They are at it again. Following what can only be described as a butt-kicking win for Gov. Scott Walker in Wisconsin's recall election, many media pundits were trying to "split the baby" by acknowledging Walker's win, but pointing to exit polls that show President Obama with a seven-point lead against Mitt Romney in that state. There is just one problem with their story: It fails to acknowledge just how far off the exit poll was in the gubernatorial contest.
To add insult to injury, the elite media bashed Matt Drudge for supposedly posting information saying the exit polls showed Walker safely winning, declaring the polls to show the race as very tight. Turns out the exit polls they referred to were wrong. Perhaps we should just go to The Drudge Report to get the accurate information from now on.
I continue to contend that exit polls are darn close to useless. I prefer polling a base model and then taking key county or even precinct results as early numbers arrive in order to project a winner. It works better and actually can often be released with a better sense of accuracy than these horrible exit polls that the media are often afraid to release.
Considering the exit polls the media relied on showed a razor-thin difference between Walker and his Democratic opponent, Milwaukee Mayor Tom Barrett, the logic behind some huge lead for Obama, produced by the same exit polls, melts away. Walker defeated Barrett by seven-point margin.
Apply that same analysis to Obama's seven-point lead in the same exit polls and the race in Wisconsin is actually closer to being dead even. Wisconsin is hardly a bastion of Republican power or conservative might, yet it is even up for grabs. And this says much more about all of the polls and predictions we have seen to date.
In order for an exit poll to truly have any value, it must somehow take the various selected precincts from which information is gathered and attempt to place it in some "model" that reflects party identification, age, race and gender proportions normally seen or expected for a certain statewide race. My guess is that these polls likely skewed stronger towards younger voters and perhaps Democrats than was the actual turnout.
This would suggest that pollsters who are "weighting" their polls in other states, particularly the critically close swing states such as North Carolina, Florida and Ohio, might want to reexamine their allocation of young voters as well as consider increasing their proportion of "independent" voters that they use as they process the raw data they collect.
What might that mean in terms of these states? It might mean that Florida and North Carolina, traditionally Republican presidential states, are not as close as recent polls suggest, with Romney leading Obama. It would also make states like Missouri and Colorado, considered potential swing states that lean Obama, more than likely toss-ups that could easily go Republican in November.
And I am not a GOP rah-rah pollster (yes, my opinion is more rah-rah!). In fact, when polling for one of the nation's top sources of political news in 2008, my results showed Obama carrying both Florida and North Carolina. But a few things have changed in four years that cause me to believe that, if the election were held today, Romney would carry both states. First, we do not see the "on fire" energy for President Obama in 2012 among voters under 30 that existed for candidate Obama in 2008.
Second and more importantly, the percentage of those who view themselves as moderates or who say they are "independent" as to their party affiliation in a general election (regardless of how they might be registered for primary purposes) is not only high, but is leaning more Republican than in 2008. Many of these same "independent" voters desperately wanted change in 2008 and supplied Obama with the margin of victory in the major swing states.
We are still far from November, and as will remain my contention, the result of the presidential debates will likely determine who wins the contest for the presidency. But my Wisconsin-based interpretation of the national electoral mood is based on something more tangible than a guess about the strength of unions or weariness over recalls. My evaluation is tangible based on the exit polls -- as compared to reality.