(Photo courtesy of UML Center for Public Opinion)
Sabine Laurent de Cannon
Connector Editor
You could hear a pin drop in the New Hampshire field office as staffers and campaign workers poured over the new Iowa polling results from the Des Moines Registrar. Then, the silence was broken when a staffer exclaimed, “She’s ahead by three percentage points!” while excitedly displaying Vice President Harris’s 47% lead over former President Trump’s 44%. “We’re going to win Iowa,” another worker declared emphatically. Behind the surprising news was J. Ann Selzer, one of the industry’s most reputable and trustworthy pollsters, whose firm MSNBC has called the “gold standard.” Seltzer’s numbers, though within the margin of error, were like a five-course meal to the starving NH Democratic team.
Polls often have this power: the ability to dictate confidence, change moods, and guide the news cycle. However, the accuracy of any given poll frequently comes under scrutiny, especially when the numbers are surprising, as with the latest Iowa results, or when certain groups feel disappointed or discouraged. To mitigate these uncertainties, FiveThirtyEight—a website dedicated to analyzing opinion polls, politics, and the economy—debuted a ranking system for pollsters in Jan. 2024. These pollster ratings are based on each firm’s track record for accuracy and methodological transparency. And, as of Sept. 12th, UMass Lowell’s Center for Public Opinion places ninth out of more than 282 firms—two spots above J. Ann’s Selzer & Co.
Elliot Morris, FiveThirtyEight’s editorial director of data and analytics, explained to ABC News how the site determines its top-rated pollsters. Morris claims that FiveThirtyEight believes a polling organization is strong if it meets two main criteria: accuracy and transparency. First, is the pollster empirically accurate? Has it performed well in past elections, closely matching the actual results? And, does the pollster show bias relative to others? If FiveThirtyEight finds that a poll appears accurate primarily due to bias while others were simply unlucky, it is taken into account. Additionally, accuracy points are deducted if a pollster simply copies other results. Concerning transparency, Morris said, “We want to know if you are showing us your work, just like your eighth-grade math test…I want to know the questions you asked and the sentiments of the different population groups. We add all that up together and if you’re accurate historically and you’re showing me your work, you’re a pretty darn good pollster.”
UMass Lowell’s ninth-place rank includes a 2.9 out of 3-star rating and a transparency score of 8.7 out of 10. FiveThirtyEight analyzed 31 polls conducted by the Center for Public Opinion to determine the university’s placement. Suffolk University and Emerson College also appeared on the Top-15 list.
According to its website, UMass Lowell’s Center for Public Opinion exists to provide students, faculty, and visiting scholars with easy access to public opinion research. The center is a member of the American Association for Public Opinion Research (AAPOR) Transparency Initiative, which aims to help survey organizations share their research methods more openly and easily. Through the initiative, AAPOR encourages pollsters to commit to being transparent in how they conduct surveys and report their findings, a commitment that UMass Lowell appears to have embraced.
The center’s opinion polls are available on its website. They feature recent presidential polling from crucial battleground states like Michigan, North Carolina, and Pennsylvania, as well as past data on primary elections, mid-terms, and gubernatorial races.
