Bell Curve The Law Talking Guy Raised by Republicans U.S. West
Well, he's kind of had it in for me ever since I accidentally ran over his dog. Actually, replace "accidentally" with "repeatedly," and replace "dog" with "son."

Tuesday, July 13, 2010

Worst Political Science Article Ever

So here is an article by Sean Trende for RealClearPolitics, an outfit that is becoming increasingly partisan (GOP). It's appallingly bad. It seeks to explain Obama's approval ratings by saying it's not the economy, but his poor handling of health care and spending. His own data shows that models based on economic performance nicely track the President's approval ratings, but he claims that it is significant that Obama's performance is consistently slightly below the prediction. The chart above - which he also prints - shows even more dramatically the same for Clinton, with no explanation. In fact, this article proves the opposite of what it is intended to do: regardless of how people answer pollsters, the economic news is the best predictor of the President's performance. Here on this blog I predicted a rosy summer after the good First Quarter. I was right for a month. Then the Second Quarter looked really bad. His approval ratings have fallen back to dead even with disapproval, and Dems everywhere are struggling in polls again. What will the third quarter bring? We'll know it when we see it. If the third quarter shows strong economic growth (ending Sept 30) then Sept and Oct will be good months for Democrats. If the economy starts growing again more strongly in July and August, the public will 'feel' it in September. They will know it when they get the stats in early October. If that doesn't happen, if the economy continues to stall out, it will be hard Fall.

The Health Care debate may have dominated 2009, but it was as much effect as cause.


Robert said...

I'm glad I'm not the only one who thinks RCP is a conservative website. To think how good it used to be...

Raised By Republicans said...


I agree with the substance of your criticism of Trende's blog post.

But I'd like to quibble with your title. Trende's article is not "political science." It is political polemic with a lot of statistics thrown in for flavor. It wasn't peer reviewed or (probably) thoroughly researched.

Also, as you point out, the statistics are not very well analyzed. The graphs he shows certainly seem to show a strong correlation between the economy and Obama's approval even if the purely economic model is not fully deterministic.

The methods he utilizes aren't well matched to the assertion he trying to substantiate. He makes no attempt to support his methodological choices. These are all hallmarks of polemic not social scientific research.

Finally, saying that economic variables are significant but deterministic predictors of Presidential approval is hardly news to real political scientists. We've know that for decades. It is however, probably news to partisan blogger/columnists like Trende.

Raised By Republicans said...


I think we're in good company thinking that Real Clear Politics is getting increasingly partisan. I've read stuff in the blogosphere about Real Clear Politics being a right leaning if not an outright GOP site since 2008.

The Law Talking Guy said...

Okay, RBR, I get it. The article is presented as scientific data, so I thought I'd demote it from mediocre commentary to worst political science ever. =)

The Law Talking Guy said...

What I've noticed about RCP is (1) the choices of articles to highlight are now overwhelmingly right-leaning and (2) the polling data makes way too much use of Rasmussen data.