Saturday, September 13, 2008

Why the Ohio Poll May or May Not be Accurate

The University of Cincinnati's Institute for Policy Research released its latest poll on the presidential race in Ohio. This poll was taken of 775 "likely" voters. It has a error rate of + or - 3.5%. The Institute doesn't give the internal demographic breakdown of the poll online, but according to the Institute's Director, the breakdown was 48% Democratic respondents, 43% Republican respondents, and 9% independent respondents. He did not have other demographic information available, although the Institute will be glad to send it out by email.

The problem for the Institute is that no one really knows who will be a "likely" voter in this election. Traditionally young voters, IE, those under 30, and minority voters, IE, African-American, Hispanic, Arabic, and Asian, don't always vote at the same rate as non-Hispanic white voters. Further, single non-Hispanic white women vote at a different rate than married white non-Hispanic women.

Given that this is the first time in American history that a African-American candidate has been nominated by a major political party, and given the fact that this candidate polls significantly higher than his opponent among voters under 30, it is hard to predict the outcome. Throw in the fact that there is a woman on the GOP ticket and that she has apparently has great emotional appeal to evangelical, Christian voters, and prediction of turn-out becomes even more difficult.

What Democrats can't do, though, is lose sight of the fact that the election is still approximately eight weeks away. Eight weeks is an eternity in a political campaign.

In 2000, in early October, there was a poll in Ohio showing Bush 10% ahead of Gore. Gore pulled his resources out of Ohio, and moved them to Florida. On election day, however, Gore lost Ohio by around 3-4%. Imagine what might have happened if he had stayed in Ohio, running ads, and campaigning here personally. The whole past eight years might not have been the nightmare they have been under Bush.


Anonymous said...

Something else to keep in mind: younger voters, who are breaking for Obama in a big way, often do not even have a landline phone anymore. They rely on their cellphones, which are not used for polling. That fact alone could give us 1-2 percentage points, especially if younger voters actually get out and vote.

Anonymous said...

Great point. Not just the young voters though. We got rid of a land line years ago ourselves. We are middle aged.2 cell phones, so who needs a land line anyway?

I think Barack is going to win in reality by a land slide. America cannot be that stupid to really vote for Palin/McCain...can they?