? re polling folks with no land line

[Shanti Mama]

It is my understanding that one of our largest, best supporter groups -- the young -- are often misrepresented in polls because they don't have land lines and polling companies call from land line lists. Does anyone know if this is still the case?
Thanks.

[OmahaBlueDog]

1. I asked a question along this lines a couple of weeks back

The conventional wisdom is that the analysts have to weight results to account for the skew brought on by the aging demographic with land lines. It is a good question, though

[Happyhippychick]

2. Yes it is true. And newly registered voters are not polled either and the Dems are ahead by

millions in that department.

[supernova]

3. Yep, true

They poll what that term "likely voters." These people have a history of voting, so are likely to vote this time too.

Most polls do not include young people or others voting for the first time. One of those reasons is that a lot of young people only have cell phones now.

[Shanti Mama]

4. GREAT!

for those worrying about the polls, they need to be reminded of this.

[protect our future]

10. Another reason is that newly registered voters are not included

in the databases the pollsters use.

[skids]

5. Was a problem in 2004. Not a problem now...

...they've run some surveys into the cord-cutter demographic to test the waters, and are able to factor in the effect. (Which was that cord-cutters lean about 2% more towards democrats.)

The problem they have this year is vastly underestimating what constitutes a likely voter, because the pollsters have never seen a GOTV effort like what is happening this year.

[MadBadger]

6. Yes, but Pollsters weight their polls to make up for this fact.

Its not perfect, but good enough.

[Captain Hilts]

7. It just takes more time. Pollsters have ALWAYS had a hard time getting

For example, many pollsters ask this question first: "May I speak to the youngest male registered to vote at this address?" because young men have always been difficult to get on the phone. My firm would just ask for the "youngest voter registered to vote at this address."

So, it's always been difficult to get this demographic on the phone. It's nothing new. They've also had lousy turnout. In 2004, something like 17% of those under 30 eligible to vote, voted. Pathetic. I think it will be higher this time, but I'm wondering how many young folks registered to vote, went to college and are now finding out that absentee voting is a total pain in the ass.

So, it's always been difficult to get young people on the phone and pollsters just need to devote more time to get them and then balance the data accordingly.

[Captain Hilts]

8. Duplicate. Here's a nice photo instead...

[qanda]

9. It's not just the young getting rid of their landlines

I know a lot of people who have a cell phone as their primary phone and they are not young by any stretch of the imagination. For a lot of people, it just makes more sense.

[Captain Hilts]

11. The broader the population that relies on cellular, the easier it is to statistically correct for it

[Clio the Leo]

12. I call this age group for a living and I know this is the case ....

... RARELY do I call one of my customers (I work in a bridal boutique) on a land line. If I'm callin a land line, their mama is answering.

[Tansy_Gold]

13. NO IT IS NOT STILL THE CASE

1. It is more expensive for polling companies to call cell phones, but they do occasionally do it.

2. Polling is based on a pre-determined sample, usually based on prior voting patterns. IT IS TRUE that current polls MAY NOT account for the advantage Dems have in new registrations. But that has NOTHING to do with cell phones.

3. As pollsters get more historical information on voting patterns, cell phone vs. landline ownership, etc., they will be able to refine their sampling demographic accordingly. With insufficient data on which to base their sample, they can't place any reliability on the results of such a poll.

4. Sample demographics are usually based on "likely" voters, which can be either/or respondents who self-identify as "likely" and respondents who claim to have voted in one or more previous elections. A voter is "more likely" to vote in an upcoming election if they have voted in the past. Newly registered voters with no voting history have little reliability in polls.

It may indeed be true that the Dems, with a huge advantage in newly registered, first-time voters will deliver a resounding win to Obama. But the pollsters, in trying to be as scientific and reliable as they can be in a very inexact science, want to be as cautious as possible and rely only on "more likely" respondents.


Tansy Gold