Why have polls been so wrong

John Diaz, of the San Francisco Chronicle, had an piece in Sunday’s Crossroads asking why the polls have been so long.

As one who is skeptical of the value of polling, this, of course, held great interest for me. Some of his reasons mirror my own.

1. Incorrect “weighting” of certain demographic groups. He notes one of John Zogby’s polls that was so off base in California was light on Latinos and heavy on younger voters.
2. Absentee ballots. As much as 50% of California’s votes were cast absentee.
3. Unprecedented turnout. “This year’s surge in participation, especially on the Democratic side, has made it extremely difficult to anticipate the allegiances and demographic makeup of those who cast votes.”
4. The number of people who won’t talk to pollsters is going up.
5. The influence of the 24/7 news cycle. He notes as an example the when Clinton had her tearful moment, “most pollsters had completed their work in the state.”

John Zogby said, “There is artwork involved here.”

More reason than ever to take polls with a grain of salt.

When will Fred begin to attract voters?

UPDATE: OK, the comments here go with the post above on Huckabee. I split this post up to separate the subjects. I thought of trying to copy the comments to that post, but didn’t want to risk it. That’s why they look out of place.

Most of the conservative Cheddarsphere here in southeastern Wisconsin is supporting Fred Thompson. I like Fred. I’ve said so before. I just wonder how long he will be in the race.

Let me back up a bit.

The blogosphere was all atwitter last week over a report that Fred would withdraw from the race if he didn’t do well in Iowa. This was denied by the Thompson camp and indeed was probably a dirty trick by some other campaign.

Or was it? Fred did OK in Iowa, good enough for encouragement among his supporters. South Carolina is the rallying cry.

But how is Fred doing? Actually?

OK, I don’t believe in polls either and in this environment they may quickly change and there’s no guarantee that any primary or caucus will turn out the way the polls are.

But just pretend with me that polls mean something. I mean candidates do. The media do.

According to Professor Franklin’s calculations, Fred is second in SC. Not great but within margin of errors. He’s no higher than third anywhere else. Looking at RealClearPolitics.com averages, Thompson is third in SC and no better than fourth anywhere else.

The numbers have to move sometime for Fred to stay in it. You can’t keep coming in third and win a nomination.

Most Fred heads seem believe that McCain, Romney and Huckabee will knock each other out. Huckabee won’t last and, presumably, McCain and Romney will fade because they’re not “true” conservatives.

Time will tell and it may start Tuesday.

Obama campaign manager: Early polls don’t mean a thing

From an e-mail I received from the Obama for America team (also found here):

In mid-September 2003, national polls showed Joe Lieberman to be the frontrunner for the Democratic presidential nomination. Then John Kerry won the Iowa caucus and the New Hampshire primary, and the shift in momentum carried him to a decisive victory.

The lesson: early polls don’t mean a thing and success in crucial early-state contests will win the Democratic nomination.

I think we’ve also found out that later polls don’t mean much either.

Surprise!

A poll commissioned by a left leaning organization finds that a right leaning, as yet unannounced candidate may be vulnerable (PDF file) in the next election.

Coulda knocked me over with a feather.

Polls, schmolls

A poll conducted by a Democratic group favorable to Healthy Wisconsin discovered wide approval for the health care initiative, while a poll by a group opposed (PDF file) to it found widespread disapproval.

Surprise, surprise.

Now it seems pretty clear that the vast majority of Americans want everyone covered, even if it means higher taxes. The rub might come in how much tax, but that’s another question.

And health care reform is not the question here either. The question is our incessant dependence on polling numbers and the need to conduct dueling polls.

Economy growth is best in a year

From Yahoo! News:

The economy snapped out of a lethargic spell and grew at a 3.4 percent pace in the second quarter, the strongest showing in more than a year. A revival in business spending was a main force behind the energized performance.

Inflation — outside a burst in energy and food prices — moderated.

However, consumers, whose spending largely prevented the economy from stalling out in the first three months of this year, lost energy in the second quarter. They boosted spending at a pace of just 1.3 percent, the smallest since the final quarter of 2005.

Still, a solid jobs climate — the nation’s unemployment rate is at a relatively low 4.5 percent — should help cushion some of these negative forces.

But then comes the amazing fact that,

Even with the steady employment climate and the economic rebound, President Bush continues to shoulder weak public-approval ratings for his economic stewardship. Only 37 percent approve of his performance, close to a record low, according to a recent AP-Ipsos poll.

A wicked good economy and only 37 approve of Bush’s economic performance? Makes me think something else is at work here.

Like polls are a poor barometer of, well, anything.