You sexy correlational bitch!*

So it looks like teenagers who watch a lot of sexual themed TV are more likely to get pregnant. Let's see what the media says. First, here's the actual researchers:

The study found that frequent exposure to TV sexual content was associated with a significantly greater likelihood of teen pregnancy in the following three years.

This, obviously, is a simple correlation (with some other variables controlled, thus likely some variation of a partial or semi-partial, but still, simple correlation): As Variable 1 (amount of TV sexual content viewed) increases, Variable 2 (likelihood of pregnancy) also increases. Will the media describe the findings as correlational? Here's the NYT:

Shows that highlight only the positive aspects of sexual behavior without the risks can lead teens to have unprotected sex (emphasis mine)...

Ah yes, one of the most basic science errors (lies?): Variable 1 is correlated with Variable 2, thus Variable 2 leads to Variable 2. Damnit, man! Anyone who has ever taken a freshman level social science class knows correlation does not mean causation (although they may not know, as I didn't until just now, that this was called the cum hoc ergo propter hoc logical fallacy)! How did NPR do?

Still, the authors say the study has limitations — that they can't rule out other factors that may influence the findings. For instance, it's possible that teens with advanced sexual attitudes are more likely to seek out more TV shows with sexual content.

Good job, NPR. There are plenty of other reasons why Variable 1 might be correlated with Variable 2 aside from Variable 1 causing Variable 2. This shouldn't be hard!
*The title is for you, Leslie.