Don't Hype Those Poll Blips
I like the Vox page aggregating the aggregators. I'm less thrilled about some tweets the site sent out this morning hyping one-day changes in the polls, or the headline on today's update.
It's hard to know how to react to, say, a seven-point swing in chances for a majority in the Senate in the Upshot forecasting model. To some extent, the problem is that the estimates of the various prediction models don't include an easy-to-interpret margin of error. The modelers generally do a good job of showing the uncertainty surrounding their forecasts (see Drew Linzer's here, or Nate Silver's here), but it's not presented in terms of the plus or minus numbers that most reporters and readers are familiar with in interpreting single polls. Although anyone who looks at those presentations carefully could see the uncertainty and the range of probable outcomes.
To back up a bit: All of the modelers who have addressed the issue, as far as I've seen, agree that anything between a 40 percent and a 60 percent chance is very close to being a toss-up. That means daily fluctuations of five or 10 points in the estimates aren't that big a deal, and shouldn't be treated as one.
Indeed, collectively, the models continue to say pretty much the same thing they've said since they were introduced: majority in the Senate is a toss-up, perhaps with a small tilt toward the Republicans.
So remember, polls roll out every day now. Some of them are going to be outliers. Even if every everything is done correctly, the nature of probability says that some odd results will show up. Go by polling averages, not single polls. Still, those outliers get thrown into the various prediction models' mixes, and they might nudge the forecasts a bit. Focus on the main questions. Including whether the battle for the majority remains a toss-up.
Of course, another sensible course is to ignore all of this, and just wait for the outcome. Unless you happen to be a campaign worker in charge of allocating resources between states, that's probably the smartest thing to do. For those of us who are political junkies and can't help ourselves, let's not get carried away by small and generally meaningless day-to-day changes in election forecasts.
What's the best wording for having the votes to organize the Senate? I'm uncomfortable with "controlling" the Senate; this isn't the House, and even a 55-45 advantage doesn't amount to party "control." Talking about chances of Republicans getting 51 or more seats is fine, though wordy and sometimes awkward. I'm just going to use majority sometimes, keeping in mind that Republicans need 51 senators to reach that majority and Democrats only need 50, with Vice President Joe Biden breaking the tie in their favor.
This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.
To contact the author on this story:
Jonathan Bernstein at firstname.lastname@example.org
To contact the editor on this story:
Max Berley at email@example.com