The outcome of the Democratic primary in New York's 14th House District last month made headlines worldwide. The upset of a 20-year incumbent renewed debate over the validity of political polling .
People took notice when 28-year-old Alexandria Ocasio-Cortez bumped off 10-term incumbent Congressman Joe Crowley 58 to 42 percent. Crowley was the No. 4 Democrat in the House. Ocasio-Cortez ran a modestly financed campaign as a progressive socialist Democrat with an agenda that includes racial diversity and Medicare for all. Crowley, the party favorite, was an assumed shoo-in who was perhaps in line to succeed Nancy Pelosi.
Jefrey Pollock is Founding Partner and President of Global Strategy Group, which released a poll three weeks before the primary that had Crowley ahead by 35 points. "None of us know the answer as to what happened. We'll know a little bit more when all of the voter file data is in meaning who voted and who didn't. As for the polling, one of the things that people have to understand is that in campaign polling, we don't walk in and say you have a 30-point lead, 40-point lead or 10-point lead. That's not the way it works. In fact, the first poll that was done for the Congressman showed him below 50 percent in the polling, and for a long -erm incumbent, that tends to signal that there is a problem, and so in fact, from the earliest poll that was done for the campaign, the notion was, that we were not in nearly as good enough shape as we should have been."
Don Levy, Director of the Siena College Research Institute, says primaries are especially difficult to poll. "Before anyone criticizes the polls, we should look at what events took place, how big was the turnout, what may have changed between the moment at which the polling was done and the election took place. And the one that you're talking about, in Queens and a portion of the Bronx, it was an especially low-turnout election, and certainly when the turnout is quite low, a small, highly-motivated cadre of voters can have a dramatic effect on the election, and that is certainly what appears to have happened in that race."
27,658 people voted in that race. Pollock doesn't believe one outcome should shake citizens' faith in polls. "There's also races like the upstate race for Joe Morelli, where our numbers were dead on. And so, I do think we may be overinterpreting one thing. So far, there's been one Congressional incumbent to lose. And yes, it was a surprise, but you know there's lots and lots of data over the course of this year that's been right on, including all over the country. Is it possible that turnout patterns can change polls? Of course. We know that because polls are very bad predictors of who's going to turn out. They are generally modeled on past behavior."
Some observers have noted that Ocasio-Cortez was all but ignored by the influential New York Times, which added to the surprise of her victory.
Was the winner of New York 14 a result of a "perfect storm" of events? Levy seems to lean in that direction. "That is a case where the Crowley people, who maybe always vote, you know you could be a Crowley voter who you vote in virtually every election, a historically reliable voter, you just didn't vote. You know, it was in June, you believed that 'Oh, this was no contest,' and you may have even told the pollsters 'oh sure sure I know Joe Crowley, he's great, he's great for the district, I'm gonna vote for him,’ and you just didn't vote. And this was a case where a small but passionate group of people did rally their forces, did turn out and did vote. I just don't think that that's a condition that we're gonna see in a statewide primary and/or a statewide general election. I think it was unique circumstance or not necessarily unique but a circumstance that is particular to these early summer primaries."
In general, pollsters have pointed to changing voter demographics and the rise of cellphones as some of the factors making polling more difficult.