Follow the FreeStater Blog by Email

Wednesday, October 24, 2012

No, Robo-Polls and Landline Only Polls are not Inaccurate

With the race for the White House so close it seems that whichever side is down in the polls seeks some way to dismiss polling accuracy. At present it is Team Obama that is trying mightily to convince folks that the polls are misleading and their typical line of attack either focuses on so-called Robo-Polls where the survey is conducted via computer instead of a live operator or the lack of cell phone only households in a survey sample. Robo-Polls are not permitted to dial cell phone numbers and the argument is this exclusion of cell phone only households introduces a bias to the detriment of the President.

Unfortunately, there's no support for the anti-Robo-Call or cell phone exclusion argument. A paper written for the American Association for Public Opinion Research (AAPOR) by an ad hoc committee formed to study polling accuracy found no problems with automated polls. They addressed both the cell phone issue and the "robo-poll" question and found: "The exclusion of cell phone only (CPO) individuals from the samples did not seem to have an effect" and "The use of either computerized telephone interviewing (CATI) techniques or interactive voice response (IVR) techniques made no difference to the accuracy of estimates."

Mark Blumenthal, the polling guru from Pollster.com has written often on the issue of robo-polls and concludes there is no evidence they are are any more or less accurate than live operator polls: "automated telephone surveys did at least as well as those from conventional live-interviewer surveys in predicting election outcomes."

Then there's the Pew Foundation, in a study they conducted in 2009 they determined "pre-election polls conducted by telephone did very well in forecasting the outcome of the election in 2008. This was true for polls using live interviewers and those conducted with recorded voices. It was true for those based only on landline interviews and those that included cell phones."

So there you have it - there is simply no evidence to support the argument that robo-polls and the exclusion of cell phone only households introduces any bias into polling results.