fb-pixel Skip to main content
2014 Primaries | Evan Horowitz

Why were the primary polls so far off?

The results are in. Martha Coakley is now the Democratic candidate for governor, Maura Healey won the chance to run for attorney general, and Seth Moulton ousted nine-term congressman John Tierney.

Only Moulton’s victory can be called a real surprise. Polls had predicted victories for Coakley and Healey. And yet, those polls were wrong in other, important ways. Coakley wasn’t supposed to win by a mere 6 points, she was supposed to win big, by 15 or 20 points. Healey’s race was supposed to be much closer, yet she won with 62 percent of the vote. Why were the polls so off? The answer is different for each race.


Why was Coakley’s win so tight?

There wasn’t a single poll all year that showed Coakley in a tight race. She maintained a double-digit lead until primary night, when her margin of victory seemed to shrink.

Perhaps the most likely explanation is turnout. Only about 16 percent of registered voters showed up this year, as compared to 25 percent in 2006. And lower turnout doesn’t just mean fewer people. It means that more of the people who do show up are “highly motivated voters,” the kind of voters who are passionate about individual candidates and knowledgeable about the issues. Earlier polls had shown that these highly motivated voters were more likely to vote for Steve Grossman and Don Berwick. So as the day wore on, and less-motivated voters stayed away, Coakley lost some of her support and found herself in a much more competitive race.

Why was Healey’s margin so big?

Healey caught a late wave. For most of the summer, her race with Warren Tolman looked like a toss-up. In the final weeks, though, the Globe released a poll showing Healey with a 16 point lead. Initially, it was hard to tell whether that one poll was a fluke, or whether Healey was in fact pulling away. The final vote count showed that the race really had tilted in her favor.


Why was Moulton’s victory a surprise?

In Moulton’s case, it’s not that the polls were wrong. It’s that there weren’t any good polls. The few that came out used were either tainted by the fact that they came direct from the candidates or made use of less reliable methodologies.

Were there any other interesting patterns?

The regional breakdown was starker than you generally see in a gubernatorial primary. While most districts went for Coakley (green), Grossman (purple) won a substantial number in Greater Boston while Berwick (blue) dominated a large swath of Western Massachusetts.

Jerold Duquette, a political scientist at Central Connecticut State, posed a question to me this morning that I thought best to pass along as a thought-experiment. To what degree did the inaccurate polling actually affect the outcome of this race? That is, did the drumbeat of Coakley’s inevitability dissuade potential voters from treating other candidates as viable alternatives? And did that sway some voters away from Grossman or Berwick, making the crucial difference in a race that proved much tighter than experts had thought?

Evan Horowitz digs through data to find information that illuminates the policy issues facing Massachusetts and the United States. He can be reached at evan.horowitz@globe.com. Follow him on Twitter @GlobeHorowitz