Assessing Your Divorce Risk

What are the odds of your marriage ending in divorce?

This is a risk with some pretty important consequences, but chances are, you don’t have the foggiest idea on how to quantify it. Until now.

My favorite economist (and my significant other), Betsey Stevenson, has put together a neat online widget for the folks at The widget crunches recent marital history data to assess your chances of getting divorced, and it does this by taking account of some very simple demographics. Learn your divorce risk here.

It is hard to overstate how big of a leap forward this calculator is. Those of us who are optimists may not want the facts — we just assume that the risk is zero. And those who want to be realists have probably learned that without usable data, being realistic isn’t easy. Perhaps the realists have a vague memory of reading that one in two marriages end in divorce. But this just isn’t true, even if it was true for my (divorced) parents’ generation.

And the national average just isn’t that informative anyway, as there is enormous variation in divorce risk across demographic groups. My graduate-school friends — say, male college grads who married in their early 30’s and have been married for five years — are in a demographic where only 5 percent have divorced so far, and perhaps another 7 percent may divorce in the next five years. By contrast, among male high-school grads who married in their early 20’s, around 19 percent divorced in the first five years of marriage, with another 37 percent likely to separate over the next five years.

The point is that factors like age at first marriage and education tell us a lot about divorce risk. Let’s not confuse correlation and causation though — these divorce risks are useful as statistical forecasts (even if they can’t answer the “what if” question) of how divorce risks change if you delay your marriage.

I’m interested in learning whether your divorce risk was higher or lower than you expected. What did you find?

More detail here; F.A.Q.’s are here. The academic research underpinning this widget is available here (and the aggregate trends are discussed here and here).

Shaun Dakin

I ran through this a week ago when I discovered this website and was surprised that my marriage risk was much lower than expected (50%) and was around 4%.

The problem is, the only number that really maters is 1, you and your marriage.


After 26 years of comfortable monogamy, my husband and I are tiptoeing towards an open marriage. With trepidation, I checked the online assessment, but, predictably the US Census doesn't ask how many of us try open marriages. So far, I can cite far more benefits than problems. Trust liberates.


According to that widget, my chance of divorce in the next five years is 18%, while my husband's chance of divorce in the next five years is 6%. Somehow, I think this is problematic - the unit of analysis should be the marriage, not the person.


"People with similar backgrounds who will be divorced over the next five years: N/A "


Professor Wolfers - How did your own results compare with your expectations?

I am single, so no comment from me...


Seems to me race, socio-economic class, and number of previous marriages are also relevant. If the number of marriages ending in divorce approaches 50% then over 7% risk can't be high risk because 50% should be average risk. Some couples have to defy odds by staying together with an 85% risk of divorce or something. I also agree with the above comment

Jon Bohlinger

I believe John Gottman has done a significant amount of research into this and come to some amazing conclusions. He has come up with some real, quantifiable things that can be done to decrease the odds of divorce considerably. Real, scholarly research that is really well funded.


Chances are, if you're going to a "will I get a divorce" calculator, you're going to get a divorce.



To understand the depths of human existence; I think more than 5 questions would be needed.

I flipped a coin and it gave me heads!


eh. 1% already divorced; 8% chance in next five years. it must not be taking something into account or I wouldn't have contemplated divorce almost daily for the past year or more. being well educated and marrying in your 30s do not make up for making the wrong choice.


Much, much lower than I would have expected (around 5%). jz, don't you mean eliminating trust from the equation liberates? After all, if the marriage is open, what are you trusting your husband to do or not do?


I can confidently state that my risk of divorce is 0%. If I have any say in the matter, it will stay that way. Where's the shotgun wedding calculator?

Tomislav Najdovski

It hardly can be done trough pure demographic segmentation.

Maybe some factors that contribute such as religious and political views, education etc can be measured, but it still involves a lot of personal characteristics which hardly or not at all can be generalized.


My wife's risk was 14% mine was 17%. I guess you could say OUR probability is (.14*.17)=.2% which according to the guide we are at average risk for divorce. We have been married for 4 years have a 1 year old girl, I am in Grad School, parents divorced. She had some college and parents are going on 50 years of marriage.

Joe D

It rates "my" risk as 8% and my wife's as 9% (interestingly, it asks women if they have children, but not men). That seems high to me.

Why don't they ask if your parents ever divorced? That would seem to be an obvious risk factor to me. In our case, both of us have parents who have been married forty-two years, and we've been married seventeen (we got married between our parents' respective twenty-fifth anniversaries). That example shows us that it's worth the work to get through any rough patches.

Andrew Helms

People with similar backgrounds who will be divorced over the next five years: 7%

which seems reasonable, I would have guessed between 5% and 10%


I got married this summer. I'm 28, and have a grad degree. It gave me a 4% chance of divorce, I think. Seems low, but then, you really go into marriage expecting less, so maybe from that perspective, 4% is high.


Getting the stats right, but missing the story. Sigh.


It's all about "business."

Family court is the busiest court in the nation, and yet, the one court not covered by reporters. It's also the court used most often for the longest period of time.



I wonder... do they consider the fact that you are someone who is Assessing Your Divorce Risk when they assess your divorce risk?


I'm surprised they didn't include questions about geography or religion.

Then again, such questions would probably make the widget more controversial. And indeed, it would be interesting to see if the differences seen across states and religions were adequately explained by, say, age at marriage. (i.e., because people whose religious beliefs proscribe pre-marital sex are more likely to marry young.)