Transition from Cohabitation to Marriage. The Role of Marital Attitudes in Seven Western and Eastern European Countries
Transition from Cohabitation to Marriage
The Role of Marital Attitudes in Seven Western and Eastern European
Countries
Zuzana Žilinčíková, Nicole Hiekel
Abstract: Using longitudinal panel data from the Generations and Gender Surveys
on 2,847 cohabiters from seven countries, we examine the role of marital attitudes
in the transition from cohabitation to marriage and compare the strength of this
association between Western and Eastern Europe. We expect a positive attitude
towards marriage to increase the likelihood of cohabiters marrying. We also expect
the association between personal attitudes and marriage formation to be weaker
among cohabiters from Eastern Europe, due to stronger normative pressure to mar-
ry in contexts where cohabitation is less prevalent.
In both Eastern and Western European countries, we fi nd a clear positive as-
sociation between favourable views on marriage among cohabiters and their entry
into marriage. Contrary to our expectations, we fi nd evidence that this association is
weaker in Western Europe. We discuss this fi nding in light of the greater postpone-
ment of marriage among Western European cohabiters, even among those with a
positive attitude towards marriage, as well as the potentially greater signifi cance
of life course events and transitions that infl uence their decision to marry more
strongly than is the case for their Eastern European counterparts.
This study extends the literature on the relationship trajectories of cohabiters by
drawing attention to the normative context that may shape cohabiters’ opportuni-
ties and constrain behavioural choice in the marriage formation process. Ultimately,
it contributes to an understanding of the consequences of the societal diffusion of
cohabitation in Europe.
Keywords: Marriage formation · Cohabitation · Cross-national comparison · Discrete
time event history analysis · Generations and Gender Surveys
Comparative Population Studies
Vol. 43 (2018): 3-30 (Date of release: 17.04.2018)
Federal Institute for Population Research 2018 URL: www.comparativepopulationstudies.de
DOI: 10.12765/CPoS-2018-04en
URN: urn:nbn:de:bib-cpos-2018-04en4
• Zuzana Žilinčíková, Nicole Hiekel4
1 Introduction
The need to better understand cohabiters’ attitudes, beliefs and values towards
marriage, and the association of these with the transition to marriage, has recently
been emphasised within life course research (Ortyl 2013; Perelli-Harris/Bernardi
2015). The decision to marry may strongly depend on how much cohabiters per-
sonally value the institution of marriage, particularly in countries where virtually all
unions start as unmarried cohabitation and cohabiters make similar relationship-
specifi c investments to those of married couples. The strong association between
positive attitudes towards marriage and marriage formation that has been found
among Swedish cohabiters (Duvander 1999; Moors/Bernhardt 2009) may, however,
not be as evident in contexts where cohabitation is less prevalent and accepted
because, regardless of their personal feelings about marriage, cohabiters may not
only be stigmatised for cohabiting in the fi rst place but may also be confronted with
social pressure to get married.
There is great variety in Europe regarding the prevalence and acceptance of un-
married cohabitation. In Western Europe cohabitation has virtually replaced direct
marriage as the start of a union and has become the norm in the process of union
formation (Billari/Liefbroer 2010). The group of cohabiters may not only be more
diverse in their marital attitudes but also face less normative pressure to marry
quickly. In Central and Eastern Europe by contrast, cohabitation is less prevalent
and thus entered into by a more selective part of the population, is contested by
normative beliefs about the institution of marriage, and is often associated with
economic disadvantage (Philipov/Dobritz 2003). We formulated the following re-
search question: Does the strength of the effect of personal attitudes towards the
institution of marriage on cohabiters’ transition to marriage differ in high and low
diffusion contexts of cohabitation?
Our study contributes to the literature in three important ways. First, we extend
a growing body of literature aimed at understanding the consequences of the so-
cietal diffusion of unmarried cohabitation by examining the marriage trajectories
of cohabiters. This line of research has predominantly focused on cohabiters’ life
course stage (i.e. union duration, age, parental status) or economic opportunities
(education, employment, race) and their association with marriage formation (Du-
vander 1999; Kennedy/Bumpass 2008; Kuo/Raley 2016; Lichter et al. 2006; Manlove
et al. 2012; Manning/Smock 2002; McClain 2011; Musick 2007; Oppenheimer 2003;
Perelli-Harris et al. 2012; Smock et al. 2005). We argue that by drawing attention
to the role of attitudes in cohabiters’ marital decisions, we complement the socio-
demographic and economic explanations of behavioural choices.
Second, we add a cross-national perspective and compare this association in
countries of low and high diffusion of unmarried cohabitation. We thus learn about
how the context shapes the opportunities and constraints of behavioural choice in
the process of union formation (Liefbroer/Billari 2010).
Third, we make use of prospective survey data on the partnership trajectories of
cohabiters. This enables us to capture attitudes towards the institution of marriage
prior to the demographic behavioural choice (marriage or continued cohabitation)
Transition from Cohabitation to Marriage • 5
and thereby to disentangle the reciprocal association between selection and expo-
sure effects; something that many prior studies could not effectively address due to
the cross-sectional or retrospective nature of their data (Surkyn/Lesthaeghe 2004).
Our data stem from seven Western, Central and Eastern European countries that
participated in two waves of the Generations and Gender Surveys (Austria, Bulgaria,
France, Georgia, Germany, Hungary and Russia). We studied 2,847 cohabiting men
and women between 18 and 45 years of age. We assessed their attitude towards the
institution of marriage in the fi rst wave of data collection and obtained information
on their relationship status roughly three years later, when some of them had mar-
ried and others were still cohabiting or had dissolved their union.
2 Theoretical Background
Since the onset of the societal diffusion of unmarried cohabitation as early as in
the 1970s in Scandinavia and as recently as the early 1990s in Central and Eastern
Europe, living together without being married has become an increasingly common
experience for European adults. The rising popularity of unmarried cohabitation
accompanied the decline in marriage rates and the postponement of marriage to
an older age. The increased prevalence of divorce has also been popularly framed
within the narrative of the so-called Second Demographic Transition (SDT hereaf-
ter, Lesthaeghe 1995, 2010; van de Kaa 1987, 1993). Adherents of the SDT concept
see ideational change as the driving force behind demographic shifts in partner-
ships and family life. At the onset of the diffusion of cohabitation, unfavourable at-
titudes towards the institution of marriage have been identifi ed as strong predictors
of people’s choice to cohabit as an alternative to marriage (Axinn/Thornton 1992;
Clarkberg et al. 1995; Liefbroer 1991; Manting 1996). At later stages of the SDT, co-
habitation was increasingly also adopted by people who viewed this union type as
a stepping-stone on the way to marriage and thus held more favourable attitudes
towards the institution. As a result, relationship trajectories prior to marriage have
become more heterogeneous and, in many European countries, most adults have
cohabited prior to their fi rst marriage (Perelli-Harris/Lyons-Amos 2015). Simultane-
ously, norms concerning cohabitation and marriage relaxed at the societal level,
hence larger groups of society approved unmarried cohabitation (Treas et al. 2014).
Cherlin (2004) coined the term “deinstitutionalisation” of marriage to describe these
changing cultural perceptions about cohabitation and marriage and the increased
complexity of union formation trajectories. Building on this thesis that marriage
has become a choice rather than a necessity, and that cohabitation is becoming
a nearly universal experience, a growing body of literature studying the marriage
trajectories of established cohabiting unions emerged in an effort to understand the
consequences of the changing composition of the cohabiting population in differ-
ent normative contexts of partnership and family life.
• Zuzana Žilinčíková, Nicole Hiekel6
2.1 Marital Attitudes and the Transition to Marriage
Symbolic interaction theory (Blumer 1969) would suggest that, through interaction
with their social environment, people form a personal attitude towards the institu-
tion of marriage that infl uences their decision whether to marry or not. Attitudes
here refl ect a positive or negative disposition towards a given person, object or
idea (Ajzen 1988). Ajzen’s theory of planned behaviour (Ajzen 1991) suggests that
marital attitudes have both a direct and an indirect effect on the transition (from co-
habitation) to marriage. Direct, in that individuals tend to exhibit behaviour that is in
line with the values they hold, and indirect in that the intention to marry that people
form is based on their attitude towards marriage. Attitudes towards the institution
of marriage have been widely used to explain the choice between cohabitation and
marriage as a selection process at the moment of union formation (Axinn/Thorn-
ton 1992; Bernhardt 2004; Bradatan/Kulcsar 2008; Clarkberg et al. 1995; Gerber/
Berman 2010; Hoem et al. 2009; Liefbroer 1991; Puur et al. 2012). This research
demonstrated that individuals who hold disapproving attitudes towards the institu-
tion of marriage choose cohabitation, whereas individuals who consider marriage
an important social institution would rather get married directly. With the unfolding
diffusion of unmarried cohabitation, the selection into cohabitation based on mari-
tal attitudes has become weaker and premarital cohabitation may even become a
normative step in the union formation process (Billari/Liefbroer 2010), hence also
widely adopted by people with a positive attitude towards marriage. It has further-
more been shown that cohabitation is not necessarily a conscious decision but
that people “slide” instead into co-residence with their partner (Manning/Smock
1995; Stanley et al. 2006). Apart from convenience as a reason to start living with
a partner, qualitative work also identifi ed housing needs or the benefi ts of shared
fi nances as frequent reasons to start cohabiting (Sassler 2004). Particularly in cases
where cohabiters already benefi t from increased returns to scale, attitudes emerge
as an explanation of why some cohabiters marry whereas others stay cohabiting
or break up (Moors/Bernhardt 2009). There is however surprisingly little research
on the infl uence of marital attitudes on cohabiters’ transition to marriage that com-
plements explanations of the transition from cohabitation to marriage based on
economic and life stage determinants (for exceptions on the marriage formation
of Swedish cohabiters, see Duvander 1999; Moors/Bernhardt 2009). In accordance
with the SDT framework that explains individual behavioural choice between co-
habitation and marriage at the moment of union formation as a selection process,
we formulate a similar hypothesis on the transition from cohabitation to marriage:
cohabiters who hold a positive attitude towards marriage are more likely to marry
than their counterparts who hold a negative attitude (Hypothesis 1).
2.2 Contextual variation in the association between marital attitudes
and marriage formation
Beyond cohabiters’ personal attitudes about marriage, their marital decisions may
also be infl uenced by what others consider to be appropriate behaviour. Buchmann
Transition from Cohabitation to Marriage • 7
(1989) argued that such social norms shape demographic behaviour, particularly in
less individualised societies. European countries differ in the degree of individuali-
sation, hence the extent to which individuals experience increased freedom from
normative constraints to conform to traditional demographic patterns enforced by
their family of origin, the church and society at large (Beck/Beck-Gernsheim 2002).
These constraints manifest themselves in the form of norms, more specifi cally social
rules regarding the timing and sequencing of life events. These can be formal (i.e.
laws that prohibit marriage below a certain age) or implicit (i.e. age norms for mar-
riage). Acts which violate these norms, called deviance, result in sanctions (Ajzen
1991). There are two forms of sanctions, namely legal consequences and stigma-
tisation. In a context where cohabitation is not prevalent, married unions are often
legally favoured over cohabitation, and cohabiters may also provoke derogatory
remarks about not being married. It is argued that norms infl uence demographic
outcomes through (anticipated) sanctions (Neugarten et al. 1965; Settersten/Mayer
1997). Concerning the transition from cohabitation to marriage, we argue that social
norms go beyond regarding unmarried cohabitation instead of marriage as a devia-
tion from expected behaviour; instead there is also social pressure to transform a
cohabiting into a married union (Ajzen 1991). Treas and colleagues (2014) showed
that although social norms concerning both premarital cohabitation and cohabita-
tion as a permanent alternative relaxed between the mid-1990 and the early 2000s
in all European regions, norms concerning cohabitation in the absence of marital
intentions are more disapproving, and more so in Eastern than in Western Europe.
Hence, in their decision of whether to marry or not, cohabiters also respond to so-
cial norms independently of their own attitudes by conforming to widely accepted
behaviour in their social environment (Liefbroer et al. 1994). Countries with different
levels of cohabitation prevalence may thus vary in the extent to which conformism
is widespread. We would consequently expect that cohabiters in countries with
low cohabitation prevalence are more likely to marry even if they have a negative
attitude towards marriage. Contrary to this, and in line with the deinstitutionalisa-
tion of marriage thesis, the marriage norm may be less rigid in countries with a high
prevalence of cohabitation, leaving the decision whether or not to marry largely up
to the couple themselves. Consequently, the extent to which individuals value the
institution of marriage may be a strong predictor of the decision to marry. In sum,
we hypothesised that the strength of the association between marital attitudes and
marriage formation is weaker in Central and Eastern European countries (Hypoth-
esis 2).
2.3 The Context of Cohabitation in Western and Eastern European
Countries Selected for this Study
Cohabitation has long historical antecedents in both Western and Eastern Europe
(Kok/Leinarte 2015). The persistent differences in patterns and meanings of cohabi-
tation in Eastern and Western Europe (Bradatan 2012) have been linked to histori-
cally distinct nuptiality patterns in these two regions since the middle ages (Hajnal
1965), as characterised by late and non-universal marriage west of an imaginary
• Zuzana Žilinčíková, Nicole Hiekel8
line running from St. Petersburg to Trieste, and early and universal marriage east
of that line.
For the purpose of this study, survey data from seven countries were selected
from the Generations and Gender Survey Programme (refer to the data section for
more information on available survey data and the sample selection): Austria, Ger-
many (East and West) and France, geographically located in Western Europe (WE),
and Bulgaria, Georgia, Hungary and Russia, geographically located in Central and
Eastern Europe (CEE). It is important to note that the selection of countries was not
based on theoretical considerations, but on the objective of including all countries
for which longitudinal information from two waves of data collection was available.
In Western European countries, represented in this study by Austria, Germany
and France, unmarried cohabitation has become a common choice for starting a co-
resident relationship, although childbearing within cohabitation is more prevalent
in France compared to Austria and Germany (Heuveline/Timberlake 2004; Hiekel et
al. 2014; Sobotka/Toulemon 2008). The majority of respondents (18-45 years) who
were married in the year when the GGS data were collected had cohabited with
their spouse before marriage (Austria: 77 percent, France: 81 percent, Germany:
69 percent), on average about 34 months in Austria and France and 22 months in
Germany (own calculations based on GGS data).
Prior to the fall of the Iron Curtain, but dramatically accelerated by the turmoil
of the transition from communism to democracy, unmarried cohabitation also
spread across countries in Central and Eastern Europe but was of shorter duration
and more rarely involved childbearing compared to Western European countries
(Philipov/Jasilioniene 2008; Puur et al. 2012). The proportion of respondents aged
18 to 45 in the year of data collection who had cohabited with their spouse prior to
marriage varied widely across the set of countries that here represent Central and
Eastern Europe, but is lower compared to Western Europe: whereas in Hungary
(25 percent) and Russia (38 percent) less than half of all married respondents had
cohabited prior to marriage, this proportion was larger in Georgia (53 percent) and
Bulgaria (64 percent) (own calculations based on GGS data). The majority of them
married within one year of starting to cohabit. Hence, the mean duration (in months)
of premarital cohabitation was signifi cantly shorter compared to Western Europe
(Bulgaria: 6.9, Georgia: 8.7, Hungary: 7.7, Russia: 6.8, own calculations based on
GGS data).
The East-West differences in the demographic patterns related to cohabitation
suggest a different normative context in both regions. Among a national represent-
ative sample of the entire population in the countries representing Western Europe
in this study, the majority agrees (strongly) that it would be “alright for two people to
live together without getting married” (Austria: 81 percent, France: 92 percent, Ger-
many: 80 percent, own calculations based on data from the 2008 European Value
Study).
Although norms concerning marriage and cohabitation have also relaxed in
CEE countries (Treas et al. 2014), Stankuniene and Maslauskaite (2008) reported
widespread resistance towards family change. The majority of the population sup-
port traditional patterns of family formation, although there is substantial variation
Transition from Cohabitation to Marriage • 9
across countries. The acceptance of unmarried cohabitation within a national rep-
resentative sample of the whole population also varies across CEE countries and is
lower compared to WE countries. In Hungary 70 percent of the population agrees
that cohabitation is acceptable, in Russia and Bulgaria it is about 61 percent, and
only 32 percent of Georgians approve of cohabitation (own calculations, 2008 EVS
data).
There is an ongoing scholarly debate surrounding the question of whether the
spread of cohabitation in Eastern Europe has more to do with the economic cri-
sis and social anomy than with value change (Aassve et al. 2006; Coleman 2004;
Kalmijn 2007). Cohabitation as a sign of economic disadvantage may have emerged
as a result of people coping with increased economic uncertainty during the shift
from communism to democracy by replacing hard-to-reverse life events (i.e. mar-
riage) with reversible ones (i.e. cohabitation) (Kohler et al. 2006). Hence, starting
to live together without being married may imply a deviation from the behavioural
norm to a lesser extent than it did in the past. However, given the short time horizon
between forming a cohabiting union and institutionalising it through marriage, the
true deviant behaviour may be violating norms by postponing (or foregoing) mar-
riage. Using the same data as in the present study, Hiekel et al. (2014) showed that
more cohabiters in these countries who considered the institution of marriage as
outdated reported plans to marry than their Western European counterparts. This
suggests that cohabiters may conform more readily to predominant and traditional
family (marriage) formation patterns in these countries.
There are also major differences within Europe in the way in which unmarried
cohabitation is treated in the legislation of individual countries. It goes beyond our
knowledge and the scope of the paper to provide an exhaustive overview of the
legal approaches to cohabitation in the different contexts. Broadly speaking, no
country gives completely equal status to marriage and cohabitation in all policy
dimensions, but also no country privileges marriage vis-à-vis cohabitation in all
respects (Perelli-Harris/Sánchez Gassen 2012; Sanchéz Gassen/Perelli-Harris 2015).
Since 1999, cohabiters in France have been able to register their union to ensure
similar rights vis-à-vis their partner compared to married couples. This is not pos-
sible in Austria or Germany (unless for same-sex couples). All three countries (to-
gether with all Western European countries) have abolished laws that discriminated
against children born outside of wedlock and grant equal rights to children from
cohabiting parents (Perelli-Harris/Sánchez Gassen 2012). After the break-up of the
Soviet system, the previously unifi ed legal jurisdiction fragmented. In none of the
former Soviet countries included in our sample is cohabitation considered a conju-
gal union under law (Lithuania and Ukraine are exceptions in the area) and most of
the few efforts to achieve legal harmonisation between cohabitation and marriage
relate to couples with dependent children (Isupova/Perelli-Harris 2012).
• Zuzana Žilinčíková, Nicole Hiekel10
3 Data and Sample
Our sample was drawn from two waves of the Generations and Gender Surveys
(GGS). The GGS is a set of comparative surveys of a nationally representative sam-
ple of the 18-79-year-old resident population in each of the 19 participating coun-
tries (Vikat et al. 2007). In the fi rst wave, the overall size of the samples differed in
each country but in most cases was about 10,000 respondents. The response rate
varied and ranged from 35 percent in Lithuania to 84 percent in Romania (Fokkema
et al. 2016). At the time this study was conducted, data from a second wave col-
lected three years after the fi rst wave were available for ten countries: Austria, Bul-
garia, Czech Republic, France, Georgia, Germany, Hungary, Lithuania, the Nether-
lands and Russia. Three countries were excluded from the sample ‒ the Netherlands
because the data did not contain a measurement for marital attitudes in Wave 1,
Lithuania and the Czech Republic because the marginal prevalence of cohabitation
and a panel attrition rate of 80 (70) percent made meaningful statistical analyses im-
possible. This left us with a selection of seven countries located in Western, Central
and Eastern Europe.
We selected individuals who were cohabiting with a partner of the opposite sex
in Wave 1, were between 18 and 45 years of age in Wave 11 and were observed
roughly three years later in Wave 2 of the panel. Our sample thus did not include
same-sex couples. We had to exclude n=28 respondents for whom consistency
checks based on sex and birth year reported in both waves raised serious doubts
as to whether the same respondent had been re-interviewed. Moreover, n=492 re-
spondents were excluded because they did not report their partnership and marital
status consistently across the waves. It is conceivable that they misremembered
when they started living with their partner, got married or whether they had been
interviewed before or after these events occurred. This makes it particularly unfor-
tunate that, in Wave 2, respondents’ retrospective information on their partnership
situation in Wave 1 had been used during data collection as fi lters for questions
on subsequent marriage or separation. We could not therefore derive any reliable
information on what had happened to these relationships, and dropped them from
the analytical sample. The total number of our analytical sample was n=2,847 co-
habiters from seven European countries.
The sample selection process faced two potential biases. First, sample attrition
can bias the results if it is non-random but related to characteristics of the respond-
ents. Second, cohabiters who marry quickly are less likely to be selected in our
sample. In the following sections we elaborate on these potential sources of bias
and describe how we dealt with them in the analyses.
The panel attrition rate between waves was 32 percent overall; lowest in Geor-
gia and Hungary (less than 20 percent), higher in Austria (26 percent), France and
1 The Hungarian sample comprised only individuals aged 21 years and older. The proportion of
cohabiters under 21 was, however, very low in all other countries (1 percent of the total sample),
meaning that there was no serious underestimation of cohabitation in the Hungarian sample.
Transition from Cohabitation to Marriage • 11
Bulgaria (32 percent) as well as Russia (37 percent), and highest in Germany (72 per-
cent).2 Panel attrition may seriously bias our results, particularly if cohabiters with
positive (negative) attitudes towards the institution of marriage are more likely to
leave the panel during the observation period and therefore be excluded from our
analytical sample. We therefore compared respondents who were cohabiting in
Wave 1 and participated in both waves of data collection ‒ hence our analytical
sample ‒ with their counterparts who dropped out of the panel between the waves
regarding their attitude towards the institution of marriage. The t-test of the dis-
tributions of the marital attitude measure (not shown) revealed that respondents
who dropped out did not differ from respondents who participated in both waves.
Our analytical sample is thus not selective in terms of our variable of interest. We
also examined potential bias due to panel attrition with regard to our control vari-
ables. These analyses showed that higher (compared to lower) educated as well as
employed (compared to unemployed) respondents were more likely to participate
in both waves. Compared to their childless counterparts, cohabiting parents were
also more likely to participate in both waves. Western Europeans were more likely
to drop out of the panel (caused above all by the high panel attrition in the German
survey). Besides our substantial interest in controlling for potentially spurious ef-
fects between these characteristics and the timing of marriage, we also decided to
include these variables in the multivariate analyses for the purpose of controlling for
panel attrition selectivity.
In line with the survey design, we selected those who were not married from all
respondents in a co-residential relationship in Wave 1. Hence, the sample selection
took place at an arbitrary moment in respondents’ relationship trajectory. The vast
majority of respondents between the ages of 18 and 45, however, were married at
Wave 1 and their proportion was higher in Eastern European countries (66.5 per-
cent) compared to Western European countries (33.5 percent). 39 percent (Western
Europe) and 58 percent (Eastern Europe) of the married respondents had cohabited
previously with their spouse. This suggests that cohabiters in our sample might ex-
hibit certain characteristics that made them choose to cohabit instead of marrying
directly on the one hand, and to postpone marriage for longer on the other. We thus
used a Heckman selection model that is described in further detail in the discussion
of our analytical approach.
4 Measurements
Given our event-history analytical framework (elaborated below), the dependent
variable was the hazard rate of transition from cohabitation to marriage between
2 Since the panel attrition rate was extraordinarily high in the German panel data we repeated
these analyses excluding these data. We drew identical conclusions from the model on the
reduced sample and therefore decided to include the German sample despite its high attrition
rate.
• Zuzana Žilinčíková, Nicole Hiekel12
Wave 1 and Wave 2. The question from which we derived the information on mar-
riage formation was posed at Wave 2 and was as follows: “Did you marry your part-
ner after the last interview?” to which respondents answered either “yes” or “no”,
providing the date (month and year) of marriage if applicable.
Marital attitudes were measured at Wave 1 by the following question: “To what
extent do you agree or disagree with the following statement: “Marriage is an out-
dated institution.”? The level of agreement was measured using a fi ve-point Likert
scale. Respondents who answered 1=strongly disagree or 2=disagree were clas-
sifi ed as holding a positive attitude towards marriage and considered it a relevant
social institution in modern times. By contrast, respondents who answered 3=nei-
ther agree nor disagree, 4= agree or 5=strongly agree were classifi ed as holding an
indifferent or negative attitude towards marriage.3
With one exception, the control variables were measured in Wave 1 and were
treated as time-constant (see Table A1 in Appendix). These included plans to get
married (1=yes), as they were found to be positively correlated with transition to
marriage (Manning/Smock 2002); age at start of the union, which refl ects the norm
related to family transitions; union duration and its squared term; gender (1 = fe-
male); level of educational attainment (1= lower secondary or lower, 2 = upper
secondary and post-secondary, 3 = tertiary) which was found to be positively as-
sociated with cohabiters’ transition to marriage in the process of family formation
(Perelli-Harris et al. 2010); employment status (1 = employed, 2 = not employed,
including unemployed, enrolled in education) and ability to make ends meet (scale
ranging from 1 = with great diffi culty to 6= very easily), which are proxies for eco-
nomic deprivation often linked to the postponement of marriage (Clarkberg et al.
1995; Duvander 1999; Gibson-Davis 2009); and previous partnership (1=no previ-
ous partnership, 2=previous cohabitation, 3=previous marriage)4 which is associ-
ated with lower odds of marriage (Lichter/Qian 2008; Steele et al. 2005). Addition-
ally, in Wave 2, respondents reported whether they had had a(nother) child since
Wave 1 and if so, provided the date of birth (month and year). The birth of a child can
be a reason for a cohabiting couple to marry (Perelli-Harris et al. 2012; Perelli-Harris
et al. 2010). We treated this event as a time-varying covariate. The corresponding
dummy variable was given the value 1 nine months prior to the birth to capture the
conception, and retained that value until the event or censoring.
3 We also experimented with the categorical operationalisation of the variable that further distin-
guished between the two negative and the indifferent attitude towards marriage. Results indi-
cated that cohabiters with an indifferent attitude towards marriage had very similar transition
rates to marriage as cohabiters who (strongly) agreed that marriage was an outdated institution
(results available upon request).
4 All respondents who were married to at least one of their prior partners were considered as
previously married even though they might have cohabited with the partner they married or
with a different partner.
Transition from Cohabitation to Marriage • 13
5 Analytical Strategy
Two waves of the GGS allowed us to study the association between marital atti-
tudes and marriage formation among cohabiters from a longitudinal perspective.
Our statistical modelling strategy was based on event-history-analysis because it
allowed us to study the association between a set of independent variables and
the subsequent timing of marriage formation. Specifi cally, a Cox regression model
(Cox 1972) was used to exploit the information of marriage formation, detailed to
the month. Cox regression is an appropriate tool for modelling the discrete-time
data with time-dependent covariates; it allows the inclusion of a correction term for
selection bias; and it can be extended to the data with non-proportional hazard (Al-
lison 1995). Respondents left the risk set at different time points. Respondents who
married between waves left the risk set in the month of marriage. Respondents who
dissolved their union between waves left the risk set in the month of separation and
respondents who were still cohabiting at the end of the observation period were
right-censored at Wave 2.5
As discussed in the data section, marriage-prone respondents may be under-
represented in our analytical sample. Consequently we had to take potential sam-
ple selection into account. Mirroring studies with similar selectivity challenges (e.g.
Billari/Liefbroer 2007) and in line with Heckman (1976), we employed a two-stage
modelling strategy.
The fi rst stage of the analysis assessed potential factors for sample selectiv-
ity, such as already being married in Wave 1. This probit model was applied to
all respondents (n=20,365), excluding those who dropped from the panel between
waves. The predictors in this model were attitudes towards marriage, age, gender,
education, employment, diffi culty making ends meet, and prior co-residential rela-
tionship. In order to avoid collinearity of the correction term and predictors in the
second step of our analyses, we included two additional variables in the selection
model. The fi rst variable indicates the diffusion of cohabitation in the given society
(measured by the standardised proportion of all individuals 18-45 who have ever
cohabited) and thus addresses potential selection into cohabitation versus direct
marriage at the moment of union formation. The second variable, having a joint
5 A potential concern is that specifying a Cox-proportional hazard regression of our event of
interest may lead to an upward bias in the estimator of the event of interest (i.e. exiting co-
habitation via marriage), if the presence of a competing risk (i.e. exiting cohabitation via un-
ion dissolution) is not accounted for. From a theoretical perspective, we argue that predicting
union dissolution goes beyond the focal interest of our study that aims to explain individual
and contextual differences in cohabiters’ marital attitudes and entry into marriage. Theoreti-
cal arguments predicting marriage cannot be applied to union dissolution in a straightforward
manner. From a methodological perspective, we argue that the availability of information on
the timing of union dissolution, detailed to the month, implies that cohabiters who split up no
longer contributed to the risk set which reduces the risk of biased estimates. However, union
dissolution is relatively prevalent, particularly in our Eastern European sample. As a robustness
analysis, we tested our hypotheses on the transition to marriage, specifying union dissolution
as a competing exit route from initial cohabitation (results available upon request). From these
fi ndings we drew identical conclusions regarding our hypotheses.
• Zuzana Žilinčíková, Nicole Hiekel14
biological child with the current partner, was chosen because of the strong empiri-
cal association between marriage and family formation (Musick 2007) and thus ad-
dresses potential selection regarding the timing of marriage. Both measures were
signifi cantly associated with the probability of being selected in our analytical sam-
ple but not with the transition from cohabitation to marriage between the waves.
The second stage of the analysis comprised two Cox regression models which
tested our hypotheses, and both included covariate accounting for the selection
bias, namely the inverse Mill’s ratio derived from the selection model. Tests of mul-
ticollinearity showed that there was no risk of obtaining incorrect estimates. We
used 200 bootstrap replications for each model to account for additional uncer-
tainty stemming from the potential measurement error of the inverse Mill’s ratio.
The fi rst model assessed the independent statistical association between our key
explanatory variables (marital attitude and region) and the cohabiters’ transition to
marriage. The second model included an interaction term between marital attitude
and region in order to test the hypothesis regarding East-West differences in the
effect of marital attitude on marriage transition. We pooled data from the seven
countries. Preliminary country-specifi c Cox regressions (not shown) revealed simi-
lar results in all countries but only had low statistical power due to the small sample
sizes and number of events of interest. The preliminary analyses nevertheless con-
fi rmed our confi dence in the fact that pooling the data and contrasting two broader
regions would not mask important country-specifi c mechanisms. However, as the
descriptive analysis shows, the greater heterogeneity in our sample of cohabiters
from Central and Eastern Europe is mainly driven by Georgia being an outlier on
most indicators. Georgian demographers have suggested that contemporary young
Georgians increasingly tend to postpone or forego legal marriage after a religious
wedding and are thus not covered by marriage statistics based on the Civil Regis-
tration Agency (Badurashvili et al. 2008). Also, marriage status in the harmonised
GGS data refers to legal marriage. Using the pre-harmonised data from Georgia, we
could identify 30 percent of Georgian respondents in the 18-45 age group who were
not legally married to their current partner in Wave 1 but who married that partner
in a religious wedding ceremony. Virtually none of them made the transition to legal
marriage between the waves. Given the ambiguity surrounding marital behaviour in
Georgia, we tested the robustness of our hypotheses-testing models by additionally
specifying the same models but omitting the data from Georgia.
6 Results
Table 1 shows transitions from cohabitation to marriage between Wave 1 and Wave
2. In all, we observed 481 marriages, hence 17 percent of all cohabiters married dur-
ing the observation period. One in seven (WE) and one in eight (CEE) respondents
were still cohabiting at the end of the observation period. 298 cohabiters ended
their relationship between both waves (285 separations, 13 partners died) and were
considered censored in the month of union dissolution. Most marriages occurred
among Austrian cohabiters, of whom one in four married (26,5 percent). The few-
Transition from Cohabitation to Marriage • 15
est marriages occurred among Georgian (5,7 percent) and Bulgarian cohabiters
(7,4 percent). We thus observed a higher share of marriages among cohabiters in
Western Europe than in Central and Eastern Europe. Given the higher prevalence
and acceptance of cohabitation in Western Europe, these results may appear coun-
terintuitive. However, as the selectivity analysis shows, these fi ndings are explained
by CEE cohabiters marrying faster, which decreases their probability of being part
of our analytical sample.
In the following, we present separate descriptive results for each of the countries
in order to gain further insight into the composition of our analytical sample. The
top of Table 2 shows the distribution of our main independent variable, i.e. whether
marriage is considered an outdated institution, as measured in the fi rst interview. In
Austria, France and Germany, around half of all cohabiters disagreed (strongly) with
that statement. In CEE countries, we observed large differences: merely 21 percent
of Bulgarian cohabiters held a positive attitude towards marriage, whereas 83 per-
cent of cohabiters in Georgia did so. In Hungary, this proportion was 36 percent,
and 50 percent in Russia.
At least half of all cohabiters planned to marry in the near future, and their pro-
portion was larger in CEE compared to WE. The overall median duration of cohabi-
tating relationships at the time of the fi rst interview was 56 months. It was longest in
Georgia (79 months) and Bulgaria (62 months) and shortest in Hungary (39 months).
Our sample included more women than men in all countries except Georgia. Most
respondents had had a secondary or post-secondary education (58 percent), fol-
lowed by respondents with tertiary education (26 percent) or lower than secondary
Wave 1 Wave 2
Cohabiting Married Still cohabiting Dissolved union
n n % n % n %
High-cohabitation countries (Western Europe)
Austria 604 160 26.5 383 63.4 61 10.1
France 611 101 16.5 465 76.1 45 7.4
Germany 128 33 25.8 81 63.3 14 10.9
Low-cohabitation countries (Central and Eastern Europe)
Bulgaria 323 24 7.4 280 86.7 19 5.9
Georgia 368 21 5.7 339 92.1 8 2.2
Hungary 475 84 17.7 296 62.3 95 20
Russia 338 58 17.2 224 66.3 56 16.6
Western Europe 1,343 294 21.9 929 69.2 120 8.9
Eastern Europe 1,504 187 12.4 1,139 75.7 178 11.8
Total 2,847 481 16.9 2,068 72.6 298 10.5
Tab. 1: Cohabiters’ relationship transitions during observation period
Source: Generations and Gender Surveys (GGS 2002-2009). Waves 1 and 2, authors’ cal-
culations
• Zuzana Žilinčíková, Nicole Hiekel16
T
a
b
.
2
:
D
e
sc
ri
p
ti
v
e
r
e
su
lt
s
o
f
in
d
e
p
e
n
d
e
n
t
v
a
ri
a
b
le
s
A
U
S
F
R
A
G
E
R
B
U
L
G
E
O
H
U
N
R
U
S
W
E
C
E
E
To
ta
l
M
a
rr
ia
g
e
i
s
a
n
o
u
td
a
te
d
i
n
st
it
u
ti
o
n
(
m
a
ri
ta
l
a
tt
it
u
d
e
)
A
g
re
e
(
st
ro
n
g
ly
)/
n
e
it
h
e
r
a
g
re
e
,
n
o
r
d
is
a
g
re
e
5
7.
8
4
8
.1
4
4
.5
7
9
.3
16
.9
6
3
.6
5
0
.0
5
2
.1
5
2
.5
5
2
.3
D
is
a
g
re
e
(
st
ro
n
g
ly
)
4
2
.2
51
.9
5
5
.5
2
0
.7
8
3
.2
3
6
.4
5
0
.0
4
7.
9
4
7.
5
4
7.
7
In
te
n
ti
o
n
t
o
m
a
rr
y
w
it
h
in
t
h
re
e
y
e
a
rs
Y
e
s
5
0
.0
4
0
.3
4
0
.6
4
5
.2
8
4
.8
6
4
.4
5
2
.7
4
4
.7
6
2
.6
5
4
.2
N
o
5
0
.0
5
9
.7
5
9
.4
5
4
.8
15
.2
3
5
.6
4
7.
3
5
5
.3
3
7.
4
4
5
.8
M
e
d
ia
n
r
e
la
ti
o
n
sh
ip
d
u
ra
ti
o
n
i
n
m
o
n
th
s
5
7.
0
5
8
.0
4
3
.0
6
2
.0
7
9
.0
3
9
.0
4
6
.0
5
6
.0
5
6
.0
5
6
.0
M
a
le
3
9
.6
3
9
.4
4
0
.6
4
5
.5
5
2
.2
4
3
.2
4
0
.5
3
9
.6
4
5
.3
4
2
.6
F
e
m
a
le
6
0
.4
6
0
.6
5
9
.4
5
4
.5
4
7.
8
5
6
.8
5
9
.5
6
0
.4
5
4
.7
5
7.
4
Le
ve
l
o
f
e
d
u
ca
ti
o
n
a
tt
a
in
m
e
n
t
(I
S
C
E
D
)
L
o
w
e
r
se
co
n
d
a
ry
o
r
b
e
lo
w
7.
0
14
.1
13
.3
4
8
.6
7.
6
15
.4
11
.8
10
.8
19
.8
15
.6
S
e
co
n
d
a
ry
,
p
o
st
-s
e
co
n
d
a
ry
71
.7
4
8
.6
5
3
.9
4
0
.9
6
2
.8
6
3
.4
5
7.
4
5
9
.5
5
7.
1
5
8
.2
Te
rt
ia
ry
21
.4
3
7.
3
3
2
.8
10
.5
2
9
.6
21
.3
3
0
.8
2
9
.7
2
3
.1
2
6
.2
E
m
p
lo
ym
e
n
t
E
m
p
lo
y
e
d
9
4
.9
8
8
.4
8
0
.5
5
5
.1
5
3
.5
9
0
.5
81
.4
9
0
.5
71
.8
8
0
.7
N
o
t
e
m
p
lo
y
e
d
5
.1
11
.6
19
.5
4
4
.9
4
6
.5
9
.5
18
.6
9
.5
2
8
.2
19
.4
M
e
a
n
d
if
fi
cu
lt
y
in
m
a
ki
n
g
e
n
d
s
m
e
e
t1
4
.4
3
.3
3
.7
1.
8
2
.4
3
.3
2
.4
3
.8
2
.5
3
.2
C
o
n
ce
p
ti
o
n
o
f
ch
il
d
b
e
tw
e
e
n
t
h
e
w
a
ve
s
3
0
.3
2
6
.7
21
.1
2
4
.5
2
6
.1
2
8
.0
12
.7
2
7.
8
2
3
.3
2
5
.4
P
ri
o
r
u
n
io
n
s
N
o
p
ri
o
r
u
n
io
n
6
4
.4
7
0
.7
6
0
.9
7
6
.5
9
4
.8
7
0
.5
4
9
.4
6
6
.9
7
3
.0
7
0
.1
P
ri
o
r
co
h
a
b
it
a
ti
o
n
2
7.
0
2
0
.8
21
.9
6
.5
1.
9
10
.7
9
.5
2
3
.7
7.
4
15
.1
P
ri
o
r
m
a
rr
ia
g
e
8
.6
8
.5
17
.2
17
.0
3
.3
18
.7
41
.1
9
.4
19
.6
14
.8
To
ta
l n
6
0
4
61
1
12
8
3
2
3
3
6
8
4
7
5
3
3
8
13
4
3
15
0
4
2
8
4
7
A
U
S
=
A
u
st
ri
a;
F
R
A
=
F
ra
n
ce
;
G
E
R
=
G
e
rm
an
y
;
B
U
L=
B
u
lg
ar
ia
;
G
E
O
=
G
e
o
rg
ia
;
H
U
N
=
H
u
n
g
ar
y
;
R
U
S
=
R
u
ss
ia
;
W
E
=
W
e
st
e
rn
E
u
ro
p
e
,
C
E
E
=
C
e
n
tr
al
a
n
d
E
as
te
rn
E
u
ro
p
e
;
1
m
e
as
u
re
d
o
n
s
ca
le
1
(
w
it
h
g
re
at
d
if
fi
cu
lt
y
)
to
6
(
v
e
ry
e
as
ily
)
S
o
u
rc
e
:
G
e
n
e
ra
ti
o
n
s
an
d
G
e
n
d
e
r
S
u
rv
e
y
s,
W
av
e
s
1
a
n
d
2
,
au
th
o
rs
’ c
al
cu
la
ti
o
n
s
Transition from Cohabitation to Marriage • 17
education (16 percent). More cohabiters in WE countries had higher-level education
compared to cohabiters in CEE countries; this difference was most pronounced
in Bulgaria where, together with Georgia, the largest proportion of non-employed
cohabiters could be found. Similarly, on average, CEE cohabiters found it more dif-
fi cult to make ends meet in comparison to WE cohabiters. About a quarter of the
cohabiters conceived a child between Wave 1 and Wave 2, with a slightly higher
proportion of cohabiters doing so in WE countries (28 percent) than in CEE coun-
tries (23 percent). About a quarter of cohabiters had previously lived together with
a partner or a spouse. The proportion of previously cohabiting respondents was
higher in WE countries, whereas the proportion of previously married respondents
was higher in CEE countries, however we again observed great differences among
CEE countries. In sum, the descriptive analyses confi rmed the well-documented
heterogeneity of cohabiters across societies. Also consistent with previous studies
was the fact that cohabiters in the selected Western European countries were some-
what more homogeneous compared to cohabiters in selected Eastern European
countries, with Georgia being an outlier in many of the selected indicators.
Table 3 presents results from a probit model that assessed the selectivity of
our analytical sample. Cohabiters in our sample clearly differ from those who were
already married in Wave 1. Respondents who held a positive attitude towards the in-
stitution of marriage were more likely to be married in Wave 1 and thus less likely to
be included in our analytical sample. Please note that the attitude towards marriage
of these respondents was assessed after they had married. Given the reciprocal as-
sociation between attitudes and behaviour, we could not deduct exposure effects
from the selection and thus have potentially overestimated the difference in co-
habiters’ and spouses’ marital attitudes. Eastern European respondents were more
likely to be married in Wave 1 and thus not part of our analytical sample, refl ecting
the lower prevalence of cohabitation in these countries but also the fact that cohab-
iters may be quicker to transform their union into marriage. Parents, older and less
well educated individuals are more likely already to be married at fi rst interview. Al-
though we have not observed selection in the sample based on employment status,
individuals with diffi culty in making ends meet are more likely to cohabit than be
married in Wave 1. Respondents who had been cohabiting or married prior to the
current relationship are also more likely to be selected, as well as those with shorter
union durations. However, as the positive effect of the squared term of union dura-
tion indicates, the probability of selecting individuals increases at a given (longer)
union duration.
Table 4 shows the results of the second stage of the statistical modelling, namely
Cox regression models, with the timing of marriage as the dependent variable. In all
models, the Inverse Mills Ratio controlling for sample selection bias is statistically
signifi cant. In a fi rst step, we tested the association between an attitude towards
marriage and the transition to marriage. A model containing only marital attitude
(results not shown) revealed that cohabiters who expressed a positive attitude to-
wards marriage had 3.1 higher odds of marrying than cohabiters who were indif-
ferent or negative about the institution of marriage. Model 1 (Table 4) shows the
results of a model that included other predictor variables. The positive association
• Zuzana Žilinčíková, Nicole Hiekel18
between a positive marital attitude and the timing of marriage remained statisti-
cally signifi cant (OR=2.27). Cohabiters who were positive about marriage in Wave
1 were more than twice as likely to be married in Wave 2. This estimate is robust
when controlling for sample selection and other factors associated with the timing
of marriage. This fi nding thus supports Hypothesis 1. Western European cohabiters
had a slightly higher likelihood of entering into marriage than cohabiters from CEE
countries, but this effect was only marginally statistically signifi cant.
We briefl y discuss the effects of other predictor variables on the timing of mar-
riage formation. Plans to get married accounted for a substantial decrease in the
effect of marital attitude on marriage formation. This is because respondents who
were positive about marriage were also more likely to have plans to get married.
Interestingly, marital attitudes were not strongly correlated with marital intention
and instead had their own separate effect. Other factors were also associated with
marriage formation. Individuals with secondary and post-secondary education, or
tertiary education had higher odds of entering marriage than cohabiters with lower
than secondary education but did not differ signifi cantly from each other. Cohabit-
ers who were employed had higher odds of entering marriage compared to their
non-employed counterparts, and similarly, the more easily a respondent was able to
make ends meet, the more likely (s)he was to make the transition to marriage. Prior
cohabitation decreased the odds of transition to marriage; union duration and its
squared term did not infl uence the transition to marriage signifi cantly. Lastly, con-
Tab. 3: Estimates (odds ratio) for a probit model predicting sample selection
(cohabiting instead of married in Wave 1) (N=20,365)
Odds ratio
Positive about marriage (Ref=neutral/negative) 0.48***
Proportion of 18-45 year old population that ever cohabited (standardised) 1.19***
Age 0.99*
Female (Ref=male) 0.98
Education (Ref=lower secondary or below)
secondary, post-secondary 0.90**
tertiary 0.80***
Not employed (Ref=employed) 1.04
Biological child with current partner (Ref=no) 0.53***
Diffi culty in making ends meet1 0.95***
Prior co-resident union (Ref=no) 1.47***
Union duration 0.91***
Union duration (squared) 1.00*
Constant 3.07***
***p<.001 **p<.01 *p<.05 +p<0.1
1 measured on scale 1 (with great diffi culty) to 6 (very easily)
Source: Generations and Gender Surveys, Waves 1 and 2, authors’ calculations
Transition from Cohabitation to Marriage • 19
T
a
b
.
4:
R
e
su
lt
s
(o
d
d
s
ra
ti
o
s)
f
ro
m
C
o
x
r
e
g
re
ss
io
n
m
o
d
e
ls
p
re
d
ic
ti
n
g
e
n
tr
y
in
to
m
a
rr
ia
g
e
b
e
tw
e
e
n
W
av
e
1
a
n
d
W
av
e
2
,
w
it
h
c
o
n
tr
o
ls
f
o
r
sa
m
p
le
s
e
le
c
ti
o
n
a
n
d
w
it
h
b
o
o
ts
tr
a
p
p
a
ra
m
e
te
r
e
st
im
a
te
s
F
u
ll
sa
m
p
le
S
a
m
p
le
e
x
cl
u
d
in
g
G
e
o
rg
ia
M
o
d
e
l 1
M
o
d
e
l 2
M
o
d
e
l 3
M
o
d
e
l 4
P
o
si
ti
ve
a
b
o
u
t
m
a
rr
ia
g
e
2
.2
7
**
*
2
.6
1
**
*
2
.1
0
**
*
2
.7
2
**
*
L
iv
in
g
i
n
W
e
st
e
rn
E
u
ro
p
e
(
R
e
f=
E
a
st
e
rn
E
.)
1.
2
7
+
1.
4
4
*
1.
11
1.
4
0
*
In
te
ra
ct
io
n
p
o
si
ti
ve
a
tt
it
u
d
e*
W
e
st
0
.8
1
0
.6
7
*
In
te
n
ti
o
n
t
o
m
a
rr
y
(R
e
f=
n
o
)
3
.1
0
**
*
3
.1
2
**
*
3
.1
9
**
*
3
.2
3
**
*
U
n
io
n
d
u
ra
ti
o
n
1.
0
0
1.
0
0
1.
0
0
1.
0
0
U
n
io
n
d
u
ra
ti
o
n
(
sq
u
a
re
d
)
1.
0
0
1.
0
0
1.
0
0
1.
0
0
A
g
e
a
t
th
e
u
n
io
n
f
o
rm
a
ti
o
n
0
.9
7
*
0
.9
7
*
0
.9
8
*
0
.9
8
*
F
e
m
a
le
(
R
e
f=
m
a
le
)
1.
13
1.
14
1.
0
8
1.
0
9
E
d
u
ca
ti
o
n
(
R
e
f=
lo
w
e
r
se
co
n
d
a
ry
o
r
b
e
lo
w
)
1.
0
0
1.
0
0
1.
0
0
1.
0
0
se
co
n
d
a
ry
,
p
o
st
-s
e
co
n
d
a
ry
1.
8
6
**
1.
8
3
**
2
.0
1
**
*
1.
9
7
**
te
rt
ia
ry
2
.4
3
**
*
2
.4
1
**
*
2
.6
5
**
*
2
.6
2
**
*
N
o
t
e
m
p
lo
ye
d
(
R
e
f=
e
m
p
lo
y
e
d
)
0
.5
8
**
0
.5
8
**
*
0
.6
4
*
0
.6
4
*
D
if
fi
cu
lt
y
in
m
a
ki
n
g
e
n
d
s
m
e
e
t1
1.
0
9
*
1.
0
9
*
1.
0
7
+
1.
0
7
+
C
o
n
ce
p
ti
o
n
o
f
ch
il
d
b
e
tw
e
e
n
t
h
e
w
a
ve
s
(R
e
f=
n
o
)
1.
6
6
**
*
1.
6
6
**
*
1.
6
8
**
*
1.
6
8
**
*
P
ri
o
r
u
n
io
n
s
(R
e
f=
n
o
p
ri
o
r
u
n
io
n
)
1.
0
0
1.
0
0
1.
0
0
1.
0
0
p
re
v
io
u
sl
y
c
o
h
a
b
it
in
g
0
.7
2
*
0
.7
2
*
0
.7
6
+
0
.7
5
+
p
re
v
io
u
sl
y
m
a
rr
ie
d
1.
0
2
1.
0
2
0
.9
9
0
.9
8
In
ve
rs
e
M
il
ls
’ R
a
ti
o
0
.4
0
**
*
0
.3
9
**
*
0
.6
3
*
0
.6
1
*
n
2
,8
4
7
2
,4
7
9
**
*p
<
.0
0
1
*
*p
<
.0
1
*
p
<
.0
5
+
p
<
0
.1
1
m
e
as
u
re
d
o
n
s
ca
le
1
(
w
it
h
g
re
at
d
if
fi
cu
lt
y
)
to
6
(
v
e
ry
e
as
ily
)
S
o
u
rc
e
:
G
e
n
e
ra
ti
o
n
s
an
d
G
e
n
d
e
r
S
u
rv
e
y
s,
W
av
e
s
1
a
n
d
2
,
au
th
o
rs
’ c
al
cu
la
ti
o
n
s
• Zuzana Žilinčíková, Nicole Hiekel20
ceiving a child between waves had a signifi cant effect on the transition to marriage,
confi rming the expected link between timing of childbearing and marriage.
Model 2 (Table 4) includes an interaction term between marital attitude and re-
gion in order to test Hypothesis 2 on expected country differences in the associa-
tion between the marital attitude and marriage formation. With this interaction term
included, the effect of a positive attitude towards marriage increasing the risk of
marriage by 2.61 applies to Eastern European cohabiters who were thus more than
twice as likely to marry compared to their counterparts who held an indifferent
or negative attitude towards marriage. The effect of the interaction term indicates
that this positive association between marital attitude and transition to marriage
was reduced by 0.81 in Western European countries. This effect is, however, not
statistically signifi cant. The effect of region indicates that the odds of marriage for
Western European cohabiters increased on average by 1.44, which is not explained
by differences in their average marital attitude. This effect may refl ect the stronger
recuperation of postponed marriages among Western European cohabiters that we
discussed in the selectivity analysis.
Given the particularities in the composition of the Georgian population of co-
habiters we estimated the same two models omitting the data from Georgia (Table
4, Model 3 and Model 4). These analyses lead to almost identical results in Model
3 (i.e., the equivalent to Model 1), thus providing support for Hypothesis 1. Differ-
ing from the fi ndings in the full sample, Model 4 (i.e., the equivalent to Model 2)
reveals that the negative interaction between marital attitude and region regarding
entry into marriage is statistically signifi cant at the 0.05 level in the analysis omit-
ting Georgia. This fi nding suggests that the link between a positive attitude towards
marriage and marriage formation is weaker in Western Europe compared to Eastern
Europe. Hence, neither the fi nding from the full model nor the model omitting Geor-
gia support Hypothesis 2 in which we formulated the expectation that the attitude-
behaviour link would be weaker in Eastern Europe because of the stronger societal
norms regarding getting married that push cohabiters into marriage beyond their
own attitudes towards the institution.
We reject Hypothesis 2 based on the direction of the effect; this points actually
to a weaker association between marital attitudes and transition to marriage among
Western European cohabiters. Based on these analyses we draw two conclusions.
First, positive attitudes towards the institution of marriage were associated with
higher odds of entering marriage. Second, although with great caution because the
fi nding is not robust when Georgia is included in the Eastern European sample of
cohabiters, the strength of the association between marital attitude and entry into
marriage is weaker in Western European countries characterised by high diffusion
of cohabitation and less normative pressure to transform cohabitation into a mar-
ried union.
Transition from Cohabitation to Marriage • 21
7 Conclusion and Discussion
This study began by arguing that a comparative perspective on the marriage trajec-
tories of established cohabiting relationships deserves greater academic interest.
We argued that the merit of theoretical considerations derived from the SDT frame-
work lies in the attention that is drawn to attitudinal change and its intertwining with
social norms in explaining changes and variation in demographic behaviour related
to union formation. We reviewed country-specifi c literature which showed that co-
habitation in some countries had already become the predominant way of start-
ing a co-resident relationship, whereas in other countries, onset of the diffusion of
cohabitation had been more recent. Related to the stage of diffusion of cohabita-
tion are predominant social norms about the importance of getting married. The
SDT framework suggests cross-national differences in the proportion of cohabiters
holding (dis)approving attitudes about the institution of marriage as well as variation
in the strength of social norms about marriage. We derived two hypotheses from
these theoretical considerations. We expected fi rst, that marital attitudes are as-
sociated with cohabiters’ transition to marriage (H1), and second, that the strength
of this association is weaker in Central and Eastern European countries with a low
prevalence of cohabitation compared to Western European countries with a high
prevalence (H2). Using panel data from two waves of the Generations and Gender
Surveys, we compared two European regions comprised of pooled data from seven
European countries and examined cohabiters’ risk of marriage as a function of the
extent to which they consider marriage an outdated institution, along with other
socio-demographic and economic predictors associated with marriage formation.
Our fi rst key fi nding was that, in line with studies that have investigated the se-
lection of individuals into either cohabitation or marriage as the start of their union,
cohabiters with a favourable attitude towards marriage were more likely to marry
than cohabiters with a negative attitude. Though the present study only used one
indicator, it was, in our view, a particularly interesting attitudinal measure for our
research purpose. A cohabiter’s (dis)agreement with a statement such as “Marriage
is an outdated institution” indicates cohabiters’ motives for living in a non-marital
union and (future) relationship expectations but also refl ects societal norms about
marriage.
Our second key fi nding was that we identifi ed country differences in the pro-
portion of cohabiters who agreed with the statement that marriage is an outdated
social institution. Agreement with that statement was generally higher among co-
habiters in three of the four Central and Eastern European countries compared to
cohabiters in the three Western European countries we examined. In line with the
SDT framework, this suggests that in countries where cohabitation is less preva-
lent and the cohabiting population is more selective, cohabiters predominantly hold
negative attitudes towards marriage, as such attitudes may have selected them
into cohabitation in the fi rst place. However, the underrepresentation of Central
and Eastern European cohabiters in our analytical sample due to their transition
to marriage prior to the fi rst interview (Table 3) also suggests that many cohabit-
ers in CEE countries have internalised widely shared ideas about the superiority of
• Zuzana Žilinčíková, Nicole Hiekel22
marriage over cohabitation (Mureşan 2008; Sobotka 2008). Although more people
are increasingly opting for cohabitation, cohabitation is largely viewed as inferior
to marriage, and childbearing remains closely linked to the transition to marriage,
particularly for individuals with greater socio-economic resources (Haragus 2015).
Moreover, we also found larger cross-country diversity in Central and Eastern Eu-
rope compared to Western Europe in the level of (dis)agreement with the statement
that marriage is outdated. This greater country variation within Central and Eastern
Europe is usually explained by different challenges in the transition from state so-
cialism to democracy and a market economy. Bulgaria and Romania experienced
a more profound economic collapse and particularly slow recuperation with great
social inequalities (Haragus 2015; Thornton/Philipov 2009). This suggests that entry
into cohabitation in these countries is associated with economic constraints and
particularly long-term cohabitating unions may be understood as a “poor man’s
marriage” for those who cannot afford to marry (Oppenheimer 1988, 2003).
Our third and main fi nding was that the multivariate analyses did not reveal the
context dependencies of the association between marital attitudes and marriage
formation that we expected to fi nd. The interaction effect provided no evidence
that association between marital attitudes and transition to marriage is weaker in
Central and Eastern Europe where cohabitation is less prevalent and less accepted
because of cohabiters’ exposure to strong social pressure to get married, irrespec-
tive of their personal attitudes towards the institution of marriage. On the contrary,
the direction of the interaction effect suggests that the association between marital
attitudes and marriage formation is actually weaker in Western Europe where co-
habitation is more common. This effect does, however, reach statistical signifi cance
in a model that omitted the Georgian sample. As readers will recall, the Georgian
data exhibited a large number of particularities in the composition of the cohabiting
population and the transition rate to legal marriage compared to other (Eastern) Eu-
ropean countries in our sample. As noted earlier, research from Georgia suggested
a stronger reluctance among young Georgians to register a church wedding with
the legal authorities (Badurashvili et al. 2008). The fact that the country comparative
GGS surveys consider legal marriage (and not religious marriages) means that they
may not fully refl ect actual marital behaviour and the meaning of marriage in the
Georgian sample.
We can envisage three reasonable explanations for the weaker association be-
tween marital attitudes and transition to marriage among cohabiters in Western
compared to Eastern Europe and invite future research to address them. First, this
result might indicate that Western European cohabiters – even those with positive
attitudes towards marriage – can postpone marrying without being stigmatised.
Second, the virtual absence of strong social pressure to transform a cohabiting into
a marital union could increase the importance of “pragmatic” reasons to marry, such
as tax benefi ts or the purchase of residential property given that in many Western
European countries, cohabiters and spouses are treated similarly in terms of taxa-
tion, property and inheritance rights (Perelli-Harris/Sánchez Gassen 2012). Finally,
in such contexts, the relationship history itself may be more predictive of the entry
to marriage than people’s general attitudes towards the institution of marriage. For
Transition from Cohabitation to Marriage • 23
instance, the conception of a child was more strongly associated with entry into
marriage among Western European cohabiters compared to their Eastern European
counterparts (results not shown). Ideally, measures which reveal how important re-
spondents view such “pragmatic” reasons or life events for their decision to marry
would allow explicit testing of such mechanisms. Unfortunately, these measures
have not been included in the GGS and we are not aware of any longitudinal, coun-
try comparative survey which collects both attitudinal measures and demographic
outcomes that includes such measures.
Of course, there are certain limitations to this study. Only a minority of our co-
habiting sample entered into marriage during the observation period of roughly
three years. Keeping the overrepresentation of long-term cohabiting unions in our
analytical sample in mind, this indicates the increasing longevity of cohabiting un-
ions, predominantly driven by a postponement of marriage to later in life. It is not
caused by the cohabiters not being observed for long enough. The median duration
of the relationship was fi ve years at the fi rst interview and we recorded transitions
to marriage during the following three years. We identifi ed stark differences in co-
habiters’ timing of marriage formation that are explained by differences in their
approval of the institution of marriage. We feel that this fi nding underlines the im-
portance of taking marital attitudes into account in order to refl ect heterogeneity in
cohabiters’ marital trajectories.
We are aware that pooling data from countries in two broad European regions
masks the diversity of cohabitation patterns within both regions that have been em-
phasised in the literature (Sobotka 2003; Sobotka/Toulemon 2008) and have been
confi rmed in our descriptive analyses. However, small sample sizes in some of the
countries in this study prevented us from conducting statistically meaningful analy-
ses for separate countries, and the number of countries did not permit the applica-
tion of a multilevel analytical framework either. Future analyses should include more
European countries to see whether the diversity within Eastern Europe observed
in the current study was predominantly due to the limited and non-random selec-
tion of countries for which we had suitable data available. A multilevel analytical
framework would also allow testing of the infl uence of country-level indicators in
the normative and economic sphere (e.g. unemployment, housing market charac-
teristics, legal arrangements) that reveal such contextual effects and formally test
whether they explain country differences in the marriage transition among cohabit-
ers. Moreover, a study on regional differences within countries could reveal differ-
ences beyond national borders.
Acknowledgements
Zuzana Žilinčíková’s contribution to this research received fi nancial support from
the Czech Science Foundation (project num. 17-18235S).
The authors would like to thank Tom Emery for his valuable advice on an earlier
version of this manuscript and express their gratitude to the editor and two anony-
mous reviewers for their constructive comments and suggestions during the review
process.
• Zuzana Žilinčíková, Nicole Hiekel24
References
Aassve, Arnstein; Iacovou, Maria; Mencarini, Letizia 2006: Youth poverty and transition
to adulthood in Europe. In: Demographic Research 15,2: 21-49 [doi: 10.4054/Dem-
Res.2006.15.2].
Ajzen, Icek 1988: Attitudes, Personality and Behavior. Milton Keynes: Open University
Press.
Ajzen, Icek 1991: The theory of planned behavior. In: Organizational Behavior and Hu-
man Decision Processes 50,2: 179-211 [doi: 10.1016/0749-5978(91)90020-T].
Allison, Paul D. 1995: Survival analysis using SAS: A practical guide. Cary, NC: SAS
Institute.
Axinn, William G.; Thornton, Arland 1992: The Relationship between cohabitation
and divorce: Selectivity or causal infl uence. In: Demography 29,3: 357-374 [doi:
10.2307/2061823].
Badurashvili, Irina et al. 2008: Gender relations in modern Georgian society. Tbilisi.
Beck, Ulrich; Beck-Gernsheim, Elisabeth 2002: Individualization: Instituitionalized Indi-
vidualism and Its Social and Political Consequences. Thousand Oaks, CA: Sage.
Bernhardt, Eva 2004: Cohabitation or marriage? Preferred living arrangements in Swe-
den. In: sdf puzzle 04: 1-10.
Billari, Francesco C.; Liefbroer, Aart C. 2007: Should I stay or should I go? The im-
pact of age norms on leaving home. In: Demography 44,1: 181-198 [doi: 10.1353/
dem.2007.0000].
Billari, Francesco C.; Liefbroer, Aart C. 2010: Towards a new pattern of transition to
adulthood? In: Advances in Life Course Research 15,2-3: 59-75 [doi: 10.1016/j.
alcr.2010.10.003].
Blumer, Herbert 1969: Symbolic interactionism: Perspective and method. Berkeley: Uni-
versity of California Press.
Bradatan, Cristina 2012: Introduction to the Special Issue Marriage as a System: Insights
from the Hajnal’s Line and Current “East-European” Family Patterns. In: Journal of
Comparative Family Studies 43,3.
Bradatan, Cristina; Kulcsar, Laszlo 2008: Choosing between marriage and cohabitation:
Women’s fi rst union patterns in Hungary. In: Journal of Comparative Family Studies
39,4: 491-507.
Buchmann, Marlis 1989: The script of life in modern society. Entry into adulthood in a
changing world. Chicago: University of Chicago Press.
Cherlin, Andrew J. 2004: The deinstitutionalization of American marriage. In: Journal of
Marriage and Family 66,4: 848-861 [doi: 10.1111/j.0022-2445.2004.00058.x].
Clarkberg, Marin; Stolzenberg, Ross M.; Waite, Linda J. 1995: Attitudes, values, and en-
trance into cohabitational versus marital unions. In: Social Forces 74,2: 609-632 [doi:
10.1093/sf/74.2.609].
Coleman, David 2004: Why we don’t have to believe without doubting in the “Second
Demographic Transition”– some agnostic comments. In: Vienna Yearbook of Popula-
tion Research. Vienna: Vienna Institute of Demography.
Cox, David R. 1972: Regression Models and Life-Tables. In: Journal of the Royal Statisti-
cal Society. Series B (Methodological) 34,2: 187-220.
Transition from Cohabitation to Marriage • 25
Duvander, Ann-Zofi e E. 1999: The Transition from Cohabitation to Marriage: A Longi-
tudinal Study of the Propensity to Marry in Sweden in the Early 1990s. In: Journal of
Family Issues 20: 698-717 [doi: 10.1177/019251399020005007].
Fokkema, Tineke et al. 2016: Generations and Gender Programme Wave 1 data collec-
tion: An overview and assessment of sampling and fi eldwork methods, weighting pro-
cedures, and cross-sectional representativeness. In: Demographic Research 34,18:
499-524 [doi: 10.4054/DemRes.2016.34.18].
Gerber, Theodore P.; Berman, Danielle 2010: Entry to marriage and cohabitation in Rus-
sia, 1985-2000: Trends, correlates, and implications for the Second Demographic Tran-
sition. In: European Journal of Population 26,1: 3-31 [doi: 10.1007/s10680-009-9196-8].
Gibson-Davis, Christina 2009: Money, marriage, and children: Testing the fi nancial ex-
pectations and family formation theory. In: Journal of Marriage and Family 71,1: 146-
161 [doi: 10.1111/j.1741-3737.2008.00586.x].
Hajnal, John 1965: European marriage pattern in historical perspective. In: Glass, David
Victor; Eversley, David Edward Charles (Eds.): Population in history. London: Arnold.
Haragus, Mihaela 2015: From Cohabitation to marriage when a child is on the way. A
comparison of three former socialist countries: Romania, Bulgaria and Hungary. In:
Journal of Comparative Family Studies 46,3: 329-350.
Heckman, James J. 1976: The common structure of statistical models of truncation,
sample selection, and limited dependent variables and a simple estimator for models.
In: Annals of Economic and Social Measurement 5: 475-492.
Heuveline, Patrick; Timberlake, Jeffrey M. 2004: The role of cohabitation in family for-
mation: The United States in comparative perspective. In: Journal of Marriage and the
Family 67,2: 1214-1230 [doi: 10.1111/j.0022-2445.2004.00088.x].
Hiekel, Nicole; Liefbroer, Aart C.; Poortman, Anne-Rigt 2014: Understanding diversity in
the meaning of cohabitation across Europe. In: European Journal of Population 30,4:
391-410 [doi: 10.1007/s10680-014-9321-1].
Hoem, Jan M. et al. 2009: Traces of the second demographic transition in four selected
countries in Central and Eastern Europe: Union formation as a demographic manifes-
tation. In: European Journal of Population 25,3: 239-255 [doi: 10.1007/s10680-009-
9177-y].
Isupova, Olga; Perelli-Harris, Brienna 2012: Cohabitation, legal policies, and cultural
systems: divergence across post-Soviet space. European Population Conference.
Stockholm, Sweden.
Kalmijn, Matthijs 2007: Explaining cross-national differences in marriage, cohabita-
tion, and divorce in Europe, 1990-2000. In: Population Studies 61,3: 243-263 [doi:
10.1080/00324720701571806].
Kennedy, Sheela; Bumpass, Larry L. 2008: Cohabitation and children’s living arrange-
ments: New estimates from the United States. In: Demographic Research 19: 1663-
1692 [doi: 10.4054/DemRes.2008.19.47].
Kohler, Hans-Peter; Billari, Francesco C.; Ortega, José Antonio 2006: Low fertility in
Europe: Causes, implications and policy options. In: Harris, Fred R. (Ed.): The Baby
Bust: Who will do the work? Who will pay the taxes? Lanham, MD: Rowman & Little-
fi eld Publishers.
Kok, Jan; Leinarte, Dalia 2015: Cohabitation in Europe: a revenge of history. In: The His-
tory of the Family 20,4: 489-514 [doi: 10.1080/1081602X.2015.1067240].
• Zuzana Žilinčíková, Nicole Hiekel26
Kuo, Janet Chen-Lan; Raley, Kelly R. 2016: Diverging patterns of union transition among
cohabitors by race/ethnicity and education: Trends and marital intentions in the Unit-
ed States. In: Demography 53,4: 921-935 [doi: 10.1007/s13524-016-0483-9].
Lesthaeghe, Ron 1995: The Second Demographic Transition in Western countries: An
interpretation. In: Mason, Karen O.; Jensen, An-Magritt (Eds.): Gender and family
change in industrialized countries. Oxford: Clarendon Press.
Lesthaeghe, Ron 2010: The unfolding story of the Second Demographic Transi-
tion. In: Population and Development Review 36,2: 211-251 [doi: 10.1111/j.1728-
4457.2010.00328.x].
Lichter, Daniel T.; Qian, Zhenchao C. 2008: Serial cohabitation and the marital life
course. In: Journal of Marriage and Family 70,4: 861-878 [doi: 10.1111/j.1741-
3737.2008.00532.x].
Lichter, Daniel T.; Qian, Zhenchao C.; Mellott, Leannam M. 2006: Marriage or dissolu-
tion? Union transitions among poor cohabiting women. In: Demography 43,2: 223-240
[doi: 10.1353/dem.2006.0016].
Liefbroer, Aart C. 1991: The Choice between a Married or Unmarried First Union by
Young Adults: A Competing Risk Analysis. In: European Journal of Population 7,3:
273-298.
Liefbroer, Aart C.; Billari, Francesco C. 2010: Bringing norms back in: A theoretical and
empirical discussion of their importance for understanding demographic behaviour.
In: Population, Space and Place 16,4: 287-305 [doi: 10.1002/psp.552].
Liefbroer, Aart C.; Gerritsen, Liesbet; De Jong Gierveld, Jenny 1994: The infl uence of
intentions and life course factors on union formation behavior of young adults. In:
Journal of Marriage and Family 56,1: 193-203 [doi: 10.2307/352713].
Manlove, Jennifer et al. 2012: Union transitions following the birth of a child to cohabit-
ing parents. In: Population research and policy review 31,3: 361-386 [doi: 10.1007/
s11113-012-9231-z].
Manning, Wendy D.; Smock, Pamela J. 2002: First comes cohabitation and then
comes marriage? A research note. In: Journal of Family Issues 23,8: 1065-1087 [doi:
10.1177/019251302237303].
Manning, Wendy D.; Smock, Pamela J. 1995: Why marry? Race and the transition to
marriage among cohabitors. In: Demography 32,4: 509-520 [doi: 10.2307/2061671].
Manting, Dorien 1996: The changing meaning of cohabitation and marriage. In: Euro-
pean Sociological Review 12,1: 53-65 [doi: 10.1093/oxfordjournals.esr.a018177].
McClain, Lauren Rinelli 2011: Better parents, more stable partners: Union transitions
among cohabiting parents. In: Journal of Marriage and Family 73,5: 889-901 [doi:
10.1111/j.1741-3737.2011.00859.x].
Moors, Guy; Bernhardt, Eva 2009: Splitting up or getting married? Competing risk analy-
sis of transitions among cohabiting couples in Sweden. In: Acta Sociologica 52,3: 227-
247 [doi: 10.1177/0001699309339800].
Mureşan, Cornelia 2008: Cohabitation, an alternative for marriage in contemporary Ro-
mania: A life table description. In: Demográfi a English Edition 51,5: 36-65.
Musick, Kelly 2007: Cohabitation, nonmarital childbearing, and the marriage process. In:
Demographic Research 16,9: 249-286 [doi: 10.4054/DemRes.2007.16.9].
Neugarten, Bernice L.; Moore, Joan W.; Lowe, John C. 1965: Age Norms, Age Con-
straints, and Adult Socialization. In: American Journal of Sociology 70,6: 710-717 [doi:
10.1086/223965].
Transition from Cohabitation to Marriage • 27
Oppenheimer, Valerie Kincade 1988: A theory of marriage timing. In: American Journal
of Sociology 94,3: 563-591 [doi: 10.1086/229030].
Oppenheimer, Valerie Kincade 2003: Cohabiting and marriage during young men’s career-
development process. In: Demography 40,1: 127-149 [doi: 10.1353/dem.2003.0006].
Ortyl, Timothy A. 2013: Long-term heterosexual cohabiters and attitudes towards mar-
riage. In: The Sociological Quarterly 54,4: 584-609 [doi: 10.1111/tsq.12043].
Perelli-Harris, Brienna; Bernardi, Laura 2015: Exploring social norms around cohabita-
tion: the life course, individualization, and culture. In: Demographic Research 33,25:
701-732 [doi: 10.4054/DemRes.2015.33.25].
Perelli-Harris, Brienna; Lyons-Amos, Mark 2015: Changes in partnership patterns across
the life course: An examination of 14 countries in Europe and the United States. In:
Demographic Research 33,6: 145-178 [doi: 10.4054/DemRes.2015.33.6].
Perelli-Harris, Brienna; Sánchez Gassen, Nora 2012: How similar are cohabitation and
marriage? Legal approaches to cohabitation across Western Europe. In: Population
and Development Review 38,3: 435-467 [doi: 10.1111/j.1728-4457.2012.00511.x].
Perelli-Harris, Brienna et al. 2010: The educational gradient of childbearing within co-
habitation in Europe. In: Population and Development Review 36,4: 775-801 [doi:
10.1111/j.1728-4457.2010.00357.x].
Perelli-Harris, Brienna et al. 2012: Changes in union status during the transition to par-
enthood in eleven European countries, 1970s to early 2000s. In: Population Studies
66,2: 167-182 [doi: 10.1080/00324728.2012.673004].
Philipov, Dimiter; Dobritz, Jürgen 2003: Demographic consequences of economic tran-
sition in countries of central and Eastern Europe. Strasbourg: Council of Europe.
Philipov, Dimiter; Jasilioniene, Aiva 2008: Union formation and fertility in Bulgaria and
Russia: A life table description of recent trends. In: Demographic Research 19,62:
2057-2114 [doi: 10.4054/DemRes.2008.19.62].
Puur, Allan et al. 2012: Transformation of partnership formation in Eastern Europe: The
legacy of the past demographic divide. In: Journal of Comparative Family Studies
43,3: 389-417.
Sanchéz Gassen, Nora; Perelli-Harris, Brienna 2015: The increase in cohabitation and
the role of union status in family policies: A comparison of 12 European countries.
In: Journal of European Social Policy 25,4: 431-449 [doi: 10.1177/0958928715594561].
Sassler, Sharon 2004: The process of entering into cohabiting unions. In: Journal of
Marriage and Family 66,2: 491-505 [doi: 10.1111/j.1741-3737.2004.00033.x].
Settersten, Richard A. Jr.; Mayer, Karl Ulrich 1997: The Measurement of Age, Age Struc-
turing, and the Life Course. In: Annual Review of Sociology 23,1: 233-261 [doi: 10.1146/
annurev.soc.23.1.233].
Smock, Pamela J.; Manning, Wendy D.; Porter, Meredith 2005: “Everything’s there ex-
cept money”: How money shapes decisions to marry among cohabitors. In: Journal of
Marriage and Family 67,3: 680-696 [doi: 10.1111/j.1741-3737.2005.00162.x].
Sobotka, Tomáš 2008: Overview Chapter 6: The diverse faces of the Second Demo-
graphic Transition in Europe. In: Demographic Research 19,8: 171-224 [doi: 10.4054/
DemRes.2008.19.8].
Sobotka, Tomáš 2003: Re-emerging diversity: Rapid fertility changes in Central and
Eastern Europe after the collapse of the communist regimes. In: Population and De-
velopment Review 58,4-5: 451-486.
• Zuzana Žilinčíková, Nicole Hiekel28
Sobotka, Tomáš; Toulemon, Laurent 2008: Changing family and partnership behaviour:
Common trends and persistent diversity across Europe. In: Demographic Research
19,6: 85-138 [doi: 10.4054/DemRes.2008.19.6].
Stankuniene, Vlada; Maslauskaite, Ausra 2008: Family transformations in the post-com-
munist countries: Attitudes towards changes. In: Höhn, Charlotte; Avramov, Dragana;
Kotowska, Irene E. (Eds.): People, population change and policy acceptance study.
Berlin: Springer.
Stanley, Scott M.; Rhoades, Galena Kline; Markman, Howard J. 2006: Sliding vs. decid-
ing: Inertia and the premarital cohabitation effect. In: Family Relations 55: 499-509.
Steele, Fiona et al. 2005: The relationship between childbearing and transitions from
marriage and cohabitation in Britain. In: Demography 42,4: 647-673 [doi: 10.1353/
dem.2005.0038].
Surkyn, John J.; Lesthaeghe, Ron 2004: Value orientations and the second demographic
transition (SDT) in Northern, Western and Southern Europe: An update. In: Demo-
graphic Research 3,3: 47-86 [doi: 10.4054/DemRes.2004.S3.3].
Thornton, Arland; Philipov, Dimiter 2009: Sweeping changes in marriage, cohabitation,
and childbearing in Central and Eastern Europe: New insights from the Developmental
Idealism framework. In: European Journal of Population 25,2: 123-156 [doi: 10.1007/
s10680-009-9181-2].
Treas, Judith; Lui, Jonathan; Gubernskaya, Zoya 2014: Attitudes on marriage and new
relationships: Cross-national evidence on the deinstitutionalization of marriage. In:
Demographic Research 30,54: 1495-1526 [doi: 10.4054/DemRes.2014.30.54].
van de Kaa, Dirk J. 1993: Second demographic transition revisited: Theories and expec-
tations. In: Beets, Gijs C.N. (Ed.): Population and family in the Low Countries 1993.
Lisse: Swets and Zeitungen.
van de Kaa, Dirk J. 1987: Europe’s Second Demographic Transition. In: Population Bul-
letin 42,1: 1-59.
Vikat, Andres et al. 2007: Generation and Gender Survey: Towards a better understand-
ing of relationships and processes in the life course. In: Demographic Research 17,14:
389-440 [doi: 10.4054/DemRes.2007.17.14].
Date of submission: 27.06.2017 Date of acceptance: 06.03.2018
Zuzana Žilinčíková (). Masaryk University. Brno, Czech Republic.
E-mail: zilincikova@mail.muni.cz
URL: https://www.muni.cz/en/people/333003-zuzana-zilincikova
Dr. Nicole Hiekel. University of Cologne, Germany. E-mail: hiekel@wiso.uni-koeln.de
URL: http://www.iss-wiso.uni-koeln.de/institut/personen/h/dr-nicole-hiekel/
Netherlands Interdisciplinary Demographic Institute (NIDI). The Hague, The Netherlands
URL: https://www.nicole-hiekel.com
Transition from Cohabitation to Marriage • 29
Appendix
Tab. A1: Descriptive statistics for variables used in the analyses
Cases Not Selected Cases Selected
Married in Wave 1 Cohabiting in Wave 1
(n=17,518) (n=2,847)
Percentage Mean (SD) Percentage Mean (SD)
Marital attitudes
Neutral/negative 23.4 52.3
Positive 76.6 47.7
Geographical location
Living in Eastern Europe 66.2 52.8
Living in Western Europe 33.9 47.2
Intention to marry
yes n.a. 54.2
no n.a. 45.8
Union duration in months at Wave 1 157.0 (79.8) 73.0 (60.5)
Age 36.1 (6.1) 31.7 (6.5)
Male 38.8 42.6
Female 61.2 57.4
Level of education
Lower secondary or below 12.1 15.6
Secondary, post-secondary 59.7 58.2
Tertiary 28.2 26.2
Employment
Employed 78.3 80.7
Not employed 21.8 19.4
Diffi culty in making ends meet1 3.0 (1.3) 3.2 (1.4)
Biological child with current partner 87.4 50.0
Prior co-resident union 9.6 29.9
Note: n.a. = not applicable
1 measured on scale 1 (with great diffi culty) to 6 (very easily)
Source: Generations and Gender Surveys, Waves 1 and 2, authors’ calculations
Published by
Prof. Dr. Norbert F. Schneider
Federal Institute for Population Research
D-65180 Wiesbaden / Germany
2018
Managing Editor
Dr. Katrin Schiefer
Copy Editor
Dr. Evelyn Grünheid
Dr. Katrin Schiefer
Editorial Assistant
Beatriz Feiler-Fuchs
Wiebke Hamann
Layout
Beatriz Feiler-Fuchs
E-mail: cpos@bib.bund.de
Scientifi c Advisory Board
Karsten Hank (Cologne)
Michaela Kreyenfeld (Berlin)
Marc Luy (Vienna)
Peter Preisendörfer (Mainz)
Nikola Sander (Wiesbaden)
Zsolt Spéder (Budapest)
Rainer Wehrhahn (Kiel)
Comparative Population Studies
www.comparativepopulationstudies.de
ISSN: 1869-8980 (Print) – 1869-8999 (Internet)
Board of Reviewers
Martin Abraham (Erlangen)
Laura Bernardi (Lausanne)
Hansjörg Bucher (Bonn)
Claudia Diehl (Konstanz)
Andreas Diekmann (Zurich)
Gabriele Doblhammer-Reiter (Rostock)
Jürgen Dorbritz (Wiesbaden)
Anette Eva Fasang (Berlin)
E.-Jürgen Flöthmann (Bielefeld)
Alexia Fürnkranz-Prskawetz (Vienna)
Beat Fux (Salzburg)
Joshua Goldstein (Berkeley)
Sonja Haug (Regensburg)
Hill Kulu (Liverpool)
Aart C. Liefbroer (The Hague)
Kurt Lüscher (Konstanz)
Emma Lundholm (Umeå)
Nadja Milewski (Rostock)
Dimiter Philipov (Vienna)
Roland Rau (Rostock)
Tomáš Sobotka (Vienna)
Jeroen Spijker (Barcelona)
Olivier Thévenon (Paris)
Helga de Valk (Brussels)
Heike Trappe (Rostock)
Michael Wagner (Cologne)