As you all know, correlation doesn't imply causation, but I was curious so I ran the regression anyway. Any thoughts on this?
Sources here and here.
Update: I've re-run the regression using more up-to-date data, you'll find the new sources above. Ultimately, the results are pretty much identical, but I thought I'd be more careful with the data. In my second run-through, I tried to find out if Excel could label the data points for each state, but apparently the only way is to manually input the labels, which is just horribly time-consuming. In any case, I've uploaded my spreadsheet with all of the data, for those of you interested in each individual state's data. Also, if you hate Google's formatting of the file online, you can go to the top right of the screen after opening up the spreadsheet and hit File > Download to open it on your computer.