Robot babies lead to more pregnancies and more births

Ann Althouse brought some interesting information to my attention

“… but a new study using lifelike simulated babies in Western Australian schools had a surprising result: girls enrolled in the Virtual Infant Parenting Program (VIP) were twice as likely to give birth in their teens.”

Now, the first problem is that it’s a study in the Lancet, so you’re just about guaranteed that it’s crap.

The second is that they’re passing around some really noisy data.  This American Life had a segment about robot babies, and the producer of that segment wrote a blog post about the release of the study (which was in Australia vs America for “This American Life”)):

Here are the numbers:

17% of the intervention (robot babies) group had teen pregnancies; while 11% of the control group had teen pregnancies.

Thanks to the magic of the internet, I was able to listen in on Dr. Brinkman’s press briefing in Australia last night, and ask her a couple of questions.

The first thing I asked was, of course, motivated by my observations of Rachel: Was there evidence that the simulators made teens interested in becoming moms? Or less afraid of accidental pregnancy?

Brinkman said there was no way to know the answer to this question. The study was designed to track pregnancy, not whether the pregnancies were intended or unintended. But, she added, they did study the pregnancy termination rate in both groups. And the group that got the infant simulators had a 6% lower proportion of abortions, compared with the control group. But, of course, there’s no way to really know if that lower rate means the girls who experienced the infant simulators felt more comfortable with the idea of becoming moms.

I decided to run the numbers myself (it only makes sense to compare the last 4 rows, which are normalized, as opposed to the first 4, which aren’t):

Study Control Compare
Girls 1267 1567
Live Births 97 67
Abortions 113 101
Pregnancies 210 168
Birth Rate 7.66% 4.28% 1.790561792
Abortion Rate 8.92% 6.45% 1.383723929
Pregnancy Rate 16.57% 10.72% 1.545974743
Abortion as % of Pregnancy rate 53.81% 60.12% 0.895049505

So the 17% and 11% are combination of abortion and live births (I wonder if there were any miscarriages?).

Key facts: The girls in the study were 79% more likely to give birth to a baby, 38% more likely to have an abortion, and overall 55% more likely to get pregnant than the control girls.

The only place where the abortion rate went “down” for the study girls was that 54% of the 210 study girls who got pregnant had an abortion, whereas 60% of the 168 control girls who got pregnant had an abortion. With numbers that small, a 6% difference is meaningless.

Personally, I think those girls are discovering that a teenage girl with a robot baby is a lot more interesting, and gets a lot more positive feedback, than a teenage girl with a real baby. But I could be wrong.