Teen pregnancy in Ohio has seen a dramatic decline since 2006, as well as a drop in the abortion rate. Ridiculously, Ohio Medical Board member (and anti-abortion activist) Mike Gonidakis is crediting the War on Women for the decline in abortion even though the abortion rate fell faster under Strickland than under Kasich’s first year… which is the most recent data we have.

image 1

Gonidakis was probably thinking of the drop in teen pregnancy, though we yet don’t have any Kasich-era statistics for that. What we see is huge decreases in 2000 and 2001, then a plateau followed by huge decreases in 2009 and 2010.

What explains this?

I think we can rule out behavioral factors. We see year-over-year declines that are very similar across age groups, but we don’t see those declines within cohorts. Remember, Ohio’s 18-19 year-olds for 2010 were Ohio’s 16-17 year olds for 2008.

18-19 15-17 15-19 10-14
2010 9.4% 11% 9.8% 8.3%
2009 4.6% 15.8% 7.1% 7.7%
2008 2% 2.3% 1.2% 7.1%
2007 -2% 0.3% -1.2% 12.5%
2006 -3.8% -0.3% -1.4% -6.7%
2005 -0.7% -0.3% 0% 0%
2004 0% 2.9% 0.9% 0%
2003 1.4% 0% 0.8% 6.3%
2002 6.2% 4.7% 6% 5.9%


The same colors are the same kids1. 2005’s 10-14 cohort was no less likely than the year prior to become pregnant, but when they were 5 years older they were 10% less likely to get pregnant.

If a behavioral intervention (namely, sex education at a particular age) were responsible for the decline, then a given cohort would be equally less likely to get pregnant early as they are later.

We don’t see that:

image 2

This scatter chart compares the relative drop in pregnancy rate versus the prior cohort. If behavioral factors were strong determinants of teen pregnancy, there would be a clearer pattern when comparing cohorts.

Since teen pregnancy drops fairly consistently across all age groups simultaneously, environmental factors (rather than behavioral factors) will be correlated with the drop. The two main possibilities here are 1) expansion of birth control access and 2) delay in sexual activity.

Here’s a graph of growth in CHIP enrollment against decline in teen pregnancy:

image 3

Growth in CHIP enrollment since 2007 is predictive of the relative decline in pregnancy among 15-17 year olds, and slightly less so among 18-19 year olds (who aren’t eligible for CHIP).

Prior to 2007, we don’t get that correlation:

image 4

Starting in 2007, CHIP covers birth control at no cost, including Long Acting Reversible Contraceptives (LARCs) like IUDs and Nexplanon. Here’s a 2007-vintage Heritage article bemoaning the prospect of teens getting their hands on contraception like they do in Massachusetts and California2.

image 5

What we see is that states that maintained contraceptive coverage for minors had an earlier and more constant drop than the national as a whole, which saw a 2003-2007 increase followed by a more recent drop.

Bringing in CDC data about self-reported sexual behavior, we can compare birth control usage to teen pregnancy in both 2005 and 20103:

image 6

image 7

In both years, states where more teens used birth control had lower teen pregnancy rates; states with fewer teens using birth control had higher teen pregnancy rates. After the addition of LARC coverage to CHIP, the correlation was slightly stronger.

But what if the Crisis Pregnancy Centers are right, and delaying sexual activity deserves credit for the drop in teen pregnancy? Well, since teen sexual activity only dropped 0.2% between 2005 and 2011, that pretty much rules it out.

It also, FWIW, basically discredits Abstinence Only in its entirely.

2011 2009 2007 2005 2003 2001 1999 1997 1995
33.7% 34.2% 35% 33.9% 34.3% 33.4% 36.3% 34.8% 37.9%

Here’s a graph of sexual activity vs. teen pregnancy.

image 8

image 9

Since the expansion of CHIP, teen sexual activity is less predictive and has a weaker effect on teen pregnancy.

This is very strong evidence that increased access to birth control is responsible for the decrease in Ohio’s teen pregnancy rate. The same logic predicts that expanded access to birth control (through no co-pay coverage and Medicaid expansion) will reduce the unintended pregnancy rate for 20-29 year-olds, and greatly reduce the abortion rate in Ohio.

That is, it would greatly reduce the abortion rate if the GOP were expanding Medicaid and not restricting access to birth control. Which kind of shows you where their priorities lie.


 Not quite. ODH breaks it into 18-19 and 15-17, so the cohorts overlap somewhat.

2 Irritatingly, I can only find teen birth rates for California, not teen pregnancy. Those are here:

3 This is extrapolated from teen births. We don’t have national teen pregnancy data from 2010 because reportage of non-births varies widely from state to state. I assumed a constant abortion ratio in 2010 compared to 2005.

That probably isn’t the case, but abortion ratios are so strongly determined by the unemployment rate that I would imagine the fluctuations were similar enough across the board, particularly among 15-19 year olds.

Using birth rates rather than pregnancy rates lowers the r2 value to 0.24 (in 2005) and 0.17 (in 2010), which is still a stronger correlation than sexual activity.