Cell phones a problem for pollsters

[question everything]

Cell phones a problem for pollsters

More mobile-only users could undermine surveys

By Will Lester
ASSOCIATED PRESS

May 16, 2006

(snip)

But the rapid growth of the cell-only crowd isn't so simple for pollsters. Their surveys depend on contacting random samples of households with landlines. They worry that if the trend continues, they could miss a significant number of people, undermining 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. Currently, 7.8 percent of adults live in households that have only a cell phone, according to research released Friday by the Centers for Disease Control and Prevention's National Center for Health Statistics. And those who only have cell phones are significantly different in many ways – typically younger, less affluent, more likely to be single, and more liberal on many political issues – from those who can be reached by landline, the AP-AOL-Pew survey finds.

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 for the Democrat, the poll found. Those differences virtually disappeared once the cell-phone-only sample was blended into the total. But the cell-phone-only group is growing at about 1 percentage point every six months.

(snip)

Landline pollsters hope to minimize the effect of missing cell-only people such as Globus through industry-standard statistical adjustments. Results from a random sample typically are weighted to known demographic parameters of the population, such as age, race and sex data collected by the Census Bureau. Since younger people are harder to reach in landline phone surveys, responses from that age group tend to get weighted up. Blending the opinions of cell-only respondents into a poll of all adults changed the findings on various political questions in the AP-AOL-Pew survey by 1 percentage point at the most.

(snip)


Find this article at:
http://www.signonsandiego.com/uniontrib/20060516/news_1n16polling.html

[napi21]

1. They've been crying about this since before the 2000 election!

I'm surprised they haven't figured out a way to solve it by now.