Saturday, August 26, 2006

Causation vs. Correlation Part II

Over at Hot Air, Allahpundit highlights a paper by a Northwestern law professor on statistics, he believes, show that exposure to pornography decreases the incidence of rape.

Here is the link to the paper itself.

Of course, the professor has to remind any who read the paper that correlation and causation are not the same things, but that doesn't stop him from assuming causation anyhow.

His theory is born from a statistical analysis: The incidence of rape has dropped by 87% since the 70s, while access to the internet (and thus pornography) has increased tremendously.

The strongest hand he plays is an analysis of 8 states, four with the greatest access to the internet and four with the least internet access. He then compares the rape statistics between the 1980 and 2000. Overall, states with the greatest net access in 2000 decreased from the 80s, while states with the least access in 2000 increased.

It's not perfect. Individual data from the states is inconsistent. Some of the states with the least access decreased, while some of the states with the most access increased. The data would have been much more useful if public vs. private access had been compared as well. People are unlikely to look at porn at work or the library (though, as I've shown, it does happy).

Unfortunately, the professor offers only his theory of porn access. He floats a few alternative theories, but just long enough to dismiss them out of hand.

I would love to see a paper that actually analyzes alternative theories. But, once again, it's just easier to assume what we want and go from there.

(Hm . . . does this mean that religion makes us rapists, too?)

2 comments:

Hal said...

Sometimes, Steve, I can't tell whether or not you're joking.

Hal said...

If you're convinced of this, Steve, then I have a bridge in New York you might be interested in.