Patterns of Family Life Courses in Europe – between Standardisation and Diversity A Cross-national Comparison of Family Trajectories and Life Course Norms in European Countries Dirk Hofäcker, Jana C haloupková Abstract: Throughout recent decades, a gradual shift away from an early contract- ed and simple life course pattern which dominated in the 1950s and 1960s to late protracted and more complex patterns could be observed within European coun- tries. Yet, despite multiple cross-national similarities in the changes of individual life course patterns, there exist considerable differences in the form and frequency of these changes. We argue that one possible way of better understanding these variations is to examine the connection between family formation choices and value orientations. Using data from the European Social Survey 2006 we empirically in- vestigate to what extent the family trajectories have changed across generations and how these practiced family trajectories correspond to cross-cohort changes in socially established norms about family transitions. Our results corroborate the assumption of an increasing restandardisation of family lives: Even though family trajectories have become more turbulent involving more stages and stage changes for the younger generation, “deviations” from traditional family patterns (such as unmarried cohabitation) are turning into majority behaviour, i.e. into a “new stand- ard”. Contrasting these trends with developments in family-related norms reveals that the liberalisation in norms appears to precede such changes in actual demo- graphic behaviour, even though European countries differ in the degree and pace to which such normative and behavioural changes have yet taken place. Keywords: Family trajectories · Norms · Cross-national comparison · European Social Survey · Sequence analysis Comparative Population Studies Vol. 39, 3 (2014): 559-586 (Date of release: 29.09.2014) © Federal Institute for Population Research 2014 URL: www.comparativepopulationstudies.de DOI: 10.12765/CPoS-2014-11en URN: urn:nbn:de:bib-cpos-2014-11en3 • Dirk Hofäcker, Jana C haloupková560 1 Introduction Throughout recent decades, European countries have undergone substantial chang- es in the timing and sequencing of family transitions. Childbearing, union formation and getting married are on average being experienced at a later age than several decades ago (Sobotka/Toulemon 2008; Corijn/Klijzing 2001). At the same time, the previously close linkage between the individual transitions, especially between marriage and family formation, has loosened. There is also an increasing decou- pling between leaving parental home and union formation. As a result, more young people live alone before they enter into a relationship, and out-of-wedlock births are becoming more common across Europe; in some countries in Northern and West- ern Europe they even predominate (Billari/Liefbroer 2010; Sobotka/Toulemon 2008). Empirical data on demographic markers of transitions to adulthood suggest that these changes in family formation from one cohort to the next refl ect a gradual shift away from an early, contracted, and simple life course pattern, which dominated in the 1950s and 1960s to late, protracted, and more complex patterns (Billari/Lief- broer 2010). These changes are often summarised as a de-standardisation of previ- ously dominant standard life course models, implying that a long-lasting standard social “blueprint” of family structures and the timing of family transitions is losing signifi cance and is gradually being replaced by the emergence of a broad variety of more heterogeneous “choice biographies” (Furlong/Cartmel 1997). Although similar changes in individual life courses are observed across Europe, there exist considerable differences in the form and rhythm of these changes across different institutional and cultural backgrounds (cp. section 3.1). A number of different explanations have been put forward to explain these cross-national differences. In this paper, we shall argue that one possible way of better understanding these variations is to examine the connection between family formation choices and value orientations. In demography, an internationally com- mon pattern of major changes of demographic behaviour since the late 1960s has been described as the “Second Demographic Transition” (e.g. Van de Kaa 1987) and has been explained as the expression of a more general value change. In this sense, changes in family structures can be considered as part of a process of “individuali- sation” and a weakening of social norms that organise and structure the life course (Beck 1992). Findings from life course research, however, challenge the assumption of a gradual demise of family-related norms. Life course theorists argue that despite trends of diversifi cation, the “standard family course” still serves as an orientation for individual life planning (Kohli 2007). Values within the family sphere have been remarkably stable over time (Scott/Braun 2006), suggesting that “the demise of tra- ditional family values has been exaggerated” (Scott 2006: 16). More than in other spheres of life (e.g. in employment or occupational careers), social norms about the timing and sequencing of family transitions still represent internalised “scripts” about appropriate behaviours (Buchmann 1989; Settersten/Hagestad 1996) exert- ing infl uence on individual action (Billari/Liefbroer 2007). Cross-national variations in norms (and their development over time) thus may provide a helpful analytical Patterns of Family Life Courses in Europe – between Standardisation and Diversity • 561 tool to better understand the different forms and “rhythms” of changes in family life courses across countries. Against this background, the aims of our contribution are twofold. First, we em- pirically investigate to what extent the family trajectories of young adults in Europe are becoming more de-standardised in terms of timing and sequencing of four fam- ily transitions: leaving parental home, starting a union, entry into fi rst marriage and becoming a parent. Using sequence analysis, we explore to what extent family trajectories of young adults in Europe aged between 18 and 35 are becoming more turbulent and dissimilar across cohorts as compared to a “traditional family pat- tern”, defi ned as entry into marriage preceding childbirth without a previous period of unmarried cohabitation. Secondly, we analyse how these practiced family tra- jectories correspond to cross-cohort changes in socially established norms about family transitions. Our results contribute to the existing research by updating the description of change of family-related social norms and of family trajectories as well as by linking both spheres and processes of change across birth cohorts in an international perspective (cp. section 3.3). For our analyses, we refer back to Wave 3 of the European Social Survey from 2006/2007 that contains comprehensive retrospective data on family life course transitions and related social norms in a separate questionnaire. Furthermore, its pan-European scope allows us to study the relevant life course transitions across a wide set of different institutional and cultural backgrounds including those of the new EU member states of Central and Eastern Europe, where family forms and de- mographic processes have changed most noticeably in recent decades. Due to the cross-sectional nature of the ESS, we cannot test for causal effects between individ- ual attitudes and norms on individual decisions and behaviour. Our analytical inter- est thus is mainly in describing changes in family behaviour and in social norms as well as elaborating associations between these two processes on the macro-level. The macro-level is of interest to us because it is plausible that individual behaviour is not only linked to individual attitudes but also to social norms that are based on dominant attitudes within a given society. In the following, we shall outline the key theoretical concepts underlying our analysis in more detail (section 2). Section 3 subsequently presents a concise over- view of the current state-of-the-art in comparative research on family trajectories, its main results and persistent limitations. Against this background, we develop an analytical framework and derive hypotheses for our empirical analyses (section 4). Subsequent to a short description of the ESS data (section 5), section 6 presents empirical evidence on developments in family trajectories and family norms in Eu- rope based on descriptive statistics as well as a sequence analyses of retrospective survey data. Finally, we critically discuss our results and draw conclusions for fur- ther analyses in this fi eld (section 7). • Dirk Hofäcker, Jana C haloupková562 2 Theoretical background 2.1 Changes in family transitions: standardisation vs. destandardisation Transformations in family life trajectories have been in the focus of various studies in the social sciences throughout previous decades. The analysis of such changes relates to more general efforts to conceptualise dimensions of life course trajecto- ries (Brückner/Mayer 2005) and to develop statistical measures for them (Barban/ Billari 2012; Elzinga/Liefbroer 2007; Fussell 2005; Fussell et al. 2007). Earlier re- search has described changes in life course patterns using various concepts like “de-institutionalisation”, “de-standardisation”, “pluralisation” and “differentiation” often used interchangeably. However, these processes entail multiple dimensions, which can occur independently or even incoherently. De-institutionalisation refers to a progressive weakening of normative, legal or organisational rules that regulate the organisation of sequences in the life course. The term de-standardisation of life courses refl ects the fact that either certain events in life are experienced by smaller and smaller shares of the population or occur at more diverse ages and for durations that vary more widely (Brückner/Mayer 2005). The processes of de-standardisation thus include both changes in the patterns of family life courses and of their volatility or turbulence, which can take different directions. For instance, the pluralisation of family forms (the increase in the number of synchronous states in a population) and the differentiation of the life course (the increase in the number of distinct stages within individual life courses, e.g. premarital cohabitation) may be accompanied by standardisation processes (e.g. when premarital cohabitation becomes an almost universal part of family trajectories) and may result in the emergence of a “new standard life course” (Brückner/Mayer 2005). At the same time, the processes of accompanying normative changes (de-institutionalisation) – to which we shall turn subsequently – can occur independently from the changes of demographic behav- iour or it can precede them. 2.2 Family norms Previous research has pointed to a number of macro-social factors that may jointly have triggered recent transformations in demographic behaviour, including chang- es in the economic and social structure, cultural changes and technological inno- vations (cp. section 3.1). Yet, while research results show that social norms about the timing and especially sequencing of family formation are still strong and differ widely across countries (e.g. Billari/Liefbroer 2010), surprisingly little research has considered these normative patterns as potential correlates of cross-national dif- ferences of changes in family trajectories. As Liefbroer and Merz (2009: 1) argue, cross-national differences in demographic behaviour may actually stem from “dif- ferent norms being dominant in different countries.” Individuals may thus not only orient their behaviour at the structural and/or material opportunities and constraints that economic and social structures create; they may equally take the expectable Patterns of Family Life Courses in Europe – between Standardisation and Diversity • 563 external evaluation of their actions through their more immediate or wider social surroundings into account and therefore avoid acting in stark contrast to the convic- tions of people with whom they interact or to general social norms. Expectations by relatives or friends may, for example, infl uence young people’s readiness to leave parental home at a specifi c age (Billari/Liefbroer 2007). From a role theory perspec- tive, individuals seek positive sanctions by signifi cant others and try to avoid nega- tive sanctions such as social exclusion. From a rational choice perspective, indi- viduals may not only be acting “rationally” in their family decisions by seeking for material gains respectively by avoiding material costs. They may, at the same time, act in line with existing norms in order to receive “social consent” from relevant oth- ers, respectively to avoid social costs. Following this line of argument, individuals may in times of rapid social and eco- nomic changes not directly adjust their demographic behaviour to changing struc- tural conditions. Instead, they may still tend to adhere to more “conservative” be- haviour as norms are changing at a lesser pace. Changes in demographic behaviour thus may then be regarded as a gradual process in which “obsolete” values hinder individuals in the adaptation of the new behaviours for some time, until the values change themselves and catch up with the innovation” (Rossier 2010: 8). It thus can be assumed that demographic behaviour will change with a certain “time lag” after norms have changed up to a certain “threshold point” at which deviations from the previous behavioural standard do no longer imply a violation of a dominant norm. 3 Previous studies Following the theoretical overview, we shall now turn to a schematic synopsis of previous research on both changes in family transitions as well as the most recent developments in family norms and values. Based on this synopsis, we subsequently identify research gaps and outline the additional contribution provided by our work. 3.1 Diversity of family formation patterns across Europe Although pathways to family formation are changing into the same direction in most parts of Europe, signifi cant differences across Europe still persist and there is yet little sign of convergence in the timing and sequencing of family events (Billari/ Liefbroer 2010; Blossfeld et al. 2005; Corijn/Klijzing 2001; Sobotka/Toulemon 2008). While there are some signs of convergence between Eastern European countries on the one hand and Northern and Western European states on the other hand, liv- ing arrangements in Northern and Western Europe and in Mediterranean countries still differ considerably (Fokkema/Liefbroer 2008). In Southern Europe, for example, leaving home is still a much more protracted process and there is a smaller propor- tion of young people living on their own than in Northern and Western Europe. Con- sequently, alternative family forms (such as unmarried cohabitation, non-marital parenthood, lone parenthood etc.) as “intermediate steps” in family formation have spread differently across the two groups of countries (Beier et al. 2010). • Dirk Hofäcker, Jana C haloupková564 Until now, there are only few comparative studies examining the demographic markers1 of the transition to adulthood in a holistic way – as a series of transitions (Gauthier 2007). Only throughout the last years, such studies were enabled through methodological developments in sequence analysis (cp. section 5) which allow studying entire life course trajectories as a whole (Abbott/Tsay 2000; Aisenbrey/Fa- sang 2007; Lesnard 2006). Using this method, Elzinga and Liefbroer (2007) explored data from the Family and Fertility Survey from 1998/1999 on 19 European countries. Chanvril and her colleagues (2009) applied sequence analysis to identify patterns of early family trajectories in European countries, using ESS round 3 data from 2006. These studies revealed that, rather than being a general and uniform trend, de- standardisation of family trajectories proceeds at a distinct pace. Comparing data for women born in 1945-1964, Elzinga and Liefbroer (2007) observed increased dis- similarity of women’s family trajectories and increased variation in types of family trajectories across cohorts in most Western European countries.2 In contrast, in the former communist countries of Central and Eastern Europe, cross-cohort differ- ences were rather small. However, these countries experienced a transformation in institutional and economic conditions only in the 1990s, affecting mainly family behaviour of those born in 1970 and later. For Sweden (and partly the Netherlands) Elzinga and Liefbroer (2007) found decreasing variation in types of family trajec- tories amongst the youngest cohort which might suggest an emergence of “new standard family trajectories” in these countries that have been the pioneers of the second demographic transitions. However, while much research acknowledges the persistence of cross-national differences in family trajectories, only little research has provided explanations for this asymmetric spread of de-standardisation (respectively re-standardisation). Elz- inga and Liefbroer (2007) refer to the varying nature of welfare state benefi ts in dif- ferent welfare regimes: In social-democratic countries, generous benefi ts are being provided irrespective of a person’s social position and thus are assumed to weaken reliance on the family, thereby lowering family infl uence and promoting autonomous behaviour. In contrast, family benefi ts in liberal, conservative and Southern Euro- pean regimes are rather marginal, leading to a stronger reliance on the family and thus to stronger adherence to “traditional” family behaviour. A similar conclusion is reached for post-socialist countries, where initial state support throughout commu- nism was high, but was counteracted by “strong normative constraints on individual behaviour” (Elzinga/Liefbroer 2007: 229). Furthermore, in a synthetic work on family formation patterns in Europe, Liefbroer points to the signifi cance of changes in the economic and social structure (e.g. educational expansion, the increasing labour force participation of women and most recent welfare state retrenchment), cultural 1 These demographic markers usually include leaving parental home, entry to fi rst union or mar- riage and having a fi rst child. From a broader perspective, the transition to adulthood typically also entails the transition from studying full-time to entering the labour market (Modell et al. 1976). 2 Their study focused on changes in following states: single, married, unmarried cohabitation and (not) having child(ren) (Elzinga/Liefbroer 2007). Patterns of Family Life Courses in Europe – between Standardisation and Diversity • 565 changes (e.g. the rise of individualism, the re-emergence of feminism and the de- crease in normative controls) and technological innovations (such as the spread of contraceptives and the increasing range of modern mass media; see Liefbroer 2009 for an overview). However, the data underlying these analyses often originate from the late 1990s and it remains an open question how the above may still serve as adequate explanations of demographic behaviour in times of increasing economic, cultural and technological convergence among European countries. 3.2 Previous research on family norms across Europe While there has been a plethora of research on developments in actual family tran- sitions, and despite a growing overall interest in values and culture within the so- cial sciences (e.g. van Oorschot et al. 2008), research on family-related norms has remained rather scarce. The reasons for this relative “ignorance of social norms in current demographic thinking” (Liefbroer/Billari 2010: 288) may be twofold: On the one hand, from a substantial perspective, theorists of the “Second Demographic Transition” have frequently assumed that the signifi cance of social norms in govern- ing individual life courses has weakened (see Van de Kaa 1987 and the above dis- cussion). On the other hand, from a methodological perspective, the idea of “family norms” has proven to be rather diffi cult to operationalise. Most generally, norms can be regarded as prescriptive, proscriptive or pro- hibitive statements about the “right” or appropriate behaviour, shared by a certain group of actors and related to some sort of sanctions in case of their violation (e.g. Billari/Liefbroer 2010; Liefbroer/Merz 2009). Norms may become internalised as at- titudes or “individual norms” but it can be assumed that regarding their explanatory power for group behaviour, it is rather norms shared by a sizeable proportion of the population (“societal norms”) that play a role. Previous research on family-related norms had largely focused on “age norms”, i.e. prescriptive ages at which specifi c transitions “should” (or should not) happen. Based on a regional US sample, Set- tersten and Hagestad (1996), for example, show that there exist defi ned and socially shared values about maximum ages until which specifi c family transitions such as leaving parental home, cohabitation or entry into parenthood should occur. Norms may, however, also entail prescriptions about the occurrence (“quantum norms”) and the sequencing of specifi c stages of family life (“sequencing norms”). They provide a somewhat “weaker” type of family norms, as they do not explicitly focus on specifi c age limits but investigate how far more general deviations from a spe- cifi c “standard” family model would be tolerated. Existing studies point towards an increasing tolerance of deviations from a “traditional family life course”. Contrasting various indicators of family-related behaviour included into the European Values Studies 1981 to 1999, Halman (2009), for example, fi nds indications of a gradual but persistent trend of “de-traditionalisation”, i.e. a shift away from traditional norms and an increasing importance of individual choice in agency. However, there appear to be considerable cross-national differences in family-related norms (Liefbroer/ Merz 2009; Wernhart/Neuwirth 2007). One extreme is represented by Scandinavian countries which show a high tolerance for “deviations” from the standard life course • Dirk Hofäcker, Jana C haloupková566 model (such as unmarried cohabitation or childlessness). At the other end, there are countries especially in Southern and to a lesser degree in Central Europe in which the acceptance of deviant “alternative” family forms (such as one-parent families or unmarried couples), of childlessness and family dissolution (e.g. through divorce) is still low (Liefbroer/Merz 2009). While much previous research focuses on developments in family-related norms per se, only a few recent studies try to systematically link the developments in norms to actual family behaviour. Billari and Liefbroer (2007) use a longitudinal Dutch sur- vey to reconstruct the infl uence of individual family norms on the process of leaving home and indeed fi nd that the timing of young people in leaving home is infl uenced by the social expectations their parents seem to have. In a similar way, Liefbroer and Billari (2010) argue based on Dutch results that relevant social expectations related to the timing of family transitions and their sequencing still exist. Their fi ndings also suggest the emergence of a new set of normative expectations on young adulthood that include living on one’s own and unmarried cohabitation. However, both studies remain restricted to the normatively rather “liberal” case of the Netherlands, thus allowing for only limited generalisations. 3.3 Aim and scope of the article Our study draws on the existing research in both the analysis of family trajectories and family norms but goes beyond it in several respects: First, we turn from the analysis of single life transitions to a holistic perspective of entire family trajectories. We focus on a series of demographic markers of the transition to adulthood ranging from leaving the parental home, starting a fi rst re- lationship, making the transition to fi rst marriage up to having a fi rst child for both women and men. Second, our study combines the analysis of demographic behaviour with a com- parative investigation of their correspondence to family-related social norms. Ow- ing to the increasing availability of adequate data sets, cross-national comparative research into gender- and family-related norms has amplifi ed signifi cantly through- out recent years, but patterns and developments in norms and values have hardly been systematically connected to actual changes in family trajectories. The system- atic juxtaposition of family-related norms with actual family transitions using one unique data set, as done in this paper, thus can make a substantial contribution to a better understanding of the interrelation between the two spheres. Finally, unlike much earlier studies, we contrast the development in family trajec- tories as well as social norms across birth cohorts in order to empirically investigate the cohort-specifi c interplay between structural and attitudinal changes. Earlier studies, using for example data from the Family and Fertility Surveys often were restricted up to the late 1990s. Most recent data from the 2006 wave of the Euro- pean Social Survey allows us to extend our analyses from the post-war generation to more recent cohorts (i.e. born in the 1970s) which partly experienced their family trajectories after the fall of the iron curtain. This extended “observation window” is of special importance for post-communist countries where family forms and demo- Patterns of Family Life Courses in Europe – between Standardisation and Diversity • 567 graphic processes have changed noticeably in the last decades. Until now, existing research has pointed to varying developments across this group of countries, but there has been only limited empirical proof for this claim.3 Our study intends to provide insights into developments within this still largely under-researched group of countries. 4 Hypotheses Based on the theoretical outline as well as the research synopsis, we now develop hypotheses for the subsequent data analyses. As outlined above, empirical research using data up to the late 1990s has point- ed to a rise in non-standard family forms and “intermediate” stages in family devel- opment (see Elzinga/Liefbroer 2007). We assume that this process may not yet be completed and thus expect to fi nd a similar increase in the overall number of states in family trajectories and/or increasing variation in durations of these states, i.e. a rise in the “turbulence” of family trajectories, in our study of family sequences. These changes will be less signifi cant in countries where the level of turbulence was already relatively high among older cohorts. At the same time we expect that family trajectories are becoming more dissimilar from traditional family formation patterns (Hypothesis 1). However, we assume that, as in the past, developments in family trajectories have not occurred unidirectionally in all modern societies. In line with earlier re- search (e.g. Elzinga/Liefbroer 2007: 227-229), we assume that both the level of de- standardisation as well as the direction and speed of changes in family trajectories will vary and that this variation is shaped by welfare regimes as well as by social norms. On the one hand, following Elzinga and Liefbroer (2007), we assume that given different welfare state support for autonomous behaviour, the level of destandardi- sation will be most pronounced in social-democratic countries while in both con- servative and liberal countries the relatively low involvement of state policies into the family sphere will delimit autonomy and foster a higher adherence to more tradi- tional family forms, i.e. a somewhat lesser degree of turbulence and destandardisa- tion of family trajectories. For the largely familialistic Southern European countries, we expect this pattern of a rather low level of destandardisation and turbulence in family trajectories to be even more pronounced. Finally, in post-socialist countries 3 A recent study on Polish women’s work-family trajectories after fi nishing full-time education (Baranowska 2008) as well as a study on family trajectories of Czechs between age 18 and 35 (Chaloupková 2010) have found an increase of heterogeneity in young adult’s life-course trajectories after the fall of communism. However, using entropy analysis, the Czech analyses revealed that the increase in the heterogeneity of age-specifi c family situations is not uniform, as for young people born in 1980s, a reversal trend of declining diversity in family statuses was found in their early adulthood due to a postponement of family-related transitions (Chaloup- ková 2010). • Dirk Hofäcker, Jana C haloupková568 we expect that the initial strong state support, suggesting a similar development as in Scandinavian countries, has been largely counteracted by rather prohibitive norms through both public policies and traditionally conservative religious values (Hypothesis 2). On the other hand, concerning patterns of changes in family behaviour over time, we assume that different “norms” about family behaviour will moderate the level and speed of change. While welfare regimes defi ne the general institutional opportunities and constraints under which individuals will make their (family) deci- sions, behavioural changes may be diffi cult to implement when they are in stark contrast to the norms and expectations of signifi cant others or the “general public”. Existing norms thus may mediate the speed at which demographic behaviour may alter or reverse. In contexts in which deviations from a “standard” family model are being tolerated by large shares of the population, we assume that trends towards a “destandardisation” of family trajectories over time will be most dynamic. On the contrary, in contexts where a sizeable share of the population rejects such devia- tions and : a certain limit of minimum acceptance is not surpassed, processes of destandardisation will be rather slow or even reversed (Hypothesis 3). 5 Data and methods For testing our hypotheses, we draw upon cross-national data from the European Social Survey (ESS) Wave 3 which was fi elded in 2006/2007 and includes 23 Euro- pean countries.4 This wave included a special module on “The Timing of Life: The organization of the life course in Europe”, including various indicators on the organi- sation of family life courses using both behavioural as well as attitudinal questions. To explore changes in family structures we employ sequence analysis. Sequence analysis is a set of methods which focus on complex descriptions of ordered lists of elements (e.g. events, statuses) as an entity. Particularly, it concentrates on the comparison of similarity and dissimilarity of sequences and the identifi cation of typical patterns (Brzinsky-Fay et al. 2006; Macindoe/Abbott 2006). This approach enables to study the changes in a whole sequence of family tran- sitions – and to take into account changes in the timing and order of family transi- tions. For reconstructing actual life course transitions, we draw upon a set of ret- rospective questions through which respondents were asked in which year (i) they fi rst left their parents for at least two months or started to live separately from them (“leaving home”– L), (ii) they fi rst lived with a spouse or a partner for three months or more (“cohabitation”– U), (iii) in which year they fi rst married (“marriage” – M) and 4 Namely: Austria, Ireland, the Netherlands, Ukraine, Belgium, Bulgaria, Cyprus, Denmark, Es- tonia, Finland, France, Germany, Hungary, Norway, Poland, Portugal, the Russian Federation, Slovakia, Slovenia, Spain, Sweden, Switzerland, and the United Kingdom. Patterns of Family Life Courses in Europe – between Standardisation and Diversity • 569 (iv) in which year they fi rst gave birth to respectively fathered a child (“childbirth”– C).5 We restrict our analyses to respondents who were 35 years old or older at the time of data collection. This design permits us to analyse full trajectories of men and women aged between 18 and 35 years who were born up to the early 1970s – these cohorts could not be considered in earlier studies such as the previous work of Elzinga and Liefbroer (2007). Using information on the respondent’s age when all “fi rst” family-related transitions happened, a set of 18 count variables representing each year of age from 18 to 35 years6 was created, and all events that respondents had already experienced by that age were assigned to each new age variable.7 In to- tal, our data enable to reconstruct 20,486 trajectories. Consequently, our sequences do not take into account possible reversibility of states and higher complexity of family patterns of young adults (e.g. the return to parental home or partnership un- ion dissolution), but only „fi rst“ events. As an illustration, Table 1 displays the family trajectory of a respondent who left parental home at age 20, started to live with a partner at 24, married at 26 and had a child at 27 years of age. In a shortened way, it can be reported as: 0(2) – L(4) – LU(2) – LUM(1) – LUMC (9), where the number in brackets indicates the number of years spent in a given state. To analyse changes in family trajectories of young European adults, we employ two indicators of destandardisation of family trajectories: the turbulence index and the dissimilarity index. The turbulence index proposed by Elzinga and Liefbroer (2007) is a composite measure that accounts for the number of distinct sub-se- quences and the variance of the consecutive times spent in each state (for details see Elzinga/Liefbroer 2007; Gabadinho et al. 2009). It not only measures the increase in the overall number of distinct states in individual family trajectories, but also in- creases in their volatility, i.e. an increasing variation in the durations of these states. Given our state defi nitions, increasing values of the turbulence index indicate that family trajectories encompass more so-called “intermediate” states between leav- 5 All these single questions were preceded by a fi lter question to establish whether the respec- tive event had actually taken place. Consistency checks were undertaken to exclude wrong or unrealistic answers. 6 Transitions that occurred before the age of 18 (such as leaving the home earlier, teenage preg- nancies etc.) are coded as if they had happened at the age of 18. Nevertheless, these transitions were marginal. At the age of 18 only 3 percent of the respondents already had a fi rst child and 4.9 percent were married. 7 As we are observing four events, there exist 16 (24) possible combinations of already experi- enced family stages. Tab. 1: Example of a family trajectory between the age of 18 and 35 Age 18 19 20 21 22 23 24 25 26 27 ... 35 State 0 0 L L L L LU LU LUM LUMC .... LUMC Source: Own design • Dirk Hofäcker, Jana C haloupková570 ing the parental home and parenthood (such as living alone or unmarried cohabita- tion) and that their durations are increasingly fl exible. To map the macro level changes in family patterns across European countries, the second measure – the dissimilarity index – compares all observed trajectories with a reference sequence representing a “traditional” family life course, defi ned as leaving parental home when married and having a child afterwards. The tradi- tional family life course thus neither includes any periods of long-term unmarried cohabitation nor a period of living alone. Technically, our reference trajectory not only includes a 0-LM-LMC trajectory (leaving parental home and marriage in the same year, without unmarried cohabitation) but also 0-LUM-LUMC, i.e. a trajectory in which also a short phase of premarital cohabitation has occurred in the same year, which may have lasted close to 12 months or only a few days.8 The refer- ence length of the family life course phases were defi ned based on the highest observed frequencies in the whole data set: (0,3)-(LM,2)-(LMC,13) and (0,2)-(LUM,1)- (LUMC,15). The fi nal distance for each trajectory was determined as the lowest val- ue of the two distances to these two references. The comparison of observed and reference trajectories is based on the length of the longest common sub-sequence (LCS distance) (Gabadinho et al. 2009). The normed LCS distance has a value of 0 when all sub-sequences occur in both sequences, and reaches its maximum of 1 if there is no common sub-sequence between these two trajectories. In our study, we measure the extent of destandardisation across cohorts by comparing the average normed LCS distance between empirically observed trajectories and the predefi ned “traditional family trajectory”. If the average dissimilarity is close to 1, the level of de-standardisation thus is high. The sequence analysis was done using the module TraMineR in R (Gabadinho et al. 2009). For contrasting transition behaviour with related norms, we draw upon a series of questions which focus on possible deviations from a traditional family life tra- jectory. Although the ESS contains a series of respective questions differentiating between ideal transition ages of men and women for a number of family-related transitions, their comparison with actual transition ages is problematic. The fi rst descriptive analysis showed that across countries individuals tend to choose “round ages” in 5-year intervals, such as 20, 25 or 30 years as ideal transition ages.9 Moreo- ver, the ESS contains no explicit question on the ideal age of leaving the parental home, thus making a direct comparison of actual and normative transition ages on all relevant dimensions impossible. For our analyses, we therefore rely on three different questions using 5-point Lik- ert scales, ranging from “strongly approve” to “strongly disapprove”, which meas- ure in how far individuals support or reject (i) never having children, (ii) living with a partner without marrying and (iii) having a child with a partner one is not married to. These indicators allow us to compare the actual incidence of non-traditional 8 As data were only collected for the years when the family transitions happened, we cannot make a more detailed distinction. 9 Based on own analyses; results are available on request. Patterns of Family Life Courses in Europe – between Standardisation and Diversity • 571 family formation patterns with their perceived societal acceptance using standard descriptive statistics. We compare trends – in family behaviour and in family-related norms – between the three birth cohorts: 1941-1950, 1951-1960 and 1961-1972. They allow us to at least partly capture trends between the war and the post-war generation which are known to have grown up under rather different circumstances with the more re- cent cohorts having been subject to a number of major social transformations (e.g. educational expansion, gradually converging labour market participation of both spouses) as discussed in earlier literature (e.g. Gradstein/Nikitin 2004; Blossfeld/ Hofmeister 2006). For post-socialist countries, this span also allows us to compare cohorts that spend more or less their entire family cycle under socialism with more recent partially “post-socialist” cohorts. We are aware that life courses as well as social norms differ within societies and age cohorts by various factors, especially by education (Sobotka 2008). Yet, for reasons of analytical clarity and parsimonious- ness of our multi-country analysis, we do not differentiate our cohorts further and follow a more descriptive research interest. The major strength of the ESS undoubtedly lies in its comparative character including European countries of very different origin. One major restriction of our analysis, however, is the availability of cross-sectional data only. While retrospec- tive life calendars incorporated into the ESS allow for a rather accurate reconstruc- tion of family-related events, this cannot straightforwardly be achieved for family- related norms that are only available for the time of the interview. It thus needs to be assumed that family norms may change across generations but remain largely stable within a generation, so that the norms measured at the time of the interview may be taken as overall proxies for norms held at earlier points in time. Given that family-related norms relate to very basic conceptions about how to organise one’s own life course (unlike, for example, political norms) we assume that this approxi- mation may be plausible. More detailed analyses with panel data (as, for example, provided by Moors and Lesthaeghe 2000 for Belgium) will be needed to ascertain the assumptions laid out in this analysis. 6 Results 6.1 Changes in family trajectories across European countries Our fi rst set of questions relates to cohort changes in family trajectories across European countries. We expected to fi nd an increase in turbulence in family trajec- tories and an increase of dissimilarity compared to traditional family patterns across cohorts. Table 2 summarizes the number of distinct sub-sequences and the vari- ance of the respective time spent in each state in an overall index of “turbulence” in family trajectories. Based on Table 2, we can draw several conclusions. The turbu- lence index reaches its highest values in Scandinavian countries, Switzerland, the Netherlands, Belgium and France implying that young people in these countries on average experience a higher number of distinct states and/or a higher variation in • Dirk Hofäcker, Jana C haloupková572 their duration. While in most Western European countries the index of turbulence increases signifi cantly across cohorts, it remains quite stable and even decreases for the youngest cohort in some countries of Southern and Central and Eastern Eu- rope. Unexceptionally, the results do not support the hypothesis of an increase of turbulence in family trajectories (H1), but rather point to cross-national differences largely in line with our regime hypothesis (H2). Tab. 2: Average turbulence per country and cohort 1941-1950 1951-1960 1961-1972 Denmark** 4.63 5.10 5.38 Netherlands* 4.73 4.72 5.36 Sweden 5.03 5.10 5.22 France** 4.57 4.54 5.19 Switzerland* 4.76 5.06 5.18 Finland** 4.63 4.86 5.18 Norway* 4.51 5.04 5.15 Belgium* 4.52 4.79 5.04 United Kingdom* 4.61 4.55 4.89 Germany* 4.62 4.57 4.88 Austria 4.58 4.68 4.79 Ireland 4.44 4.54 4.79 Spain 4.53 4.41 4.62 Portugal* 4.42 4.01 4.42 Estonia 4.32 4.51 4.37 Cyprus 4.17 4.43 4.37 Slovenia 4.49 4.48 4.30 Poland 4.40 4.27 4.29 Hungary 4.12 4.08 4.27 Russian Federation 4.27 4.35 4.23 Ukraine 4.30 4.05 4.05 Slovakia 4.33 4.16 4.03 Bulgaria 4.05 4.04 3.89 Note: Values of turbulences are between 1-9.5; mean 4.62. Cohort differences signifi cance *p< 0.05; **p< 0.01. Turbulence is a composite measure that accounts for the number of distinct sub-sequenc- es and the variance of the consecutive times spent in each state (for details see Elzinga/ Liefbroer 2007; Gabadinho et al. 2009). Increasing values indicate that family trajectories are more volatile and encompass more distinct states. Source: European Social Survey Wave 3 (own calculations) Patterns of Family Life Courses in Europe – between Standardisation and Diversity • 573 In order to investigate whether family trajectories in young adulthood are indeed becoming more diverse across cohorts, we compute average distances between observed family trajectories and the “traditional” reference trajectory based on the longest common sub-sequence for every country and cohort (Table 3). In line with our previous analyses, the fi ndings suggest that family trajectories indeed are be- coming more dissimilar compared to traditional family trajectories across cohorts. Tab. 3: Average distances to “traditional family trajectory” per country and cohort a cohort differences are signifi cant at 0.05 level. The table shows normed distances of observed trajectories to reference traditional family trajectory (see above for defi nition) based LCS distance (Gabadinho et al. 2009). It reaches values from 0 (similarity of sequences) to 1 (no common subsequence). The closer the value is to 1, the higher is the average dissimilarity to traditional family trajectory. Source: European Social Survey Wave 3 (own calculations) 1941-1950 1951-1960 1961-1972 Swe dena 0.50 0.63 0.70 Denmarka 0.39 0.58 0.66 Irelanda 0.48 0.54 0.65 Netherlandsa 0.41 0.57 0.65 Norwaya 0.39 0.47 0.64 Switzerlanda 0.51 0.58 0.64 United Kingdoma 0.41 0.52 0.63 Francea 0.40 0.46 0.62 Finlanda 0.45 0.50 0.61 Germanya 0.46 0.47 0.57 Spaina 0.40 0.47 0.57 Austriaa 0.43 0.51 0.56 Portugala 0.37 0.38 0.51 Sloveniaa 0.41 0.44 0.49 Belgiuma 0.32 0.42 0.48 Bulgariaa 0.39 0.42 0.47 Estonia 0.43 0.40 0.44 Poland 0.38 0.42 0.43 Russian Federationa 0.35 0.35 0.42 Hungary 0.38 0.35 0.40 Ukraine 0.37 0.36 0.40 Cyprus 0.34 0.37 0.39 Slovakia 0.31 0.33 0.36 • Dirk Hofäcker, Jana C haloupková574 The highest dissimilarity is observed in Scandinavian countries, Ireland, Switzer- land and in the Netherlands, while in post-communist countries, the dissimilarity is lower and cohort differences are less pronounced. Obviously, there is a clear correlation between these two measures. In coun- tries where family trajectories in young adulthood are more turbulent, the family trajectories are also more dissimilar compared to the “traditional” pattern.10 These processes, however, do not necessarily lead away from universal family models as refl ected in the idea of “de-standardisation”. If, for example, all young adults would follow similar new patterns of family developments “standard family trajectories” may prevail despite an increase in stages passed through. In principle, one can dis- tinguish two alternative scenarios: Firstly, in line with some earlier fi ndings (Wagner/Franzmann 2000) one may ex- pect that despite the observed changes family life courses still remain rather uni- form within societies. This scenario of continuing intra-cohort standardisation im- plies that there is no signifi cant increase in intra-cohort differences in trajectories and there is continuing dominance of one “standard” family model rather than a pluralisation of family forms, even if the “standard” family model changes between cohorts and becomes more turbulent. The alternative restandardisation scenario moves away from the assumption of a unilateral development by linking the ideas of de-standardisation and (re)stand- ardisation as two subsequent phases. It assumes that while initial family life tra- jectories may become more diversifi ed, this divergence may only be temporary. In the long-term, a new (more turbulent) dominant model of family trajectories will be established in the youngest cohort. To differentiate these scenarios, we compare the share of individuals who ex- perienced specifi c family states and events until the age of 35. Table 4 shows the share of individuals who have ever (before the age of 35): (i) lived alone (neither living in a parental home, nor with the partner); (ii) lived in unmarried cohabitation (without subsequent marriage); (iii) experienced premarital unmarried cohabitation (lived in unmarried cohabitation at least one year before marriage); (iv) married; (v) had a child out of wedlock; (vi) married after childbirth; and (vii) who have remained childless until the age of 35. Although the share of young people who left parental home and are not living with a partner is increasing across cohorts, there are pronounced differences across European countries. While in Scandinavia and in most Western European countries, more than 50 percent of individuals have left parental home without starting a un- ion particularly in the youngest cohort, their proportion is much smaller in Central and Eastern European countries. Similar differences arise with regard to the experi- ence of premarital and unmarried cohabitation. Cohabitation before marriage has become the standard pattern of family trajectories in most Western European coun- 10 In some countries of Eastern Europe, the at fi rst sight surprising dissimilarity to the defi ned reference trajectory might be infl uenced by the fact that a notable proportion of those who married and have children continue to live in their parental home. Patterns of Family Life Courses in Europe – between Standardisation and Diversity • 575 Tab. 4: Share of respondents who have experienced specifi c family states (in percent), by country and birth cohort Had ever (until age of 35) lived alone lived in unmarried cohabitation lived in premarital cohabitation (% of married) married child out of wedlock marriage after childbirth childless Austria 1941-1950 44.6 31.8 35.7 92.6 9.3 7.4 19.0 1951-1960 42.1 49.2 56.9 81.4 17.1 12.5 23.0 1961-1972 46.3 64.2 72.0 74.4 23.4 16.1 28.2 Belgium 1941-1950 15.2 5.7 8.6 95.9 2.0 1.6 11.5 1951-1960 30.2 21.3 25.6 88.9 4.6 2.3 17.7 1961-1972 28.9 39.5 46.4 81.4 11.4 4.8 19.4 Bulgaria 1941-1950 26.1 8.4 9.7 98.2 1.8 1.8 5.3 1951-1960 24.2 7.8 11.9 95.0 1.8 1.4 11.9 1961-1972 25.9 10.3 19.5 88.6 2.2 2.2 16.8 Cyprus 1941-1950 22.1 13.0 17.6 95.4 0.0 0.0 13.7 1951-1960 30.5 17.5 23.2 93.2 2.8 2.3 10.2 1961-1972 27.7 26.5 35.5 86.7 3.6 3.0 17.5 Denmark 1941-1950 68.5 42.6 45.9 88.9 12.6 7.8 14.1 1951-1960 68.5 84.4 87.9 71.6 30.4 17.1 25.3 1961-1972 63.4 87.1 91.9 66.8 40.3 22.0 21.7 Estonia 1941-1950 65.1 14.3 22.2 88.4 9.5 4.8 12.7 1951-1960 57.0 25.7 32.2 85.2 16.1 8.7 10.9 1961-1972 52.7 39.2 43.8 80.0 20.4 9.2 12.3 Finland 1941-1950 61.4 29.1 36.2 86.1 8.0 2.7 17.8 1951-1960 58.6 57.3 66.4 74.2 10.8 4.7 24.4 1961-1972 58.0 77.7 85.0 68.5 22.9 9.9 26.1 France 1941-1950 46.5 21.7 25.4 89.3 6.1 3.4 14.7 1951-1960 51.4 41.6 48.0 76.1 11.6 3.7 22.0 1961-1972 52.0 74.2 81.0 64.6 28.3 10.3 22.7 Germany 1941-1950 41.9 26.6 31.3 88.3 8.4 4.5 19.9 1951-1960 45.9 41.1 48.3 83.0 12.0 7.9 22.4 1961-1972 51.4 59.5 69.9 69.7 17.3 9.0 32.4 Hungary 1941-1950 16.5 7.4 13.6 90.9 4.9 3.7 13.6 1951-1960 16.6 12.4 16.2 91.7 4.6 3.3 13.3 1961-1972 21.1 26.3 32.9 82.6 7.5 2.8 17.8 Ireland 1941-1950 50.5 4.8 20.7 82.4 2.1 1.6 23.4 1951-1960 52.5 13.8 33.6 72.8 5.5 4.1 30.9 1961-1972 53.7 38.3 53.7 66.9 15.6 4.6 34.4 Netherlands 1941-1950 38.2 15.4 22.9 88.9 3.6 2.6 17.6 1951-1960 53.4 37.7 50.3 74.3 7.5 3.4 36.0 1961-1972 54.5 64.6 76.5 65.7 11.7 3.5 35.2 • Dirk Hofäcker, Jana C haloupková576 Tab. 4: Continuation Had ever (until age of 35) lived alone lived in unmarried cohabitation lived in premarital cohabitation (% of married) married child out of wedlock marriage after childbirth childless Norway 1941-1950 70.8 27.6 33.6 89.2 6.4 5.6 12.8 1951-1960 74.5 57.7 61.6 83.9 18.1 11.0 12.6 1961-1972 70.5 74.6 80.1 65.8 37.2 19.7 23.0 Poland 1941-1950 23.5 4.1 6.6 96.9 3.6 3.6 5.6 1951-1960 25.0 6.2 13.4 91.1 5.8 4.1 11.6 1961-1972 33.5 12.5 21.4 85.6 8.2 4.7 14.0 Portugal 1941-1950 24.5 2.0 7.6 93.4 7.0 5.6 10.9 1951-1960 19.1 4.7 13.0 91.0 4.7 3.6 16.2 1961-1972 27.1 14.1 24.9 81.7 10.8 5.5 23.0 Russian Federation 1941-1950 47.4 5.6 11.7 92.1 5.3 2.6 8.6 1951-1960 48.8 8.5 12.6 93.9 6.4 3.5 8.5 1961-1972 42.6 16.2 24.2 87.2 10.3 5.3 12.0 Slovakia 1941-1950 27.7 6.5 9.8 95.7 3.3 1.6 7.6 1951-1960 29.7 5.6 13.5 91.0 5.3 3.4 10.5 1961-1972 22.1 14.0 23.7 86.3 7.2 5.0 15.6 Slovenia 1941-1950 34.3 14.8 20.5 92.4 10.5 8.6 9.0 1951-1960 32.1 29.1 33.8 86.5 15.2 7.6 10.5 1961-1972 29.8 40.8 51.7 70.2 26.1 10.1 16.4 Spain 1941-1950 24.1 5.6 14.9 89.2 4.1 3.6 13.3 1951-1960 27.2 11.8 23.5 82.4 3.7 1.8 25.0 1961-1972 29.9 27.0 38.8 75.1 9.1 4.0 28.6 Sweden 1941-1950 60.7 59.8 63.7 80.1 28.0 16.1 12.5 1951-1960 60.2 84.4 90.5 65.3 40.1 24.1 22.4 1961-1972 66.9 84.0 90.6 53.9 45.6 16.3 21.5 Switzerland 1941-1950 66.5 24.8 33.5 83.8 5.4 1.8 24.1 1951-1960 67.1 56.6 62.7 79.3 6.1 2.7 34.2 1961-1972 65.5 70.0 78.4 71.6 7.9 4.5 35.8 Ukraine 1941-1950 57.4 5.9 8.2 94.5 5.1 2.3 7.4 1951-1960 45.6 9.4 14.2 93.2 3.6 2.3 7.4 1961-1972 44.3 15.5 21.4 89.7 8.1 5.5 11.1 United Kingdom 1941-1950 44.0 12.9 18.2 90.6 8.3 5.9 16.6 1951-1960 47.7 30.2 37.8 80.3 7.4 3.1 29.8 1961-1972 55.6 58.8 66.4 69.8 19.6 7.1 30.2 Source: European Social Survey Wave 3 (own calculations) Patterns of Family Life Courses in Europe – between Standardisation and Diversity • 577 tries. Although marriage still represents the dominant “fi nal stage” in family trajec- tories, there is an increasing proportion of individuals who have lived in unmarried cohabitation. The share of people who had a child born out-of-wedlock and who enter into marriage after childbirth are also increasing. Comparing the proportion of married people and individuals who experienced premarital cohabitation shows that premarital cohabitation lasting at least one year is experienced by most of the people who entered into marriage in the Scandinavian countries, the UK, France, the Netherlands, Germany and in Austria – thus signalling the emergence of a new standard transition pattern – a process of restandardisation. The share of childless people at the age of 35 differs considerably across European countries. More than 30 percent of the youngest cohort stays childless at the age of 35 in Switzerland, the Netherlands, Ireland and in Germany. The lowest proportion of childless people at the age of 35, although increasing across cohorts, is found in East European coun- tries like Russia, Ukraine, and Estonia indicating relatively low ages when entering into parenthood in a European comparison. To sum up, leaving home without marrying and an unmarried cohabitation are becoming new “standard” stages in family trajectories in the youngest cohorts and unmarried cohabitation is becoming a standard prerequisite of marriage. These changes mainly occurring in Western and Northern European countries can be in- terpreted as signs of restandardisation towards a new family model, while particu- larly in Eastern Europe we rather observe a pattern of continued standardisation. 6.2 Family norms related to family state sequencing/quantum Following the study of actual family transitions, we now turn to the investigation of relevant family norms. Table 5 displays individuals who “strongly disagree” or “disagree” with (i) the decision to remain childless, (ii) the decision to enter into non-marital cohabitation and (iii) the birth of a child out of wedlock, broken down by age cohorts. As further analyses (results not shown here) revealed that gender differences within countries are rather modest, these are aggregated results for both men and women. Table 5 indicates that there are pronounced differences in the disapproval of different types of deviations from a standard life course pattern. In most countries, disapproval is clearly higher for childlessness than for the other two indicators. Only in some countries (Denmark, Sweden, Finland, the UK, Ireland and Switzerland) respondents are similarly or even more in opposition to births out of wedlock. Yet, the degree to which Europeans disapprove of these deviations from a traditional family life course differs markedly between countries: Rejection is most pronounced in Eastern European countries as well as Cyprus where between half (Poland, Slovakia) and four fi fths (Russia, Ukraine) of the population disapprove of childlessness. Rejection of unmarried cohabitation as well as births out of wedlock is markedly lower but still remains at around 30 to 50 percent. Slovenia and Hun- gary partly deviate from this pattern as only slightly less than half of all respondents reject childlessness and less than one fi fth disapproves of unmarried cohabitation and childbirth out of wedlock. • Dirk Hofäcker, Jana C haloupková578 T a b . 5: D is a p p ro v a l* o f d e v ia ti o n s fr o m a “ st an d a rd f am ily li fe c o u rs e ” in p e rc e n t, t o ta l a n d b y b ir th c o h o rt D is ap p ro v al c h ild le ss n e ss D is ap p ro v al U n m ar ri e d C o h ab it at io n D is ap p ro v al o u t- o f- w e d lo ck b ir th 1 9 4 1 -1 9 5 0 1 9 5 1 -1 9 6 0 1 9 6 1 -1 9 7 2 1 9 4 1 -1 9 5 0 1 9 5 1 -1 9 6 0 1 9 6 1 -1 9 7 2 1 9 4 1 -1 9 5 0 1 9 5 1 -1 9 6 0 1 9 6 1 -1 9 7 2 A u st ri a 2 9 .9 2 4 .8 2 4 .0 1 6 .1 1 0 .2 7 .5 2 1 .4 1 5 .3 1 3 .2 B e lg iu m 2 2 .4 1 3 .7 1 1 .9 1 0 .0 6 .8 4 .1 1 4 .4 1 1 .1 7 .0 B u lg ar ia 8 2 .3 8 0 .7 7 7 .7 4 4 .4 3 4 .3 2 7 .0 4 4 .0 3 2 .8 2 8 .9 C y p ru s 6 6 .7 5 8 .5 6 1 .9 5 5 .3 2 8 .2 1 9 .2 6 6 .7 4 2 .3 3 9 .8 D e n m ar k 4 .7 2 .2 5 .0 8 .6 2 .9 3 .6 7 .8 8 .5 5 .6 E st o n ia 8 1 .2 7 7 .2 6 5 .7 3 8 .0 2 4 .4 1 8 .4 3 6 .8 2 6 .1 2 1 .1 Fi n la n d 1 7 .8 1 3 .7 1 1 .6 6 .8 3 .3 3 .5 1 2 .7 6 .6 6 .3 Fr an ce 3 7 .3 2 3 .6 2 4 .4 9 .6 1 1 .0 7 .4 1 3 .7 1 2 .8 9 .2 G e rm an y 2 8 .6 2 0 .0 1 8 .0 8 .6 5 .6 6 .2 1 7 .6 1 0 .7 9 .8 H u n g ar y 5 6 .7 4 6 .4 4 3 .1 1 4 .2 9 .7 9 .9 1 5 .1 1 0 .9 1 0 .6 Ir e la n d 1 8 .7 1 1 .3 1 1 .3 2 5 .2 1 3 .2 1 1 .8 3 3 .9 2 0 .1 1 8 .2 N e th e rl an d s 1 3 .1 1 0 .1 8 .2 9 .7 8 .5 8 .0 1 2 .6 1 2 .1 9 .6 N o rw ay 6 .8 6 .2 3 .5 6 .8 3 .4 4 .6 7 .6 4 .7 4 .8 P o la n d 6 2 .3 5 3 .7 5 0 .4 4 4 .1 3 5 .5 2 9 .8 3 3 .2 3 2 .2 2 5 .4 P o rt u g al 2 6 .1 2 2 .0 1 6 .2 1 6 .1 1 1 .6 7 .1 1 6 .4 1 1 .6 7 .1 R u ss ia n F e d e ra ti o n 8 6 .4 8 2 .1 8 1 .5 4 7 .0 3 4 .6 3 0 .3 4 6 .3 3 6 .2 3 3 .6 S lo v ak ia 6 0 .8 6 1 .6 4 7 .6 4 6 .4 3 5 .8 2 6 .6 5 0 .0 4 0 .1 3 5 .1 S lo v e n ia 4 9 .8 3 9 .1 3 2 .7 2 3 .1 1 6 .9 1 3 .5 1 9 .5 1 8 .4 1 2 .7 S p ai n 3 8 .6 1 8 .4 1 5 .0 1 6 .2 8 .5 8 .5 2 4 .3 1 1 .9 9 .5 S w e d e n 3 .8 5 .8 5 .2 2 .9 2 .0 3 .3 4 .1 3 .7 5 .8 S w it ze rl an d 1 1 .8 1 3 .3 1 2 .9 7 .9 5 .4 1 0 .7 1 9 .5 1 3 .2 1 5 .7 U kr ai n e 8 6 .7 8 3 .6 7 5 .3 5 5 .3 5 2 .1 3 3 .3 5 6 .3 4 9 .4 3 6 .4 U n it e d K in g d o m 3 .1 5 .6 5 .7 1 0 .8 7 .9 9 .1 2 1 .7 1 6 .2 1 5 .6 * D is ap p ro v al s u m m ar is e s th e a n sw e rs “ S tr o n g ly d is ap p ro v e ”/ “d is ap p ro v e ” o n a 5 -p o in t Li ke rt s ca le . S o u rc e : E u ro p e an S o ci al S u rv e y W av e 3 ( o w n c al cu la ti o n s) Patterns of Family Life Courses in Europe – between Standardisation and Diversity • 579 On the other hand, tolerance toward deviations from a standard family life trajec- tory is clearly highest in Scandinavia where less than one tenth of all respondents object to any of the aforementioned items (with Finland where 14.5 percent reject childlessness as the only exception). This group is joined by the UK, the Netherlands and Switzerland, where similarly “liberal” norms prevail. Most Southern European as well as Central European countries occupy a position in between these two ex- tremes with around a quarter of respondents rejecting childlessness but less than a fi fth expressing scepticism towards the other two items. In most of the countries, individuals in the oldest cohort (born 1941-1950) exhibit the most “traditional” family values while the two younger cohorts (and especially those born 1961-1972) seem to follow more “liberal” family norms. Value changes appear to be most pronounced regarding the stigmatisation or acceptance of un- married cohabitation where disapproval values among the youngest cohorts often only amount to half of that among the older cohort. Only for countries where rejec- tion of life course deviations was already low in the fi rst cohort, we fi nd no clear and systematic pattern across birth cohorts. Liberalisation trends appear to be most pronounced in Southern European countries. In many Eastern European countries, however, the picture of largely conservative family norms remains persistent. To investigate the relationship between family norms and family structures in greater detail (H3), we select unmarried cohabitation as one exemplary life stage for which both actual occurrence and normative acceptance have increased most signifi cantly. Figure 1 contrasts the incidence of unmarried cohabitation and its nor- mative rejection for the oldest cohort (i.e. those born 1941-1950) with that of the youngest cohort (i.e. born 1961-1972). Looking fi rst at results for the oldest cohort (1940-1951) three country group- ings can be differentiated. On the one hand, there is a group of countries mostly consisting of Northern European nations in which unmarried cohabitation is largely accepted (with less than 10 percent rejection) and its incidence is rather high (rang- ing between 30 and 60 percent). On the other hand, there is a group of countries mostly consisting of Eastern Europe as well as Cyprus where unmarried cohabita- tion is still rejected by a majority of the population (i.e. 40-60 percent) and only oc- curs very occasionally. In between the two clusters, there is a heterogeneous group of nations (including Continental and Southern European countries as well as Great Britain, Hungary and Slovenia) where less than a third of the population rejects the idea of unmarried cohabitation but where it simultaneously remained a behaviour of a minority practiced by less than a third of the cohort under study. The comparison of these results with those for the youngest cohort (i.e. 1961- 1972) suggests that both behaviour and norms are changing. Especially in Scandi- navian countries, where normative rejection of unmarried cohabitation has been virtually non-existent, it has turned into a new “institutionalised” norm. In this case, family behaviour seems to have temporally followed normative ideas (not neces- sarily in the sense of a unilateral causal relationship), as transitional phases such as unmarried cohabitation or births out of wedlock have turned into widespread behavioural practices, especially among younger cohorts. This pattern is also in- creasingly found in conservative and liberal countries where family behaviour has • Dirk Hofäcker, Jana C haloupková580 Fig. 1: Actual incidence (y-axis) and normative rejection (x-axis) of unmarried cohabitation (percentages, cohorts 1941-1950 vs. 1961-1970) Source: European Social Survey Wave 3 (own calculations) Cohort 1941-1950 Lived in unmarried cohabitation; in % Cohort 1961-1972 Lived in unmarried cohabitation; in % 0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 SWE DK POR NO FI CH D NL UK BE HU ES A SLV IRL EE PL BUL SLK RUS UKR CYP 0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 DK SWE FI F NO CHA NL UKD BE IRL SLV EE PO CYHUN ES SLK BUL RUS POL UKR Disapproval of unmarried cohabitation; in % Disapproval of unmarried cohabitation; in % Patterns of Family Life Courses in Europe – between Standardisation and Diversity • 581 gradually converged with family norms on non-marital cohabitation among the younger cohort. The de and subsequent reinstitutionalisation of family norms thus was accompanied by a de and subsequent restandardisation of family trajectories across generations. For Southern European countries, our fi ndings point to notable tensions be- tween the patterns of family trajectories and their normative equivalence (at least with respect to the exemplary family type of unmarried cohabitation), as family norms have liberalised while family life and family trajectories still appear to follow rather traditional lines. Change of family norms in these countries has progressed signifi cantly while family life has rather remained standardised. Normative deinstitutionalisation towards unmarried cohabitation can also be ob- served among the majority of Eastern European countries. While among those born in the 1940s, about half of the individuals rejected unmarried cohabitation; resist- ance only prevails among a third of the 1960s cohort. At the same time, however, behaviour has remained remarkably stable: less than a third in fact live in unmarried cohabitation. The dynamics in family trajectories have been clearly less pronounced than those with regard to normative change. These results reveal a process in which conformity between traditional family norms and practices among older birth co- horts has turned into an ambivalent situation of non-traditional family norms de- spite traditional family practices among the younger population. The situation in these two regimes appears to be well in line with the descrip- tion of “gradual” transitions from one dominant demographic behavioural pattern to another. Even though norms about family behaviour have been relaxed among the younger population, there still is a largely prevalent cultural orientation at a “traditional” family model within the overall population, refl ected in around 50 per- cent rejection of unmarried cohabitation. An individual may be affected primarily by dominant social norms within the same cohort (from which peers are typically recruited) but also by acceptance among people from other age groups. This “nor- mative climate” may still present an obstacle for younger cohorts when it comes to adapting their behaviour. Examples from Central European countries indeed seem to indicate that only once a “critical point” of tolerance is surpassed – in the exam- ple chosen here: an open rejection by only around 10 percent of the population or less – traditional norms do no longer form an obstacle for individual adaptation of demographic behaviour among the majority of the population. Even though the existence and identifi cation of such a “threshold” is not verifi ed by our analysis, the assembly of countries in fi gure 1 may be interpreted as a typical pathway of change. 7 Summary In this paper, we have used data from the third wave of the ESS to investigate in how far family trajectories and family-related social norms have changed over time as well as across European countries. Our results provide support for the assumption of an increasing restandardisation of family lives. In most of the analysed countries, family trajectories have become more turbulent involving more stages and stage • Dirk Hofäcker, Jana C haloupková582 changes for the younger generation. This increasing turbulence has led to a ris- ing “dissimilarity” from traditional family life courses in most European countries (destandardisation). While marriage still constitutes the dominant “fi nal stage” in family trajectories, there is an increasing share of individuals across birth cohorts who experiences “intermediate” family phases such as unmarried cohabitation or parenthood before marriage, furthermore, an ever greater share of younger cohorts appears to either postpone fi rst birth beyond the age of 35 (the age window used in this study) or to entirely forego parenthood. In some countries, we fi nd fi rst evi- dence that previous “deviations” from traditional family patterns (such as unmarried cohabitation) are turning into majority behaviour, i.e. into a “new standard”, thus suggesting a “restandardisation of family life courses”. However, in line with our regime and social norms hypotheses (H2, H3), we fi nd that the degree and speed of these changes is very different across European re- gions. Destandardisation appears to be most developed in Northern and Central Eu- ropean countries where family trajectories had moved away from standard “ideal” family trajectories with a higher degree of turbulence in family transitions. These countries also show fi rst signs of a restandardisation. On the other hand, Eastern European countries display a higher proximity to traditional standard life courses and lesser turbulence in family lives and only rather modest signs of destandardi- sation. Most Southern European countries take an intermediate position between these two regimes with persistently moderate degrees of destandardisation and turbulence in family transitions. Contrasting the results above showing changes in family trajectories with shifts in family norms allowed us to fi nd indications for the mutual interrelations between these two dynamics from a comparative perspective. Our results point to a notable, yet not complete correspondence between norms and demographic behaviour. In fact, the liberalisation in norms appears to precede changes in actual demographic behaviour. Given the cross-sectional and macro-oriented nature of the analyses conducted, this does not prove a unilateral causality; yet it provides some plausi- bility for the assumption that normative change (that might be fostered by innova- tive family life trajectories of a minority as visible examples) may provide the basis for change in family trajectories among the majority of the population. As long as norms stand in stark contrast to specifi c types of family behaviour, individuals will shy away from such behaviour in order to minimise social cost. However, once re- jection of specifi c types of behaviour decreases below a certain critical level, chang- es in demographic behaviour may occur quickly across cohorts. Our results indicate that European countries appear to have reached different stages of this continuum, with especially social democratic countries having established non-marital relation- ships and parenthood as a new “standard type” of the life course, while especially Eastern European countries still largely follow a traditional pattern. Despite the fact that our study helped to shed some light on the relationship between family structures and family norms and their regime-specifi c development over time, it apparently also entails some limitations. As indicated, family norms could only be collected at the time of the interview; this implies that we probably underscore generational differences in norms at the time of family formation at least Patterns of Family Life Courses in Europe – between Standardisation and Diversity • 583 in part due to possible attitudinal adaptation across time. Furthermore, due to the lack of longitudinal data on normative orientations, we were not able to establish causality between changes in norms and changes in family behaviour. These limita- tions call for a more comprehensive coverage of attitudinal data in cross-national longitudinal surveys in order to substantiate the relationship between the structural and normative dimension of social change. Furthermore, given our objective to cov- er entire family trajectories of individuals from legal age to their mid-30s, we could only consider birth cohorts up to the early 1970s. Especially for Eastern European countries but also for some post-despotism Southern European states such as Por- tugal or Spain, country-study evidence suggests that substantial transformations in family trajectories have taken place among younger birth cohorts which may only be reliably ascertained by future studies. Acknowledgement This collaborative research was made possible through a Short Visit Grant of the European Science Foundation (ESF) Social Science Research Network TransEurope (no.3596). Jana Klímová Chaloupková also thanks for the support from the Grant agency of the Academy of Science of the Czech Republic (project no. KJB700280802) and the Grant Agency of the Czech Republic (project no. P404/11/0145). References A bbott, Andrew; Tsay, Angela 2000: Sequence Analysis and Optimal Matching Methods in Sociology: Review and Prospect. In: Sociological Methods Research 29,3: 3-31 [doi: 10.1177/0049124100029001001]. A isenbrey, Silke; Fasang, Anette E. 2007: Beyond Optimal Matching: The “Second Wave” of Sequence analysis. New Haven: Yale University, The Center for Research on Inequalities and the Life Course. B aranowska, Anna 2008: Changing patterns of entry into employment and motherhood in Poland a cross cohort comparison. Social Sciences Research Network TransEurope. Working Paper No. 5. Bamberg: Otto-Friedrich-University Bamberg. B arban, Nicola; Billari, Francesco 2012: Classifying life course trajectories: a com- parison of latent class and sequence analysis. In: Applied Statistics 61,5: 765-784 [doi: 10.1111/j.1467-9876.2012.01047.x]. Beck, Ulrich 1992: Risk Society: Towards a New Modernity. London: Sage. Be ier, Loreen; Hofäcker, Dirk; Marchese, Elisa; Rupp, Marina 2010: Familyplatform. Ex- istantial Field 1: Family Structures & Family Forms in the European Union – An over- view of major trends and developments. Working Report – Summary (April 2010). State Institute for Family Research (ifb), University of Bamberg: Bamberg. Bil lari, 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]. Bil lari, Francesco C.; Liefbroer, Aart C. 2007: Should I stay or should I go? The impact of age norms on leaving home. In: Demography 44: 181-198 [doi: 10.1353/dem.2007.0000]. • Dirk Hofäcker, Jana C haloupková584 Blossfeld, Hans-Peter; Klijzing, Erik; Mills, Melinda; Kurz, Karin (Eds.) 2005: Globaliza- tion, Uncertainty and Youth in Society. London: Routledge. Blossfeld, Hans-Peter; Hofmeister, Heather 2006: Globalization, Uncertainty and Wom- en’s Careers: An International Comparison. Cheltenham and Northampton: Edward Elgar. Brückner, Hannah; Mayer, Karl Ulrich 2005: De-standardization of the life course: what it might mean? And if it means anything, whether it actually took place? In: Advances in Life Course Research 9: 27-53 [doi: 10.1016/S1040-2608(04)09002-1]. Brzinsky-Fay, Christian; Kohler, Ulrich; Luniak, Magdalena 2006: Sequence analysis with Stata. In: Stata Journal 6,4: 435-460 [URL: http://www.stata-journal.com/article. html?article=st0111, 08.09.2014]. Buchmann, Marlis 1989: The Script of Life in Modern Society. Entry into Adulthood in a Chan ging World. Chicago: The University of Chicago Press. C haloupková, Jana 2010: De-standardization of Early Family Trajectories in the Czech Republic: A Cross-cohort Comparison. In: Sociologický časopis/Czech Sociological Review 46,3: 427-451 [URL: http://sreview.soc.cas.cz/en/issue/112-sociologicky-caso- pis-czech-sociological-review-3-2010/2035, 08.09.2014]. Chanvril, Flora et al. 2009: La parentalité en Europe. Analyse séquentielle des trajec- toires d’entrée dans l’âge adulte à partir de l’Enquête sociale européenne. Dossiers d’Etudes 122. Luxembourg: Caisse Nationale d’Allocation Familiale. Corijn, Martine; Klijzing, Erik 2001: Transitions to adulthood in Europe. Dordrecht: Klu- wer Academic Publishers. Elzinga, Cees H.; Liefbroer, Aart C. 2007: De-standardization of Family-Life Trajectories of Young Adults: A Cross-National Comparison Using Sequence Analysis. In: Euro- pean Journal of Population 23,3-4: 225-250 [doi: 10.1007/s10680-007-9133-7]. Esping-Andersen, Gosta 1990: The Three Worlds of Welfare Capitalism. Cambridge: Pol- ity Press. Fokkema, Tineke; Liefbroer Aart. C. 2008: Trends in living arrangements in Europe: Con- vergence or divergence? In: Demographic Research 19: 1351-1417 [doi: 10.4054/Dem- Res.2008.19.36]. Furlong, Andy; Cartmel, Fred 1997: Young People and Social Change: Individualization and Risk in Late Modernity. Buckingam: Open University Press. Fussell, Elizabet h 2005: Measuring the early adult life course in Mexico: an application of the entropy index. In: Advances in Life Course Research 9: 91-122 [doi: 10.1016/ S1040-2608(04)09004-5]. Fussell, Elizabet h; Gauthier, Anne H.; Evans, Ann 2007: Heterogeneity in the Transition to Adulthood: The Cases of Australia, Canada, and the United States. In: European Journal of Population 23: 389-414 [doi: 10.1007/s10680-007-9136-4]. Gabadinho, Alexis; Ritschard, Gilbert; Studer, Matthias; Müller, Nicalas S. 2009: Min- ing sequence data in R with the TraMineR package: A user ś guide for version 1.2. Geneva: University of Geneva. Gauthier, Anne 200 7: Becoming a Young Adult: An International Perspective on the Transitions to Adulthood. In: European Journal of Population/Revue européenne de Démographie 23,3: 217-223 [doi: 10.1007/s10680-007-9130-x]. Gradstein, Mark; Nikitin, Denis 2004: Educational Expansion: Evidence and Interpreta- tion. World Bank Policy Research Working Paper 3245. Washington: The World Bank. Patterns of Family Life Courses in Europe – between Standardisation and Diversity • 585 Halman, Loek C.J.M. 2009: Value change in Western European societies: Results from the European values study. In: Kwansei Gakuin University School of Sociology Journal 107: 35-48. Kohli, Martin 2007: The institutionalization of the life course: Looking back to look ahead. In: Research in Human Development 4,3-4: 253-271 [doi: 10.1080/15427600701663122]. Lesnard, Laurent 200 6: Optimal Matching and the Social Sciences. Document de travail du Crest. Paris: INSEE. Liefbroer, Aart C. 2009: From youth to adulthood: understanding changing patterns of family formation from a life course perspective. In: Heinz, Walter R.; Huinink, Jo- hannes; Weymann, Ansgar (Eds.): The life course reader: individuals and societies across time. Frankfurt/Main et al.: Campus Reader: 311-337. 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.; Merz, Eva-Maria 2009: Reproductive decision-making in a macro- micro perspective: report on analysis of ESS data on cross-national differences in perceived norms concerning fertility-related behaviour to the European Commission within the 7th Framework Programme. The Hague: Netherlands Interdisciplinary De- mographic Institute (NIDI). Modell, John; Furstenberg Jr., Frank F.; Hershberg, Theodore 1976: Social Change and Transitions to Adulthood in Historical Perspective. In: Journal of Family History 1,1: 7-32. Macindoe, Heather; Abbott, Andrew 2006: Sequence Analysis and Optimal Matching technique for Social Science Data. In: Hardy, Mellissa; Bryman, Alan (Eds.): Handbook of Data Analysis. London: Sage Publications: 387-406. Moors, Guy B.D.; Lesthaeghe, Ron J. 2000: Life course transition and value orientations: selection and adaptation. In: Demografi a 4,XLIII: 405-445. Rossier, Clémentine 2 010: Variation in social norms and practices of social infl uences in different family and fertility cultures and specifi c political economies. REPRO report. Vienna: Austrian Academy of Sciences. Scott, Jacqueline; Braun, Michael 2006: Individualisation of family values? In: Ester, Pe- ter; Braun, Michael; Mohler, Peter (Eds): Globalization, Value Change, and Genera- tions: A Cross-National and Intergenerational Perspective. Brill: Leiden: 61-87. Scott, Jacqueline 2006: Family and Gender Roles: How Attitudes are Changing. In: GeNet Working Paper 21. Settersten, Richard A.; Hagestad, Gunhild O. 1996: What’s the Latest? II. Cultural Age Deadlines for Educational and Work Transitions. In: The Gerontologist 36,2: 602-613. Sobotka, Tomáš; Toulemon, Laurent 2008: Changing family and partnership behaviour: Common trends and persistent diversity across Europe. In: Demographic Research 19: 85-138 [doi: 10.4054/DemRes.2008.19.6]. Sobotka, Tomáš 2008: The diverse faces of the second demographic transition in Eu- rope. In: Demographic Research 19,8: 171-224 [doi: 10.4054/DemRes.2008.19.8]. van de Kaa, Dirk 1987: Europe ś Second Demographic Transition. In: Population Bulletin 42,1: 1-57. van Oorschot, Wim; Opielka, Michael; Pfau-Effi nger, Birgit (Eds.) 2008: Culture and wel- fare state: Values and social policy in comparative perspective. London: Edward Elgar. • Dirk Hofäcker, Jana C haloupková586 Wagner, Michael; Franzmann, Gabr iele 2000: Die Pluralisierung der Lebensformen. In: Zeitschrift für Bevölkerungswissenschaft 25: 151-173. Wernhart, Georg; Neuwirth, Norbert 2007: Geschlechterrollenwandel und Familien- werte (1998-2002): Österreich im europäischen Vergleich. Ergebnisse auf Basis des ISSP 1998, 2002. oif Working Paper Nr. 54/2007. Vienna: Austrian Institute for Family Studies (oif). Widmer, Eric D.; R itschard, Gilbert 2009: The de-standardization of the life course: Are men and women equal? In: Advances in Life Course Research 14,1-2: 28-39 [doi: 10.1016/j.alcr.2009.04.001]. Date of submission: 21.02.2011 Date of acceptance: 02.07.2014 Prof. Dr. Dirk Hofäcker (). Universität Duisburg-Essen, Institut für Soziale Arbeit und Sozialpolitik. Essen, Germany. E-Mail: dirk.hofaecker@uni-due.de URL: https://www.uni-due.de/biwi/hofaecker/zurperson.php Mgr. Jana Klímová Chaloupková, Ph.D. Institute of Sociology, Academy of Science of the Czech Republic, Prague, The Czech Republic. E-mail: jana.chaloupkova@soc.cas.cz URL: http://www.soc.cas.cz/en/lide/jana-klimova-chaloupkova Published by / Herausgegeben von Prof. Dr. Norbert F. Schneider Federal Institute for Population Research D-65180 Wiesbaden / Germany Managing Editor / Verantwortlicher Redakteur Frank Swiaczny Assistant Managing Editor / Stellvertretende Redakteurin Katrin Schiefer Language & Copy Editor (English) / Lektorat & Übersetzungen (englisch) Amelie Franke Copy Editor (German) / Lektorat (deutsch) Dr. Evelyn Grünheid Layout / Satz Beatriz Feiler-Fuchs E-mail: cpos@bib.bund.de Comparative Population Studies www.comparativepopulationstudies.de ISSN: 1869-8980 (Print) – 1869-8999 (Internet) Scientifi c Advisory Board / Wissenschaftlicher Beirat Paul Gans (Mannheim) Johannes Huinink (Bremen) Michaela Kreyenfeld (Rostock) Marc Luy (Wien) Clara H. Mulder (Groningen) Notburga Ott (Bochum) Peter Preisendörfer (Mainz) Zsolt Spéder (Budapest) Board of Reviewers / Gutachterbeirat Martin Abraham (Erlangen) Laura Bernardi (Lausanne) Hansjörg Bucher (Bonn) Claudia Diehl (Konstanz) Andreas Diekmann (Zürich) Gabriele Doblhammer-Reiter (Rostock) E.-Jürgen Flöthmann (Bielefeld) Alexia Fürnkranz-Prskawetz (Wien) Beat Fux (Salzburg) Joshua Goldstein (Berkeley) Karsten Hank (Köln) Sonja Haug (Regensburg) Aart C. Liefbroer (Den Haag) Kurt Lüscher (Konstanz) Dimiter Philipov (Wien) Tomáš Sobotka (Wien) Heike Trappe (Rostock) © Federal Institute for Population Research 2014 – All rights reserved