The subject of this dissertation is trends in family formation among highly educated men and women in Sweden. The highly educated have typically differed from other educational groups in their patterns of childbearing. This has particularly been

the case for highly educated women, who used to be in the minority among the highly educated and who were much more likely to be childless than other women. The goal of this dissertation is to understand how the expansion of higher education has transformed  the formation of childbearing unions among the highly educated group. The context for the dissertation is the dramatic expansion of higher education which has occurred in Sweden over the last half century. As the share of cohorts graduating from post-secondary education has grown, diversity among the highly educated has also increased. This dissertation draws upon rich Swedish administrative register data to answer questions about changes in the behavior of the highly educated group, as well as emerging stratification within the group. This dissertation consists of five studies and an introductory chapter.

In Study 1, we examine changes in geographical distances between young couples and their parents. We find that among younger cohorts, generations live further apart. The expansion of higher education contributes to these distances, though the introduction of regional colleges has mediated the impact of educational expansion to some extent. In Study 2, we consider how effective colleges are as partner markets. We follow one birth cohort (1970), and examine the likelihood that they form a childbearing union with someone who attended the same university at the same time. We find that colleges are an important potential meeting place for childbearing partners, and examine how the likelihood of partnering with a fellow student is related to the college composition.

In Study 3, I assess changes in partner choice among the highly educated, by comparing the likelihood that highly

educated men and women born in 1940-2, 1950-2, 1960-2, and 1970-2 form a childbearing union, and whether they do

so with a highly or a lower educated partner. I find that female graduates are much more likely to enter unions, and to

“partner down”. Men’s likelihood of forming a childbearing union hasn’t changed across cohorts, but men from later cohorts are much more likely to find a highly educated partner than men from earlier cohorts. I show that partnership outcomes for graduates are related to social class background, university experience (degree length and institution type), and post-graduation income. In Study 4, we study unions with at least one highly educated partner, including men and women born in 1950-2, 1960-2, 1970-2, and 1980-2. We examine the extent to which educational (in)equality is mirrored in other measures of status such as social class background, income, and occupational prestige. We conclude that although the number of women “partnering down” in terms of education has increased dramatically, these unions are not necessarily characterized by female status-dominance more generally. In Study 5, I focus on highly educated men who do not form any childbearing union, studying men born in the years 1945-1974. I find that the consistent levels of childlessness among highly educated men may best be explained by changes in the composition of graduates in terms of field of study and post-graduation income.