National Media Wakes Up To The Polling Problem

As I predicted, the polling error issue is becoming a big issue in 2008. Amazingly, the pollsters blew the Democratic vote for New Hampshire.

Candidate	Predicted % Of Vote (Rasmussen)	% of NH Vote	Within Margin of Error (95% confidence)?	% Difference
Barack Obama	37%	36%	Yes	-2.7%
Hillary Clinton	30%	39%	No	30.0%
John Edwards	19%	17%	Yes	-10.5%
Bill Richardson	8%	5%	No	-37.5%
Dennis Kucinich	3%	1%	Yes	-66.7%
Mike Gravel	0%	0%	Yes	0.0%

Polling Results Vs. Real Results For The First Primary

Look at the predictions from Rasmussen (a highly respected pollster), versus the actual results. The margin of sampling error for this poll was +/-4 percentage points with a 95% level of confidence.

Candidate	Rasmussen Poll 12/17/2007 (%)	Actual Result (%)	% Difference
Mike Huckabee	28	34	21.43
Mitt Romney	27	25	-7.41
John McCain	14	13	-7.14
Rudy Giuliani	8	3	-62.50
Fred Thompson	8	13	62.50
Ron Paul	6	10	66.67
Tom Tancredo	2	0	-100.00
Duncan Hunter	1	0	-100.00

Giuliani's, Huckabee's, and Thompson's results all fell outside of the margin of error. Paul was on the very edge of it.

I Hear Whispers of The Polling Crisis

A story in the Associated Press today says that 2007 is likely to be the first year where Americans spent more money on cell phone bills than land line phones. A quote from the article:

"What we're finding is there's a huge move of people giving up their land line service altogether and using cell phones exclusively," said Allyn Hall, consumer research director for market research firm In-Stat.

Yup. I'm one of those people. I know many others. And no one with just a cell phone is being polled. As I mentioned earlier in December, telephone polling, which is pretty much the basis of all the polls you hear or read about, is seriously broken. How will it be fixed? Without a change in federal law allowing pollsters to call cell phones, it's uncertain. What is certain, there will be a shock in 2008 when the mainstream media realizes their polling tools are worthless. To be sure, the biggest gainer in this debacle will be Ron Paul. Whether or not the error is significant enough to win the presidency, I'm not willing to say. But I will say that it will be big enough to make the pollsters worry about their job security.

I can't wait until those "same 10 Ron Paul spammers (i.e. supporters)" spam the voting booths.

Prediction: 2008 Will Be The Year Of The Pollster Crisis

Finally, a prediction of my own. It is simply this: In the next year, pollsters will realize that telephone polls are no longer predictive of reality. What evidence do I base my prediction on? Well, other than my 'gut instinct', I'm basing my prediction on the few items I briefly outlined in this post. Mainly that:

• Telephone polls rely solely on traditional 'wired' telephones. They do not call cellphones or VOIP phones (for various reason, some of which are legal restrictions). Nearly 20% of all Europeans have just a cellphone, and I think the US is not far behind that figure.
• Caller ID is in such widespread use that more than 50% had it in 2004. Today, I'm sure that a much larger percentage has Caller ID, and nearly all cellphone only households have it.

Both of these facts indicate that a large number of people are not being called at all (cellphone-only users) and those people are likely to be younger (as most early adopters of technology are). Another large segment of people are not answering calls from a telephone number they don't recognize, primarily younger people who understand how to use Caller ID to avoid solicitation calls. Which of course begs the question, who does answer the telephone when they see an unknown number or "Number Not Available" on their Caller ID?

These two effects will combine to grossly underestimate the number of young people (15-35) who are being polled.

In the past, the effect of excluding that group might have been muted by the fact that their opinions were largely represented by the public at large. However, the more age-specific the poll issue is, the larger the possibility for a huge error. I think 2008 election is the perfect set up for a 'young vs. old' polling issue to finally bring this error to everyone's attention. We'll see.

The Republican Party: Invasion Of The Body Snatchers?

I've watched most of the Republican debates over the last few weeks and I am just about as confused as the RNC must be right now. Which of these guys is actually a Republican? Or more accurately, "How did all these Democratic candidates get into the Republican debates?" A story in the Politico suggests that Sen. John McCain and Rudy Giuliani are the best ticket for the Repubs. Wow. That would be a "good" ticket?. First of all, I am still somewhat astounded that a RINO like Rudy Giuliani is the leader of the pack, considering that he is pro-abortion (favors taxpayer funded abortions no less), is for same-sex unions, and is anti-gun. Overall, he's looks to be somewhat to the left of Hilary Clinton. Am I missing something here? Which "Republicans" are the pollsters calling anyway?

Will Iraq be the big issue in 2008? Survey says:....Yes...well, maybe

A recent Rasmussen poll says that 68% of Americans want to be out of Iraq within a year. Yet all the front runners, both Democrat and Republican, want to be more into Iraq. Hillary and Giuliani are the exact same type of war hawks as Bush, with both of them mirroring his current policies with regard to Iran in their campaign rhetoric. Only three candidates, Ron Paul, Mike Gravel and Dennis Kucinich are suggesting that the whole thing was a huge mistake. And all of them are polling in the single digits.

And how is Giuliani polling so high anyway? The whole thing begs the question that was brought up in the 2004 election. How big an effect are cell phone-only households having on telephone polling. In 2004 the alarm was sounded over the rise of cell phone only households. This caused great consternation for the pollsters as federal law prohibits automated dialing of cell phone numbers (used almost exclusively in telephone polls). In 2004, the error introduced by these people was non-existent: Polls matched actual results. However back then, cell only households made up just 3% of all households. Today 13% of all adults are using only a cell phone or VOIP. More importantly, 33% of 18-29 year olds only have a cell phone. None of these people are being polled. That's a huge concern, specifically for pollsters trying to gauge issues that revolve around a particular age group (elections anyone?). Those opinions might now show up in a telephone poll, but may at the ballot box. In June 2007, the Pew Research Center wrote:

The picture is not entirely positive, however. While the cell-only problem is currently not biasing polls based on the entire population, it may very well be damaging estimates for certain subgroups in which the use of only a cell phone is more common. This concern is particularly relevant for young adults. According to the most recent government estimate, more than 25% of those under age 30 use only a cell phone.

This might help to explain Ron Paul's huge internet support, near total dominance in Republican straw polls, but weak telephone polling. Interestingly enough, Paul has also done very well on text message polls, which would favor a younger more technically savvy audience.