Landmark study reveals how predictors of divorce have transformed since the 1950s

In a study recently published in the Journal of Marriage and Family, researchers embarked on an ambitious project to assess how the predictors of marital dissolution, commonly known as divorce, have evolved from the 1950s through the 2010s in the United States. The research sheds light on whether certain factors traditionally associated with higher divorce rates have maintained their predictive power or if new patterns have emerged in the face of growing economic and social inequalities.

Previous research on marital dissolution has often focused on singular predictors such as education levels, premarital cohabitation, and the impact of parental divorce. The last comprehensive review before this study, conducted by Teachman in 2002, noted a general stability in these predictors up to the mid-1990s, with significant changes primarily in the convergence of marital dissolution rates between Black and White women.

Since then, America has witnessed a period marked by increasing economic disparities and changes in social norms, prompting a need for an updated examination.

“I study marriage and divorce. I like to understand the topic as broadly as I can,” said study author Michael J. Rosenfeld, a professor of sociology at Stanford University and author of The Rainbow After the Storm.

“Most of the academic literature on the predictors divorce only examines one predictor of divorce. Academic literature can sometimes be so narrow that the big picture is missing. Also, it is a lot easier to publish papers that document changes than it is to publish papers that show that something, such as a predictor of divorce, has been stable over time. So I found the previous literature to be insufficiently broad.”

Free daily newsletter

Rosenfeld and his co-author, Katharina Roesler, utilized data from ten cycles of the National Survey of Family Growth (NSFG) spanning from 1973 to 2017, focusing on first marriages of women aged 15–44 (expanded to 15–49 in the later waves). This data set, encompassing 47,390 women and resulting in 14,236 marital dissolutions, provided a rich basis for analysis. By employing Cox proportional hazards regression, the study could account for both static and evolving factors over time, such as race, education, premarital cohabitation, and family origin’s intactness.

One of the study’s most significant revelations pertains to the relationship between race and marital dissolution. Initially, the researchers observed a convergence in divorce rates between Black and non-Black women from the 1950s through the 1990s, aligning with the optimistic views of the “Converging Destinies” hypothesis. This trend was seen as a positive outcome of the Civil Rights movement, suggesting that societal progress towards equality had a stabilizing effect on marriages.

However, post-2000, this trend reversed, with a noticeable divergence in marital dissolution rates between Black and non-Black women. This latter finding resonates with the “Diverging Destinies” hypothesis, suggesting that the recent decades’ increasing social and economic inequalities have begun to exert a divisive effect on marital stability, particularly disadvantaging Black women.

“There is a fundamental question about whether increasing inequality in the United States has lead to more inequality in the predictors of divorce, and I find that yes it has on the dimensions of race, education, and age at marriage,” Rosenfeld told PsyPost. “The difference in divorce rate between Black and White women narrowed after the Civil Rights revolution but has widened again in the 2000s. The greater divorce risk of marrying as a teenager has increased over time.”

Furthermore, Rosenfeld and Roesler closely examined the impact of education on marital dissolution. The findings indicate a widening gap in divorce rates between women with and without Bachelor’s degrees, affirming the Diverging Destinies hypothesis in this regard. This suggests that higher education, often associated with better economic prospects and social standing, continues to be a protective factor against marital dissolution, while those without such degrees face increasing risks.

Moreover, the findings highlighted the evolving impact of marrying at a young age and coming from a non-intact family. Both factors showed increasing association with higher divorce rates over time, particularly for those who married as teenagers. This trend underscores the continuing and perhaps growing challenges faced by individuals from these backgrounds in maintaining marital stability.

“Having a college degree and marrying later in life (mid-twenties at least) are two historically consistent protectors against later divorce that a person can control,” Rosenfeld said. “It is easier to have a stable marriage if you get a college degree and don’t marry too young. The protective aspects (against divorce) of a college degree and later marriage have increased over time.”

Interestingly, the study found that the impact of premarital cohabitation on marital dissolution did not follow the expected patterns of either the Diverging or Converging Destinies hypotheses. Instead, the association between premarital cohabitation and marital stability remained relatively stable over time, challenging assumptions that changes in societal norms around cohabitation would significantly impact marital dissolution rates.

This study provides invaluable insights into how the landscape of marital dissolution in the United States has shifted amid broader societal transformations. But, like all research, the study includes some limitations, including the lack of data on income over time and the potential biases introduced by selective marriage patterns among different demographic groups.

“Most people don’t understand that the divorce rate in the United States is actually quite low: less than 2% of married couples get divorced each year,” Rosenfeld explained. “The reason we hear about the 50% lifetime risk of divorce is that we live so long. One down side of the low divorce rate for researchers is that we always need more data to test our theories. This paper used the National Survey of Family Growth, a series of surveys going back to the 1970s, but even this substantial sample size was not enough to answer every question we had.”

The study, “Stability and change in predictors of marital dissolution in the US 1950–2017,” was published August 14, 2023.