Pollsters beware: phone cord cutting still en vogue
One trend that continues unabated is the growing number of Americans who don't have home telephones, according to newly released government data.
A report published Tuesday by the Center for Disease Control shows fully 25 percent of adults now have only cellphones. The proportion of "cell-only" adults is up two percentage points from the previous six month period, and is about double what it was just three years ago.
The new numbers are a second, stark reminder this fall that surveys that do not include cellphone samples might be missing the mark. A Pew Research Center report on pre-election polls released in late November demonstrated the "increasing likelihood that public opinion polls conducted only by landline telephone will be biased."
It's more than a theoretical problem: Pew found "support for Republican candidates was significantly higher in samples based only on landlines" than in ones including cellphone interviews. (Pew, The Washington Post and others now routinely include both conventional and cellular telephones in their polls.)
For its part, the CDC also warns that "[t]he potential for bias due to undercoverage remains a real and growing threat to surveys conducted only on landline telephone." While the new government report doesn't detail unavoidable problems from landline-only polls, the cataloged demographic and behavioral differences between adults who do and don't have home telephones are significant.
Consider: 1) More than half - 52 percent - of those aged 25 to 29 are cell-only, compared with a slender 5 percent of those 65 and older; 2) Adults living in poverty (39 percent) or near poverty (33 percent) are much more apt to be cell-only than those above the poverty line (22 percent); 3) A third of all wireless-only adults said they've had five or more drinks in a single day within the year, compared with 19 percent of those in families with home telephones.
Traditionally, these differences are smoothed through by standard post-survey statistical adjustments. For example, younger adults are more likely to both have had a big drinking night out and to be cell only so factoring that in might "fix" the overall numbers. But the CDC's new warning is that these weighting procedures may no longer be enough.
The American Association for Public Opinion Research (AAPOR) has a new report out detailing the current state of research on surveying on cell phones. Here's a link to that report. (Disclosure: I am an AAPOR member, currently serving on the association's executive council.)
%adults in wireless-only households
Now (Jan-June 2010) 24.9% Second half 2009 22.9 First half 2009 21.1 Second half 2008 18.4 First half 2008 16.1 Second half 2007 14.5 First half 2007 12.6
| December 21, 2010; 11:30 AM ET
Categories: Ask the Pollsters, Polls, methodology | Tags: Polls, cellphones, methodology, polling
Save & Share: Previous: Public support for components of health care law
Next: Americans split on how to interpret Constitution
The comments to this entry are closed.