As the “Farmer’s Almanac” predicts the weather for the upcoming winter, it is time to think about the data we use to predict political seasons. Since 1990 the Minnesota electorate has been very volatile. If it were under psychoanalysis, it might be considered schizophrenic or bi-polar. This year appears to be no different.
Despite early polls that showed Barack Obama winning the presidential race and Sen. Norm Coleman, R-Minn., winning his Senate race by wide margins, those gaps are closing. Is that because voters in Minnesota are changing their minds or is it because early polls are not strong predictors of future voter behavior? Or is it simply because in Minnesota, the weather and electoral politics are as predictable as the other?
Our recent electoral history has been generally unguided by polls. Just ask Skip Humphrey, Roger Moe and Mike Hatch, all of whom were leading in late polls in their runs for governor. Or ask George W. Bush, who twice thought that Minnesota was winnable for a Republican running for president.
Instead our polls are perplexing and Minnesota voters are sometimes unpredictable.
The Quinnipiac/Wall Street Journal/Washington Post poll in June showed Coleman leading opponent Al Franken by 10 points, and Obama leading John McCain by 17 points. That illustrates the Minnesota ticket-splitting legacy as well as anything. The July poll showed Obama’s lead shrinking to 2 points, and Coleman’s lead expanding to 15 points. That makes sense, you say, it’s a GOP trend.
But now polls are showing closer races.
The Humphrey Institute/MPR poll released this week calls the Senate race a “toss up” and shows Franken with a 1 point lead. In the Obama-McCain race, it shows Obama up by 8 points.
The most recent Rasmussen poll reported a 4 point Obama lead in the presidential race in Minnesota, compared to 13 points in July. It also gave Coleman a 3 point lead, compared to a 3 point to lead for Franken in July. Confused yet?
Now take Survey USA, which partners locally with KSTP-TV (Channel 5). In June it had the presidential race a dead heat in Minnesota with Obama edging McCain by 1 point, and Franken trailing Coleman by 12 points in the Senate race. This week it shows Coleman’s lead shrinking to 7, and Obama leading McCain by 4.
This shouldn’t be surprising, but it’s a good reminder of how there’s little if any science to politics.
What poll was the closest to actual voting results in Minnesota’s Senate race in 2006? Survey USA had the actual voting within 2 percentage points for both the Democratic and GOP candidates. It put Democrat Amy Klobuchar at 56 percent and Republican Mark Kennedy at 40 percent. The actual vote total: Klobuchar won with 58 percent of the vote and Kennedy had 38 percent. The now defunct Star Tribune Minnesota Poll had the spread almost dead on, 21 points.
In 2004, John Kerry beat Bush by 3.5 points in Minnesota. The closest poll in that race? Rasmussen had Kerry winning by 1 point. But the Real Clear Politics average had Kerry by more than 3 points.
The current Real Clear Politics average has Obama leading by 4.5 points, and Coleman leading Franken by 6.
The big difference in 2004 and 2006 compared to today is that the top issues in the previous elections were related to war and national security. The economy began to emerge as an important issue in 2006. It now ranks as the top issue for more than 50 percent of voters in the Quinnipiac poll. The second most important issue is energy, and in Minnesota, poll respondents trust Obama’s approach to energy more than McCain’s.
Meanwhile, the most recent Rasmussen poll shows national security as the top issue among 23 percent of respondents; 42 percent rated the economy as the top issue.
What explains the differences in the polls? One reason could be that Rasmussen under polled some groups or that the lower gas prices in the last couple weeks have made a lot of people less concerned about the economy. It also appears Rasmussen isn’t even asking the question about energy. Hmm.
Confused? Welcome to Punditville.
Predicting Minnesota voters is like predicting Minnesota weather. Be it Doppler radar or all the polls in the world, no one really knows if it’s going to rain or snow — or if it’s going to be Obama or McCain, or Franken or Coleman.