Google+
Close

The Campaign Spot

Election-driven news and views . . . by Jim Geraghty.

Be Wary of Pollsters’ Likely-Voter Screens! (UPDATED)



Text  



For much of the week, I’ve thought that momentum for House Republicans continues to gallop (no pun intended) ahead, but that Senate Republicans have stalled. I figured this was because when the White House, the DNC, the DSCC, the DCCC, the unions, MoveOn.org, and every other liberal organization puts millions of dollars and millions of man-hours into motivating their base, sooner or later, parts of the base wake up. But the Democrats’ base isn’t distributed evenly geographically; it’s concentrated in big cities and college towns. When Obama does a rally in Seattle, that helps Patty Murray’s Senate reelection bid more than it helps Denny Heck’s bid for the House district that includes Olympia, Vancouver, and the southwestern counties.

But the sharp-as-a-knife Sean Trende offers another observation: Some of the state-level polls are using a thoroughly unrealistic screen for determining a “likely voter”: 

For example, simply asking respondents if they are certain to vote (used by Suffolk) will sometimes let more than 90 percent of respondents through a screen. In such a situation, nearly half of the respondents who are counted will not actually vote. Even tighter voter screens tend to let through a significant number of respondents who will not actually vote.

In most years, this isn’t such a big deal. If these “excess voters” are more-or-less spread evenly between the supporters of both candidates, then they will cancel each other out.

But what happens if one party’s supporters are significantly more enthusiastic about voting in an election?

Suffolk’s polls in recent weeks:

  • CT-SEN: Richard Blumenthal 51, Linda McMahon 32. (92 percent likely (actually self-described “almost certain”) voters)
  • NV-SEN: Harry Reid 46, Sharron Angle 43. (96 percent likely voters)
  • NV-GOV: Brian Sandoval 50, Rory Reid 39. (96 percent likely voters)
  • OH-SEN: Rob Portman 47, Lee Fisher 37. (89 percent likely voters)
  • OH-GOV: John Kasich 46, Ted Strickland 42. (89 percent likely voters)
  • FL-GOV: Alex Sink 45, Rick Scott 38. (94 percent likely voters)
  • FL-SEN: Marco Rubio 39, Charlie Crist 31, Kendrick Meek 22. (94 percent likely voters)
  • IL-GOV: Pat Quinn 43, Bill Brady 37. (84 percent likely voters)
  • IL-SEN: Mark Kirk 42, Alexi Giannoulias 41. (84 percent likely voters)
  • PA-SEN: Pat Toomey 45, Joe Sestak 40. (89 percent likely voters)
  • PA-GOV: Tom Corbett 47, Dan Onorato 40. (89 percent likely voters)

Just about all of those have the Democrat doing at least a little better than other polls. Now we know why.

UPDATE: Suffolk contends Trende is misreading their data; they say the numbers on the sheet represent the sample after they’ve weeded out the unlikely voters. That would explain why such a high percentage of the respondents describe themselves as “almost certain to vote.”

Having said that, I’m not quite sure that someone saying they’re likely to vote makes them an actual likely voter. How many folks want to admit they’re probably or certainly not going to vote out of laziness or disinterest? How many people tell a pollster they’re going to vote and then don’t?


Tags: Polling


Text  


Sign up for free NRO e-mails today:

Subscribe to National Review