For hardcore political junkies, I’ll recommend this fivethirtyeight.com post from Harry Enten taking a look at Senate races, trying to factor in both polling and presidential approval ratings. My main takeaway is that it’s too early to be making predictions!
Enten focuses on states, such as Arkansas, with a “disparity between what the polls are telling us and the state’s opinion of Democrats.” In addition to Arkansas, the polls are close in Kentucky, Georgia, Alaska, Louisiana and North Carolina, despite all of them voting for Mitt Romney over Barack Obama in 2012. Enten creates a model that seeks to balance the state of the current polling, which he finds is the more historically predictive measure, with the low approval ratings of the sitting president, which he finds is a secondary, but also important, predictive factor.
Here’s Enten:
More than six months from the midterm elections, current polling and past precedent are competing for our trust. I analyzed which measure is more indicative come November, and it turns out that polls are a more robust metric even though their numbers are still sparse and there’s still so much time remaining before the election. That’s not to say that a president’s approval rating is useless: It can help refine early polls to make them more accurate. This year, when we factor in both, it doesn’t look promising for Democrats in Alaska, Arkansas, Kentucky or Louisiana.
Basically, Enten’s conclusion is that while Sen. Mark Pryor is polling strong of late, he nevertheless faces an uphill battle because, by Enten’s estimate, Obama’s approval rating in Arkansas is somewhere between 30 percent and 33 percent .
If you just look at the polls, Enten’s model finds that the Pryor-Cotton race is a tossup, but when he used Obama’s approval rating in his model, Pryor’s chances fall to just 31 percent (Nate Silver’s model, which incorporates additional factors, suggested Pryor had a 30 percent chance last month, but that was before the recent spate of good polling for Pryor).