Can We Really Trust Political Polls?

Political polling’s widespread use in this election is drawing attention to what those numbers really mean.

Before the first presidential debate, Governor Mitt Romney was behind President Obama in the Polls. But afterwards, he jumped a few points to catch up with the president. We also watched the polls before and after the Vice Presidential debate, but there was not much change.

What is it that polls really measure? CU Journalism Professor Sandra Fish said that “polls give an idea of the horserace aspect of the campaign, but they don’t necessarily predict the outcome.”

The first public opinion poll recorded was in 1824 in The Harrisburg Pennsylvanian newspaper to predict who would win the presidential election between Andrew Jackson and John Quincy Adams. When Jackson won both the popular vote and the electorate vote (but not the presidency due to other issues), polls gradually became more popular.

Polls have also made some serious mistakes over the years. In 1936, after years of correctly predicting presidential victories through postcard mailings, the Literary Digest predicted that Alf Landon would win over Franklin D. Roosevelt. Obviously, they were wrong. The poll was skewed because the cards were being sent to mostly affluent Republicans.

At the same time, George Gallop conducted his own scientifically based poll, and predicted a Roosevelt win.

Today polls are much more scientifically based than postcard mailings. But are they really a good predictor of elections and do they really tell us the truth about different regions' political preferences?

According to Associate Professor of Journalism, Elizabeth Skewes, “polls can gauge [regional] opinions in the moment but lots of things will change over the next few weeks that will cause voters to shift... how they feel today may not be how they feel two days from now.”

In 2000, Florida exit polling caused several news stations to jump the gun and declare an Al Gore victory over Bush.

In 2008, the polls predicted an Obama win over Hillary Clinton in the New Hampshire primary, but they were wrong again.

1948’s national polls of the election between Thomas Dewey and Harry Truman produced headlines reading “Dewey Defeats Truman."

After all of these mistakes, how much credibility should polls be given?

Senior Art History student Rebecca Winterfield believes, “if they’re done correctly, or if they’re done to survey the widest demographic possible,” then they can be.

Freshman Krista Marks says, “if it is varied enough with the people that it asks it can be credible without an actual voting taking place."

Even though polls have made a number of mistakes over the years and only 2-percent of voters changed their mind after watching the debate, Pew Research Center’s Scott Keeter says, “polls are by no means a perfect instrument, but they do make it possible for more opinions, held by a broader and more representative range of citizens, to be known to the government and potentially be heeded.”