Governors: Democrats' Best Case Scenario
Yesterday, I used Rasmussen polls to find the worst case scenario for Democrats in gubernatorial races. Last night I was thinking, well, what's the ...
Yesterday, I used Rasmussen polls to find the worst case scenario for Democrats in gubernatorial races. Last night I was thinking, well, what's the best case scenario for Democrats?
The easiest thing to do would be to look at a pollster with Democratic-leaning house effect equivalent to Rasmussen's and see what they're finding. The pollster that fits that description this cycle is Research 2000, which does polling for the Democratic blog Daily Kos, among others.
The problem is that while Research 2000 is a fairly prolific pollster, it's not as prolific Rasmussen. No one is. Research 2000 hasn't recently polled the gubernatorial races in most states.
However, here's what Nate Silver says:
Rasmussen, thus far, has a Republican-leaning house effect of about 5 and 1/2 points. So if Rasmussen, for example, has a Republican leading by 7 points in a particular race, an average pollster would have the Republican ahead by only 1 or 2 points. Research 2000, on the other hand, has a Democratic-leaning house effect of about 4 1/2 points. If they show an R+7 in a particular race, it would be the equivalent of an R+11 or an R+12 from an average pollster. These are, obviously, very large differences: it implies that if Research 2000 and Rasmussen were to poll the same race, we'd expect about a 10 point difference between them.
So, we can use Rasmussen polls to theoretically guess what Research 2000's result would be on a state-by-state basis. If the Republican is leading by much less than 10 points, Research 2000 would show the Democrat leading. If it's much more than 10 points, then even Research 2000 would place the Republican in the lead.
That's basically what I did in the chart below, although I also took into account the results of other pollsters (if Quinnipiac or Public Policy Polling, survey firms with smaller house effects, show a Democrat with the lead, that probably means Research 2000 would too) and actual Research 2000 polls. Here's the result:
Yesterday, with Rasmussen, Republicans were winning 23 races and Democrats 9, with the rest unknown or up for grabs. Today, I'd have changed that to 24-9 because of the new Public Policy Polling survey showing Republican candidates leading in Alabama (that's also why Alabama is Republican above).
In contrast, under the Research 2000 scenario, Republicans win 16 races and Democrats also win 16 races, with the rest close or, in the case of Maine, lacking polling. Republicans' net gain would only be a couple of seats. Democrats would win more gubernatorial races than they did in 1998 or 2002.
What's more, if I'd really been a stickler for following the 10-point rule precisely, I'd actually have given three of the question marks above to the Democrats. According to Rasmussen, the strongest Republican leads the strongest Democrat by six points in Wisconsin, seven points in Vermont and seven points in South Carolina.
At this point, I should note something that I should have noted from the start. What I'm really talking about here is the best case scenario for each party if the election were held right now. That is, after all, the only thing that any poll measures. Polls, by their nature, are a snapshot of public opinion at a certain point in time.
That's also what makes the difference between the two scenarios so striking. Rasmussen indicated that Republicans would win a landslide on par with 1994. Research 2000 says that Democrats might well win a majority of the governor's races. But, the actual best case scenario for Republicans actually is better than what Rasmussen indicates because Rasmussen's model could be right and, additionally, events could strengthen the G.O.P.'s position between now and November. And, the same logic means that Democrats could do even better than the Research 2000 perspective.
Join the Discussion
After you comment, click Post. You can enter an anonymous Display Name or connect to a social profile.