Is Cell Phone Bias Skewing Polls?

[AmericanMan1958]

This is why Liberals doesn't trust polls.



http://politicalwire.com/archives/2010/10/13/is_cell_phone_bias_skewing_polls.html?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+PoliticalWire+%28Political+Wire%29

[bob4460]

1. HELL YES they are

VERY few young people have land lines and I think they poll areas that are heavily REpug

[dragonlady]

2. Young people and minorities

Both are embracing cell phones much more than the older while folks. People who move around a lot find it easer to have a phone that's always with them rather than paying to reconnect a land line every time they move.

As far as the polls, I believe the answers of any young person they do happen to reach are weighted more heavily to achieve the desired proportion of the sample. But the young people who talk on land lines very well may not be representative of their peers, so the results are still questionable.

[BurtWorm]

3. I suspect there may be surprises on election day.

[Scuba]

4. I've seen this while doing phone bank work. Many have had...

... their land-lines disconnected.

[SoCalDem]

7. or they leave them on silent & check messages occasionally

(that's what I do)

[4lbs]

5. It sure is. Any polling outfit that doesn't do the majority of its polling using cell phones isn't

likely to be too accurate.

http://pewresearch.org/pubs/1601/assessing-cell-phone-challenge-in-public-opinion-surveys


According to that article, which references the CDC research document released earlier this year:

23% of US Households don't have a landline phone at all but did have cell phone service.

30% of Latinos/Hispanics use only cell phones.

49% of adults 25-29 use only cell phones.

66% of adults over 50 have landlines.

Thus, any polling that calls landlines exclusively, or for the majority of their data, is much more likely to reach adults 50+ years of age. Adults in their 20s, which often support Democrats and President Obama, are then more likely to be underrepresented.

[SoCalDem]

6. and most of the 23% with landlines would probably be older people

who "may" skew right-wing.

[4lbs]

9. True. As the CDC study showed, 2/3 of adults over 50 had landlines.

People over 50 tend to be Conservative in voting patterns.

[NoPasaran]

13. "May" also skew toward people who actually vote in midterms.

[SoCalDem]

14. double whammy

:(

[BzaDem]

11. But they already weight for age. If they don't get enough young voters on a landline, they will

boost their weighting in the poll to reflect census data. That's why polls have historically been accurate, even in Scott Brown's special election in 2010, the NJ/VA governor elections in 2009, and election 2008.

[4lbs]

12. I'd rather they actually redo their calling patterns to reflect the percentages.

Since 23% of the US population has only cell phones, then 23% of their polling should be from cell phones.

So on and so forth.

[nadinbrzezinski]

8. Yes... Dewey defeats Truman

For the same reasons actually. The tech is ahead of pollsters.

So yes the polls are accurate, insofar as you look at the universe of voters they're reaching.

[BzaDem]

10. That's a pretty dumb reason why not to trust polls. Polls have been incredibly accurate historically

if you aggregate them together. The polling average of Obama in 2008 was 7.6 points, and in reality it was 7.3.

The cell phone effect doesn't necessarily produce a bias. It only produces a bias if the normal demographic weightings pollsters already do are not ENOUGH to overcome any bias in landline only samples. This might swing polls a few points in either direction (as the article says), but not more than that.

[deaniac21]

15. Yep

[madrchsod]

16. i have`t had a land line in 8 yrs.

[Steepler0t]

17. A phone that is stuck to a wall?

Sounds like the charger is stuck or something, how odd.