Kent Bush: Exit polls are notoriously unreliable

Kent Bush More Content Now

I normally don’t answer letters I receive from readers. In my columns, I express my opinion. In letters to the editor, readers express theirs. Just because I work at a newspaper, my opinion isn’t worth any more than theirs. I try to express my opinion on many topics without having this space transform into a bully pulpit.

But I received a very kind and thoughtful letter after making a reference to Ross Perot making two Bill Clinton presidencies possible by splitting the conservative vote. I think if Donald Trump runs as a third-party candidate, he would damage the Republican candidate again in 2016.

My reader pointed out that exit polls showed that many people who identified themselves as Perot voters also said if they had not voted for Perot they would have voted for Clinton.

Exit polls are thought to be a good way to figure out how and why elections are won and lost. Did minorities vote for a candidate? What issues were important?

The reader asked if I had seen that data.

Yes I have. I disagree. I have no evidence. I just believe my own two eyes more than voters who often try to make a point with pollsters.

There is a reason votes need to be made in secret. Ask voters in 2008 and 2012 who voted for John McCain and Mitt Romney if Barack Obama’s race played any part in their decision. Those polls would show that race played no part in anyone voting against Obama. If you believe that, you should also know that I am considering a career as a plus-sized marathon runner.

In 2004, exit polls showed John Kerry upsetting incumbent George W. Bush by a 53-46 margin – that is seven percentage points.

Bush actually won by 3,000,000 votes.

Is it possible that exit polls were more accurate when people explained why they voted for Perot and who their second choice was? Sure.

Is it more likely that Perot was far more appealing to Republicans who were disenchanted with George H.W. Bush and Bob Dole? I think so.

We won’t ever know for sure. The only data we have says Perot didn’t help Clinton. However, that data is notoriously unreliable and seems to contradict logic and the actual results of the elections.

When I hear hoofbeats, I think horse, not zebra. That logic isn’t always right, but I will take my chances.

Kent Bush is publisher of Shawnee (Oklahoma) News-Star and can be reached at kent.bush@news-star.com.