Jämförande studier av utbildning och familjebildning
Detta är ett komparativt projekt om betydelsen av utbildning för sannolikheten att uppnå ett stabilt parförhållande med barn, ett viktigt livsmål för de flesta människor. Jämförelser görs mellan Sverige, Frankrike och USA, tre länder med mycket olika nivåer på samhällsstöd till
barnfamiljer. Ett stabilt familjeliv ökar sannolikheten för ett högt individuellt välbefinnande, vilket i sin tur bidrar till ett samhälles produktivitet och välfärd. Att skapa ett stabilt familjeliv kräver dock resurser. Utbildning är det första steget för att uppnå dessa resurser. Kunskapen om den totala effekten av utbildning på ett stabilt familjeliv är mycket knapphändig, likaså under vilka livscykler utbildning har störst effekt. Komparativ forskning om detta fenomen saknas nästan helt. Vi undersöker utbildningens betydelse för fem av familjebildningens nyckelhändelser: första parförhållandet, första- och andrabarnsfödslar, övergången från samboende till äktenskap, samt separation och skilsmässa. Vi använder oss av livshistoriedata från den svenska levnadsnivåundersökningen (2000), den franska familjeundersökningen från 1999 samt USAs Family Growth Survey från 2002, samt använder analysmetoder som kontrollerar för icke-observerbar heterogenitet. Slutligen ägnar vi särskild uppmärksamhet åt hur statliga stöd riktade till barnfamiljer och familjens betydelse i Sverige kan reducera effekten av utbildning på sannolikheten att uppnå ett stabilt parförhållande med barn.
Slutredovisning
Project goals
The goals of the project were to study educational differentials in the family building process, including union formation, childbearing, and union dissolution. The project was designed to be comparative, investigating educational differentials in countries with distinctly different welfare regimes – Sweden, France and the United States. We planned to study educational differentials in the particular transitions using event-history methods and also to document the cumulative family experience of different birth cohorts by age 40. We also planned a study of educational policy and consequent enrolment on Swedish childbearing. Swedish data come from the 2000 Level of Living Survey, the Young Adult Panel Study, and Swedish register data. We also used the 1999 French Family Survey and the 1995 U.S. National Survey of Family Growth. In the course of our research, we discovered an undocumented error in the computer-assisted interview for the 2002 NSFG, the first such survey to include male respondents and were therefore unable to compare the family life course experiences of U.S. men Swedish and French men.
The three most important results from our studies are summarized as follows:
In Sweden and France, about half of women born in the 1950s experienced what we term a ‘standard’ family life course – having had two or more children with a single partner and remaining in the partnership by age 40. Women with higher education were less likely to have had any children at all, but if they did so, they were more likely than less well educated women to have their children in a stable partnership. High educated French and Swedish men were slightly more likely to have experienced a standard family life course by age 40 than the lowest educated men. The primary source of differences between women’s and men’s experience is due to the fact that better educated men are more likely to become fathers, in contrast to the negative differential for educated women. Experiences of Swedish and French women contrast markedly with those of women in the United States where less than 40% of women have experienced the standard family life course by age 40 and where educational differences are marked. Although highly educated U.S. women had relatively high rates of childlessness, that experience was overbalanced by their greater propensity to have children in stable unions, compared to low-educated women.
Given increases in separation and divorce, it is not surprising that older cohorts (born in the 1930s and 1940s) in Sweden and France were more likely than their younger counterparts to experience a standard family life course. What is surprising is the size of the difference – only 5-10% across cohorts. Even for older cohorts, at least 40% had not experienced a standard family life course by age 40. Educational differences (or lack thereof) were virtually the same for all three cohorts in France. In Sweden, educational differences found for the 1930s cohort were reduced over time because of declines among the less well educated in non-union childbearing and increases among the better educated in parental separation.
Despite the massive inflow of students to tertiary education in the 1990s, student status continued to dramatically reduce the risk of childbearing. This result means that low fertility among students is not a matter of selection (common determinants of the wish to postpone family building and the desire for higher education), but is an effect of the conditions of students’ lives. We found that birth risks among female students in all age-groups depended quite heavily on income. Thus, it is the reduced economic circumstances of students – not their study time or social norms about the ‘right’ time to have children – that produce later childbearing among those who obtain higher education in their early 20s.
New research questions
Although the three countries we studied have quite different welfare regimes with implications for inequality in family life, they have in common relatively high rates of childbearing and parental separation compared to other wealthy countries. We plan therefore to extend this research with studies of countries with ‘lowest-low’ fertility in which parental separation is often also very unlikely. In such countries, the influence of education on childlessness may be the primary mechanism through which educational differentials in the standard life course arise, and this may have implications for differentials among women compared to men in their experiencing a standard family life course. We have just joined a new European network for the study of relationships between union formation or dissolution and childbearing through which we can obtain the appropriate data.
Throughout our analyses of Swedish and U.S. data we have been hindered by relatively small sample sizes. The French survey produced much more robust estimates because of the large number of respondents in each birth cohort. We are in process of solving this problem to some extent for the Swedish case. We are generating a ‘pooled’ data set using old and new surveys with similar information on education, birth and union histories. And we have developed a register-based dataset with Statistics Sweden and colleagues at the Swedish Institute for Social Research that can be used to estimate non-union childbearing and parental separation for the entire Swedish population from 1968 through the most recent year available.
Finally, recent research suggesting that not only educational level but also type of education influences family building. Educational programs leading to occupations characterized by care work and/or work schedules that are more compatible with child care appear to be associated with higher fertility. This work can be developed further to examine the relationship between educational type and parental separation and also to determine whether it is indeed one’s later occupational conditions that matter or whether some other process produces the association. The aforementioned register-based dataset will enable investigation of these questions for Sweden, at least from 1985.
Two most important publications
Thomson, Elizabeth, Maria Winkler-Dworak, and Sheela Kennedy. Forthcoming. Education and the Family Life Course. In Ann Evans and Janeen Baxter (eds.), Negotiating the Life Course. Springer. Also Stockholm Research Reports in Demography 2009:3 (www.suda.su.se) This chapter provides the argument for studying the standard family life course and presents the comparative data on education for the 1950s cohort.
Thalberg, Sara. 2009. Childbearing of Students. The Case of Sweden. Institute for Futures Studies, Working paper 2009:X. This report provides an overview of changes in tertiary education and its relationship to childbearing across periods of differential levels and types of study support and through the economic crisis of the 1990s. It is being developed into three chapters for Thalberg’s dissertation.