One of the postscripts of Wisconsin Gov. Scott Walker’s recall election victory yesterday was the news media’s prodigious reporting of the associated exit polling. Along with coverage of Walker being re-elected was their secondary story that President Obama leads Mitt Romney within this same voting universe, 51-44 percent. But, just how accurate are typical exit poll findings?
As we have seen in many elections, early exit poll reporting has been wrong. Most famously was the networks calling the state of Florida for Al Gore back in 2000, based upon exit polling, and announcing that the then-vice president had defeated George W. Bush. But, there have been many other incorrect predictions based upon this type of electoral polling over the years, just not as famous as in the Bush-Gore contest.
Exit data is sometimes flawed because the polling samples are not randomly selected. This lessens the reliability factor. Survey firms will choose polling places based upon electoral voting history, selecting precincts that accurately reflect the statewide voting patterns. While at the polling stations, it is the voters themselves that volunteer to answer the survey questions instead of the pollsters conducting more random sorting procedures within the sample cell. Therefore, the data found in these polls should not be considered in the same reliability category as benchmark or even tracking surveys that are conducted over a long period of time. That’s not to say exit polling has little value. One simply must take into account that their reliability factor is much lower.