Photo: Scott Eisen (Getty)

On Tuesday, Boston city councilor Ayanna Pressley became the second progressive insurgent to topple an incumbent member of Congress in a Democratic primary this year, trouncing Congressman Mike Capuano by 18 points.

But unlike other stunning left-leaning primary winners like Alexandria Ocasio-Cortez in New York and Florida gubernatorial nominee Andrew Gillum, Pressley had considerable establishment support, getting endorsements from Massachusetts Attorney General Maura Healey and the Boston Globe editorial board. So why didn’t any polls predict that she would win?

The little polling that was done consistently showed Capuano in the lead. One poll for WBUR released in early August showed Capuano with a comfortable 13-point lead, which was virtually unchanged from February. This follows on the heels of Gillum’s win in the Democratic primary for governor last week, even after no poll in that race taken this year, according to RealClearPolitics, showed Gillum with a lead; many showed him in fourth place. In Ocasio-Cortez’s race, internal polling by the Crowley campaign three weeks before the election had the incumbent with a commanding 36-point lead. Ocasio-Cortez ultimately won by 14 points.

While the reputation of polling itself took a considerable hit after the 2016 election, the shock of left wins in primaries against well-moneyed and/or entrenched opponents has only been exacerbated by the conventional Data Clearly Shows thinking that these wins weren’t supposed to happen. But the most recent WBUR Pressley/Capuano poll gives one indication of what the discrepancy might be.

Advertisement

In February, shortly after Pressley announced her run for office, the same outlet noted that the majority of the district’s voters were white, even though the district’s population was minority-majority. In WBUR’s July poll, Pressley’s lead among voters ages 18-49 was eight points; among people of color in this minority-majority district, her lead was 18 points—coincidentally the margin she eventually won by.

Polls failing to predict these historic progressive wins could indicate several shifts. One is that the people who have historically been expected to be the most likely to vote—older white voters—are not showing up in as strong of numbers this year. Another, which seems even more likely given the surge of enthusiasm among Democratic voters this year, is that young people and people of color are voting in stronger numbers.

Still another possibility could simply be that supporters of people like Capuano and Crowley thought their candidates had it in the bag and decided not to bother with voting, although that doesn’t exactly explain Gillum’s situation, which was always expected to be a close primary (although the frontrunners were Gwen Graham and Philip Levine, not Gillum) in an election where the state’s primary turnout record was shattered.

Advertisement

What we’ve seen this summer is that Democratic primary voters are showing a clear preference for candidates who are women and people of color, even more so when the candidate is young and facing off against aging white candidates. And given that yet another candidate just defied predictions that she would lose handily, the Democratic incumbents in the few major primaries we have left have got to be feeling a little more worried today.