The father of my friend’s five-year-old daughter proposed to her, and she said yes. They haven’t set a wedding date, and she told me that she hopes he sets one before she changes her mind.

Even though I’ve told you very little about this couple, I doubt you’d rate the chance of success for their marriage very highly. I was concerned, too, when I learned almost simultaneously about his proposal and her ambivalence, and so I asked, “Why did you say yes?” Her reply: “We’ve been together off-and-on for almost a decade, and Kylie [yes, it is a pseudonym] would have both of her parents.” There are far worse reasons to get married, but even those of us who have outgrown fairy tales would still hope for a little more: some indication of a spark between them, something about liking or loving (ideally both), something about readiness for commitment.

New research on the marital prospects of single mothers published in the August issue of the journal Demography has good and bad news for my friend. The good news is that biological father marriages are more enduring than stepfather marriages. In other words, saying yes to her daughter’s father means that she is more likely to stay married than if she had said yes to someone else. The bad news is that even though she opted for the better bet, the odds are still stacked against her marriage lasting.

I stared at the numbers in Christina Gibson-Davis’s article, and I didn’t like them. “Among marriages that involved a black mother and a biological father,” she found, “20% had dissolved within five years, and 47% had dissolved within 10 years.” That gives five-year-old Kylie about a flip-of-the-coin chance of having still-married parents by her fifteenth birthday (if her father does set a date before her mother changes her mind). I reminded myself that her mother is 41 years old and has a college degree. As any good demographer knows, both older age and higher educational attainment are associated with marital stability, so I checked another table in the article to adjust Kylie’s parents’ odds. Unfortunately, I found that maternal age and education are insignificant predictors of divorce risk for black women marrying after childbirth. So I didn’t get to adjust their odds. They stayed bad.

Very few personal characteristics affect the divorce risk of women who have had a child before marrying.

Then I noticed that very few personal characteristics affected the divorce risk of women of any race in the study. Their husbands’ characteristics didn’t matter, either. Even though factors like age and education have been shown to matter quite a bit for ten-year divorce rates overall, they didn’t make a difference among the women that Gibson-Davis studied: those who gave birth outside of marriage.

I keep thinking about what that means. It means that the normal advantages that my friend should enjoy get swamped by her single mother status. It means that her chances of a successful marriage aren’t statistically different from those of a kid who drops out of high school, has a baby, and then marries the father as soon as she turns 18.

All my demography training has taught me that aggregate statistics do not predict individual outcomes—that I shouldn’t assume that I know whether Kylie’s parents will stay together based on Gibson-Davis’s tables. But knowing Kylie’s mom gives me some insight into why the numbers in the table are so depressing. She hasn’t married Kylie’s dad yet. I suspect that one major reason is that fidelity is not something that Kylie’s dad expects of himself. Couples who did not marry before their child was born have opted out for some reason. Some of those reasons can change over time, as when someone gets a job or both partners become mature enough to commit. But some of those reasons don’t change over time; some of the factors that cause people to forgo marriage are the same factors that cause people to divorce. I would love to see Kylie grow up in a home with both parents, but I would hate seeing my friend endure adultery to make that possible.

By showing that a single mother is most likely to marry her child’s biological father shortly after giving birth, and that such a marriage is less likely to end in divorce than marriage to another man, Gibson-Davis has provided some limited, encouraging evidence to those championing marriage promotion programs. Nevertheless, she correctly stresses that much work remains for those seeking to encourage single mothers to marry because enduring unions are relatively rare, even among biological parents. I have one window showing me why that is true. There are many others. The bottom line is that post-birth marriages are likely to be more difficult and less stable than others, and strengthening them is indeed an uphill battle.