12/03/2012 10:02 am ET Updated Feb 02, 2013

How Technology Changed Public Opinion Research in 2012

Among the many takeaways from the 2012 election cycle is that public opinion research must evolve with the American electorate. As more Americans use the Internet and mobile devices as part of their daily lives, harnessing these technologies is increasingly among the best ways to get an accurate read on voters' complex attitudes and opinions.

Any discussion of how changing patterns in technology use are affecting opinion research begins with voters' increased reliance on cell-phones.

The need for sufficient samples of cell-phone-only households in surveys has been widely documented. While effective statistical models can compensate for a lack of cell phones in a sample, failure to take into account the appropriate number of cell phone users in a sample can lead to an underrepresentation of young voters and communities of color, most of whom tend to vote for Democrats and made the difference for Obama in key battlegrounds.

Republican pollsters have pointed to a shortage of cell phone interviews to explain some of their inaccurate polls in 2012.

Underscoring this point, PNA polling -- conducted with Grove Insight in nine battleground states over the final three weeks of the campaign -- shows that 54 percent of respondents without landlines were voting for President Obama. Just 36 percent of these cell phone-only voters said they were voting for Mitt Romney.

Setting quotas for the number of cell phone-only households in each tracking poll was one reason why our tracking polls were among the most accurate of the 2012 cycle.

Harnessing new technology, however, is important for reasons beyond ensuring representative samples in polls. By reaching voters where they increasingly spend their time -- on the Internet -- researchers can more easily reach a broad swath of voters, and conduct more accurate qualitative research.

In September of this year, PNA worked with Keating Research to conduct a groundbreaking online survey of undecided presidential voters in Western battleground states. In order to find a representative sample of 503 undecided voters across four states, PNA contacted over 4,500 voters to find out whether they'd made up their minds in the presidential race.

This volume of contacts would not have been possible via traditional telephone surveys without expending a vast amount of time and resources. The ability to survey a statistically-significant sample of this tiny sliver of the electorate allowed for a uniquely in-depth dive into who these voters were, and how they would make their decisions. This research effectively anticipated one of the key dynamics of the final month of the race, in which Romney was exposed as untrustworthy and shifty.

The research also suggested that undecided voters would choose their candidate late, and that -- contrary to conventional wisdom -- these voters wouldn't necessarily break for the challenger. Since election day, pollsters have acknowledged that last-second deciders didn't break for Romney as anticipated.

In addition to allowing us to contact a broader swath of voters, online research gives voters an opportunity to answer open-ended questions at their own pace, in the privacy of their own homes. That's why PNA -- teaming with Benenson Strategy Group -- used individual online interviews in conducting our nationwide research on attitudes women's health issues in 2011.

Using this system, PNA was able to identify 100 voters who had staked out moderate positions on abortion rights, and conduct private in-depth conversations with individuals about these sensitive topics. This approach to capture many personal thoughts and feelings from respondents that they may have otherwise been reluctant to share in a more traditional focus group settings.

These conversations were instrumental in informing our nationwide survey, which showed that views on abortion are far more complex than the simple pro-choice/pro-life dichotomy. While voters harbor complex and often conflicting views on the issue, the research showed wide consensus around the idea that decisions on women's health should be made by a woman, her family and her faith, not government bureaucrats.

This frame was widely used by candidates and advocacy organizations to fight back against extremism on women's health throughout the 2012 cycle.

Tracking the increasingly complex, diverse and dynamic public opinion landscape in years to come will require keeping even closer pace with the changing ways that Americans are using technology. PNA is proud of its work developing in the latest technologies in 2012, and looks forward to continuing this work in future cycles.