Impact of cell-onlys starts to show on presidential polls

US-- The latest survey on the impact of ‘cell-onlys' on polling has for the first time found that voter preferences were “modestly affected” by whether cellphones were included in the sample.

Just six months ago the Pew Research Center said the inclusion of cell-onlys was still not affecting poll results – but it now says differences are starting to be seen.

Support for Democratic presidential candidate Barack Obama among registered voters was two points higher in the sample that included cellphones (48% compared to 46%), while support for Republican rival John McCain was one point lower (40% compared to 41%).

However, Scott Keeter, the Center's director of survey research, said the difference still barely exceeds typical margins of error, and in many other examples there is still no noticeable difference between results from the combined and landline-only samples. Keeter said it was too early to start drawing conclusions from the results, and that the Center will continue to track the effect of cell-onlys in upcoming polls. “We're going to keep an eye on this,” he said.

The Center estimates that 37% of US adults are now cell-only or ‘cell-mostly' – living in homes with landlines, but rarely using them.

Phone surveys in the US are typically based on lists of landlines, and including cellphones can be costly and problematic. Respondents usually incur a cost for taking a call, and many consider their cellphone to be for private use only. The law also bans the use of autodiallers for cellphones, meaning interviewers must enter every number by hand.

Although cell-onlys and cell-mostlys can be seen to be different from the general population in a number of ways, Pew said this is largely due to demographic rather than attitudinal factors, and survey results have until now not been affected once samples were combined and weighted to reflect the population as a whole.

Pew said that despite these latest findings, “the more serious challenge to survey research posed by cellphones is the declining absolute numbers of certain types of respondents, most notably the young”. Although weighting can help to correct this, Pew said “pollsters are limited in their ability to analyse differences within this age group”.

Author: Robert Bain

Related links:

Industry must make cellphone research ‘more acceptable'

Presidential polls unaffected by cellphone issue, says Pew

Gallup adds cell-only homes to poll sample

26m Americans now cellphone only