Completed U.S. Fertility by Sex,Cohort,and Education Level

Abstract

This visualization shows the relationship between completed fertility and education across five decades of cohorts in the United States using data from the General Social Survey. Across these cohorts, the educational gradient of fewer children for those with more education is stronger for women than for men. The gradient has become less steep across cohorts as the gap between people with medium and high levels of education has narrowed. Those without high school degrees have the highest fertility rates at all times, but this group represents a decreasing share of the population as educational levels rise over time. Thus, the educational gradient in completed fertility persists but has become less important for overall fertility patterns.

Educational groups differ in how many children they have. For instance, people attending higher education often start a family later, which reduces their overall fertility (Trimarchi and Van Bavel 2017). People with higher education tend to earn more, so they have more to lose when reducing their working hours to care for children—but they also may be more able to afford the family size they desire (Guzzo and Hayford 2024). The patterns are complicated by unintended pregnancies, which are much more common in the United States than in other rich countries (Bearak et al. 2022), especially among those with lower levels of education (England 2016).

Fertility rates and their educational gradients differ by sex (Dudel and Klüsener 2021). We have less data on how many children men father because birth records often do not include information about the father and because men’s fertile years extend to older ages than women’s. In the Nordic countries, male fertility is higher among those with more education, and the gap has risen (Jalovaara et al. 2019; Kravdal and Rindfuss 2008). On the contrary, female fertility has been higher among the lower educated, but the gap is declining. This is partly because men with less education have low and declining marriage prospects, whereas the historical marriage penalty for highly educated women is diminishing (Hudde and Engelhardt 2023).

This visualization (Figure 1) provides a dynamic view of the trend in completed fertility (all the children a person has had in their lifetime) by educational level and across five decades of cohorts. Data are from the General Social Survey (GSS). Since 1972, the survey has asked: “How many children have you ever had?” We report the mean value of this variable by decade of birth, sex, and education, using the 32 surveys administered from 1974 through 2022, for respondents who are at least 40 years old (cf. Kravdal and Rindfuss 2008).

Figure 1.

Mean completed fertility (children ever born) by birth cohort in decades, sex, and education level. The height of the bars indicates the relative size of each education group within the sex cohort. Completed fertility is from the General Social Survey for respondents age 40 or older; cohort education composition is calculated from the decennial census (1950–1990) and American Community Survey (2000–2021) data from IPUMS.org. All data are weighted.

The GSS has the advantage of consistently asking a completed fertility question to both men and women (in the survey, sex is binary only) as well as educational degree attained. Although the sample for each sex cohort is large (from >6,000 for those born in the 1930s down to >2,000 for those born in the 1970s), cut by education, they get smaller, down to N = 97 (men with some college born in the 1930s) versus N = 2,218 (for women with high school degrees born in the 1940s). (We maximize sample size by including all respondents over age 40, which means earlier cohorts in our sample include more older people than the most recent cohort, but as Supplemental Figure 1 shows, this does not substantially affect the estimated completed fertility.)

Fertility trends depend on changes in both average group behavior and relative group sizes. The visualization accounts for this: The height of the bars shows the relative size of each education group within the sex cohort (from census data). The figure shows that the educational gradient in fertility for U.S. men and women—fewer children for those with more education—is more consistent for women. The gradient also has become less steep. (Supplemental Figure 2 provides a disaggregation by parity and shows that this reflects in both completed fertility and in the percentage having at least one child.) At the same time, the more highly educated groups increased their population share. For instance, the share of women with at least some college has doubled across cohorts. Those without high school degrees continue to stand out because they have considerably more children than the other groups, but their contribution to overall fertility levels decreases as their proportion of the population shrinks.

In conclusion, the educational gradient in completed fertility persists, unlike, for example, in the Nordic countries, but it has weakened and become less influential for overall fertility levels.

Supplemental Material

sj-docx-1-srd-10.1177_23780231241261610 – Supplemental material for Completed U.S. Fertility by Sex, Cohort, and Education Level

Supplemental material, sj-docx-1-srd-10.1177_23780231241261610 for Completed U.S. Fertility by Sex, Cohort, and Education Level by Ansgar Hudde and Philip N. Cohen in Socius