Thursday, July 15, 2010

Working with de Rugy's Data

In the comment section of this post, Xenophon brought up someone that I personally don't feel has much good to add on the fiscal stimulus debate: Veronique de Rugy. She regularly confuses the issues and doesn't seem to even understand her opponent's arguments.

Xenophon mentioned that de Rugy was apparently vindicated in her disagreement with Nate Silver over the importance of state capitals in the determination of the receipt of stimulus funds. I was curious and checked out her data. You can see my results in this google document (please let me know if the link isn't working).

To summarize,

(1.) Both the state capital variable and the Republican variable are significant. However, the state capital variable seems to explain about 30% of the variation in stimulus dollars received, while the Republican variable only explains about 1% of the variation. So having a Republican representative meant less stimulus money, but the impact was trivial. Nate Silver identified a much, much, much more improtant factor, and de Rugy is making a big deal about a statistically significant but relatively small factor.

(2.) The biggest difference between Rs and Ds is between Rs and Ds that represent state capitals. Rs and Ds that don't represent state capitals aren't all the different from each other in stimulus receipt.

(3.) Among the Rs and Ds representing state capitals, the biggest difference is between districts with both a Republican representative and a Republican governor, and everyone else. It's that governor/representative combination that makes you considerably less likely to receive stimulus. This suggests to me that a lot of the Republican story isn't about what's going on in Cognress - it's about whether Republican governors are accepting stimulus money or not (many weren't).

It doesn't take much poking around in the data to see that de Rugy's framing of the issue was a little misleading.

Please check out the google document for more detail.

No comments:

Post a Comment

All anonymous comments will be deleted. Consistent pseudonyms are fine.