Mathematical Malpractice Watch: IUDs and Teens

Right now, the liberal blogosphere is erupting over Republican plans to not fund a program to give free IUDs to low income women:

Republican legislators in Colorado will not authorize funding for a program that gives free IUDs to low-income women — an effort that many believe was responsible for hugely driving down teen births.

Colorado has recently experienced a stunning decline in its teen birth rate. Between 2007 and 2012, federal data shows that births declined 40 percent — faster than any other state in the country.

State officials attributed part of this success to the Colorado Family Planning Initiative, which provided free IUDs to low-income women seen at 68 family planning clinics across the state. Last year, state officials estimated that young women served by those family planning clinics accounted for about three-fourths of the overall decline in Colorado’s teen birth rate.

I disagree with the Republicans on this. But the idea that free IUD program cut Colorado’s teen birth rate by 40% or 3/4 of 40% or anywhere close to 40% is high-test nonsense.

Here is the data from the CDC on teen brith rates. From the first graph, you’ll see that teen birth rates have been steadily falling for seventy years. Like most positive social trends, it has many, um, parents, each of which are flogged by whomever supports that particular issue. Availability of contraception has certainly played a role. The legalization of abortion played a role (although abortion rates peaked in the early 80’s). As social and professional barrier have fallen, many more women are delaying pregnancy for college and jobs. And there is some evidence that teenagers are waiting longer to have sex (that would be the dreaded “abstinence”).

Since 2007, however, the teen birth rate has fallen off a cliff. But not just in Colorado. It’s fallen everywhere, by an average of 30%. If anything, it’s fallen faster in red states than in blue ones (see Figure 9 of the CDC’s report). Colorado has seen the steepest decline (39%), but just behind it are the red states of Arizona (37%), Georgia (37%), North Carolina (34%), Utah (34%) and Virginia (33%).

Is Colorado’s IUD program so awesome that it dropped the teen birth rate for the entire country?

Given the extent of the program and Colorado having the largest reduction, it’s very probable that the IUD program did play a role here. But I would ballpark it at maybe 10% at the most.1 That’s not nothing and it’s probably worth continuing the program. But let’s not pretend the reduction is due only to that.

So what is causing the large reduction? Availability of contraception is playing a role, yes, but there’s something else going on. Birth rates have fallen for all women since 2007, not just teenagers. I don’t think it’s coincidence (and neither does the CDC) that the teen birth rate plunged when we hit the worst recession since the Great Depression. If you look at historical birth rates, you’ll see a similar plunge in during the 1930’s. And that was long before almost the entirety of modern birth control, least of all free birth control.

I think that’s the story here. Colorado’s program was fortuitously timed in that regard and there is likely some synergy between the economic downturn and the IUD program (i.e., the program kicked in right when a bunch of women were more eager for birth control).

One of the difficult things about Mathematical Malpractice Watch is that I frequently end up attacking people I fundamentally agree with. I think Colorado should extend their IUD program (although I’m old enough to remember, in the 90’s, when Republican governors offering incentives for low-income women to use Norplant was denounced as eugenics). But the claim that it has produced a “huge” reduction in the teen birth rate is just not true.

Actually, there is a chance that the effect is 0%. Colorado had the sharpest reduction in teen pregnancy rates. It’s easy to go in, post facto, and identify a pet policy to pin it on while ignoring the thousand other factors occurring in fifty states. It’s called the Texas Sharpshooter Fallacy. Colorado might just be a statistical outlier and we’re crediting a policy for that outlierness because we like the policy. Colorado’s barely two standard deviations from the mean. I think it’s likely the IUD fund had an effect, but I’d be pressed to prove it statistically.

Tags: , , ,

Comments are closed.