Who is going to win?

[Asher]

Too expensive? Then they shouldn't be doing their job.

Polls are a joke, let's not make them even moreso.

The Rasmussen polls have always overstated GOP voters this election compared to other polls. The fact that they only do landline polling explains it all.

They could do all of the complicated studies they want, but their fundamentals are flawed. Any idiot could tell them that.

[snoopy369]

Too expensive? Then they shouldn't be doing their job.

Polls are a joke, let's not make them even moreso.

The Rasmussen polls have always overstated GOP voters this election compared to other polls. The fact that they only do landline polling explains it all.

They could do all of the complicated studies they want, but their fundamentals are flawed. Any idiot could tell them that.

I guess the fact that windows 7 doesn't boot in 0.1 seconds means MS should quit doing their jobs also?

The polls aren't perfect, as I have been repeatedly saying; but they are better than nothing. To get better, re you prepared to change the laws to permit RDD cellphone studies with predictive dialing?

[Asher]

I guess the fact that windows 7 doesn't boot in 0.1 seconds means MS should quit doing their jobs also?

Why do you think they spent millions making sure Windows 8 boots faster than Windows 7? Why do you think Windows 7 boots faster than Windows Vista? Why do you think Windows Vista boots faster than Windows XP?

It's expensive, but MS is spending the money to make it happen. It's their business.

Whining that it's too expensive to get the fundamentals of your business right is ridiculous. Those companies deserve to fail. And they probably will if they keep this up.

[snoopy369]

Why do you think they spent millions making sure Windows 8 boots faster than Windows 7? Why do you think Windows 7 boots faster than Windows Vista? Why do you think Windows Vista boots faster than Windows XP?

It's expensive, but MS is spending the money to make it happen. It's their business.

Sure, and polling firms are spending tons to figure out how to poll in the modern age. That doesn't mean they can do it perfectly with limited resources. Cost is a business detail and is relevant, no matter how much you might belittle it. Cell phone studies are (iirc) 4-5x as expensive as landline - and even with that you aren't getting true RDD usually. You have to pay someone to sit there, dial by hand, and wait while it rings - and even when it goes to voicemail.

[Hauldren Collider]

US households with no landline
May 2010 = 1 in 4
Dec 2011 = 1 in 3
Oct 2012 (estimated) over 40%

For an organization wanting a representative sample, the whole "does not call cellphones" thing is pretty laughable.

But they do call cellphones. Like I said, Oerdin Is Always Wrong.

[Asher]

Sure, and polling firms are spending tons to figure out how to poll in the modern age. That doesn't mean they can do it perfectly with limited resources.

I'm not asking them to do it perfectly.

I am asking them to do the basics.

Why not just put a ****ing poll on TMZ.com and run those numbers? It's cheaper!

[Asher]

But they do call cellphones.

Nate Silver says they don't.

What's your source?

[Hauldren Collider]

Hm, I remember them saying it on their website for their poll of registered voters.

[Asher]

All of the pollsters were hilariously wrong in the Alberta provincial election earlier this year. Their oldschool methods just don't work in 2012. They're not accurate.

http://www.calgaryherald.com/news/polls+wrong/6507496/story.html

How did the polls get it so wrong?

Pollsters and pundits were wondering what happened after Monday’s election ended with a strong Progressive Conservative majority, contrary to an array of recent surveys.

Every poll conducted in recent weeks predicted a Wildrose win, although one released Sunday night suggested the PCs were closing the gap in the final 24 hours.

But many pollsters had the Wildrose leading the long-ruling Tories in popular support heading into the vote on Monday, and several suggested a Wildrose majority was at hand.

“We were all wrong,” said pollster Ian Large of Leger Marketing. “We were all equally wrong.”

Bruce Cameron of Return on Insight told a CBC panel Monday evening that every pollster missed the mark, but suggested his firm saw some movement coming back to the Tories in the past week.

Premier-designate Alison Redford, whose Tories were leading in 62 seats at press time, said Monday night she never focused on the polls.

“I’ll say exactly what I said during the campaign, the only poll that matters is election night,” she told reporters after her victory speech.

Former Tory cabinet minister Lyle Oberg, who switched his support over to the Wildrose last year, said the results flowing in “are quite surprising when you’ve seen the polls for the last four weeks.”

“What it shows is the Wildrose vote was very fragile. And those last couple of issues in the last week just flipped the voters over.”

Oberg said he does believe the move occurred because of controversial social issues raised by Wildrose candidates Ron Leech and Allan Hunsperger.

“I think it was a (well) run campaign with the exception of a couple of things and I think that was enough just to flip the voters, I really do.”

He added, “you know, all those polls can’t be completely wrong.”

Large said the contentious comments by the Wildrose candidates were never captured in polls and other surveys conducted heading into the final week.

“We never got the chance to measure the last week with the Hunsperger and Leech comments,” he said, noting both could have painted the Wildrose as too much of a risk in the minds of many voters.

In addition, he suggested there could have been two other key factors at play that made the spate of surveys miss the mark.

“Strategic voting actually materialized,” said Large, pointing to the Liberal collapse as a sign of a move to the Tories to keep the Wildrose out of power.

And Large said there was also the “black hole of the undecided vote,” who seemingly moved toward the Tories come election day.

Several polls showed that up to 20 per cent of people were undecided last week before voting Monday.

[rah]

Just like any other business, they'll do the best they can depending on how much they're paid.
If a client comes up and says we're only going to pay you this much, we have to consider that in the methodology. We'll hint that it might lead to shortcomings in the accuracy, but we're not going to tell them the results will be crap and not worth it. We do want the revenue.
The cheapness of using the internet has really changed the industry. The wide use of cellphone has had similar impact for other reasons.

[-Jrabbit]

Rah, yes weighting is the obvious workaround. But if you don't know the demographics of the unsampled group, how do you weight that? Do you just throw darts and hope?

If there's a well-reasoned model for this, fine. But I haven't seen one, and 40% of the population is, um, significant. If this group skews strongly (in any direction, on any demographic), I would think the pollsters would want to know - and not just for politics, but for consumer research as well.

[kentonio]

Just like any other business, they'll do the best they can depending on how much they're paid.
If a client comes up and says we're only going to pay you this much, we have to consider that in the methodology. We'll hint that it might lead to shortcomings in the accuracy, but we're not going to tell them the results will be crap and not worth it. We do want the revenue.

How much revenue are you likely to get long term if you are a widely recognized polling firm who is consistently polling inaccurately?

[rah]

The demographics of the unsampled groups are known. These can be determined using Census and other data sources. The issue is knowing the specific demographics of those surveyed to correctly apply the weights. If collected as part of the survey, there aren't any problems. As with anything else some models are better than others. A lot of effort is being put into it. Those that do it best will generate more business. But for everyone that spends a lot of effort on it, there with be those that will just take the money and run and not care about the quality of their work.
I'll be the first to admit that it not always perfect, but that doesn't mean it's all crap.

[rah]

How much revenue are you likely to get long term if you are a widely recognized polling firm who is consistently polling inaccurately?

Reputation for the larger companies is quite important, but there will always be low cost providers.
Working for one of the largest research companies in the world, I'd like to think that all our research is of the highest quality, (and most of it is) but I know there are exceptions.

[notyoueither]

Large said the contentious comments by the Wildrose candidates were never captured in polls and other surveys conducted heading into the final week.

“We never got the chance to measure the last week with the Hunsperger and Leech comments,” he said, noting both could have painted the Wildrose as too much of a risk in the minds of many voters.

In addition, he suggested there could have been two other key factors at play that made the spate of surveys miss the mark.

“Strategic voting actually materialized,” said Large, pointing to the Liberal collapse as a sign of a move to the Tories to keep the Wildrose out of power.

And Large said there was also the “black hole of the undecided vote,” who seemingly moved toward the Tories come election day.

Several polls showed that up to 20 per cent of people were undecided last week before voting Monday.

That's exactly what happened. Even Liberal and NDP die-hards were voting PC in close ridings to keep Wild Rose out.