The Diminishing Effect of Cellphones

By: TJHalva | Comments [0] | Category: Demographics | 8/5/2008 11:35:21 AM CT

The key demographic of 2008 election is undoubtedly the youth vote; but is the youth vote represented in the current polling figures. It appears the answer may be tied to the inclusion or exclusion of cellphone sampling; younger adults are more likely to be overlooked by landline sampling due to their higher adoption rate of cellular phones. The inclusion or exclusion of cellphones from the sampling model does however vary by pollster. I looked through our polls and picked the four most prominent institutes: SurveyUSA, Rasmussen, Gallup and Quinnipiac University to delve deeper into their polling methodologies.

The results surprised me. Perhaps the most interesting item of note comes directly from the original report on the cellphone effect published by Pew Research. In the first line of the report it states that "Pollsters are continuing to monitor changes in telephone use by the U.S. public, since most surveys are still conducted using only landline telephones." That is just not the case; of the four main pollsters they all appear to incorporate a random digit dialing process that would insure the inclusion of cellphones into the sample. But that's not even the best part, Pew Research goes on to cite Survey Sampling, Incorporated as their data provider; the very same sample provider that SurveyUSA uses. To further the point, I've listed direct quotes from the four pollsters explicitly identifying their use of random phone numbers; this of course is contrary to the Pew report which appears to have neglected the research part.

SurveyUSA:
"We do know that Survey Sampling, Inc., of Fairfield CT, is the largest and most respected provider of 'random sample' to opinion research companies."

Rasmussen:
"Calls are placed to randomly-selected phone numbers through a process that insures appropriate geographic representation."

Gallup:
"Interviews are conducted with respondents on land-line telephones (for respondents with a land-line telephone) and cellular phones (for respondents who are cell-phone only)."

Quinnipiac University:
"Professionally trained students and non-students conduct the interviews using a CATI (Computer Assisted Telephone Interviewing system). For a typical public opinion survey a randomly-selected sample of about 1,000 adults aged 18 and over are interviewed over a 5-6 day period."