Why Early Senate Polling Is Usually Useless
February 13, 2015 · 1:20 PM EST
I never pay too much attention to early polls, since snapshots of a race more than 18 months before Election Day can be misleading.
And political parties ought to be careful about crowing too loudly about early polls for fear someone will look too closely into them.
A Feb. 10 “rapid response” press release from the National Republican Senatorial Committee didn’t waste any time noting a series of new Quinnipiac University polls in Florida, Ohio and Pennsylvania found Republican Sens. Marco Rubio, Rob Portman and Patrick J. Toomey in decent shape for re-election.
Toomey’s numbers caught my eye.
A plurality of the 881 Pennsylvania voters in the Jan. 22-Feb. 1, 2015 survey, 43 percent, approved of the job that Toomey is doing, while only 25 percent disapproved. And 37 percent said that Toomey deserved to be re-elected, while only 29 percent said he did not.
Those are certainly not terrible numbers these days, given the low regard in which Congress is generally viewed. But they are hardly evidence the incumbent is headed for re-election.
The problem is that Toomey’s numbers in the Quinnipiac survey aren’t very different from then-Sen. Mark Udall’s numbers in a June 5-10, 2013 Quinnipiac University survey of 1,065 registered voters in Colorado.
Udall’s job performance ratings were 45 percent approve, while 31 percent disapprove. Forty percent of respondents said that he deserved to be re-elected, and only 33 percent said that he did not. And Udall lost.
I’m certainly not saying that Toomey will lose, or that Democrat Joe Sestak, who lost narrowly to the Republican in 2010 and wants a re-match next year, is anywhere near as good a candidate as was Colorado Republican Cory Gardner, who defeated Udall.
But it’s wise not to read too much into early surveys, remembering that national atmospherics and state-specific factors will have a substantial impact on how races develop.
Don’t ignore where candidates start. But always remember that where they end can be very different from where they begin.