Innovative Data Gathering with Emerson College Polling

Avid political watchers have likely seen Emerson College Polling in media outlets far and wide this election season, from local TV affiliates, to The New York Times and Newsweek. They also may know that ECP is considered one of the more reliable polls around by the likes of the Times and ABC News.

What they may not know is how Emerson College Polling gathers its data, why they use different collection methods depending on the geography and demography of those they’re polling, and why that matters.

“Over the last 10 years, we’ve been able to work with all of these different modes of data collection – text-to-web, online, email – and from that, we’ve been able to learn these new modes of data collection, which is essentially the future of survey research,” Emerson College Polling Director Spencer Kimball said in an explainer video last election cycle.

Last week, Kimball appeared on Nexstar Media’s Political Power Ranker to talk about Emerson’s latest polling on two major contests this election cycle: the races for New York City mayor and New Jersey governor.

Kimball was there to talk about the results of the most recent polls, of course. (ECP has New York mayoral candidate Zohran Mamdani with 50 percent of the vote. New Jersey Democrat Mikie Sherrill has just a 1 percentage point lead over Republican Jack Ciattarelli as of October 30, but the margin of error is +/-3%.) But Kimball also explained the differences between how ECP polled for the two races.

In New York City, where not many constituents have landlines, Kimball and team used a two-pronged approach: online polling, consisting of people who opted into the polls, and texting randomly selected people. The yield from texting is much lower than online polling, where respondents opt in, but they netted more than 650 respondents total from the two groups, which could be analyzed in relation to one another.

“We compare those data sets to be able to verify if the online is skewed or if the texting is skewed,” Kimball said.

For the New Jersey race, where residents are much more likely to still have landlines, Emerson College Polling again used texting, but replaced online polling with phone calls. But even the methods themselves can bias the results, Kimball said: Data from landlines will skew right, data from texts will skew left.

“You’ve got to be careful of how that data works,” Kimball said. “We’ve got 20 years of experience looking at this, and that’s why, when we get that data back, it may look that way, but I said, ‘No, it’s biased because of this,’ and that’s how we get to our results.”

To see all of Emerson College Polling’s latest results, including the campaign for California’s Proposition 50, which would allow the state to temporarily change state congressional maps; and the Virginia gubernatorial race, visit emersoncollegepolling.com.