The culprit of the 2012 US election polling inaccuracies, according to two professors in the department of statistics at Oklahoma State University (OSU), was cellphones. Pollsters need to recommit to stricter methodologies or risk further poll inaccuracies in the future.
Prior to 2012, state election polling tended to be accurate, but for the 2012 presidential cycle public polls had Mitt Romney as the winner of many battleground states. A post-election analysis noted that public opinion pollsters who only called landlines performed poorly, generally showing results that skewed more Republican than Democratic. It turns out that cellphone-only households don’t poll the same as the landlines, which leads to more bias in general.
In the 2012 election cycle, 40% of all US households were cellphone-only (CPO Households) up from 8% in 2006. Those households tend to lean more towards Democratic support (let your imagination determine the reasons), and so their underrepresentation is greatly skewing the results. It turns out pollsters are partially ignoring a defined portion of the population who vote decidedly differently from the segment who are being polled. And if we keep following this increasingly-invalid sampling methodology, polls will become even more skewed in the future – in just 3 years, it is estimated that CPO households are at nearly 60%.
Of course as political pollsters try harder to reach CPO households, FCC restrictions and the increased costs associated with stricter cellphone-calling are making it harder to follow standardized telephone survey methodology by equally weighting CPO and landline households. The OSU professors recommend that sampling methodology is adjusted “to reflect the new realities and account for a segment of the CPO voting population that tends to vote for Democratic candidates.”
Full regulations surrounding the new ruling by the FCC on TCPA changes has been released and we are working to provide a summary of the meaning soon. Keep an eye on the blog for what it means and how it will affect telephone surveying of CPO households.