RSS Feed

Graphs and False Correlations

Posted on

In February, BuzzFeed posted this article describing teen pregnancy rates in 2008 (which came from a paper from the Guttmacher Institute). The article (it’s really short: go look) has two maps.

The first is from the paper, showing color-coded states according to teen pregnancy rates:

Teen Pregnancy Rates, by State, 2008

Teen Pregnancy Rates, by State, 2008, Guttmacher Institute

The second is a map of states that require contraception education in schools:
States that Teach Contraceptive Use in Sex Ed, 2013

States that Teach Contraceptive Use in Sex Ed, 2013, BuzzFeed

I have a problem with BuzzFeed’s implications from the second map.

BuzzFeed is trying to imply that contraception education is the answer to teenage pregnancy rate increases. Besides the fact that the years of the studies/data are off, the bigger problem is that the maps are not aligned very well to make this correlation true.

According to the paper, the states with highest pregnancy rates in 2008 were NM, MS, TX, NV, AR, and AZ. The lowest were NH, VT, MN, ND, and MA.

According to the contraceptive map, these are the states that require sex ed in classrooms: WA, OR, CA, HI, CO, NM, AL, NC, SC, VA, WV, MD, DE, NJ, RI, VT, ME (and Washington DC). That means that Minnesota, North Dakota, New Hampshire, and Massachusetts don’t have contraception education but have low pregnancy rates, and New Mexico has contraception education, but has the highest pregnancy rate in the country. BuzzFeed wasn’t the only one complaining about the lack of sex ed; the Washington Post also made a quick implication that missing contraception education (and use) might be to blame.

What’s the big deal with posting two maps that may or may not have anything to do with each other? It sets up a false correlation between the two issues, which can be misleading. I have no data, but I’d guess that if kids do see contraception education, they probably have a greater chance of using it (as opposed to never hearing about it). But after skimming the article and its implied ties, some people may have come away with a sense of, “If we only increase sex ed in schools, teen pregnancy rates will go down!”

Okay, so what’s a better correlation? As a teacher, I see a better link between the teen pregnancy rate map and this:

By State Map, SAT and ACT scores combined

By State Map, SAT and ACT scores combined

Education levels. In general, the low-pregnancy rate states have higher test scores than the higher-rate states. This is a lovely map of ACT and SAT scores (deeper blue means higher scores). I think the map looks much more like the original.

But here’s the real issue: I also cherry-picked the map I wanted you to see. Linked to that ACT/SAT map are the individual maps for each test, and I didn’t think they worked as well to prove my point. In other words, you, the reader, are manipulated all the time by statistics and people who claim to know statistics. I could have also used this map:

USA Annual Mean Temperature

USA Annual Mean Temperature, By State Map

This map also works pretty well to my eye, and would seem to indicate that the hotter the mean temperature, the more teen pregnancies you can expect (except for a glaring problem with Nevada, and New York’s not great either).

What to do about it? Know how you’re manipulated through statistics and false connections. This article from Cracked is a decent start. Check out Number One: Correlation does not equal Causation.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: