Because the prevalence of nonmarital childbearing remains so low in Japan, bridal pregnancy is perhaps a more informative indicator of changing relationships between pregnancy and marriage. We again use information on the timing of marriage and first birth to define bridal pregnancy as cases in which the first birth occurs within the first seven months of marriage. This is the conventional approach used to define “shotgun marriages” (England et al. 2012).
Figure 4.4 presents trends, by educational attainment, in the predicted probability of bridal pregnancy. This pattern of change depicted in this figure is quite similar to the pattern of bifurcation in family outcomes described by McLanahan (2004) in her analyses of U.S. data. In the 1940s birth cohort, bridal pregnancy was relatively uncommon in all educational groups. University graduates were less likely than high school graduates to be pregnant at marriage, but the magnitude of the difference was small (3 vs. 6 %, respectively). The prevalence of bridal pregnancy has increased markedly across the educational distribution, but this change is most pronounced among women with a vocational school education or less. The gap between university graduates and high school graduates in the 1970s birth cohort was 15 % points, five times larger (in absolute terms) than the gap for the 1940s birth cohort. However, change in the relative difference is small and not
Fig. 4.4 Predicted probability of bridal pregnancy, by education and birth cohort
statistically different from zero, except for a significant growth in the gap for the 1960s birth cohort (relative to the gap for the 1950s birth cohort). This is one of the only examples in our analyses of a significant change across successive cohorts that does not correspond to a significant difference with the earliest cohort. Again, it appears that the 1960s birth cohort experienced particularly pronounced socioeconomic bifurcation in family behavior.
It is also interesting to note the pattern of bifurcation among women with intermediate levels of education. Those who completed vocational school closely resemble high school graduates (at least until the 1970s birth cohort), while junior college graduates more closely resemble university graduates. We are not aware of any research that carefully considers the composition of these two groups (with respect to background compositional characteristics and aspirations with respect to work and family), but further consideration of differences and similarities in family (and work) behavior relative to those with higher and lower education could shed light on this. We revisit this point below in the discussion section.