Leading pollsters’ discussions at a recent conference augur the most significant shift in American polling in decades.
Published papers and discussions at the American Association for Public Opinion Research’s annual meeting foreshadowed the most significant shift in election polling since Americans began abandoning landlines in favor of cell phones.
NCSES is participating in the 77th Annual Conference of the American Association for Public Opinion Research! We’re proud to have our experts share innovative survey research and data. Click here to learn more about the event: https://t.co/oCc65MMxWj @AAPOR #AAPOR #StatsAreCool pic.twitter.com/WDT9zDdn61
— National Center for Science & Engineering Stats (@NCSESgov) May 12, 2022
As the polling industry attempts to recover from what veteran Pew Research Center pollster Scott Keeter termed a “disaster for our field” in the 2020 election, a few of these new methods are infiltrating the mainstream.
Major media surveys now reach some participants via SMS or mail solicitations, substituting or complementing the phone calls that ruled polling for decades. However, potential respondents are now less likely than ever to respond.
To contact the most difficult-to-reach Americans, pollsters more and more favor combining multiple contact methods in a single survey.
David Kanevsky, a Republican pollster who attended the event, stated, “The future will be multimodal.”
It is already occurring everywhere.
After ending its more than 30-year collaboration with NBC News after the 2020 election, the Wall Street Journal’s new poll reaches a quarter of its survey participants via text message.
The poll is a joint effort between Trump and Biden’s pollsters.
Text-to-web is a technique that sends a link to an online survey; these interviews are added to those conducted via voice-over landlines and cell phones.
National surveys of the 2020 presidential contest were the least accurate in 40 years, while the state polls were the worst in at least two decades, according to the new, comprehensive report from the American Association for Public Opinion Research. https://t.co/K1vaDWQ3eF
— Alexandra Jaffe (@ajjaffe) July 19, 2021
In the past year, CNN’s polling has utilized a variety of methodologies.
Some polls are conducted by respondents who decided to join a panel maintained by SSRS, a Pennsylvania-based firm that conducts polls for CNN and other media outlets. These polls are conducted over the phone and on the internet.
In contrast, the samples for two other CNN polls were obtained by mailing invitations to people’s homes and asking them to participate via phone or the internet.
Prior to last summer, CNN conducted all of its national polling via telephone.
At the moment, the CBS News/New York Times poll using web pollster YouGov was contentious.
There is no longer any such hesitation regarding new methods. The AAPOR conference featured panel discussions with headlines like “The Horse Race: We Can’t Tell Who Wins, But We Know Pollsters Lose” and “Innovations in Election Polling.”
This is primarily due to the realization that as Americans become harder to reach, conventional means are becoming untenable.
One pollster, who requested anonymity to assess the industry’s current state candidly, stated, “These tasks are only becoming more difficult. Therefore, if you continue doing the same thing, it will only worsen.
The innovations extend beyond data collection and sampling. However, developing new weighting parameters—adjusting the results to more accurately reflect the electorate—is more challenging.
Nonresponse bias appears to be one of the leading causes of the 2020 election’s miss. This group was more likely to support Trump, making it more difficult for polls to measure his support accurately.
Some pollsters were not weighted by educational attainment prior to the 2016 election, which was one of AAPOR’s primary recommendations. However, almost every pollster addressed this before the 2020 election, and Trump was still underestimated.
Most of the research now focuses on determining which voters pollsters are overlooking. It is insufficient to simply “weight up” Republicans because, for example, those who speak with pollsters are less willing to back Trump.