RealClearPolitics launched its “Polling Accountability Initiative” Friday in an effort to restore voters’ trust in the polling industry.
The news website and polling aggregator, whose polling averages are frequently used by media outlets, launched the project just days before the 2022 midterms in an effort to “increase accountability and restore public trust in political and election polling.” Pollsters are divided into single-state and multi-state pollsters, and ranked by the average error between their polls and the final election results.
“Since pioneering the RCP Poll Average in 2002, the public opinion survey industry has undergone momentous changes and faced a series of challenges in a profoundly altered political landscape,” RealClearPolitics said in a press release announcing the initiative. “Twenty years ago, Facebook and Twitter did not exist. The iPhone did not exist. The percentage of households with landlines has dropped from 90% in 2004 to just 40% today. Over the same period, the number of Americans owning cell phones has skyrocketed to 97%.”
“There are many quality pollsters and media outlets that are doing excellent work in a constantly changing technological and political environment,” RCP continued. “However, there are also pollsters and news organizations that are doing less-than-stellar work and, unfortunately, many of the polls from these organizations receive a disproportionate amount of attention.”
“RCP’s goal here is simple,” the outlet concluded. “Accuracy is the foundational bedrock of public trust. To that end we will be evaluating pollsters almost exclusively on one metric – accuracy in reflecting the actual results… Our hope is that by bringing attention to the most accurate polling firms and organizations, as well as shining a light on firms releasing less accurate surveys, we will help bring accountability to political polling that has been lacking in recent election cycles.”
RCP released its initial rankings Friday; the rankings were broken down by single and multi-state or national pollsters, as well by the 2016 election and 2018 midterms, and the 2020 election. The best national pollster in 2016/2018 was Selzer & Co., with an average error of just 2 points; the best single-state pollster was New Hampshire’s St. Anselm College, with an average error of just 0.8 points. The worst national pollster was Quinnipiac University, with an average error of 5.7 points; the worst single-state pollster was Franklin & Marshall College, with a massive average error of 10 points.
In 2020, the best national pollster was the Trafalgar Group, with a 2.5-point average error; the worst was Monmouth University, with a 7.6-point error. The best state pollster was Stockton University, with a 0.5-point error; the worst was Detroit News/Glengariff, with a 5.8-point error.
One independent pollster opined on the state of his industry earlier this week. “Media, they used to be a safeguard against the misuse of polling as a weapon of information war, and unfortunately now I think it’s very clear that big media has become one of those abusers, and they’re the repeat offenders of it,” Big Data Poll director Richard Baris said in an interview with the Epoch Times. “And the most [supposedly] credible news outlets out there are the very ones that are doing this time and time again.”
“Many [pollsters] no longer even understand the people that they’re trying to learn about, trying to gauge,” he continued. “And polling is essentially attempting to predict human behavior… How could you do that if you don’t know that much about the subject, and maybe even dislike them? And that’s the truth, most people in my field dislike them.”
“I think we have really really serious problems, ethically and methodologically,” Baris added, later praising RCP’s initiative to increase accountability.