Nate Silver contradicts the conventional wisdom and demonstrates that Democrats are actually more enthused about the upcoming elections than they are normally in midterm cycles. The problem is that the Republicans are off-the-charts excited about voting.
I have one question about how these pollsters are looking at things. During the last two cycles, voters dropped their Republican affiliation in droves and become independents and Democrats. How are these pollsters accounting for this? Registered Republicans make up a much smaller slice of the pie than they did in 2006.
With that caveat, I still think we’re in trouble in these upcoming elections, and I have thought that for a while. My great hope is that many of the Republican candidates wither under close scrutiny. But, in the meantime, we have to figure out how to do a combination of exciting our own base and depressing the Republican base, We’re doing okay with registered voters, but the likely voter models are showing this big turnout gap. We can’t allow that to happen. If our base is already more excited than normal, how do we get them even more fired up?
My sense is that pollsters have folks self-identify as Republican, Democrat, or Independent. And self-identify as a “likely voter” through a variety of different questions, depending on the pollster.
There is another variable to consider. Are the people who see that it is a pollster on the caller-ID and don’t answer randomly distributed among Democrats, Republicans, and Independents?
Polls don’t get people out to vote in contentious elections. GOTV campaigns do. And so do folks in their personal network asking them to go vote.
Okay. Let me clear about what I am wondering about.
Take a simple telephone poll.
Type 1 if you are a registered Republican
Type 2 if you are a registered Democrat
Type 3 if you are registered to vote with an independent or Third Party
Then you ask them to Type 1 if they will definitely vote, etc.
Once you have that data, you have to model the electorate. You have to figure out what percentage of the electorate are registered Republicans. And if that number is significantly lower than it was in 2006, then you can wind up overrepresenting them in your poll. So, my question is whether or not pollsters are doing a good job of accounting for this.
In Pennsylvania, for example, we had a five-week period with no primaries between Mississippi and our election. We got oodles of people to register as Democrats so that they could vote for Obama or Clinton, and those people are mostly still on the rolls as Democrats. The Republicans lost probably more than 100,000 registered voters here since 2006. But if your baseline is the last midterm election, you’re going to have a bad model. See what I mean?
My guess is that different pollsters have modeled it different ways. I’m not sure that the 2006 election is the sole model of how pollsters are modeling. Nate Silver’s forecasting model takes a whole lot of elections into account and does a pretty sophisticated process of adjustment. And it looks miserable for Democrats. But Nate’s model is a derivative model which depends on the underlying polls. And if all of the underlying polls are skewed, so will the forecast.
The meta discussion of this in the media indicates that they might be off. But we won’t know until November 3.
And the bigger error might be the “as of today” assumption. There is more movement now than usual. And the trend for now seems to be toward more Democrats turning out.
And Nate Silver today correctly relabeling the “enthusiasm gap” as what it is — an expected turnout gap.
Two words: negative advertising. We have the advantage that ours can be totally fact-based and theirs can’t. Even in America I think that will have some weight in the end.
That was my first thought too, DaveW. If likely Democratic voters are more likely to vote than in a typical off-year election, then maybe the best (cheapest, most effective) way to narrow the “enthusiasm gap” is to decrease the excitement likely Republican voters have to actually turn out and vote.
As a subset of that, highlighting Republican plans to privatize Social Security might be a good way to accomplish that goal.
On this score, the US Chamber of Commerce has handed Democrats an excellent issue. Are foreign corporations laundering money through the US Chamber of Commerce to buy Republican seats in the election?
Perfect issue, and it seems like some Dems are picking up on it.
Two words, twice.
Public Option
Dawn Johnsen
Because all the Democrats I know, that’s all they ever talk about.
(Do you think just maybe I spend too much time online?)
If that’s all the Democrats you know talk about, you are in a progressive stronghold, which won’t be in play in this election.
Neither of those are major issues in much of the country because of the economy.
It was 100% snark.
Before my second cup of coffee.
Thanks.
I think we would be totally screwed if a) the economy wasn’t seeing signs of life and b)OFA wasn’t involved in these elections. The Dem candidates don’t know how to make the historic legislation work for them in ads and in their talking points but the economy speaks for itself (and it is coming around) and OFA is targeting all the new Obama voters who probably are being undercounted in the polls.
So my back of the napking analysis, (which possibly isn’t worth the napkin it’s written on!) is that the close ones we will pull out if Obama won in those districts/states. That’s why I’m not believing that Feingold is going to lose. The polls had the presidential election wrong on that by at least 5-10 points.
A big problem is getting the Democratic message out. I was watching CNN this morning (only news on) and they were positively salivating over the “certainty” that the Republicans will take both the House and the Senate. And every station break had another smear from Mark Kirk or Bill Brady.
I’ve seen the Debbie Halvorsen ads on another channel. Why is anyone even thinking about voting for her opponent? Anti-minimum wage. Anti-Social Security. Hers is not a rich district. It reminds me of the mechanic from West Virginia that was a classmate back in ’07. Obviously from a dirt poor family in a dirt poor state. He told me “I believe that George Bush is a good decent Christian man.”
1) Modeling is absolutely crucial. I share your suspicion that models based on 2006 may be out of date. 2) The daily grind from cable “news” gets those of us upset who pay attention, but most people have real lives and don’t. 3) By my limited observation, some of the GOP incumbents are being complacent and smug. Let’s see if a few of them don’t get surprised.
All GOP candidates need to be asked, for the record,
“Have you ever dabbled in witchcraft?”