How to poll with diminishing home phones?

[Emp]

Ok, I've recently heard the premise that cel phone only households and individuals are being under represented in the polls (presumably understating the positions of younger and/or more tech savvy voters), and don't know what to think about that.

But how DO the pollsters conduct reliable polls as we gradually abandon home phone lines?

[Snipe]

I never answer polls anymore. If you do they keep your number and keep calling back. It is just easier on them. So you have a certain demographic that likes polls and participates a lot. You then have another demographic that doesn't answer the phone or has caller ID. As you pointed out they also can't poll people who have abandoned phone lines.

I think that young people with just cell phones are more likely to vote for Obama. They are also more likely not to vote.

I like polls just before election day that get it wrong. I like it when the result is off by a margin over the stated "margin of error". Pollsters really think a lot of themselves when they get to define their own margin of error.

[Kahns Krazy]

Pollsters really think a lot of themselves when they get to define their own margin of error.

Polls are stupid and pollsters are stupider, but they do not define their own margin of error. It is a computation based on sample size and standard deviations.

[Snipe]

I was just taking a shot at pollsters. I do understand that it is a computation. I remember statistics class. But that is not the way it is explained to the public. They act like it is the real margin for error which simply isn't true. Margin of errors and confidence intervals are based on standard deviations and as such they have a statistical chance of being completely wrong. Even a confidence interval of 95% is going to give you statistically inaccurate information 1 out of every twenty times. And that doesn't take into account the problems of sample composition, who is asked (& who is not asked as the Emperor points out), wording of questions and a host of other factors.

This is never explained to the general public. I doubt that most of the people in the press really understand it enough to explain it anyway.

[drudy23]

I dont even know why they quote the "margin of error" on TV shows because, you're right, most people don't have a clue what it means anyway.

Basically the "cover your ass" metric.

[Kahns Krazy]

And that doesn't take into account the problems of sample composition

Clearly, this is the biggest problem. It's one thing when you are taking a sample of thicknesses of metal coming off a production line, but quite another when you are talking about subjective preferences. Not to mention the human factor, i.e. people that lie in polls.

[drudy23]

Clearly, this is the biggest problem. It's one thing when you are taking a sample of thicknesses of metal coming off a production line, but quite another when you are talking about subjective preferences. Not to mention the human factor, i.e. people that lie in polls.

Or surveyors that aren't the best at "reading"

[Emp]

Yet we poll the death out of everything. There is a thread here dealing with a poll. We cite polls on approval ratings, persons likely to vote. The Board itself does polling, albeit silly in most cases.

Everyone cites polls as if they are statistics. Dismissing polls outright is wishful thinking. The assertion that young people with cell phones are less likely to vote implicitly flows from some poll or other.

I'm still interested in how pollsters will compensate for the loss of home phones and the lack of access to persons with cel phones only.

[MADXSTER]

Yes, but some polls I don't mind.

http://www.topnews.in/files/pole-dancing1.jpg

[Snipe]

Everyone cites polls as if they are statistics. Dismissing polls outright is wishful thinking. The assertion that young people with cell phones are less likely to vote implicitly flows from some poll or other.



Young people are less likely to vote period. That doesn't come from a poll it comes from historical statistics of voter turnout.

And I wish my wife would take a pole dancing class. That would be awesome.

[Kahns Krazy]

Your wife dances on my pole with plenty of class.

[Fred Garvin]

Young people and Snipe don't vote. I know Snipe doesn't vote because he is afraid of jury duty. What is the excuse for young people?

[Emp]

Young people are less likely to vote period. That doesn't come from a poll it comes from historical statistics of voter turnout.
.

I'm serious, how would you find that statistic for say the last Hamilton County presidential election in 2004? I dont recall giving my age when I voted.

Registration, I can see that. But of registered voters who do vote, how do you get demographics. Is my vote tallied under my age, sex, race, party? Or is it exit poll sampling?

[Snipe]

I'm serious, how would you find that statistic for say the last Hamilton County presidential election in 2004? I dont recall giving my age when I voted.

Registration, I can see that. But of registered voters who do vote, how do you get demographics. Is my vote tallied under my age, sex, race, party? Or is it exit poll sampling?

Check out these spreadsheets (http://www.census.gov/population/www/socdemo/voting/cps2006.html) from this study (http://www.census.gov/prod/2008pubs/p20-557.pdf).

The data comes from the census.

[Emp]

Check out these spreadsheets (http://www.census.gov/population/www/socdemo/voting/cps2006.html) from this study (http://www.census.gov/prod/2008pubs/p20-557.pdf).

The data comes from the census.

and as the link points out, based on telephone polls, not on actual voting records, so it is not voting data at all, but polling data.

One might posit that the older and more educated you are, the less likely you are to admit that you aren't registered or didn't vote. Conversely, the younger you are, the less likely you have a home phone to be surveyed, or care whether someone knows you voted or not.

Which brings us back to the original premise, how does the Census Bureau or anyone deal with this changing demographic?