Why The Polls Got The Democratic Caucus Right

The same methods that were wrong on the Republican side were more accurate for the Democrats.

While the Iowa caucuses provided a polling upset on the Republican side, most pollsters -- often in the same surveys -- fared much better with the Democratic results.

Every poll released in the last week before the caucus showed businessman Donald Trump winning on the Republican side. On Monday night, however, Texas Sen. Ted Cruz won by a 4-percent margin.

There will be plenty of focus on why pollsters were wrong in the Republican caucuses, but it's worth a second look at the Democratic side to see what went right.

Polls projected a very close race. HuffPost Pollster's final average put former Secretary of State Hillary Clinton 3 points over Sen. Bernie Sanders (I-Vt.). Clinton seems to have won the caucuses with 49.8 percent of the delegate count to Sanders' 49.6 percent. The Democratic Party reports results as delegate counts, making a comparison to the polls more difficult, but entrance poll estimates show the raw vote just about in line with Clinton's 3-point polling lead.

Final polls from the Des Moines Register/Bloomberg/Selzer poll and the NBC News/Wall Street Journal/Marist poll were especially close -- both had Clinton up on Sanders by 3 percent. Monmouth University showed Clinton up by 5 percentage points. The final results are easily within those polls' margins of error.

How could pollsters be wrong on the Republicans and right about the Democrats? Lee Miringoff, director of the Marist Institute for Public Opinion, explains: “The Democratic race was between two contenders evenly matched ... I think all the polls picked up that stability.”

Pollster J. Ann Selzer also cited the size of the Republican field as a major differentiating factor. She noted that the final DMR/Bloomberg/Selzer poll got all the demographics of the caucusgoers right on both the Democratic and Republican sides.

The large Republican field meant less certainty. “Our poll, six days out, showed the Republican side with more variability,” Miringoff said via email. “More voters said they might vote differently than on the Democratic side, and voters had a comfort level with Cruz and [Sen. Marco] Rubio as second choice. A Rubio surge was evident as he had gone from 12 to 18 points.”

Pollsters had agreed that in order for Sanders to win, he needed young voters to turn out at levels close to the record-breaking numbers President Barack Obama motivated in 2008. Pollsters and pundits alike were skeptical that would happen, and in this case they were right. Around 236,000 Democrats caucused in 2008, far above Monday night's turnout of about 171,000.

The other wild card for the Democrats was where former Maryland Gov. Martin O'Malley's supporters would go when he inevitably didn't make it to the 15 percent threshold to earn delegates in each precinct. Although there's no solid data on what happened, it seems that his supporters might have swung toward Sanders, but not in large enough numbers to shift the outcome.

Even with the question of O’Malley’s supporters reallocating, the Democratic campaign simply seemed more traditional, making it easier to poll. “In reflecting upon the Democratic and Republican contests, the Democratic campaigns played out truer to form,” Miringoff said. “Voters were more baked in, candidates delivered consistent messages, and there were fewer eleventh hour controversies (Clinton emails are the exception).”

Iowa caucuses are notoriously difficult to poll, but Patrick Murray, director of the Monmouth University poll, warns that the situation might not improve even with Iowa in the rearview mirror.

“I don’t anticipate New Hampshire being a lot smoother,” he said. “The issue is undecided voters and voters who make a choice but are open to changing their minds.”

Also on HuffPost:

Popular in the Community

Close

What's Hot