Who is going to win?

[SlowwHand]

Because Democrats stutter.

[MikeH]

To be fair, I look back on my posts when I was HC's age and I think that as well.
Although, to be fair again, I always posted with much more humour and style than he does. I was probably more right-wing than he is now, too.

And more modesty.

Ditto. Man, I'm just glad there aren't many people still around from when I started posting here back in 1998.

You were always more compassionate than HC is now though.

[kentonio]

Rasmussen was the most accurate pollster four years ago due to its sophisticated model for likely voters.

Nate Silver disagrees.

Rasmussen’s polls — after a poor debut in 2000 in which they picked the wrong winner in 7 key states in that year’s Presidential race — nevertheless had performed quite strongly in in 2004 and 2006. And they were about average in 2008.

Rasmussen’s polls have come under heavy criticism throughout this election cycle, including from FiveThirtyEight. We have critiqued the firm for its cavalier attitude toward polling convention. Rasmussen, for instance, generally conducts all of its interviews during a single, 4-hour window; speaks with the first person it reaches on the phone rather than using a random selection process; does not call cellphones; does not call back respondents whom it misses initially; and uses a computer script rather than live interviewers to conduct its surveys. These are cost-saving measures which contribute to very low response rates and may lead to biased samples.

http://fivethirtyeight.blogs.nytimes...rmed-strongly/

[Asher]

If they don't call cell phones, that alone torpedoes their accuracy.

A huge portion of younger voters don't have landlines. Vast majority of my friends don't. And we aren't GOP voters.

My dad, who would be a GOP voter, has a landline. Of course.

[-Jrabbit]

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.

[rah]

Ever hear of weighting?

[Wezil]

What weight do you apply to a group of which you have no responses?

edit - I'm not trying to be confrontational, I just don't understand.

[snoopy369]

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.

Yep. Unfortunately unless the US changes the law re: auto dialing, it's unlikely to improve any time soon.

Some of those other issues are almost as bad - first person that answers is really biasing, and the four hour window can be really bad in some areas (in particular where there is lower unemployment or higher retiree numbers).

[snoopy369]

What weight do you apply to a group of which you have no responses?

edit - I'm not trying to be confrontational, I just don't understand.

I don't think anyone will argue that the no-landline group is a demographic in and of itself; instead what happens is you miss younger voters and, increasingly, middle class voters of variou ages. So you would need to weight younger voters up, for example (typical weighting scheme would be age, income, maybe housing type). But you are right in that there might well be features of cell phone only households that weighting would not capture.

[snoopy369]

In solidarity with the east coast, my train is going to be really late, lol... Rah, 10:15 arrival =/

[rah]

Just because a lot of younger people don't have land lines doesn't mean that no younger people don't have them. You find those that do and weight them higher. This is standard in survey research. While it's always best to pull a representative sample, it's not always possible. This is not something new to the industry.

In the old days, all telephone samples were biased towards middle class white people.
Not using random digit dialing biases samples towards poorer people.
All internet surveys bias samples against poorer minorities.
The internet has caused the biggest problem since it's so cheap to use but introduces considerable sample bias. A lot of research has gone into how to best weight internet surveys especially when demographics are not readily available. The era of mail surveys is almost completely over.

[Asher]

Just because a lot of younger people don't have land lines doesn't mean that no younger people don't have them. You find those that do and weight them higher. This is standard in survey research. While it's always best to pull a representative sample, it's not always possible. This is not something new to the industry.

It's patently absurd though, as young people who own landlines are kind of...odd. Old-school even. You might even say conservative.

Think about it. You're using an old-style of communication to survey people for political views. It's implicitly biased towards people who are traditional. Even if you find a couple young people with landlines and weight them higher, they're more likely to be conservative.

[rah]

Yes, it's a challenge to work around, but if you don't consider sample bias then your results will be crap to begin with. That's why I always like to review the methodology of a survey.
And yes, some of the challenges are more difficult to overcome than others. Sometimes alternate survey delivery methods are required. (combine internet with phone) But in the current instant and cheap response times many shortcuts are taken which leads to bad research. It's been very difficult for the industry to adapt and keep up high standards. Results should always be questioned.

[giblets]

If you assume young people like Al B. Sure who probably own landlines are representative of young people who do not then you could very easily be wrong.

[snoopy369]

It's patently absurd though, as young people who own landlines are kind of...odd. Old-school even. You might even say conservative.

Think about it. You're using an old-style of communication to survey people for political views. It's implicitly biased towards people who are traditional. Even if you find a couple young people with landlines and weight them higher, they're more likely to be conservative.

I highly doubt it is thy straightforward. Its not impossible that Rasmussen has done studies to determine the intrinsic effects and does adjust for them; we certainly do look at the effects of cellphone only households and know something of how to adjust our results to adapt. It's not perfect, and not nearly as good as including cellphones, but for most pollsters it is too expensive to regularly poll cell phones.