Cell-Phone-Only Voters in the 2008 Exit Poll and Implications for Future Noncoverage Bias

Abstract

National pre-election polls again performed well in 2008, although evidence began to emerge of minor but consistent non-coverage bias if samples excluded cell phones. As in 2004, the National Election Pool's national exit poll in 2008 provided an opportunity to reach Election Day voters regardless of telephone status. The exit poll found a sharp increase in cell-only incidence, comparable to trends in government surveys, with cell-only status approaching the norm for voters under age 30. But the growth rate of cell-only adoption and difference in presidential preference were even greater among voters age 30 and up, suggesting that typical post-stratification weighting adjustments for age may be less likely to mitigate non-coverage bias in future landline-only random digit dial surveys.

Spanish exit polls. Sampling error or nonresponse bias?

“…On the other hand, early voting is almost anecdotic in Spain. It represents less than 3% of votes, when for example early voting reached 32.7% in the 2008 US Presidential election (Mokrzycki, Keeter and Kennedy 2009). Nevertheless, even in the event that the distributions of mailing and face voters were significantly different, the possible divergences triggered by this could again be contemplated as a part of nonresponse bias.…”

“…Rather, an exit poll is specifically addressed to the voting population and, what is more, can collect very large samples in a very cost-effective manner. However, in the same way as other surveys, exit polls are still exposed to powerful potential sources of error, such as interviewer effects (Blom, de Leeuw and Hox 2011), sampling design effects (Pavía and García-Cárceles 2012), noncoverage by early voting (Mokrzycki, Keeter and Kennedy 2009), measurement errors (Pavía and Larraz 2012), or nonresponse (Groves et al 2002), which can still seriously compromise their inferences.…”

Making Inferences Using Incidentally Collected Data

“…Given the widespread concerns about traditional data collection techniques such as polling data (Mellon, 2013b;Mokrzycki, Keeter, & Kennedy, 2009;Sturgis et al, 2016), it is doubtful whether we truly have a gold standard for many public opinion measures, but traditional techniques at least have a set of established standards for assessing their likely quality and have unambiguous tests of their accuracy around elections.…”

Exit Poll Sponsorship and Response Intentions

“…In the 2008 presidential election cycle, nearly one-third (32.7 percent) of all votes cast were cast early or by mail before Election Day (Mokrzycki, Keeter, & Kennedy, 2009). These developments have led exit pollsters to supplement traditional precinct-based surveying with surveys conducted via telephone in order to maximize coverage.…”

“…These developments have led exit pollsters to supplement traditional precinct-based surveying with surveys conducted via telephone in order to maximize coverage. In 2008, Edison Research conducted over 15,000 telephone interviews in 18 states across the country to measure early or mail-in voting (Mokrzycki, Keeter, & Kennedy, 2009). …”