sig.<0.05)
0.025
(sig.<0.64)
0.140
(sig.<0.01)
Work-centred 1.34 1.20 1.74 1.81 1.77 1.85
Adaptive 2.01 1.88 2.28 2.22 2.09 2.09
Family-centred 2.06 2.11 2.25 2.37 2.15 2.19
Inconsistent (ambivalent) 1.43 1.68 1.85 2.04 1.73 1.93
Source: Data set Male Reproductive Behaviour 2011; authors’ calculations.
Note: The level of religiousness was assessed using the question: ‘Irrespective of whether you go to church or not, would you say
you are a religious person, not a religious person or a convinced atheist?’
Beatrice Chromková Manea – Ladislav Rabušic
Male Fertility in the Czech Republic – First Empirical Evidence
ArticleS2013 55 (4)
284
18)	Categories of the preference typology: work-centred, adaptive, family-centred and inconsistent (ambivalent). These categories
were constructed based on two questions: (1) There has been much debate lately about the changing traditional roles of a man
and a woman in the family. I will now present to you six models of the family. Which of them corresponds best with your own
ideas about an ideal family arrangement? (see appendix – models, first column), and (2) There are many ways in which people
can organise their lives in terms of starting a family with children, having a job, pursuing a career and pursuing hobbies. Let
me give four examples. Which of them comes closest to your own views? We coded as ‘committed to work’ those respondents
who chose the answer ‘Most important for me is work – to this I subordinate my family life as well as hobbies and interests’.
Those who answered ‘Most important for me is to have a family and children – to this I subordinate my work as well as my
hobbies and interests’ are coded as ‘committed to family’. ‘Adaptive’ are respondents who answered ‘Both family and work are
important for me, so I try to reconcile them’. Respondents were also offered a fourth option: ‘My interests and hobbies are
most important for me, so I prefer them to the family and work’. We recoded these answers as ‘committed to work’ because
quite often personal hobbies and work are correlated, and preferences of hobbies instead of family indicate work orientations
(for more information on the preference typology, see Chromková Manea – Rabušic – forthcoming in October 2013, in Crespi,
I. – Miller, T. (eds.) ‘Family, Care and Work in Europe: Gender Issues’).
19)	We use standardised beta coefficients for interpretation as b coefficients are measured in different units and the effects
	 of the various independent variables cannot be compared otherwise.
with higher education levels did not significantly differ.
However, the value of the association coefficient
between completed education and the completed
fertility rate is much weaker for women than it is for
men (Kendall’s Tau for females = –0.11, sig. 0.00).
Similarly to men, female respondents who considered
themselves religious did not differ from those who did
not, in terms of their completed number of children.
However, if conditions were ideal, religious women
would desire to have more children than women who
are not religious (2.49 : 2.17; the t-test showed, however,
that this difference is not statistically significant).
The typology based on the preference theory showed
a more pronounced effect in the case of women than
it did in the case of men: work-centred women had
a much lower completed fertility rate (1.20) than any
other category, while women oriented towards family
had on average the highest level of completed fertility
rate (2.11). It is worth mentioning that adaptive
and family-centred women are quite close in their
preferences for the ideal number of children at both
the individual and societal levels.
Further on, the variables in Table 3 were put into
a multiple linear regression analysis to find out
the net effects of predictor variables (those in rows)
to outcome variables (those in columns). The results
are presented in Tables 4a and 4b, where the
standardised beta coefficients are presented for men
and women separately.19)
The outcome (dependent)
variables – completed fertility rate, the ideal number
of children at the individual level and the ideal number
of children at the general societal level – are all measured
at the interval level; the predictor (independent)
variables other than the interval (continuous) ones,
were entered as dummies.
The covariates included in our analysis were
as follows:
•	 Age of respondent – continuous variable, respondents
	 aged 40+
•	 Marital status – categorical variable, where 1 means
	 married, 2 widowed, 3 cohabiting, 4 single – recorded
	 into 3 dichotomous/dummy variables and using
	 a reference coding scheme with ‘single’ as the reference.
•	 Educational level – categorical variable, where 1
	 means primary education (ISCED level 1 and 2),
	 2 is lower secondary/vocational education
	 (ISCED 3B and 3C), 3 stands for upper secondary
	 education (with GCSE – ISCED level 3A),
	 and 4 is for completed university/tertiary education
	 (ISCED level 5A, 5B or 6) – recorded into 3
	 dichotomous/dummies variables and using
	 a reference coding scheme with ‘university
	 education’ as reference.
•	 Religion – dichotomous variable coded as 1
	 ‘religious’ and 2 ‘not religious’
•	 Preference typology – categorical variable, where
	 1 means work-oriented, 2 adaptive, 3 family-oriented
	 and 4 inconsistent – recorded into 3 dichotomous/
285
Table 4a Multiple linear regression models for dependent variables completed fertility level, ideal number
of children at the individual level and ideal number of children at the general societal level by selected sociodemographic
characteristics, standardized beta coef. – male respondents aged 40–55 (N=1,251)
Completed fertility
Ideal number of children
at the individual level
Ideal number of children
at the general societal level
Beta coef. Sig. Beta coef. Sig. Beta coef. Sig.
Age 0.034 0.282 0.016 0.649 0.044 0.217
Family status (ref. cat. Single)
Married 0.575 0.000 0.197 0.004 0.146 0.028
Widowed 0.173 0.000 0.077 0.042 0.074 0.053
Divorced 0.520 0.000 0.175 0.006 0.156 0.013
Completed education (ref. cat. University)
Primary 0.156 0.000 0.113 0.002 0.038 0.302
Lower secondary / vocational 0.164 0.001 0.063 0.230 –0.029 0.585
Upper secondary 0.136 0.004 0.050 0.343 0.002 0.969
Religious belief (ref. cat. Not a religious person)
Religious person 0.068 0.030 0.116 0.000 0.087 0.013
Preference typology (ref. cat. Inconsistent)
Work-centred –0.121 0.000 –0.183 0.001 –0.154 0.000
Adaptive excluded excluded excluded
Family-centred 0.022 0.509 –0.035 0.334 0.047 0.206
Nagelkerke R square 0.195 0.087 0.066
Source: Data set Male Reproductive Behaviour 2011; authors’ calculations.
	 dummies variables and using a reference coding
	 scheme with ‘inconsistent’ as reference.
The first model for the completed fertility rate
has an R2
of 0.195. This means that the predictor
variables explain only 19.5% of the variation
in the outcome variable. The p value for the F statistic
is lower than 0.05 and it implies that at least one
of the independent variables is a significant predictor
of the completed fertility rate. Variable age is not a significant
determinant of male completed fertility rate.
All other variables do play statistically significant role,
though. Being married or divorced have the strongest
impact relative to the single respondents on the male
completed fertility rate when controlling for other
variables (0.575 and 0.520, respectively, sig. < 0.00).
Religiousness, although statistically significant, has
the lowest value for the standardised beta coefficient,
which indicates a very week impact of being religious
on our dependent variable.
The second model aimed to determine the covariates
having significant impact on the male ideal
preferred number of children at the individual level
explains only 9% of the variance (R2
is 0.087). We
learn that the ideal number of children for men does
not vary by age when all other variables are controlled
for. As in the model for completed fertility rate, indicators
for marital status are significant, but the value
of the standardised beta coefficients is quite low (close
to 0). As for completed education, it seems that primary
education have a significant influence on the ideal
preferred number of children compared to university
education when other covariates are controlled for.
However, the beta coefficient is low. The same effect
has religiousness. As far as the preference typology
is concerned, the results come to confirm that being
oriented towards work as compared to being inconsistent
(ambivalent) decreases the ideal preferred
number of children men would like to have when
controlling for the rest of determinants included
in the analysis. Being family-centred has no effect
in this model, however.
The third model for the ideal preferred number
of children at the general societal level has an even
lower explanatory power than the previous two
Beatrice Chromková Manea – Ladislav Rabušic
Male Fertility in the Czech Republic – First Empirical Evidence
ArticleS2013 55 (4)
286
Table 4b Multiple linear regression models for dependent variables completed fertility level, ideal number
of children at the individual level and ideal number of children at the general societal level by selected sociodemographic
characteristics, standardised beta coef. – female respondents aged 40–55 (N=1,070)
Completed fertility
Ideal number of children
at the individual level
Ideal number of children
at the general societal level
Beta coef. Sig. Beta coef. Sig. Beta coef. Sig.
Age –0.005 0.891 0.024 0.500 0.030 0.830
Family status (ref. cat. Single)
Married 0.636 0.000 0.275 0.000 0.207 0.010
Widowed 0.258 0.000 0.091 0.044 0.056 0.227
Divorced 0.536 0.000 0.207 0.005 0.133 0.080
Completed education (ref. cat. University)
Primary 0.157 0.000 0.203 0.000 0.054 0.186
Lower secondary / vocational 0.137 0.010 0.128 0.024 0.105 0.072
Upper secondary 0.055 0.303 0.025 0.665 –0.002 0.969
Religious belief (ref. cat. Not a religious person)
Religious person 0.087 0.008 0.152 0.000 0.143 0.000
Preference typology (ref. cat. Inconsistent)
Work-centred –0.104 0.003 –0.070 0.064 –0.067 0.086
Adaptive excluded excluded excluded
Family-centred 0.063 0.110 0.069 0.061 0.072 0.055
Nagelkerke R square 0.183 0.119 0.070
Source: Data set Male Reproductive Behaviour 2011; authors’ calculations.
models (R2
only 0.066). Marital status, religious belief
and to some extent preference typology have
a significant impact on the dependent variable, but
the beta coefficients have a very low value (close to zero).
The next table (see Table 4b) presents the same
regression analysis but for female respondents.
The results are quite similar to the male ones for
the completed fertility rate, where being married or
divorced as compared to single increases the chances
of having a higher realised fertility rate (beta coefficients
above 0.5). As for men, lower levels of completed education
for women seem to have a significant impact on female
fertility rate when all other variables are controlled
for, but the values of beta coefficient are low (close
to 0). A similar effect is observed for religiousness.
The same pattern can be found for the ideal desired
number of children at the individual level, where beta
coefficients for marital status, education, religiousness
and additionally preference typology are statistically
significant, but their value is low (close to 0).
As for the reported ideal number of children
at the general societal level, the regression model for
women contains only two variables that are statistically
significant, but again the beta coefficients are low:
being married (as compared to being single) or being
a religious person have a growing impact on the ideal
desired number of children when all other variables
are controlled for in the equation.
4. Final remarks and discussion
This article presented new empirical data on male
reproductive behaviour in the Czech Republic,
building on our previously published overview paper
on this topic. It examines the completed fertility
rates of men and women, as well as their attitudes
concerning the number of children, fertility intentions,
and other determinants of fertility. As regards
completed fertility, we have seen that men remain
childless more often than women – this is particularly
evident in the case of people living outside
of marriage, even when in a stable relationship.
Overall, the comparison showed only small differences
between male and female fertility:
287
women had on average 1.84 children and men had
1.75 children. The number of children perceived
to be ideal at the individual level is on average
higher than what is perceived to be ideal at the general
societal level. In comparing the ideal number
of children to the actual number we can see that
the ideal is not being lived out in the case of people
who are childless or have only one child – there
were many more respondents with no or only one
child than the expressed ideal would suggest.
It follows that the decision not to have a child
or to have only one is not necessarily made in advance.
By the same token, there were fewer respondents
with two children than would be expected given
the declared ideal.
We also examined the possible effects of basic sociodemographic
characteristics such as the respondent’s
age, family status, education, and religiousness,
as well as the possible impact of a synthetic indicator
measuring work and family-related preferences.
For the most part, the influence of the independent
variables has proven to be very weak or none
at all. Single respondents had a markedly lower
completed fertility than the others and their subjective
ideal number of children was also significantly
lower than was the case with the other family status
categories. Religiosity has no effect on either completed
fertility or the ideal of the number of children.
The indicator ‘work and family-related preferences’
has revealed some partial effects in the group of workcentred
men who have both a lower completed fertility
rate and who regard as ideal a smaller number of children.
The female sample showed similar results. Our
data captures reproductive behaviour of people born
between 1956 and 1971 that is in the period of the old,
so-called socialist, demographic regime with largely
homogenised living conditions. The differentiation
of lifestyles of the ensuing generations that we have
been witnessing is likely to produce greater variation.
This survey is only the first step and will be followed
by further research: for example, we intend
to analyse the timing of life events in relation
to male reproduction. We will also seek to explore
the phenomenon described as ‘baby fever’ in international
literature (see e.g. Rotkirch – Basten, 2012)
to see whether it is present in the Czech population
– specifically the male population – and what
possible effect it has on the completed male fertility
level. We also plan to analyse the role that men
have in sexual activities and the use of contraceptives,
as this is also an influential attribute of male
reproductive behaviour.
References
•	 Bledsoe, C. – Lerner, S. – Guyer, J. I. 2000. Fertility and the Male Life-Cycle in the Era of Fertility Decline. New York: Oxford University Press.
•	 Demografická ročenka ČR 2011. 2012. Kód publikace e-4019-12. Praha: ČSÚ. Available at:
	<http://www.czso.cz/csu/2012edicniplan.nsf/p/4019-12>.
•	 Greene, M. E. – Biddlecom A. E., 2000. Absent and Problematic Men: Demographic Accounts of Male Reproductive Roles: Population
	 and Development Review, 26(1), pp. 81–115.
•	 Goldscheider, F. K. – Kaufman, G. 1996. Fertility and commitment: Bringing men back in. In J. Casterline and R. Lee (Eds.), Fertility
	 in the United States: New patterns, new theories. Population and Development Review, Special Supplement to Vol. 22, pp. 87–99.
•	 Hakim, C. 2000. Work-lifestyle choices in the 21st century: Preference theory. Oxford: Oxford University Press.
•	 Chromková Manea, B. – Rabušic, L. 2013. Preference for family and work in the Czech Republic. In Crespi, I. and Miller, T. (eds.). Family,
	 care and work in Europe: gender issues. Pp. 15–46. Macerata: eum Edizioni Università di Macerata, 2013. Prima edizione.
•	 Chromková Manea, B. – Fučík, P. 2007. Teorie konfliktních preferencí a plodnost v České republice. Demografie, 49, 4, s. 244–252. Praha: ČSÚ.
•	 Miller, W. B. – Pasta, D. J. 1995a. Behavioral Intentions: Which Ones Predict Fertility Behavior in Married Couples? Journal of Applied
	 Social Psychology, 25(6), pp. 530–555.
•	 Miller, W. B. – Pasta, D. J. 1995b. How Does Childbearing Affect Fertility Motivations and Desires? Social Biology, 42, pp. 185–98.
•	 Miller, W. B. 1994: Childbearing Motivations, Desires, and Intentions: A Theoretical Framework. Genetic, Social, and General P
	 sychological Monographs, 120, pp. 223–258.
•	 Pakosta, P. 2008. Rozvodovost a vzdělání obou manželů. Demografie, 5 (1), pp. 64–70.
Beatrice Chromková Manea – Ladislav Rabušic
Male Fertility in the Czech Republic – First Empirical Evidence
ArticleS2013 55 (4)
288
•	 Průměrný věk matek při narození dítěte se zvyšuje, a co věk otců? 2011. Praha: ČSÚ. Available at:
	 < http://www.czso.cz/xb/redakce.nsf/i/prumerny_vek_matek_pri_narozeni_ditete_se_zvysuje_a_co_vek_otcu>.
•	 Rabušic, L. – Chromková Manea, B.-E. 2007. Jednodětnost v českých rodinách (kdo jsou ti, kdo mají nebo plánují pouze jedno dítě).
	 Sociologický časopis/ Czech Sociological Review, 43 (4), pp. 699–719.
•	 Rabušic, L. – Chromková Manea, B. 2007. Preferenční teorie Hakimové v českém kontextu. Demografie, 49 (2), pp. 77–86.
•	 Rabušic, L. – Chromková Manea, B. 2011. Řekni, kde ti muži jsou? O chybějících mužích ve studiích reprodukce. Sociální studia, 8(4), pp. 47–66.
•	 Rotkirch, A. – Basten, S. 2010. Men Longing for Children: Social and Situational Characteristics. VID Conference ‘From Intentions
	 to Behaviour: Reproductive Decision-Making in a Macro-Micro Perspective’.
•	 Rychtaříková, J. 2004. Změny v generační plodnosti v České republice se zaměřením na vzdělání žen. Demografie, 46, 2, s. 77–90. Praha: ČSÚ.
•	 Sonenstein, F. L. – Stewart, K. – Duberstein Lindberg, L. – Pernas, M. – Williams, S. 1997. Involving Males in Preventing Teen Pregnancy:
	 A Guide for Program Planners. Advice and strategies from successful programs that teach young men about responsible sexual behavior.
	 Washington, D.C.: Urban Institute.
•	 Tölke, A. – Diewald, M. 2003. Insecurities in employment and occupational careers and their impact on the transition to fatherhood
	 in Western Germany. Demographic Research, vol. 9(3), pp. 41–68. Rostock: Max Planck Institute for Demographic Research.
•	 Toulemon, L. 2001. Men's fertility and family size as compared to women's. XXIVth IUSSP General Population Conference, Session S-13.
•	 van de Kaa, D. J. 1987. Europe’s Second Demographic Transition. Population Bulletin,42(1). Washington: Population Reference Bureau.
•	 van Peer, Ch. 2002. Desired and realized fertility in selected FFS-countries. In Macura, M. – Beets, G. (eds.). Dynamics of Fertility
	 and Partnership in Europe: Insights and Lessons from Comparative Research. Vol. 1. Pp. 117–142. New York, Geneva: United Nations.
•	 van Peer, Ch. – Rabušic, L. 2008. Shall we witness an upturn in European fertility in the near future? In Avramov, D. – Hoehn, C. – Kotowska,
	I. People, population change and policies – lessons from the population policy acceptance study – volume 1: family change.
	 Pp. 215–241. The Hague: Springer. European Studies of Population, 16.
•	 Vývoj obyvatelstva České republiky (2011). 2012. Kód publikace y-4007-12. Praha: ČSÚ. Avalable at:
	<http://www.czso.cz/csu/2012edicniplan.nsf/t/6A002ECDC2/$File/400712.pdf>.
•	 Zeman, K. 2007. Nemanželská plodnost – demografický přehled. In.: Dana Hamplová (ed.). Děti na psí knížku? Mimomanželská plodnost
	 v ČR. Pp. 17–28. Praha: SOÚ AV ČR.
•	 Zhang, Li. 2011. Male fertility patterns and determinants. Dordrecht: Springer.
289
Appendix
Table 5 Main sample distribution
N – absolute values % col
Gender
Male 1,251 50.0
Female 1,249 50.0
Age
20–3920) 153 6.1
40–47 1,198 47.9
48+21) 1,142 45.7
Marital status
Married 1,764 70.6
Widowed 47 1.9
Divorced, separated 420 16.8
Single 245 9.8
Completed education
Primary 77 3.1
Lower secondary / vocational 970 38.8
Upper secondary 1,110 44.4
University 333 13.3
Size of residence
up to 2,000 inhabitants 472 18.9
2,000–4,999 412 16.5
5,000–9,999 312 12.5
10,000–19,999 261 10.4
20,000–49,999 264 10.6
50,000–99,999 213 8.5
100,000–499,999 111 4.4
500,000 and more inhabitants 455 18.2
Source: Data set Male Reproductive Behaviour 2011; authors’ calculations.
20)	All respondents in the age group 20–39 are women.
21)	Men only up to 55.
Beatrice Chromková Manea – Ladislav Rabušic
Male Fertility in the Czech Republic – First Empirical Evidence
ArticleS2013 55 (4)
290
Table 6 Typological scheme of lifestyle preferences
A. Committed
to work
B. Adaptive
C. Committed
to family
1) A family where the two partners each have an
equally demanding job and where housework
and the care of the children are shared equally
between them
Work-oriented Adaptive Adaptive
2) A family where the wife has a less demanding
job than her husband and where she does the
larger share of housework and caring for the
children
Inconsistent Adaptive Family-oriented
3) A family where the wife has a more demanding
job than her husband and where she does the
larger share of housework and caring for the
children
Work-oriented Adaptive Inconsistent
4) A family where only the husband has a job and
the wife runs the home
Inconsistent Family-oriented Family-oriented
5) A family where the two partners each have an
equally demanding job and where woman does
the larger share of housework and caring for the
children
Work-oriented Adaptive Family-oriented
6) A family where only the husband has a job and
where housework and the care of the children are
shared equally between them
Inconsistent Family-oriented Family-oriented
Source: Data set Male Reproductive Behaviour 2011; authors’ calculations.