Race and Cohort Differences in Family Status in the United States

Abstract

In this visualization, the authors show changes in family patterns by different race groups across two cohorts. Using data from the National Longitudinal Survey of Youth 1979 (born from 1957 to 1965) and 1997 (born from 1980 to 1984), the authors visualize the relationship-parenthood state distributions at each age between 15 and 35 years by race and cohort. The results suggest the rise of cohabiting mothers and the decline of married and divorced mothers among women born from 1980 to 1984. Black women born from 1980 to 1984 were more likely to experience single/childless and single/parent status compared with Black women born from 1957 to 1965. Although with some visible postponement in the recent cohort, white women in both cohorts were more likely to experience married/parent status than other race groups. The decline in married/parent status across the two generations was sharpest among Hispanic women. These descriptive findings highlight the importance of identifying race when discussing changes in family formation and dissolution trends across generations.

Studies suggest that recent family formation and dissolution patterns in the United States are different from those of earlier generations: women are less likely to marry and less likely to become mothers (Schaeffer and Aragão 2023). In place of marriage, cohabitation has become the most common union status among young adults (Manning, Smock, and Fettro 2019). Given that marriage rates have declined, it is not surprising that divorce rates have also declined (Cohen 2019). Most visualizations of these demographic trends have been based on cross-sectional data sources, such as the American Community Survey (Cohen 2019; Kennedy and Ruggles 2014). In prior studies, less attention has been paid to how these patterns are diversified by race and across generations in the United States from a longitudinal perspective. We visualize these trends by taking a longitudinal approach (group-based sequence density plots) to highlight race and cohort differences in family formation and dissolution patterns in the United States.

Data and Method

Two longitudinal data sources are used to tease out the “cohort change” between those born from 1957 to 1965 and those born from 1980 to 1984. The first comprises the 1979 to 1998 waves of the National Longitudinal Survey of Youth 1979 (NLSY79) and the second includes the 1997 to 2019 waves of the National Longitudinal Survey of Youth 1997 (NLSY97). We use the most recent interview year of the NLSY97 and capture the same age observation window for all respondents in the two datasets. For data comparability, we include young adult women living in the United States, aged between 15 and 35 years. This age observation window also allows us to capture the average age of first-time mothers, which ranged from 21.4 to 27.0 years between 1979 and 2019 (Martin et al. 2021; Mathews and Hamilton 2002). After appending both samples, we observe 10,668 women. After selecting only those women who have complete imputed histories and excluding those with racial affiliations other than Black, Hispanic, or white, our final sample ¹ consists of 6,125 biographies, composed of 3,643 from NLSY79 and 2,482 from NLSY97.

We use sequence density plots to visualize the most frequent relationship-parenthood status observed during the early stages of women’s lives across two generations (women born from 1957 to 1965 and women born from 1980 to 1984) by race. Visualizations of categorical sequences are widely used in social science, especially in life-course research (Cornwell 2015; Fasang and Liao 2014). By looking at the distributions by age in each group, and organizing our “state categories” such that we can also observe the percentages that have transitioned to parenthood, we are both able to describe group changes in the tempo (timing of first childbirth) and the quantity of the observed relationship-parenthood statuses. The change in the timing of the transition to parenthood is observable when 50 percent of the population are already parents, and the status frequency is measured as a percentage of person-years. We use the package TraMiner in R, developed by Gabadinho et al. (2011) to visualize life sequences. Using the seqdplot() function with a grouping variable, we can see the state distributions for each race and cohort group. By using this approach, we describe how family behavior has evolved between racial groups and cohorts.

Findings

The sequence density plots provide the age-specific relationship-parenthood status distributions by race (Black, Hispanic, and white) and cohort (women born from 1957 to 1965 and women born from 1980 to 1984) groups in the United States. On the x-axis, we observe the age categories in yearly progression from 15 to 35 years. On the y-axis, each color represents partnership-parenthood status, measuring the proportion of women in each status at each age unit for each group.

The categories we observe as possible partnership-parenthood status types are (1) single/childless, (2) cohabiting/childless, (3) married/childless, (4) divorced/childless, (5) single/parent, (6) cohabiting/parent, (7) married/parent, and (8) divorced/parent. ²

Figure 1 displays the distribution of these status categories at each age by group.

Figure 1.

Sequence density plots for women born between 1957 and 1964 versus women born between 1980 and 1984 by race.

In the top row, the recent cohort of Black women’s sequence density plot shows that for this observed age window, a higher proportion of Black women (32 percent of all the person-years) compared with the older cohort of Black women (27 percent of the person-years) experienced single/parent status. Yet a lower proportion of the recent cohort of Black women (3 percent of person-years) compared with the older cohort of Black women (10 percent of the person-years) had experienced, so far, divorced/parent status. A significant number of Black women born from 1980 to 1984 experienced spells of married parenthood, particularly in their 30s, but married parenthood was not the most prevalent parental status among them, compared with Black women born from 1957 to 1965. Single parenthood became more prevalent among this group in their late 20s and early 30s. Regarding the timing of first birth, this changed only slightly from one cohort to another among this group: 50 percent of the women born from 1957 to 1965 had children by age 21, while 50 percent of women born from 1980 to 1984 had children by age 22.

Among Hispanic women, the most prevalent family status for both cohorts was married/parent, although the proportion of married/parent status was lower among Hispanic women born from 1980 to 1984. As the proportion of married/parent status decreased across generations, the percentages of women in cohabiting parenthood (5 percent to 12 percent of the person-years) and single parenthood (8 percent to 10 percent of the person-years) were higher among Hispanic women born from 1980 to 1984. The decline in married parenthood has been more pronounced among Hispanic women born from 1980 to 1984 compared with Black counterparts, but it has been similar in level to that of white women born from 1980 to 1984. As for the timing of the transition to parenthood, we observe some postponement. Although the median age for women born from 1957 to 1965 was 22 years, for women born from 1980 to 1984, it was 24 years.

White women born from 1980 to 1984 show the highest percentage of married status (34 percent of the person-years) across different race groups born from 1980 to 1984 and a similar level of married/parent status as Hispanic women in the same birth cohort (24 percent of the person-years). Before age 30, white women born from 1980 to 1984 were more likely to experience cohabiting/childless or married/childless status, which was rare among white women born from 1957 to 1965. These trajectories reflected the rise of childlessness in the United States (Hayford 2013). However, only a small proportion of white women born from 1980 to 1984 remained childless while cohabiting or married by age 35. We can also observe postponement behavior among white women born from 1980 to 1984: although white women born from 1957 to 1965 had a median age of first birth of 26, the majority of white women born from 1980 to 1984 waited until at least the age of 28.

In sum, transitions to divorce became rarer across all race groups among women born from 1980 to 1984, as marriage rates declined, but cohabitation increased across all race groups in the recent cohort of women born from 1980 to 1984.

Summary

Our visualization highlights the diverse relationship-parenthood living arrangements in the recent cohort of women (born from 1980 to 1984) compared with the older cohort of women (born from 1957 to 1965) and the importance of understanding the different patterns by race among recent cohorts of women. Despite the common trends of family trajectories across cohorts, such as the decline in marriage and divorce and the rise of cohabitation and singleness, we find racial differences in the magnitude of these changing trends, as the decline in married and divorced parenthood has been more pronounced among Black and Hispanic women born from 1980 to 1984 and the postponement in the transition to parenthood is more prevalent among white and Hispanic women born from 1980 to 1984.

Supplemental Material

sj-docx-1-srd-10.1177_23780231241241041 – Supplemental material for Race and Cohort Differences in Family Status in the United States

Supplemental material, sj-docx-1-srd-10.1177_23780231241241041 for Race and Cohort Differences in Family Status in the United States by Misun Lim and Cristina Samper Mejia in Socius