Divorce Rate in the United States
By Matthew Theisen, Ph.D.
Updated April 19, 2018
Divorce rate in the United States
Using data from the U.S. Census, we can estimate the divorce rate based on some simple demographic characteristics. Our analysis uses data from 2015 on people who are married or recently divorced to estimate the present probability of divorce. Thus, our data is a 'snapshot' of marriage in 2015, rather than a long-term, longitudinal analysis of trends over time.
Calculating divorce rate
Estimated divorce rates vary from 1% to about 13% over five years depending on your characteristics. Education, the presence of children and length of marriage strongly influence the likelihood of divorce. Interestingly, age at marriage did not strongly impact divorce rate when these other factors are taken into account. We consider only data from first marriages.
Divorce rate facts
Using the data, we can also look at the marital status of the population by age in 2015. Here we split the population into never married, 1st marriage ended in divorce, or in 1st marriage/widowed. Here we see some distinct phases of life. Fraction married increases sharply from age 20-30. In the 30s, 40s, and 50s, the fraction of people whose first marriage ended in divorce increases. In this analysis, we do not consider remarriage. In later ages, we see the married fraction increase again. This could be because of generational effects (people of older generations got and stayed married), or lifespan effects (married people lived longer than divorced), or some combination.
We can also look at the rate of first marriages and first divorces by age, rather than the overall fraction of population. In this analysis, we look at the fraction of population at each age that got married or divorced for the first time within the past 12 months. Here we see the rate of first marriage peaks in the mid 20s, while the rate of divorce peaks and overtakes first marriage rate at age 40.
Matthew Theisen is a software engineer and data scientist at Leaf Group. He is the lead researcher for careertrend.com. He holds a PhD from UCLA in Bioengineering, specializing in numerical modeling.