Toughest job this election: Pollster

By Jay Warren
WSLS10 Anchor
Published: October 20, 2008

I would hate to be a pollster this election season. There are just too many variables to consider.

Already the job of a pollster has been complicated by modern technology. Polling firms have struggled to figure out how to include people who only use cell phones in their samples. Leaving out this important group is make any poll instantly invalid. But I digress.

There are other problems this election season like the historic nature of both campaigns and how that could impact turnout.

Will race play a positive or negative role? Will Obama’s presence on the ticket increase African American support and by how much? Or, is there a Bradley or Wilder Effect, in which people will tell a pollster they are going to vote for Obama, but in the voting booth racist tendencies come out and they vote for McCain? How does a pollster account for this?

Will all the newly registered voters actually show up to the polls? First time voters are notoriously unreliable. The Obama campaign has registered millions of new voters this election season. Will they actually vote?

And while we’re at, the only group more unreliable than new voters could be young voters. Polls show college aged adults swinging dramatically to Obama. Anecdotally, I have found much the same from the class I teach at Virginia Tech and my time on campuses across the area. College kids are excited about this election. They appear engaged and are even involved in the campaign. But, John Kerry was relying on a heavy turnout among this group. It didn’t happen. Obama is being smarter. He isn’t betting the farm on the youth vote. Still, how do pollsters factor in this group of unreliable voters, many of whom don’t have a land line and use cell phones exclusively.

To account for these variables, each pollster comes up with an elaborate formula. Each formula is different which accounts for some of the wide swings we see in the polls. Rarely do they all line up 100%. Each is pollster is trying to guess the real turnout for African Americans or young voters or new voters or women or “closeted’ racists. It’s part science, part research, and part guessing.

You see now why I wouldn’t want to be a pollster this election season!