Patterns of marriage and divorce from ages 15 to 55: Evidence from the NLSY79

This article is an extension of previous work by Alison Aughinbaugh, Omar Robles, and Hugette Sun.¹ The earlier article examined patterns of marriage and divorce up to age 46 using the National Longitudinal Survey of Youth 1979 (NLSY79), a nationally representative sample of men and women who were born from 1957 to 1964 and were first interviewed in 1979. The previous article considered how marital history differed by educational attainment and age at marriage. The current article uses the same dataset, extending the previous analysis in two ways. First, we examine marriage and divorce patterns over a longer age range, from ages 15 to 55. Second, we consider how marital patterns varied by an additional dimension, work history. Like the previous work, this article is descriptive and does not seek to explain the underlying reasons for differences across groups.

Over the last 60 years, marriage and divorce rates in the United States have changed substantially. Marriage rates rose through the 1960s, then remained relatively stable until the mid-1980s, and have fallen since.² Divorce rates rose markedly from 1960 to 1980.³ After 1990, divorce rates continued to rise for older couples, but leveled off for younger couples. Divorce rates for adults over age 50 doubled between 1990 and 2010.⁴

The youngest of the baby boom generation, the cohort that makes up the NLSY79, lived in a dynamic time. They were young adults during the 1970s and 1980s when legal constraints around divorce loosened. And women’s labor force participation rose substantially as they reached adulthood, from about 40 percent in the 1960s to about 60 percent by the mid-1990s. Maternal labor force participation rose by even larger amounts, from about 20 percent in the 1960s to 70 percent by 1995.⁵

For earlier cohorts, in which couples were less likely to divorce and women were less likely to work, research found that marital history was related to economic security and health at older ages.⁶ For women born from 1911 to 1958, marital history affected women’s labor force participation at ages 50 to 74.⁷

Research based on the NLSY79 data showed that men who marry worked earlier in their working lives compared with those who never marry, and that marriage was associated with increased hours worked.⁸ Analysis by Professors Sarah Damaske and Adrianne Frech showed that substantial variation existed in women’s work histories, and that women from more advantaged backgrounds were more likely to work consistently throughout adulthood.⁹

For the men and women in the NLSY79 cohort, we anticipate that marital history will be related to economic well-being and work decisions as they reach traditional retirement ages. In this article, we show that marital history and work history, both measured up to age 55, were related.¹⁰

About 87 percent of men and women born from 1957 to 1964 had married by age 55. Among those who married, about 27 percent married more than once.¹¹ By age 35, 11 percent were in their second marriage. The percentage of men and women who remarried continued to rise as the cohort aged. For example, at age 35, 14 percent had married two or more times and by age 55, 27 percent had married two or more times. Women experienced more marital changes than men. Among men at age 55, 85 percent had married at least once, and 37 percent had ever divorced. In contrast, among women at age 55, 90 percent had married at least once, and 45 percent had ever divorced.¹²

Marriage patterns varied by educational attainment, with those with less than a high school diploma having had the lowest marriage levels. College-educated men and women first married at older ages compared with those with less schooling. Men and women who received a bachelor’s degree or higher were about equally likely to have married by age 55 (89 percent of men and 91 percent of women); the marriage rates for men and women with less than a high school diploma were significantly lower (71 percent of men and 83 percent of women). The likelihood that a marriage ended in divorce was lower for men and women with a bachelor’s degree or higher. Among those who did not earn a bachelor’s degree, at least half of first marriages ended in divorce compared with approximately 27 and 41 percent of first marriages of men and women with bachelor’s degrees, respectively.

Marriage patterns also varied by work history. Among men, those who spent more years in full-time work were more likely to marry and less likely to divorce. Among men who worked full time in at least 80 percent of the years from ages 25 to 54, 95 percent married, with 33 percent divorcing, by age 55. In contrast, women who worked more tended to be less likely to marry. Women who spent at least 80 percent of the same years working full time were about 7 to 8 percentage points less likely to have married compared with women who spent less than 60 percent of the years in full-time work. In addition, women who worked full time in 60 to 80 percent of their prime-age years (ages 25 to 54) were less likely than other women to remain in a first marriage and, conditional on divorcing, were less likely to remarry than women who worked full time in 15 to 30 percent or 30 to 60 percent of the 40 years examined. Women may have worked more in anticipation of divorce or in response to divorce.¹³

Data

The NLSY79 comprises men and women born from 1957 to 1964 and living in the United States when the U.S. Bureau of Labor Statistics (BLS) began the survey. The sample members were first interviewed in 1979 when they were ages 14 to 22. They were interviewed annually through 1994 and have been interviewed every 2 years since. We use data collected through the 2020–21 interview to document the patterns of marriage and divorce through midlife, when the youngest sample members were age 56.¹⁴

The NLSY79 is well suited for studying marriage and divorce patterns over the lifecycle. The NLSY79 collects detailed information on marital transitions in an event-history format, which allows one to determine the dates of marriage, divorce, and widowhood/widowerhood. In addition, the NLSY79 collects event-history data on other areas of respondents’ lives (including schooling, employment, and fertility) and also collects repeated measures of health, thus permitting researchers to study relationships between marital changes and other aspects of life.

The richness of the NLSY79 allows us to form a longitudinal marital history for each respondent. At each interview, NLSY79 respondents reported whether their marital status had changed since the date of their last interview. Respondents who had experienced a change in marital status were asked to list each change and report the type and date of that change. Using the respondent reports, survey staff created start dates and end dates for all marriages.

The current study examines marriage and divorce patterns from ages 15 to 55 using a sample of 5,766 men and women who had 6,658 marriages over those ages. The sample criteria used in this study required that a sample member participated in the survey at age 55 or older, reported valid dates for the start and end of all marriages, and reported his or her educational attainment in round 9 (1988) or later.¹⁵ We took the most recent reports of highest grade and degree completed and used them to measure educational attainment. In all results presented, the data were weighted using custom weights to make the sample representative of the population from which the NLSY79 is drawn.¹⁶

Tables 1, 2, and 3 provide information about the sample composition. The sample was composed of about 51 percent men and 49 percent women. Non-Black non-Hispanic men and women made up 80 percent of the sample, with Black and Hispanic men and women composing the remainder at 14 and 7 percent, respectively (see table 1).¹⁷ For the remainder of the article, the term “White” is used as shorthand for non-Black, non-Hispanic sample members; included in the White group are Asian, Pacific Islander, and Native American men and women as well as White men and women. Note that in the NLSY79 more than 90 percent of non-Black, non-Hispanic men and women were White.

The educational distribution in table 1 shows that 7 percent of the NLSY79 sample did not complete high school, 38 percent completed high school but did not go on to college, 24 percent attended some college including earning an associate degree, and 31 percent earned a bachelor’s degree or higher. We use five categories to describe work or employment history from ages 25 to 54, the ages over which individuals are prime-age workers, classifying by percentage of years that the respondent worked at least 2,000 hours, which can be thought of as working at least 40 hours per week, year round. The thresholds were chosen so that 15 to 25 percent of the sample falls into each category. The categories are working 2,000 or more hours in

• (1) less than 15 percent of the years,
• (2) 15 percent to less than 30 percent of the years,
• (3) 30 percent to less than 60 percent of the years,
• (4) 60 percent to less than 80 percent of the years,
• (5) and at least 80 percent of the years.¹⁸

The subsequent tables show that marital outcomes were strongly related to educational attainment and employment. Tables 2 and 3 provide additional breakdowns of the sample by educational attainment and employment history for men, women, and each of the racial and ethnic groups. The 31 percent of the sample that earned a bachelor’s degree or higher was evenly divided between men and women. However, sizeable differences in the proportion earning a bachelor’s degree existed across racial and ethnic groups. Among White men and women, 34 percent earned a bachelor’s degree or higher, but the percentages were lower for Hispanic and Black men and women at 19 percent and 21 percent, respectively (see table 2).¹⁹

Employment history differed substantially between men and women, with men spending more years working full time. (See table 3.) Among women, 31 percent were in the lowest employment-history category, working at least 2,000 hours in less than 15 percent of the 40 years considered.²⁰ The proportion of women in an employment-history category declined as the proportion of years in full-time work rose. That is, as we move up from the lowest employment-history category to the highest, the proportion of women in each level was smaller than in the previous level, ending with 9 percent of women working 2,000 hours or more in at least 80 percent of the years considered. The pattern was reversed for men. The percentage of men rose as the thresholds for percentage of years worked increased. Approximately 35 percent of men fell into the highest category of having worked full time in at least 80 percent of the years from ages 25 to 54 and 8 percent fell in the lowest category.²¹ For all three racial and ethnic groups, men outnumbered women in the two highest employment categories, with the margin greater for the highest employment category. This pattern was most prominent among White men and women. In the lowest employment category, the percentage of White women was roughly 5.5 times larger than the percentage of White men.

Results

In this section we examine marital histories, including timing of marriage and divorce, and their relationship to employment and educational attainment.

Marriage outcomes by age 55

Table 4 displays statistics for marriage, marriage duration, and divorce by sex, race, and ethnicity. By age 55, 87 percent of men and women had ever been married, 40 percent had ever been divorced, and 46 percent of those who had married had ever divorced. Men were less likely to have married than women (85 versus 90 percent), and conditional on having ever married, men were also less likely to have been divorced than women (43 versus 50 percent). The average age at first marriage is 26 for men and 24 for women.

At age 55, Black, non-Hispanic men and women were less likely to have been married (69 percent) than White (91 percent) or Hispanic (86 percent) men and women. Conditional on having been married, Black men and women were more likely to have divorced by age 55 compared with White men and women (51 versus 46 percent).

On average, Black men and women married later (age 27) than White (age 25) and Hispanic (age 24) men and women. Among those who had ever married, 55 percent of men and 46 percent of women were still in their first marriage at age 55. Black men and women who ever married were less likely to be in their first marriage at age 55 (43 percent) than White (52 percent) or Hispanic (48 percent) men and women.

Marriage outcomes differed within sex by education, particularly for men (see table 5). Men with less than a high school diploma were less likely to have ever married (71 percent) than those with higher levels of education (84 to 89 percent). The educational differences were smaller for women, with 83 percent of those with less than a high school diploma having ever married, compared with 90 to 91 percent for those with higher levels of education. Conditional on having married, men with a bachelor’s degree or higher were more than 20 percentage points less likely to get divorced than men with lower levels of education. About 27 percent of men with a bachelor’s degree or higher were ever divorced compared with 50 to 53 percent of men with lower levels of education. The divorce differential between ever-married women with a bachelor’s degree or higher and women with lower levels of education was smaller, with the largest differential at 15 percentage points; about 41 percent of those with a bachelor’s degree or higher ever divorced compared with 53 to 56 percent of women with lower levels of education.

Among those who married, the average age of first marriage increased with educational level. For example, men with a bachelor’s degree or higher had an average age at first marriage of 28 compared with 24 for men with less than a high school diploma. The average age at first marriage for women with a bachelor’s degree or higher was 26 compared with 21 for women with less than a high school diploma.

Table 6 displays marriage outcomes by sex and work history, in which work history is measured as the percentage of years from ages 25 to 54 that the individual worked at least 2,000 hours. The percentage of men who married at least once varied by work history; more full-time work was associated with a higher rate of marriage. At the extremes, men who worked at least 80 percent of years with full-time hours were much more likely to have ever married compared with those who worked less than 15 percent of years with full-time hours (95 versus 56 percent). Among men who married at least once, those with 60 percent or more of years with full-time hours were much less likely to get divorced than men who worked less than 60 percent of years with full-time hours (34 to 43 percent versus 54 to 57 percent). Conditional on divorce, among men, the likelihood of remarrying increased with the percentage of years spent working full-time. Among men, conditional on a second marriage, the likelihood of a second marriage ending in divorce decreased with more years of work.

Compared with men, the relationship between marriage and divorce outcomes and employment history for women was weaker and moves in the opposite direction. Women who worked the most were less likely to marry, with 84 percent of them marrying, compared to those who worked full time in less than 60 percent of the years, who married at rates of 91 to 92 percent. Conditional on marriage, women who worked from 60 percent to less than 80 percent of the years were at least 6 percentage points more likely to divorce compared with women in the highest work-history category and with women in the lower two work categories. Conditional on divorce, women in the lower two work categories remarried about 68 percent of the time compared with 62 to 63 percent of the time for the higher categories.

Number of marriages by age 55

Table 7 shows the number of marriages from age 15 through age 55 for people born from 1957 to 1964, by demographic characteristics and education. At age 55, only 13 percent of men and women in the NLSY79 cohort had never married, 60 percent married once, 22 percent married twice, and 6 percent married three or more times. On average, the cohort members had 1.2 marriages by age 55. Those with less than a high school diploma or equivalent were more likely to have never been married by age 55 (24 percent), than those with higher levels of education (10 percent for those with a bachelor’s degree or higher, for example).

Marriage rates differed by sex: 15 percent of men had never been married by age 55 compared with 10 percent of women. Comparable percentages of men and women had married one time, but women were more likely than men to have been married two or more times.

Marriage rates by race and ethnicity and sex also showed differences. Only 12 percent of White men had never been married by age 55, compared with 33 percent of Black men and 17 percent of Hispanic men. Similarly, 7 percent of White women had never been married, compared with 28 percent of Black women, and 12 percent of Hispanic women. White and Hispanic women were more likely to have one marriage only (60 to 61 percent) compared with Black women (53 percent). Black women were less likely to have had two or more marriages (18 percent) compared with White (30 percent) and Hispanic (27 percent) women.

Marriage rates varied by education, within sex and race categories, with higher education generally associated with a lower likelihood of having had not married and a higher likelihood of having had one marriage. Over 71 percent of White men with a bachelor’s degree or higher had married once, compared with 44 percent of White men with less than a high school diploma or equivalent. For White women, the difference was slightly smaller, with 69 percent with a bachelor’s degree or higher having had one marriage by age 55, compared with 51 percent of White women with less than a high school diploma.

On average, Black men with a bachelor’s degree or higher had 1.2 marriages by age 55, whereas Black men with less than a high school diploma only averaged 0.5 marriages by age 55. The difference in the average number of marriages among Black men by education was primarily due to differences in the probability of marrying. Almost 58 percent of Black men with less than a high school diploma had never been married by age 55, compared with 15 percent of Black men with a bachelor’s degree or higher. Conditional on marrying at least once, Black men with less than a high school diploma had 1.2 marriages, compared with 1.4 marriages for Black men with a bachelor’s degree or higher.²²

Black women with a bachelor’s degree or higher averaged 1.5 marriages by age 55, whereas Black women with less than a high school diploma averaged 0.7 marriages. Over half of Black women with less than a high school diploma had never been married, compared with about a quarter of Black women with a bachelor’s degree or higher. Conditional on marrying, Black women with less than a high school diploma had 1.4 marriages by age 55, compared with 1.3 for Black women with a bachelor’s degree or higher.

For Hispanic and White women, the probability of never having married was similar across educational levels. For example, 11 percent of Hispanic women and 8 percent of White women in both the bachelor’s degree or higher and less than a high school diploma categories had never been married by age 55.

Among both Hispanic men and White men, the likelihood of never marrying decreased with educational attainment, but the differences were less steep and the likelihoods much lower compared with Black men. Additionally, the percentages who never marry by educational group were similar for Hispanic men and White men, with 22 to 23 percent of Hispanic and White men with less than a high school diploma having never been married, compared with 11 to 12 percent of Hispanic and White men with a bachelor’s degree or higher.

Hispanic women with a bachelor’s degree or higher were less likely to have only one marriage (59 percent) and more likely to have two or more marriages (31 percent) than those with less than a high school diploma (72 and 17 percent, respectively). For Hispanic men the opposite was true: those with a bachelor’s degree or higher were more likely to have one marriage (74 percent) and less likely to have two or more marriages (14 percent) than were Hispanic men with less than a high school diploma or equivalent (60 and 17 percent, respectively).

Table 8 shows that the percentage of men who did not marry by age 55 decreased in the percentage of years working full time in each racial and ethnic group. That is, more work was associated with a higher rate of marriage. For example, at the extremes, only 5 percent of White men, 3 percent of Hispanic men, and 14 percent of Black men who worked at least 2,000 hours in 80 percent or more of the years were never married by age 55. In contrast, 36 percent of White men, 56 percent of Black men, and 51 percent of Hispanic men who rarely worked full time (fewer than 15 percent of years from ages 25 to 54) were never married by age 55.

For women overall, there was less variation by work history. However, White women who worked at least 80 percent of years full time were more likely to have zero marriages by age 55 (14 percent) than White women who worked fewer than 80 percent of years full time (4 to 9 percent). In contrast, Black women who worked full time in fewer than 15 percent of years were more likely to have zero marriages by age 55 (38 percent) compared with Black women who worked more than 15 percent of years full time (20 to 27 percent).

White, Black, and Hispanic men who worked full time at least 80 percent of years had similar numbers of marriages by age 55 (1.2 to 1.3). By comparison, Black women who worked full time at least 80 percent of years had a lower number of marriages by age 55 (0.9) than White (1.2) or Hispanic (1.5) women in the same work-history category.

Marital history by age

Charts 1, 2, and 3 display marital status through the lifecycle, for the full sample and separately for men and women. Chart 1 shows a steep decline in the percent of men and women who never married from age 23 (61 percent) to age 27 (39 percent). By age 35 only 20 percent of men and women had never married, and by age 55 only 13 percent of men and women had never married. Chart 1 shows that by the members’ mid-30s and beyond, divorce rates from a first marriage increased as did the percentage of the sample who entered into a second marriage.

A comparison of chart 2 (men) and chart 3 (women) shows that women married at younger ages than men, with a steeper increase in entry into marriage in the early to mid-20s. By age 27, 55 percent of women were in a first marriage and almost 15 percent were no longer in their first marriage. By comparison, at age 27, 45 percent of men were in a first marriage and almost 8 percent were no longer in their first marriage.

Charts 4, 5, 6, 7, 8, and 9 display estimates of the cumulative percentage of men and women who had begun a first marriage by age, beginning at 15.²³ Each population subgroup started at age 15 with virtually zero people having begun a first marriage; the cumulative percentage who had a first marriage increased as they aged, at different rates for different population subgroups. Chart 4 shows that White and Hispanic men had very similar likelihoods of entering a first marriage as they aged, with a small divergence after age 35 or so. But Black men were less likely to get married than White or Hispanic men at each age, with the gap widening as they got older. A similar pattern emerged for women; Black women had consistently lower rates of entry into a first marriage than White and Hispanic women as they aged (see chart 5).

Chart 6 shows that the men’s rates of entry into a first marriage varied by educational attainment. As throughout the article, educational attainment is measured at the time of the most recent survey. Before age 25, men who earned less than a high school diploma, who have a high school diploma, and who have some college had entered a first marriage at similar rates. However, after age 25 the curve for men with less than a high school diploma remained relatively flat compared with the curves for those with higher levels of education. Men who earned a bachelor’s degree or higher were slower to enter a first marriage but around age 28 overtook those with less than a high school diploma and at age 31 caught up to men with a high school diploma or some college. From about age 30 to 35, the estimated cumulative percentage who have ever married remains fairly unchanged for men with less than a high school diploma while the cumulative percentage of men with higher levels of education who enter into a first marriage continued to increase.

Chart 7 shows that women with less than a high school diploma were less likely to ever marry by age 55, though the cumulative percentage who first married at younger ages was higher compared with those with more education. After age 35, the curves that show the cumulative percentage that had ever married was stable for each level of education.

Charts 8 and 9 illustrate the relationship, stratified by age, between the propensity to enter into a first marriage and the employment history for men and women, respectively. Men who were more consistently employed full time and year round, married at higher rates. By their mid-20s, they were more likely to have ever married than their counterparts who worked less intensively over their lives. In contrast, among women, work rates made smaller differences in the propensity to marry, with the differences narrowing as the women age.

Charts 10 and 11 show estimates of the cumulative percentage of men and women who have divorced since their first marriage began. The charts show that in the first years after a first marriage, Black men and women divorced at higher rates than White and Hispanic men and women.

Chart 12 shows that men with a bachelor’s degree or higher by age 55 were much less likely to get divorced from a first marriage than men with lower levels of education. By year 15 of marriage, the slope became very shallow for men with a bachelor’s degree or higher, but those with lower levels of education continued to leave their first marriages. Thus, the gap in the probability of not having divorced between those with a bachelor’s degree or higher and those with lower levels of education increased with time. Though the difference was smaller, women who earned a bachelor’s degree or higher by age 55 were less likely to divorce beginning 3 years after a first marriage, with the difference in the probability of exiting a first marriage increasing over time. (See chart 13.)

The relationship between work history and divorce was different for men versus women. Among men, the likelihood of divorce from a first marriage was lower for men who work in a greater percentage of years. (See chart 14.) Compared with men, women’s propensities to divorce varied less by employment history, with women who work more having a higher propensity to leave their first marriages. (See chart 15.)

Conclusion and discussion

This article illustrates marriage and divorce patterns from ages 15 to 55 for a cohort of Americans born between 1957 and 1964. This work is descriptive and does not explain the underlying reasons for differences across groups.

The data show that, on average, men and women in this cohort had 1.2 marriages by age 55. By age 55, most men and women had ever been married (87 percent), and conditional on having married, 46 percent had ever divorced. The article shows an array of differences in marital patterns by sex, race, ethnicity, educational attainment, and work history. For example, Black men and women were less likely to marry than White and Hispanic men and women. In addition, Black men and women married later (age 27) than White (age 25) and Hispanic (age 24) men and women. Differences by education were stark. Men with less than a high school diploma were less likely to have ever married (71 percent) than those with higher levels of education (84 to 89 percent). Conditional on marriage, divorce rates were much lower for those with a bachelor’s degree or higher (27 percent) compared with those with less than a bachelor’s degree (50 to 53 percent). For men and women, marital history was tied to employment history. Among men, more full-time work was associated with a higher rate of marriage.

Recent work that examines the economic impacts of divorce by the intersection of race, ethnicity, and sex found important differences by race and ethnicity for men and for women.²⁴ That may be the case with respect to the relationships we describe here, too. We stop short of fully examining the patterns between education and work history and marital outcomes in part because of the relatively small sample sizes of the groups for some of the combinations of race, ethnicity, and sex. For example, small numbers of Hispanic women and Black women worked at least 2,000 hours in at least 80 percent of years, and small numbers of Hispanic men and White men worked at least 2,000 hours in less than 15 percent of years from ages 25 to 54.

Differences between the marriage rates of groups might be explained by couples choosing to live together, without marrying. This article does not examine patterns of cohabitation. The NLSY79 did not begin collecting detailed information on cohabitation until 2002, when the sample members were ages 37 to 45.²⁵

The patterns illustrated in this article imply that economic well-being and retirement decisions of the NLSY79 cohort may depend on marriage and divorce outcomes over their lives.