Demystifying Divorce Statistics

Statistics say that half of all marriages will fail. Is it really a 50-50 chance that if you choose to marry, you will also choose to divorce? Possibly not. Your personal risk of divorce might be lower than you realize. It is important to look beyond the numbers and at the data, factors, and calculations statisticians used to determine this high risk of divorce in America.

The widely publicized 50% U.S. divorce rate and high international divorce rate are not a calculation of the number of people who married and subsequently divorced. The percentage arose from a mathematical formula, calculated in way that is similar to how life expectancy is predicted. This projection is based upon the number of marriages and divorces reported in a particular year. It is not based on the same people or the same marriages. It does not distinguish between first marriages and second marriages, the duration of the marriage, age, or any other important characteristics.

According to the National Marriage Project of the University of Virginia, there are several flaws in the 50% divorce statistic. First, this percentage includes an estimation of the number of couples who will separate. As a result, it may ignore the possibility of reconciliation. Second, researchers are not consistently using the same method of calculation over time. The National Marriage Project also notes that this type of formula is most helpful in analyzing past rates of divorce. Its usefulness in predicting future divorce is limited because the calculation must assume that divorce and death rates will remain exactly the same for the entire duration of every marriage entered during the calculated year.

In fact, there are some researchers who say that the 50% rate of divorce is just plain wrong. They suggest that divorce has never as prevalent as to reach 1 in every 2 marriages. Two primary problems have been cited as the cause of misinformation: first, inconsistencies in data collection; and second, the type of data needed to accurately predict divorce is unavailable.

Other researchers claim a 50% divorce rate is limited to a very specific subset of the population. The likelihood of divorce is based on many complex factors. Certain risk factors that increase the chance of divorce have been identified, and the non-existence of these risk factors has been found to significantly decrease the chance of divorce. As more accurate data emerges, researchers are starting to see that there are really different divorce probabilities for different subsets of the population. To understand divorce rates, it is necessary to look at the personal, social, and cultural characteristics of spouses. Click here for a graphical view of divorce based on specific factors.


In “State of Our Unions: Marriage in America, 2010” published by the National Marriage Project, the following factors were identified as lowering the chance of divorce.


Percent Decrease in Risk of Divorce

Making over $50,000 annually (vs. under $25,000)




Having graduated college (vs. not completed high school)




Having a baby seven months or more after marriage (vs. before marriage)




Marrying over 25 years of age (vs. under 18)




Coming from an intact family of origin (vs. divorced parents)




Religious affiliation (vs. none)


The belief that the 50% divorce rate is an accurate prediction of whether a specific couple will divorce may have a detrimental effect on marriage. Joshua Goldstein, professor of sociology and public affairs at Princeton’s Office of Population Research, believes “expectations of high divorce are in some ways self-fulfilling.” If that is true, then hopefully, publication of the realities and shortcomings of divorce statistics can help create a self-fulfilling prophecy towards lifelong marriages.

This entry was posted in News and tagged , , . Both comments and trackbacks are currently closed.

One Comment

  1. More Barn
    Posted September 16, 2011 at 6:56 am | Permalink

    Seriously, are you joking? This entire article is spent debunking the 50 percent estimate and saying why, but offering no evidence of a more accurate estimate. I’m not demanding a smoking gun number, but certainly some learned estimates can be considered in opposition to the supposedly wack 50 percent estimate. You’re telling us not to believe the 50 percent studies, that they’re slanted and stacked one way, but you offer no educated estimate to support your claim that the other is flawed. Nice try. This should have been an opinion piece for the editorial page.

One Trackback