Cell-Phone-Only Crowd May Alter Polling (then discounts the evidence)

[UpInArms]

http://news.yahoo.com/news?tmpl=story&cid=519&e=5&u=/ap/20060516/ap_on_re_us/polling_predicament

WASHINGTON - Justin Globus is part of a fast-growing group — approaching one in 10 Americans — who have given up traditional telephones and depend only on their cell phones. That trend is making pollsters uneasy.

For Globus, a 25-year-old salesman from New York, "It was a fiscal decision — a matter of chopping down to one bill."

But the rapid growth of the cell-only crowd isn't so simple for pollsters. Their survey research depends on contacting random samples of households with landline phones. They worry that if the trend continues they could miss a significant number of people and that could undermine their ability to accurately measure public opinion. There could be implications for politics, government policy, academia, business and journalism.

So far, the differences aren't so great and the cell-only group isn't large enough to affect survey accuracy, according to an AP-AOL-Pew study, one of the most extensive news surveys of cell phone users yet.

<snip>

For example, 51 percent of the cell-phone-only group believes gay marriage should be allowed, while 37 percent of a standard polling sample felt that way. And 53 percent of the cell phone only sample said they would vote for the Democratic Party's candidate, while 47 percent of a standard sample said they would vote Democratic, the poll found. Those differences virtually disappeared once the cell-phone-only sample was blended into the total.

...more...

Why this stupid stenographer - oops! I meant typist...no. no. I think they prefer to be called reporters :eyes: even bothered to submit this self-contradictory piece of crap is beyond me - but the spin was so thick, I just had to post it as an example of what passes for snews.

[bluestateguy]

1. I'm part of that demographic

Although if pollsters use random digit dialing, wouldn't that fix the problem?

[slackmaster]

2. Random dialing excludes the possibility of stratifying your sample

It's one approach to achieving a representative sample, but sacrifices the ability to excercise appropriate controls over the set.

[snooper2]

3. the number of business lines..

Far outnumbers that of individual phone lines....

[slackmaster]

4. Fax machines have no opinions

;-)

[snooper2]

13. or modems :)

lol

[pimpbot]

5. I think its illegal to blind call cell phones

After all it will be using your plan minutes. Maybe they should go back to door to door polling :)

I am part of that cell phone only demographic and I sure as hell dont want anyone cold calling me. Leave my cell phone number private. Its been so nice not having any telemarketers the last 5 years.

[jrw14125]

10. ditto that

Yeah, keep polling 70-year-olds that get their news from the local paper and cable. what an informed demographic that is.

[Chovexani]

6. I've just become part of that demographic

I just moved in with two people, with at least one more housemate coming in the summer, and more than one land line just isn't economically feasible for us.

I find this is a common situation for people around my age (20-something crowd).

[Sunny_Sunshine]

12. I'm 50 and no land line

When the last kid left for college, the land line didn't make sense, cable modem and everyone has a mobil. Works for us but we aren't 20 somethings (although my children are now)

[Mythsaje]

7. My wife and I

are also part of that demographic. Who needs a land line?

[slackmaster]

I have a land line but live alone and am rarely at home

Nobody ever calls me since I put it on the No Call List.

:argh:

[slackmaster]

9. I have a land line but live alone and am rarely at home

Nobody ever calls me since I put it on the No Call List.

:argh:

[katty]

8. have2 landlines-one for fax, almost never call from them

[sarcasmo]

11. I am part of the one in ten and have been for two years now.

[athena]

14. There is no contradiction.

10% of the population has only cellphones. 51% of those are for gay marriage
90% of the population has a home phone. 37% of those are for gay marriage.

If, instead of only polling the 90% with home phones, you were able to include the 10% with cell phones, you would get (0.10 x 0.51) + (0.90 x 0.37) = .384 = 38% in favor of gay marriage.

Similarly, for the percentage of democrats, the calculation is (0.10 x 0.53) + (0.90 x 0.47) = 0.476 = 48%.

So there is a ~1% bias in the polls. Most polls have a stated error of 3%, so the pollsters can claim the effect is not significant. It is, however, a systematic, signed error, so one shouldn't be so casual about ignoring it. One other concern should be whether people who have a cellphone in addition to a home phone are less likely to be reached at their home phone number than those with only home phones. If the former crowd is more liberal, this would introduce yet another bias.

[high density]

15. I'm cell-phone-only as of May 1

Spending on money for landline service doesn't make sense for me given the amount of calls I make and receive. Tracfone was the right choice for me.

[Raine]

16. I have both, yesterday

when I got home I found a message on my land phone from the local ABC news saying they would like me to participate in their poll and would have the results on the evening news. Course I didn't see the message till way passed the time, don't even know what the poll was for. :-( I think telephone answering machines (for land lines) had an effect on polls anyway, I don't think there has been accurate polls for ages if ever.