The undecideds break for the incumbent and other polling myths

In the last couple of weeks, the Democratic Congressional Campaign Committee (DCCC) and Congressman David Cicilline have released figures from a set of three polls suggesting that Cicilline has a six-to-11-point lead over his Republican challenger Brendan Doherty.

Neither the DCCC nor Cicilline agreed to release the full polls. So naturally, there's been some skepticism. And that's healthy. But there's also been some commentary of the even-if-the-numbers-are-right-this-doesn't-look-great-for-Cicilline variety

Doherty's campaign has focused on the incumbent's failure to get 50 percent of voters in any of the three polls. "The fact that Cicilline can’t win 50% of the vote under those circumstances in an overwhelmingly Democratic district shows just how little people trust him,” Doherty's campaign manager Ian Prior said in a statement, at one point.

Prior didn't explicitly say that Cicilline's failure to get to 50 percent in the polls suggests he could lose on election day. But that's the implication. And it fits in with an oft-repeated bit of political wisdom: any incumbent who can't get to that 50 percent figure in the run-up to the big day is in serious trouble.

Venerable political analyst Jennifer Duffy leaned on another old saw, telling WPRI-TV that voters who are still undecided, this late in the game, tend to break for the challenger. Cicilline's apparent lead, she was suggesting, may not be as safe as it seems.

The only trouble with these arguments is that, well, they're wrong. The numbers don't bear them out.

Two years ago New York Times blogger Nate Silver looked at 83 gubernatorial, Senate, and Congressional incumbents who, with 30 days to go before the election, led their opponents in the polls but claimed less than 50 percent of the vote. It turns out 80 percent of them went on to win re-election. They weren't exactly safe, but certainly safer than the doomsayers would suggest.

And there's not much evidence, as Silver reports, that the majority of undecideds broke for the challenger in these races. The incumbents he studied led in the polls by an average of 8.1 points with 30 days to go and wound up with an average victory margin of 7.2 points. Not much of a shift at all.

Silver offers up this caveat: it is worth noting the sheer size of the undecided pool. A large cohort of undecideds suggests a broader volatility in the race. And that makes any poll - including a poll that favors the incumbent - less than reliable.

The last poll figures the Democrats released - Cicilline's - gave the incumbent a 46-36 lead, with independent David Vogel claiming support from 7 percent. Assuming the numbers are correct, we are left with a relatively large number of undecideds.

That should give the Doherty camp hope. But the arguments trotted out to date - the 50 percent threshhold, the undecideds breaking to the challenger - should not. 



| More

 Friends' Activity   Popular 
All Blogs
Follow the Phoenix
  • newsletter
  • twitter
  • facebook
  • youtube
  • rss
Latest Comments
Search Blogs
Not For Nothing Archives