How many kids will college women likely end up having?

Impacts of religiosity, sociosexuality, and race.

As the first Millennials reach their early 40s, we can start to get a look at women’s near-completed fertility rates in the National Longitudinal Survey of Youth 1997, which has been tracking a large group of early Millennials as they moved from their teens through their 30s.

In this post, I’m focusing on prototypical college-educated women who go from high school to college and get a 4-year degree more-or-less on schedule without first having any children. While this probably describes practically 100% of the people that I and my readers usually hang out with, it’s a pattern shared by fewer than 30% of Millennials.

The question I’m asking is: Limiting oneself to info available about these women when they were 21, what predicts how many kids they ended up having by age 39? (Age 39 is about as far as you can reasonably determine with the NLSY97 data released to date.)

The items I have available include: whether they lived with their biological mother and/or father in their early teens; parents’ education, income, and frequency of religious service attendance (all measured in 1997); standardized test scores and grades in high school; and a bunch of items that have been measured over the course of the longitudinal study, including education, income, sexual activity, relationships, drinking and drug usage, BMI, health problems, and religious service attendance.

Looking through this predictor set at what we know at age 21 — and keeping in mind these are all women who had not had any children yet at that time — the most impactful items in predicting how many kids they had by age 39 were:

• Race: Black women had fewer children (which, as we’ll see, seems primarily related to their far lower marriage rates).

• Religiosity: Women who were going to church more frequently in their mid-teens had more children than those attending less frequently.

• Sexual patterns: Women who didn’t have sex at all with men around age 21 had fewer children, as did women who had sex with lots of men around age 21. In a kind of Goldilocks pattern, those who were in between — women who had sex with men but not lots of men around age 21 — had higher average near-completed fertility.

Let’s take a look at the details of the patterns.

I defined the prototypical college group as those who had no children at age 21 and who got a 4-year degree by their mid-20s. I then started splitting the sample based on those features known at age 21 that were most highly predictive of the number of children they had by age 39. The first split was by race into black vs. non-black — black women in the prototypical college group ended up with substantially fewer children than non-black women.

For the non-black women, the next-most-impactful split was between those whose frequency of religious service attendance at age 15 was in the top 40% vs. the bottom 60%. This is essentially a split between those who went to services at least weekly and those who did not.

The next splits were based on a variable I created to track sociosexuality. It’s a composite measure that is highest for those who combine higher numbers of recent sex partners and having sex with strangers with higher rates of drinking (especially binge drinking) and smoking pot. So, you know, the higher levels are folks who at a given age are partying a lot and having sex with a lot of partners, including ones they just met. The lower ends are folks who rarely drink or use recreational drugs and either have been sexually abstinent or have had only one partner recently.

The chart below shows average number of children for the various sub-groups of prototypical college women (along with, for reference, the average number for women not in the prototypical college group, i.e., the other 70%+ of the cohort). For the college group, the highest near-completed fertility came from non-black women who were in church at least weekly in their mid-teens and who, at age 21, were not partying and hooking up very much — these women averaged a bit over 2 children by age 39.

The lowest fertility came from three groups: (1) non-black women who didn’t go to church weekly at age 15 and who were not having sex with men at age 21, (2) black women, and (3) non-black women who didn’t go to church weekly at age 15 and who were partying and hooking up quite a bit at age 21. These groups ended up with between 1 and 1.25 kids on average.

And then two groups landed in the middle at around 1.6 kids on average: (1) non-black women who combined frequent teenage religious service attendance with higher levels of partying and hooking up in college, and (2) non-black women who combined infrequent teenage religious service attendance with lower levels of partying and hooking up in college.

In short: Among college women, the highest fertility came from religious women who didn’t hook up much, and the lowest came from black women along with less-religious women who either tended to hook up with lots of men or tended to avoid sexual activity with men altogether in college.

Marriage rates for college women

For college women, fertility is very closely associated with marriage. The next chart shows the percentage of women who were married at different ages. The highest marriage rates come from the same group that had the highest fertility — non-black women who were religious in their teens and sociosexually reserved at 21. Half were married by age 24.

The lowest marriage rates come from black women. This is likely due primarily to the substantial sex ratio imbalance caused by the markedly higher rates of incarceration and death among young adult black men, which not only means there aren’t enough men to go around, but also destabilizes mating markets by decreasing available men’s incentives to commit to only one woman.

For these college women, marriages were rising most quickly in their mid-to-late 20s, while births followed by a few years, rising most quickly in their early 30s.

In general, the three low-fertility groups are also the slowest and least likely to marry, though it looks to be due to three independent sets of reasons. For black women, the key issue likely involves the shortage of available men. For women who party and hook up a lot, the key issue likely involves a lack of interest in monogamy along with a preference for the kinds of men who aren’t likely to remain monogamous. For women who aren’t having sex at all with men, it really depends on why.

Most of the sexually-abstinent-at-21 women are actually up in the group of women who were religious teens and then didn’t party or hook up much in college — the group of college women with the highest near-completed fertility. For them, not sleeping with men before commitment is a tactical move, motivated by a desire to achieve the kind of stable marriages that support having more children. For non-religious teens who don’t have sex with men in college, it appears to be more a matter of the women just not being too interested in men or marrying or having children in the first place. For example, LGBTQ sexual orientations are closely linked with both reduced religiosity and reduced likelihood of having or wanting children.

Three dimensions

If you want to better understand what we’re seeing here, let’s take a step back. In general, there are three key dimensions when it comes to individual differences in relationships and childbearing (at least, this is true of the contemporary U.S.). One dimension is socioeconomic status (SES), where some people tend to get longer educations and wait longer to form stable relationships and have children, which typically means having fewer children in the end but much higher incomes and more stable marriages than those folks who truncate their educations and start having children earlier.

Another dimension is sociosexuality, where, at one end, there are people who party and hook up a lot, are less likely to marry and more likely to divorce, and end up with fewer children (i.e., what I’ve called Freewheelers), and at the other end, there are people who don’t party or hook up much, have stable long-term relationships, and end up with more children (i.e., what I’ve called Ring-Bearers). This dimension is a central correlate of religiosity in developed countries.

Then there’s a third dimension, where, at one end, there are people who are enthusiastically in the heterosexual mating game (let’s call them Participators), who are sexually active most of their adult lives, spend more time married or at least cohabiting, and have more children. At the other end, there are people who just really aren’t in the heterosexual mating game at all (let’s call them Sideliners), who are sexually abstinent for long stretches, are unlikely to marry or cohabit, don’t have many kids — none of it.

Of course, each of the three dimensions is a continuum and not a simple contrast. Further, not everyone sticks with the same pattern throughout their lives. A common flip — and this should be pretty intuitive — involves a sizable chunk of folks who have Freewheeler patterns as young people and subsequently shift to Ring-Bearer patterns as they settle down into committed relationships and children. I’ve called this a Mixer pattern, with this kind of one-time adult flip especially prevalent among higher-SES folks. Another Mixer pattern involves being married or cohabiting for most of one’s adult life but with a higher number of partners, with each pretty quickly following the last. This kind of serial monogamy is more often seen among lower-middle-SES whites.

Another example involves how a prototypical high-SES Ring-Bearer pattern plays out. It actually begins with a very strong Sideliner pattern, with these folks pretty much abstaining entirely from sex as they pursue college educations. But then they fully switch to being Participators by getting and staying married and having a bunch of kids.

The tremendous diversity of modern fertility and relationship patterns is found in how the three dimensions of reproductive variance — High vs. Low SES, Freewheeler vs. Ring-Bearer, and Participator vs. Sideliner — interact with each other. You can see this with my series on different clusters of early Millennial women, which included High-SES Mixers, High-SES Freewheelers, Higher Ring-Bearers, Mid-SES Mixers, Mid-SES Freewheelers, Sideliners, Lower Ring-Bearers, Low-SES Mixers, and Low-SES Single Moms – each of these groups has a distinct prototypical set of outcomes based on the different components, timing, and sequencing of relationships and fertility.

A notable point made clear by these varying life patterns is that there are multiple pathways to low fertility and multiple pathways to high fertility. In this post, I’ve been focusing on prototypical college-educated women, who overwhelmingly cut off one pathway to high fertility — namely, starting to have kids early in life without necessarily waiting until marriage.

The way I think about it is: College women typically behave as though they would rather have no children than poor children, while non-college women more often behave as though they would rather have poor children than no children. For college women, then, having children is something they typically do only after getting married to men who have some combination of higher likelihood of marital commitment and higher incomes. Both features reduce the risk that they’ll have children who end up being poor.

So, in the end, there are two categories of college women in the NLSY97 sample who ended up having children on average: Higher Ring-Bearers and High-SES Mixers. The former typically started out highly religious, didn’t party or hook up much in college, and then entered stable marriages, having two or more kids and upper-middle-class incomes. The latter typically were not religious by the time they reached college, partied hard in college, but dialed it back a bit in their mid-to-late-20s, and then got married just in time to have one or two children in their 30s while enjoying very high incomes. The other two groups of college women – Sideliners and High-SES Freewheelers – typically had no children.

What we’ve seen in this post is that by age 21 we can already get a decent sense of how many children they will have, even though most didn’t begin to have them until their mid-late 20s or early 30s. Early life patterns set the stage for what happens downstream, making some outcomes substantially more likely than others.